dumb-http walker has been updated to share more error recovery
strategy with the normal codepath.
* jk/http-walker-status-fix:
http: use normalize_curl_result() instead of manual conversion
http: normalize curl results for dumb loose and alternates fetches
http: factor out curl result code normalization
"git multi-pack-index verify" did not scale well with the number of
packfiles, which is being improved.
* jh/midx-verify-too-many-packs:
midx: during verify group objects by packfile to speed verification
midx: add progress indicators in multi-pack-index verify
trace2:data: add trace2 data to midx
progress: add sparse mode to force 100% complete message
The completion helper code now pays attention to repository-local
configuration (when available), which allows --list-cmds to honour
a repository specific setting of completion.commands, for example.
* tz/completion:
completion: use __git when calling --list-cmds
completion: fix multiple command removals
t9902: test multiple removals via completion.commands
git: read local config in --list-cmds
Dev support update to make it easier to compare two formatted
results from our documentation.
* ma/doc-diff-doc-vs-doctor-comparison:
doc-diff: add `--cut-header-footer`
doc-diff: support diffing from/to AsciiDoc(tor)
doc-diff: let `render_tree()` take an explicit directory name
Doc: auto-detect changed build flags
A corner-case object name ambiguity while the sequencer machinery
is working (e.g. "rebase -i -x") has been (half) fixed.
* js/get-short-oid-drop-cache:
get_oid(): when an object was not found, try harder
sequencer: move stale comment into correct location
sequencer: improve error message when an OID could not be parsed
rebase -i: demonstrate obscure loose object cache bug
* ma/asciidoctor-fixes-more:
Documentation: turn middle-of-line tabs into spaces
git-svn.txt: drop escaping '\' that ends up being rendered
git.txt: remove empty line before list continuation
config/fsck.txt: avoid starting line with dash
config/diff.txt: drop spurious backtick
Help developers by making it easier to run most of the tests under
different versions of over-the-wire protocols.
* jt/test-protocol-version:
t5552: compensate for v2 filtering ref adv.
tests: fix protocol version for overspecifications
t5700: only run with protocol version 1
t5512: compensate for v0 only sending HEAD symrefs
t5503: fix overspecification of trace expectation
tests: always test fetch of unreachable with v0
t5601: check ssh command only with protocol v0
tests: define GIT_TEST_PROTOCOL_VERSION
"git worktree add" used to do a "find an available name with stat
and then mkdir", which is race-prone. This has been fixed by using
mkdir and reacting to EEXIST in a loop.
* ra/t3600-test-path-funcs:
t3600: use helpers to replace test -d/f/e/s <path>
t3600: modernize style
test functions: add function `test_file_not_empty`
"git rebase" uses the refs/rewritten/ hierarchy to store its
intermediate states, which inherently makes the hierarchy per
worktree, but it didn't quite work well.
* nd/rewritten-ref-is-per-worktree:
Make sure refs/rewritten/ is per-worktree
files-backend.c: reduce duplication in add_per_worktree_entries_to_dir()
files-backend.c: factor out per-worktree code in loose_fill_ref_dir()
When the "clean" filter can reduce the size of a huge file in the
working tree down to a small "token" (a la Git LFS), there is no
point in allocating a huge scratch area upfront, but the buffer is
sized based on the original file size. The convert mechanism now
allocates very minimum and reallocates as it receives the output
from the clean filter process.
* jh/resize-convert-scratch-buffer:
convert: avoid malloc of original file size
http: use normalize_curl_result() instead of manual conversion
When we switched off CURLOPT_FAILONERROR in 17966c0a63 (http: avoid
disconnecting on 404s for loose objects, 2016-07-11), the fetch_object()
function started manually handling 404's. Since we now have
normalize_curl_result() for use elsewhere, we can use it here as well,
shortening the code.
