]> git.ipfire.org Git - thirdparty/git.git/log
thirdparty/git.git
16 months agoMerge branch 'jk/drop-hg-to-git'
Junio C Hamano [Thu, 28 Mar 2024 21:13:51 +0000 (14:13 -0700)] 
Merge branch 'jk/drop-hg-to-git'

Remove an ancient and not well maintained Hg-to-git migration
script from contrib/.

Acked-by: Stelian Pop <stelian@popies.net>
cf. <37e4cd61-b370-437e-bd42-f98f47d3ad32@popies.net>

* jk/drop-hg-to-git:
  contrib: drop hg-to-git script

16 months agoMerge branch 'rs/t-prio-queue-fixes'
Junio C Hamano [Thu, 28 Mar 2024 21:13:51 +0000 (14:13 -0700)] 
Merge branch 'rs/t-prio-queue-fixes'

Test clean-up.

* rs/t-prio-queue-fixes:
  t-prio-queue: check result array bounds
  t-prio-queue: shorten array index message

16 months agoMerge branch 'bt/fuzz-config-parse'
Junio C Hamano [Thu, 28 Mar 2024 21:13:51 +0000 (14:13 -0700)] 
Merge branch 'bt/fuzz-config-parse'

A new fuzz target that exercises config parsing code has been
added.

* bt/fuzz-config-parse:
  fuzz: add fuzzer for config parsing

16 months agoMerge branch 'jc/show-untracked-false'
Junio C Hamano [Thu, 28 Mar 2024 21:13:50 +0000 (14:13 -0700)] 
Merge branch 'jc/show-untracked-false'

The status.showUntrackedFiles configuration variable had a name
that tempts users to set a Boolean value expressed in our usual
"false", "off", and "0", but it only took "no".  This has been
corrected so "true" and its synonyms are taken as "normal", while
"false" and its synonyms are taken as "no".

* jc/show-untracked-false:
  status: allow --untracked=false and friends
  status: unify parsing of --untracked= and status.showUntrackedFiles

16 months agoMerge branch 'ph/diff-src-dst-prefix-config'
Junio C Hamano [Thu, 28 Mar 2024 21:13:50 +0000 (14:13 -0700)] 
Merge branch 'ph/diff-src-dst-prefix-config'

"git diff" and friends learned two extra configuration variables,
diff.srcPrefix and diff.dstPrefix.

* ph/diff-src-dst-prefix-config:
  diff.*Prefix: use camelCase in the doc and test titles
  diff: add diff.srcPrefix and diff.dstPrefix configuration variables

16 months agoMerge branch 'eb/hash-transition'
Junio C Hamano [Thu, 28 Mar 2024 21:13:50 +0000 (14:13 -0700)] 
Merge branch 'eb/hash-transition'

Work to support a repository that work with both SHA-1 and SHA-256
hash algorithms has started.

* eb/hash-transition: (30 commits)
  t1016-compatObjectFormat: add tests to verify the conversion between objects
  t1006: test oid compatibility with cat-file
  t1006: rename sha1 to oid
  test-lib: compute the compatibility hash so tests may use it
  builtin/ls-tree: let the oid determine the output algorithm
  object-file: handle compat objects in check_object_signature
  tree-walk: init_tree_desc take an oid to get the hash algorithm
  builtin/cat-file: let the oid determine the output algorithm
  rev-parse: add an --output-object-format parameter
  repository: implement extensions.compatObjectFormat
  object-file: update object_info_extended to reencode objects
  object-file-convert: convert commits that embed signed tags
  object-file-convert: convert commit objects when writing
  object-file-convert: don't leak when converting tag objects
  object-file-convert: convert tag objects when writing
  object-file-convert: add a function to convert trees between algorithms
  object: factor out parse_mode out of fast-import and tree-walk into in object.h
  cache: add a function to read an OID of a specific algorithm
  tag: sign both hashes
  commit: export add_header_signature to support handling signatures on tags
  ...

16 months agoThe eleventh batch
Junio C Hamano [Mon, 25 Mar 2024 23:16:20 +0000 (16:16 -0700)] 
The eleventh batch

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoMerge branch 'bl/doc-key-val-sep-fix'
Junio C Hamano [Mon, 25 Mar 2024 23:16:35 +0000 (16:16 -0700)] 
Merge branch 'bl/doc-key-val-sep-fix'

The documentation for "%(trailers[:options])" placeholder in the
"--pretty" option of commands in the "git log" family has been
updated.

* bl/doc-key-val-sep-fix:
  docs: adjust trailer `separator` and `key_value_separator` language
  docs: correct trailer `key_value_separator` description

16 months agoMerge branch 'bl/doc-config-fixes'
Junio C Hamano [Mon, 25 Mar 2024 23:16:35 +0000 (16:16 -0700)] 
Merge branch 'bl/doc-config-fixes'

A few typoes in "git config --help" have been corrected.

* bl/doc-config-fixes:
  docs: fix typo in git-config `--default`
  docs: clarify file options in git-config `--edit`

16 months agoMerge branch 'ja/doc-formatting-fix'
Junio C Hamano [Mon, 25 Mar 2024 23:16:34 +0000 (16:16 -0700)] 
Merge branch 'ja/doc-formatting-fix'

Documentation mark-up fix.

* ja/doc-formatting-fix:
  doc: fix some placeholders formating
  doc: format alternatives in synopsis

16 months agoMerge branch 'bb/sh-scripts-cleanup'
Junio C Hamano [Mon, 25 Mar 2024 23:16:34 +0000 (16:16 -0700)] 
Merge branch 'bb/sh-scripts-cleanup'

Shell scripts clean-up.

