4 Backward compatibility notes and other notable changes.
6 * Use of an empty string as a pathspec element that is used for
7 'everything matches' is still warned and Git asks users to use a
8 more explicit '.' for that instead. The hope is that existing
9 users will not mind this change, and eventually the warning can be
10 turned into a hard error, upgrading the deprecation into removal of
11 this (mis)feature. That is now scheduled to happen in the upcoming
14 * Git now avoids blindly falling back to ".git" when the setup
15 sequence said we are _not_ in Git repository. A corner case that
16 happens to work right now may be broken by a call to die("BUG").
17 We've tried hard to locate such cases and fixed them, but there
18 might still be cases that need to be addressed--bug reports are
21 * "branch --set-upstream" that has been deprecated in Git 1.8 has
28 UI, Workflows & Features
30 * An example that is now obsolete has been removed from a sample hook,
31 and an old example in it that added a sign-off manually has been
32 improved to use the interpret-trailers command.
34 * The advice message given when "git rebase" stops for conflicting
35 changes has been improved.
37 * The "rerere-train" script (in contrib/) learned the "--overwrite"
38 option to allow overwriting existing recorded resolutions.
40 * "git contacts" (in contrib/) now lists the address on the
41 "Reported-by:" trailer to its output, in addition to those on
42 S-o-b: and other trailers, to make it easier to notify (and thank)
43 the original bug reporter.
45 * "git rebase", especially when it is run by mistake and ends up
46 trying to replay many changes, spent long time in silence. The
47 command has been taught to show progress report when it spends
48 long time preparing these many changes to replay (which would give
49 the user a chance to abort with ^C).
51 * "git merge" learned a "--signoff" option to add the Signed-off-by:
52 trailer with the committer's name.
54 * "git diff" learned to optionally paint new lines that are the same
55 as deleted lines elsewhere differently from genuinely new lines.
57 * "git interpret-trailers" learned to take the trailer specifications
58 from the command line that overrides the configured values.
60 * "git interpret-trailers" has been taught a "--parse" and a few
61 other options to make it easier for scripts to grab existing
62 trailer lines from a commit log message.
64 * "gitweb" shows a link to visit the 'raw' contents of blbos in the
65 history overview page.
67 * "[gc] rerereResolved = 5.days" used to be invalid, as the variable
68 is defined to take an integer counting the number of days. It now
71 * The code to acquire a lock on a reference (e.g. while accepting a
72 push from a client) used to immediately fail when the reference is
73 already locked---now it waits for a very short while and retries,
74 which can make it succeed if the lock holder was holding it during
75 a read-only operation.
77 * "branch --set-upstream" that has been deprecated in Git 1.8 has
80 * The codepath to call external process filter for smudge/clean
81 operation learned to show the progress meter.
83 * "git rev-parse" learned "--is-shallow-repository", that is to be
84 used in a way similar to existing "--is-bare-repository" and
87 * "git describe --match <pattern>" has been taught to play well with
90 * "git branch" learned "-c/-C" to create a new branch by copying an
93 * Some commands (most notably "git status") makes an opportunistic
94 update when performing a read-only operation to help optimize later
95 operations in the same repository. The new "--no-optional-locks"
96 option can be passed to Git to disable them.
99 Performance, Internal Implementation, Development Support etc.
101 * Conversion from uchar[20] to struct object_id continues.
103 * Start using selected c99 constructs in small, stable and
104 essentialpart of the system to catch people who care about
105 older compilers that do not grok them.
107 * The filter-process interface learned to allow a process with long
108 latency give a "delayed" response.
110 * Many uses of comparision callback function the hashmap API uses
111 cast the callback function type when registering it to
112 hashmap_init(), which defeats the compile time type checking when
113 the callback interface changes (e.g. gaining more parameters).
114 The callback implementations have been updated to take "void *"
115 pointers and cast them to the type they expect instead.
117 * Because recent Git for Windows do come with a real msgfmt, the
118 build procedure for git-gui has been updated to use it instead of a
119 hand-rolled substitute.
121 * "git grep --recurse-submodules" has been reworked to give a more
122 consistent output across submodule boundary (and do its thing
123 without having to fork a separate process).
125 * A helper function to read a single whole line into strbuf
126 mistakenly triggered OOM error at EOF under certain conditions,
127 which has been fixed.
128 (merge 642956cf45 rs/strbuf-getwholeline-fix later to maint).
130 * The "ref-store" code reorganization continues.
132 * "git commit" used to discard the index and re-read from the filesystem
133 just in case the pre-commit hook has updated it in the middle; this
134 has been optimized out when we know we do not run the pre-commit hook.
135 (merge 680ee550d7 kw/commit-keep-index-when-pre-commit-is-not-run later to maint).
