i18n: only extract comments marked with "TRANSLATORS:"
When extract l10n messages, we use "--add-comments" option to keep
comments right above the l10n messages for references. But sometimes
irrelevant comments are also extracted. For example in the following
code block, the comment in line 2 will be extracted as comment for the
l10n message in line 3, but obviously it's wrong.
{ OPTION_CALLBACK, 0, "ignore-removal", &addremove_explicit,
NULL /* takes no arguments */,
N_("ignore paths removed in the working tree (same as
--no-all)"),
PARSE_OPT_NOARG, ignore_removal_cb },
Since almost all comments for l10n translators are marked with the same
prefix (tag): "TRANSLATORS:", it's safe to only extract comments with
this special tag. I.E. it's better to call xgettext as:
xgettext --add-comments=TRANSLATORS: ...
Also tweaks the multi-line comment in "init-db.c", to make it start with
the proper tag, not "* TRANSLATORS:" (which has a star before the tag).
Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
i18n: fix uncatchable comments for translators in date.c
Comment for l10n translators can not be extracted by xgettext if it
is not right above the l10n tag. Moving the comment right before
the l10n tag will fix this issue.
Reported-by: Brian Gesiak <modocache@gmail.com> Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The workaround 99855ddf introduced to deal with problematic
"return" statements in scripts run by "dot" commands located
inside functions only handles one part of the problem. The
issue has now been addressed by not using "return" statements
in this way in the git-rebase--*.sh scripts.
This workaround is therefore no longer necessary, so clean
up the code by reverting it.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com> Acked-by: Matthieu Moy <Matthieu.Moy@imag.fr> Signed-off-by: Junio C Hamano <gitster@pobox.com>
rebase: avoid non-function use of "return" on FreeBSD
Since a1549e10, 15d4bf2e and 01a1e646 (first appearing in v1.8.4)
the git-rebase--*.sh scripts have used a "return" to stop execution
of the dot-sourced file and return to the "dot" command that
dot-sourced it. The /bin/sh utility on FreeBSD however behaves
poorly under some circumstances when such a "return" is executed.
In particular, if the "dot" command is contained within a function,
then when a "return" is executed by the script it runs (that is not
itself inside a function), control will return from the function
that contains the "dot" command skipping any statements that might
follow the dot command inside that function. Commit 99855ddf (first
appearing in v1.8.4.1) addresses this by making the "dot" command
the last line in the function.
Unfortunately the FreeBSD /bin/sh may also execute some statements
in the script run by the "dot" command that appear after the
troublesome "return". The fix in 99855ddf does not address this
problem.
For example, if you have script1.sh with these contents:
run_script2() {
. "$(dirname -- "$0")/script2.sh"
_e=$?
echo only this line should show
[ $_e -eq 5 ] || echo expected status 5 got $_e
return 3
}
run_script2
e=$?
[ $e -eq 3 ] || { echo expected status 3 got $e; exit 1; }
And script2.sh with these contents:
if [ 5 -gt 3 ]; then
return 5
fi
case bad in *)
echo always shows
esac
echo should not get here
! :
When running script1.sh (e.g. '/bin/sh script1.sh' or './script1.sh'
after making it executable), the expected output from a POSIX shell
is simply the single line:
only this line should show
However, when run using FreeBSD's /bin/sh, the following output
appears instead:
should not get here
expected status 3 got 1
Not only did the lines following the "dot" command in the run_script2
function in script1.sh get skipped, but additional lines in script2.sh
following the "return" got executed -- but not all of them (e.g. the
"echo always shows" line did not run).
These issues can be avoided by not using a top-level "return" in
script2.sh. If script2.sh is changed to this:
main() {
if [ 5 -gt 3 ]; then
return 5
fi
case bad in *)
echo always shows
esac
echo should not get here
! :
}
main
Then it behaves the same when using FreeBSD's /bin/sh as when using
other more POSIX compliant /bin/sh implementations.
We fix the git-rebase--*.sh scripts in a similar fashion by moving
the top-level code that contains "return" statements into its own
function and then calling that as the last line in the script.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com> Acked-by: Matthieu Moy <Matthieu.Moy@imag.fr> Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Windows, absolute paths might start with a DOS drive prefix,
which these two checks failed to recognize.
