* tr/cache-tree:
t0090: be prepared that 'wc -l' writes leading blanks
reset: update cache-tree data when appropriate
commit: write cache-tree data when writing index anyway
Refactor cache_tree_update idiom from commit
Test the current state of the cache-tree optimization
Add test-scrap-cache-tree
* cb/maint-t5541-make-server-port-portable:
t5541: check error message against the real port number used
remote-curl: Fix push status report when all branches fail
* tr/maint-bundle-boundary:
bundle: keep around names passed to add_pending_object()
t5510: ensure we stay in the toplevel test dir
t5510: refactor bundle->pack conversion
* tr/maint-bundle-long-subject:
t5704: match tests to modern style
strbuf: improve strbuf_get*line documentation
bundle: use a strbuf to scan the log for boundary commits
bundle: put strbuf_readline_fd in strbuf.c with adjustments
The 'name' field passed to add_pending_object() is used to later
deduplicate in object_array_remove_duplicates().
git-bundle had a bug in this area since 18449ab (git-bundle: avoid
packing objects which are in the prerequisites, 2007-03-08): it passed
the name of each boundary object in a static buffer. In other words,
all that object_array_remove_duplicates() saw was the name of the
*last* added boundary object.
The recent switch to a strbuf in bc2fed4 (bundle: use a strbuf to scan
the log for boundary commits, 2012-02-22) made this slightly worse: we
now free the buffer at the end, so it is not even guaranteed that it
still points into addressable memory by the time object_array_remove_
duplicates looks at it. On the plus side however, it was now
detectable by valgrind.
The fix is easy: pass a copy of the string to add_pending_object.
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The last test descended into a subdir without ever re-emerging, which
is not so nice to the next test writer.
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's not so much a conversion as a "strip everything up to and
including the first blank line", but it will come in handy again.
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-am.sh's check_patch_format function would attempt to preview
the patch to guess its format, but would go into an infinite loop
when the patch file happened to be empty. The solution: exit the
loop when "read" fails, not when the line var, "$l1" becomes empty.
Signed-off-by: Jim Meyering <meyering@redhat.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The test did not adhere to the current style on several counts:
. empty lines around the test blocks, but within the test string
. ': > file' or even just '> file' with an extra space
. inconsistent indentation
. hand-rolled commits instead of using test_commit
Fix all of them.
There's a catch to the last point: test_commit creates a tag, which the
original test did not create. We still change it to test_commit, and
explicitly delete the tags, so as to highlight that the test relies on not
having them.
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The first part of the bundle header contains the boundary commits, and
could be approximated by
# v2 git bundle
$(git rev-list --pretty=oneline --boundary <ARGS> | grep ^-)
git-bundle actually spawns exactly this rev-list invocation, and does
the grepping internally.
There was a subtle bug in the latter step: it used fgets() with a
1024-byte buffer. If the user has sufficiently long subjects (e.g.,
by not adhering to the git oneline-subject convention in the first
place), the 'oneline' format can easily overflow the buffer. fgets()
then returns the rest of the line in the next call(s). If one of
these remaining parts started with '-', git-bundle would mistakenly
insert it into the bundle thinking it was a boundary commit.
Fix it by using strbuf_getwholeline() instead, which handles arbitrary
line lengths correctly.
Note that on the receiving side in parse_bundle_header() we were
already using strbuf_getwholeline_fd(), so that part is safe.
Reported-by: Jannis Pohlmann <jannis.pohlmann@codethink.co.uk>
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Otherwise the test cannot be run with custom port set to LIB_HTTPD_PORT.
Signed-off-by: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Starting from GNU Make 3.82 $(wildcard ...) no longer sorts the result
(from NEWS):
* WARNING: Backward-incompatibility!
Wildcards were not documented as returning sorted values, but the results
have been sorted up until this release.. If your makefiles require sorted
results from wildcard expansions, use the $(sort ...) function to request
it explicitly.
2a59dc32aa
I usually watch test progress visually, and if tests are sorted, even
with make -j4 they go more or less incrementally by their t number. On
the other side, without sorting, tests are executed in seemingly random
order even for -j1. Let's please maintain sane tests order for perceived
prettyness.
Another note is that in GNU Make sort also works as uniq, so after sort
being removed, we might expect e.g. $(wildcard *.sh a.*) to produce
duplicates for e.g. "a.sh". From this point of view, adding sort could
be seen as hardening t/Makefile from accidentally introduced dups.
It turned out that prevous releases of GNU Make did not perform full
sort in $(wildcard), only sorting results for each pattern, that's why
explicit sort-as-uniq is relevant even for older makes.
