The diff code represents paths using the diff_filespec
struct. This struct has a sha1 to represent the sha1 of the
content at that path, as well as a sha1_valid member which
indicates whether its sha1 field is actually useful. If
sha1_valid is not true, then the filespec represents a
working tree file (e.g., for the no-index case, or for when
the index is not up-to-date).
The diff_filespec is only used internally, though. At the
interfaces to the diff subsystem, callers feed the sha1
directly, and we create a diff_filespec from it. It's at
that point that we look at the sha1 and decide whether it is
valid or not; callers may pass the null sha1 as a sentinel
value to indicate that it is not.
We should not typically see the null sha1 coming from any
other source (e.g., in the index itself, or from a tree).
However, a corrupt tree might have a null sha1, which would
cause "diff --patch" to accidentally diff the working tree
version of a file instead of treating it as a blob.
This patch extends the edges of the diff interface to accept
a "sha1_valid" flag whenever we accept a sha1, and to use
that flag when creating a filespec. In some cases, this
means passing the flag through several layers, making the
code change larger than would be desirable.
One alternative would be to simply die() upon seeing
corrupted trees with null sha1s. However, this fix more
directly addresses the problem (while bogus sha1s in a tree
are probably a bad thing, it is really the sentinel
confusion sending us down the wrong code path that is what
makes it devastating). And it means that git is more capable
of examining and debugging these corrupted trees. For
example, you can still "diff --raw" such a tree to find out
when the bogus entry was introduced; you just cannot do a
"--patch" diff (just as you could not with any other
corrupted tree, as we do not have any content to diff).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we are leaving a detached HEAD, we do a revision traversal to
check whether we are orphaning any commits, marking the commit we're
leaving as the start of the traversal, and all existing refs as
uninteresting.
Prior to commit 468224e5, we did so by calling for_each_ref, and
feeding each resulting refname to setup_revisions. Commit 468224e5
refactored this to simply mark the pending objects, saving an extra
lookup.
However, it confused the "flags" parameter to the each_ref_fn
clalback, which is about the flags we found while looking up the ref
with the object flag. Because REF_ISSYMREF ("this ref is a symbolic
ref, e.g. refs/remotes/origin/HEAD") happens to be the same bit
pattern as SEEN ("we have picked this object up from the pending
list and moved it to revs.commits list"), we incorrectly reported
that a commit previously at the detached HEAD will become
unreachable if the only ref that can reach the commit happens to be
pointed at by a symbolic ref.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The identity of the committer will ultimately be pulled from
the ident code by commit_tree(). However, we make an attempt
to check the author and committer identity early, before the
user has done any manual work like inputting a commit
message. That lets us abort without them having to worry
about salvaging the work from .git/COMMIT_EDITMSG.
The early check for committer ident does not use the
IDENT_STRICT flag, meaning that it would not find an empty
name field. The motivation was presumably because we did not
want to be too restrictive, as later calls might be more lax
(for example, when we create the reflog entry, we do not
care too much about a real name). However, because
commit_tree will always get a strict identity to put in the
commit object itself, there is no point in being lax only to
die later (and in fact it is harmful, because the user will
have wasted time typing their commit message).
Incidentally, this bug was masked prior to 060d4bb, as the
initial loose call would taint the later strict call. So the
commit would succeed (albeit with a bogus committer line in
the commit object), and nobody noticed that our early check
did not match the later one.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
79a9312 (commit-tree: update the command line parsing, 2011-11-09)
updated the command line parser to understand the usual "flags first
and then non-flag arguments" order, in addition to the original and
a bit unusual "tree comes first and then zero or more -p <parent>".
Unfortunately, ba3c69a (commit: teach --gpg-sign option, 2011-10-05)
broke it by mistake. Resurrect it, and protect the feature with a
test from future breakages.
Noticed by Keshav Kini
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We should be letting the user's umask take care of
restricting permissions. Even though this is a temporary
file and probably nobody would notice, this brings us in
line with other temporary file creations in git (e.g.,
choosing "e"dit from git-add--interactive).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The original version of the git-clone just used mkdir(1) to create
the working directories. The version rewritten in C creates all
directories inside the working tree by using the mode argument of
0777 when calling mkdir(2) to let the umask take effect.
But the top-level directory of the working tree is created by
passing the mode argument of 0755 to mkdir(2), which results in an
overly tight restriction if the user wants to make directories group
writable with a looser umask like 002.
Signed-off-by: Alex Riesen <raa.lkml@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git commit --amend" used on a commit with an empty message fails
unless -m is given, whether or not --allow-empty-message is
specified.
