* jc/show-sig:
log --show-signature: reword the common two-head merge case
log-tree: show mergetag in log --show-signature output
log-tree.c: small refactor in show_signature()
commit --amend -S: strip existing gpgsig headers
verify_signed_buffer: fix stale comment
gpg-interface: allow use of a custom GPG binary
pretty: %G[?GS] placeholders
test "commit -S" and "log --show-signature"
log: --show-signature
commit: teach --gpg-sign option
Conflicts:
builtin/commit-tree.c
builtin/commit.c
builtin/merge.c
notes-cache.c
pretty.c
Any existing commit signature was made against the contents of the old
commit, including its committer date that is about to change, and will
become invalid by amending it.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* nd/war-on-nul-in-commit:
commit_tree(): refuse commit messages that contain NULs
Convert commit_tree() to take strbuf as message
merge: abort if fails to commit
Conflicts:
builtin/commit.c
commit.c
commit.h
Current implementation sees NUL as terminator. If users give a message
with NUL byte in it (e.g. editor set to save as UTF-16), the new commit
message will have NULs. However following operations (displaying or
amending a commit for example) will not keep anything after the first NUL.
Stop user right when they do this. If NUL is added by mistake, they have
their chance to fix. Otherwise, log messages will no longer be text "git
log" and friends would grok.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There wan't a way for commit_tree() to notice if the message the caller
prepared contained a NUL byte, as it did not take the length of the
message as a parameter. Use a pointer to a strbuf instead, so that we can
either choose to allow low-level plumbing commands to make commits that
contain NUL byte in its message, or forbid NUL everywhere by adding the
check in commit_tree(), in later patches.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This teaches the "log" family of commands to pass the GPG signature in the
commit objects to "gpg --verify" via the verify_signed_buffer() interface
used to verify signed tag objects. E.g.
$ git show --show-signature -s HEAD
shows GPG output in the header part of the output.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This uses the gpg-interface.[ch] to allow signing the commit, i.e.
$ git commit --gpg-sign -m foo
You need a passphrase to unlock the secret key for
user: "Junio C Hamano <gitster@pobox.com>"
4096-bit RSA key, ID 96AFE6CB, created 2011-10-03 (main key ID 713660A7)
[master 8457d13] foo
1 files changed, 1 insertions(+), 0 deletions(-)
The lines of GPG detached signature are placed in a new multi-line header
field, instead of tucking the signature block at the end of the commit log
message text (similar to how signed tag is done), for multiple reasons:
- The signature won't clutter output from "git log" and friends if it is
in the extra header. If we place it at the end of the log message, we
would need to teach "git log" and friends to strip the signature block
with an option.
- Teaching new versions of "git log" and "gitk" to optionally verify and
show signatures is cleaner if we structurally know where the signature
block is (instead of scanning in the commit log message).
- The signature needs to be stripped upon various commit rewriting
operations, e.g. rebase, filter-branch, etc. They all already ignore
unknown headers, but if we place signature in the log message, all of
these tools (and third-party tools) also need to learn how a signature
block would look like.
- When we added the optional encoding header, all the tools (both in tree
and third-party) that acts on the raw commit object should have been
fixed to ignore headers they do not understand, so it is not like that
new header would be more likely to break than extra text in the commit.
A commit made with the above sample sequence would look like this:
$ git cat-file commit HEAD
tree 3cd71d90e3db4136e5260ab54599791c4f883b9d
parent b87755351a47b09cb27d6913e6e0e17e6254a4d4
author Junio C Hamano <gitster@pobox.com> 1317862251 -0700
committer Junio C Hamano <gitster@pobox.com> 1317862251 -0700
gpgsig -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
iQIcBAABAgAGBQJOjPtrAAoJELC16IaWr+bL4TMP/RSe2Y/jYnCkds9unO5JEnfG
...
=dt98
-----END PGP SIGNATURE-----
foo
but "git log" (unless you ask for it with --pretty=raw) output is not
cluttered with the signature information.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After running "git pull $there for-linus" to merge a signed tag, the
integrator may need to amend the resulting merge commit to fix typoes
in it. Teach --amend option to read the existing extra headers, and
carry them forward.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now MERGE_HEAD records the tag objects without peeling, we could record
the result of manual conflict resolution via "git commit" without losing
the tag information. Introduce a new "mergetag" multi-line header field to
the commit object, and use it to store the entire contents of each signed
tag merged.
A commit header that has a multi-line payload begins with the header tag
(e.g. "mergetag" in this case), SP, the first line of payload, LF, and all
the remaining lines have a SP inserted at the beginning.
In hindsight, it would have been better to make "merge --continue" as the
way to continue from such an interrupted merge, not "commit", but this is
a backward compatibility baggage we would need to carry around for now.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The merge-recursive code uses the commit->util field directly to annotate
the commit objects given from the command line, i.e. the remote heads to
be merged, with a single string to be used to describe it in its trace
messages and conflict markers.
