4 Backward compatibility notes
5 ----------------------------
10 UI, Workflows & Features
12 * "git pull --rebase --verify-signature" learned to warn the user
13 that "--verify-signature" is a no-op when rebasing.
15 * An upstream project can make a recommendation to shallowly clone
16 some submodules in the .gitmodules file it ships.
18 * "git worktree add" learned that '-' can be used as a short-hand for
19 "@{-1}", the previous branch.
21 * Update the funcname definition to support css files.
23 * The completion script (in contrib/) learned to complete "git
26 * Messages that are generated by auto gc during "git push" on the
27 receiving end are now passed back to the sending end in such a way
28 that they are shown with "remote: " prefix to avoid confusing the
31 * "git add -i/-p" learned to honor diff.compactionHeuristic
32 experimental knob, so that the user can work on the same hunk split
35 * "upload-pack" allows a custom "git pack-objects" replacement when
36 responding to "fetch/clone" via the uploadpack.packObjectsHook.
37 (merge b738396 jk/upload-pack-hook later to maint).
39 * Teach format-patch and mailsplit (hence "am") how a line that
40 happens to begin with "From " in the e-mail message is quoted with
41 ">", so that these lines can be restored to their original shape.
42 (merge d9925d1 ew/mboxrd-format-am later to maint).
44 * "git repack" learned the "--keep-unreachable" option, which sends
45 loose unreachable objects to a pack instead of leaving them loose.
46 This helps heuristics based on the number of loose objects
48 (merge e26a8c4 jk/repack-keep-unreachable later to maint).
50 * "log --graph --format=" learned that "%>|(N)" specifies the width
51 relative to the terminal's left edge, not relative to the area to
52 draw text that is to the right of the ancestry-graph section. It
53 also now accepts negative N that means the column limit is relative
56 * A careless invocation of "git send-email directory/" after editing
57 0001-change.patch with an editor often ends up sending both
58 0001-change.patch and its backup file, 0001-change.patch~, causing
59 embarrassment and a minor confusion. Detect such an input and
60 offer to skip the backup files when sending the patches out.
61 (merge 531220b jc/send-email-skip-backup later to maint).
63 * "git submodule update" that drives many "git clone" could
64 eventually hit flaky servers/network conditions on one of the
65 submodules; the command learned to retry the attempt.
67 * The output coloring scheme learned two new attributes, italic and
68 strike, in addition to existing bold, reverse, etc.
70 * "git log" learns log.showSignature configuration variable, and a
71 command line option "--no-show-signature" to countermand it.
72 (merge fce04c3 mj/log-show-signature-conf later to maint).
74 * A couple of "git svn" updates.
76 * More markings of messages for i18n, with updates to various tests
77 to pass GETTEXT_POISON tests.
79 * "git archive" learned to handle files that are larger than 8GB and
80 commits far in the future than expressible by the traditional US-TAR
82 (merge 5caeeb8 jk/big-and-future-archive-tar later to maint).
84 * A new configuration variable core.sshCommand has been added to
85 specify what value for GIT_SSH_COMMAND to use per repository.
87 * "git worktree prune" protected worktrees that are marked as
88 "locked" by creating a file in a known location. "git worktree"
89 command learned a dedicated command pair to create and remove such
90 a file, so that the users do not have to do this with editor.
92 * A handful of "git svn" updates.
94 * "git push" learned to accept and pass extra options to the
95 receiving end so that hooks can read and react to them.
97 * "git status" learned to suggest "merge --abort" during a conflicted
98 merge, just like it already suggests "rebase --abort" during a
100 (merge b0a61ab mm/status-suggest-merge-abort later to maint).
102 * "git jump" script (in contrib/) has been updated a bit.
103 (merge a91e692 jk/git-jump later to maint).
105 * "git push" and "git clone" learned to give better progress meters
106 to the end user who is waiting on the terminal.
108 * An entry "git log --decorate" for the tip of the current branch is
109 shown as "HEAD -> name" (where "name" is the name of the branch);
110 paint the arrow in the same color as "HEAD", not in the color for
114 Performance, Internal Implementation, Development Support etc.
116 * "git fast-import" learned the same performance trick to avoid
117 creating too small a packfile as "git fetch" and "git push" have,
118 using *.unpackLimit configuration.
