Git 2.29 Release Notes
======================
Updates since v2.28
-------------------
UI, Workflows & Features
* "git help log" has been enhanced by sharing more material from the
documentation for the underlying "git rev-list" command.
* "git for-each-ref --format=<>" learned %(contents:size).
* "git merge" learned to selectively omit " into <branch>" at the end
of the title of default merge message with merge.suppressDest
configuration.
* The component to respond to "git fetch" request is made more
configurable to selectively allow or reject object filtering
specification used for partial cloning.
* Stop when "sendmail.*" configuration variables are defined, which
could be a mistaken attempt to define "sendemail.*" variables.
* The existing backends for "git mergetool" based on variants of vim
have been refactored and then support for "nvim" has been added.
* "git bisect" learns the "--first-parent" option to find the first
breakage along the first-parent chain.
* "git log --first-parent -p" showed patches only for single-parent
commits on the first-parent chain; the "--first-parent" option has
been made to imply "-m". Use "--no-diff-merges" to restore the
previous behaviour to omit patches for merge commits.
* The commit labels used to explain each side of conflicted hunks
placed by the sequencer machinery have been made more readable by
humans.
* The "--batch-size" option of "git multi-pack-index repack" command
is now used to specify that very small packfiles are collected into
one until the total size roughly exceeds it.
* The recent addition of SHA-256 support is marked as experimental in
the documentation.
* "git fetch" learned --no-write-fetch-head option to avoid writing
the FETCH_HEAD file.
* Command line completion (in contrib/) usually omits redundant,
deprecated and/or dangerous options from its output; it learned to
optionally include all of them.
* The output from the "diff" family of the commands had abbreviated
object names of blobs involved in the patch, but its length was not
affected by the --abbrev option. Now it is.
* "git worktree" gained a "repair" subcommand to help users recover
after moving the worktrees or repository manually without telling
Git. Also, "git init --separate-git-dir" no longer corrupts
administrative data related to linked worktrees.
* The "--format=" option to the "for-each-ref" command and friends
learned a few more tricks, e.g. the ":short" suffix that applies to
"objectname" now also can be used for "parent", "tree", etc.
* "git worktree add" learns that the "-d" is a synonym to "--detach"
option to create a new worktree without being on a branch.
* "format-patch --range-diff=<prev> <origin>..HEAD" has been taught
not to ignore <origin> when <prev> is a single version.
* "add -p" now allows editing paths that were only added in intent.
* The 'meld' backend of the "git mergetool" learned to give the
underlying 'meld' the '--auto-merge' option, which would help
reduce the amount of text that requires manual merging.
* "git for-each-ref" and friends that list refs used to allow only
one --merged or --no-merged to filter them; they learned to take
combination of both kind of filtering.
* "git maintenance", a "git gc"'s big brother, has been introduced to
take care of more repository maintenance tasks, not limited to the
object database cleaning.
* "git receive-pack" that accepts requests by "git push" learned to
outsource most of the ref updates to the new "proc-receive" hook.
* "git push" that wants to be atomic and wants to send push
certificate learned not to prepare and sign the push certificate
when it fails the local check (hence due to atomicity it is known
that no certificate is needed).
* "git commit-graph write" learned to limit the number of bloom
filters that are computed from scratch with the --max-new-filters
option.
* The transport protocol v2 has become the default again.
* The installation procedure learned to optionally omit "git-foo"
executable files for each 'foo' built-in subcommand, which are only
required by old timers that still rely on the age old promise that
prepending "git --exec-path" output to PATH early in their script
will keep the "git-foo" calls they wrote working.
* The command line completion (in contrib/) learned that "git restore
-s <TAB>" is often followed by a refname.
* "git shortlog" has been taught to group commits by the contents of
the trailer lines, like "Reviewed-by:", "Coauthored-by:", etc.
* "git archive" learns the "--add-file" option to include untracked
files into a snapshot from a tree-ish.
* "git fetch" and "git push" support negative refspecs.
* "git format-patch" learns to take "whenAble" as a possible value
for the format.useAutoBase configuration variable to become no-op
when the automatically computed base does not make sense.
* Credential helpers are now allowed to terminate lines with CRLF
line ending, as well as LF line ending.