Correct this short-signtedness by redefining the field to be a pointer to
a structure "struct merge_remote_desc" that later enhancements can add
more information. Store the original objects we were told to merge in a
field "obj" in this struct, so that we can recover the tag we were told to
merge.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* rs/pending:
commit: factor out clear_commit_marks_for_object_array
checkout: use leak_pending flag
bundle: use leak_pending flag
bisect: use leak_pending flag
revision: add leak_pending flag
checkout: use add_pending_{object,sha1} in orphan check
revision: factor out add_pending_sha1
checkout: check for "Previous HEAD" notice in t2020
Conflicts:
builtin/checkout.c
revision.c
* nd/maint-autofix-tag-in-head:
Accept tags in HEAD or MERGE_HEAD
merge: remove global variable head[]
merge: use return value of resolve_ref() to determine if HEAD is invalid
merge: keep stash[] a local variable
Conflicts:
builtin/merge.c
Factor out the code to clear the commit marks for a whole struct
object_array from builtin/checkout.c into its own exported function
clear_commit_marks_for_object_array and use it in bisect and bundle
as well. It handles tags and commits and ignores objects of any
other type.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
HEAD and MERGE_HEAD (among other branch tips) should never hold a
tag. That can only be caused by broken tools and is cumbersome to fix
by an end user with:
$ git update-ref HEAD $(git rev-parse HEAD^{commit})
which may look like a magic to a new person.
Be easy, warn users (so broken tools can be fixed if they bother to
report) and move on.
Be robust, if the given SHA-1 cannot be resolved to a commit object,
die (therefore return value is always valid).
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* nd/decorate-grafts:
log: Do not decorate replacements with --no-replace-objects
log: decorate "replaced" on to replaced commits
log: decorate grafted commits with "grafted"
Move write_shallow_commits to fetch-pack.c
Add for_each_commit_graft() to iterate all grafts
decoration: do not mis-decorate refs with same prefix
This function produces network traffic and should be in fetch-pack. It
has been in commit.c because it needs to iterate (private) graft
list. It can now do so using for_each_commit_graft().
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* pn/commit-autosquash:
add tests of commit --squash
commit: --squash option for use with rebase --autosquash
add tests of commit --fixup
commit: --fixup option for use with rebase --autosquash
pretty.c: teach format_commit_message() to reencode the output
commit: helper methods to reduce redundant blocks of code
Conflicts:
Documentation/git-commit.txt
t/t3415-rebase-autosquash.sh
Bad graft data is noticed in several places in read_graft_line(), and in
each case we go back to the first site of detection. It in general is a
better style to have an exception handling out of line from the main
codepath and make error codepath jump forward.
Signed-off-by: Ralf Thielow <ralf.thielow@googlemail.com>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add commit_list prefix to insert_by_date function and to sort_by_date,
so it's clear that these functions refer to commit_list structure.
Signed-off-by: Thiago Farina <tfransosi@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* builtin/commit.c: Replace block of code with a one-liner call to
logmsg_reencode().
* commit.c: new function for looking up a comit by name
* pretty.c: helper methods for getting output encodings
Add helpers get_log_output_encoding() and
get_commit_output_encoding() that eliminate some messy and duplicate
if-blocks.
Signed-off-by: Pat Notz <patnotz@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* cc/find-commit-subject:
blame: use find_commit_subject() instead of custom code
merge-recursive: use find_commit_subject() instead of custom code
bisect: use find_commit_subject() instead of custom code
revert: rename variables related to subject in get_message()
revert: refactor code to find commit subject in find_commit_subject()
revert: fix off by one read when searching the end of a commit subject
Until now, this has been part of the commit-tree builtin.
However, it is already used by other builtins (like commit,
merge, and notes), and it would be useful to access it from
library code.
The check_valid helper has to come along, too, but is given
a more library-ish name of "assert_sha1_type".
Otherwise, the code is unchanged. There are still a few
rough edges for a library function, like printing the utf8
warning to stderr, but we can address those if and when they
come up as inappropriate.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Caught by valgrind in t5500, but it is pretty obvious from
reading the code that this is shifting elements of an array
to the left, which needs memmove.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* 'jh/notes' (early part):
Add selftests verifying concatenation of multiple notes for the same commit
Refactor notes code to concatenate multiple notes annotating the same object
Add selftests verifying that we can parse notes trees with various fanouts
Teach the notes lookup code to parse notes trees with various fanout schemes
Teach notes code to free its internal data structures on request
Add '%N'-format for pretty-printing commit notes
Add flags to get_commit_notes() to control the format of the note string
t3302-notes-index-expensive: Speed up create_repo()
fast-import: Add support for importing commit notes
Teach "-m <msg>" and "-F <file>" to "git notes edit"
Add an expensive test for git-notes
Speed up git notes lookup
Add a script to edit/inspect notes
Introduce commit notes
Conflicts:
.gitignore
Documentation/pretty-formats.txt
pretty.c
* sp/smart-http: (37 commits)
http-backend: Let gcc check the format of more printf-type functions.
http-backend: Fix access beyond end of string.
http-backend: Fix bad treatment of uintmax_t in Content-Length
t5551-http-fetch: Work around broken Accept header in libcurl
t5551-http-fetch: Work around some libcurl versions
http-backend: Protect GIT_PROJECT_ROOT from /../ requests
Git-aware CGI to provide dumb HTTP transport
http-backend: Test configuration options
http-backend: Use http.getanyfile to disable dumb HTTP serving
test smart http fetch and push
http tests: use /dumb/ URL prefix
set httpd port before sourcing lib-httpd
t5540-http-push: remove redundant fetches
Smart HTTP fetch: gzip requests
Smart fetch over HTTP: client side
Smart push over HTTP: client side
Discover refs via smart HTTP server when available
http-backend: more explict LocationMatch
http-backend: add example for gitweb on same URL
http-backend: use mod_alias instead of mod_rewrite
...