Unfortunately, we cannot simply use the file_name_is_absolute
helper in File::Spec::Functions, because Git for Windows has an
MSYS-based Perl, where this helper doesn't grok DOS
drive-prefixes.
So let's manually check for these in that case, and fall back to
the File::Spec-helper on other platforms (e.g Win32 with native
Perl)
Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This can help avoid -Wuninitialized false positives in
git_config_int and git_config_ulong, as the compiler now
knows that we do not return "ret" if we hit the error
codepath.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Multi-threaing of index-pack was disabled with c0f8654
(index-pack: Disable threading on cygwin - 2012-06-26), because
pread() implementations for Cygwin and MSYS were not thread
safe. Recent Cygwin does offer usable pread() and we enabled
multi-threading with 103d530f (Cygwin 1.7 has thread-safe pread,
2013-07-19).
Work around this problem on platforms with a thread-unsafe
pread() emulation by opening one file handle per thread; it
would prevent parallel pread() on different file handles from
stepping on each other.
Also remove NO_THREAD_SAFE_PREAD that was introduced in c0f8654
because it's no longer used anywhere.
This workaround is unconditional, even for platforms with
thread-safe pread() because the overhead is small (a couple file
handles more) and not worth fragmenting the code.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Tested-by: Johannes Sixt <j6t@kdbg.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
curl-config should always be installed alongside a curl distribution,
and its purpose is to provide flags for building against libcurl, so
use it instead of guessing flags and dependent libraries.
Allow overriding CURL_CONFIG to a custom path to curl-config, to
compile against a curl installation other than the first in PATH.
Depending on the set of features curl is compiled with, there may be
more libraries required than the previous two options of -lssl and
-lidn. For example, with a vanilla build of libcurl-7.36.0 on Mac OS X
10.9:
When a remote helper crashes while pushing we should revert back to the
state before the push, however, it's possible that `git fast-export`
already finished its job, and therefore has exported the marks already.
This creates a synchronization problem because from that moment on
`git fast-{import,export}` will have marks that the remote helper is not
aware of and all further commands fail (if those marks are referenced).
The fix is to tell `git fast-export` to export to a temporary file, and
only after the remote helper has finishes successfully, move to the
final destination.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
zsh seems to have a bug while redirecting the stderr of the 'read'
command:
% read foo 2>/dev/null <foo
zsh: no such file or directory: foo
Which causes errors to be displayed when certain files are missing.
Let's add a convenience function to manually check if the file is
readable before calling "read".
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since fd0a8c2e (first appearing in v1.7.0), the
t/t5560-http-backend-noserver.sh test has used a backslash escape
inside a ${} expansion in order to specify a literal '?' character.
Unfortunately the FreeBSD /bin/sh does not interpret this correctly.
In a POSIX compliant shell, the following:
x='one?two?three'
echo "${x#*\?}"
Would be expected to produce this:
two?three
When using the FreeBSD /bin/sh instead you get this:
one?two?three
In fact the FreeBSD /bin/sh treats the backslash as a literal
character to match so that this:
y='one\two\three'
echo "${y#*\?}"
Produces this unexpected value:
wo\three
In this case the backslash is not only treated literally, it also
fails to defeat the special meaning of the '?' character.
Instead, we can use the [...] construct to defeat the special meaning
of the '?' character and match it exactly in a way that works for the
FreeBSD /bin/sh as well as other POSIX /bin/sh implementations.
Changing the example like so:
x='one?two?three'
echo "${x#*[?]}"
Produces the expected output using the FreeBSD /bin/sh.
Therefore, change the use of \? to [?] in order to be compatible with
the FreeBSD /bin/sh which allows t/t5560-http-backend-noserver.sh to
pass on FreeBSD again.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 11502468 and 04c1ee57 (both first appearing in v1.8.5), the
t7001-mv test has used "cp -a" to perform a copy in several of the
tests.
However, the "-a" option is not required for a POSIX cp utility and
some platforms' cp utilities do not support it.
The POSIX equivalent of -a is -R -P -p.