Allow it to proceed to the editor with an empty commit message.
Unless --allow-empty-message is in force, it will still abort later
if an empty message is saved from the editor (this check was
already necessary to prevent a non-empty commit message being edited
to an empty one).
Add a test for --amend --edit of an empty commit message which fails
without this fix, as it's a rare case that won't get frequently
tested otherwise.
Signed-off-by: Chris Webb <chris@arachsys.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Julia Lawall noticed that in linux-next repository the commit object
60d5c9f5 (shown with the default abbreviation width baked into "git
blame") in output from
$ git blame -L 3675,3675 60d5c9f5b -- \
drivers/staging/brcm80211/brcmfmac/wl_iw.c
is no longer unique in the repository, which results in "short SHA1
60d5c9f5 is ambiguous".
Compute the minimum abbreviation width that ensures uniqueness when
the user did not specify the --abbrev option to avoid this.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A path containing a space must be quoted when used as an
argument to either the copy or rename commands (because
unlike other commands, the path is not the final thing on
the line for those commands).
Commit 6280dfdc3b (fast-export: quote paths in output,
2011-08-05) previously attempted to fix fast-export's
quoting by passing all paths through quote_c_style().
However, that function does not consider the space to be a
character which requires quoting, so let's special-case the
space inside print_path(). This will cause space-containing
paths to also be quoted in other commands where such quoting
is not strictly necessary, but it does not hurt to do so.
The test from 6280dfdc3b did not detect this because, while
it does introduce renames in the export stream, it does not
actually turn on rename detection, so they were presented as
pairs of deletions/adds. Using "-M" reveals the bug.
Signed-off-by: Jay Soffian <jaysoffian@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The Cygwin implementation of pread() is not thread-safe since, just
like the emulation provided by compat/pread.c, it uses a sequence of
seek-read-seek calls. In order to avoid failues due to thread-safety
issues, commit b038a61 disables threading when NO_PREAD is defined.
(ie when using the emulation code in compat/pread.c).
We introduce a new build variable, NO_THREAD_SAFE_PREAD, which allows
use to disable the threaded index-pack code on cygwin, in addition to
the above NO_PREAD case.
Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
abe199808c (git checkout -b: allow switching out of an unborn branch)
introduced a bug demonstrated by
git checkout --orphan foo
git checkout --detach
git symbolic-ref HEAD
which gives 'refs/heads/(null)'.
This happens because we strbuf_addf(&branch_ref, "refs/heads/%s",
opts->new_branch) when opts->new_branch can be NULL for --detach.
Catch and forbid this case, adding a test to t2017 to catch it in
future.
Signed-off-by: Chris Webb <chris@arachsys.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
- do not fetch HEAD
- do not also fetch refs following "xxx"
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
verify_filename() can be called in two different contexts. Either we
just tried to interpret a string as an object name, and it fails, so
we try looking for a working tree file (i.e. we finished looking at
revs that come earlier on the command line, and the next argument
must be a pathname), or we _know_ that we are looking for a
pathname, and shouldn't even try interpreting the string as an
object name.
For example, with this change, we get:
$ git log COPYING HEAD:inexistant
fatal: HEAD:inexistant: no such path in the working tree.
Use '-- <path>...' to specify paths that do not exist locally.
$ git log HEAD:inexistant
fatal: Path 'inexistant' does not exist in 'HEAD'
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-diff does not rely on the git wrapper to setup its
pager; instead, it sets it up on its own after seeing
whether --quiet or --exit-code has been specified. After
diff_no_index was split off from cmd_diff, commit b3fde6c
(git diff --no-index: default to page like other diff
frontends, 2008-05-26) duplicated the one-liner from
cmd_diff to turn on the pager.
Later, commit 8f0359f (Allow pager of diff command be
enabled/disabled, 2008-07-21) taught the the version in
cmd_diff to respect the pager.diff config, but the version
in diff_no_index was left behind. This meant that
git -c pager.diff=0 diff a b
would not use a pager, but
git -c pager.diff=0 diff --no-index a b
would. Let's fix it by factoring out a common function.
While we're there, let's update the antiquated comment,
which claims that the pager interferes with propagating the
exit code; this has not been the case since ea27a18 (spawn
pager via run_command interface, 2008-07-22).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In commit e01105 Linus introduced gitlinks to update-index. He explains
that he thinks it is not the right thing to replace a gitlink with
something else.
That commit is from the very first beginnings of submodule support.
Since then we have gotten a lot closer to being able to remove a
submodule without losing its history. This check prevents such a use
case, so I think this assumption has changed.