120 * When "git daemon" is run without --[init-]timeout specified, a
121 connection from a client that silently goes offline can hang around
122 for a long time, wasting resources. The socket-level KEEPALIVE has
123 been enabled to allow the OS to notice such failed connections.
125 * "git upload-pack" command has been updated to use the parse-options
128 * The "git apply" standalone program is being libified; this is the
129 first step to move many state variables into a structure that can
130 be explicitly (re)initialized to make the machinery callable more
133 * HTTP transport gained an option to produce more detailed debugging
135 (merge 73e57aa ep/http-curl-trace later to maint).
137 * Instead of taking advantage of a struct string_list that is
138 allocated with all NULs happens to be STRING_LIST_INIT_NODUP kind,
139 initialize them explicitly as such, to document their behaviour
141 (merge 2721ce2 jk/string-list-static-init later to maint).
143 * HTTPd tests learned to show the server error log to help diagnosing
145 (merge 44f243d nd/test-lib-httpd-show-error-log-in-verbose later to maint).
147 * The ownership rule for the piece of memory that hold references to
148 be fetched in "git fetch" was screwy, which has been cleaned up.
150 * "git bisect" makes an internal call to "git diff-tree" when
151 bisection finds the culprit, but this call did not initialize the
152 data structure to pass to the diff-tree API correctly.
154 * Further preparatory clean-up for "worktree" feature continues.
155 (merge 0409e0b nd/worktree-cleanup-post-head-protection later to maint).
157 * Formats of the various data (and how to validate them) where we use
158 GPG signature have been documented.
160 * A new run-command API function pipe_command() is introduced to
161 sanely feed data to the standard input while capturing data from
162 the standard output and the standard error of an external process,
163 which is cumbersome to hand-roll correctly without deadlocking.
165 The codepath to sign data in a prepared buffer with GPG has been
166 updated to use this API to read from the status-fd to check for
167 errors (instead of relying on GPG's exit status).
168 (merge efee955 jk/gpg-interface-cleanup later to maint).
170 * Allow t/perf framework to use the features from the most recent
171 version of Git even when testing an older installed version.
173 * The commands in the "log/diff" family have had an FILE* pointer in the
174 data structure they pass around for a long time, but some codepaths
175 used to always write to the standard output. As a preparatory step
176 to make "git format-patch" available to the internal callers, these
177 codepaths have been updated to consistently write into that FILE*
180 * Conversion from unsigned char sha1[20] to struct object_id
183 * Improve the look of the way "git fetch" reports what happened to
184 each ref that was fetched.
186 * The .c/.h sources are marked as such in our .gitattributes file so
187 that "git diff -W" and friends would work better.
188 (merge e82675a rs/help-c-source-with-gitattributes later to maint).
190 * Code clean-up to avoid using a variable string that compilers may
191 feel untrustable as printf-style format given to write_file()
194 * "git p4" used a location outside $GIT_DIR/refs/ to place its
195 temporary branches, which has been moved to refs/git-p4-tmp/.
197 * Existing autoconf generated test for the need to link with pthread
198 library did not check all the functions from pthread libraries;
199 recent FreeBSD has some functions in libc but not others, and we
200 mistakenly thought linking with libc is enough when it is not.
201 (merge a9b02de ew/autoconf-pthread later to maint).
203 * When "git fsck" reports a broken link (e.g. a tree object contains
204 a blob that does not exist), both containing object and the object
205 that is referred to were reported with their 40-hex object names.
206 The command learned the "--name-objects" option to show the path to
207 the containing object from existing refs (e.g. "HEAD~24^2:file.txt").
209 * Allow http daemon tests in Travis CI tests.
210 (merge d9d1426 ls/travis-enable-httpd-tests later to maint).
212 * Makefile assumed that -lrt is always available on platforms that
213 want to use clock_gettime() and CLOCK_MONOTONIC, which is not a
214 case for recent Mac OS X. The necessary symbols are often found in
215 libc on many modern systems and having -lrt on the command line, as
216 long as the library exists, had no effect, but when the platform
217 removes librt.a that is a different matter--having -lrt will break
220 This change could be seen as a regression for those who do need to
221 specify -lrt, as they now specifically ask for NEEDS_LIBRT when
222 building. Hopefully they are in the minority these days.
224 * Further preparatory work on the refs API before the pluggable
225 backend series can land.
227 * Error handling in the codepaths that updates refs has been
230 * The API to iterate over all the refs (i.e. for_each_ref(), etc.)