Change "cp -a" to "cp -R -P -p" so that the t7001-mv test works
on systems with a cp utility that only implements the POSIX
required set of options and not the "-a" option.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
read-cache.c: verify index file before we opportunistically update it
Before we proceed to opportunistically update the index (often done
by an otherwise read-only operation like "git status" and "git diff"
that internally refreshes the index), we must verify that the
current index file is the same as the one that we read earlier
before we took the lock on it, in order to avoid a possible race.
In the example below git-status does "opportunistic update" and
git-rebase updates the index, but the race can happen in general.
1. process A calls git-rebase (or does anything that uses the index)
2. process A applies 1st commit
3. process B calls git-status (or does anything that updates the index)
4. process B reads index
5. process A applies 2nd commit
6. process B takes the lock, then overwrites process A's changes.
7. process A applies 3rd commit
As an end result the 3rd commit will have a revert of the 2nd commit.
When process B takes the lock, it needs to make sure that the index
hasn't changed since step 4.
Signed-off-by: Yiannis Marangos <yiannis.marangos@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
It is a common mistake to call read(2)/pread(2) and forget to
anticipate that they may return error with EAGAIN/EINTR when the
system call is interrupted.
We have xread() helper to relieve callers of read(2) from having to
worry about it; add xpread() helper to do the same for pread(2).
Update the caller in the builtin/index-pack.c and the mmap emulation
in compat/.
Signed-off-by: Yiannis Marangos <yiannis.marangos@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
completion: fix completing args of aliased "push", "fetch", etc.
Some commands need the first word to determine the actual action that is
being executed, however, the command is wrong when we use an alias, for
example 'alias.p=push', if we try to complete 'git p origin ', the
result would be wrong because __git_complete_remote_or_refspec() doesn't
know where it came from.
So let's override words[1], so the alias 'p' is override by the actual
command, 'push'.
Reported-by: Aymeric Beaumet <aymeric.beaumet@gmail.com> Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
remote-bzr: include authors field in pushed commits
Tests-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
remote-helpers: allow all tests running from any dir
Commit d3243d7 (test-bzr.sh, test-hg.sh: allow running from any dir)
allowed the tests to run from any directory, however, it didn't update
all the tests.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jl/nor-or-nand-and:
code and test: fix misuses of "nor"
comments: fix misuses of "nor"
contrib: fix misuses of "nor"
Documentation: fix misuses of "nor"
Merge branch 'cn/fetch-prune-overlapping-destination' into maint
* cn/fetch-prune-overlapping-destination:
fetch: handle overlaping refspecs on --prune
fetch: add a failing test for prunning with overlapping refspecs
Unicode 6.3 defines more code points as combining or accents. For
example, the character "ö" could be expressed as an "o" followed by
U+0308 COMBINING DIARESIS (aka umlaut, double-dot-above). We should
consider that such a sequence of two codepoints occupies one display
column for the alignment purposes, and for that, git_wcwidth()
should return 0 for them. Affected codepoints are:
Earlier unicode standards had defined these as "reserved".
Only the range 0..U+07FF has been checked to see which codepoints
need to be marked as 0-width while preparing for this commit; more
updates may be needed.
Signed-off-by: Torsten Bögershausen <tboegi@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Both MSVC and MINGW have alloca(3) definitions in malloc.h, so by moving
win32-compat alloca.h from compat/vcbuild/include/ to compat/win32/ ,
which is included by both MSVC and MINGW CFLAGS, we can make alloca()
work on both those Windows environments.
In MINGW, malloc.h has explicit check for GNUC and if it is so, defines
alloca to __builtin_alloca, so it looks like we don't need to add any
code to here-shipped alloca.h to get optimum performance.
Compile-tested on Windows in MSysGit.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru> Acked-by: Erik Faye-Lund <kusmabite@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* bp/commit-p-editor:
run-command: mark run_hook_with_custom_index as deprecated
merge hook tests: fix and update tests
merge: fix GIT_EDITOR override for commit hook
commit: fix patch hunk editing with "commit -p -m"
test patch hunk editing with "commit -p -m"
merge hook tests: use 'test_must_fail' instead of '!'
merge hook tests: fix missing '&&' in test
Eradicate mistaken use of "nor" (that is, essentially "nor" used
not in "neither A nor B" ;-)) from in-code comments, command output
strings, and documentations.
