We're testing that trying to --track a ref that is not covered by any remote
refspec should fail. For that, we want to have refs/remotes/local/master
present, but we also want the remote.local.fetch refspec to NOT match
refs/remotes/local/master (so that the tracking setup will fail, as intended).
However, when doing "git fetch local" to ensure the existence of
refs/remotes/local/master, we must not already have changed remote.local.fetch
so as to cause refs/remotes/local/master not to be fetched. Therefore, set
remote.local.fetch to refs/heads/*:refs/remotes/local/* BEFORE we fetch, and
then reset it to refs/heads/s:refs/remotes/local/s AFTER we have fetched
(but before we test --track).
Signed-off-by: Johan Herland <johan@herland.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is similar in spirit to "make -C dir ..." and "tar -C dir ...".
It takes more keypresses to invoke git command in a different
directory without leaving the current directory:
1. (cd ~/foo && git status)
git --git-dir=~/foo/.git --work-dir=~/foo status
GIT_DIR=~/foo/.git GIT_WORK_TREE=~/foo git status
2. (cd ../..; git grep foo)
3. for d in d1 d2 d3; do (cd $d && git svn rebase); done
The methods shown above are acceptable for scripting but are too
cumbersome for quick command line invocations.
With this new option, the above can be done with fewer keystrokes:
1. git -C ~/foo status
2. git -C ../.. grep foo
3. for d in d1 d2 d3; do git -C $d svn rebase; done
A new test script is added to verify the behavior of this option with
other path-related options like --git-dir and --work-tree.
Signed-off-by: Nazri Ramliy <ayiehere@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When it was originally added, the git_remote_helpers library was used as
part of the tests of the remote-helper interface, but since commit
fc407f9 (Add new simplified git-remote-testgit, 2012-11-28) a simple
shell script is used for this.
A search on Ohloh [1] indicates that this library isn't used by any
external projects and even the Python remote helpers in contrib/ don't
use this library, so it is only used by its own test suite.
Since this is the only Python library in Git, removing it will make
packaging easier as the Python scripts only need to be installed for one
version of Python, whereas the library should be installed for all
available versions.
[1] http://code.ohloh.net/search?s=%22git_remote_helpers%22
Signed-off-by: John Keeping <john@keeping.me.uk>
Acked-by: Sverre Rabbelier <srabbelier@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
List of files in other sections ("Changes to be committed", ...) end with
a blank line. It is not the case with the "Untracked files" and "Ignored
files" sections. The issue become particularly visible after the #-prefix
removal, as the last line (e.g. "nothing added to commit but untracked
files present") seems mixed with the untracked files.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous commit set status.displayCommentPrefix file-wide in
t7060-wtstatus.sh, t7508-status.sh and t/t7512-status-help.sh to make the
patch small. However, now that status.displayCommentPrefix is not the
default, it is better to disable it in tests so that the most common
situation is also the most tested.
While we're there, move the "cat > expect << EOF" blocks inside the
tests.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Historically, "git status" needed to prefix each output line with '#' so
that the output could be added as comment to the commit message. This
prefix comment has no real purpose when "git status" is ran from the
command-line, and this may distract users from the real content.
Disable this prefix comment by default, and make it re-activable for
users needing backward compatibility with status.displayCommentPrefix.
Obviously, "git commit" ignores status.displayCommentPrefix and keeps the
comment unconditionnaly when writing to COMMIT_EDITMSG (but not when
writing to stdout for an error message or with --dry-run).
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The --for-status option was an undocumented option used only by
wt-status.c, which inserted a header and commented out the output. We can
achieve the same result within wt-status.c, without polluting the
submodule command-line options.
This will make it easier to disable the comments from wt-status.c later.
The --for-status is kept so that another topic in flight
(bc/submodule-status-ignored) can continue relying on it, although it is
currently a no-op.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The recent "short-cut clone connectivity check" topic broke a shallow
repository when a fetch operation tries to auto-follow tags.
* nd/fetch-pack-shallow-fix:
fetch-pack: do not remove .git/shallow file when --depth is not specified
Replace 'committish' in documentation and comments with 'commit-ish'
to match gitglossary(7) and to be consistent with 'tree-ish'.
The only remaining instances of 'committish' are:
* variable, function, and macro names
* "(also committish)" in the definition of commit-ish in
gitglossary[7]
Signed-off-by: Richard Hansen <rhansen@bbn.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If a user is working on master, and has merged in their feature branch, but now
has to "git pull" because master moved, with pull.rebase their feature branch
will be flattened into master.
This is because "git pull" currently does not know about rebase's preserve
merges flag, which would avoid this behavior, as it would instead replay just
the merge commit of the feature branch onto the new master, and not replay each
individual commit in the feature branch.
Add a --rebase=preserve option, which will pass along --preserve-merges to
rebase.
Also add 'preserve' to the allowed values for the pull.rebase config setting.
Signed-off-by: Stephen Haberman <stephen@exigencecorp.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fixes a minor bug in "git rebase -i" (there could be others, as the
root cause is pretty generic) where the code feeds a random, data
dependeant string to 'echo' and expects it to come out literally.
* mm/no-shell-escape-in-die-message:
die_with_status: use "printf '%s\n'", not "echo"
* jk/fast-import-empty-ls:
fast-import: allow moving the root tree
fast-import: allow ls or filecopy of the root tree
fast-import: set valid mode on root tree in "ls"
t9300: document fast-import empty path issues
Commit 68be2fea (receive-pack, fetch-pack: reject bogus pack that
records objects twice, 2011-11-16) taught index-pack to notice and
reject duplicate objects if --strict is given (which it is for
incoming packs, if transfer.fsckObjects is set). However, it never
tested the code, because we did not have an easy way of generating
such a bogus pack.