Note that we lose the check for http/https in the URL here. None of the
other result-normalizing code paths bother with this. Looking at
missing_target(), which checks specifically for an FTP-specific CURLcode
and "http" code 550, it seems likely that git-over-ftp has been subtly
broken since 17966c0a63. This patch does nothing to fix that, but nor
should it make anything worse (in fact, it may be slightly better
because we'll actually recognize an error as such, rather than assuming
CURLE_OK means we actually got some data).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
http: normalize curl results for dumb loose and alternates fetches
If the dumb-http walker encounters a 404 when fetching a loose object,
it then looks at any http-alternates for the object. The 404 check is
implemented by missing_target(), which checks not only the http code,
but also that we got an http error from the CURLcode.
That broke when we stopped using CURLOPT_FAILONERROR in 17966c0a63
(http: avoid disconnecting on 404s for loose objects, 2016-07-11), since
our CURLcode will now be CURLE_OK. As a result, fetching over dumb-http
from a repository with alternates could result in Git printing "Unable
to find abcd1234..." and aborting.
We could probably fix this just by loosening missing_target(). However,
there's other code which looks at the curl result, and it would have to
be tweaked as well. Instead, let's just normalize the result the same
way the smart-http code does.
There's a similar case in processing the alternates (where we failover
from "info/http-alternates" to "info/alternates"). We'll give it the
same treatment.
After this patch, we should be hitting all code paths that need this
normalization (notably absent here is the http_pack_request path, but it
does not use FAILONERROR, nor missing_target()).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We make some requests with CURLOPT_FAILONERROR and some without, and
then handle_curl_result() normalizes any failures to a uniform CURLcode.
There are some other code paths in the dumb-http walker which don't use
handle_curl_result(); let's pull the normalization into its own function
so it can be reused.
Arguably those code paths would benefit from the rest of
handle_curl_result(), notably the auth handling. But retro-fitting it
now would be a lot of work, and in practice it doesn't matter too much
(whatever authentication we needed to make the initial contact with the
server is generally sufficient for the rest of the dumb-http requests).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
refs/files-backend: don't look at an aborted transaction
When deleting refs, we hold packed-refs.lock and prepare a packed
transaction to drop the refs from the packed-refs file. If it turns out
that we don't need to rewrite the packed refs (e.g., because none of the
deletions were present in the file), then we abort the transaction.
If that abort succeeds, then the transaction struct will have been
freed, and we set our local pointer to NULL so we don't look at it
again.
However, if it fails, then the struct will _still_ have been freed
(because ref_transaction_abort() always frees). But we don't clean up
the pointer, and will jump to our cleanup code, which will try to abort
it again, causing a use-after-free.
It's actually impossible for this to trigger in practice, since
packed_transaction_abort() will never return anything but success. But
let's fix it anyway, since that's more than we should assume about the
packed-refs code (after all, we are already bothering to check for an
error result which cannot be triggered).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
In files_transaction_prepare(), if we have to delete some refs, we use a
subordinate packed_transaction to do so. It's rare for that
sub-transaction's prepare step to fail, since we hold the packed-refs
lock. But if it does, we trigger a BUG() due to these steps:
- we've attached the packed transaction to the files transaction as
backend_data->packed_transaction
- when the prepare step fails, the packed transaction cleans itself
up, putting itself into the CLOSED state
- the error value from preparing the packed transaction lets us know
in files_transaction_prepare() that we should also clean up and
return an error. We call files_transaction_cleanup(), which tries to
abort backend_data->packed_transaction. Since it's already CLOSED,
that triggers an assertion in ref_transaction_abort().
We can fix that by disconnecting the packed transaction from the outer
files transaction, and then free-ing (not aborting!) it ourselves.
A few other options/alternatives I considered:
- we could just make it a noop to abort a CLOSED transaction. But that
seems less safe, since clearly this code expects (and enforces) a
particular set of state transitions.
- we could have files_transaction_cleanup() selectively call abort()
vs free() based on the state of the on the packed transaction.
That's basically a more restricted version of the above, but also
potentially unsafe.
- instead of disconnecting backend_data->packed_transaction on error,
we could wait to install it until we successfully prepare. That
might make the flow a little simpler, but it introduces a hassle.
Earlier parts of files_transaction_prepare() that encounter an error
will jump to the cleanup label, and expect that cleaning up the
outer transaction will clean up the packed transaction, too. We'd
have to adjust those sites to clean up the packed transaction.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
midx: during verify group objects by packfile to speed verification
Teach `multi-pack-index verify` to sort the set of object by
packfile so that only one packfile needs to be open at a time.