* bb/sh-scripts-cleanup: (22 commits)
  git-quiltimport: avoid an unnecessary subshell
  contrib/coverage-diff: avoid redundant pipelines
  t/t9*: merge "grep | sed" pipelines
  t/t8*: merge "grep | sed" pipelines
  t/t5*: merge a "grep | sed" pipeline
  t/t4*: merge a "grep | sed" pipeline
  t/t3*: merge a "grep | awk" pipeline
  t/t1*: merge a "grep | sed" pipeline
  t/t9*: avoid redundant uses of cat
  t/t8*: avoid redundant use of cat
  t/t7*: avoid redundant use of cat
  t/t6*: avoid redundant uses of cat
  t/t5*: avoid redundant uses of cat
  t/t4*: avoid redundant uses of cat
  t/t3*: avoid redundant uses of cat
  t/t1*: avoid redundant uses of cat
  t/t0*: avoid redundant uses of cat
  t/perf: avoid redundant use of cat
  t/annotate-tests.sh: avoid redundant use of cat
  t/lib-cvs.sh: avoid redundant use of cat
  ...

16 months agoMerge branch 'jc/index-pack-fsck-levels'
Junio C Hamano [Mon, 25 Mar 2024 23:16:34 +0000 (16:16 -0700)] 
Merge branch 'jc/index-pack-fsck-levels'

Test fix.

* jc/index-pack-fsck-levels:
  t5300: fix test_with_bad_commit()

16 months agoMerge branch 'js/bugreport-no-suffix-fix'
Junio C Hamano [Mon, 25 Mar 2024 23:16:34 +0000 (16:16 -0700)] 
Merge branch 'js/bugreport-no-suffix-fix'

"git bugreport --no-suffix" was not supported and instead
segfaulted, which has been corrected.

* js/bugreport-no-suffix-fix:
  bugreport.c: fix a crash in `git bugreport` with `--no-suffix` option

16 months agoMerge branch 'rj/restore-plug-leaks'
Junio C Hamano [Mon, 25 Mar 2024 23:16:33 +0000 (16:16 -0700)] 
Merge branch 'rj/restore-plug-leaks'

Leaks from "git restore" have been plugged.

* rj/restore-plug-leaks:
  checkout: plug some leaks in git-restore

16 months agoThe tenth batch
Junio C Hamano [Thu, 21 Mar 2024 21:54:56 +0000 (14:54 -0700)] 
The tenth batch

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoMerge branch 'bb/t0006-negative-tz-offset'
Junio C Hamano [Thu, 21 Mar 2024 21:55:14 +0000 (14:55 -0700)] 
Merge branch 'bb/t0006-negative-tz-offset'

More tests on showing time with negative TZ offset.

* bb/t0006-negative-tz-offset:
  t0006: add more tests with a negative TZ offset

16 months agoMerge branch 'bb/iso-strict-utc'
Junio C Hamano [Thu, 21 Mar 2024 21:55:14 +0000 (14:55 -0700)] 
Merge branch 'bb/iso-strict-utc'

The output format for dates "iso-strict" has been tweaked to show
a time in the Zulu timezone with "Z" suffix, instead of "+00:00".

* bb/iso-strict-utc:
  date: make "iso-strict" conforming for the UTC timezone

16 months agoMerge branch 'jw/doc-show-untracked-files-fix'
Junio C Hamano [Thu, 21 Mar 2024 21:55:14 +0000 (14:55 -0700)] 
Merge branch 'jw/doc-show-untracked-files-fix'

The status.showUntrackedFiles configuration variable was
incorrectly documented to accept "false", which has been corrected.

* jw/doc-show-untracked-files-fix:
  doc: status.showUntrackedFiles does not take "false"

16 months agoMerge branch 'dg/user-manual-hash-example'
Junio C Hamano [Thu, 21 Mar 2024 21:55:13 +0000 (14:55 -0700)] 
Merge branch 'dg/user-manual-hash-example'

User manual (the original one) update.

* dg/user-manual-hash-example:
  Documentation/user-manual.txt: example for generating object hashes

16 months agoMerge branch 'ja/doc-markup-fixes'
Junio C Hamano [Thu, 21 Mar 2024 21:55:13 +0000 (14:55 -0700)] 
Merge branch 'ja/doc-markup-fixes'

Mark-ups used in the documentation has been improved for
consistency.

* ja/doc-markup-fixes:
  doc: git-clone: format placeholders
  doc: git-clone: format verbatim words
  doc: git-init: rework config item init.templateDir
  doc: git-init: rework definition lists
  doc: git-init: format placeholders
  doc: git-init: format verbatim parts

16 months agoMerge branch 'pb/ci-win-artifact-names-fix'
Junio C Hamano [Thu, 21 Mar 2024 21:55:13 +0000 (14:55 -0700)] 
Merge branch 'pb/ci-win-artifact-names-fix'

CI update.

* pb/ci-win-artifact-names-fix:
  ci(github): make Windows test artifacts name unique

16 months agoMerge branch 'ps/reftable-reflog-iteration-perf'
Junio C Hamano [Thu, 21 Mar 2024 21:55:13 +0000 (14:55 -0700)] 
Merge branch 'ps/reftable-reflog-iteration-perf'

The code to iterate over reflogs in the reftable has been optimized
to reduce memory allocation and deallocation.

Reviewed-by: Josh Steadmon <steadmon@google.com>
cf. <Ze9eX-aaWoVaqsPP@google.com>

* ps/reftable-reflog-iteration-perf:
  refs/reftable: track last log record name via strbuf
  reftable/record: use scratch buffer when decoding records
  reftable/record: reuse message when decoding log records
  reftable/record: reuse refnames when decoding log records
  reftable/record: avoid copying author info
  reftable/record: convert old and new object IDs to arrays
  refs/reftable: reload correct stack when creating reflog iter

16 months agoMerge branch 'jc/safe-implicit-bare'
Junio C Hamano [Thu, 21 Mar 2024 21:55:13 +0000 (14:55 -0700)] 
Merge branch 'jc/safe-implicit-bare'

Users with safe.bareRepository=explicit can still work from within
$GIT_DIR of a seconary worktree (which resides at .git/worktrees/$name/)
of the primary worktree without explicitly specifying the $GIT_DIR
environment variable or the --git-dir=<path> option.