Now that we have test infrastructure to handle this, let's confirm
that it works.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Complete the <rev>^{<type>} family of object descriptors by having
<rev>^{tag} dereference <rev> until a tag object is found (or fail if
unable).
At first glance this may not seem very useful, as commits, trees, and
blobs cannot be peeled to a tag, and a tag would just peel to itself.
However, this can be used to ensure that <rev> names a tag object:
$ git rev-parse --verify v1.8.4^{tag}
04f013dc38d7512eadb915eba22efc414f18b869
$ git rev-parse --verify master^{tag}
error: master^{tag}: expected tag type, but the object dereferences to tree type
fatal: Needed a single revision
Users can already ensure that <rev> is a tag object by checking the
output of 'git cat-file -t <rev>', but:
* users may expect <rev>^{tag} to exist given that <rev>^{commit},
<rev>^{tree}, and <rev>^{blob} all exist
* this syntax is more convenient/natural in some circumstances
Signed-off-by: Richard Hansen <rhansen@bbn.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Save the reader from learning specialized t6* setup functions
where familiar commands like test_commit, "git checkout --orphan",
and "git merge" will do.
While at it, wrap the setup commands in a test assertion so errors can
be caught and stray output suppressed when running without --verbose
as in other tests.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use test_cmp instead of passing two command substitutions to the
"test" builtin. This way:
- when tests fail, they can print a helpful diff if run with
"--verbose"
- the argument order "test_cmp expect actual" feels natural,
unlike test <known> = <unknown> that seems backwards
- the exit status from invoking git is checked, so if rev-parse
starts segfaulting then the test will notice and fail
Use a custom function for this instead of test_cmp_rev to emphasize
that we are testing the output from "git rev-parse" with certain
arguments, not checking that the revisions are equal in abstract.
Reported-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This way, if rev-parse segfaults then the test will fail instead
of treating it the same way as a controlled failure.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of cramming everything in one line, put the test body in an
indented block after the opening test_expect_success line and quote
and put the closing quote on a line by itself.
Use single-quote instead of double-quote to quote the test body
for more useful --verbose output.
Signed-off-by: Felipe Contreras <felipe.contreras@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 664059fb (transport-helper: update remote helper namespace,
2013-04-17), a 'push' operation on a remote helper updates the
private ref by default. This is often a good thing, but it can also
be desirable to disable this update to force the next 'pull' to
re-import the pushed revisions.
Allow remote-helpers to disable the automatic update by introducing a new
capability.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This makes reinitializing on a .git file repository work.
This is probably the only case that setup_git_env() (via
set_git_dir()) is called on a .git file. Other cases in
setup_git_dir_gently() and enter_repo() both cover .git file case
explicitly because they need to verify the target repo is valid.
Reported-by: Ximin Luo <infinity0@gmx.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some people still use rather old versions of bash, which cannot
grok some constructs like 'printf -v varname' the prompt and
completion code started to use recently.
* bc/completion-for-bash-3.0:
contrib/git-prompt.sh: handle missing 'printf -v' more gracefully
t9902-completion.sh: old Bash still does not support array+=('') notation
git-completion.bash: use correct Bash/Zsh array length syntax
The recent "short-cut clone connectivity check" topic broke a
shallow repository when a fetch operation tries to auto-follow tags.
* nd/fetch-pack-shallow-fix:
fetch-pack: do not remove .git/shallow file when --depth is not specified
Commit 4337b58 (do not write null sha1s to on-disk index,
2012-07-28) added a safety check preventing git from writing
null sha1s into the index. The intent was to catch errors in
other parts of the code that might let such an entry slip
into the index (or worse, a tree).
Some existing repositories may have invalid trees that
contain null sha1s already, though. Until 4337b58, a common
way to clean this up would be to use git-filter-branch's
index-filter to repair such broken entries. That now fails
when filter-branch tries to write out the index.
Introduce a GIT_ALLOW_NULL_SHA1 environment variable to
relax this check and make it easier to recover from such a
history.
It is tempting to not involve filter-branch in this commit
at all, and instead require the user to manually invoke
GIT_ALLOW_NULL_SHA1=1 git filter-branch ...
to perform an index-filter on a history with trees with null
sha1s. That would be slightly safer, but requires some
specialized knowledge from the user. So let's set the
GIT_ALLOW_NULL_SHA1 variable automatically when checking out
the to-be-filtered trees. Advice on using filter-branch to
remove such entries already exists on places like
stackoverflow, and this patch makes it Just Work again on
recent versions of git.
Further commands that touch the index will still notice and
fail, unless they actually remove the broken entries. A
filter-branch whose filters do not touch the index at all
will not error out (since we complain of the null sha1 only
on writing, not when making a tree out of the index), but
this is acceptable, as we still print a loud warning, so the
problem is unlikely to go unnoticed.
Signed-off-by: Jeff King <peff@peff.net>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The read_mailmap_buf function reads each line of the mailmap
using strchrnul, like:
const char *end = strchrnul(buf, '\n');
unsigned long linelen = end - buf + 1;
But that's off-by-one when we actually hit the NUL byte; our
line does not have a terminator, and so is only "end - buf"
bytes long. As a result, when we subtract the linelen from
the total len, we end up with (unsigned long)-1 bytes left
in the buffer, and we start reading random junk from memory.
