A behavior change, but a worthwhile one: "git submodule foreach"
was treating its arguments as part of a single command to be
concatenated and passed to a shell, making writing buggy
scripts too easy.
This patch preserves the old "just pass it to the shell" behavior
when a single argument is passed to 'git submodule foreach' and
moves to a new "skip the shell and use the arguments passed
unmolested" behavior when more than one argument is passed.
The old behavior (always concatenating and passing to the shell)
was similar to the 'ssh' command, while the new behavior (switching
on the number of arguments) is what 'xterm -e' does.
May need more thought to make sure this change is advertised well
so that scripts that used multiple arguments but added their own
extra layer of quoting are not broken.
* ak/submodule-foreach-quoting:
submodule foreach: skip eval for more than one argument
In t5000, we test the built-in ".tar.gz" config for
git-archive. To make our tests portable, we check that we
have a way to both gzip and gunzip, and we respected
environment variables to point to alternate commands for
doing these operations.
However, the $GZIP variable did not actually do anything, as
changing it would not affect the baked-in value in
archive-tar.c. Moreover, setting the variable $GZIP
influences gzip itself. From the gzip man page:
The environment variable GZIP can hold a set of default
options for gzip. These options are interpreted first and
can be overwritten by explicit command line parameters.
We could rename this variable, and use it to set up custom
config (or even have a Makefile knob to affect the built
binary), but it is not worth the trouble; nobody has ever
reported a problem with the baked-in default, and they can
always change it via config if they need to. Let's just drop
the variable and use "gzip" in the test (keeping the
prerequisite, of course).
While we're at it, we can drop the GUNZIP variable and
prerequisite; it uses "gzip -d", so if we have GZIP, we
will have both.
We can also use test_lazy_prereq for the gzip prerequisite,
which is simpler and behaves more consistently with the rest
of git (e.g., by making output available when the test is
run with "-v").
Noticed-by: Christian Hesse <mail@eworm.de>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Bug 6530 [1] in glibc causes "git show v0.99.6~1" to fail with error
"your vsnprintf is broken". The workaround avoids that, but it
corrupts system error messages in non-C locales.
The bug has been fixed since 2.17. We could know running glibc version
with gnu_get_libc_version(). But version is not a sure way to detect
the bug because downstream may back port the fix to older versions. Do
a runtime test that immitates the call flow that leads to "your
vsnprintf is broken". Only enable the workaround if the test fails.
Tested on Gentoo Linux, glibc 2.16.0 and 2.17, amd64.
[1] http://sourceware.org/bugzilla/show_bug.cgi?id=6530
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add more tests testing all the combinations:
-IPv4 or IPv6
-path starting with "/" or with "/~"
-with and without the ssh:// scheme
Some tests fail; they need updates in connect.c
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 8d3d28f5 added test cases for URLs which should be ssh.
Remove the function clear_ssh, use test_when_finished to clean up.
Introduce the function setup_ssh_wrapper, which could be factored
out together with expect_ssh.
Tighten one test and use "foo:bar" instead of "./foo:bar",
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the user is using the 'upstream' mode, these commands:
$ git push
$ git push origin
would find the 'upstream' branch for the current branch, and then
push the current branch to update it. However, pushing a single
branch explicitly, i.e.
$ git push origin $(git symbolic-ref --short HEAD)
would not go through the same ref mapping process, and ends up
updating the branch at 'origin' of the same name, which may not
necessarily be the upstream of the branch being pushed.
In the spirit similar to the previous one, map a colon-less refspec
using the upstream mapping logic.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since f2690487 (fetch: opportunistically update tracking refs,
2013-05-11), we stopped taking a non-storing refspec given on the
command line of "git fetch" literally, and instead started mapping
it via remote.$name.fetch refspecs. This allows
$ git fetch origin master
from the 'origin' repository, which is configured with
[remote "origin"]
fetch = +refs/heads/*:refs/remotes/origin/*
to update refs/remotes/origin/master with the result, as if the
command line were
$ git fetch origin +master:refs/remotes/origin/master
to reduce surprises and improve usability. Before that change, a
refspec on the command line without a colon was only to fetch the
history and leave the result in FETCH_HEAD, without updating the
remote-tracking branches.