137 * Updates to the HTTP layer we made recently unconditionally used
138 features of libCurl without checking the existence of them, causing
139 compilation errors, which has been fixed. Also migrate the code to
140 check feature macros, not version numbers, to cope better with
141 libCurl that vendor ships with backported features.
143 * The API to start showing progress meter after a short delay has
145 (merge 8aade107dd jc/simplify-progress later to maint).
147 * Code clean-up to avoid mixing values read from the .gitmodules file
148 and values read from the .git/config file.
150 * We used to spend more than necessary cycles allocating and freeing
151 piece of memory while writing each index entry out. This has been
154 * Platforms that ship with a separate sha1 with collision detection
155 library can link to it instead of using the copy we ship as part of
158 * Code around "notes" have been cleaned up.
159 (merge 3964281524 mh/notes-cleanup later to maint).
161 * The long-standing rule that an in-core lockfile instance, once it
162 is used, must not be freed, has been lifted and the lockfile and
163 tempfile APIs have been updated to reduce the chance of programming
166 * Our hashmap implementation in hashmap.[ch] is not thread-safe when
167 adding a new item needs to expand the hashtable by rehashing; add
168 an API to disable the automatic rehashing to work it around.
170 * Many of our programs consider that it is OK to release dynamic
171 storage that is used throughout the life of the program by simply
172 exiting, but this makes it harder to leak detection tools to avoid
173 reporting false positives. Plug many existing leaks and introduce
174 a mechanism for developers to mark that the region of memory
175 pointed by a pointer is not lost/leaking to help these tools.
177 * As "git commit" to conclude a conflicted "git merge" honors the
178 commit-msg hook, "git merge" that records a merge commit that
179 cleanly auto-merges should, but it didn't.
181 * The codepath for "git merge-recursive" has been cleaned up.
183 * Many leaks of strbuf have been fixed.
185 * "git imap-send" has our own implementation of the protocol and also
186 can use more recent libCurl with the imap protocol support. Update
187 the latter so that it can use the credential subsystem, and then
188 make it the default option to use, so that we can eventually
189 deprecate and remove the former.
191 * "make style" runs git-clang-format to help developers by pointing
192 out coding style issues.
194 * A test to demonstrate "git mv" failing to adjust nested submodules
196 (merge c514167df2 hv/mv-nested-submodules-test later to maint).
198 * On Cygwin, "ulimit -s" does not report failure but it does not work
199 at all, which causes an unexpected success of some tests that
200 expect failures under a limited stack situation. This has been
203 * Many codepaths have been updated to squelch -Wimplicit-fallthrough
204 warnings from Gcc 7 (which is a good code hygiene).
206 * Add a helper for DLL loading in anticipation for its need in a
209 * "git status --ignored", when noticing that a directory without any
210 tracked path is ignored, still enumerated all the ignored paths in
211 the directory, which is unnecessary. The codepath has been
212 optimized to avoid this overhead.
214 * The final batch to "git rebase -i" updates to move more code from
215 the shell script to C has been merged.
217 * Operations that do not touch (majority of) packed refs have been
218 optimized by making accesses to packed-refs file lazy; we no longer
219 pre-parse everything, and an access to a single ref in the
220 packed-refs does not touch majority of irrelevant refs, either.
222 Also contains various documentation updates and code clean-ups.
228 * "%C(color name)" in the pretty print format always produced ANSI
229 color escape codes, which was an early design mistake. They now
230 honor the configuration (e.g. "color.ui = never") and also tty-ness
231 of the output medium.
233 * The http.{sslkey,sslCert} configuration variables are to be
234 interpreted as a pathname that honors "~[username]/" prefix, but
235 weren't, which has been fixed.
237 * Numerous bugs in walking of reflogs via "log -g" and friends have
240 * "git commit" when seeing an totally empty message said "you did not
241 edit the message", which is clearly wrong. The message has been
244 * When a directory is not readable, "gitweb" fails to build the
245 project list. Work this around by skipping such a directory.
247 * Some versions of GnuPG fails to kill gpg-agent it auto-spawned
248 and such a left-over agent can interfere with a test. Work it
249 around by attempting to kill one before starting a new test.
251 * A recently added test for the "credential-cache" helper revealed
252 that EOF detection done around the time the connection to the cache
253 daemon is torn down were flaky. This was fixed by reacting to
254 ECONNRESET and behaving as if we got an EOF.
256 * "git log --tag=no-such-tag" showed log starting from HEAD, which
257 has been fixed---it now shows nothing.
259 * The "tag.pager" configuration variable was useless for those who
260 actually create tag objects, as it interfered with the use of an
261 editor. A new mechanism has been introduced for commands to enable
262 pager depending on what operation is being carried out to fix this,
263 and then "git tag -l" is made to run pager by default.