We could fix it with:
unsigned long linelen = end - buf + !!*end;
but let's take a step back for a moment. It's questionable
in the first place for a function that takes a buffer and
length to be using strchrnul. But it works because we only
have one caller (and are only likely to ever have this one),
which is handing us data from read_sha1_file. Which means
that it's always NUL-terminated.
Instead of tightening the assumptions to make the
buffer/length pair work for a caller that doesn't actually
exist, let's let loosen the assumptions to what the real
caller has: a modifiable, NUL-terminated string.
This makes the code simpler and shorter (because we don't
have to correlate strchrnul with the length calculation),
correct (because the code with the off-by-one just goes
away), and more efficient (we can drop the extra allocation
we needed to create NUL-terminated strings for each line,
and just terminate in place).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is a testcase that checks for a problem where, during a specific
shallow fetch where the client does not have any commits that are a
successor of the new shallow root (i.e., the fetch creates a new
detached piece of history), the server would simply send over _all_
objects, instead of taking into account the objects already present in
the client.
The actual problem was fixed by a recent patch series by Nguyễn Thái
Ngọc Duy already.
Signed-off-by: Matthijs Kooijman <matthijs@stdin.nl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
upload-pack has a special revision walking code for shallow
recipients. It works almost like the similar code in pack-objects
except:
1. in upload-pack, graft points could be added for deepening;
2. also when the repository is deepened, the shallow point will be
moved further away from the tip, but the old shallow point will be
marked as edge to produce more efficient packs. See 6523078 (make
shallow repository deepening more network efficient - 2009-09-03).
Pass the file to pack-objects via --shallow-file. This will override
$GIT_DIR/shallow and give pack-objects the exact repository shape
that upload-pack has.
mark edge commits by revision command arguments. Even if old shallow
points are passed as "--not" revisions as in this patch, they will not
be picked up by mark_edges_uninteresting() because this function looks
up to parents for edges, while in this case the edge is the children,
in the opposite direction. This will be fixed in an later patch when
all given uninteresting commits are marked as edges.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When core.precomposeunicode was introduced in 76759c7d,
it was set to false on a unicode decomposing file system like HFS+
to be compatible with older versions of Git.
The Mac OS users need to find out that this configuration exist
and change it manually from false to true.
A smoother workflow can be achieved,
so set core.precomposeunicode to true on a decomposing file system.
Signed-off-by: Torsten Bögershausen <tboegi@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In order to see what the current branch is tracking, one way is using
"git branch -v -v", but branches other than the current are also
reported. Another way is using "git status", such as:
$ git status
# On branch master
# Your branch is ahead of 'origin/master' by 1 commit.
...
But this will not work if there is no change between the current
branch and its upstream. Always report upstream tracking info
even if there is no difference, so that "git status" is consistent
for checking tracking info for current branch. E.g.
$ git status
# On branch feature1
# Your branch is up-to-date with 'github/feature1'.
...
$ git status -bs
## feature1...github/feature1
...
$ git checkout feature1
Already on 'feature1'
Your branch is up-to-date with 'github/feature1'.
...
Also add some test cases in t6040.
Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Command "git branch -vv" will report tracking branches, but invalid
tracking branches are also reported. This is because the function
stat_tracking_info() can not distinguish invalid tracking branch
from other cases which it would not like to report, such as
there is no upstream settings at all, or nothing is changed between
one branch and its upstream.
Junio suggested missing upstream should be reported [1] like:
$ git branch -v -v
master e67ac84 initial
* topic 3fc0f2a [topicbase: gone] topic
$ git status
# On branch topic
# Your branch is based on 'topicbase', but the upstream is gone.
# (use "git branch --unset-upstream" to fixup)
...
$ git status -b -s
## topic...topicbase [gone]
...
In order to do like that, we need to distinguish these three cases
(i.e. no tracking, with configured but no longer valid tracking, and
with tracking) in function stat_tracking_info(). So the refactored
function stat_tracking_info() has three return values: -1 (with "gone"
base), 0 (no base), and 1 (with base).
If the caller does not like to report tracking info when nothing
changed between the branch and its upstream, simply checks if
num_theirs and num_ours are both 0.
[1]: http://thread.gmane.org/gmane.comp.version-control.git/231830/focus=232288
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'todo' sheet for interactive rebase shows abbreviated SHA-1's and
then performs its operations upon those shortened values. This can lead
to an abort if the SHA-1 of a reworded or edited commit is no longer
unique within the abbreviated SHA-1 space and a subsequent SHA-1 in the
todo list has the same abbreviated value.
For example:
edit f00dfad first
pick badbeef second
If, after editing, the new SHA-1 of "first" also has prefix badbeef,
then the subsequent 'pick badbeef second' will fail since badbeef is no
longer a unique SHA-1 abbreviation:
error: short SHA1 badbeef is ambiguous.
fatal: Needed a single revision
Invalid commit name: badbeef
Fix this problem by expanding the SHA-1's in the todo list before
performing the operations.
[es: also collapse & expand SHA-1's for --edit-todo; respect
core.commentchar in transform_todo_ids(); compose commit message]
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'todo' sheet for interactive rebase shows abbreviated SHA-1's and
then performs its operations upon those shortened values. This can lead
to an abort if the SHA-1 of a reworded or edited commit is no longer
unique within the abbreviated SHA-1 space and a subsequent SHA-1 in the
todo list has the same abbreviated value.