This is a performance improvement. Previously, objects were
verified in OID order. This essentially requires all packfiles
to be open at the same time. If the number of packfiles exceeds
the open file limit, packfiles would be LRU-closed and re-opened
many times.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
progress: add sparse mode to force 100% complete message
Add new start_sparse_progress() and start_delayed_sparse_progress()
constructors and "sparse" flag to struct progress.
Teach stop_progress() to force a 100% complete progress message before
printing the final "done" message when "sparse" is set.
Calling display_progress() for every item in a large set can
be expensive. If callers try to filter this for performance
reasons, such as emitting every k-th item, progress would
not reach 100% unless they made a final call to display_progress()
with the item count before calling stop_progress().
Now this is automatic when "sparse" is set.
Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
checkout.txt: note about losing staged changes with --merge
If you have staged changes in path A and perform 'checkout
--merge' (which could result in conflicts in a totally unrelated path
B), changes in A will be gone. Which is unexpected. We are supposed
to keep all changes, or kick and scream otherwise.
This is the result of how --merge is implemented, from the very first
day in 1be0659efc (checkout: merge local modifications while switching
branches., 2006-01-12):
1. a merge is done, unmerged entries are collected
2. a hard switch to a new branch is done, then unmerged entries added
back
There is no trivial fix for this. Going with 3-way merge one file at a
time loses rename detection. Going with 3-way merge by trees requires
teaching the algorithm to pick up staged changes. And even if we detect
staged changes with --merge and abort for safety, an option to continue
--merge is very weird. Such an option would keep worktree changes, but
drop staged changes.
Because the problem has been with us since the introduction of --merge
and everybody has been pretty happy (except Phillip, who found this
problem), I'll just take a note here to acknowledge it and wait for
merge wizards to come in and work their magic. There may be a way
forward [1].
unpack-trees: fix oneway_merge accidentally carry over stage index
Phillip found out that 'git checkout -f <branch>' does not restore
conflict/unmerged files correctly. All tracked files should be taken
from <branch> and all non-zero stages removed. Most of this is true,
except that the final file could be in stage one instead of zero.
"checkout -f" (among other commands) does this with one-way merge, which
is supposed to take stat info from the index and everything else from
the given tree. The add_entry(.., old, ...) call in oneway_merge()
though will keep stage index from the index.
This is normally not a problem if the entry from the index is
normal (stage #0). But if there is a conflict, stage #0 does not exist
and we'll get stage #1 entry as "old" variable, which gets recorded in
the final index. Fix it by clearing stage mask.
This bug probably comes from b5b425074e (git-read-tree: make one-way
merge also honor the "update" flag, 2005-06-07). Before this commit, we
may create the final ("dst") index entry from the one in index, but we
do clear CE_STAGEMASK.
I briefly checked two- and three-way merge functions. I think we don't
have the same problem in those.
Reported-by: Phillip Wood <phillip.wood@dunelm.org.uk> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
As we made --list-cmds read the local configuration file in an
earlier step, the completion.commands variable respects repo-level
configuration. Use __git which ensures that the proper repo config
is consulted if the command line contains 'git -C /some/other/repo'.
Suggested-by: SZEDER Gábor <szeder.dev@gmail.com> Signed-off-by: Todd Zullinger <tmz@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 6532f3740b ("completion: allow to customize the completable
command list", 2018-05-20) tried to allow multiple space-separated
entries in completion.commands. To do this, it copies each parsed token
into a strbuf so that the result is NUL-terminated.
However, for tokens starting with "-", it accidentally passes the
original non-terminated string, meaning that only the final one worked.
Switch to using the strbuf.
Reported-by: Todd Zullinger <tmz@pobox.com> Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t9902: test multiple removals via completion.commands
6532f3740b ("completion: allow to customize the completable command
list", 2018-05-20) added the completion.commands config variable.
Multiple commands may be added or removed, separated by a space.
Demonstrate the failure of multiple removals.