* jc/safe-implicit-bare:
  setup: notice more types of implicit bare repositories

16 months agoMerge branch 'fs/find-end-of-log-message-fix'
Junio C Hamano [Thu, 21 Mar 2024 21:55:12 +0000 (14:55 -0700)] 
Merge branch 'fs/find-end-of-log-message-fix'

The code to find the effective end of log message can fall into an
endless loop, which has been corrected.

* fs/find-end-of-log-message-fix:
  wt-status: don't find scissors line beyond buf len

16 months agoMerge branch 'ps/reftable-block-search-fix'
Junio C Hamano [Thu, 21 Mar 2024 21:55:12 +0000 (14:55 -0700)] 
Merge branch 'ps/reftable-block-search-fix'

The reftable code has its own custom binary search function whose
comparison callback has an unusual interface, which caused the
binary search to degenerate into a linear search, which has been
corrected.

* ps/reftable-block-search-fix:
  reftable/block: fix binary search over restart counter
  reftable/record: fix memory leak when decoding object records

16 months agoMerge branch 'ps/reftable-stack-tempfile'
Junio C Hamano [Thu, 21 Mar 2024 21:55:12 +0000 (14:55 -0700)] 
Merge branch 'ps/reftable-stack-tempfile'

The code in reftable backend that creates new table files works
better with the tempfile framework to avoid leaving cruft after a
failure.

* ps/reftable-stack-tempfile:
  reftable/stack: register compacted tables as tempfiles
  reftable/stack: register lockfiles during compaction
  reftable/stack: register new tables as tempfiles
  lockfile: report when rollback fails

16 months agoMerge branch 'rs/opt-parse-long-fixups'
Junio C Hamano [Thu, 21 Mar 2024 21:55:12 +0000 (14:55 -0700)] 
Merge branch 'rs/opt-parse-long-fixups'

The parse-options code that deals with abbreviated long option
names have been cleaned up.

Reviewed-by: Josh Steadmon <steadmon@google.com>
cf. <ZfDM5Or3EKw7Q9SA@google.com>

* rs/opt-parse-long-fixups:
  parse-options: rearrange long_name matching code
  parse-options: normalize arg and long_name before comparison
  parse-options: detect ambiguous self-negation
  parse-options: factor out register_abbrev() and struct parsed_option
  parse-options: set arg of abbreviated option lazily
  parse-options: recognize abbreviated negated option with arg

16 months agocontrib: drop hg-to-git script
Jeff King [Wed, 20 Mar 2024 09:48:24 +0000 (05:48 -0400)] 
contrib: drop hg-to-git script

The hg-to-git script is full of command injection vulnerabilities
against malicious branch and tag names. It's also old and largely
unmaintained; the last commit was over 4 years ago, and the last code
change before that was from 2013. Users are better off with a modern
remote-helper tool like cinnabar or remote-hg.

So rather than spending time to fix it, let's just get rid of it.

Reported-by: Matthew Rollings <admin@stealthcopter.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoThe ninth batch
Junio C Hamano [Mon, 18 Mar 2024 20:02:12 +0000 (13:02 -0700)] 
The ninth batch

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoMerge branch 'jk/doc-remote-helpers-markup-fix'
Junio C Hamano [Mon, 18 Mar 2024 20:04:25 +0000 (13:04 -0700)] 
Merge branch 'jk/doc-remote-helpers-markup-fix'

Doc mark-up fix.

* jk/doc-remote-helpers-markup-fix:
  doc/gitremote-helpers: fix missing single-quote

16 months agoMerge branch 'jh/trace2-missing-def-param-fix'
Junio C Hamano [Mon, 18 Mar 2024 20:04:25 +0000 (13:04 -0700)] 
Merge branch 'jh/trace2-missing-def-param-fix'

Some trace2 events that lacked def_param have learned to show it,
enriching the output.

Reviewed-by: Josh Steadmon <steadmon@google.com>
cf. <ZejkVOVQBZhLVfHW@google.com>

* jh/trace2-missing-def-param-fix:
  trace2: emit 'def_param' set with 'cmd_name' event
  trace2: avoid emitting 'def_param' set more than once
  t0211: demonstrate missing 'def_param' events for certain commands

16 months agoMerge branch 'pw/rebase-i-ignore-cherry-pick-help-environment'
Junio C Hamano [Mon, 18 Mar 2024 20:04:25 +0000 (13:04 -0700)] 
Merge branch 'pw/rebase-i-ignore-cherry-pick-help-environment'

Code simplification by getting rid of code that sets an environment
variable that is no longer used.

* pw/rebase-i-ignore-cherry-pick-help-environment:
  rebase -i: stop setting GIT_CHERRY_PICK_HELP

16 months agodocs: adjust trailer `separator` and `key_value_separator` language
Brian Lyles [Mon, 18 Mar 2024 05:38:02 +0000 (00:38 -0500)] 
docs: adjust trailer `separator` and `key_value_separator` language

The language describing the trailer separator and key-value separator
default value is overly complicated.

Indicate the default with simpler "Defaults to ..." language.

Suggested-by: Linus Arver <linusa@google.com>
Signed-off-by: Brian Lyles <brianmlyles@gmail.com>
Acked-by: Linus Arver <linusa@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodocs: correct trailer `key_value_separator` description
Brian Lyles [Mon, 18 Mar 2024 05:38:01 +0000 (00:38 -0500)] 
docs: correct trailer `key_value_separator` description

The description for `key_value_separator` incorrectly states that this
separator is inserted between trailer lines, which appears likely to
have been incorrectly copied from `separator` when this option was
added.