For example:
edit f00dfad first
pick badbeef second
If, after editing, the new SHA-1 of "first" also has prefix badbeef,
then the subsequent 'pick badbeef second' will fail since badbeef is no
longer a unique SHA-1 abbreviation:
error: short SHA1 badbeef is ambiguous.
fatal: Needed a single revision
Invalid commit name: badbeef
Demonstrate this problem with a couple of specially crafted commits
which initially have distinct abbreviated SHA-1's, but for which the
abbreviated SHA-1's collide after a simple rewording of the first
commit's message.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As its very first action, t3404 installs (via set_fake_editor) a
specialized $EDITOR which simplifies automated 'rebase -i' testing. Many
tests rely upon this setting, thus tests which need a different editor
must take extra care upon completion to restore $EDITOR in order to
avoid breaking following tests. This places extra burden upon such tests
and requires that they undesirably have extra knowledge about
surrounding tests. Ease this burden by having each test install the
$EDITOR it requires, rather than relying upon a global setting.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
fetch_pack() can remove .git/shallow file when a shallow repository
becomes a full one again. This behavior is triggered incorrectly when
tags are also fetched because fetch_pack() will be called twice. At
the first fetch_pack() call:
- shallow_lock is set up
- alternate_shallow_file points to shallow_lock.filename, which is
"shallow.lock"
- commit_lock_file is called, which sets shallow_lock.filename to "".
alternate_shallow_file also becomes "" because it points to the
same memory.
At the second call, setup_alternate_shallow() is not called and
alternate_shallow_file remains "". It's mistaken as unshallow case and
.git/shallow is removed. The end result is a broken repository.
Fix this by always initializing alternate_shallow_file when
fetch_pack() is called. As an extra measure, check if args->depth > 0
before commit/rollback shallow file.
Reported-by: Kacper Kornet <kornet@camk.edu.pl>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Avoid command substitution and pipes to ensure that the exit status
from each git command is tested (and in particular that any segfaults
are caught).
Maintain the test setup (no commits, one file named "a", another named
"b") even after the last test, to make it easier to rearrange tests or
add new tests after the last in the future.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The test 'reset -p' uses git-reset -p, so it depends on the perl code.
Signed-off-by: Kacper Kornet <draenog@pld-linux.org>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It is too easy to forget to add the PERL prerequisite for new
"add -i" tests, especially given that many people do not test with
NO_PERL so the missing prereq is not always noticed quickly.
The test had used the skip_all mechanism since 1b19ccd2 (2009-04-03)
but switched to explicit PERL prereqs in f0459319 (2010-10-13) in hope
of helping people see how many tests were skipped, perhaps to motivate
them to tweak their platform or tests to improve test coverage. That
didn't pan out much in practice, so let's move back to the simpler
skip_all method.
Reported-by: Kacper Kornet <draenog@pld-linux.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the test 'using invalid commit with -C' git-commit would have failed
even if the -C option had been given the correct commit, as there was
nothing to commit. Pass --allow-empty to make sure it would make a commit,
were there no issues with the argument given to the -C option.
Signed-off-by: Kacper Kornet <draenog@pld-linux.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The previous commit added tests to show that index-pack
correctly bails in unrecoverable situations. There are some
situations where the data could be recovered, but it is not
currently:
1. If we can break the cycle using an object from another
pack via --fix-thin.
2. If we can break the cycle using a duplicate of one of
the objects found in the same pack.
Note that neither of these is particularly high priority; a
delta cycle within a pack should never occur, and we have no
record of even a buggy git implementation creating such a
pack.
However, it's worth adding these tests for two reasons. One,
to document that we do not currently handle the situation,
even though it is possible. And two, to exercise the code
that runs in this situation; even though it fails, by
running it we can confirm that index-pack detects the
situation and aborts, and does not misbehave (e.g., by
following the cycle in an infinite loop).
In both cases, we hit an assert that aborts index-pack.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If we receive a broken or malicious pack from a remote, we
will feed it to index-pack. As index-pack processes the
objects as a stream, reconstructing and hashing each object
to get its name, it is not very susceptible to doing the
wrong with bad data (it simply notices that the data is
bogus and aborts).
However, one question raised on the list is whether it could
be susceptible to problems during the delta-resolution
phase. In particular, can a cycle in the packfile deltas
cause us to go into an infinite loop or cause any other
problem?
The answer is no.
We cannot have a cycle of delta-base offsets, because they
go only in one direction (the OFS_DELTA object mentions its
base by an offset towards the beginning of the file, and we
explicitly reject negative offsets).
We can have a cycle of REF_DELTA objects, which refer to
base objects by sha1 name. However, index-pack does not know
these sha1 names ahead of time; it has to reconstruct the
objects to get their names, and it cannot do so if there is
a delta cycle (in other words, it does not even realize
there is a cycle, but only that there are items that cannot
be resolved).
Even though we can reason out that index-pack should handle
this fine, let's add a few tests to make sure it behaves
correctly.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The sha1_entry_pos function tries to be smart about
selecting the middle of a range for its binary search by
looking at the value differences between the "lo" and "hi"
constraints. However, it is unable to cope with entries with
duplicate keys in the sorted list.
We may hit a point in the search where both our "lo" and
"hi" point to the same key. In this case, the range of
values between our endpoints is 0, and trying to scale the
difference between our key and the endpoints over that range
is undefined (i.e., divide by zero). The current code
catches this with an "assert(lov < hiv)".
Moreover, after seeing that the first 20 byte of the key are
the same, we will try to establish a value from the 21st
byte. Which is nonsensical.
Instead, we can detect the case that we are in a run of
duplicates, and simply do a final comparison against any one
of them (since they are all the same, it does not matter
which). If the keys match, we have found our entry (or one
of them, anyway). If not, then we know that we do not need
to look further, as we must be in a run of the duplicate
key.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git commit --author=$name" sets the author to one whose name
matches the given string from existing commits, when $name is not in
the "Name <e-mail>" format. However, it does not honor the mailmap
to use the canonical name for the author found this way.