Conflicts:
.gitignore
remote-curl.c
This change is being offered as a refactoring to make later
commits in the smart HTTP series easier.
By changing the enabled capabilities to be formatted in a strbuf
it is easier to add a new capability to the set of supported
capabilities.
By formatting the want portion of the request into a strbuf and
writing it as a whole block we can later decide to hold onto
the req_buf (instead of releasing it) to recycle in stateless
communications.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit notes are blobs which are shown together with the commit
message. These blobs are taken from the notes ref, which you can
configure by the config variable core.notesRef, which in turn can
be overridden by the environment variable GIT_NOTES_REF.
The notes ref is a branch which contains "files" whose names are
the names of the corresponding commits (i.e. the SHA-1).
The rationale for putting this information into a ref is this: we
want to be able to fetch and possibly union-merge the notes,
maybe even look at the date when a note was introduced, and we
want to store them efficiently together with the other objects.
This patch has been improved by the following contributions:
- Thomas Rast: fix core.notesRef documentation
- Tor Arne Vestbø: fix printing of multi-line notes
- Alex Riesen: Using char array instead of char pointer costs less BSS
- Johan Herland: Plug leak when msg is good, but msglen or type causes return
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Tor Arne Vestbø <tavestbo@trolltech.com>
Signed-off-by: Johan Herland <johan@herland.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
get_commit_notes(): Plug memory leak when 'if' triggers, but not because of read_sha1_file() failure
When creating an info/grafts under windows, one typically gets a CRLF file.
There is no good reason to forbid trailing CR at the end of the line (for
that matter, any trailing whitespaces); the code allowed only LF simply
because that was good enough for the platforms with LF line endings.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In 2d14d65 (Use a clearer style to issue commands to remote helpers,
2009-09-03) I happened to notice two changes like this:
- write_in_full(helper->in, "list\n", 5);
+
+ strbuf_addstr(&buf, "list\n");
+ write_in_full(helper->in, buf.buf, buf.len);
+ strbuf_reset(&buf);
IMHO, it would be better to define a new function,
static inline ssize_t write_str_in_full(int fd, const char *str)
{
return write_in_full(fd, str, strlen(str));
}
and then use it like this:
- strbuf_addstr(&buf, "list\n");
- write_in_full(helper->in, buf.buf, buf.len);
- strbuf_reset(&buf);
+ write_str_in_full(helper->in, "list\n");
Thus not requiring the added allocation, and still avoiding
the maintenance risk of literal string lengths.
These days, compilers are good enough that strlen("literal")
imposes no run-time cost.
Transformed via this:
perl -pi -e \
's/write_in_full\((.*?), (".*?"), \d+\)/write_str_in_full($1, $2)/'\
$(git grep -l 'write_in_full.*"')
Signed-off-by: Jim Meyering <meyering@redhat.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The variable named 'n' was initially declared to be of type int. The name
'n' was reused inside inner blocks as a different type. Rename the uses
within inner blocks to avoid confusion and give them a slightly more
descriptive name.
Signed-off-by: Brandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* js/maint-graft-unhide-true-parents:
git repack: keep commits hidden by a graft
Add a test showing that 'git repack' throws away grafted-away parents
Conflicts:
git-repack.sh
When you have grafts that pretend that a given commit has different
parents than the ones recorded in the commit object, it is dangerous
to let 'git repack' remove those hidden parents, as you can easily
remove the graft and end up with a broken repository.
So let's play it safe and keep those parent objects and everything
that is reachable by them, in addition to the grafted parents.
As this behavior can only be triggered by git pack-objects, and as that
command handles duplicate parents gracefully, we do not bother to cull
duplicated parents that may result by using both true and grafted
parents.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It is common practice to use the Unix epoch as a fallback date
when a suitable date is not available. This is true of git svn
and possibly other importing tools that import non-git history
into git.
Instead of clobbering established strtoul() error reporting
semantics with our own, preserve the strtoul() error value
of ULONG_MAX for fsck.c to handle.
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The goal of this patch series is to check if good revisions are
ancestor of the bad revision without forking a process to launch
"git rev-list $good ^$bad".
This new version of this patch series does not use an "unparse_commit"
function anymore, we use "clear_commit_marks" instead.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This patch adds the "unparse_commit" function that returns a commit
into an unparsed state by freeing its data and resetting its fields
to 0.
Its parents are recursively unparsed too, because they might have
been changed. But its tree is not unparsed as it should not have
been modifed.
Note that as the "flags" and "used" fields may be used even if the
object is not parsed, we have to reset them anyway.
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* js/notes:
git-notes: fix printing of multi-line notes
notes: fix core.notesRef documentation
Add an expensive test for git-notes
Speed up git notes lookup
Add a script to edit/inspect notes
Introduce commit notes
Conflicts:
pretty.c
Move has_commit() from branch to a common location, in preparation for
using it in "git-tag". Rename it to is_descendant_of() to make it more
unique and descriptive.
Signed-off-by: Jake Goulding <goulding@vivisimo.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit notes are blobs which are shown together with the commit
message. These blobs are taken from the notes ref, which you can
configure by the config variable core.notesRef, which in turn can
be overridden by the environment variable GIT_NOTES_REF.