Update the description to correctly indicate that it is a separator that
appears between the key and the value of each trailer.

Signed-off-by: Brian Lyles <brianmlyles@gmail.com>
Acked-by: Linus Arver <linusa@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot-prio-queue: check result array bounds
René Scharfe [Sat, 16 Mar 2024 21:09:47 +0000 (22:09 +0100)] 
t-prio-queue: check result array bounds

Avoid reading past the end of the "result" array, which could otherwise
happen if the prio-queue were to yield more items than were put into it
due to an implementation bug, or if the array has not enough entries due
to a test bug.

Also check at the end whether all "result" entries were consumed, which
would not be the case if the prio-queue forgot some entries or the test
definition contained too many.

Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot-prio-queue: shorten array index message
René Scharfe [Sat, 16 Mar 2024 20:45:51 +0000 (21:45 +0100)] 
t-prio-queue: shorten array index message

If we get an unexpected result, the prio-queue unit test reports it like
this:

 # check "result[j++] == show(get)" failed at t/unit-tests/t-prio-queue.c:43
 #    left: 5
 #   right: 1
 # failed at result[] index 0

That last line repeats "failed" and "result" from the first line.
Shorten it to resemble a similar one in t-ctype and also remove the
incrementation from the first line to avoid possible distractions from
the message of which comparison went wrong where:

 # check "result[j] == show(get)" failed at t/unit-tests/t-prio-queue.c:43
 #    left: 5
 #   right: 1
 #       j: 0

Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodiff.*Prefix: use camelCase in the doc and test titles
Junio C Hamano [Sat, 16 Mar 2024 05:57:22 +0000 (22:57 -0700)] 
diff.*Prefix: use camelCase in the doc and test titles

We added documentation for diff.srcPrefix and diff.dstPrefix with
their names properly camelCased, but the diff.noPrefix is listed
there in all lowercase.  Also these configuration variables, both
existing ones and the {src,dst}Prefix we recently added, were
spelled in all lowercase in the tests in t4013.

Now we are done with the main change, clean these up.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agogit-quiltimport: avoid an unnecessary subshell
Beat Bolli [Fri, 15 Mar 2024 19:46:19 +0000 (20:46 +0100)] 
git-quiltimport: avoid an unnecessary subshell

Use braces for the compound command.

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agocontrib/coverage-diff: avoid redundant pipelines
Beat Bolli [Fri, 15 Mar 2024 19:46:18 +0000 (20:46 +0100)] 
contrib/coverage-diff: avoid redundant pipelines

Merge multiple sed and "grep | awk" invocations, finally use "sort -u"
instead of "sort | uniq".

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t9*: merge "grep | sed" pipelines
Beat Bolli [Fri, 15 Mar 2024 19:46:17 +0000 (20:46 +0100)] 
t/t9*: merge "grep | sed" pipelines

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t8*: merge "grep | sed" pipelines
Beat Bolli [Fri, 15 Mar 2024 19:46:16 +0000 (20:46 +0100)] 
t/t8*: merge "grep | sed" pipelines

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t5*: merge a "grep | sed" pipeline
Beat Bolli [Fri, 15 Mar 2024 19:46:15 +0000 (20:46 +0100)] 
t/t5*: merge a "grep | sed" pipeline

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t4*: merge a "grep | sed" pipeline
Beat Bolli [Fri, 15 Mar 2024 19:46:14 +0000 (20:46 +0100)] 
t/t4*: merge a "grep | sed" pipeline

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t3*: merge a "grep | awk" pipeline
Beat Bolli [Fri, 15 Mar 2024 19:46:13 +0000 (20:46 +0100)] 
t/t3*: merge a "grep | awk" pipeline

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t1*: merge a "grep | sed" pipeline
Beat Bolli [Fri, 15 Mar 2024 19:46:12 +0000 (20:46 +0100)] 
t/t1*: merge a "grep | sed" pipeline

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t9*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:11 +0000 (20:46 +0100)] 
t/t9*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t8*: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:10 +0000 (20:46 +0100)] 
t/t8*: avoid redundant use of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t7*: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:09 +0000 (20:46 +0100)] 
t/t7*: avoid redundant use of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t6*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:08 +0000 (20:46 +0100)] 
t/t6*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t5*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:07 +0000 (20:46 +0100)] 
t/t5*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t4*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:06 +0000 (20:46 +0100)] 
t/t4*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t3*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:05 +0000 (20:46 +0100)] 
t/t3*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t1*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:04 +0000 (20:46 +0100)] 
t/t1*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/t0*: avoid redundant uses of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:03 +0000 (20:46 +0100)] 
t/t0*: avoid redundant uses of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/perf: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:02 +0000 (20:46 +0100)] 
t/perf: avoid redundant use of cat

Take care to redirect stdin, otherwise the output of wc would also contain
the file name.

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/annotate-tests.sh: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:01 +0000 (20:46 +0100)] 
t/annotate-tests.sh: avoid redundant use of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot/lib-cvs.sh: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:46:00 +0000 (20:46 +0100)] 
t/lib-cvs.sh: avoid redundant use of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agocontrib/subtree/t: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:45:59 +0000 (20:45 +0100)] 
contrib/subtree/t: avoid redundant use of cat

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodoc: avoid redundant use of cat
Beat Bolli [Fri, 15 Mar 2024 19:45:58 +0000 (20:45 +0100)] 
doc: avoid redundant use of cat

The update-hook-example.txt script uses this anti-pattern twice. Call grep
with the input file name directy. While at it, merge the two consecutive
grep calls.

