Junio C Hamano [Mon, 14 Jun 2021 04:33:26 +0000 (13:33 +0900)]
Merge branch 'ga/send-email-sendmail-cmd'
"git send-email" learned the "--sendmail-cmd" command line option
and the "sendemail.sendmailCmd" configuration variable, which is a
more sensible approach than the current way of repurposing the
"smtp-server" that is meant to name the server to instead name the
command to talk to the server.
* ga/send-email-sendmail-cmd:
git-send-email: add option to specify sendmail command
Junio C Hamano [Mon, 14 Jun 2021 04:33:25 +0000 (13:33 +0900)]
Merge branch 'zh/ref-filter-atom-type'
The code to handle the "--format" option in "for-each-ref" and
friends made too many string comparisons on %(atom)s used in the
format string, which has been corrected by converting them into
enum when the format string is parsed.
René Scharfe [Sun, 6 Jun 2021 01:01:57 +0000 (03:01 +0200)]
parallel-checkout: avoid dash local bug in tests
Dash bug https://bugs.launchpad.net/ubuntu/+source/dash/+bug/139097
lets the shell erroneously perform field splitting on the expansion of a
command substitution during declaration of a local variable. It causes
the parallel-checkout tests to fail e.g. when running them with
/bin/dash on MacOS 11.4, where they error out like this:
./t2080-parallel-checkout-basics.sh: 33: local: 0: bad variable name
That's because the output of wc -l contains leading spaces and the
returned number of lines is treated as another variable to declare, i.e.
as in "local workers= 0".
Work around it by enclosing the command substitution in quotes.
Helped-by: Matheus Tavares Bernardino <matheus.bernardino@usp.br> Helped-by: SZEDER Gábor <szeder.dev@gmail.com> Helped-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: René Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Fri, 4 Jun 2021 01:36:11 +0000 (10:36 +0900)]
fsync(): be prepared to see EINTR
Some platforms, like NonStop do not automatically restart fsync()
when interrupted by a signal, even when that signal is setup with
SA_RESTART.
This can lead to test breakage, e.g., where "--progress" is used,
thus SIGALRM is sent often, and can interrupt an fsync() syscall.
Make sure we deal with such a case by retrying the syscall
ourselves. Luckily, we call fsync() fron a single wrapper,
fsync_or_die(), so the fix is fairly isolated.
Reported-by: Randall S. Becker <randall.becker@nexbridge.ca> Helped-by: Jeff King <peff@peff.net> Helped-by: Taylor Blau <me@ttaylorr.com>
[jc: the above two did most of the work---I just tied the loose end] Helped-by: René Scharfe <l.s.r@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
David Aguilar [Tue, 1 Jun 2021 20:52:29 +0000 (13:52 -0700)]
contrib/completion: fix zsh completion regression from 59d85a2a05
A recent change to make git-completion.bash use $__git_cmd_idx
in more places broke a number of completions on zsh because it
modified __git_main but did not update __git_zsh_main.
Notably, completions for "add", "branch", "mv" and "push" were
broken as a result of this change.
In addition to the undefined variable usage, "git mv <tab>" also
prints the following error:
__git_count_arguments:7: bad math expression:
operand expected at `"1"'
_git_mv:[:7: unknown condition: -gt
Remove the quotes around $__git_cmd_idx in __git_count_arguments
and set __git_cmd_idx=1 early in __git_zsh_main to fix the
regressions from 59d85a2a05.
This was tested on zsh 5.7.1 (x86_64-apple-darwin19.0).
Suggested-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: David Aguilar <davvid@gmail.com> Acked-by: Felipe Contreras <felipe.contreras@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
builtin/fsck.c: don't conflate "int" and "enum" in callback
Fix a warning on AIX's xlc compiler that's been emitted since my a1aad71601a (fsck.h: use "enum object_type" instead of "int",
2021-03-28):
"builtin/fsck.c", line 805.32: 1506-068 (W) Operation between
types "int(*)(struct object*,enum object_type,void*,struct
fsck_options*)" and "int(*)(struct object*,int,void*,struct
fsck_options*)" is not allowed.
I.e. it complains about us assigning a function with a prototype "int"
where we're expecting "enum object_type".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Many places in the code were doing
while ((d = readdir(dir)) != NULL) {
if (is_dot_or_dotdot(d->d_name))
continue;
...process d...