The notes ref is a branch which contains "files" whose names are
the names of the corresponding commits (i.e. the SHA-1).
The rationale for putting this information into a ref is this: we
want to be able to fetch and possibly union-merge the notes,
maybe even look at the date when a note was introduced, and we
want to store them efficiently together with the other objects.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function is not called by any other file.
Signed-off-by: Nanako Shiraishi <nanako3@lavabit.com>
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
We used to set the TOPOSORT flag of commits during the topological
sorting, but we can just as well use the member "indegree" for it:
indegree is now incremented by 1 in the cases where the commit used
to have the TOPOSORT flag.
This is the same behavior as before, since indegree could not be
non-zero when TOPOSORT was unset.
Incidentally, this fixes the bug in show-branch where the 8th column
was not shown: show-branch sorts the commits in topological order,
assuming that all the commit flags are available for show-branch's
private matters.
But this was not true: TOPOSORT was identical to the flag corresponding
to the 8th ref. So the flags for the 8th column were unset by the
topological sorting.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Because we do not try computing merge base with itself for obvious
reasons, the code was not prepared for an arguably insane case of
the caller feeding the same commit twice to it.
Noticed and test written by Sverre Hvammen Johansen
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When comparing two commit objects for equality, it is sufficient to
compare their in-core pointers because the object layer guarantees the
uniqueness. However, comparing pointers to two "struct commit_list"
instances that point at the same commit does not make any sense.
Spotted by Sverre Hvammen Johansen who wrote an additional test to expose
the problem, fixed by Miklos Vajna.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The new function reduce_heads() is given a list of commits, and removes
ones that can be reached from other commits in the list. It is useful for
reducing the commits randomly thrown at the git-merge command and remove
redundant commits that the user shouldn't have given to it.
The implementation uses the get_merge_bases_many() introduced in the
previous commit. If the merge base between one commit taken from the list
and the remaining commits is the commit itself, that means the commit is
reachable from some of the other commits.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This introduces a new function get_merge_bases_many() which is a natural
extension of two commit merge base computation. It is given one commit
(one) and a set of other commits (twos), and computes the merge base of
one and a merge across other commits.
This is mostly useful to figure out the common ancestor when iterating
over heads during an octopus merge. When making an octopus between
commits A, B, C and D, we first merge tree of A and B, and then try to
merge C with it. If we were making pairwise merge, we would be recording
the tree resulting from the merge between A and B as a commit, say M, and
then the next round we will be computing the merge base between M and C.
o---C...*
/ .
o---B...M
/ .
o---o---A
But during an octopus merge, we actually do not create a commit M. In
order to figure out that the common ancestor to use for this merge,
instead of computing the merge base between C and M, we can call
merge_bases_many() with one set to C and twos containing A and B.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is like get_merge_bases() but it works for multiple heads, like
show-branch --merge-base.
Signed-off-by: Miklos Vajna <vmiklos@frugalware.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function is useful outside builtin-merge-recursive, for example in
builtin-merge.
Signed-off-by: Miklos Vajna <vmiklos@frugalware.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The n_refs variable is no longer really used in this function, so there
is no reason to keep it.
It was introduced in 27dedf0c and the code that really used it was
removed in 7914053.
Signed-off-by: Miklos Vajna <vmiklos@frugalware.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
parse_commit ignores parent commits with certain errors
(eg. a non commit object is already loaded under the sha1 of
the parent). To make fsck reports such errors, it has to compare
the nummer of parent commits returned by parse commit with the
number of parent commits in the object or in the graft/shallow file.
Signed-off-by: Martin Koegler <mkoegler@auto.tuwien.ac.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some codepaths (eg. builtin-rev-parse -> get_merge_bases -> parse_commit)
can pass NULL.
Signed-off-by: Martin Koegler <mkoegler@auto.tuwien.ac.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This tightens the parsing of a commit object in a couple of ways.
- The "tree " header must end with a LF (earlier we did not
check this condition).
- Make sure parsing of timestamp on the "committer " header
does not go beyond the buffer, even when (1) the "author "
header does not end with a LF (this means that the commit
object is malformed and lacks the committer information) or
(2) the "committer " header does not have ">" that is the end
of the e-mail address, or (3) the "committer " header does
not end with a LF.
We however still keep the existing behaviour to return a parsed
commit object even when non-structural headers such as committer
and author are malformed, so that tools that need to look at
commits to clean up a history with such broken commits can still
get at the structural data (i.e. the parents chain and the tree
object).
Signed-off-by: Martin Koegler <mkoegler@auto.tuwien.ac.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* lt/rev-list-interactive:
Fix parent rewriting in --early-output
revision walker: mini clean-up
Enhance --early-output format
Add "--early-output" log flag for interactive GUI use
Simplify topo-sort logic
The file commit.c got quite large, but it does not have to be: the
code concerning pretty printing is pretty well contained. In fact,
this commit just splits it off into pretty.c, leaving commit.c with
just 672 lines.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
.. by not using quite so much indirection.