Fix it by telling the logic to find a matching existing author to
honor the mailmap, and use the name and email after applying the
mailmap.
Signed-off-by: Antoine Pelisse <apelisse@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
directory_exists_in_index() takes pathname and its length, but its
helper function directory_exists_in_index_icase() reads one byte
beyond the end of the pathname and expects there to be a '/'.
This needs to be fixed, as that one-byte-beyond-the-end location may
not even be readable, possibly by not registering directories to
name hashes with trailing slashes. In the meantime, update the new
caller added recently to treat_one_path() to make sure that the path
buffer it gives the function is one byte longer than the path it is
asking the function about by appending a slash to it.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Old Bash (3.0) which is distributed with RHEL 4.X and other ancient
platforms that are still in wide use, does not understand the
array+=() notation. Let's use an explicit assignment to the new array
element which works everywhere, like:
array[${#array[@]}+1]=''
The right-hand side '' is not strictly necessary, but in this case
I think it is more clear.
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "merge.log" config is set, "rebase --preserve-merges" will add
the log lines to the message of the rebased merge commit. A rebase
should not modify a commit message automatically.
Teach "git-rebase" to ignore that configuration by passing
"--no-log" to the git-merge call.
Signed-off-by: Ralf Thielow <ralf.thielow@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Checking out 2GB or more through an external filter (see test) fails
on Mac OS X 10.8.4 (12E55) for a 64-bit executable with:
error: read from external filter cat failed
error: cannot feed the input to external filter cat
error: cat died of signal 13
error: external filter cat failed 141
error: external filter cat failed
The reason is that read() immediately returns with EINVAL when asked
to read more than 2GB. According to POSIX [1], if the value of
nbyte passed to read() is greater than SSIZE_MAX, the result is
implementation-defined. The write function has the same restriction
[2]. Since OS X still supports running 32-bit executables, the
32-bit limit (SSIZE_MAX = INT_MAX = 2GB - 1) seems to be also
imposed on 64-bit executables under certain conditions. For write,
the problem has been addressed earlier [6c642a].
Address the problem for read() and write() differently, by limiting
size of IO chunks unconditionally on all platforms in xread() and
xwrite(). Large chunks only cause problems, like causing latencies
when killing the process, even if OS X was not buggy. Doing IO in
reasonably sized smaller chunks should have no negative impact on
performance.
The compat wrapper clipped_write() introduced earlier [6c642a] is
not needed anymore. It will be reverted in a separate commit. The
new test catches read and write problems.
Note that 'git add' exits with 0 even if it prints filtering errors
to stderr. The test, therefore, checks stderr. 'git add' should
probably be changed (sometime in another commit) to exit with
nonzero if filtering fails. The test could then be changed to use
test_must_fail.
Thanks to the following people for suggestions and testing:
Johannes Sixt <j6t@kdbg.org>
John Keeping <john@keeping.me.uk>
Jonathan Nieder <jrnieder@gmail.com>
Kyle J. McKay <mackyle@gmail.com>
Linus Torvalds <torvalds@linux-foundation.org>
Torsten Bögershausen <tboegi@web.de>
[1] http://pubs.opengroup.org/onlinepubs/009695399/functions/read.html
[2] http://pubs.opengroup.org/onlinepubs/009695399/functions/write.html
[6c642a] commit 6c642a8786
compate/clipped-write.c: large write(2) fails on Mac OS X/XNU
Signed-off-by: Steffen Prohaska <prohaska@zib.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git fails due to a segmentation fault if a submodule path is empty.
Here is an example .gitmodules that will cause a segmentation fault:
[submodule "foo-module"]
path
url = http://host/repo.git
$ git status
Segmentation fault (core dumped)
This is because the parsing of "submodule.*.path" is not prepared to
see a value-less "true" and assumes that the value is always
non-NULL (parsing of "ignore" has the same problem).
Fix it by checking the NULL-ness of value and complain with
config_error_nonbool().
Signed-off-by: Jharrod LaFon <jlafon@eyesopen.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Newlines in the path to a git repository were not an issue for the
git-specific bash prompt before commit efaa0c1532 (bash prompt:
combine 'git rev-parse' executions in the main code path, 2013-06-17),
because the path returned by 'git rev-parse --git-dir' was directly
stored in a variable, and this variable was later always accessed
inside double quotes.
Newlines are not an issue after commit efaa0c1532 either, but it's
more subtle. Since efaa0c1532 we use the following single 'git
rev-parse' execution to query various info about the repository:
git rev-parse --git-dir --is-inside-git-dir \
--is-bare-repository --is-inside-work-tree
The results to these queries are separated by a newline character in
the output, e.g.:
/home/szeder/src/git/.git
false
false
true
A newline in the path to the git repository could potentially break
the parsing of these results and ultimately the bash prompt, unless
the parsing is done right. Commit efaa0c1532 got it right, as I
consciously started parsing 'git rev-parse's output from the end,
where each record is a single line containing either 'true' or 'false'
or, after e3e0b9378b (bash prompt: combine 'git rev-parse' for
detached head, 2013-06-24), the abbreviated commit object name, and
all what remains at the beginning is the path to the git repository,
no matter how many lines it is.