233 * A few tests that specifically target "git rebase -i" have been
236 * Dumb http transport on the client side has been optimized.
237 (merge ecba195 ew/http-walker later to maint).
239 * Users of the parse_options_concat() API function need to allocate
240 extra slots in advance and fill them with OPT_END() when they want
241 to decide the set of supported options dynamically, which makes the
242 code error-prone and hard to read. This has been corrected by tweaking
243 the API to allocate and return a new copy of "struct option" array.
244 (merge 023ff39 jk/parse-options-concat later to maint).
246 * "git fetch" exchanges batched have/ack messages between the sender
247 and the receiver, initially doubling every time and then falling
248 back to enlarge the window size linearly. The "smart http"
249 transport, being an half-duplex protocol, outgrows the preset limit
250 too quickly and becomes inefficient when interacting with a large
251 repository. The internal mechanism learned to grow the window size
252 more aggressively when working with the "smart http" transport.
254 * Tests for "git svn" have been taught to reuse the lib-httpd test
255 infrastructure when testing the subversion integration that
256 interacts with subversion repositories served over the http://
258 (merge a8a5d25 ew/git-svn-http-tests later to maint).
260 * "git pack-objects" has a few options that tell it not to pack
261 objects found in certain packfiles, which require it to scan .idx
262 files of all available packs. The codepaths involved in these
263 operations have been optimized for a common case of not having any
264 non-local pack and/or any .kept pack.
266 * The t3700 test about "add --chmod=-x" have been made a bit more
267 robust and generally cleaned up.
268 (merge 766cdc4 ib/t3700-add-chmod-x-updates later to maint).
270 * The build procedure learned PAGER_ENV knob that lists what default
271 environment variable settings to export for popular pagers. This
272 mechanism is used to tweak the default settings to MORE on FreeBSD.
273 (merge 995bc22 ew/build-time-pager-tweaks later to maint).
276 Also contains various documentation updates and code clean-ups.
282 Unless otherwise noted, all the fixes since v2.8 in the maintenance
283 track are contained in this release (see the maintenance releases'
286 * The commands in `git log` family take %C(auto) in a custom format
287 string. This unconditionally turned the color on, ignoring
288 --no-color or with --color=auto when the output is not connected to
289 a tty; this was corrected to make the format truly behave as
292 * "git rev-list --count" whose walk-length is limited with "-n"
293 option did not work well with the counting optimized to look at the
296 * "git show -W" (extend hunks to cover the entire function, delimited
297 by lines that match the "funcname" pattern) used to show the entire
298 file when a change added an entire function at the end of the file,
299 which has been fixed.
301 * The documentation set has been updated so that literal commands,
302 configuration variables and environment variables are consistently
303 typeset in fixed-width font and bold in manpages.
305 * "git svn propset" subcommand that was added in 2.3 days is
308 * The documentation tries to consistently spell "GPG"; when
309 referring to the specific program name, "gpg" is used.
311 * "git reflog" stopped upon seeing an entry that denotes a branch
312 creation event (aka "unborn"), which made it appear as if the
313 reflog was truncated.
315 * The git-prompt scriptlet (in contrib/) was not friendly with those
316 who uses "set -u", which has been fixed.
318 * compat/regex code did not cleanly compile.
320 * A codepath that used alloca(3) to place an unbounded amount of data
321 on the stack has been updated to avoid doing so.
323 * "git update-index --add --chmod=+x file" may be usable as an escape
324 hatch, but not a friendly thing to force for people who do need to
325 use it regularly. "git add --chmod=+x file" can be used instead.
327 * Build improvements for gnome-keyring (in contrib/)
329 * "git status" used to say "working directory" when it meant "working
332 * Comments about misbehaving FreeBSD shells have been clarified with
333 the version number (9.x and before are broken, newer ones are OK).
335 * "git cherry-pick A" worked on an unborn branch, but "git
336 cherry-pick A..B" didn't.
338 * Fix an unintended regression in v2.9 that breaks "clone --depth"
339 that recurses down to submodules by forcing the submodules to also
340 be cloned shallowly, which many server instances that host upstream
341 of the submodules are not prepared for.
343 * Fix unnecessarily waste in the idiomatic use of ': ${VAR=default}'
344 to set the default value, without enclosing it in double quotes.