Signed-off-by: Kirill Smelkov <kirr@navytux.spb.ru>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The protocol between transport-helper.c and remote-curl requires
remote-curl to always print a blank line after the push command
has run. If the blank line is ommitted, transport-helper kills its
container process (the git push the user started) with exit(128)
and no message indicating a problem, assuming the helper already
printed reasonable error text to the console.
However if the remote rejects all branches with "ng" commands in the
report-status reply, send-pack terminates with non-zero status, and
in turn remote-curl exited with non-zero status before outputting
the blank line after the helper status printed by send-pack. No
error messages reach the user.
This caused users to see the following from git push over HTTP
when the remote side's update hook rejected the branch:
$ git push http://... master
Counting objects: 4, done.
Delta compression using up to 6 threads.
Compressing objects: 100% (2/2), done.
Writing objects: 100% (3/3), 301 bytes, done.
Total 3 (delta 0), reused 0 (delta 0)
$
Always print a blank line after the send-pack process terminates,
ensuring the helper status report (if it was output) will be
correctly parsed by the calling transport-helper.c. This ensures
the helper doesn't abort before the status report can be shown to
the user.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The pathspec structure has a few bits of data to drive various operation
modes after we unified the pathspec matching logic in various codepaths.
For example, max_depth field is there so that "git grep" can limit the
output for files found in limited depth of tree traversal. Also in order
to show just the surface level differences in "git diff-tree", recursive
field stops us from descending into deeper level of the tree structure
when it is set to false, and this also affects pathspec matching when
we have wildcards in the pathspec.
The diff-index has always wanted the recursive behaviour, and wanted to
match pathspecs without any depth limit. But we forgot to do so when we
updated tree_entry_interesting() logic to unify the pathspec matching
logic.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add commit message to avoid commit's aborting due to the lack of
commit message, not because there are INTENT_TO_ADD entries in index.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we prepare the attribute stack for a lookup on a path,
we start with the cached stack from the previous lookup
(because it is common to do several lookups in the same
directory hierarchy). So the first thing we must do in
preparing the stack is to pop any entries that point to
directories we are no longer interested in.
For example, if our stack contains gitattributes for:
foo/bar/baz
foo/bar
foo
but we want to do a lookup in "foo/bar/bleep", then we want
to pop the top element, but retain the others.
To do this we walk down the stack from the top, popping
elements that do not match our lookup directory. However,
the test do this simply checked strncmp, meaning we would
mistake "foo/bar/baz" as a leading directory of
"foo/bar/baz_plus". We must also check that the character
after our match is '/', meaning we matched the whole path
component.
There are two special cases to consider:
1. The top of our attr stack has the empty path. So we
must not check for '/', but rather special-case the
empty path, which always matches.
2. Typically when matching paths in this way, you would
also need to check for a full string match (i.e., the
character after is '\0'). We don't need to do so in
this case, though, because our path string is actually
just the directory component of the path to a file
(i.e., we know that it terminates with "/", because the
filename comes after that).
Helped-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When gitweb URL does not provide action explicitly, e.g.
http://git.example.org/repo.git/branch
dispatch() tries to guess action (view to be used) based on remaining
parameters. Among others it is based on the type of requested object,
which gave problems when asking for non-existent branch or file (for
example misspelt name).
Now undefined $action from dispatch() should not result in problems.
Signed-off-by: Jakub Narebski <jnareb@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Subsequently we assume that there is only one pack. Currently this is
true only by accident. Pass '-a -d' to repack in order to guarantee that
assumption to hold true.
The prune-packed command is now redundant since repack -d already calls
it.
Signed-off-by: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-repo if interupted at the exact wrong time will generate zero
length bundles- literal empty files. git-repo is wrong here, but
git fetch shouldn't effectively spin loop if pointed at a zero
length bundle.
Signed-off-by: Brian Harring <ferringb@chromium.org>
Helped-by: Johannes Sixt
Helped-by: Nguyen Thai Ngoc Duy
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/checkout-m-twoway:
t/t2023-checkout-m.sh: fix use of test_must_fail
checkout_merged(): squelch false warning from some gcc
Test 'checkout -m -- path'
checkout -m: no need to insist on having all 3 stages
* jn/maint-sequencer-fixes:
revert: stop creating and removing sequencer-old directory
Revert "reset: Make reset remove the sequencer state"
revert: do not remove state until sequence is finished
revert: allow single-pick in the middle of cherry-pick sequence
revert: pass around rev-list args in already-parsed form
revert: allow cherry-pick --continue to commit before resuming
revert: give --continue handling its own function
* jk/fetch-no-tail-match-refs:
connect.c: drop path_match function
fetch-pack: match refs exactly
t5500: give fully-qualified refs to fetch-pack
drop "match" parameter from get_remote_heads
Change an invocation of test_must_fail() to be inside a
test_expect_success() as is our usual pattern. Having it outside
caused our tests to fail under prove(1) since we wouldn't print a
newline before TAP output:
CONFLICT (content): Merge conflict in both.txt
# GETTEXT POISON #ok 2 - -m restores 2-way conflicted+resolved file
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use 'printf %d $(whatever|wc -l)' so that the shell removes the blanks
for us.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Acked-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When receive-pack advertises its list of refs, it generally hides the
capabilities information after a NUL at the end of the first ref.