Performance, Internal Implementation, Development Support etc.
* The changed-path Bloom filter is improved using ideas from an
independent implementation.
* Updates to the changed-paths bloom filter.
* The test framework has been updated so that most tests will run
with predictable (artificial) timestamps.
* Preliminary clean-up of the refs API in preparation for adding a
new refs backend "reftable".
* Dev support to limit the use of test_must_fail to only git commands.
* While packing many objects in a repository with a promissor remote,
lazily fetching missing objects from the promissor remote one by
one may be inefficient---the code now attempts to fetch all the
missing objects in batch (obviously this won't work for a lazy
clone that lazily fetches tree objects as you cannot even enumerate
what blobs are missing until you learn which trees are missing).
* The pretend-object mechanism checks if the given object already
exists in the object store before deciding to keep the data
in-core, but the check would have triggered lazy fetching of such
an object from a promissor remote.
* The argv_array API is useful for not just managing argv but any
"vector" (NULL-terminated array) of strings, and has seen adoption
to a certain degree. It has been renamed to "strvec" to reduce the
barrier to adoption.
* The final leg of SHA-256 transition plus doc updates. Note that
there is no interoperability between SHA-1 and SHA-256
repositories yet.
* CMake support to build with MSVC for Windows bypassing the Makefile.
* A new helper function has_object() has been introduced to make it
easier to mark object existence checks that do and don't want to
trigger lazy fetches, and a few such checks are converted using it.
* A no-op replacement function implemented as a C preprocessor macro
does not perform as good a job as one implemented as a "static
inline" function in catching errors in parameters; replace the
former with the latter in <git-compat-util.h> header.
* Test framework update.
(merge d572f52a64 es/test-cmp-typocatcher later to maint).
* Updates to "git merge" tests, in preparation for a new merge
strategy backend.
* midx and commit-graph files now use the byte defined in their file
format specification for identifying the hash function used for
object names.
* The FETCH_HEAD is now always read from the filesystem regardless of
the ref backend in use, as its format is much richer than the
normal refs, and written directly by "git fetch" as a plain file..
* An unused binary has been discarded, and a bunch of commands
have been turned into built-in.
* A handful of places in in-tree code still relied on being able to
execute the git subcommands, especially built-ins, in "git-foo"
form, which have been corrected.
* When a packfile is removed by "git repack", multi-pack-index gets
cleared; the code was taught to do so less aggressively by first
checking if the midx actually refers to a pack that no longer
exists.
* Internal API clean-up to handle two options "diff-index" and "log"
have, which happen to share the same short form, more sensibly.
* The "add -i/-p" machinery has been written in C but it is not used
by default yet. It is made default to those who are participating
in feature.experimental experiment.
* Allow maintainers to tweak $(TAR) invocations done while making
distribution tarballs.
* "git index-pack" learned to resolve deltified objects with greater
parallelism.
* "diff-highlight" (in contrib/) had a logic to flush its output upon
seeing a blank line but the way it detected a blank line was broken.
* The logic to skip testing on the tagged commit and the tag itself
was not quite consistent which led to failure of Windows test
tasks. It has been revamped to consistently skip revisions that
have already been tested, based on the tree object of the revision.
Fixes since v2.28
-----------------
* The "mediawiki" remote backend which lives in contrib/mw-to-git/
and is not built with git by default, had an RCE bug allowing a
malicious MediaWiki server operator to inject arbitrary commands
for execution by a cloning client. This has been fixed.
The bug was discovered and reported by Joern Schneeweisz of GitLab
to the git-security mailing list. Its practical impact due to the
obscurity of git-remote-mediawiki was deemed small enough to forgo
a dedicated security release.
* "git clone --separate-git-dir=$elsewhere" used to stomp on the
contents of the existing directory $elsewhere, which has been
taught to fail when $elsewhere is not an empty directory.
(merge dfaa209a79 bw/fail-cloning-into-non-empty later to maint).
* With the base fix to 2.27 regresion, any new extensions in a v0
repository would still be silently honored, which is not quite
right. Instead, complain and die loudly.
(merge ec91ffca04 jk/reject-newer-extensions-in-v0 later to maint).
* Fetching from a lazily cloned repository resulted at the server
side in attempts to lazy fetch objects that the client side has,
many of which will not be available from the third-party anyway.
