7 Backward incompatible and other important changes
9 * The "pack-redundant" command, which has been left stale with almost
10 unusable performance issues, now warns loudly when it gets used, as
11 we no longer want to recommend its use (instead just "repack -d"
14 * The development community has adopted Contributor Covenant v2.0 to
15 update from v1.4 that we have been using.
17 * The support for deprecated PCRE1 library has been dropped.
20 UI, Workflows & Features
22 * The "--format=%(trailers)" mechanism gets enhanced to make it
23 easier to design output for machine consumption.
25 * When a user does not tell "git pull" to use rebase or merge, the
26 command gives a loud message telling a user to choose between
27 rebase or merge but creates a merge anyway, forcing users who would
28 want to rebase to redo the operation. Fix an early part of this
29 problem by tightening the condition to give the message---there is
30 no reason to stop or force the user to choose between rebase or
31 merge if the history fast-forwards.
33 * The configuration variable 'core.abbrev' can be set to 'no' to
34 force no abbreviation regardless of the hash algorithm.
36 * "git rev-parse" can be explicitly told to give output as absolute
37 or relative path with the `--path-format=(absolute|relative)` option.
39 * Bash completion (in contrib/) update to make it easier for
40 end-users to add completion for their custom "git" subcommands.
42 * "git maintenance" learned to drive scheduled maintenance on
43 platforms whose native scheduling methods are not 'cron'.
45 * After expiring a reflog and making a single commit, the reflog for
46 the branch would record a single entry that knows both @{0} and
47 @{1}, but we failed to answer "what commit were we on?", i.e. @{1}
49 * "git bundle" learns "--stdin" option to read its refs from the
50 standard input. Also, it now does not lose refs whey they point
53 * "git log" learned a new "--diff-merges=<how>" option.
55 * "git ls-files" can and does show multiple entries when the index is
56 unmerged, which is a source for confusion unless -s/-u option is in
57 use. A new option --deduplicate has been introduced.
59 * `git worktree list` now annotates worktrees as prunable, shows
60 locked and prunable attributes in --porcelain mode, and gained
63 * "git clone" tries to locally check out the branch pointed at by
64 HEAD of the remote repository after it is done, but the protocol
65 did not convey the information necessary to do so when copying an
66 empty repository. The protocol v2 learned how to do so.
68 * There are other ways than ".." for a single token to denote a
69 "commit range", namely "<rev>^!" and "<rev>^-<n>", but "git
70 range-diff" did not understand them.
72 * The "git range-diff" command learned "--(left|right)-only" option
73 to show only one side of the compared range.
75 * "git mergetool" feeds three versions (base, local and remote) of
76 a conflicted path unmodified. The command learned to optionally
77 prepare these files with unconflicted parts already resolved.
79 * The .mailmap is documented to be read only from the root level of a
80 working tree, but a stray file in a bare repository also was read
81 by accident, which has been corrected.
83 * "git maintenance" tool learned a new "pack-refs" maintenance task.
85 * The error message given when a configuration variable that is
86 expected to have a boolean value has been improved.
88 * Signed commits and tags now allow verification of objects, whose
89 two object names (one in SHA-1, the other in SHA-256) are both
92 * "git rev-list" command learned "--disk-usage" option.
94 * "git {diff,log} --{skip,rotate}-to=<path>" allows the user to
95 discard diff output for early paths or move them to the end of the
98 * "git difftool" learned "--skip-to=<path>" option to restart an
99 interrupted session from an arbitrary path.
101 * "git grep" has been tweaked to be limited to the sparse checkout
104 * "git rebase --[no-]fork-point" gained a configuration variable
105 rebase.forkPoint so that users do not have to keep specifying a
109 Performance, Internal Implementation, Development Support etc.
111 * A 3-year old test that was not testing anything useful has been
114 * Retire more names with "sha1" in it.
116 * The topological walk codepath is covered by new trace2 stats.
118 * Update the Code-of-conduct to version 2.0 from the upstream (we've
119 been using version 1.4).
121 * "git mktag" validates its input using its own rules before writing
122 a tag object---it has been updated to share the logic with "git
125 * Two new ways to feed configuration variable-value pairs via
126 environment variables have been introduced, and the way
127 GIT_CONFIG_PARAMETERS encodes variable/value pairs has been tweaked
128 to make it more robust.
130 * Tests have been updated so that they do not to get affected by the
131 name of the default branch "git init" creates.
133 * "git fetch" learns to treat ref updates atomically in all-or-none
134 fashion, just like "git push" does, with the new "--atomic" option.
136 * The peel_ref() API has been replaced with peel_iterated_oid().