* jl/nor-or-nand-and:
code and test: fix misuses of "nor"
comments: fix misuses of "nor"
contrib: fix misuses of "nor"
Documentation: fix misuses of "nor"
Make sure that the help text given to describe the "<param>" part
of the "git cmd --option=<param>" does not contain SP or _,
e.g. "--gpg-sign=<key-id>" option for "git commit" is not spelled
as "--gpg-sign=<key id>".
* jc/rev-parse-argh-dashed-multi-words:
parse-options: make sure argh string does not have SP or _
update-index: teach --cacheinfo a new syntax "mode,sha1,path"
parse-options: multi-word argh should use dash to separate words
git-p4: explicitly specify that HEAD is a revision
'git p4 rebase' fails with the following message if there is a file
named HEAD in the current directory:
fatal: ambiguous argument 'HEAD': both revision and filename
Use '--' to separate paths from revisions, like this:
'git <command> [<revision>...] -- [<file>...]'
Take the suggestion above and explicitly state that HEAD should be
treated as a revision.
Signed-off-by: Vlad Dogaru <vdogaru@ixiacom.com> Acked-by: Pete Wyckoff <pw@padd.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
combine-diff: speed it up, by using multiparent diff tree-walker directly
As was recently shown in "combine-diff: optimize
combine_diff_path sets intersection", combine-diff runs very slowly. In
that commit we optimized paths sets intersection, but that accounted
only for ~ 25% of the slowness, and as my tracing showed, for linux.git
v3.10..v3.11, for merges a lot of time is spent computing
diff(commit,commit^2) just to only then intersect that huge diff to
almost small set of files from diff(commit,commit^1).
In previous commit, we described the problem in more details, and
reworked the diff tree-walker to be general one - i.e. to work in
multiple parent case too. Now is the time to take advantage of it for
finding paths for combine diff.
The implementation is straightforward - if we know, we can get generated
diff paths directly, and at present that means no diff filtering or
rename/copy detection was requested(*), we can call multiparent tree-walker
directly and get ready paths.
(*) because e.g. at present, all diffcore transformations work on
diff_filepair queues, but in the future, that limitation can be
lifted, if filters would operate directly on combine_diff_paths.
Timings for `git log --raw --no-abbrev --no-renames` without `-c` ("git log")
and with `-c` ("git log -c") and with `-c --merges` ("git log -c --merges")
before and after the patch are as follows:
linux.git v3.10..v3.11
log log -c log -c --merges
before 1.9s 16.4s 15.2s
after 1.9s 2.4s 1.1s
The result stayed the same.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru> Signed-off-by: Junio C Hamano <gitster@pobox.com>
tree-diff: rework diff_tree() to generate diffs for multiparent cases as well
Previously diff_tree(), which is now named ll_diff_tree_sha1(), was
generating diff_filepair(s) for two trees t1 and t2, and that was
usually used for a commit as t1=HEAD~, and t2=HEAD - i.e. to see changes
a commit introduces.
In Git, however, we have fundamentally built flexibility in that a
commit can have many parents - 1 for a plain commit, 2 for a simple merge,
but also more than 2 for merging several heads at once.
For merges there is a so called combine-diff, which shows diff, a merge
introduces by itself, omitting changes done by any parent. That works
through first finding paths, that are different to all parents, and then
showing generalized diff, with separate columns for +/- for each parent.
The code lives in combine-diff.c .
There is an impedance mismatch, however, in that a commit could
generally have any number of parents, and that while diffing trees, we
divide cases for 2-tree diffs and more-than-2-tree diffs. I mean there
is no special casing for multiple parents commits in e.g.
revision-walker .
That impedance mismatch *hurts* *performance* *badly* for generating
combined diffs - in "combine-diff: optimize combine_diff_path
sets intersection" I've already removed some slowness from it, but from
the timings provided there, it could be seen, that combined diffs still
cost more than an order of magnitude more cpu time, compared to diff for
usual commits, and that would only be an optimistic estimate, if we take
into account that for e.g. linux.git there is only one merge for several
dozens of plain commits.