When you are simulating a fetch from you by your mothership with a
push by you into your mothership, instead of having:
[remote "satellite"]
fetch = +refs/heads/*:refs/remotes/satellite/*
on the mothership repository and running:
mothership$ git fetch satellite
you would have:
[remote "mothership"]
push = +refs/heads/*:refs/remotes/satellite/*
on your satellite machine, and run:
satellite$ git push mothership
Because we so far did not make the corresponding change to the push
side, this command:
satellite$ git push mothership master
does _not_ allow you on the satellite to only push 'master' out but
still to the usual destination (i.e. refs/remotes/satellite/master).
Implement the logic to map an unqualified refspec given on the
command line via the remote.$name.push refspec. This will bring a
bit more symmetry between "fetch" and "push".
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git used to trim the trailing slash, and make the command equivalent
to 'git mv file no-such-dir', which created the file no-such-dir
(while the trailing slash explicitly stated that it could only be a
directory).
This patch skips the trailing slash removal for the destination
path. The path with its trailing slash is passed to rename(2),
which errors out with the appropriate message:
$ git mv file no-such-dir/
fatal: renaming 'file' failed: Not a directory
Original-patch-by: Duy Nguyen <pclouds@gmail.com>
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When --smtp-encryption=ssl, we use a Net::SMTP::SSL connection,
passing its ->new all the options that would otherwise go to
Net::SMTP->new (most options) and IO::Socket::SSL->start_SSL (for the
SSL options).
However, while Net::SMTP::SSL replaces the underlying socket class
with an SSL socket, it does nothing to allow passing options to that
socket. So the SSL-relevant options are lost.
Fortunately there is an escape hatch: we can directly set the options
with IO::Socket::SSL::set_client_defaults. They will then persist
within the IO::Socket::SSL module.
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
35035bb (send-email: be explicit with SSL certificate verification,
2013-07-18) forgot to specify that --smtp-ssl-cert-path takes a string
argument. This means that the option could not actually be used as
intended. Presumably noone noticed because it's much easier to set it
through configs anyway.
Add the required "=s".
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We forgot to pass the Debug option through to Net::SMTP::SSL->new --
which is the same as Net::SMTP->new. This meant that with security
set to SSL, we would never enable debug output.
Pass through the flag.
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The command line arguments given to "git push" are massaged into
a list of refspecs in set_refspecs() function. This was implemented
using xmalloc, strcpy and friends, but it is much easier to read if
done using strbuf.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the built-in "git tar-tree" command (a thin wrapper around "git
archive") was removed in 925ceccf (tar-tree: remove deprecated
command, 2013-11-10), the build continued to install a non-functioning
git-tar-tree command in gitexecdir by mistake:
$ PATH=$(git --exec-path):$PATH
$ git-tar-tree -h
fatal: cannot handle tar-tree internally
The list of links in gitexecdir is populated from BUILTIN_OBJS, which
includes builtin/tar-tree.o to implement "git get-tar-commit-id".
Rename the get-tar-commit-id source file to builtin/get-tar-commit-id.c
to reflect its purpose and fix 'make install'.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The usage sample of add_submodule_odb() function in the Submodules
section expects non-zero return value for success, but the function
actually reports success with zero.
Helped-by: René Scharfe <l.s.r@web.de>
Reviewed-by: Heiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: Nick Townsend <nick.townsend@mac.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When submodule.$name.update is given as hint from the upstream in
the .gitmodules file, we used to blindly copy it to .git/config,
unless there already is a value defined for the submodule.
However, there is no reason to expect that the update mode hinted by
the upstream is available in the version of Git the user is using,
and a really custom "!cmd" prepared by an upstream person running on
Linux may not even be available to a user on Windows. It is simply
irresponsible to copy the setting blindly and to attempt to use it
during a later "submodule update" without validating it first.
Just show the suggested value to the diagnostic output, and set the
value to 'none' in the configuration, if it is not one of the ones
that are known to be supported by this version of Git.
Helped-by: Jens Lehmann <Jens.Lehmann@web.de>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When allocating the slab, the code accidentally computed the array
size from s->slab (an elemtype**). The slab is an array of elemtype*,
however, so we should take the size of *s->slab.
Noticed-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This gives line-log support to gitk, by exploiting the new support for
processing and showing "inline" diffs straight from the git-log
output.