(merge 77aa0941ce jt/avoid-lazy-fetching-upon-have-check later to maint).
* Fix to an ancient bug caused by an over-eager attempt for
optimization.
(merge a98f7fb366 rs/add-index-entry-optim-fix later to maint).
* Pushing a ref whose name contains non-ASCII character with the
"--force-with-lease" option did not work over smart HTTP protocol,
which has been corrected.
(merge cd85b447bf bc/push-cas-cquoted-refname later to maint).
* "git mv src dst", when src is an unmerged path, errored out
correctly but with an incorrect error message to claim that src is
not tracked, which has been clarified.
(merge 9b906af657 ct/mv-unmerged-path-error later to maint).
* Fix to a regression introduced during 2.27 cycle.
(merge cada7308ad en/fill-directory-exponential later to maint).
* Command line completion (in contrib/) update.
(merge 688b87c81b mp/complete-show-color-moved later to maint).
* All "mergy" operations that internally use the merge-recursive
machinery should honor the merge.renormalize configuration, but
many of them didn't.
* Doc cleanup around "worktree".
(merge dc9c144be5 es/worktree-doc-cleanups later to maint).
* The "git blame --first-parent" option was not documented, but now
it is.
(merge 11bc12ae1e rp/blame-first-parent-doc later to maint).
* The logic to find the ref transaction hook script attempted to
cache the path to the found hook without realizing that it needed
to keep a copied value, as the API it used returned a transitory
buffer space. This has been corrected.
(merge 09b2aa30c9 ps/ref-transaction-hook later to maint).
* Recent versions of "git diff-files" shows a diff between the index
and the working tree for "intent-to-add" paths as a "new file"
patch; "git apply --cached" should be able to take "git diff-files"
and should act as an equivalent to "git add" for the path, but the
command failed to do so for such a path.
(merge 4c025c667e rp/apply-cached-with-i-t-a later to maint).
* "git diff [<tree-ish>] $path" for a $path that is marked with i-t-a
bit was not showing the mode bits from the working tree.
(merge cb0dd22b82 rp/ita-diff-modefix later to maint).
* Ring buffer with size 4 used for bin-hex translation resulted in a
wrong object name in the sequencer's todo output, which has been
corrected.
(merge 5da69c0dac ak/sequencer-fix-find-uniq-abbrev later to maint).
* When given more than one target line ranges, "git blame -La,b
-Lc,d" was over-eager to coalesce groups of original lines and
showed incorrect results, which has been corrected.
(merge c2ebaa27d6 jk/blame-coalesce-fix later to maint).
* The regexp to identify the function boundary for FORTRAN programs
has been updated.
(merge 75c3b6b2e8 pb/userdiff-fortran-update later to maint).
* A few end-user facing messages have been updated to be
hash-algorithm agnostic.
(merge 4279000d3e jc/object-names-are-not-sha-1 later to maint).
* "unlink" emulation on MinGW has been optimized.
(merge 680e0b4524 jh/mingw-unlink later to maint).
* The purpose of "git init --separate-git-dir" is to initialize a
new project with the repository separate from the working tree,
or, in the case of an existing project, to move the repository
(the .git/ directory) out of the working tree. It does not make
sense to use --separate-git-dir with a bare repository for which
there is no working tree, so disallow its use with bare
repositories.
(merge ccf236a23a es/init-no-separate-git-dir-in-bare later to maint).
* "ls-files -o" mishandled the top-level directory of another git
working tree that hangs in the current git working tree.
(merge ab282aa548 en/dir-nonbare-embedded later to maint).
* Fix some incorrect UNLEAK() annotations.
(merge 3e19816dc0 jk/unleak-fixes later to maint).
* Use more buffered I/O where we used to call many small write(2)s.
(merge a698d67b08 rs/more-buffered-io later to maint).
* The patch-id computation did not ignore the "incomplete last line"
marker like whitespaces.
(merge 82a62015a7 rs/patch-id-with-incomplete-line later to maint).
* Updates into a lazy/partial clone with a submodule did not work
well with transfer.fsckobjects set.
* The parser for "git for-each-ref --format=..." was too loose when
parsing the "%(trailers...)" atom, and forgot that "trailers" and
"trailers:<modifiers>" are the only two allowed forms, which has
been corrected.