}
Introduce a readdir_skip_dot_and_dotdot() helper to make that a one-liner:
while ((d = readdir_skip_dot_and_dotdot(dir)) != NULL) {
...process d...
}
This helper particularly simplifies checks for empty directories.
Also use this helper in read_cached_dir() so that our statistics are
consistent across platforms. (In other words, read_cached_dir() should
have been using is_dot_or_dotdot() and skipping such entries, but did
not and left it to treat_path() to detect and mark such entries as
path_none.)
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Derrick Stolee [Thu, 27 May 2021 04:53:55 +0000 (04:53 +0000)]
dir: update stale description of treat_directory()
The documentation comment for treat_directory() was originally written
in 095952 (Teach directory traversal about subprojects, 2007-04-11)
which was before the 'struct dir_struct' split its bitfield of named
options into a 'flags' enum in 7c4c97c0 (Turn the flags in struct
dir_struct into a single variable, 2009-02-16). When those flags
changed, the comment became stale, since members like
'show_other_directories' transitioned into flags like
DIR_SHOW_OTHER_DIRECTORIES.
Update the comments for treat_directory() to use these flag names rather
than the old member names.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano [Thu, 27 May 2021 03:36:58 +0000 (12:36 +0900)]
Merge branch 'ab/pack-linkage-fix'
"ld" on Solaris fails to link some test helpers, which has been
worked around by reshuffling the inline function definitions from a
header file to a source file that is the only user of them.
* ab/pack-linkage-fix:
pack-objects: move static inline from a header to the sole consumer
pack-objects: move static inline from a header to the sole consumer
Move the code that is only used in builtin/pack-objects.c out of
pack-objects.h.
This fixes an issue where Solaris's SunCC hasn't been able to compile
git since 483fa7f42d9 (t/helper/test-bitmap.c: initial commit,
2021-03-31).
The real origin of that issue is that in 898eba5e630 (pack-objects:
refer to delta objects by index instead of pointer, 2018-04-14)
utility functions only needed by builtin/pack-objects.c were added to
pack-objects.h. Since then the header has been used in a few other
places, but 483fa7f42d9 was the first time it was used by test helper.
Since Solaris is stricter about linking and the oe_get_size_slow()
function lives in builtin/pack-objects.c the build started failing
with:
Undefined first referenced
symbol in file
oe_get_size_slow t/helper/test-bitmap.o
ld: fatal: symbol referencing errors. No output written to t/helper/test-tool
On other platforms this is presumably OK because the compiler and/or
linker detects that the "static inline" functions that reference
oe_get_size_slow() aren't used.
Let's solve this by moving the relevant code from pack-objects.h to
builtin/pack-objects.c. This is almost entirely a code-only move, but
because of the early macro definitions in that file referencing some
of these inline functions we need to move the definition of "static
struct packing_data to_pack" earlier, and declare these inline
functions above the macros.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Matheus Tavares [Wed, 26 May 2021 23:58:56 +0000 (20:58 -0300)]
t2080: fix cp invocation to copy symlinks instead of following them
t2080 makes a few copies of a test repository and later performs a
branch switch on each one of the copies to verify that parallel checkout
and sequential checkout produce the same results. However, the
repository is copied with `cp -R` which, on some systems, defaults to
following symlinks on the directory hierarchy and copying their target
files instead of copying the symlinks themselves. AIX is one example of
system where this happens. Because the symlinks are not preserved, the
copied repositories have paths that do not match what is in the index,
causing git to abort the checkout operation that we want to test. This
makes the test fail on these systems.
Fix this by copying the repository with the POSIX flag '-P', which
forces cp to copy the symlinks instead of following them. Note that we
already use this flag for other cp invocations in our test suite (see
t7001). With this change, t2080 now passes on AIX.
Reported-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
send-email: move "hooks_path" invocation to git-send-email.perl
Move the newly added "hooks_path" API in Git.pm to its only user in
git-send-email.perl. This was added in c8243933c74 (git-send-email:
Respect core.hooksPath setting, 2021-03-23), meaning that it hasn't
yet made it into a non-rc release of git.