This currently grows the "struct commit" a bit, which could be avoided by
using a union for "util" and "indegree" (the topo-sort used to use "util"
anyway, so you cannot use them together), but for now the goal of this was
to simplify, not optimize.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When the body of the commit log message contains a non-ASCII character,
format-patch correctly emitted the encoding header to mark the resulting
message as such. However, if the original message was fully ASCII, the
command line switch "-s" was given to add a new sign-off, and
the signer's name was not ASCII only, the resulting message would have
contained non-ASCII character but was not marked as such.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before this patch, clear_commit_marks() recursed for each parent. This
could be potentially very expensive in terms of stack space. Probably
the only reason that this did not lead to problems is the fact that we
typically call clear_commit_marks() after marking a relatively small set
of commits.
Use (sort of) a tail recursion instead: first recurse on the parents
other than the first one, and then continue the loop with the first
parent.
Noticed by Shawn Pearce.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Lars Hjemli <hjemli@gmail.com>
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
For that purpose, the ->buf is always initialized with a char * buf living
in the strbuf module. It is made a char * so that we can sloppily accept
things that perform: sb->buf[0] = '\0', and because you can't pass "" as an
initializer for ->buf without making gcc unhappy for very good reasons.
strbuf_init/_detach/_grow have been fixed to trust ->alloc and not ->buf
anymore.
as a consequence strbuf_detach is _mandatory_ to detach a buffer, copying
->buf isn't an option anymore, if ->buf is going to escape from the scope,
and eventually be free'd.
API changes:
* strbuf_setlen now always works, so just make strbuf_reset a convenience
macro.
* strbuf_detatch takes a size_t* optional argument (meaning it can be
NULL) to copy the buffer's len, as it was needed for this refactor to
make the code more readable, and working like the callers.
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This changes the interporate() to replace entries with NULL values
by the empty string, and uses it to interpolate missing fields in
custom format output used in git-log and friends. It is most useful
to avoid <unknown> output from %b format for a commit log message
that lack any body text.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add strbuf_remove, change strbuf_insert:
As both are special cases of strbuf_splice, implement them as such.
gcc is able to do the math and generate almost optimal code this way.
Add strbuf_swap:
Exchange the values of its arguments.
Use it in fast-import.c
Also fix spacing issues in strbuf.h
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Careful profiling shows that we spend more time guessing what pattern
allocation will have, whereas we can delay it only at the point where
add_rfc2047 will be used and don't allocate huge memory area for the many
cases where it's not.
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* Be more clever in how we search for "encoding ...\n": parse for real
instead of the sloppy strstr's.
* use strbuf_splice to do the substring replacements.
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Also remove the "len" parameter, as:
(1) it was used as a max boundary, and every caller used ~0u
(2) we check for final NUL no matter what, so it doesn't help for speed.
As a result most of the pp_* function takes 3 arguments less, and we need
a lot less local variables, this makes the code way more readable, and
easier to extend if needed.
This patch also fixes some spacing and cosmetic issues.
This patch also fixes (as a side effect) a memory leak intoruced in
builtin-archive.c at commit df4a394f (fmt was xmalloc'ed and not free'd)
Signed-off-by: Pierre Habouzit <madcoder@debian.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Drop the parameter "msg" of format_commit_message() (as it can be
inferred from the parameter "commit"), add a parameter "template"
in order to avoid accessing the static variable user_format
directly and export the result.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a commit message doesn't have encoding information
and encoding output is utf-8 (default) then an useless
xstrdup() of commit message is done.
If we assume most of users live in an utf-8 world, this
useless copy is the common case.
Performance issue found with KCachegrind.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These days, show_date() takes a date_mode parameter to specify
the output format, and a separate specialized function for dates
in E-mails does not make much sense anymore.
This retires show_rfc2822_date() function and make it just
another date output format.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Support output of full ISO 8601 style dates in e.g. git log
and other places that use interpolation for formatting.
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
add_user_info() possibly adds way more than just the commit header line.
In fact, it sometimes needs so much more space that there is a buffer
overrun, leading to an ugly crash. For example, the date is printed in its
own line, and usually takes up more space than the equivalent Unix epoch.
So, for good measure, add 80 characters (a full line) to the allocated
space, in addition to the header line length.
Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
so that an ugly commit message like this can be
handled sanely.
Currently, --pretty=oneline and --pretty=email (hence
format-patch) take and use only the first line of the commit log
message. This changes them to:
- Take the first paragraph, where the definition of the first
paragraph is "skip all blank lines from the beginning, and
then grab everything up to the next empty line".
- Replace all line breaks with a whitespace.
This change would not affect a well-behaved commit message that
adheres to the convention of "single line summary, a blank line,
and then body of message", as its first paragraph always
consists of a single line. Commit messages from different
culture, such as the ones imported from CVS/SVN, can however get
chomped with the existing behaviour at the first linebreak in
the middle of sentence right now, which would become much easier
to see with this change.
The Subject: and --pretty=oneline output would become very long
and unsightly for non-conforming commits, but their messages are
already ugly anyway, and thischange at least avoids the loss of
information.
The Subject: line from a multi-line paragraph is folded using
RFC2822 line folding rules at the places where line breaks were
in the original.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Traditionally we had 16kB limit when formatting log messages for
output, because it was easier to arrange for the caller to have
a reasonably big buffer and pass it down without ever worrying
about reallocating.
This changes the calling convention of pretty_print_commit() to
lift this limit. Instead of the buffer and remaining length, it
now takes a pointer to the pointer that points at the allocated
buffer, and another pointer to the location that stores the
allocated length, and reallocates the buffer as necessary.