(merge 2c22e102f8 hv/ref-filter-trailers-atom-parsing-fix later to maint).
* Long ago, we decided to use 3 threads by default when running the
index-pack task in parallel, which has been adjusted a bit upwards.
(merge fbff95b67f jk/index-pack-w-more-threads later to maint).
* "git restore/checkout --no-overlay" with wildcarded pathspec
mistakenly removed matching paths in subdirectories, which has been
corrected.
(merge bfda204ade rs/checkout-no-overlay-pathspec-fix later to maint).
* The description of --cached/--index options in "git apply --help"
has been updated.
(merge d064702be3 rp/apply-cached-doc later to maint).
* Feeding "$ZERO_OID" to "git log --ignore-missing --stdin", and
running "git log --ignore-missing $ZERO_OID" fell back to start
digging from HEAD; it has been corrected to become a no-op, like
"git log --tags=no-tag-matches-this-pattern" does.
(merge 04a0e98515 jk/rev-input-given-fix later to maint).
* Various callers of run_command API have been modernized.
(merge afbdba391e jc/run-command-use-embedded-args later to maint).
* List of options offered and accepted by "git add -i/-p" were
inconsistent, which have been corrected.
(merge ce910287e7 pw/add-p-allowed-options-fix later to maint).
* "git diff --stat -w" showed 0-line changes for paths whose changes
were only whitespaces, which was not intuitive. We now omit such
paths from the stat output.
(merge 1cf3d5db9b mr/diff-hide-stat-wo-textual-change later to maint).
* It was possible for xrealloc() to send a non-NULL pointer that has
been freed, which has been fixed.
(merge 6479ea4a8a jk/xrealloc-avoid-use-after-free later to maint).
* "git status" has trouble showing where it came from by interpreting
reflog entries that record certain events, e.g. "checkout @{u}", and
gives a hard/fatal error. Even though it inherently is impossible
to give a correct answer because the reflog entries lose some
information (e.g. "@{u}" does not record what branch the user was
on hence which branch 'the upstream' needs to be computed, and even
if the record were available, the relationship between branches may
have changed), at least hide the error and allow "status" to show its
output.
* "git status --short" quoted a path with SP in it when tracked, but
not those that are untracked, ignored or unmerged. They are all
shown quoted consistently.
* "git diff/show" on a change that involves a submodule used to read
the information on commits in the submodule from a wrong repository
and gave a wrong information when the commit-graph is involved.
(merge 85a1ec2c32 mf/submodule-summary-with-correct-repository later to maint).
* Unlike "git config --local", "git config --worktree" did not fail
early and cleanly when started outside a git repository.
(merge 378fe5fc3d mt/config-fail-nongit-early later to maint).
* There is a logic to estimate how many objects are in the
repository, which is meant to run once per process invocation, but
it ran every time the estimated value was requested.
(merge 67bb65de5d jk/dont-count-existing-objects-twice later to maint).
* "git remote set-head" that failed still said something that hints
the operation went through, which was misleading.
(merge 5a07c6c3c2 cs/don-t-pretend-a-failed-remote-set-head-succeeded later to maint).
* "git fetch --all --ipv4/--ipv6" forgot to pass the protocol options
to instances of the "git fetch" that talk to individual remotes,
which has been corrected.
(merge 4e735c1326 ar/fetch-ipversion-in-all later to maint).
* The "unshelve" subcommand of "git p4" incorrectly used commit^N
where it meant to say commit~N to name the Nth generation
ancestor, which has been corrected.
(merge 0acbf5997f ld/p4-unshelve-fix later to maint).
* "git clone" that clones from SHA-1 repository, while
GIT_DEFAULT_HASH set to use SHA-256 already, resulted in an
unusable repository that half-claims to be SHA-256 repository
with SHA-1 objects and refs. This has been corrected.
* Adjust sample hooks for hash algorithm other than SHA-1.
(merge d8d3d632f4 dl/zero-oid-in-hooks later to maint).
* "git range-diff" showed incorrect diffstat, which has been
corrected.
* Earlier we taught "git pull" to warn when the user does not say the
histories need to be merged, rebased or accepts only fast-
forwarding, but the warning triggered for those who have set the
pull.ff configuration variable.