However, when we have an empty repository, there are no refs, and
therefore receive-pack writes a fake ref "capabilities^{}" with the
capabilities afterwards.
On the client side, git reads the result with get_remote_heads(). We pick
the capabilities from the end of the line, and then call check_ref() to
make sure the ref name is valid. We see that it isn't, and don't bother
adding it to our list of refs.
However, the call to check_ref() is enabled by passing the REF_NORMAL flag
to get_remote_heads. For the regular git transport, we pass REF_NORMAL in
get_refs_via_connect() if we are doing a push (since only receive-pack
uses this fake ref). But in remote-curl, we never use this flag, and we
accept the fake ref as a real one, passing it back from the helper to the
parent git-push.
Most of the time this bug goes unnoticed, as the fake ref won't match our
refspecs. However, if "--mirror" is used, then we see it as remote cruft
to be pruned, and try to pass along a deletion refspec for it. Of course
this refspec has bogus syntax (because of the ^{}), and the helper
complains, aborting the push.
Let's have remote-curl mirror what the builtin get_refs_via_connect() does
(at least for the case of using git protocol; we can leave the dumb
info/refs reader as it is).
This also fixes pushing with --mirror to a smart-http remote that uses
alternates. The fake ".have" refs the server gives to avoid unnecessary
network transfer has a similar bad interactions with the machinery.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before commit 986bbc08, git was proactive about asking for
http passwords. It assumed that if you had a username in
your URL, you would also want a password, and asked for it
before making any http requests.
However, this could interfere with the use of .netrc (see
986bbc08 for details). And it was also unnecessary, since
the http fetching code had learned to recognize an HTTP 401
and prompt the user then. Furthermore, the proactive prompt
could interfere with the usage of .netrc (see 986bbc08 for
details).
Unfortunately, the http push-over-DAV code never learned to
recognize HTTP 401, and so was broken by this change. This
patch does a quick fix of re-enabling the "proactive auth"
strategy only for http-push, leaving the dumb http fetch and
smart-http as-is.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We don't currently test this case at all, and instead just
test the DAV mechanism over an unauthenticated push. That
isn't very realistic, as most people will want to
authenticate pushes.
Two of the tests expect_failure as they reveal bugs:
1. Pushing without a username in the URL fails to ask for
credentials when we get an HTTP 401. This has always
been the case, but it would be nice if it worked like
smart-http.
2. Pushing with a username fails to ask for the password
since 986bbc0 (http: don't always prompt for password,
2011-11-04). This is a severe regression in v1.7.8, as
authenticated push-over-DAV is now totally unusable
unless you have credentials in your .netrc.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we are determining the list of refs to fetch via
fetch-pack, we have two sets of refs to compare: those on
the remote side, and a "match" list of things we want to
fetch. We iterate through the remote refs alphabetically,
seeing if each one is wanted by the "match" list.
Since def88e9 (Commit first cut at "git-fetch-pack",
2005-07-04), we have used the "path_match" function to do a
suffix match, where a remote ref is considered wanted if
any of the "match" elements is a suffix of the remote
refname.
This enables callers of fetch-pack to specify unqualified
refs and have them matched up with remote refs (e.g., ask
for "A" and get remote's "refs/heads/A"). However, if you
provide a fully qualified ref, then there are corner cases
where we provide the wrong answer. For example, given a
remote with two refs:
refs/foo/refs/heads/master
refs/heads/master
asking for "refs/heads/master" will first match
"refs/foo/refs/heads/master" by the suffix rule, and we will
erroneously fetch it instead of refs/heads/master.
As it turns out, all callers of fetch_pack do provide
fully-qualified refs for the match list. There are two ways
fetch_pack can get match lists:
1. Through the transport code (i.e., via git-fetch)
2. On the command-line of git-fetch-pack
In the first case, we will always be providing the names of
fully-qualified refs from "struct ref" objects. We will have
pre-matched those ref objects already (since we have to
handle more advanced matching, like wildcard refspecs), and
are just providing a list of the refs whose objects we need.