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodoc: fix some placeholders formating
Jean-Noël Avila [Sat, 16 Mar 2024 15:21:41 +0000 (15:21 +0000)] 
doc: fix some placeholders formating

Signed-off-by: Jean-Noël Avila <jn.avila@free.fr>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodoc: format alternatives in synopsis
Jean-Noël Avila [Sat, 16 Mar 2024 15:21:40 +0000 (15:21 +0000)] 
doc: format alternatives in synopsis

This is a list of various fixes on malformed alternative in commands
and option syntax.

Signed-off-by: Jean-Noël Avila <jn.avila@free.fr>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodocs: fix typo in git-config `--default`
Brian Lyles [Sat, 16 Mar 2024 05:01:25 +0000 (00:01 -0500)] 
docs: fix typo in git-config `--default`

Signed-off-by: Brian Lyles <brianmlyles@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodocs: clarify file options in git-config `--edit`
Brian Lyles [Sat, 16 Mar 2024 05:01:24 +0000 (00:01 -0500)] 
docs: clarify file options in git-config `--edit`

The description for the `-e`/`--edit` option references scopes
inconsistently: system and global are referenced by their option name
(`--system`/`--global`), but repository (`--local` is not. Additionally,
neither `--worktree` nor `--file` are referenced at all, despite also
being a valid options.

Update the description to mention all four available scopes as well as
`--file`, referencing each consistently by their option name.

Signed-off-by: Brian Lyles <brianmlyles@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agobugreport.c: fix a crash in `git bugreport` with `--no-suffix` option
Jiamu Sun [Thu, 14 Mar 2024 04:00:16 +0000 (04:00 +0000)] 
bugreport.c: fix a crash in `git bugreport` with `--no-suffix` option

`git bugreport` does not complain when `--no-suffix` is given, but
it leads to a segmentation fault as the it is not prepared to see a
NULL assigned to the option_suffix variable.

Signed-off-by: Jiamu Sun <barroit@linux.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoThe eighth batch
Junio C Hamano [Fri, 15 Mar 2024 22:20:53 +0000 (15:20 -0700)] 
The eighth batch

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoMerge branch 'hd/config-mak-os390'
Junio C Hamano [Fri, 15 Mar 2024 23:06:00 +0000 (16:06 -0700)] 
Merge branch 'hd/config-mak-os390'

Platform specific tweaks for OS/390 has been added to
config.mak.uname.

* hd/config-mak-os390:
  build: support z/OS (OS/390).

16 months agoMerge branch 'vm/t7301-use-test-path-helpers'
Junio C Hamano [Fri, 15 Mar 2024 23:06:00 +0000 (16:06 -0700)] 
Merge branch 'vm/t7301-use-test-path-helpers'

GSoC practice to replace "test -f" with "test_path_is_file".

* vm/t7301-use-test-path-helpers:
  t7301: use test_path_is_(missing|file)

16 months agoMerge branch 'jc/xwrite-cleanup'
Junio C Hamano [Fri, 15 Mar 2024 23:06:00 +0000 (16:06 -0700)] 
Merge branch 'jc/xwrite-cleanup'

Uses of xwrite() helper have been audited and updated for better
error checking and simpler code.

* jc/xwrite-cleanup:
  repack: check error writing to pack-objects subprocess
  sideband: avoid short write(2)
  unpack: replace xwrite() loop with write_in_full()

16 months agoMerge branch 'ag/t0010-modernize'
Junio C Hamano [Fri, 15 Mar 2024 23:05:59 +0000 (16:05 -0700)] 
Merge branch 'ag/t0010-modernize'

GSoC practice to modernize a test script.

* ag/t0010-modernize:
  tests: modernize the test script t0010-racy-git.sh

16 months agoMerge branch 'as/option-names-in-messages'
Junio C Hamano [Fri, 15 Mar 2024 23:05:59 +0000 (16:05 -0700)] 
Merge branch 'as/option-names-in-messages'

Error message updates.

* as/option-names-in-messages:
  revision.c: trivial fix to message
  builtin/clone.c: trivial fix of message
  builtin/remote.c: trivial fix of error message
  transport-helper.c: trivial fix of error message

16 months agoMerge branch 'kh/branch-ref-syntax-advice'
Junio C Hamano [Fri, 15 Mar 2024 23:05:59 +0000 (16:05 -0700)] 
Merge branch 'kh/branch-ref-syntax-advice'

When git refuses to create a branch because the proposed branch
name is not a valid refname, an advice message is given to refer
the user to exact naming rules.

* kh/branch-ref-syntax-advice:
  branch: advise about ref syntax rules
  advice: use double quotes for regular quoting
  advice: use backticks for verbatim
  advice: make all entries stylistically consistent
  t3200: improve test style

16 months agofuzz: add fuzzer for config parsing
Brian C Tracy [Fri, 15 Mar 2024 05:47:31 +0000 (05:47 +0000)] 
fuzz: add fuzzer for config parsing

Add a new fuzz target that exercises the parsing of git configs.
The existing git_config_from_mem function is a perfect entry point
for fuzzing as it exercises the same code paths as the rest of the
config parsing functions and offers an easily fuzzable interface.

Config parsing is a useful thing to fuzz because it operates on user
controlled data and is a central component of many git operations.

Signed-off-by: Brian C Tracy <brian.tracy33@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot5300: fix test_with_bad_commit()
John Cai [Fri, 15 Mar 2024 04:57:26 +0000 (04:57 +0000)] 
t5300: fix test_with_bad_commit()

0f8edf7317 (index-pack: --fsck-objects to take an optional argument for
fsck msgs, 2024-02-01) added a test function test_with_bad_commit() that
contained two bugs. test_expect_fail was used instead of test_must_fail,
and a && was not included at the end of the line.

Fix these two issues in the test.