That slowness comes from the fact that currently, while generating
combined diff, a lot of time is spent computing diff(commit,commit^2)
just to only then intersect that huge diff to almost small set of files
from diff(commit,commit^1).
That's because at present, to compute combine-diff, for first finding
paths, that "every parent touches", we use the following combine-diff
property/definition:
D(A,P1...Pn) is combined diff between commit A, and parents Pi
and
D(A,Pi) is usual two-tree diff Pi..A
So if any of that D(A,Pi) is huge, tracting 1 n-parent combine-diff as n
1-parent diffs and intersecting results will be slow.
And usually, for linux.git and other topic-based workflows, that
D(A,P2) is huge, because, if merge-base of A and P2, is several dozens
of merges (from A, via first parent) below, that D(A,P2) will be diffing
sum of merges from several subsystems to 1 subsystem.
The solution is to avoid computing n 1-parent diffs, and to find
changed-to-all-parents paths via scanning A's and all Pi's trees
simultaneously, at each step comparing their entries, and based on that
comparison, populate paths result, and deduce we could *skip*
*recursing* into subdirectories, if at least for 1 parent, sha1 of that
dir tree is the same as in A. That would save us from doing significant
amount of needless work.
Such approach is very similar to what diff_tree() does, only there we
deal with scanning only 2 trees simultaneously, and for n+1 tree, the
logic is a bit more complex:
A B
- -
|a| |b| a < b -> a ∉ B -> D(A,B) += +a a↓
|-| |-| a > b -> b ∉ A -> D(A,B) += -b b↓
| | | | a = b -> investigate δ(a,b) a↓ b↓
|-| |-|
|.| |.|
. .
. .
~~~~~~~~
This patch generalizes diff tree-walker to work with arbitrary number of
parents as described above - i.e. now there is a resulting tree t, and
some parents trees tp[i] i=[0..nparent). The generalization builds on
the fact that usual diff
D(A,B)
is by definition the same as combined diff
D(A,[B]),
so if we could rework the code for common case and make it be not slower
for nparent=1 case, usual diff(t1,t2) generation will not be slower, and
multiparent diff tree-walker would greatly benefit generating
combine-diff.
What we do is as follows:
1) diff tree-walker ll_diff_tree_sha1() is internally reworked to be
a paths generator (new name diff_tree_paths()), with each generated path
being `struct combine_diff_path` with info for path, new sha1,mode and for
every parent which sha1,mode it was in it.
2) From that info, we can still generate usual diff queue with
struct diff_filepairs, via "exporting" generated
combine_diff_path, if we know we run for nparent=1 case.
(see emit_diff() which is now named emit_diff_first_parent_only())
3) In order for diff_can_quit_early(), which checks
DIFF_OPT_TST(opt, HAS_CHANGES))
to work, that exporting have to be happening not in bulk, but
incrementally, one diff path at a time.
For such consumers, there is a new callback in diff_options
introduced:
->pathchange(opt, struct combine_diff_path *)
which, if set to !NULL, is called for every generated path.
(see new compat ll_diff_tree_sha1() wrapper around new paths
generator for setup)
4) The paths generation itself, is reworked from previous
ll_diff_tree_sha1() code according to "D(A,P1...Pn) calculation
scheme" provided above:
On the start we allocate [nparent] arrays in place what was
earlier just for one parent tree.
then we just generalize loops, and comparison according to the
algorithm.
Some notes(*):
1) alloca(), for small arrays, is used for "runs not slower for
nparent=1 case than before" goal - if we change it to xmalloc()/free()
the timings get ~1% worse. For alloca() we use just-introduced
xalloca/xalloca_free compatibility wrappers, so it should not be a
portability problem.
2) For every parent tree, we need to keep a tag, whether entry from that
parent equals to entry from minimal parent. For performance reasons I'm
keeping that tag in entry's mode field in unused bit - see S_IFXMIN_NEQ.
Not doing so, we'd need to alloca another [nparent] array, which hurts
performance.
