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 20b20a2 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.
88 Performance, Internal Implementation, Development Support etc.
90 * "git fast-import" learned the same performance trick to avoid
91 creating too small a packfile as "git fetch" and "git push" have,
92 using *.unpackLimit configuration.
94 * When "git daemon" is run without --[init-]timeout specified, a
95 connection from a client that silently goes offline can hang around
96 for a long time, wasting resources. The socket-level KEEPALIVE has
97 been enabled to allow the OS to notice such failed connections.
99 * "git upload-pack" command has been updated to use the parse-options
102 * The "git apply" standalone program is being libified; this is the
103 first step to move many state variables into a structure that can
104 be explicitly (re)initialized to make the machinery callable more
107 * HTTP transport gained an option to produce more detailed debugging
109 (merge 73e57aa ep/http-curl-trace later to maint).
111 * Instead of taking advantage of a struct string_list that is
112 allocated with all NULs happens to be STRING_LIST_INIT_NODUP kind,
113 initialize them explicitly as such, to document their behaviour
115 (merge 2721ce2 jk/string-list-static-init later to maint).
117 * HTTPd tests learned to show the server error log to help diagnosing
119 (merge 44f243d nd/test-lib-httpd-show-error-log-in-verbose later to maint).
121 * The ownership rule for the piece of memory that hold references to
122 be fetched in "git fetch" was screwy, which has been cleaned up.
124 * "git bisect" makes an internal call to "git diff-tree" when
125 bisection finds the culprit, but this call did not initialize the
126 data structure to pass to the diff-tree API correctly.
128 * Further preparatory clean-up for "worktree" feature continues.
129 (merge 0409e0b nd/worktree-cleanup-post-head-protection later to maint).
131 * Formats of the various data (and how to validate them) where we use
132 GPG signature have been documented.
134 * A new run-command API function pipe_command() is introduced to
135 sanely feed data to the standard input while capturing data from
136 the standard output and the standard error of an external process,
137 which is cumbersome to hand-roll correctly without deadlocking.
139 The codepath to sign data in a prepared buffer with GPG has been
140 updated to use this API to read from the status-fd to check for
141 errors (instead of relying on GPG's exit status).
142 (merge efee955 jk/gpg-interface-cleanup later to maint).
144 * Allow t/perf framework to use the features from the most recent
145 version of Git even when testing an older installed version.
147 * The commands in the "log/diff" family have had an FILE* pointer in the
148 data structure they pass around for a long time, but some codepaths
149 used to always write to the standard output. As a preparatory step
150 to make "git format-patch" available to the internal callers, these
151 codepaths have been updated to consistently write into that FILE*
154 * Conversion from unsigned char sha1[20] to struct object_id
157 * Improve the look of the way "git fetch" reports what happened to
158 each ref that was fetched.
159 (merge bc437d1 nd/fetch-ref-summary later to maint).
161 * The .c/.h sources are marked as such in our .gitattributes file so
162 that "git diff -W" and friends would work better.
163 (merge e82675a rs/help-c-source-with-gitattributes later to maint).
165 * Code clean-up to avoid using a variable string that compilers may
166 feel untrustable as printf-style format given to write_file()
169 * "git p4" used a location outside $GIT_DIR/refs/ to place its
170 temporary branches, which has been moved to refs/git-p4-tmp/.
172 * Existing autoconf generated test for the need to link with pthread
173 library did not check all the functions from pthread libraries;
174 recent FreeBSD has some functions in libc but not others, and we
175 mistakenly thought linking with libc is enough when it is not.
176 (merge a9b02de ew/autoconf-pthread later to maint).
178 * When "git fsck" reports a broken link (e.g. a tree object contains
179 a blob that does not exist), both containing object and the object
180 that is referred to were reported with their 40-hex object names.
181 The command learned the "--name-objects" option to show the path to
182 the containing object from existing refs (e.g. "HEAD~24^2:file.txt").
184 * Allow http daemon tests in Travis CI tests.
185 (merge d9d1426 ls/travis-enable-httpd-tests later to maint).