Signed-off-by: John Cai <johncai86@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodiff: add diff.srcPrefix and diff.dstPrefix configuration variables
Peter Hutterer [Fri, 15 Mar 2024 01:03:10 +0000 (11:03 +1000)] 
diff: add diff.srcPrefix and diff.dstPrefix configuration variables

Allow the default prefixes "a/" and "b/" to be tweaked by the
diff.srcPrefix and diff.dstPrefix configuration variables.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoThe seventh batch
Junio C Hamano [Thu, 14 Mar 2024 21:03:22 +0000 (14:03 -0700)] 
The seventh batch

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agoMerge branch 'rj/complete-worktree-paths-fix'
Junio C Hamano [Thu, 14 Mar 2024 21:05:25 +0000 (14:05 -0700)] 
Merge branch 'rj/complete-worktree-paths-fix'

The logic to complete the command line arguments to "git worktree"
subcommand (in contrib/) has been updated to correctly honor things
like "git -C dir" etc.

* rj/complete-worktree-paths-fix:
  completion: fix __git_complete_worktree_paths

16 months agoMerge branch 'rj/complete-reflog'
Junio C Hamano [Thu, 14 Mar 2024 21:05:24 +0000 (14:05 -0700)] 
Merge branch 'rj/complete-reflog'

The command line completion script (in contrib/) learned to
complete "git reflog" better.

* rj/complete-reflog:
  completion: reflog subcommands and options
  completion: factor out __git_resolve_builtins
  completion: introduce __git_find_subcommand
  completion: reflog show <log-options>
  completion: reflog with implicit "show"

16 months agoMerge branch 'jc/test-i18ngrep'
Junio C Hamano [Thu, 14 Mar 2024 21:05:24 +0000 (14:05 -0700)] 
Merge branch 'jc/test-i18ngrep'

With release 2.44 we got rid of all uses of test_i18ngrep and there
is no in-flight topic that adds a new use of it.  Make a call to
test_i18ngrep a hard failure, so that we can remove it at the end
of this release cycle.

* jc/test-i18ngrep:
  test_i18ngrep: hard deprecate and forbid its use

16 months agoMerge branch 'gt/core-bare-in-templates'
Junio C Hamano [Thu, 14 Mar 2024 21:05:24 +0000 (14:05 -0700)] 
Merge branch 'gt/core-bare-in-templates'

Code simplification.

* gt/core-bare-in-templates:
  setup: remove unnecessary variable

16 months agoMerge branch 'la/trailer-api'
Junio C Hamano [Thu, 14 Mar 2024 21:05:24 +0000 (14:05 -0700)] 
Merge branch 'la/trailer-api'

Trailer API updates.

Acked-by: Christian Couder <christian.couder@gmail.com>
cf. <CAP8UFD1Zd+9q0z1JmfOf60S2vn5-sD3SafDvAJUzRFwHJKcb8A@mail.gmail.com>

* la/trailer-api:
  format_trailers_from_commit(): indirectly call trailer_info_get()
  format_trailer_info(): move "fast path" to caller
  format_trailers(): use strbuf instead of FILE
  trailer_info_get(): reorder parameters
  trailer: move interpret_trailers() to interpret-trailers.c
  trailer: reorder format_trailers_from_commit() parameters
  trailer: rename functions to use 'trailer'
  shortlog: add test for de-duplicating folded trailers
  trailer: free trailer_info _after_ all related usage

16 months agoMerge branch 'kh/doc-commentchar-is-a-byte'
Junio C Hamano [Thu, 14 Mar 2024 21:05:23 +0000 (14:05 -0700)] 
Merge branch 'kh/doc-commentchar-is-a-byte'

The "core.commentChar" configuration variable only allows an ASCII
character, which was not clearly documented, which has been
corrected.

* kh/doc-commentchar-is-a-byte:
  config: document `core.commentChar` as ASCII-only

16 months agoMerge branch 'jh/fsmonitor-icase-corner-case-fix'
Junio C Hamano [Thu, 14 Mar 2024 21:05:23 +0000 (14:05 -0700)] 
Merge branch 'jh/fsmonitor-icase-corner-case-fix'

FSMonitor client code was confused when FSEvents were given in a
different case on a case-insensitive filesystem, which has been
corrected.

Acked-by: Patrick Steinhardt <ps@pks.im>
cf. <ZehofMaSZyUq8S1N@tanuki>

* jh/fsmonitor-icase-corner-case-fix:
  fsmonitor: support case-insensitive events
  fsmonitor: refactor bit invalidation in refresh callback
  fsmonitor: trace the new invalidated cache-entry count
  fsmonitor: return invalidated cache-entry count on non-directory event
  fsmonitor: remove custom loop from non-directory path handler
  fsmonitor: return invalidated cache-entry count on directory event
  fsmonitor: move untracked-cache invalidation into helper functions
  fsmonitor: refactor untracked-cache invalidation
  dir: create untracked_cache_invalidate_trimmed_path()
  fsmonitor: refactor refresh callback for non-directory events
  fsmonitor: clarify handling of directory events in callback helper
  fsmonitor: refactor refresh callback on directory events
  t7527: add case-insensitve test for FSMonitor
  name-hash: add index_dir_find()

16 months agoMerge branch 'ps/reftable-iteration-perf-part2'
Junio C Hamano [Thu, 14 Mar 2024 21:05:23 +0000 (14:05 -0700)] 
Merge branch 'ps/reftable-iteration-perf-part2'

The code to iterate over refs with the reftable backend has seen
some optimization.

