1 Git v2.46 Release Notes
2 =======================
4 Backward Compatibility Notes
8 UI, Workflows & Features
10 * The "--rfc" option of "git format-patch" learned to take an
11 optional string value to be used in place of "RFC" to tweak the
12 "[PATCH]" on the subject header.
14 * The credential helper protocol, together with the HTTP layer, have
15 been enhanced to support authentication schemes different from
16 username & password pair, like Bearer and NTLM.
18 * Command line completion script (in contrib/) learned to complete
19 "git symbolic-ref" a bit better (you need to enable plumbing
20 commands to be completed with GIT_COMPLETION_SHOW_ALL_COMMANDS).
22 * When the user responds to a prompt given by "git add -p" with an
23 unsupported command, list of available commands were given, which
24 was too much if the user knew what they wanted to type but merely
25 made a typo. Now the user gets a much shorter error message.
27 * The color parsing code learned to handle 12-bit RGB colors, spelled
28 as "#RGB" (in addition to "#RRGGBB" that is already supported).
30 * The operation mode options (like "--get") the "git config" command
31 uses have been deprecated and replaced with subcommands (like "git
34 * "git tag" learned the "--trailer" option to futz with the trailers
35 in the same way as "git commit" does.
37 * A new global "--no-advice" option can be used to disable all advice
38 messages, which is meant to be used only in scripts.
40 * Updates to symbolic refs can now be made as a part of ref
43 * The trailer API has been reshuffled a bit.
45 * Terminology to call various ref-like things are getting
48 * The command line completion script (in contrib/) has been adjusted
49 to the recent update to "git config" that adopted subcommand based
52 * The knobs to tweak how reftable files are written have been made
53 available as configuration variables.
55 * When "git push" notices that the commit at the tip of the ref on
56 the other side it is about to overwrite does not exist locally, it
57 used to first try fetching it if the local repository is a partial
58 clone. The command has been taught not to do so and immediately
61 * The promisor.quiet configuration knob can be set to true to make
62 lazy fetching from promisor remotes silent.
64 * The inter/range-diff output has been moved to the end of the patch
65 when format-patch adds it to a single patch, instead of writing it
66 before the patch text, to be consistent with what is done for a
67 cover letter for a multi-patch series.
69 * A new command has been added to migrate a repository that uses the
70 files backend for its ref storage to use the reftable backend, with
73 * "git diff --exit-code --ext-diff" learned to take the exit status
74 of the external diff driver into account when deciding the exit
75 status of the overall "git diff" invocation when configured to do
78 * "git update-ref --stdin" learned to handle transactional updates of
81 * "git format-patch --interdiff" for multi-patch series learned to
82 turn on cover letters automatically (unless told never to enable
83 cover letter with "--no-cover-letter" and such).
85 * The "--heads" option of "ls-remote" and "show-ref" has been been
86 deprecated; "--branches" replaces "--heads".
88 * For over a year, setting add.interactive.useBuiltin configuration
89 variable did nothing but giving a "this does not do anything"
90 warning. The warning has been removed.
92 * The http transport can now be told to send request with
93 authentication material without first getting a 401 response.
95 * A handful of entries are added to the GitFAQ document.
97 * "git var GIT_SHELL_PATH" should report the path to the shell used
98 to spawn external commands, but it didn't do so on Windows, which
102 Performance, Internal Implementation, Development Support etc.
104 * Advertise "git contacts", a tool for newcomers to find people to
105 ask review for their patches, a bit more in our developer
108 * In addition to building the objects needed, try to link the objects
109 that are used in fuzzer tests, to make sure at least they build
110 without bitrot, in Linux CI runs.
112 * Code to write out reftable has seen some optimization and
115 * Tests to ensure interoperability between reftable written by jgit
116 and our code have been added and enabled in CI.
118 * The singleton index_state instance "the_index" has been eliminated
119 by always instantiating "the_repository" and replacing references
120 to "the_index" with references to its .index member.
122 * Git-GUI has a new maintainer, Johannes Sixt.