In the second case, users could in theory be providing
non-qualified refs on the command-line. However, the
fetch-pack documentation claims that refs should be fully
qualified (and has always done so since it was written in
2005).
Let's change this path_match call to simply check for string
equality, matching what the callers of fetch_pack are
expecting.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The fetch-pack documentation is very clear that refs given
on the command line are to be full refs:
<refs>...::
The remote heads to update from. This is relative to
$GIT_DIR (e.g. "HEAD", "refs/heads/master"). When
unspecified, update from all heads the remote side has.
and this has been the case since fetch-pack was originally documented in
8b3d9dc ([PATCH] Documentation: clone/fetch/upload., 2005-07-14).
Let's follow our own documentation to set a good example,
and to avoid breaking when this restriction is enforced in
the next patch.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit 95eb88d8ee, which
was a UI experiment that did not reflect how "git reset" actually gets
used. The reversion also fixes a test, indicated in the patch.
Encouraged-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As v1.7.8-rc0~141^2~4 (2011-08-04) explains, git cherry-pick removes
the sequencer state just before applying the final patch. In the
single-pick case, that was a good thing, since --abort and --continue
work fine without access to such state and removing it provides a
signal that git should not complain about the need to clobber it ("a
cherry-pick or revert is already in progress") in sequences like the
following:
git cherry-pick foo
git read-tree -m -u HEAD; # forget that; let's try a different one
git cherry-pick bar
After the recent patch "allow single-pick in the middle of cherry-pick
sequence" we don't need that hack any more. In the new regime, a
traditional "git cherry-pick <commit>" command never looks at
.git/sequencer, so we do not need to cripple "git cherry-pick
<commit>..<commit>" for it any more.
So now you can run "git cherry-pick --abort" near the end of a
multi-pick sequence and it will abort the entire sequence, instead of
misbehaving and aborting just the final commit.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After messing up a difficult conflict resolution in the middle of a
cherry-pick sequence, it can be useful to be able to
git checkout HEAD . && git cherry-pick that-one-commit
to restart the conflict resolution. The current code however errors out
saying that another cherry-pick is already in progress.
Suggested-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 7e2bfd3f (revert: allow cherry-picking more than one commit,
2010-07-02), the pick/revert machinery has kept track of the set of
commits to be cherry-picked or reverted using commit_argc and
commit_argv variables, storing the corresponding command-line
parameters.
Future callers as other commands are built in (am, rebase, sequencer)
may find it easier to pass rev-list options to this machinery in
already-parsed form. Teach cmd_cherry_pick and cmd_revert to parse
the rev-list arguments in advance and pass the commit set to
pick_revisions() as a rev_info structure.
Original patch by Jonathan, tweaks and test from Ram.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Improved-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git cherry-pick ..bar" encounters conflicts, permit the operator
to use cherry-pick --continue after resolving them as a shortcut for
"git commit && git cherry-pick --continue" to record the resolution
and carry on with the rest of the sequence.
This improves the analogy with "git rebase" (in olden days --continue
was the way to preserve authorship when a rebase encountered
conflicts) and fits well with a general UI goal of making "git cmd
--continue" save humans the trouble of deciding what to do next.
Example: after encountering a conflict from running "git cherry-pick
foo bar baz":
CONFLICT (content): Merge conflict in main.c
error: could not apply f78a8d98c... bar!
hint: after resolving the conflicts, mark the corrected paths
hint: with 'git add <paths>' or 'git rm <paths>'
hint: and commit the result with 'git commit'
We edit main.c to resolve the conflict, mark it acceptable with "git
add main.c", and can run "cherry-pick --continue" to resume the
sequence.
$ git cherry-pick --continue
[editor opens to confirm commit message]
[master 78c8a8c98] bar!
1 files changed, 1 insertions(+), 1 deletions(-)
[master 87ca8798c] baz!
1 files changed, 1 insertions(+), 1 deletions(-)
This is done for both codepaths to pick multiple commits and a single
commit.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the case of --mixed and --hard, we throw away the old index and
rebuild everything from the tree argument (or HEAD). So we have an
opportunity here to fill in the cache-tree data, just as read-tree
did.
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In prepare_index(), we refresh the index, and then write it to disk if
this changed the index data. After running hooks we re-read the index
and compute the root tree sha1 with the cache-tree machinery.
This gives us a mostly free opportunity to write up-to-date cache-tree
data: we can compute it in prepare_index() immediately before writing
the index to disk.
If we do this, we were going to write the index anyway, and the later
cache-tree update has no further work to do. If we don't do it, we
don't do any extra work, though we still don't have have cache-tree
data after the commit.
The only case that suffers badly is when the pre-commit hook changes
many trees in the index. I'm writing this off as highly unusual.
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