Additionally in the git add codepath we do not have such a check, so for
consistency reasons I think removing this check is the correct thing to
do.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The submaintainer credit is not something you can compute purely by
looking at the history and its shape, especially in the presense of
fast-forward merges, and this observation makes the information on
the "via" line unreliable. Let's leave the final determination of
credits up to whoever is making the merge and show them as comments.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This only happens in --ignore-missing --dry-run codepath which
presumably nobody should care, but is for completeness.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It was stupid of me to make the API too much cache-entry specific;
the caller may want to check arbitrary pathname without having a
corresponding cache-entry to see if a path is ignored.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git ls-files --exclude=t/ -i" does not show paths in directory t/
that have been added to the index, but it should.
The excluded() API was designed for callers who walk the tree from
the top, checking each level of the directory hierarchy as it
descends if it is excluded, and not even bothering to recurse into
an excluded directory. This would allow us optimize for a common
case by not having to check if the exclude pattern "foo/" matches
when looking at "foo/bar", because the caller should have noticed
that "foo" is excluded and did not even bother to read "foo/bar"
out of opendir()/readdir() to call it.
The code for "ls-files -i" however walks the index linearly, feeding
paths without checking if the leading directory is already excluded.
Introduce a helper function path_excluded() to let this caller
properly call excluded() check for higher hierarchies as necessary.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 4th arg of "new mode (%o) of %s does not match old mode (%o)%s%s"
is blank string or string " of ". Even mark the string " of " for a
complete i18n, this message is still hard to translate right.
Split it into two slight different messages would make l10n teams happy.
Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When adding the information from a tag, put an empty line between the
message of the tag and the commented-out signature verification
information.
At least for the kernel workflow, I often end up re-formatting the message
that people send me in the tag data. In that situation, putting the tag
message and the tag signature verification back-to-back then means that
normal editor "reflow parapgraph" command will get confused and think that
the signature is a continuation of the last message paragraph.
So I always end up having to first add an empty line, and then go back and
reflow the last paragraph. Let's just do it in git directly.
The extra vertical space also makes the verification visually stand out
more from the user-supplied message, so it looks a bit more readable to me
too, but that may be just an odd personal preference.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We can ask git_committer_info to be strict about coming up
with an email, which will die automatically on a poorly
configured machine. This is better than letting invalid
message-ids into the wild.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Callers who ask for ERROR_ON_NO_NAME are not so much
concerned that the name will be blank (because, after all,
we will fall back to using the username), but rather it is a
check to make sure that low-quality identities do not end up
in things like commit messages or emails (whereas it is OK
for them to end up in things like reflogs).
When future commits add more quality checks on the identity,
each of these callers would want to use those checks, too.
Rather than modify each of them later to add a new flag,
let's refactor the flag.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before commit 43ae9f4, we generated the tail of a message id
by calling git_committer_info and parsing the email out of
the result. 43ae9f4 changed to use ident_default_email
directly, so we didn't have to bother with parsing. As a
side effect, it meant we no longer used GIT_COMMITTER_EMAIL
at all.
In general, this is probably reasonable behavior. Either the
default email is sane on your system, or you are using
user.email to provide something sane. The exception is if
you rely on GIT_COMMITTER_EMAIL being set all the time to
override the bogus generated email.
This is unlikely to match anybody's real-life setup, but we
do use it in the test environment. And furthermore, it's
what we have always done, and the change in 43ae9f4 was
about cleaning up, not fixing any bug; we should be
conservative and keep the behavior identical.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If stderr isn't a tty, we shouldn't be printing incremental progress
messages. In particular, this affects 'git checkout -f . >&logfile'
unless you provided -q. And git-new-workdir has no way to provide -q.
It would probably be better to have progress.c check isatty(2) all the time,
but that wouldn't allow things like 'git push --progress' to force progress
reporting to on, so I won't try to solve the general case right now.
Actual fix suggested by Jeff King.
Signed-off-by: Avery Pennarun <apenwarr@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 4435968 started sorting heads fed to fetch-pack so
that later commits could use more optimized algorithms;
commit 7db8d53 switched the remove_duplicates function to
such an algorithm.
Of course, the sorting is more effective if you do it
_before_ the algorithm in question.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We have a list of refs that we want to compare against the
"match" array. The current code searches the match list
linearly, giving quadratic behavior over the number of refs
when you want to fetch all of them.
Instead, we can compare the lists as we go, giving us linear
behavior.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Having the list sorted means we can avoid some quadratic
algorithms when comparing lists.