124 * The "test-tool" has been taught to run testsuite tests in parallel,
125 bypassing the need to use the "prove" tool.
127 * The "whitespace check" task that was enabled for GitHub Actions CI
128 has been ported to GitLab CI.
130 * The refs API lost functions that implicitly assumes to work on the
131 primary ref_store by forcing the callers to pass a ref_store as an
134 * Code clean-up to reduce inter-function communication inside
135 builtin/config.c done via the use of global variables.
137 * The pack bitmap code saw some clean-up to prepare for a follow-up topic.
139 * Preliminary code clean-up for "git send-email".
141 * The default "creation-factor" used by "git format-patch" has been
142 raised to make it more aggressively find matching commits.
144 * Before discovering the repository details, We used to assume SHA-1
145 as the "default" hash function, which has been corrected. Hopefully
146 this will smoke out codepaths that rely on such an unwarranted
149 * The project decision making policy has been documented.
151 * The strcmp-offset tests have been rewritten using the unit test
154 * "git add -p" learned to complain when an answer with more than one
155 letter is given to a prompt that expects a single letter answer.
157 * The alias-expanded command lines are logged to the trace output.
159 * A new test was added to ensure git commands that are designed to
160 run outside repositories do work.
162 * A few tests in reftable library have been rewritten using the
165 * A pair of test helpers that essentially are unit tests on hash
166 algorithms have been rewritten using the unit-tests framework.
168 * A test helper that essentially is unit tests on the "decorate"
169 logic has been rewritten using the unit-tests framework.
171 * Many memory leaks in the sparse-checkout code paths have been
174 * "make check-docs" noticed problems and reported to its output but
175 failed to signal its findings with its exit status, which has been
178 * Building with "-Werror -Wwrite-strings" is now supported.
180 * To help developers, the build procedure now allows builders to use
181 CFLAGS_APPEND to specify additional CFLAGS.
183 * "oidtree" tests were rewritten to use the unit test framework.
185 * The structure of the document that records longer-term project
186 decisions to deprecate/remove/update various behaviour has been
189 * The pseudo-merge reachability bitmap to help more efficient storage
190 of the reachability bitmap in a repository with too many refs has
193 * When "git merge" sees that the index cannot be refreshed (e.g. due
194 to another process doing the same in the background), it died but
195 after writing MERGE_HEAD etc. files, which was useless for the
196 purpose to recover from the failure.
198 * The output from "git cat-file --batch-check" and "--batch-command
199 (info)" should not be unbuffered, for which some tests have been
202 * A CPP macro USE_THE_REPOSITORY_VARIABLE is introduced to help
203 transition the codebase to rely less on the availability of the
204 singleton the_repository instance.
206 * "git version --build-options" reports the version information of
207 OpenSSL and other libraries (if used) in the build.
209 * Memory ownership rules for the in-core representation of
210 remote.*.url configuration values have been straightened out, which
211 resulted in a few leak fixes and code clarification.
213 * When bundleURI interface fetches multiple bundles, Git failed to
214 take full advantage of all bundles and ended up slurping duplicated
215 objects, which has been corrected..
217 * The code to deal with modified paths that are out-of-cone in a
218 sparsely checked out working tree has been optimized.
220 * An existing test of oidmap API has been rewritten with the
223 * The "ort" merge backend saw one bugfix for a crash that happens
224 when inner merge gets killed, and assorted code clean-ups.
226 * A new warning message is issued when a command has to expand a
227 sparse index to handle working tree cruft that are outside of the
230 * The test framework learned to take the test body not as a single
231 string but as a here-document.
233 * "git push '' HEAD:there" used to hit a BUG(); it has been corrected
234 to die with "fatal: bad repository ''".
236 * What happens when http.cookieFile gets the special value "" has
237 been clarified in the documentation.
243 * "git rebase --signoff" used to forget that it needs to add a
244 sign-off to the resulting commit when told to continue after a
245 conflict stops its operation.