265 * "git push --recurse-submodules $there HEAD:$target" was not
266 propagated down to the submodules, but now it is.
268 * Commands like "git rebase" accepted the --rerere-autoupdate option
269 from the command line, but did not always use it. This has been
272 * "git clone --recurse-submodules --quiet" did not pass the quiet
273 option down to submodules.
275 * Test portability fix for OBSD.
277 * Portability fix for OBSD.
279 * "git am -s" has been taught that some input may end with a trailer
280 block that is not Signed-off-by: and it should refrain from adding
281 an extra blank line before adding a new sign-off in such a case.
283 * "git svn" used with "--localtime" option did not compute the tz
284 offset for the timestamp in question and instead always used the
285 current time, which has been corrected.
287 * Memory leak in an error codepath has been plugged.
289 * "git stash -u" used the contents of the committed version of the
290 ".gitignore" file to decide which paths are ignored, even when the
291 file has local changes. The command has been taught to instead use
292 the locally modified contents.
294 * bash 4.4 or newer gave a warning on NUL byte in command
295 substitution done in "git stash"; this has been squelched.
297 * "git grep -L" and "git grep --quiet -L" reported different exit
298 codes; this has been corrected.
300 * When handshake with a subprocess filter notices that the process
301 asked for an unknown capability, Git did not report what program
302 the offending subprocess was running. This has been corrected.
304 * "git apply" that is used as a better "patch -p1" failed to apply a
305 taken from a file with CRLF line endings to a file with CRLF line
306 endings. The root cause was because it misused convert_to_git()
307 that tried to do "safe-crlf" processing by looking at the index
308 entry at the same path, which is a nonsense---in that mode, "apply"
309 is not working on the data in (or derived from) the index at all.
312 * Killing "git merge --edit" before the editor returns control left
313 the repository in a state with MERGE_MSG but without MERGE_HEAD,
314 which incorrectly tells the subsequent "git commit" that there was
315 a squash merge in progress. This has been fixed.
317 * "git archive" did not work well with pathspecs and the
318 export-ignore attribute.
320 * In addition to "cc: <a@dd.re.ss> # cruft", "cc: a@dd.re.ss # cruft"
321 was taught to "git send-email" as a valid way to tell it that it
322 needs to also send a carbon copy to <a@dd.re.ss> in the trailer
324 (merge cc90750677 mm/send-email-cc-cruft later to maint).
326 * "git branch -M a b" while on a branch that is completely unrelated
327 to either branch a or branch b misbehaved when multiple worktree
328 was in use. This has been fixed.
329 (merge 31824d180d nd/worktree-kill-parse-ref later to maint).
331 * "git gc" and friends when multiple worktrees are used off of a
332 single repository did not consider the index and per-worktree refs
333 of other worktrees as the root for reachability traversal, making
334 objects that are in use only in other worktrees to be subject to
337 * A regression to "gitk --bisect" by a recent update has been fixed.
338 (merge 1d0538e486 mh/packed-ref-store-prep later to maint).
340 * "git -c submodule.recurse=yes pull" did not work as if the
341 "--recurse-submodules" option was given from the command line.
342 This has been corrected.
344 * Unlike "git commit-tree < file", "git commit-tree -F file" did not
345 pass the contents of the file verbatim and instead completed an
346 incomplete line at the end, if exists. The latter has been updated
347 to match the behaviour of the former.
348 (merge c818e74332 rk/commit-tree-make-F-verbatim later to maint).
350 * Many codepaths did not diagnose write failures correctly when disks
351 go full, due to their misuse of write_in_full() helper function,
352 which have been corrected.
353 (merge f48ecd38cb jk/write-in-full-fix later to maint).
355 * "git help co" now says "co is aliased to ...", not "git co is".
356 (merge b3a8076e0d ks/help-alias-label later to maint).
358 * "git archive", especially when used with pathspec, stored an empty
359 directory in its output, even though Git itself never does so.
361 (merge 4318094047 rs/archive-excluded-directory later to maint).
363 * API error-proofing which happens to also squelch warnings from GCC.
364 (merge c788c54cde tg/refs-allowed-flags later to maint).
366 * The explanation of the cut-line in the commit log editor has been
368 (merge 8c4b1a3593 ks/commit-do-not-touch-cut-line later to maint).
370 * "git gc" tries to avoid running two instances at the same time by
371 reading and writing pid/host from and to a lock file; it used to
372 use an incorrect fscanf() format when reading, which has been
374 (merge afe2fab72c aw/gc-lockfile-fscanf-fix later to maint).
376 * The scripts to drive TravisCI has been reorganized and then an
377 optimization to avoid spending cycles on a branch whose tip is
378 tagged has been implemented.
379 (merge 8376eb4a8f ls/travis-scriptify later to maint).