346 * Some platform-specific code had non-ANSI strict declarations of C
347 functions that do not take any parameters, which has been
350 * The internal code used to show local timezone offset is not
351 prepared to handle timestamps beyond year 2100, and gave a
352 bogus offset value to the caller. Use a more benign looking
353 +0000 instead and let "git log" going in such a case, instead
356 * One among four invocations of readlink(1) in our test suite has
357 been rewritten so that the test can run on systems without the
358 command (others are in valgrind test framework and t9802).
360 * t/perf needs /usr/bin/time with GNU extension; the invocation of it
361 is updated to "gtime" on Darwin.
363 * A bug, which caused "git p4" while running under verbose mode to
364 report paths that are omitted due to branch prefix incorrectly, has
365 been fixed; the command said "Ignoring file outside of prefix" for
366 paths that are _inside_.
368 * The top level documentation "git help git" still pointed at the
369 documentation set hosted at now-defunct google-code repository.
370 Update it to point to https://git.github.io/htmldocs/git.html
373 * A helper function that takes the contents of a commit object and
374 finds its subject line did not ignore leading blank lines, as is
375 commonly done by other codepaths. Make it ignore leading blank
378 * For a long time, we carried an in-code comment that said our
379 colored output would work only when we use fprintf/fputs on
380 Windows, which no longer is the case for the past few years.
382 * "gc.autoPackLimit" when set to 1 should not trigger a repacking
383 when there is only one pack, but the code counted poorly and did
386 * Add a test to specify the desired behaviour that currently is not
387 available in "git rebase -Xsubtree=...".
389 * More mark-up updates to typeset strings that are expected to
390 literally typed by the end user in fixed-width font.
392 * "git commit --amend --allow-empty-message -S" for a commit without
393 any message body could have misidentified where the header of the
396 * "git rebase -i --autostash" did not restore the auto-stashed change
397 when the operation was aborted.
399 * Git does not know what the contents in the index should be for a
400 path added with "git add -N" yet, so "git grep --cached" should not
401 show hits (or show lack of hits, with -L) in such a path, but that
402 logic does not apply to "git grep", i.e. searching in the working
403 tree files. But we did so by mistake, which has been corrected.
405 * "git blame -M" missed a single line that was moved within the file.
407 * Fix recently introduced codepaths that are involved in parallel
408 submodule operations, which gave up on reading too early, and
409 could have wasted CPU while attempting to write under a corner
412 * "git grep -i" has been taught to fold case in non-ascii locales
415 * A test that unconditionally used "mktemp" learned that the command
416 is not necessarily available everywhere.
418 * There are certain house-keeping tasks that need to be performed at
419 the very beginning of any Git program, and programs that are not
420 built-in commands had to do them exactly the same way as "git"
421 potty does. It was easy to make mistakes in one-off standalone
422 programs (like test helpers). A common "main()" function that
423 calls cmd_main() of individual program has been introduced to
424 make it harder to make mistakes.
425 (merge de61ceb jk/common-main later to maint).
427 * The test framework learned a new helper test_match_signal to
428 check an exit code from getting killed by an expected signal.
430 * General code clean-up around a helper function to write a
431 single-liner to a file.
432 (merge 7eb6e10 jk/write-file later to maint).
434 * One part of "git am" had an oddball helper function that called
435 stuff from outside "his" as opposed to calling what we have "ours",
436 which was not gender-neutral and also inconsistent with the rest of
437 the system where outside stuff is usuall called "theirs" in
440 * "git blame file" allowed the lineage of lines in the uncommitted,
441 unadded contents of "file" to be inspected, but it refused when
442 "file" did not appear in the current commit. When "file" was
443 created by renaming an existing file (but the change has not been
444 committed), this restriction was unnecessarily tight.
446 * "git add -N dir/file && git write-tree" produced an incorrect tree
447 when there are other paths in the same directory that sorts after
450 * "git fetch http://user:pass@host/repo..." scrubbed the userinfo
451 part, but "git push" didn't.
453 * "git merge" with renormalization did not work well with
454 merge-recursive, due to "safer crlf" conversion kicking in when it
456 (merge 1335d76 jc/renormalize-merge-kill-safer-crlf later to maint).
458 * The use of strbuf in "git rm" to build filename to remove was a bit
459 suboptimal, which has been fixed.
460 (merge deb8e15 rs/rm-strbuf-optim later to maint).