247 * The procedure to build multi-pack-index got confused by the
248 replace-refs mechanism, which has been corrected by disabling the
251 * The "-k" and "--rfc" options of "format-patch" will now error out
252 when used together, as one tells us not to add anything to the
253 title of the commit, and the other one tells us to add "RFC" in
256 * "git stash -S" did not handle binary files correctly, which has
259 * A scheduled "git maintenance" job is expected to work on all
260 repositories it knows about, but it stopped at the first one that
261 errored out. Now it keeps going.
263 * zsh can pretend to be a normal shell pretty well except for some
264 glitches that we tickle in some of our scripts. Work them around
265 so that "vimdiff" and our test suite works well enough with it.
267 * Command line completion support for zsh (in contrib/) has been
268 updated to stop exposing internal state to end-user shell
271 * Tests that try to corrupt in-repository files in chunked format did
272 not work well on macOS due to its broken "mv", which has been
275 * The maximum size of attribute files is enforced more consistently.
277 * Unbreak CI jobs so that we do not attempt to use Python 2 that has
278 been removed from the platform.
280 * Git 2.43 started using the tree of HEAD as the source of attributes
281 in a bare repository, which has severe performance implications.
282 For now, revert the change, without ripping out a more explicit
283 support for the attr.tree configuration variable.
285 * The "--exit-code" option of "git diff" command learned to work with
286 the "--ext-diff" option.
288 * Windows CI running in GitHub Actions started complaining about the
289 order of arguments given to calloc(); the imported regex code uses
290 the wrong order almost consistently, which has been corrected.
292 * Expose "name conflict" error when a ref creation fails due to D/F
293 conflict in the ref namespace, to improve an error message given by
295 (merge 9339fca23e it/refs-name-conflict later to maint).
297 * The SubmittingPatches document now refers folks to manpages
300 * The documentation for "git diff --name-only" has been clarified
301 that it is about showing the names in the post-image tree.
303 * The credential helper that talks with osx keychain learned to avoid
304 storing back the authentication material it just got received from
306 (merge e1ab45b2da kn/osxkeychain-skip-idempotent-store later to maint).
308 * The chainlint script (invoked during "make test") did nothing when
309 it failed to detect the number of available CPUs. It now falls
310 back to 1 CPU to avoid the problem.
312 * Revert overly aggressive "layered defence" that went into 2.45.1
313 and friends, which broke "git-lfs", "git-annex", and other use
314 cases, so that we can rebuild necessary counterparts in the open.
316 * "git init" in an already created directory, when the user
317 configuration has includeif.onbranch, started to fail recently,
318 which has been corrected.
320 * Memory leaks in "git mv" has been plugged.
322 * The safe.directory configuration knob has been updated to
323 optionally allow leading path matches.
325 * An overly large ".gitignore" files are now rejected silently.
327 * Upon expiration event, the credential subsystem forgot to clear
328 in-core authentication material other than password (whose support
329 was added recently), which has been corrected.
331 * Fix for an embarrassing typo that prevented Python2 tests from running
334 * Varargs functions that are unannotated as printf-like or execl-like
335 have been annotated as such.
337 * "git am" has a safety feature to prevent it from starting a new
338 session when there already is a session going. It reliably
339 triggers when a mbox is given on the command line, but it has to
340 rely on the tty-ness of the standard input. Add an explicit way to
341 opt out of this safety with a command line option.
342 (merge 62c71ace44 jk/am-retry later to maint).
344 * A leak in "git imap-send" that somehow escapes LSan has been
347 * Setting core.abbrev too early before the repository set-up
348 (typically in "git clone") caused segfault, which as been
351 * When the user adds to "git rebase -i" instruction to "pick" a merge
352 commit, the error experience is not pleasant. Such an error is now
353 caught earlier in the process that parses the todo list.
355 * We forgot to normalize the result of getcwd() to NFC on macOS where
356 all other paths are normalized, which has been corrected. This still
357 does not address the case where core.precomposeUnicode configuration
358 is not defined globally.