Note that we 'set allknown 0', which is a bit counterintuitive since
this is a "known" option. But that flag prevents gitk from thinking
it can optimize the view by running rev-list to see the topology; in
the -L case that doesn't work.
Signed-off-by: Thomas Rast <trast@inf.ethz.ch>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The previous commit split the diffs into a separate field. Now we
actually want to show them.
To that end we use the stored diff, and
- process it once to build a fake "tree diff", i.e., a list of all
changed files;
- feed it through parseblobdiffline to actually format it into the
$ctext field, like the existing diff machinery would.
Signed-off-by: Thomas Rast <trast@inf.ethz.ch>
Signed-off-by: Paul Mackerras <paulus@samba.org>
So far we just parsed everything after the headers into the "comment"
bit of $commitinfo, including notes and -- if you gave weird options
-- the diff.
Split out the diff, if any, into a separate field. It's easy to
recognize, since it always starts with /^diff/ and is preceded by an
empty line.
We take care to snip away said empty line. The display code already
properly spaces the end of the message from the first diff, and
leaving another empty line at the end looks ugly.
Signed-off-by: Thomas Rast <trast@inf.ethz.ch>
Signed-off-by: Paul Mackerras <paulus@samba.org>
For later use with data sources other than a pipe, refactor the big
worker part of getblobdiffline to a separate function
parseblobdiffline. Also refactor its initialization and wrap-up to
separate routines.
Signed-off-by: Thomas Rast <trast@inf.ethz.ch>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The -G option's usage is exactly analogous to that of -S, so
supporting it is easy.
Signed-off-by: Thomas Rast <trast@inf.ethz.ch>
Signed-off-by: Paul Mackerras <paulus@samba.org>
git-cherry(1)'s "description" section has never really managed
to explain to me what the command does. It contains too much
explanation of the algorithm instead of simply saying what
goals it achieves, and too much terminology that we otherwise
do not use (fork-point instead of merge-base).
Try a much more concise approach: state what it finds out, why
this is neat, and how the output is formatted, in a few short
paragraphs. In return, provide much longer examples of how it
fits into a "format-patch | am" based workflow, and how it
compares to reading the same from git-log.
Also carefully avoid using "merge" in a context where it does
not mean something that comes from git-merge(1). Instead, say
"apply" in an attempt to further link to patch workflow
concepts.
While there, also omit the language about _which_ upstream
branch we treat as the default. I literally just learned that
we support having several, so let's not confuse new users
here, especially considering that git-config(1) does not
document this.
Prompted-by: a.huemer@commend.com on #git
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The internal mercurial API expects ordinary 8-bit string objects, not
Unicode string objects. With this change, the test-hg.sh unit tests
pass again.
Signed-off-by: Richard Hansen <rhansen@bbn.com>
Reviewed-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This shuts up compiler warnings about unused functions. No such
warnings are currently triggered, but if someone were to actually
use init_NAME_with_stride() as documented, they would get a warning
about init_NAME() being unused.
While there, write a comment about why the last real declaration of
the variable is without a terminating semicolon, while another
forward declarations have one.
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The clear_$slabname() function was only documented by source code so
far. Write something about it.
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a shell snippet meant to be sourced by other shell scripts, an
opening #! line does more harm than good.
The harm:
- When the shell library is sourced, the interpreter and options from
the #! line are not used. Specifying a particular shell can
confuse the reader into thinking it is safe for the shell library
to rely on idiosyncrasies of that shell.
- Using #! instead of a plain comment drops a helpful visual clue
that this is a shell library and not a self-contained script.
- Tools such as lintian can use a #! line to tell when an
installation script has failed by forgetting to set a script
executable. This check does not work if shell libraries also start
with a #! line.
The good:
- Text editors notice the #! line and use it for syntax highlighting
if you try to edit the installed scripts (without ".sh" suffix) in
place.
The use of the #! for file type detection is not needed because Git's
shell libraries are meant to be edited in source form (with ".sh"
suffix). Replace the opening #! lines with comments.
This involves tweaking the test harness's valgrind support to find
shell libraries by looking for "# " in the first line instead of "#!"
(see v1.7.6-rc3~7, 2011-06-17).
Suggested by Russ Allbery through lintian. Thanks to Jeff King and
Clemens Buchacher for further analysis.