Signed-off-by: Todd Zullinger <tmz@pobox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Normally code that is checking config before we've decided to do
setup_git_directory() would use read_early_config(), which uses
discover_git_directory() to tentatively see if we're in a repo,
and if so to add it to the config sequence.
But list_cmds() uses the caching configset mechanism which
rightly does not use read_early_config(), because it has no
idea if it's being called early.
Call setup_git_directory_gently() so we can pick up repo-level
config (like completion.commands).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git config --type=color ..." is meant to replace "git config --get-color"
but there is a slight difference that wasn't documented, which is
now fixed.
* jk/config-type-color-ends-with-lf:
config: document --type=color output is a complete line
A recent update broke "is this object available to us?" check for
well-known objects like an empty tree (which should yield "yes",
even when there is no on-disk object for an empty tree), which has
been corrected.
* jk/virtual-objects-do-exist:
rev-list: allow cached objects in existence check
On platforms where "git fetch" is killed with SIGPIPE (e.g. OSX),
the upload-pack that runs on the other end that hangs up after
detecting an error could cause "git fetch" to die with a signal,
which led to a flakey test. "git fetch" now ignores SIGPIPE during
the network portion of its operation (this is not a problem as we
check the return status from our write(2)s).
"git fsck --connectivity-only" omits computation necessary to sift
the objects that are not reachable from any of the refs into
unreachable and dangling. This is now enabled when dangling
objects are requested (which is done by default, but can be
overridden with the "--no-dangling" option).
* jk/fsck-doc:
fsck: always compute USED flags for unreachable objects
doc/fsck: clarify --connectivity-only behavior
"git rebase" that was reimplemented in C did not set ORIG_HEAD
correctly, which has been corrected.
* js/rebase-orig-head-fix:
built-in rebase: set ORIG_HEAD just once, before the rebase
built-in rebase: demonstrate that ORIG_HEAD is not set correctly
built-in rebase: use the correct reflog when switching branches
built-in rebase: no need to check out `onto` twice
The final report from "git bisect" used to show the suspected
culprit using a raw "diff-tree", with which there is no output for
a merge commit. This has been updated to use a more modern and
human readable output that still is concise enough.
* jk/bisect-final-output:
bisect: make diff-tree output prettier
bisect: fix internal diff-tree config loading
bisect: use string arguments to feed internal diff-tree
CFLAGS now can be tweaked when invoking Make while using
DEVELOPER=YesPlease; this did not work well before.
* ab/makefile-help-devs-more:
Makefile: allow for combining DEVELOPER=1 and CFLAGS="..."
Makefile: move the setting of *FLAGS closer to "include"
Makefile: Move *_LIBS assignment into its own section
Makefile: add/remove comments at top and tweak whitespace
Makefile: move "strip" assignment down from flags
Makefile: remove an out-of-date comment
Some compilers don't allow NULL to be passed for a va_list,
and e.g. "gcc (Raspbian 6.3.0-18+rpi1+deb9u1) 6.3.0 20170516"
errors out like this:
trace2/tr2_tgt_event.c:193:18:
error: invalid operands to binary &&
(have ‘int’ and ‘va_list {aka __va_list}’)
if (fmt && *fmt && ap) {
^^
I couldn't find any hints that va_list and pointers can be mixed,
and no hints that they can't either. Morten Welinder comments:
"C99, Section 7.15, simply says that va_list "is an object type suitable for
holding information needed by the macros va_start, va_end, and
va_copy". So clearly not guaranteed to be mixable with pointers...
The portable solution is to use "va_list" everywhere in the callchain.
As a consequence, both trace2_region_enter_fl() and trace2_region_leave_fl()
now take a variable argument list.
Signed-off-by: Torsten Bögershausen <tboegi@web.de> Acked-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the rebase.useBuiltin setting, which was added as an escape
hatch to disable the builtin version of rebase first released with Git
2.20.
See [1] for the initial implementation of rebase.useBuiltin, and [2]
and [3] for the documentation and corresponding
GIT_TEST_REBASE_USE_BUILTIN option.