3) For emitted paths, memory could be reused, if we know the path was
processed via callback and will not be needed later. We use efficient
hand-made realloc-style path_appendnew(), that saves us from ~1-1.5%
of potential additional slowdown.
4) goto(s) are used in several places, as the code executes a little bit
faster with lowered register pressure.
Also
- we should now check for FIND_COPIES_HARDER not only when two entries
names are the same, and their hashes are equal, but also for a case,
when a path was removed from some of all parents having it.
The reason is, if we don't, that path won't be emitted at all (see
"a > xi" case), and we'll just skip it, and FIND_COPIES_HARDER wants
all paths - with diff or without - to be emitted, to be later analyzed
for being copies sources.
The new check is only necessary for nparent >1, as for nparent=1 case
xmin_eqtotal always =1 =nparent, and a path is always added to diff as
removal.
~~~~~~~~
Timings for
# without -c, i.e. testing only nparent=1 case
`git log --raw --no-abbrev --no-renames`
before and after the patch are as follows:
navy.git linux.git v3.10..v3.11
before 0.611s 1.889s
after 0.619s 1.907s
slowdown 1.3% 0.9%
This timings show we did no harm to usual diff(tree1,tree2) generation.
From the table we can see that we actually did ~1% slowdown, but I think
I've "earned" that 1% in the previous patch ("tree-diff: reuse base
str(buf) memory on sub-tree recursion", HEAD~~) so for nparent=1 case,
net timings stays approximately the same.
The output also stayed the same.
(*) If we revert 1)-4) to more usual techniques, for nparent=1 case,
we'll get ~2-2.5% of additional slowdown, which I've tried to avoid, as
"do no harm for nparent=1 case" rule.
For linux.git, combined diff will run an order of magnitude faster and
appropriate timings will be provided in the next commit, as we'll be
taking advantage of the new diff tree-walker for combined-diff
generation there.
P.S. and combined diff is not some exotic/for-play-only stuff - for
example for a program I write to represent Git archives as readonly
filesystem, there is initial scan with
to extract log of what was created/changed when, as a result building a
map
{} sha1 -> in which commit (and date) a content was added
that `-c` means also show combined diff for merges, and without them, if
a merge is non-trivial (merges changes from two parents with both having
separate changes to a file), or an evil one, the map will not be full,
i.e. some valid sha1 would be absent from it.
That case was my initial motivation for combined diffs speedup.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru> Signed-off-by: Junio C Hamano <gitster@pobox.com>
It used to be that ref_transaction_commit() allocated a temporary
array to hold the types of references while it is working. Instead,
add a type field to ref_update that ref_transaction_commit() can use
as its scratch space.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that we manage ref_update objects internally, we can use them to
hold some of the scratch space we need when actually carrying out the
updates. Store the (struct ref_lock *) there.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
update-ref --stdin: reimplement using reference transactions
This change is mostly clerical: the parse_cmd_*() functions need to
use local variables rather than a struct ref_update to collect the
arguments needed for each update, and then call ref_transaction_*() to
queue the change rather than building up the list of changes at the
caller side.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Build out the API for dealing with a bunch of reference checks and
changes within a transaction. Define an opaque ref_transaction type
that is managed entirely within refs.c. Introduce functions for
beginning a transaction, adding updates to a transaction, and
committing/rolling back a transaction.
This API will soon replace update_refs().
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Make (most of) the error messages for invalid input have the same
format [1]:
$COMMAND [SP $REFNAME]: $MESSAGE
Update the tests accordingly.
[1] A few error messages are left with their old form, because
$COMMAND and $REFNAME aren't passed all the way down the call
stack. Maybe those sites should be changed some day, too.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
This case wants to test passing a bad refname to the "update" command.
But it also passes too few arguments to "update", which muddles the
situation: which error should be diagnosed? So split this test into
two:
* One that passes too few arguments to update
* One that passes all three arguments to "update", but with a bad
refname.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
update-ref --stdin -z: deprecate interpreting the empty string as zeros
In the original version of this command, for the single case of the
"update" command's <newvalue>, the empty string was interpreted as
being equivalent to 40 "0"s. This shorthand is unnecessary (binary
input will usually be generated programmatically anyway), and it
complicates the parser and the documentation.