* ps/reftable-iteration-perf-part2:
  refs/reftable: precompute prefix length
  reftable: allow inlining of a few functions
  reftable/record: decode keys in place
  reftable/record: reuse refname when copying
  reftable/record: reuse refname when decoding
  reftable/merged: avoid duplicate pqueue emptiness check
  reftable/merged: circumvent pqueue with single subiter
  reftable/merged: handle subiter cleanup on close only
  reftable/merged: remove unnecessary null check for subiters
  reftable/merged: make subiters own their records
  reftable/merged: advance subiter on subsequent iteration
  reftable/merged: make `merged_iter` structure private
  reftable/pq: use `size_t` to track iterator index

16 months agoMerge branch 'so/clean-dry-run-without-force'
Junio C Hamano [Thu, 14 Mar 2024 21:05:23 +0000 (14:05 -0700)] 
Merge branch 'so/clean-dry-run-without-force'

The implementation in "git clean" that makes "-n" and "-i" ignore
clean.requireForce has been simplified, together with the
documentation.

* so/clean-dry-run-without-force:
  clean: further clean-up of implementation around "--force"
  clean: improve -n and -f implementation and documentation

16 months agocheckout: plug some leaks in git-restore
Rubén Justo [Thu, 14 Mar 2024 18:08:58 +0000 (19:08 +0100)] 
checkout: plug some leaks in git-restore

In git-restore we need to free the pathspec and pathspec_from_file
values from the struct checkout_opts.

A simple fix could be to free them in cmd_restore, after the call to
checkout_main returns, like we are doing [1][2] in the sibling function
cmd_checkout.

However, we can do even better.

We have git-switch and git-restore, both of them spin-offs[3][4] of
git-checkout.  All three are implemented as thin wrappers around
checkout_main.  Considering this, it makes a lot of sense to do the
cleanup closer to checkout_main.

Move the cleanups, including the new_branch_info variable, to
checkout_main.

As a consequence, mark: t2070, t2071, t2072 and t6418 as leak-free.

 [1] 9081a421a6 (checkout: fix "branch info" memory leaks, 2021-11-16)

 [2] 7ce4088ab7 (parse-options: consistently allocate memory in
     fix_filename(), 2023-03-04)

 [3] d787d311db (checkout: split part of it to new command 'switch',
     2019-03-29)

 [4] 46e91b663b (checkout: split part of it to new command 'restore',
     2019-04-25)

Signed-off-by: Rubén Justo <rjusto@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agot0006: add more tests with a negative TZ offset
Beat Bolli [Thu, 14 Mar 2024 08:55:12 +0000 (09:55 +0100)] 
t0006: add more tests with a negative TZ offset

This test doesn't systematically check a negative timezone offset. Add a
test for each format that outputs the offset to improve our test
coverage.

Signed-off-by: Beat Bolli <dev+git@drbeat.li>
16 months agodate: make "iso-strict" conforming for the UTC timezone
Beat Bolli [Wed, 13 Mar 2024 22:54:23 +0000 (23:54 +0100)] 
date: make "iso-strict" conforming for the UTC timezone

ISO 8601-1:2020-12 specifies that a zero timezone offset must be denoted
with a "Z" suffix instead of the numeric "+00:00". Add the correponding
special case to show_date() and a new test.

Changing an established output format which might be depended on by
scripts is always problematic, but here we choose to adhere more closely
to the published standard.

Reported-by: Michael Osipov <michael.osipov@innomotics.com>
Signed-off-by: Beat Bolli <dev+git@drbeat.li>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agostatus: allow --untracked=false and friends
Junio C Hamano [Wed, 13 Mar 2024 17:32:14 +0000 (10:32 -0700)] 
status: allow --untracked=false and friends

It is natural to expect that the "--untracked" option and the
status.showuntrackedFiles configuration variable to take a Boolean
value ("do you want me to show untracked files?"), but the current
code takes nothing but "no" as "no, please do not show any".

Allow the usual Boolean values to be given, and treat 'true' as
"normal", and 'false' as "no".

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agostatus: unify parsing of --untracked= and status.showUntrackedFiles
Junio C Hamano [Wed, 13 Mar 2024 17:32:13 +0000 (10:32 -0700)] 
status: unify parsing of --untracked= and status.showUntrackedFiles

There are two code paths that take a string and parse it to enum
untracked_status_type.  Introduce a helper function and use it.

As these two places handle an error differently, add an additional
invalid value to the enum, and have the caller of the helper handle
the error condition, instead of dying or emitting error message from
the helper.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
16 months agodoc: status.showUntrackedFiles does not take "false"
Jonas Wunderlich [Tue, 12 Mar 2024 21:34:11 +0000 (21:34 +0000)] 
doc: status.showUntrackedFiles does not take "false"

The `status.showUntrackedFiles` config option only accepts the
values "no", "normal" or "all", but not as this part of the man page
suggested "false".  While we are at it, camel-case the name of the
variable.

Signed-off-by: Jonas Wunderlich <git@03j.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
17 months agoDocumentation/user-manual.txt: example for generating object hashes
Dirk Gouders [Tue, 12 Mar 2024 10:41:56 +0000 (11:41 +0100)] 
Documentation/user-manual.txt: example for generating object hashes

Add a simple example on how object hashes can be generated manually.

Further, because the document suggests to have a look at the initial
commit, clarify that some details changed since that time.

Signed-off-by: Dirk Gouders <dirk@gouders.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
17 months agoThe sixth batch
Junio C Hamano [Mon, 11 Mar 2024 21:11:28 +0000 (14:11 -0700)] 
The sixth batch

Signed-off-by: Junio C Hamano <gitster@pobox.com>
17 months agoMerge branch 'sj/t9117-path-is-file'
Junio C Hamano [Mon, 11 Mar 2024 21:12:31 +0000 (14:12 -0700)] 
Merge branch 'sj/t9117-path-is-file'

GSoC practice to replace "test -f" with "test_path_is_file".

* sj/t9117-path-is-file:
  t9117: prefer test_path_* helper functions

17 months agoMerge branch 'kh/doc-dashed-commands-have-not-worked-for-a-long-time'
Junio C Hamano [Mon, 11 Mar 2024 21:12:31 +0000 (14:12 -0700)] 
Merge branch 'kh/doc-dashed-commands-have-not-worked-for-a-long-time'

Doc update.