138 * The .use_shell flag in struct child_process that is passed to
139 run_command() API has been clarified with a bit more documentation.
141 * Document, clean-up and optimize the code around the cache-tree
142 extension in the index.
144 * The ls-refs protocol operation has been optimized to narrow the
145 sub-hierarchy of refs/ it walks to produce response.
147 * When removing many branches and tags, the code used to do so one
148 ref at a time. There is another API it can use to delete multiple
149 refs, and it makes quite a lot of performance difference when the
152 * The "pack-objects" command needs to iterate over all the tags when
153 automatic tag following is enabled, but it actually iterated over
154 all refs and then discarded everything outside "refs/tags/"
155 hierarchy, which was quite wasteful.
157 * A perf script was made more portable.
159 * Our setting of GitHub CI test jobs were a bit too eager to give up
160 once there is even one failure found. Tweak the knob to allow
161 other jobs keep running even when we see a failure, so that we can
162 find more failures in a single run.
164 * We've carried compatibility codepaths for compilers without
165 variadic macros for quite some time, but the world may be ready for
166 them to be removed. Force compilation failure on exotic platforms
167 where variadic macros are not available to find out who screams in
168 such a way that we can easily revert if it turns out that the world
171 * Code clean-up to ensure our use of hashtables using object names as
172 keys use the "struct object_id" objects, not the raw hash values.
174 * Lose the debugging aid that may have been useful in the past, but
175 no longer is, in the "grep" codepaths.
177 * Some pretty-format specifiers do not need the data in commit object
178 (e.g. "%H"), but we were over-eager to load and parse it, which has
179 been made even lazier.
181 * Get rid of "GETTEXT_POISON" support altogether, which may or may
182 not be controversial.
184 * Introduce an on-disk file to record revindex for packdata, which
185 traditionally was always created on the fly and only in-core.
187 * The commit-graph learned to use corrected commit dates instead of
188 the generation number to help topological revision traversal.
190 * Piecemeal of rewrite of "git bisect" in C continues.
192 * When a pager spawned by us exited, the trace log did not record its
193 exit status correctly, which has been corrected.
195 * Removal of GIT_TEST_GETTEXT_POISON continues.
197 * The code to implement "git merge-base --independent" was poorly
198 done and was kept from the very beginning of the feature.
200 * Preliminary changes to fsmonitor integration.
202 * Performance improvements for rename detection.
204 * The common code to deal with "chunked file format" that is shared
205 by the multi-pack-index and commit-graph files have been factored
206 out, to help codepaths for both filetypes to become more robust.
208 * The approach to "fsck" the incoming objects in "index-pack" is
209 attractive for performance reasons (we have them already in core,
210 inflated and ready to be inspected), but fundamentally cannot be
211 applied fully when we receive more than one pack stream, as a tree
212 object in one pack may refer to a blob object in another pack as
213 ".gitmodules", when we want to inspect blobs that are used as
214 ".gitmodules" file, for example. Teach "index-pack" to emit
215 objects that must be inspected later and check them in the calling
216 "fetch-pack" process.
218 * The logic to handle "trailer" related placeholders in the
219 "--format=" mechanisms in the "log" family and "for-each-ref"
220 family is getting unified.
222 * Raise the buffer size used when writing the index file out from
223 (obviously too small) 8kB to (clearly sufficiently large) 128kB.
225 * It is reported that open() on some platforms (e.g. macOS Big Sur)
226 can return EINTR even though our timers are set up with SA_RESTART.
227 A workaround has been implemented and enabled for macOS to rerun
228 open() transparently from the caller when this happens.
234 * Diagnose command line error of "git rebase" early.
236 * Clean up option descriptions in "git cmd --help".
238 * "git stash" did not work well in a sparsely checked out working
241 * Some tests expect that "ls -l" output has either '-' or 'x' for
242 group executable bit, but setgid bit can be inherited from parent
243 directory and make these fields 'S' or 's' instead, causing test
246 * "git for-each-repo --config=<var> <cmd>" should not run <cmd> for
247 any repository when the configuration variable <var> is not defined
250 * Fix 2.29 regression where "git mergetool --tool-help" fails to list
251 all the available tools.
253 * Fix for procedure to building CI test environment for mac.
255 * The implementation of "git branch --sort" wrt the detached HEAD
256 display has always been hacky, which has been cleaned up.
258 * Newline characters in the host and path part of git:// URL are
261 * "git diff" showed a submodule working tree with untracked cruft as
262 "Submodule commit <objectname>-dirty", but a natural expectation is
263 that the "-dirty" indicator would align with "git describe --dirty",
264 which does not consider having untracked files in the working tree
265 as source of dirtiness. The inconsistency has been fixed.