Carrying the legacy version is a maintenance burden as seen in 7e097e27d3 ("legacy-rebase: backport -C<n> and --whitespace=<option>
checks", 2018-11-20) and 9aea5e9286 ("rebase: fix regression in
rebase.useBuiltin=false test mode", 2019-02-13). Since the built-in
version has been shown to be stable enough let's remove the legacy
version.
As noted in [3] having use_builtin_rebase() shell out to get its
config doesn't make any sense anymore, that was done for the purposes
of spawning the legacy rebase without having modified any global
state. Let's instead handle this case in rebase_config().
There's still a bunch of references to git-legacy-rebase in po/*.po,
but those will be dealt with in time by the i18n effort.
Even though this configuration variable only existed two releases
let's not entirely delete the entry from the docs, but note its
absence. Individual versions of git tend to be around for a while due
to distro packaging timelines, so e.g. if we're "lucky" a given
version like 2.21 might be installed on say OSX for half a decade.
That'll mean some people probably setting this in config, and then
when they later wonder if it's needed they can Google search the
config option name or check it in git-config. It also allows us to
refer to the docs from the warning for details.
1. 55071ea248 ("rebase: start implementing it as a builtin",
2018-08-07)
2. d8d0a546f0 ("rebase doc: document rebase.useBuiltin", 2018-11-14)
3. 62c23938fa ("tests: add a special setup where rebase.useBuiltin is
off", 2018-11-14)
3. https://public-inbox.org/git/nycvar.QRO.7.76.6.1903141544110.41@tvgsbejvaqbjf.bet/
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
AsciiDoc and Asciidoctor do not agree on what to write in the header and
footer of each man-page, i.e., the very first and the very last line of
*.[157]. Those differences can certainly be interesting in their own
right, but they clutter the output of `./doc-diff --from-asciidoc
--to-asciidoctor HEAD HEAD` quite a bit since the diff contains some
10-15 lines of noise per file diffed.
Teach doc-diff to cut away the first two and last two lines, i.e., the
header/footer and the empty line immediately following/preceding it.
Because Asciidoctor uses an extra empty line compared to AsciiDoc,
remove one more line at each end of the file, but only if it's empty.
An alternative approach might be to pass down `--no-header-footer`,
which both AsciiDoc and Asciidoctor understand, but it has some
drawbacks. First of all, the result doesn't build -- `xmlto` stumbles on
the resulting xml since it has multiple root elements. Second, it cuts
too much -- dropping the header loses the synopsis, which would be
interesting to diff.
Like in the previous commit, encode this option into the directory name
of the "installed" and "rendered" files. Otherwise, we wouldn't be able
to trust that what we use out of that cache actually corresponds to the
options given for this run. (We could optimize this caching a little
since this flag doesn't affect the contents of "installed" at all, but
let's punt on that.)
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Provide `--from-asciidoctor` and `--to-asciidoctor` to select that the
"from" resp. "to" commit should be built with Asciidoctor, and provide
an `--asciidoctor` shortcut for giving both. Similarly, provide
--{from-,to-,}asciidoc for explicitly selecting AsciiDoc.
Implement this using the USE_ASCIIDOCTOR flag. Let's not enforce a
default here, but instead just let the Makefile fall back on whatever is
in config.mak, so that `./doc-diff foo bar` without any of of these new
options behaves exactly like it did before this commit.
Encode the choice into the directory names of our "installed" and
"rendered" files, so that we can run `./doc-diff --from-asciidoc
--to-asciidoctor HEAD HEAD` without our two runs stomping on each other.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
doc-diff: let `render_tree()` take an explicit directory name
In `render_tree()`, `$1` is documented to be the commit-ish/oid and we
use it as that with `git checkout`, but we mostly use it to form the
name of various directories. To separate these concerns, and because we
are about to construct the directory names a bit differently, take two
distinct arguments instead.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
If you build the documentation switching between different options,
e.g., to build with both Asciidoc and Asciidoctor, you'll probably find
yourself running `make -C Documentation clean` either too often (wasting
time) or too rarely (getting mixed builds).