462 * An age old bug that caused "git diff --ignore-space-at-eol"
463 misbehave has been fixed.
465 * "git notes merge" had a code to see if a path exists (and fails if
466 it does) and then open the path for writing (when it doesn't).
467 Replace it with open with O_EXCL.
469 * "git pack-objects" and "git index-pack" mostly operate with off_t
470 when talking about the offset of objects in a packfile, but there
471 were a handful of places that used "unsigned long" to hold that
472 value, leading to an unintended truncation.
474 * Recent update to "git daemon" tries to enable the socket-level
475 KEEPALIVE, but when it is spawned via inetd, the standard input
476 file descriptor may not necessarily be connected to a socket.
477 Suppress an ENOTSOCK error from setsockopt().
479 * Recent FreeBSD stopped making perl available at /usr/bin/perl;
480 switch the default the built-in path to /usr/local/bin/perl on not
481 too ancient FreeBSD releases.
483 * "git commit --help" said "--no-verify" is only about skipping the
484 pre-commit hook, and failed to say that it also skipped the
486 (merge def480f os/no-verify-skips-commit-msg-too later to maint).
488 * "git merge" in Git v2.9 was taught to forbid merging an unrelated
489 lines of history by default, but that is exactly the kind of thing
490 the "--rejoin" mode of "git subtree" (in contrib/) wants to do.
491 "git subtree" has been taught to use the "--allow-unrelated-histories"
492 option to override the default.
493 (merge 0f12c7d da/subtree-2.9-regression later to maint).
495 * The build procedure for "git persistent-https" helper (in contrib/)
496 has been updated so that it can be built with more recent versions
498 (merge accb613 pm/build-persistent-https-with-recent-go later to maint).
500 * There is an optimization used in "git diff $treeA $treeB" to borrow
501 an already checked-out copy in the working tree when it is known to
502 be the same as the blob being compared, expecting that open/mmap of
503 such a file is faster than reading it from the object store, which
504 involves inflating and applying delta. This however kicked in even
505 when the checked-out copy needs to go through the convert-to-git
506 conversion (including the clean filter), which defeats the whole
507 point of the optimization. The optimization has been disabled when
508 the conversion is necessary.
509 (merge 06dec43 jk/diff-do-not-reuse-wtf-needs-cleaning later to maint).
511 * "git -c grep.patternType=extended log --basic-regexp" misbehaved
512 because the internal API to access the grep machinery was not
514 (merge 8465541 jc/grep-commandline-vs-configuration later to maint).
516 * Windows port was failing some tests in t4130, due to the lack of
517 inum in the returned values by its lstat(2) emulation.
518 (merge 54956df js/t4130-rename-without-ino later to maint).
520 * The reflog output format is documented better, and a new format
521 --date=unix to report the seconds-since-epoch (without timezone)
523 (merge 442f6fd jk/reflog-date later to maint).
525 * "git difftool <paths>..." started in a subdirectory failed to
526 interpret the paths relative to that directory, which has been
528 (merge 32b8c58 jk/difftool-in-subdir later to maint).
530 * The characters in the label shown for tags/refs for commits in
531 "gitweb" output are now properly escaped for proper HTML output.
532 (merge 77947bb ab/gitweb-link-html-escape later to maint).
534 * FreeBSD can lie when asked mtime of a directory, which made the
535 untracked cache code to fall back to a slow-path, which in turn
536 caused tests in t7063 to fail because it wanted to verify the
537 behaviour of the fast-path.
538 (merge 6b7728d nd/fbsd-lazy-mtime later to maint).
540 * Squelch compiler warnings for netmalloc (in compat/) library.
541 (merge c6c9e18 js/nedmalloc-gcc6-warnings later to maint).
543 * A small memory leak in the command line parsing of "git blame"
546 * The API documentation for hashmap was unclear if hashmap_entry
547 can be safely discarded without any other consideration. State
548 that it is safe to do so.
549 (merge 54ba5a1 jc/hashmap-doc-init later to maint).
551 * Other minor clean-ups and documentation updates
552 (merge 9d1ca1d jk/t4205-cleanup later to maint).
553 (merge 5f072e0 cp/completion-clone-recurse-submodules later to maint).
554 (merge bc57b9c rs/use-strbuf-addstr later to maint).
555 (merge 52db4b0 jh/clean-smudge-f-doc later to maint).
projects / git / blob