This subtlety really warrants its own test, especially since I didn't
explain it in the log message or in an in-code comment back then, so
add a test to excercise the prompt with newline characters in the path
to the repository. Guard this test with the FUNNYNAMES prerequisite,
because not all filesystems support newlines in filenames. Note that
'git rev-parse --git-dir' prints '.git' or '.' when at the top of the
worktree or the repository, respectively, and only prints the full
path to the repository when in a subdirectory, hence the need for
changing into a subdir in the test.
Signed-off-by: SZEDER Gábor <szeder@ira.uka.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An earlier draft of the previous step used cache_name_exists() to
check the directory we were looking at, which missed the second case
described in its log message. Demonstrate why it is not sufficient.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit cdfd94837b, as it
does not just apply to "@" (and forms with modifiers like @{u}
applied to it), but also affects e.g. "refs/heads/@/foo", which it
shouldn't.
The basic idea of giving a short-hand might be good, and the topic
can be retried later, but let's revert to avoid affecting existing
use cases for now for the upcoming release.
This reverts commit a73653130e, as it
has been reported that "ls-files --killed" is too time-consuming in
a deep directory with too many untracked crufts (e.g. $HOME/.git
tracking only a few files).
We'd need to revisit it later but "ls-files --killed" needs to be
optimized before it happens.
- From the beginning of push.c in 755225d, 2006-04-29, "thin" option
was enabled by default but could be turned off with --no-thin.
- Then Shawn changed the default to 0 in favor of saving server
resources in a4503a1, 2007-09-09. --no-thin worked great.
- One day later, in 9b28851 Daniel extracted some code from push.c to
create transport.c. He (probably accidentally) flipped the default
value from 0 to 1 in transport_get().
From then on --no-thin is effectively no-op because git-push still
expects the default value to be false and only calls
transport_set_option() when "thin" variable in push.c is true (which
is unnecessary). Correct the code to respect --no-thin by calling
transport_set_option() in both cases.
receive-pack learns about --reject-thin-pack-for-testing option,
which only is for testing purposes, hence no document update.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the tests, p4d is started without using "internationalized
mode". Make sure this environment variable is unset, otherwise
a mis-matched user setting would break the tests. The error
message would be "Unicode clients require a unicode enabled server."
[pw: use unset, add commit text]
Signed-off-by: Kazuki Saitoh <ksaitoh560@gmail.com>
Signed-off-by: Pete Wyckoff <pw@padd.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In t/t7407-submodule-foreach.sh there is a typo in one of the
path names given for a test step. The correct path is
nested1/nested2/.git, but nested1/nested1/nested2/.git is
given instead. The typo is hidden because this line also
accidentally omits the && chain operator. The omitted chain
also means the return values of all the previous commands in
this test are also being ignored.
Fix the path and add the chain operator so the entire test
sequence can be properly validated.
Signed-off-by: Phil Hord <hordp@cisco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is an attempt to reproduce a problem reported for a third-party
custom "connect" remote helper. The conjecture is that sometimes
"git fetch" wants to make two connections (one for the primary
transfer with 'follow-tags' option set, and then after noticing that
some tags are not packed because the primary transfer did not have
to send any commit that is pointed by them, another to explicitly
ask for the missing tags), and their "connect" helper is not called
in the second request, breaking the "fetch" as a whole.
Unfortunately this test script does not trigger the alleged failure
and happily passes when talking to upload-pack from git-core (see
patch 5/5 for details).
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some implementations of 'echo' (e.g. dash's built-in) interpret
backslash sequences in their arguments.
This triggered at least one bug: the error message of "rebase -i" was
turning \t in commit messages into actual tabulations. There may be
others.
Using "printf '%s\n'" instead avoids this bad behavior, and is the form
used by the "say" function.
Noticed-by: David Kastrup <dak@gnu.org>
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since inception, git-blame -L has been documented as accepting 1-based
line numbers. When handed a line number less than 1, -L's behavior is
undocumented and undefined; it's also nonsensical and should be
diagnosed as an error. Do so.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-blame -L is documented as accepting 1-based line numbers. When
handed a line number less than 1, -L's behavior is undocumented and
undefined; it's also nonsensical and should be rejected but is
nevertheless accepted. Demonstrate this shortcoming.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The -L:RE option of blame/log searches from the end of the previous -L
range, if any. Add new notation -L^:RE to override this behavior and
search from start of file.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
For consistency with -L/RE/, teach -L:RE to search relative to the end
of the previous -L range, if any.
The new behavior invalidates one test in t4211 which assumes that -L:RE
begins searching at start of file. This test will be resurrected in a
follow-up patch which teaches -L:RE how to override the default relative
search behavior.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The -L/RE/ option of blame/log searches from the end of the previous -L
range, if any. Add new notation -L^/RE/ to override this behavior and
search from start of file.
The new ^/RE/ syntax is valid only as the <start> argument of
-L<start>,<end>. The <end> argument, as usual, is relative to <start>.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently using "git rm" on a submodule removes the submodule's work tree
from that of the superproject and the gitlink from the index. But the
submodule's section in .gitmodules is left untouched, which is a leftover
of the now removed submodule and might irritate users (as opposed to the
setting in .git/config, this must stay as a reminder that the user showed
interest in this submodule so it will be repopulated later when an older
commit is checked out).
Let "git rm" help the user by not only removing the submodule from the
work tree but by also removing the "submodule.<submodule name>" section
from the .gitmodules file and stage both. This doesn't happen when the
"--cached" option is used, as it would modify the work tree. This also
silently does nothing when no .gitmodules file is found and only issues a
warning when it doesn't have a section for this submodule. This is because
the user might just use plain gitlinks without the .gitmodules file or has
already removed the section by hand before issuing the "git rm" command
(in which case the warning reminds him that rm would have done that for
him). Only when .gitmodules is found and contains merge conflicts the rm
command will fail and tell the user to resolve the conflict before trying
again.