Track the flags we're using in the documentation build, similar to how
the main Makefile tracks CFLAGS and prefix flags. Track ASCIIDOC_COMMON
directly rather than its individual components -- that should make it
harder to forget to update the tracking if/when we modify the build
commands.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
perf-lib.sh: rely on test-lib.sh for --tee handling
Since its inception, the perf-lib.sh script has manually handled the
"--tee" option (and other options which imply it, like "--valgrind")
with a cut-and-pasted block from test-lib.sh. That block has grown stale
over the years, and has at least three problems:
1. It uses $SHELL to re-exec the script, whereas the version in
test-lib.sh learned to use $TEST_SHELL_PATH.
2. It does an ad-hoc search of the "$*" string, whereas test-lib.sh
learned to carefully parse the arguments left to right.
3. It never learned about --verbose-log (which also implies --tee),
so it would not trigger for that option.
This last one was especially annoying, because t/perf/run uses the
GIT_TEST_OPTS from your config.mak to run the perf scripts. So if you've
set, say, "-x --verbose-log" there, it will be passed as part of most
perf runs. And while this script doesn't recognize the option, the
test-lib.sh that we source _does_, and the behavior ends up being much
more annoying:
- as the comment at the top of the block says, we have to run this
tee code early, before we start munging variables (it says
GIT_BUILD_DIR, but the problematic variable is actually
GIT_TEST_INSTALLED).
- since we don't recognize --verbose-log, we don't trigger the block.
We go on to munge GIT_TEST_INSTALLED, converting it from a relative
to an absolute path.
- then we source test-lib.sh, which _does_ recognize --verbose-log. It
re-execs the script, which runs again. But this time with an
absolute version of GIT_TEST_INSTALLED.
- As a result, we copy the absolute version of GIT_TEST_INSTALLED into
perf_results_prefix. Instead of writing our results to the expected
"test-results/build_1234abcd.p1234-whatever.times", we instead write
them to "test-results/_full_path_to_repo_t_perf_build_1234...".
The aggregate.perl script doesn't expect this, and so it prints
"<missing>" for each result (even though it spent considerable time
running the tests!).
We can solve all of these in one blow by just deleting our custom
handling, and relying on the inclusion of test-lib.sh to handle --tee,
--verbose-log, etc.
There's one catch, though. We want to handle GIT_TEST_INSTALLED after
we've included test-lib.sh, since we want it un-munged in the re-exec'd
version of the script. But if we want to convert it from a relative
to an absolute path, we must do so before we load test-lib.sh, since it
will change our working directory. So we compute the absolute directory
first, store it away, then include test-lib.sh, and finally assign to
GIT_TEST_INSTALLED as appropriate.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a definition for what overlay means in the context of git, to
clarify the recently introduced overlay-mode in git checkout.
Helped-by: Elijah Newren <newren@gmail.com> Helped-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
get_oid(): when an object was not found, try harder
It is quite possible that the loose object cache gets stale when new
objects are written. In that case, get_oid() would potentially say that
it cannot find a given object, even if it should find it.
Let's blow away the loose object cache as well as the read packs and try
again in that case.
Note: this does *not* affect the code path that was introduced to help
avoid looking for the same non-existing objects (which made some
operations really expensive via NFS): that code path is handled by the
`OBJECT_INFO_QUICK` flag (which does not even apply to `get_oid()`,
which has no equivalent flag, at least at the time this patch was
written).
This incidentally fixes the problem identified earlier where an
interactive rebase wanted to re-read (and validate) the todo list after
an `exec` command modified it.
Helped-by: Jeff King <peff@peff.net> Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
rebase -i: demonstrate obscure loose object cache bug
We specifically support `exec` commands in `git rebase -i`'s todo lists
to rewrite the very same todo list. Of course, we need to validate that
todo list when re-reading it.
It is also totally legitimate to extend the todo list by `pick` lines
using short names of commits that were created only after the rebase
started.
And this is where the loose object cache interferes with this feature:
if *some* loose object was read whose hash shares the same first two
digits with a commit that was not yet created when that loose object was
created, then we fail to find that new commit by its short name in
`get_oid()`, and the interactive rebase fails with an obscure error
message like:
error: invalid line 1: pick 6568fef
error: please fix this using 'git rebase --edit-todo'.