(merge 54200cef86 ah/pull later to maint).
* Compilation fix around type punning.
(merge 176380fd11 jk/drop-unaligned-loads later to maint).
* "git blame --ignore-rev/--ignore-revs-file" failed to validate
their input are valid revision, and failed to take into account
that the user may want to give an annotated tag instead of a
commit, which has been corrected.
(merge 610e2b9240 jc/blame-ignore-fix later to maint).
* "git bisect start X Y", when X and Y are not valid committish
object names, should take X and Y as pathspec, but didn't.
(merge 73c6de06af cc/bisect-start-fix later to maint).
* The explanation of the "scissors line" has been clarified.
(merge 287416dba6 eg/mailinfo-doc-scissors later to maint).
* A race that leads to an access to a free'd data was corrected in
the codepath that reads pack files.
(merge bda959c476 mt/delta-base-cache-races later to maint).
* in_merge_bases_many(), a way to see if a commit is reachable from
any commit in a set of commits, was totally broken when the
commit-graph feature was in use, which has been corrected.
(merge 8791bf1841 ds/in-merge-bases-many-optim-bug later to maint).
* "git submodule update --quiet" did not squelch underlying "rebase"
and "pull" commands.
(merge 3ad0401e9e td/submodule-update-quiet later to maint).
* The lazy fetching done internally to make missing objects available
in a partial clone incorrectly made permanent damage to the partial
clone filter in the repository, which has been corrected.
* "log -c --find-object=X" did not work well to find a merge that
involves a change to an object X from only one parent.
(merge 957876f17d jk/diff-cc-oidfind-fix later to maint).
* Other code cleanup, docfix, build fix, etc.
(merge 84544f2ea3 sk/typofixes later to maint).
(merge b17f411ab5 ar/help-guides-doc later to maint).
(merge 98c6871fad rs/grep-simpler-parse-object-or-die-call later to maint).
(merge 861c4ce141 en/typofixes later to maint).
(merge 60e47f6773 sg/ci-git-path-fix-with-pyenv later to maint).
(merge e2bfa50ac3 jb/doc-packfile-name later to maint).
(merge 918d8ff780 es/worktree-cleanup later to maint).
(merge dc156bc31f ma/t1450-quotefix later to maint).
(merge 56e743426b en/merge-recursive-comment-fixes later to maint).
(merge 7d23ff818f rs/bisect-oid-to-hex-fix later to maint).
(merge de20baf2c9 ny/notes-doc-sample-update later to maint).
(merge f649aaaf82 so/rev-parser-errormessage-fix later to maint).
(merge 6103d58b7f bc/sha-256-cvs-svn-updates later to maint).
(merge ac900fddb7 ma/stop-progress-null-fix later to maint).
(merge e767963ab6 rs/upload-pack-sigchain-fix later to maint).
(merge a831908599 rs/preserve-merges-unused-code-removal later to maint).
(merge 6dfefe70a9 jb/commit-graph-doc-fix later to maint).
(merge 847b37271e pb/set-url-docfix later to maint).
(merge 748f733d54 mt/checkout-entry-dead-code-removal later to maint).
(merge ce820cbd58 dl/subtree-docs later to maint).
(merge 55fe225dde jk/leakfix later to maint).
(merge ee22a29215 so/pretty-abbrev-doc later to maint).
(merge 3100fd5588 jc/post-checkout-doc later to maint).
(merge 17bae89476 pb/doc-external-diff-env later to maint).
(merge 27ed6ccc12 jk/worktree-check-clean-leakfix later to maint).
(merge 1302badd16 ea/blame-use-oideq later to maint).
(merge e6d5a11fed al/t3200-back-on-a-branch later to maint).
(merge 324efcf6b6 pw/add-p-leakfix later to maint).
(merge 1c6ffb546b jk/add-i-fixes later to maint).
(merge e40e936551 cd/commit-graph-doc later to maint).
(merge 0512eabd91 jc/sequencer-stopped-sha-simplify later to maint).
(merge d01141de5a so/combine-diff-simplify later to maint).
(merge 3be01e5ab1 sn/fast-import-doc later to maint).
Codebase Browser
git