To support the user format, the error return of interpolate()
needed to be changed. It used to return a bool telling "Ok the
result fits", or "Sorry, I had to truncate it". Now it returns
0 on success, and returns the size of the buffer it wants in
order to fit the whole result.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This uses "git-apply --whitespace=strip" to fix whitespace errors that have
crept in to our source files over time. There are a few files that need
to have trailing whitespaces (most notably, test vectors). The results
still passes the test, and build result in Documentation/ area is unchanged.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Encode ' ' using '=20' even though rfc2047 allows using '_' for
readability. Unfortunately, many programs do not understand this and
just leave the underscore in place. Using '=20' seems to work better.
[jc: with adjustment to t3901]
Signed-off-by: Kristian Høgsberg <hoegsberg@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
* maint:
format-patch: add MIME-Version header when we add content-type.
Fixed link in user-manual
import-tars: Use the "Link indicator" to identify directories
git name-rev writes beyond the end of malloc() with large generations
Documentation/branch: fix small typo in -D example
* maint:
Small correction in reading of commit headers
Documentation: fix typo in git-remote.txt
Add test for blame corner cases.
blame: -C -C -C
blame: Notice a wholesale incorporation of an existing file.
Fix --boundary output
diff format documentation: describe raw combined diff format
Mention version 1.5.1 in tutorial and user-manual
Add --no-rebase option to git-svn dcommit
Fix markup in git-svn man page
Check if a line of the header has enough characters to possibly
contain the requested prefix.
Signed-off-by: Alex Riesen <raa.lkml@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
* maint:
gitweb: use decode_utf8 directly
posix compatibility for t4200
Document 'opendiff' value in config.txt and git-mergetool.txt
Allow PERL_PATH="/usr/bin/env perl"
Make xstrndup common
diff.c: fix "size cache" handling.
http-fetch: Disable use of curl multi support for libcurl < 7.16.
This also improves the implementation to match how strndup is
specified (by GNU): if the length given is longer than the string,
only the string's length is allocated and copied, but the string need
not be null-terminated if it is at least as long as the given length.
Signed-off-by: Daniel Barkalow <barkalow@iabervon.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This adds --date={local,relative,default} option to log family of commands,
to allow displaying timestamps in user's local timezone, relative time, or
the default format.
Existing --relative-date option is a synonym of --date=relative; we could
probably deprecate it in the long run.
Signed-off-by: Junio C Hamano <junkio@cox.net>
This replaces the fairly odd "created_object()" function that did _most_
of the object setup with a more complete "create_object()" function that
also has a more natural calling convention.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
When used with '--boundary A...B', this shows the -/</> marker
you would get with --left-right option to 'git-log' family.
When symmetric diff is not used, everybody is shown to be on the
"right" branch.
Signed-off-by: Junio C Hamano <junkio@cox.net>
The function replace_encoding_header is given the whole
commit buffer, including the commit message. When looking
for the encoding header, if none was found in the header, it
would locate any line in the commit message matching
"\nencoding " and remove it.
Instead, we now make sure to search only to the end of the
header.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
It printed the header "encoding " instead of just showing
the encoding, as all other items do.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
There are two breakages in the %P/%p interpolation. It appended
an excess SP at the end of the list, and it gave uninitialized
contents of a buffer on the stack for root commits.
This fixes it, while updating the t6006 test which expected the
wrong output.
Signed-off-by: Junio C Hamano <junkio@cox.net>
A pointer arithmetic error in fill_person caused random data
from the commit object to be included with the timestamp,
which looked something like:
$ git-rev-list --pretty=format:%ct origin/next | head
commit 98453bdb3db10db26099749bc4f2dc029bed9aa9
1174977948 -0700
Merge branch 'master' into next
* master:
Bisect: Use
commit c0ce981f5e
1174889646 -0700
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
We shouldn't attempt to assign constant strings into char*, as the
string is not writable at runtime. Likewise we should always be
treating unsigned values as unsigned values, not as signed values.
Most of these are very straightforward. The only exception is the
(unnecessary) xstrdup/free in builtin-branch.c for the detached
head case. Since this is a user-level interactive type program
and that particular code path is executed no more than once, I feel
that the extra xstrdup call is well worth the easy elimination of
this warning.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
We currently have two parallel notation for dealing with object types
in the code: a string and a numerical value. One of them is obviously
redundent, and the most used one requires more stack space and a bunch
of strcmp() all over the place.
This is an initial step for the removal of the version using a char array
found in object reading code paths. The patch is unfortunately large but
there is no sane way to split it in smaller parts without breaking the
system.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
With this patch,
$ git show -s \
--pretty=format:' Ze komit %h woss%n dunn buy ze great %an'
shows something like
Ze komit 04c5c88 woss
dunn buy ze great Junio C Hamano
The supported placeholders are:
'%H': commit hash
'%h': abbreviated commit hash
'%T': tree hash
'%t': abbreviated tree hash
'%P': parent hashes
'%p': abbreviated parent hashes
'%an': author name
'%ae': author email
'%ad': author date
'%aD': author date, RFC2822 style
'%ar': author date, relative
'%at': author date, UNIX timestamp
'%cn': committer name
'%ce': committer email
'%cd': committer date
'%cD': committer date, RFC2822 style
'%cr': committer date, relative
'%ct': committer date, UNIX timestamp
'%e': encoding
'%s': subject
'%b': body
'%Cred': switch color to red
'%Cgreen': switch color to green
'%Cblue': switch color to blue
'%Creset': reset color
'%n': newline
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
--pretty=o is a valid abbreviation, --pretty=omfg is not
Noticed by: Nicolas Vilz
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
The internal function in_merge_bases(A, B) is used to make sure
that commit A is an ancestor of commit B. This changes the
signature of it to take an array of B's and updates its current
callers.