Tested by searching for non-executable scripts with #! line:
find . -name .git -prune -o -type f -not -executable |
while read file
do
read line <"$file"
case $line in
'#!'*)
echo "$file"
;;
esac
done
The only remaining scripts found are templates for shell scripts
(unimplemented.sh, wrap-for-bin.sh) and sample input used in tests
(t/t4034/perl/{pre,post}).
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A #! line in these files is misleading, since these scriptlets are
meant to be sourced with '.' (using whatever shell sources them)
instead of run directly using the interpreter named on the #! line.
Removing the #! line shouldn't hurt syntax highlighting since
these files have filenames ending with '.sh'. For documentation,
add a brief description of how the files are meant to be used in
place of the shebang line.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This way, test authors don't need to remember to source
lib-prereq-FILEMODE.sh before using the FILEMODE prereq to guard tests
that rely on the executable bit being honored when checking out files.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The git p4import documentation has suggested git p4 as a better
alternative for more than 6 years. (According to the mailing list
discussion when it was moved to contrib/, git-p4import has serious
bugs --- e.g., its incremental mode just doesn't work.) Since then,
git p4 has been actively developed and was promoted to a standard git
command alongside git svn.
Searches on google.com/trends and stackoverflow suggest that no one is
looking for git-p4import any more. Remove it.
Noticed while considering marking the contrib/p4import/git-p4import.py
script executable as part of a wider sweep.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Acked-by: Pete Wyckoff <pw@padd.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The docs in contrib/hooks/pre-auto-gc-battery suggest:
For example, if the hook is stored in
/usr/share/git-core/contrib/hooks/pre-auto-gc-battery:
chmod a+x pre-auto-gc-battery
cd /path/to/your/repository.git
ln -sf /usr/share/git-core/contrib/hooks/pre-auto-gc-battery \
hooks/pre-auto-gc
Unfortunately on multi-user systems most users do not have write
access to /usr. Better to mark the sample hooks executable in
the first place so users do not have to tweak their permissions to
use them by symlinking into .git/hooks/.
Reported-by: Olivier Berger <olivier.berger@it-sudparis.eu>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These scripts are not run directly as part of a normal build, so no
one noticed that they did not have the +x bit. Mark them executable
to make it more obvious that they can be run directly (when debugging,
for example).
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Windows the convention is to rely on filename extensions to decide
whether a file is executable so Windows users are probably not relying
on the executable bit of these scripts, but on other platforms it can
be useful documentation.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Up to now git has assumed that all servers are able to fix thin
packs. This is however not always the case.
Document the 'no-thin' capability and prevent send-pack from generating
a thin pack if the server advertises it.
Signed-off-by: Carlos Martín Nieto <cmn@elego.de>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Unbreaks a recent breakage due to use of unquote-c-style.
This may need to be cherry-picked down to 1.8.4.x series.
* 'rh/remote-hg-bzr-updates' (early part):
remote-hg: don't decode UTF-8 paths into Unicode objects
The output of git format-patch can vary with user preferences. In
particular setting diff.noprefix will break the "git apply" that
is done as part of "git p4 submit".
Acked-by: Pete Wyckoff <pw@padd.com>
Signed-off-by: Crestez Dan Leonard <cdleonard@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In git v1.4.3, we introduced a new loose object format that
encoded some object information outside of the zlib stream.
Ultimately the format was dropped in v1.5.3, but we kept the
reading side around to help people migrate objects. Each
time we open a loose object, we use a heuristic to check
whether it is in the normal loose format, or the
experimental one.
This heuristic is robust in the face of valid data, but it
tends to treat corrupted or garbage data as an experimental
object. With the regular format, we would notice quickly
that zlib's crc does not check out and complain. With the
experimental object, we are likely to extract a nonsensical
object size and try to allocate a huge buffer, resulting in
xmalloc calling "die".
This latter behavior is much worse, for two reasons. One,
git reports an allocation error when the real error is
corruption. And two, the program dies unconditionally, so
you cannot even run fsck (which would otherwise ignore the
broken object and keep going).
We could try to improve the heuristic to err on the side of
normal objects in the face of corruption, but there is
really little point. The experimental format is long-dead,
and was never enabled by default to begin with. We can
instead simply remove it. The only affected repository would
be one that explicitly set core.legacyheaders in 2007, and
then never repacked in the intervening 6 years.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