* kh/doc-dashed-commands-have-not-worked-for-a-long-time:
  gitcli: drop mention of “non-dashed form”

17 months agoMerge branch 'rs/t-ctype-simplify'
Junio C Hamano [Mon, 11 Mar 2024 21:12:30 +0000 (14:12 -0700)] 
Merge branch 'rs/t-ctype-simplify'

Code simplification to one unit-test program.

* rs/t-ctype-simplify:
  t-ctype: avoid duplicating class names
  t-ctype: align output of i
  t-ctype: simplify EOF check
  t-ctype: allow NUL anywhere in the specification string

17 months agoMerge branch 'es/config-doc-sort-sections'
Junio C Hamano [Mon, 11 Mar 2024 21:12:30 +0000 (14:12 -0700)] 
Merge branch 'es/config-doc-sort-sections'

Doc updates.

* es/config-doc-sort-sections:
  docs: sort configuration variable groupings alphabetically

17 months agoMerge branch 'js/merge-base-with-missing-commit'
Junio C Hamano [Mon, 11 Mar 2024 21:12:30 +0000 (14:12 -0700)] 
Merge branch 'js/merge-base-with-missing-commit'

Make sure failure return from merge_bases_many() is properly caught.

* js/merge-base-with-missing-commit:
  merge-ort/merge-recursive: do report errors in `merge_submodule()`
  merge-recursive: prepare for `merge_submodule()` to report errors
  commit-reach(repo_get_merge_bases_many_dirty): pass on errors
  commit-reach(repo_get_merge_bases_many): pass on "missing commits" errors
  commit-reach(get_octopus_merge_bases): pass on "missing commits" errors
  commit-reach(repo_get_merge_bases): pass on "missing commits" errors
  commit-reach(get_merge_bases_many_0): pass on "missing commits" errors
  commit-reach(merge_bases_many): pass on "missing commits" errors
  commit-reach(paint_down_to_common): start reporting errors
  commit-reach(paint_down_to_common): prepare for handling shallow commits
  commit-reach(repo_in_merge_bases_many): report missing commits
  commit-reach(repo_in_merge_bases_many): optionally expect missing commits
  commit-reach(paint_down_to_common): plug two memory leaks

17 months agosetup: notice more types of implicit bare repositories
Junio C Hamano [Sat, 9 Mar 2024 23:27:09 +0000 (15:27 -0800)] 
setup: notice more types of implicit bare repositories

Setting the safe.bareRepository configuration variable to explicit
stops git from using a bare repository, unless the repository is
explicitly specified, either by the "--git-dir=<path>" command line
option, or by exporting $GIT_DIR environment variable.  This may be
a reasonable measure to safeguard users from accidentally straying
into a bare repository in unexpected places, but often gets in the
way of users who need valid accesses to the repository.

Earlier, 45bb9162 (setup: allow cwd=.git w/ bareRepository=explicit,
2024-01-20) loosened the rule such that being inside the ".git"
directory of a non-bare repository does not really count as
accessing a "bare" repository.  The reason why such a loosening is
needed is because often hooks and third-party tools run from within
$GIT_DIR while working with a non-bare repository.

More importantly, the reason why this is safe is because a directory
whose contents look like that of a "bare" repository cannot be a
bare repository that came embedded within a checkout of a malicious
project, as long as its directory name is ".git", because ".git" is
not a name allowed for a directory in payload.

There are at least two other cases where tools have to work in a
bare-repository looking directory that is not an embedded bare
repository, and accesses to them are still not allowed by the recent
change.

 - A secondary worktree (whose name is $name) has its $GIT_DIR
   inside "worktrees/$name/" subdirectory of the $GIT_DIR of the
   primary worktree of the same repository.

 - A submodule worktree (whose name is $name) has its $GIT_DIR
   inside "modules/$name/" subdirectory of the $GIT_DIR of its
   superproject.

As long as the primary worktree or the superproject in these cases
are not bare, the pathname of these "looks like bare but not really"
directories will have "/.git/worktrees/" and "/.git/modules/" as a
substring in its leading part, and we can take advantage of the same
security guarantee allow git to work from these places.

Extend the earlier "in a directory called '.git' we are OK" logic
used for the primary worktree to also cover the secondary worktree's
and non-embedded submodule's $GIT_DIR, by moving the logic to a
helper function "is_implicit_bare_repo()".  We deliberately exclude
secondary worktrees and submodules of a bare repository, as these
are exactly what safe.bareRepository=explicit setting is designed to
forbid accesses to without an explicit GIT_DIR/--git-dir=<path>

Helped-by: Kyle Lippincott <spectral@google.com>
Helped-by: Kyle Meyer <kyle@kyleam.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
17 months agoci(github): make Windows test artifacts name unique
Philippe Blain [Sun, 10 Mar 2024 20:04:56 +0000 (20:04 +0000)] 
ci(github): make Windows test artifacts name unique

If several jobs in the windows-test or vs-test matrices fail, the
upload-artifact action in each job tries to upload the test directories
of the failed tests as "failed-tests-windows.zip", which fails for all
jobs except the one which finishes first with the following error:

    Error: Failed to CreateArtifact: Received non-retryable error:
    Failed request: (409) Conflict: an artifact with this name
    already exists on the workflow run

Make the artifacts name unique by using the 'matrix.nr' token, and
disambiguate the vs-test artifacts from the windows-test ones.

Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
17 months agodoc: git-clone: format placeholders
Jean-Noël Avila [Sun, 10 Mar 2024 19:10:30 +0000 (19:10 +0000)] 
doc: git-clone: format placeholders

With the new formatting rules, we use _<placeholders>_.

Signed-off-by: Jean-Noël Avila <jn.avila@free.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>