Signed-off-by: Junio C Hamano <junkio@cox.net>
We have a number of badly checked write() calls. Often we are
expecting write() to write exactly the size we requested or fail,
this fails to handle interrupts or short writes. Switch to using
the new write_in_full(). Otherwise we at a minimum need to check
for EINTR and EAGAIN, where this is appropriate use xwrite().
Note, the changes to config handling are much larger and handled
in the next patch in the sequence.
Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This modifies pretty_print_commit() to make the output of git-rev-list and
friends a bit more predictable.
A commit body starting with blank lines might be unheard-of, but still possible
to create using git-commit-tree (so is bound to appear somewhere, sometime).
Signed-off-by: Lars Hjemli <hjemli@gmail.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
We special case the case where encoding recorded in the commit
and the output encoding are the same and do not call iconv().
But we should drop 'encoding' header for this case as well for
consistency.
Signed-off-by: Junio C Hamano <junkio@cox.net>
After re-coding the commit message into the encoding the user
specified (either with core.logoutputencidng or --encoding
option), this drops the "encoding" header altogether. The
output is after re-coding as the user asked (either with the
config or --encoding=<encoding> option), and the extra header
becomes redundant information.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Telling the git-log family not to do any character conversion is
done with --encoding=none, which sets log_output_encoding to an
empty string. However, logmsg_reencode() confused this with
log_output_encoding and commit_encoding set to NULL. The latter
means we should use the default encoding (i.e. utf-8).
Signed-off-by: Junio C Hamano <junkio@cox.net>
* jc/utf8:
t3900: test conversion to non UTF-8 as well
Rename t3900 test vector file
UTF-8: introduce i18n.logoutputencoding.
Teach log family --encoding
i18n.logToUTF8: convert commit log message to UTF-8
Move encoding conversion routine out of mailinfo to utf8.c
Conflicts:
commit.c
It is plausible for somebody to want to view the commit log in a
different encoding from i18n.commitencoding -- the project's
policy may be UTF-8 and the user may be using a commit message
hook to run iconv to conform to that policy (and either not have
i18n.commitencoding to default to UTF-8 or have it explicitly
set to UTF-8). Even then, Latin-1 may be more convenient for
the usual pager and the terminal the user uses.
The new variable i18n.logoutputencoding is used in preference to
i18n.commitencoding to decide what encoding to recode the log
output in when git-log and friends formats the commit log message.
Signed-off-by: Junio C Hamano <junkio@cox.net>
This is to adjust to:
count-objects -v: show number of packs as well.
which will break a test in this series.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Updated commit objects record the encoding used in their
encoding header. This updates the log family to reencode it
into the encoding specified in i18n.commitencoding (or the
default, which is "utf-8") upon output.
To force a specific encoding that is different, log family takes
command line flag --encoding=<encoding>; giving --encoding=none
entirely disables the reencoding and lets you view log messges
in their original encoding.
Signed-off-by: Junio C Hamano <junkio@cox.net>
The output from "symmetric diff", i.e. A...B, does not
distinguish between commits that are reachable from A and the
ones that are reachable from B. In this picture, such a
symmetric diff includes commits marked with a and b.
x---b---b branch B
/ \ /
/ .
/ / \
o---x---a---a branch A
However, you cannot tell which ones are 'a' and which ones are
'b' from the output. Sometimes this is frustrating. This adds
an output option, --left-right, to rev-list.
rev-list --left-right A...B
would show ones reachable from A prefixed with '<' and the ones
reachable from B prefixed with '>'.
When combined with --boundary, boundary commits (the ones marked
with 'x' in the above picture) are shown with prefix '-', so you
would see list that looks like this:
git rev-list --left-right --boundary --pretty=oneline A...B
>bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb 3rd on b
>bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb 2nd on b
<aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 3rd on a
<aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 2nd on a
-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 1st on b
-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 1st on a
Signed-off-by: Junio C Hamano <junkio@cox.net>
Now, by saying "git fetch -depth <n> <repo>" you can deepen
a shallow repository.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
A shallow commit is a commit which has parents, which in turn are
"grafted away", i.e. the commit appears as if it were a root.
Since these shallow commits should not be edited by the user, but
only by core git, they are recorded in the file $GIT_DIR/shallow.
A repository containing shallow commits is called shallow.
The advantage of a shallow repository is that even if the upstream
contains lots of history, your local (shallow) repository needs not
occupy much disk space.
The disadvantage is that you might miss a merge base when pulling
some remote branch.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
I was running git show on various commits found by fsck-objects
when I found this bug. Since find_unique_abbrev() cannot find
an abbreviation for an object not in the database, it will
return NULL, which is bad to run strlen() on. So instead, we'll
just display the unabbreviated sha1 that we referenced in the
commit.
I'm not sure that this is the best 'fix' for it because the
commit I was trying to show was broken, but I don't think a
program should segfault even if the user tries to do something
stupid.