The consensus with Git.pm is that we need to be considerate of
out-of-tree users who treat it as a public documented interface. We
should therefore be less willing to add new functionality to it, least
we be stuck supporting it after our own uses for it disappear.
In this case the git-send-email.perl hook invocation will probably be
replaced by a future "git hook run" command, and in the commit
preceding this one the "hooks_path" become nothing but a trivial
wrapper for "rev-parse --git-path hooks" anyway (with no
Cwd::abs_path() call), so let's just inline this command in
git-send-email.perl itself.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
send-email: don't needlessly abs_path() the core.hooksPath
In c8243933c74 (git-send-email: Respect core.hooksPath setting,
2021-03-23) we started supporting core.hooksPath in "send-email". It's
been reported that on Windows[1] doing this by calling abs_path()
results in different canonicalizations of the absolute path.
This wasn't an issue in c8243933c74 itself, but was revealed by my ea7811b37e0 (git-send-email: improve --validate error output,
2021-04-06) when we started emitting the path to the hook, which was
previously only internal to git-send-email.perl.
The just-landed 53753a37d09 (t9001-send-email.sh: fix expected
absolute paths on Windows, 2021-05-24) narrowly fixed this issue, but
I believe we can do better here. We should not be relying on whatever
changes Perl's abs_path() makes to the path "rev-parse --git-path
hooks" hands to us. Let's instead trust it, and hand it to Perl's
system() in git-send-email.perl. It will handle either a relative or
absolute path.
So let's revert most of 53753a37d09 and just have "hooks_path" return
what we get from "rev-parse" directly without modification. This has
the added benefit of making the error message friendlier in the common
case, we'll no longer print an absolute path for repository-local hook
errors.
Junio C Hamano [Tue, 25 May 2021 20:52:34 +0000 (05:52 +0900)]
t1092: revert the "-1" hack for emulating "no progress meter"
This looked like a good idea, but it seems to break tests on 32-bit
builds rather badly. Revert to just use "100 thousands must be big
enough" for now.
Derrick Stolee [Mon, 24 May 2021 19:55:07 +0000 (19:55 +0000)]
t1092: use GIT_PROGRESS_DELAY for consistent results
The t1092-sparse-checkout-compatibility.sh tests compare the stdout and
stderr for several Git commands across both full checkouts, sparse
checkouts with a full index, and sparse checkouts with a sparse index.
Since these are direct comparisons, sometimes a progress indicator can
flush at unpredictable points, especially on slower machines. This
causes the tests to be flaky.
One standard way to avoid this is to add GIT_PROGRESS_DELAY=0 to the Git
commands that are run, as this will force every progress indicator
created with start_progress_delay() to be created immediately. However,
there are some progress indicators that are created in the case of a
full index that are not created with a sparse index. Moreover, their
values may be different as those indexes have a different number of
entries.
Instead, use GIT_PROGRESS_DELAY=-1 (which will turn into UINT_MAX)
to ensure that any reasonable machine running these tests would
never display delayed progress indicators.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Matheus Tavares [Tue, 25 May 2021 03:41:01 +0000 (00:41 -0300)]
init: fix bug regarding ~/ expansion in init.templateDir
We used to read the init.templateDir setting at builtin/init-db.c using
a git_config() callback that, in turn, called git_config_pathname(). To
simplify the config reading logic at this file and plug a memory leak,
this was replaced by a direct call to git_config_get_value() at e4de4502e6 ("init: remove git_init_db_config() while fixing leaks",
2021-03-14). However, this function doesn't provide path expanding
semantics, like git_config_pathname() does, so paths with '~/' and
'~user/' are treated literally. This makes 'git init' fail to handle
init.templateDir paths using these constructs:
$ git config init.templateDir '~/templates_dir'
$ git init
'warning: templates not found in ~/templates_dir'
Replace the git_config_get_value() call by git_config_get_pathname(),
which does the '~/' and '~user/' expansions. Also add a regression test.
Note that unlike git_config_get_value(), the config cache does not own
the memory for the path returned by git_config_get_pathname(), so we
must free() it.
Reported on IRC by rkta.
Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Fix a regression with the "the editor exited uncleanly, aborting
everything" error message going missing after my d21616c0394 (git-send-email: refactor duplicate $? checks into a
function, 2021-04-06).