So gently deprecate this usage: remove its description from the
documentation and emit a warning if it is found. But for reasons of
backwards compatibility, continue to accept it.
Helped-by: Brad King <brad.king@kitware.com> Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
update-ref.c: extract a new function, parse_next_sha1()
Replace three functions, update_store_new_sha1(),
update_store_old_sha1(), and parse_next_arg(), with a single function,
parse_next_sha1(). The new function takes care of a whole argument,
including checking whether it is there, converting it to an SHA-1, and
emitting errors on EOF or for invalid values. The return value
indicates whether the argument was present or absent, which requires
a bit of intelligence because absent values are represented
differently depending on whether "-z" was used.
The new interface means that the calling functions, parse_cmd_*(),
don't have to interpret the result differently based on the
line_termination mode that is in effect. It also means that
parse_cmd_create() can distinguish unambiguously between an empty new
value and a zeros new value, which fixes a failure in t1400.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
update-ref --stdin: make error messages more consistent
The old error messages emitted for invalid input sometimes said
"<oldvalue>"/"<newvalue>" and sometimes said "old value"/"new value".
Convert them all to the former. Update the tests accordingly.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
update-ref --stdin: improve error messages for invalid values
If an invalid value is passed to "update-ref --stdin" as <oldvalue> or
<newvalue>, include the command and the name of the reference at the
beginning of the error message. Update the tests accordingly.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
update-ref.c: extract a new function, parse_refname()
There is no reason to obscure the fact that parse_first_arg() always
parses refnames. Form the new function by combining parse_first_arg()
and update_store_ref_name().
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
parse_cmd_verify(): copy old_sha1 instead of evaluating <oldvalue> twice
Aside from avoiding a tiny bit of work, this makes it transparently
obvious that old_sha1 and new_sha1 are identical. It is arguably a
bit silly to have to set new_sha1 in order to verify old_sha1, but
that is a problem for another day.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Read the whole input into a strbuf at once, and then parse it from
there. This might also be a tad faster, but that is not the point.
The point is to decouple the parsing code from the input source (the
old parsing code had to read new data even in the middle of commands).
Add docstrings for the parsing functions.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The old signature of update_refs() required a
(const struct ref_update **) for its updates_orig argument. The
"const" is presumably there to promise that the function will not
modify the contents of the structures.
But this declaration does not permit the function to be called with a
(struct ref_update **), which is perfectly legitimate. C's type
system is not powerful enough to express what we'd like. So remove
the first "const" from the declaration.
On the other hand, the function *can* promise not to modify the
pointers within the array that is passed to it without inconveniencing
its callers. So add a "const" that has that effect, making the final
declaration
(struct ref_update * const *).
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Given that these constants are only being used when updating
references, it is inappropriate to give them such generic names as
"DIE_ON_ERR". So prefix their names with "UPDATE_REFS_".
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
parse_arg(): really test that argument is properly terminated
The old parse_arg(), when fed an argument
"refs/heads/a"master
parsed 'refs/heads/a' off of the front of the argument and considered
itself successful. It was only when parse_next_arg() tried to parse
the *next* argument that a problem was noticed. But in fact, the
definition of the input format requires arguments to be terminated by
SP or NUL, so *this* argument is already erroneous and parse_arg()
should diagnose the problem.
So teach parse_arg() to verify that C-quoted arguments are terminated
correctly. If not, emit a more specific error message.
There is no corresponding error case of a non-C-quoted argument that
is not terminated correctly, because the end of a non-quoted argument
is *by definition* a space or NUL, so there is no way to insert other
junk between the "end" of the argument and the argument terminator.