Also extend the man page to inform the user about this new feature. While
at it promote the submodule sub-section to a chapter as it made not much
sense under "REMOVING FILES THAT HAVE DISAPPEARED FROM THE FILESYSTEM".
In t7610 three uses of "git rm submod" had to be replaced with "git rm
--cached submod" because that test expects .gitmodules and the work tree
to stay untouched. Also in t7400 the tests for the remaining settings in
the .gitmodules file had to be changed to assert that these settings are
missing.
Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently using "git mv" on a submodule moves the submodule's work tree in
that of the superproject. But the submodule's path setting in .gitmodules
is left untouched, which is now inconsistent with the work tree and makes
git commands that rely on the proper path -> name mapping (like status and
diff) behave strangely.
Let "git mv" help here by not only moving the submodule's work tree but
also updating the "submodule.<submodule name>.path" setting from the
.gitmodules file and stage both. This doesn't happen when no .gitmodules
file is found and only issues a warning when it doesn't have a section for
this submodule. This is because the user might just use plain gitlinks
without the .gitmodules file or has already updated the path setting by
hand before issuing the "git mv" command (in which case the warning
reminds him that mv would have done that for him). Only when .gitmodules
is found and contains merge conflicts the mv command will fail and tell
the user to resolve the conflict before trying again.
Also extend the man page to inform the user about this new feature.
Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Using the same urlmatch_config_entry() infrastructure, add a new
mode "--get-urlmatch" to the "git config" command, to learn values
for the "virtual" two-level variables customized for the specific
URL.
git config [--<type>] --get-urlmatch <section>[.<key>] <url>
With <section>.<key> fully specified, the configuration data for
<section>.<urlpattern>.<key> for <urlpattern> that best matches the
given <url> is sought (and if not found, <section>.<key> is used)
and reported. For example, with this configuration:
[http]
sslVerify
[http "https://weak.example.com"]
cookieFile = /tmp/cookie.txt
sslVerify = false
You would get
$ git config --bool --get-urlmatch http.sslVerify https://good.example.com
true
$ git config --bool --get-urlmatch http.sslVerify https://weak.example.com
false
With only <section> specified, you can get a list of all variables
in the section with their values that apply to the given URL. E.g
$ git config --get-urlmatch http https://weak.example.com
http.cookiefile /tmp/cookie.txt
http.sslverify false
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use the urlmatch_config_entry() to wrap the underlying
http_options() two-level variable parser in order to set
http.<variable> to the value with the most specific URL in the
configuration.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Empty ranges -L,+0 and -L,-0 are nonsensical in the context of blame yet
they are accepted (in fact, both are interpreted as -L1,Y where Y is
end-of-file). Report them as invalid.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Empty ranges -L,+0 and -L,-0 are nonsensical in the context of blame yet
they are accepted. They should be errors. Demonstrate this shortcoming.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Empty ranges -LX,+0 and -LX,-0 are nonsensical in the context of blame
yet they are accepted (in fact, both are interpreted as -LX,+2). Report
them as invalid.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Empty ranges -LX,+0 and -LX,-0 are nonsensical in the context of blame
yet they are accepted. They should be errors. Demonstrate this
shortcoming.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When 12da1d1f added -L support to git-log, a broken bounds check was
copied from git-blame -L which incorrectly allows -LX to extend one line
past end of file without reporting an error. Instead, it generates an
empty range. Fix this bug.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
58960978 and 99780b0a added tests which demonstrated bugs (crashes) in
range-set and line-log when handed empty ranges specified via "log
-LX:file" where X is one greater than the last line of the file. After
these tests were added, it was realized that the ability to specify an
empty range is a loophole due to a bug in -L bounds checking. That bug
is slated to be fixed in a subsequent patch.
Unfortunately, the closure of this loophole makes it impossible to
continue checking range-set and line-log behavior with regard to empty
ranges since there is no other way to specify empty ranges via the
command-line. APIs of both facilities are private (file static) so
there likewise is no way to test their behaviors programmatically.
Consequently, retire these two tests.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A bounds checking bug allows the X in -LX to extend one line past the
end of file. For example, given a file with 5 lines, -L6 is accepted as
valid. Demonstrate this problem.
While here, also add tests to check that the remaining cases of X and Y
in -LX,Y are handled correctly at and in the vicinity of end-of-file.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since inception, -LX,Y has correctly reported an out-of-range error when
Y is beyond end of file, however, X was not checked, and an out-of-range
X would cause a crash. 92f9e273 (blame: prevent a segv when -L given
start > EOF; 2010-02-08) attempted to rectify this shortcoming but has
its own off-by-one error which allows X to extend one line past end of
file. For example, given a file with 5 lines:
git blame -L5 foo # OK, blames line 5
git blame -L6 foo # accepted, no error, no output, huh?
git blame -L7 foo # error "fatal: file foo has only 5 lines"
Fix this bug.
In order to avoid regressing "blame foo" when foo is an empty file, the
fix is slightly more complicated than changing '<' to '<='.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add boundary case tests, with and without -L, for empty file; file with
one partial line; file with one full line.
The empty file test without -L is of particular interest. Historically,
this case has been supported (empty blame output) and this test protects
against regression by a subsequent patch fixing an off-by-one bug which
incorrectly accepts -LX where X is one past end-of-file.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A bounds checking bug allows the X in -LX to extend one line past the
end of file. For example, given a file with 5 lines, -L6 is accepted as
valid. Demonstrate this problem.