I introduced a $msg variable, but did not actually use it. This caused
us to miss the optional error message supplied by the "do_edit"
codepath. Fix that, and add tests to check that this works.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Johannes Sixt [Mon, 24 May 2021 19:38:09 +0000 (21:38 +0200)]
t9001-send-email.sh: fix expected absolute paths on Windows
Git for Windows is a native Windows program that works with native
absolute paths in the drive letter style C:\dir. The auxiliary
infrastructure is based on MSYS2, which uses POSIX style /C/dir.
When we test for output of absolute paths produced by git.exe, we
usally have to expect C:\dir style paths. To produce such expected
paths, we have to use $(pwd) in the test scripts; the alternative,
$PWD, produces a POSIX style path. ($PWD is a shell variable, and the
shell is bash, an MSYS2 program, and operates in the POSIX realm.)
There are two recently added tests that were written to expect C:\dir
paths. The output that is tested is produced by `git send-email`, but
behind the scenes, this is a Perl script, which also works in the
POSIX realm and produces /C/dir style output.
In the first test case that is changed here, replace $(pwd) by $PWD
so that the expected path is constructed using /C/dir style.
The second test case sets core.hooksPath to an absolute path. Since
the test script talks to native git.exe, it is supposed to place a
C:/dir style path into the configuration; therefore, keep $(pwd).
When this configuration value is consumed by the Perl script, it is
transformed to /C/dir style by the MSYS2 layer and echoed back in
this form in the error message. Hence, do use $PWD for the expected
value.
Signed-off-by: Johannes Sixt <j6t@kdbg.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Denton Liu [Fri, 21 May 2021 10:37:47 +0000 (03:37 -0700)]
stash show: use stash.showIncludeUntracked even when diff options given
If options pertaining to how the diff is displayed is provided to
`git stash show`, the command will ignore the stash.showIncludeUntracked
configuration variable, defaulting to not showing any untracked files.
This is unintuitive behaviour since the format of the diff output and
whether or not to display untracked files are orthogonal.
Use stash.showIncludeUntracked even when diff options are given. Of
course, this is still overridable via the command-line options.
Update the documentation to explicitly say which configuration variables
will be overridden when a diff options are given.
Signed-off-by: Denton Liu <liu.denton@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff Hostetler [Thu, 20 May 2021 18:28:10 +0000 (18:28 +0000)]
simple-ipc: correct ifdefs when NO_PTHREADS is defined
Simple IPC always requires threads (in addition to various
platform-specific IPC support). Fix the ifdefs in the Makefile
to define SUPPORTS_SIMPLE_IPC when appropriate.
Previously, the Unix version of the code would only verify that
Unix domain sockets were available.
This problem was reported here:
https://lore.kernel.org/git/YKN5lXs4AoK%2FJFTO@coredump.intra.peff.net/T/#m08be8f1942ea8a2c36cfee0e51cdf06489fdeafc
Reported-by: Randall S. Becker <rsbecker@nexbridge.com> Helped-by: Jeff King <peff@peff.net> Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Regression fix for a change made during this cycle.
* cs/http-use-basic-after-failed-negotiate:
Revert "remote-curl: fall back to basic auth if Negotiate fails"
t5551: test http interaction with credential helpers
Jeff King [Wed, 19 May 2021 09:17:09 +0000 (05:17 -0400)]
doc: explain the use of color.pager
The current documentation for color.pager is technically correct, but
slightly misleading and doesn't really clarify the purpose of the
variable. As explained in the original thread which added it:
the point is to deal with pagers that don't understand colors. And hence it
being set to "true" is necessary for colorizing output to the pager, but
not sufficient by itself (you must also have enabled one of the other
color options, though note that these are set to "auto" by default these
days).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jiang Xin [Thu, 20 May 2021 04:56:10 +0000 (12:56 +0800)]
l10n: fix typos in po/TEAMS
Find typos in "po/TEAMS" file using the "git-po-helper" program. These
typos were introduced from commit v2.24.0-1-g9917eca794 (l10n: zh_TW:
add translation for v2.24.0, 2019-11-20 19:14:22 +0800).