These should typically be sorted already, but they do come
from the remote, so let's be extra careful. Our ref-sorting
implementation does a mergesort, so we do not have to care
about performance degrading in the common case that the list
is already sorted.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We remove duplicate entries from the list of refs we are
fed in fetch-pack. The original algorithm is quadratic over
the number of refs, but since the list is now guaranteed to
be sorted, we can do it in linear time.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There's no reason to preserve the incoming order of the
heads we're requested to fetch. By having them sorted, we
can replace some of the quadratic algorithms with linear
ones.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The old code cast away the constness of the strings passed to the
function in argument argv[], which could result in their being
modified by filter_refs(). Fix by copying reference names from argv
and putting them into our own array (similarly to how refnames passed
to stdin were already handled).
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If an argument that does not start with '-' is found, the loop is
terminated. So move that check into the for-loop condition.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This makes it more obvious that the code is always executed unless
there is an error, and that the first initialization of nr_heads is
unnecessary.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is no need for it to be non-const, and this avoids the need
for casting away the constness of an argv element.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The get_patch_filename function expects a commit argument
and uses it to get the sanitized subject line when making a
patch filename. However, we also want to use this same
function for the cover letter, which does not have a commit
object. The current solution is to create a fake commit with
the subject "cover letter". Instead, let's make the
get_patch_filename interface more flexibile, and allow
passing a direct subject.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We try to generate a sane message id for cover letters and
threading by appending some changing bits to the front of
the user's email address. The current code parses the email
out of the results of git_committer_info, but we can do this
much more easily by just calling ident_default_email
ourselves.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The record_person function just parses out the "name" field
of the person line in a commit and adds it to a string_list.
The only reason we need an extra buffer is that the
string_list functions require a NUL-terminated string.
Instead of the static buffer, we can just allocate a
temporary NUL-terminated copy. In addition to removing a
useless limit, this removes the only user of MAX_GITNAME
outside of ident.c.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When reading patterns from a file, we pass the lines as allocated string
buffers to append_grep_pat() and never free them. That's not a problem
because they are needed until the program ends anyway.
However, now that the function duplicates the pattern string, we can
reuse the strbuf after calling that function. This simplifies the code
a bit and plugs a minor memory leak.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The mapping that describe what ref fetched from the remote is used to
update what ref locally is called "refspec", not "respec".
Signed-off-by: Ralf Thielow <ralf.thielow@googlemail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function first decides if we want to copy data taken from existing
pack verbatim or we want to encode the data ourselves for the packfile
we are creating and then carries out the decision. Separate the latter
phase into two helper functions, one for the case the data is reused,
the other for the case the data is produced anew.
A little twist is that it can later turn out that we cannot reuse the
data after we initially decide to do so; in such a case, the "reuse"
helper makes a call to "generate" helper. It is easier to follow than
the current fallback code that uses "goto" inside a single large
function.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is because all other places do "xx > big_file_threshold"
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Plain gcc may not but sparse catches and complains about this sort of
stuff.
Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use handle_split_cb() directly as hunk_func() callback, without going
through xdi_diff_hunks().
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use blame_chunk_cb() directly as hunk_func() callback, without detour
through xdi_diff_hunks().
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In abe1998 ("git checkout -b: allow switching out of an unborn
branch"), a code-path overly-optimisticly assumed that a
branch-name was specified. This is not always the case, and as
a result a NULL-pointer was attempted printed to .git/HEAD.
This could lead to at least two different failure modes:
1) vsnprintf formated the NULL-string as something useful (e.g
"(null)")
2) vsnprintf crashed
Neither were very convenient for formatting a new HEAD-reference.
To fix this, reintroduce some strictness so we only take this
new codepath if a banch-name was specified.
Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It marks the string "...inconsistent %s filename..." where %s is either
"old" or "new" from caller. Make it two strings "...inconsistent new
filename..." and "...inconsistent old filename...".
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The current code reads the config and command-line options
into a separate "colopts" variable, and then copies the
contents of that variable into the "struct wt_status". We
can eliminate the extra variable and copy just write
straight into the wt_status struct.
This simplifies the "status" code a little bit.
Unfortunately, it makes the "commit" code one line more
complex; a side effect of the separate variable was that
"commit" did not copy the colopts variable, so any
column.status configuration had no effect.
The result still ends up cleaner, though. In the previous
version, it was unclear whether commit simply forgot to copy
the colopt variable, or whether it was intentional. Now it
explicitly turns off column options. Furthermore, if commit
later learns to respect column.status, this will make the
end result simpler. I punted on just adding that feature
now, because it was sufficiently non-obvious that it should
not go into a refactoring patch.
Signed-off-by: Jeff King <peff@peff.net>
Jeff King