360 * Earlier we stopped using the tree of HEAD as the default source of
361 attributes in a bare repository, but failed to document it. This
364 * "git update-server-info" and "git commit-graph --write" have been
365 updated to use the tempfile API to avoid leaving cruft after
368 * An unused extern declaration for mingw has been removed to prevent
369 it from causing build failure.
371 * A helper function shared between two tests had a copy-paste bug,
372 which has been corrected.
374 * "git fetch-pack -k -k" without passing "--lock-pack" (which we
375 never do ourselves) did not work at all, which has been corrected.
377 * CI job to build minimum fuzzers learned to pass NO_CURL=NoThanks to
378 the build procedure, as its build environment does not offer, or
379 the rest of the build needs, anything cURL.
380 (merge 4e66b5a990 jc/fuzz-sans-curl later to maint).
382 * "git diff --no-ext-diff" when diff.external is configured ignored
383 the "--color-moved" option.
384 (merge 0f4b0d4cf0 rs/diff-color-moved-w-no-ext-diff-fix later to maint).
386 * "git archive --add-virtual-file=<path>:<contents>" never paid
387 attention to the --prefix=<prefix> option but the documentation
388 said it would. The documentation has been corrected.
389 (merge 72c282098d jc/archive-prefix-with-add-virtual-file later to maint).
391 * When GIT_PAGER failed to spawn, depending on the code path taken,
392 we failed immediately (correct) or just spew the payload to the
393 standard output (incorrect). The code now always fail immediately
394 when GIT_PAGER fails.
395 (merge 78f0a5d187 rj/pager-die-upon-exec-failure later to maint).
397 * date parser updates to be more careful about underflowing epoch
399 (merge 9d69789770 db/date-underflow-fix later to maint).
401 * The Bloom filter used for path limited history traversal was broken
402 on systems whose "char" is unsigned; update the implementation and
403 bump the format version to 2.
404 (merge 9c8a9ec787 tb/path-filter-fix later to maint).
407 (merge 231cf7370e as/pathspec-h-typofix later to maint).
410 (merge 4b837f821e rs/simplify-submodule-helper-super-prefix-invocation later to maint).
412 * "git describe --dirty --broken" forgot to refresh the index before
413 seeing if there is any chang, ("git describe --dirty" correctly did
414 so), which has been corrected.
415 (merge b8ae42e292 as/describe-broken-refresh-index-fix later to maint).
417 * Test suite has been taught not to unnecessarily rely on DNS failing
418 a bogus external name.
419 (merge 407cdbd271 jk/tests-without-dns later to maint).
421 * GitWeb update to use committer date consistently in rss/atom feeds.
422 (merge cf6ead095b am/gitweb-feed-use-committer-date later to maint).
424 * Custom control structures we invented more recently have been
425 taught to the clang-format file.
426 (merge 1457dff9be rs/clang-format-updates later to maint).
428 * Developer build procedure fix.
429 (merge df32729866 tb/dev-build-pedantic-fix later to maint).
431 * "git push" that pushes only deletion gave an unnecessary and
432 harmless error message when push negotiation is configured, which
434 (merge 4d8ee0317f jc/disable-push-nego-for-deletion later to maint).
436 * Address-looking strings found on the trailer are now placed on the
437 Cc: list after running through sanitize_address by "git send-email".
438 (merge c852531f45 cb/send-email-sanitize-trailer-addresses later to maint).
440 * Tests that use GIT_TEST_SANITIZE_LEAK_LOG feature got their exit
441 status inverted, which has been corrected.
442 (merge 8c1d6691bc rj/test-sanitize-leak-log-fix later to maint).
444 * The http.cookieFile and http.saveCookies configuration variables
445 have a few values that need to be avoided, which are now ignored
446 with warning messages.
447 (merge 4f5822076f jc/http-cookiefile later to maint).
449 * Other code cleanup, docfix, build fix, etc.
450 (merge 493fdae046 ew/object-convert-leakfix later to maint).
451 (merge 00f3661a0a ss/doc-eol-attr-fix later to maint).
452 (merge 428c40da61 ri/doc-show-branch-fix later to maint).
453 (merge 58696bfcaa jc/where-is-bash-for-ci later to maint).