Junio C Hamano [Wed, 19 May 2021 23:55:00 +0000 (08:55 +0900)]
Merge branch 'jk/test-chainlint-softer'
The "chainlint" feature in the test framework is a handy way to
catch common mistakes in writing new tests, but tends to get
expensive. An knob to selectively disable it has been introduced
to help running tests that the developer has not modified.
* jk/test-chainlint-softer:
t: avoid sed-based chain-linting in some expensive cases
Junio C Hamano [Wed, 19 May 2021 23:55:00 +0000 (08:55 +0900)]
Merge branch 'zh/ref-filter-push-remote-fix'
The handling of "%(push)" formatting element of "for-each-ref" and
friends was broken when the same codepath started handling
"%(push:<what>)", which has been corrected.
* zh/ref-filter-push-remote-fix:
ref-filter: fix read invalid union member bug
Junio C Hamano [Wed, 19 May 2021 23:54:59 +0000 (08:54 +0900)]
Merge branch 'ew/sha256-clone-remote-curl-fix'
"git clone" from SHA256 repository by Git built with SHA-1 as the
default hash algorithm over the dumb HTTP protocol did not
correctly set up the resulting repository, which has been corrected.
* ew/sha256-clone-remote-curl-fix:
remote-curl: fix clone on sha256 repos
Junio C Hamano [Wed, 19 May 2021 23:54:58 +0000 (08:54 +0900)]
Merge branch 'en/dir-traversal'
"git clean" and "git ls-files -i" had confusion around working on
or showing ignored paths inside an ignored directory, which has
been corrected.
* en/dir-traversal:
dir: introduce readdir_skip_dot_and_dotdot() helper
dir: update stale description of treat_directory()
dir: traverse into untracked directories if they may have ignored subfiles
dir: avoid unnecessary traversal into ignored directory
t3001, t7300: add testcase showcasing missed directory traversal
t7300: add testcase showing unnecessary traversal into ignored directory
ls-files: error out on -i unless -o or -c are specified
dir: report number of visited directories and paths with trace2
dir: convert trace calls to trace2 equivalents
Jeff King [Wed, 19 May 2021 16:11:05 +0000 (12:11 -0400)]
fetch-pack: signal v2 server that we are done making requests
When fetching with the v0 protocol over ssh (or a local upload-pack with
pipes), the server closes the connection as soon as it is finished
sending the pack. So even though the client may still be operating on
the data via index-pack (e.g., resolving deltas, checking connectivity,
etc), the server has released all resources.
With the v2 protocol, however, the server considers the ssh session only
as a transport, with individual requests coming over it. After sending
the pack, it goes back to its main loop, waiting for another request to
come from the client. As a result, the ssh session hangs around until
the client process ends, which may be much later (because resolving
deltas, etc, may consume a lot of CPU).
This is bad for two reasons:
- it's consuming resources on the server to leave open a connection
that won't see any more use
- if something bad happens to the ssh connection in the meantime (say,
it gets killed by the network because it's idle, as happened in a
real-world report), then ssh will exit non-zero, and we'll propagate
the error up the stack.
The server is correct here not to hang up after serving the pack. The v2
protocol's design is meant to allow multiple requests like this, and
hanging up would be the wrong thing for a hypothetical client which was
planning to make more requests (though in practice, the git.git client
never would, and I doubt any other implementations would either).
The right thing is instead for the client to signal to the server that
it's not interested in making more requests. We can do that by closing
the pipe descriptor we use to write to ssh. This will propagate to the
server upload-pack as an EOF when it tries to read the next request (and
then it will close its half, and the whole connection will go away).
It's important to do this "half duplex" shutdown, because we have to do
it _before_ we actually receive the pack. This is an artifact of the way
fetch-pack and index-pack (or unpack-objects) interact. We hand the
connection off to index-pack (really, a sideband demuxer which feeds
it), and then wait until it returns. And it doesn't do that until it has
resolved all of the deltas in the pack, even though it was done reading
from the server long before.
So just closing the connection fully after index-pack returns would be
too late; we'd have held it open much longer than was necessary. And
teaching index-pack to close the connection is awkward. It's not even
seeing the whole conversation (the sideband demuxer is, but it doesn't
actually know what's in the packets, or when the end comes).