Signed-off-by: Eric Wong <normalperson@yhbt.net>
Signed-off-by: Junio C Hamano <junkio@cox.net>
I noticed that I was looking at the kernel gitweb output at some point
rather than just do "git log", simply because I liked seeing the
simplified date-format, ie the "5 days ago" rather than a full date.
This adds infrastructure to do that for "git log" too. It does NOT add the
actual flag to enable it, though, so right now this patch is a no-op, but
it should now be easy to add a command line flag (and possibly a config
file option) to just turn on the "relative" date format.
The exact cut-off points when it switches from days to weeks etc are
totally arbitrary, but are picked somewhat to avoid the "1 weeks ago"
thing (by making it show "10 days ago" rather than "1 week", or "70
minutes ago" rather than "1 hour ago").
[jc: with minor fix and tweak around "month" and "week" area.]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Introduces global inline:
hashcmp(const unsigned char *sha1, const unsigned char *sha2)
Uses memcmp for comparison and returns the result based on the length of
the hash name (a future runtime decision).
Acked-by: Alex Riesen <raa.lkml@gmail.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
[jc: I needed to hand merge the changes to the updated codebase,
so the result needs to be checked.]
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
format-patch previously didn't generate a newline after a subject. This
caused the diffstat to not be displayed in messages with only one line
for the commit message.
This patch fixes this by adding a newline after the headers if a body
hasn't been added.
Signed-off-by: Robert Shearman <rob@codeweavers.com>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This updates the type-enumeration constants introduced to reduce
the memory footprint of "struct object" to match the type bits
already used in the packfile format, by removing the former
(i.e. TYPE_* constant macros) and using the latter (i.e. enum
object_type) throughout the code for consistency.
Eventually we can stop passing around the "type strings"
entirely, and this will help - no confusion about two different
integer enumeration.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This removes the "contaminate the well even more" approach
taken in the current merge-base postprosessing code. Instead,
when there are more than one merge-base results, we compute the
merge-base between them and see if one is a fast-forward of the
other, in which case the ancestor is removed from the result.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Fix clear_commit_marks() enough to be usable in
get_merge_bases(), and retire now unused clear_object_marks().
Signed-off-by: Junio C Hamano <junkio@cox.net>
Earlier change broke "git describe A B" among other things.
Revert it for now, and clean the commits smudged by
get_merge_bases using clear_object_marks() function. For
complex commit ancestry graph, this is way cheaper as well.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Actually in this case we would have traversed a lot of commits, so cleaning
things up is even more important.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Change get_merge_bases() to be able to clean up after itself if
needed by adding a cleanup parameter.
We don't need to save the flags and restore them afterwards anymore;
that was a leftover from before the flags were moved out of the
range used in revision.c. clear_commit_marks() sets them to zero,
which is enough.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Acked-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Don't care if objects have been parsed or not and don't stop when we
reach a commit that is already clean -- its parents could be dirty.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Add get_merge_bases_clean(), a wrapper for get_merge_bases() that cleans
up after doing its work and make get_merge_bases() NOT clean up.
Single-shot programs like git-merge-base can use the dirty and fast
version.
Also move the object flags used in get_merge_bases() out of the range
defined in revision.h. This fixes the "66ae0c77...ced9456a
89719209...262a6ef7" test of the ... operator which is introduced with
the next patch.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <junkio@cox.net>
This creates a simple specialized object allocator for basic
objects.
This avoids wasting space with malloc overhead (metadata and
extra alignment), since the specialized allocator knows the
alignment, and that objects, once allocated, are never freed.
It also allows us to track some basic statistics about object
allocations. For example, for the mozilla import, it shows
object usage as follows:
blobs: 627629 (14710 kB)
trees: 1119035 (34969 kB)
commits: 196423 (8440 kB)
tags: 1336 (46 kB)
and the simpler allocator shaves off about 2.5% off the memory
footprint off a "git-rev-list --all --objects", and is a bit
faster too.
[ Side note: this concludes the series of "save memory in object storage".
The thing is, there simply isn't much more to be saved on the objects.
Doing "git-rev-list --all --objects" on the mozilla archive has a final
total RSS of 131498 pages for me: that's about 513MB. Of that, the
object overhead is now just 56MB, the rest is going somewhere else (put
another way: the fact that this patch shaves off 2.5% of the total
memory overhead, considering that objects are now not much more than 10%
of the total shows how big the wasted space really was: this makes
object allocations much more memory- and time-efficient).
I haven't looked at where the rest is, but I suspect the bulk of it is
just the pack-file loading. It may be that we should pack the tree
objects separately from the blob objects: for git-rev-list --objects, we
don't actually ever need to even look at the blobs, but since trees and
blobs are interspersed in the pack-file, we end up not being dense in
the tree accesses, so we end up looking at more pages than we strictly
need to.
So with a 535MB pack-file, it's entirely possible - even likely - that
most of the remaining RSS is just the mmap of the pack-file itself. We
don't need to map in _all_ of it, but we do end up mapping a fair
amount. ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Every single user actually wanted this only for commit objects, and we
have no reason to waste space on it for other object types. So just move
the structure member from the low-level "struct object" into the "struct
commit".
This leaves the commit object the same size, and removes one unnecessary
pointer from all other object allocations.
This shrinks memory usage (still at a fairly hefty half-gig, admittedly)
of "git-rev-list --all --objects" on the mozilla repo by another 5% in my
tests.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