Adjust the tests to expect the new error message. Add a docstring to
the function, incorporating the comments that were formerly within the
function plus some added information.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
actually passes an empty new value, not a zero new value. So rename
the test s/zero/empty/, and change the expected error from
fatal: create $c given zero new value
to
fatal: create $c missing <newvalue>
Of course, this makes the test fail now, because although "git
update-ref" tries to distinguish between these two errors, it does not
succeed in this situation. Fixing it is more than a one-liner, so
mark the test test_expect_failure for now. The failure will be fixed
later in this patch series.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Junio C Hamano <gitster@pobox.com>
pack-objects: do not reuse packfiles without --delta-base-offset
When we are sending a packfile to a remote, we currently try
to reuse a whole chunk of packfile without bothering to look
at the individual objects. This can make things like initial
clones much lighter on the server, as we can just dump the
packfile bytes.
However, it's possible that the other side cannot read our
packfile verbatim. For example, we may have objects stored
as OFS_DELTA, but the client is an antique version of git
that only understands REF_DELTA. We negotiate this
capability over the fetch protocol. A normal pack-objects
run will convert OFS_DELTA into REF_DELTA on the fly, but
the "reuse pack" code path never even looks at the objects.
This patch disables packfile reuse if the other side is
missing any capabilities that we might have used in the
on-disk pack. Right now the only one is OFS_DELTA, but we
may need to expand in the future (e.g., if packv4 introduces
new object types).
We could be more thorough and only disable reuse in this
case when we actually have an OFS_DELTA to send, but:
1. We almost always will have one, since we prefer
OFS_DELTA to REF_DELTA when possible. So this case
would almost never come up.
2. Looking through the objects defeats the purpose of the
optimization, which is to do as little work as possible
to get the bytes to the remote.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When pack-objects is computing the reachability bitmap to
serve a fetch request, it can erroneously die() if some of
the UNINTERESTING objects are not present. Upload-pack
throws away HAVE lines from the client for objects we do not
have, but we may have a tip object without all of its
ancestors (e.g., if the tip is no longer reachable and was
new enough to survive a `git prune`, but some of its
reachable objects did get pruned).
In the non-bitmap case, we do a revision walk with the HAVE
objects marked as UNINTERESTING. The revision walker
explicitly ignores errors in accessing UNINTERESTING commits
to handle this case (and we do not bother looking at
UNINTERESTING trees or blobs at all).
When we have bitmaps, however, the process is quite
different. The bitmap index for a pack-objects run is
calculated in two separate steps:
First, we perform an extensive walk from all the HAVEs to
find the full set of objects reachable from them. This walk
is usually optimized away because we are expected to hit an
object with a bitmap during the traversal, which allows us
to terminate early.
Secondly, we perform an extensive walk from all the WANTs,
which usually also terminates early because we hit a commit
with an existing bitmap.
Once we have the resulting bitmaps from the two walks, we
AND-NOT them together to obtain the resulting set of objects
we need to pack.
When we are walking the HAVE objects, the revision walker
does not know that we are walking it only to mark the
results as uninteresting. We strip out the UNINTERESTING flag,
because those objects _are_ interesting to us during the
first walk. We want to keep going to get a complete set of
reachable objects if we can.
We need some way to tell the revision walker that it's OK to
silently truncate the HAVE walk, just like it does for the
UNINTERESTING case. This patch introduces a new
`ignore_missing_links` flag to the `rev_info` struct, which
we set only for the HAVE walk.
It also adds tests to cover UNINTERESTING objects missing
from several positions: a missing blob, a missing tree, and
a missing parent commit. The missing blob already worked (as
we do not care about its contents at all), but the other two
cases caused us to die().
Note that there are a few cases we do not need to test:
1. We do not need to test a missing tree, with the blob
still present. Without the tree that refers to it, we
would not know that the blob is relevant to our walk.
2. We do not need to test a tip commit that is missing.
Upload-pack omits these for us (and in fact, we
complain even in the non-bitmap case if it fails to do
so).
Reported-by: Siddharth Agarwal <sid0@fb.com> Signed-off-by: Vicent Marti <tanoku@gmail.com> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Merge branch 'mh/remove-subtree-long-pathname-fix' into maint
* mh/remove-subtree-long-pathname-fix:
entry.c: fix possible buffer overflow in remove_subtree()
checkout_entry(): use the strbuf throughout the function
* jk/shallow-update-fix:
shallow: verify shallow file after taking lock
shallow: automatically clean up shallow tempfiles
shallow: use stat_validity to check for up-to-date file