Note that this close() is happening deep within the transport code. It's
possible that a caller would want to perform other operations over the
same ssh transport after receiving the pack. But as of the current code,
none of the callers do, and there haven't been discussions of any plans
to change this. If we need to support that later, we can probably do so
by passing down a flag for "you're the last request on the transport;
it's OK to close" instead of the code just assuming that's true.
The description above all discusses v2 ssh, so it's worth thinking about
how this interacts with other protocols:
- in v0 protocols, we could do the same half-duplex shutdown (it just
goes into the v0 do_fetch_pack() instead). This does work, but since
it doesn't have the same persistence problem in the first place,
there's little reason to change it at this point.
- local fetches against git-upload-pack on the same machine will
behave the same as ssh (they are talking over two pipes, and see EOF
on their input pipe)
- fetches against git-daemon will run this same code, and close one of
the descriptors. In practice, this won't do anything, since there
our two descriptors are dups of each other, and not part of a
half-duplex pair. The right thing would probably be to call
shutdown(SHUT_WR) on it. I didn't bother with that here. It doesn't
face the same error-code problem (since it's just a TCP connection),
so it's really only an optimization problem. And git:// is not that
widely used these days, and has less impact on server resources than
an ssh termination.
- v2 http doesn't suffer from this problem in the first place, as our
pipes terminate at a local git-remote-https, which is passing data
along as individual requests via curl. Probably curl is keeping the
TCP/TLS connection open for more requests, and we might be able to
tell it manually "hey, we are done making requests now". But I think
that's much less important. It again doesn't suffer from the
error-code problem, and HTTP keepalive is pretty well understood
(importantly, the timeouts can be set low, because clients like curl
know how to reconnect for subsequent requests if necessary). So it's
probably not worth figuring out how to tell curl that we're done
(though if we do, this patch is probably the first step anyway;
fetch-pack closes the pipe back to remote-https, which would be the
signal that it should tell curl we're done).
The code is pretty straightforward. We close the pipe at the right
moment, and set it to -1 to mark it as invalid. I modified the later
cleanup code to avoid calling close(-1). That's not strictly necessary,
since close(-1) is a noop, but hopefully makes things a bit more obvious
to a reader.
I suspect that trying to call more transport functions after the close()
(e.g., calling transport_fetch_refs() again) would fail, as it's not
smart enough to realize we need to re-open the ssh connection. But
that's already true when v0 is in use. And no current callers want to do
that (and again, the solution is probably a flag in the transport code
to keep things open, which can be added later).
There's no test here, as the situation it covers is inherently racy (the
question is when upload-pack exits, compared to when index-pack finishes
resolving deltas and exits). The rather gross shell snippet below does
recreate the problematic situation; when run on a sufficiently-large
repository (git.git works fine), it kills an "idle" upload-pack while
the client is resolving deltas, leading to a failed clone.
(
git clone --no-local --progress . foo.git 2>&1
echo >&2 "clone exit code=$?"
) |
tr '\r' '\n' |
while read line
do
case "$done,$line" in
,Resolving*)
echo "hit resolving deltas; killing upload-pack"
killall -9 git-upload-pack
done=t
;;
esac
done
Reported-by: Greg Pflaum <greg.pflaum@pnp-hcl.com> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King [Wed, 19 May 2021 11:17:15 +0000 (07:17 -0400)]
clone: clean up directory after transport_fetch_refs() failure
git-clone started respecting errors from the transport subsystem in aab179d937 (builtin/clone.c: don't ignore transport_fetch_refs() errors,
2020-12-03). However, that commit didn't handle the cleanup of the
filesystem quite right.
The cleanup of the directory that cmd_clone() creates is done by an
atexit() handler, which we control with a flag. It starts as
JUNK_LEAVE_NONE ("clean up everything"), then progresses to
JUNK_LEAVE_REPO when we know we have a valid repo but not working tree,
and then finally JUNK_LEAVE_ALL when we have a successful checkout.
Most errors cause us to die(), which then triggers the handler to do the
right thing based on how far into cmd_clone() we got. But the checks
added by aab179d937 instead set the "err" variable and then jump to a
new "cleanup" label, which then returns our non-zero status. However,
the code after the cleanup label includes setting the flag to
JUNK_LEAVE_ALL, and so we accidentally leave the repository and working
tree in place.