267 * When more than one commit with the same patch ID appears on one
268 side, "git log --cherry-pick A...B" did not exclude them all when a
269 commit with the same patch ID appears on the other side. Now it
272 * Documentation for "git fsck" lost stale bits that has become
275 * Doc fix for packfile URI feature.
277 * When "git rebase -i" processes "fixup" insn, there is no reason to
278 clean up the commit log message, but we did the usual stripspace
279 processing. This has been corrected.
280 (merge f7d42ceec5 js/rebase-i-commit-cleanup-fix later to maint).
282 * Fix in passing custom args from "git clone" to "upload-pack" on the
284 (merge ad6b5fefbd jv/upload-pack-filter-spec-quotefix later to maint).
286 * The command line completion (in contrib/) completed "git branch -d"
287 with branch names, but "git branch -D" offered tagnames in addition,
288 which has been corrected. "git branch -M" had the same problem.
289 (merge 27dc071b9a jk/complete-branch-force-delete later to maint).
291 * When commands are started from a subdirectory, they may have to
292 compare the path to the subdirectory (called prefix and found out
293 from $(pwd)) with the tracked paths. On macOS, $(pwd) and
294 readdir() yield decomposed path, while the tracked paths are
295 usually normalized to the precomposed form, causing mismatch. This
296 has been fixed by taking the same approach used to normalize the
297 command line arguments.
298 (merge 5c327502db tb/precompose-prefix-too later to maint).
300 * Even though invocations of "die()" were logged to the trace2
301 system, "BUG()"s were not, which has been corrected.
302 (merge 0a9dde4a04 jt/trace2-BUG later to maint).
304 * "git grep --untracked" is meant to be "let's ALSO find in these
305 files on the filesystem" when looking for matches in the working
306 tree files, and does not make any sense if the primary search is
307 done against the index, or the tree objects. The "--cached" and
308 "--untracked" options have been marked as mutually incompatible.
309 (merge 0c5d83b248 mt/grep-cached-untracked later to maint).
311 * Fix "git fsck --name-objects" which apparently has not been used by
312 anybody who is motivated enough to report breakage.
313 (merge e89f89361c js/fsck-name-objects-fix later to maint).
315 * Avoid individual tests in t5411 from getting affected by each other
316 by forcing them to use separate output files during the test.
317 (merge 822ee894f6 jx/t5411-unique-filenames later to maint).
319 * Test to make sure "git rev-parse one-thing one-thing" gives
320 the same thing twice (when one-thing is --since=X).
321 (merge a5cdca4520 ew/rev-parse-since-test later to maint).
323 * When certain features (e.g. grafts) used in the repository are
324 incompatible with the use of the commit-graph, we used to silently
325 turned commit-graph off; we now tell the user what we are doing.
326 (merge c85eec7fc3 js/commit-graph-warning later to maint).
328 * Objects that lost references can be pruned away, even when they
329 have notes attached to it (and these notes will become dangling,
330 which in turn can be pruned with "git notes prune"). This has been
331 clarified in the documentation.
332 (merge fa9ab027ba mz/doc-notes-are-not-anchors later to maint).
334 * The error codepath around the "--temp/--prefix" feature of "git
335 checkout-index" has been improved.
336 (merge 3f7ba60350 mt/checkout-index-corner-cases later to maint).
338 * The "git maintenance register" command had trouble registering bare
339 repositories, which had been corrected.
341 * A handful of multi-word configuration variable names in
342 documentation that are spelled in all lowercase have been corrected
343 to use the more canonical camelCase.
344 (merge 7dd0eaa39c dl/doc-config-camelcase later to maint).
346 * "git push $there --delete ''" should have been diagnosed as an
347 error, but instead turned into a matching push, which has been
349 (merge 20e416409f jc/push-delete-nothing later to maint).
351 * Test script modernization.
352 (merge 488acf15df sv/t7001-modernize later to maint).
354 * An under-allocation for the untracked cache data has been corrected.
355 (merge 6347d649bc jh/untracked-cache-fix later to maint).
357 * Other code cleanup, docfix, build fix, etc.
358 (merge e3f5da7e60 sg/t7800-difftool-robustify later to maint).
359 (merge 9d336655ba js/doc-proto-v2-response-end later to maint).
360 (merge 1b5b8cf072 jc/maint-column-doc-typofix later to maint).
361 (merge 3a837b58e3 cw/pack-config-doc later to maint).
362 (merge 01168a9d89 ug/doc-commit-approxidate later to maint).
363 (merge b865734760 js/params-vs-args later to maint).