Let's first demonstrate that this is actually a bug in a new regression
test, in a separate commit so that other developers who do not believe
me can cherry-pick it to confirm the problem.
This new regression test generates two commits whose hashes share the
first two hex digits (so that their corresponding loose objects live in
the same subdirectory of .git/objects/, and are therefore supposed to be
in the same loose object cache bin).
It then picks the first, to make sure that the loose object cache is
initialized and cached that object directory, then generates the second
commit and picks it, too. Since the commit was generated in a different
process than the sequencer that wants to pick it, the loose object cache
had no chance of being updated in the meantime.
Technically, we would need only one `exec` command in this regression
test case, but for ease of implementation, it uses a pseudo-recursive
call to the same script.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When cloning with --recurse-submodules a superproject with at least one
submodule with HEAD pointing to an unborn branch, the clone goes
something like this:
Cloning into 'test'...
<messages about cloning of superproject>
Submodule '<name>' (<uri>) registered for path '<submodule path>'
Cloning into '<submodule path>'...
fatal: Couldn't find remote ref HEAD
Unable to fetch in submodule path '<submodule path>'
<messages about fetching with SHA-1>
From <uri>
* branch <hash> -> FETCH_HEAD
Submodule path '<submodule path>': checked out '<hash>'
In other words, first, a fetch is done with no hash arguments (that is,
a fetch of HEAD) resulting in a "Couldn't find remote ref HEAD" error;
then, a fetch is done given a hash, which succeeds.
The fetch given a hash was added in fb43e31f2b ("submodule: try harder
to fetch needed sha1 by direct fetching sha1", 2016-02-24), and the
"Unable to fetch..." message was downgraded from a fatal error to a
notice in e30d833671 ("git-submodule.sh: try harder to fetch a
submodule", 2018-05-16).
This commit improves the notice to be clearer that we are retrying the
fetch, and that the previous messages (in particular, the fatal errors
from fetch) do not necessarily indicate that the whole command fails. In
other words:
- If the HEAD-fetch succeeds and we then have the commit we want,
git-submodule prints no explanation.
- If the HEAD-fetch succeeds and we do not have the commit we want, but
the hash-fetch succeeds, git-submodule prints no explanation.
- If the HEAD-fetch succeeds and we do not have the commit we want, but
the hash-fetch fails, git-submodule prints a fatal error.
- If the HEAD-fetch fails, fetch prints a fatal error, and
git-submodule informs the user that it will retry by fetching
specific commits by hash.
- If the hash-fetch then succeeds, git-submodule prints no
explanation (besides the ones already printed).
- If the HEAD-fetch then fails, git-submodule prints a fatal error.
It could be said that we should just eliminate the HEAD-fetch
altogether, but that changes some behavior (in particular, some refs
that were opportunistically updated would no longer be), so I have left
that alone for now.
There is an analogous situation with the fetching code in fetch_finish()
and surrounding functions. For now, I have added a NEEDSWORK.
Signed-off-by: Jonathan Tan <jonathantanmy@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Makefile: fix unaligned loads in sha1dc with UBSan
The sha1dc library uses unaligned loads on platforms that support them.
This is normally what you'd want for performance, but it does cause
UBSan to complain when we compile with SANITIZE=undefined. Just like we
set -DNO_UNALIGNED_LOADS for our own code in that case, we should set
-DSHA1DC_FORCE_ALIGNED_ACCESS.
Of course that does nothing without pulling in the patches from sha1dc
to respect that define. So let's do that, too, updating both the
submodule link and our in-tree copy (from the same commit).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
In commit 735285b403 ("am: fix signoff when other trailers are present",
2017-08-08) tests using variable $signoff were rewritten and it is no
longer used, so just remove it from the test setup.
Signed-off-by: Andrei Rybak <rybak.a.v@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the contributing guide and PR template seen by people who open pull
requests on GitHub, we mention the submitGit tool, which gives an
alternative to figuring out the mailing list. These days we also have
the similar GitGitGadget tool, and we should make it clear that this
is also an option.