One obvious option to fix this is to reorder the end of the function to
set the flag first, before cleanup code, and put the label between them.
But we can observe another small bug: the error return from
transport_fetch_refs() is generally "-1", and we propagate that to the
return value of cmd_clone(), which ultimately becomes the exit code of
the process. And we try to avoid transmitting negative values via exit
codes (only the low 8 bits are passed along as an unsigned value, though
in practice for "-1" this at least retains the property that it's
non-zero).
Instead, let's just die(). That makes us consistent with rest of the
code in the function. It does add a new "fatal:" line to the output, but
I'd argue that's a good thing:
- in the rare case that the transport code didn't say anything, now
the user gets _some_ error message
- even if the transport code said something like "error: ssh died of
signal 9", it's nice to also say "fatal" to indicate that we
considered that to be a show-stopper.
Triggering this in the test suite turns out to be surprisingly
difficult. Almost every error we'd encounter, including ones deep inside
the transport code, cause us to just die() right there! However, one way
is to put a fake wrapper around git-upload-pack that sends the complete
packfile but exits with a failure code.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Derrick Stolee [Tue, 18 May 2021 18:32:48 +0000 (18:32 +0000)]
read-cache: delete unused hashing methods
These methods were marked as MAYBE_UNUSED in the previous change to
avoid a complicated diff. Delete them entirely, since we now use the
hashfile API instead of this custom hashing code.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Derrick Stolee [Tue, 18 May 2021 18:32:47 +0000 (18:32 +0000)]
read-cache: use hashfile instead of git_hash_ctx
The do_write_index() method in read-cache.c has its own hashing logic
and buffering mechanism. Specifically, the ce_write() method was
introduced by 4990aadc (Speed up index file writing by chunking it
nicely, 2005-04-20) and similar mechanisms were introduced a few months
later in c38138cd (git-pack-objects: write the pack files with a SHA1
csum, 2005-06-26). Based on the timing, in the early days of the Git
codebase, I figured that these roughly equivalent code paths were never
unified only because it got lost in the shuffle. The hashfile API has
since been used extensively in other file formats, such as pack-indexes,
multi-pack-indexes, and commit-graphs. Therefore, it seems prudent to
unify the index writing code to use the same mechanism.
I discovered this disparity while trying to create a new index format
that uses the chunk-format API. That API uses a hashfile as its base, so
it is incompatible with the custom code in read-cache.c.
This rewrite is rather straightforward. It replaces all writes to the
temporary file with writes to the hashfile struct. This takes care of
many of the direct interactions with the_hash_algo.
There are still some git_hash_ctx uses remaining: the extension headers
are hashed for use in the End of Index Entries (EOIE) extension. This
use of the git_hash_ctx is left as-is. There are multiple reasons to not
use a hashfile here, including the fact that the data is not actually
writing to a file, just a hash computation. These hashes do not block
our adoption of the chunk-format API in a future change to the index, so
leave it as-is.
The internals of the algorithms are mostly identical. Previously, the
hashfile API used a smaller 8KB buffer instead of the 128KB buffer from
read-cache.c. The previous change already unified these sizes.
There is one subtle point: we do not pass the CSUM_FSYNC to the
finalize_hashfile() method, which differs from most consumers of the
hashfile API. The extra fsync() call indicated by this flag causes a
significant peformance degradation that is noticeable for quick
commands that write the index, such as "git add". Other consumers can
absorb this cost with their more complicated data structure
organization, and further writing structures such as pack-files and
commit-graphs is rarely in the critical path for common user
interactions.
Some static methods become orphaned in this diff, so I marked them as
MAYBE_UNUSED. The diff is much harder to read if they are deleted during
this change. Instead, they will be deleted in the following change.
In addition to the test suite passing, I computed indexes using the
previous binaries and the binaries compiled after this change, and found
the index data to be exactly equal. Finally, I did extensive performance
testing of "git update-index --force-write" on repos of various sizes,
including one with over 2 million paths at HEAD. These tests
demonstrated less than 1% difference in behavior. As expected, the
performance should be considered unchanged. The previous changes to
increase the hashfile buffer size from 8K to 128K ensured this change
would not create a peformance regression.
Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
projects / rit / log