381 * The test linter has been taught that we do not like "echo -e".
382 (merge 1a6d46895d tb/test-lint-echo-e later to maint).
384 * Code cmp.std.c nitpick.
385 (merge ac7da78ede mh/for-each-string-list-item-empty-fix later to maint).
387 * A regression fix for 2.11 that made the code to read the list of
388 alternate object stores overrun the end of the string.
389 (merge f0f7bebef7 jk/info-alternates-fix later to maint).
391 * "git describe --match" learned to take multiple patterns in v2.13
392 series, but the feature ignored the patterns after the first one
393 and did not work at all. This has been fixed.
394 (merge da769d2986 jk/describe-omit-some-refs later to maint).
396 * "git filter-branch" cannot reproduce a history with a tag without
397 the tagger field, which only ancient versions of Git allowed to be
398 created. This has been corrected.
399 (merge b2c1ca6b4b ic/fix-filter-branch-to-handle-tag-without-tagger later to maint).
401 * "git cat-file --textconv" started segfaulting recently, which
403 (merge cc0ea7c9e5 jk/diff-blob later to maint).
405 * The built-in pattern to detect the "function header" for HTML did
406 not match <H1>..<H6> elements without any attributes, which has
408 (merge 9c03caca2c ik/userdiff-html-h-element-fix later to maint).
410 * "git mailinfo" was loose in decoding quoted printable and produced
411 garbage when the two letters after the equal sign are not
412 hexadecimal. This has been fixed.
413 (merge c8cf423eab rs/mailinfo-qp-decode-fix later to maint).
415 * The machinery to create xdelta used in pack files received the
416 sizes of the data in size_t, but lost the higher bits of them by
417 storing them in "unsigned int" during the computation, which is
420 * The delta format used in the packfile cannot reference data at
421 offset larger than what can be expressed in 4-byte, but the
422 generator for the data failed to make sure the offset does not
423 overflow. This has been corrected.
425 * The documentation for '-X<option>' for merges was misleadingly
426 written to suggest that "-s theirs" exists, which is not the case.
427 (merge c25d98b2a7 jc/merge-x-theirs-docfix later to maint).
429 * "git fast-export" with -M/-C option issued "copy" instruction on a
430 path that is simultaneously modified, which was incorrect.
431 (merge b3e8ca89cf jt/fast-export-copy-modify-fix later to maint).
433 * Many codepaths have been updated to squelch -Wsign-compare
435 (merge 071bcaab64 rj/no-sign-compare later to maint).
437 * Memory leaks in various codepaths have been plugged.
438 (merge 4d01a7fa65 ma/leakplugs later to maint).
440 * Recent versions of "git rev-parse --parseopt" did not parse the
441 option specification that does not have the optional flags (*=?!)
442 correctly, which has been corrected.
443 (merge a6304fa4c2 bc/rev-parse-parseopt-fix later to maint).
445 * Other minor doc, test and build updates and code cleanups.
446 (merge f094b89a4d ma/parse-maybe-bool later to maint).
447 (merge 39b00fa4d4 jk/drop-sha1-entry-pos later to maint).
448 (merge 6cdf8a7929 ma/ts-cleanups later to maint).
449 (merge 7560f547e6 ma/up-to-date later to maint).
450 (merge 0db3dc75f3 rs/apply-epoch later to maint).
451 (merge 74f1bd912b dw/diff-highlight-makefile-fix later to maint).
452 (merge f991761eb8 jk/config-lockfile-leak-fix later to maint).
453 (merge 150efef1e7 ma/pkt-line-leakfix later to maint).
454 (merge 5554451de6 mg/timestamp-t-fix later to maint).
455 (merge 276d0e35c0 ma/split-symref-update-fix later to maint).
456 (merge 3bc4b8f7c7 bb/doc-eol-dirty later to maint).
457 (merge c1bb33c99c jk/system-path-cleanup later to maint).
458 (merge ab46e6fc72 cc/subprocess-handshake-missing-capabilities later to maint).
459 (merge f7a32dd97f kd/doc-for-each-ref later to maint).
460 (merge be94568bc7 ez/doc-duplicated-words-fix later to maint).
461 (merge 01e4be6c3d ks/test-readme-phrasofix later to maint).
462 (merge 217bb56d4f hn/typofix later to maint).
463 (merge c08fd6388c jk/doc-read-tree-table-asciidoctor-fix later to maint).
464 (merge c3342b362e ks/doc-use-camelcase-for-config-name later to maint).
465 (merge 0bca165fdb jk/validate-headref-fix later to maint).
466 (merge 93dbefb389 mr/doc-negative-pathspec later to maint).
467 (merge 5e633326e4 ad/doc-markup-fix later to maint).
projects / git / blob