While here, also add tests to check that the remaining cases of X and Y
in -LX,Y are handled correctly at and in the vicinity of end-of-file.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Checking all bogus -L syntax forms in a single test makes it difficult
to identify the offender when one case fails. Decompose this
conglomerate test in order to check each bad syntax case separately.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The URL included in the header appears to vary from curl version to
curl version. Since we only care about the final few lines, only test
them. However, make sure the blank line after the header is still
included to make sure there are no extra cookie lines.
Signed-off-by: Brian Gernhardt <brian@gernhardtsoftware.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Sub-test 42 of t8001 and t8002 ("blame -L :literal") fails on NetBSD
with the following verbose output:
git annotate -L:main hello.c
Author F (expected 4, attributed 3) bad
Author G (expected 1, attributed 1) good
This is not caused by different behaviour of git blame or annotate on
that platform, but by different test input, in turn caused by a sed
command that forgets to add a newline on NetBSD. Here's the diff of the
commit that adds "goodbye" to hello.c, for Linux:
@@ -1,4 +1,5 @@
int main(int argc, const char *argv[])
{
puts("hello");
+ puts("goodbye");
}
We see that it adds an extra TAB, but that's not a problem. Here's the
same on NetBSD:
@@ -1,4 +1,4 @@
int main(int argc, const char *argv[])
{
puts("hello");
-}
+ puts("goodbye");}
It also adds an extra TAB, but it is missing the newline character
after the semicolon.
The following patch gets rid of the extra TAB at the beginning, but
more importantly adds the missing newline at the end in a (hopefully)
portable way, mentioned in http://sed.sourceforge.net/sedfaq4.html.
The diff becomes this, on both Linux and NetBSD:
@@ -1,4 +1,5 @@
int main(int argc, const char *argv[])
{
puts("hello");
+ puts("goodbye");
}
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It is not like that our longer term desire is to someday start
accept log messages with NULs in them, so it is wrong to mark a test
that demonstrates "git commit" that correctly fails given such an
input as "expect-failure". "git commit" should fail today, and it
should fail the same way in the future given a message with NUL in it.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The test file that the UTF-16 rejection test looks for is missing, but this went
unnoticed because the test is expected to fail anyway; as a consequence, the
test fails because the file containing the commit message is missing, and not
because the test file contains a NUL byte. Fix this by including a sample text
file containing a commit message encoded in UTF-16.
Signed-off-by: Brian M. Carlson <sandals@crustytoothpaste.net>
Tested-by: Duy Nguyen <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit c334b87b (cat-file: split --batch input lines on whitespace,
2013-07-11) taught `cat-file --batch-check` to split input lines on
the first whitespace, and stash everything after the first token
into the %(rest) output format element. It claimed:
Object names cannot contain spaces, so any input with
spaces would have resulted in a "missing" line.
But that is not correct. Refs, object sha1s, and various peeling
suffixes cannot contain spaces, but some object names can. In
particular:
1. Tree paths like "[<tree>]:path with whitespace"
2. Reflog specifications like "@{2 days ago}"
3. Commit searches like "rev^{/grep me}" or ":/grep me"
To remain backwards compatible, we cannot split on whitespace by
default, hence we will ship 1.8.4 with the commit reverted.
Resurrect its attempt but in a weaker form; only do the splitting
when "%(rest)" is used in the output format. Since that element did
not exist at all before c334b87, old scripts cannot be affected.
The existence of object names with spaces does mean that you
cannot reliably do:
echo ":path with space and other data" |
git cat-file --batch-check="%(objectname) %(rest)"
as it would split the path and feed only ":path" to get_sha1. But
that command is nonsensical. If you wanted to see "and other data"
in "%(rest)", git cannot possibly know where the filename ends and
the "rest" begins.
It might be more robust to have something like "-z" to separate the
input elements. But this patch is still a reasonable step before
having that. It makes the easy cases easy; people who do not care
about %(rest) do not have to consider it, and the %(rest) code
handles the spaces and newlines of "rev-list --objects" correctly.
Hard cases remain hard but possible (if you might get whitespace in
your input, you do not get to use %(rest) and must split and join
the output yourself using more flexible tools). And most
importantly, it does not preclude us from having different splitting
rules later if a "-z" (or similar) option is added. So we can make
the hard cases easier later, if we choose to.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The reflog walking logic (git log -g) replaces the true parent list
with the preceding commit in the reflog. This results in bogus commit
diffs when combined with options such as -p; the diff is against the
reflog predecessor, not the parent of the commit.
Save the true parents on the side, extending the functions from the
previous commit. The diff logic picks them up and uses them to show
the correct diffs.
We do have to be somewhat careful about repeated calling of
save_parents(), since the reflog may list a commit more than once. We
now store (commit_list*)-1 to distinguish the "not saved yet" and
"root commit" cases. This lets us preserve an empty parent list even
if save_parents() is repeatedly called.
Suggested-by: Jeff King <peff@peff.net>
Signed-off-by: Thomas Rast <trast@inf.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We have been passing enough information to enable the
compare-and-swap logic down to the transport layer, but the
transport helper was not passing it to smart-http transport.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit c334b87b30c1464a1ab563fe1fb8de5eaf0e5bac; the
update assumed that people only used the command to read from
"rev-list --objects" output, whose lines begin with a 40-hex object
name followed by a whitespace, but it turns out that scripts feed
random extended SHA-1 expressions (e.g. "HEAD:$pathname") in which
a whitespace has to be kept.