We could continue to mention _both_ tools, but it's probably better to
pick one in order to avoid overwhelming the user with choice. After all,
one of the purposes here is to reduce friction for first-time or
infrequent contributors. And there are a few reasons to prefer GGG:
1. submitGit seems to still have a few rough edges. E.g., it doesn't
munge timestamps to help threaded mail readers handled out-of-order
delivery.
2. Subjectively, GGG seems to be more commonly used on the list these
days, especially by list regulars.
3. GGG seems to be under more active development (likely related to
point 2).
So let's actually swap out submitGit for GGG. While we're there, let's
put another link to the GGG page in the PR template, because that's
where users who are learning about it for the first time will want to go
to read more.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
While looking at the inline help for git-subtree.sh, I noticed that
git subtree split --prefix=<prefix> <commit...>
was given as an option. However, it only really makes sense to provide
one revision because of the way the commits are forwarded to rev-parse
so change "<commit...>" to "<commit>" to reflect this. In addition,
check the arguments to ensure that only one rev is provided for all
subcommands that accept a commit.
Signed-off-by: Denton Liu <liu.denton@gmail.com> Acked-by: Avery Pennarun <apenwarr@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
mingw: respect core.hidedotfiles = false in git-init again
This is a brown paper bag. When adding the tests, we actually failed
to verify that the config variable is heeded in git-init at all. And
when changing the original patch that marked the .git/ directory as
hidden after reading the config, it was lost on this developer that
the new code would use the hide_dotfiles variable before the config
was read.
The fix is obvious: read the (limited, pre-init) config *before*
creating the .git/ directory.
Please note that we cannot remove the identical-looking `git_config()`
call from `create_default_files()`: we create the `.git/` directory
between those calls. If we removed it, and if the parent directory is
in a Git worktree, and if that worktree's `.git/config` contained any
`init.templatedir` setting, we would all of a sudden pick that up.
This fixes https://github.com/git-for-windows/git/issues/789
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
We have something much better now: --rebase-merges (which is a
complete re-design --preserve-merges, with a lot of issues fixed such as
the inability to reorder commits with --preserve-merges).
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Git runs a stat loop to find a worktree name that's available and
then does mkdir on the found name. Turn it to mkdir loop to avoid
another invocation of worktree add finding the same free name and
creating the directory first.
Signed-off-by: Michal Suchanek <msuchanek@suse.de> Acked-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
asciidoctor-extensions: fix spurious space after linkgit
When we render, e.g., "linkgit:gitglossary[7]." with Asciidoctor, we get
"gitglossary(7) ." with a space between the linkgit macro expansion and
the punctuation. We can fix this by dropping the trailing newline after
we've turned `linkgit:foo[bar]` into `<citerefentry>..</citerefentry>`.
The diff produced by `USE_ASCIIDOCTOR=Yes ./doc-diff HEAD^ HEAD` is
almost 6000 lines large and shows how this fixes "git-foo(x) ,", "(see
git-bar(y) )" and so on. One might wonder whether this also turns, e.g.,
"see linkgit:foo[1] for more" into "see foo(1)for more", but no. We get
"...</citerefentry> for more" in the XML, see, e.g., git-am.xml, so the
space ends up in git-am.1 just fine. The same is true for the HTML output.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation/Makefile: add missing dependency on asciidoctor-extensions
asciidoctor-extensions.rb has never changed, but when it does -- such as
in the next commit --, it helps if the xml-files depend on it. We're
casting the net a bit too wide here, since we'll be rebuilding even with
AsciiDoc, which won't look at this file. But since this file changes so
rarely, that should be ok. It's better than missing a dependency.
Similarly, most of the html-files are produced directly from ".txt';
make the html-files too depend on asciidoctor-extensions.rb, both
the HTMLified manual pages as well as the user-manual that does use
.xml intermediary.
Signed-off-by: Martin Ågren <martin.agren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
If you use "log -L" with an output format like "--raw" or "--stat",
we'll silently ignore the format and just output the normal patch.
Let's detect and complain about this, which at least tells the user
what's going on.
The tests here aren't exhaustive over the set of all formats, but it
should at least let us know if somebody breaks the format-checking.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>