187 * Makefile assumed that -lrt is always available on platforms that
188 want to use clock_gettime() and CLOCK_MONOTONIC, which is not a
189 case for recent Mac OS X. The necessary symbols are often found in
190 libc on many modern systems and having -lrt on the command line, as
191 long as the library exists, had no effect, but when the platform
192 removes librt.a that is a different matter--having -lrt will break
195 This change could be seen as a regression for those who do need to
196 specify -lrt, as they now specifically ask for NEEDS_LIBRT when
197 building. Hopefully they are in the minority these days.
199 * Further preparatory work on the refs API before the pluggable
200 backend series can land.
202 * Error handling in the codepaths that updates refs has been
205 * The API to iterate over all the refs (i.e. for_each_ref(), etc.)
209 Also contains various documentation updates and code clean-ups.
215 Unless otherwise noted, all the fixes since v2.8 in the maintenance
216 track are contained in this release (see the maintenance releases'
219 * The commands in `git log` family take %C(auto) in a custom format
220 string. This unconditionally turned the color on, ignoring
221 --no-color or with --color=auto when the output is not connected to
222 a tty; this was corrected to make the format truly behave as
225 * "git rev-list --count" whose walk-length is limited with "-n"
226 option did not work well with the counting optimized to look at the
229 * "git show -W" (extend hunks to cover the entire function, delimited
230 by lines that match the "funcname" pattern) used to show the entire
231 file when a change added an entire function at the end of the file,
232 which has been fixed.
234 * The documentation set has been updated so that literal commands,
235 configuration variables and environment variables are consistently
236 typeset in fixed-width font and bold in manpages.
238 * "git svn propset" subcommand that was added in 2.3 days is
241 * The documentation tries to consistently spell "GPG"; when
242 referring to the specific program name, "gpg" is used.
244 * "git reflog" stopped upon seeing an entry that denotes a branch
245 creation event (aka "unborn"), which made it appear as if the
246 reflog was truncated.
248 * The git-prompt scriptlet (in contrib/) was not friendly with those
249 who uses "set -u", which has been fixed.
251 * compat/regex code did not cleanly compile.
253 * A codepath that used alloca(3) to place an unbounded amount of data
254 on the stack has been updated to avoid doing so.
256 * "git update-index --add --chmod=+x file" may be usable as an escape
257 hatch, but not a friendly thing to force for people who do need to
258 use it regularly. "git add --chmod=+x file" can be used instead.
260 * Build improvements for gnome-keyring (in contrib/)
262 * "git status" used to say "working directory" when it meant "working
265 * Comments about misbehaving FreeBSD shells have been clarified with
266 the version number (9.x and before are broken, newer ones are OK).
268 * "git cherry-pick A" worked on an unborn branch, but "git
269 cherry-pick A..B" didn't.
271 * Fix an unintended regression in v2.9 that breaks "clone --depth"
272 that recurses down to submodules by forcing the submodules to also
273 be cloned shallowly, which many server instances that host upstream
274 of the submodules are not prepared for.
276 * Fix unnecessarily waste in the idiomatic use of ': ${VAR=default}'
277 to set the default value, without enclosing it in double quotes.
279 * Some platform-specific code had non-ANSI strict declarations of C
280 functions that do not take any parameters, which has been
283 * The internal code used to show local timezone offset is not
284 prepared to handle timestamps beyond year 2100, and gave a
285 bogus offset value to the caller. Use a more benign looking
286 +0000 instead and let "git log" going in such a case, instead
288 (merge bab7483 jk/tzoffset-fix later to maint).
290 * One among four invocations of readlink(1) in our test suite has
291 been rewritten so that the test can run on systems without the
292 command (others are in valgrind test framework and t9802).
294 * t/perf needs /usr/bin/time with GNU extension; the invocation of it
295 is updated to "gtime" on Darwin.
297 * A bug, which caused "git p4" while running under verbose mode to
298 report paths that are omitted due to branch prefix incorrectly, has
299 been fixed; the command said "Ignoring file outside of prefix" for
300 paths that are _inside_.
302 * The top level documentation "git help git" still pointed at the
303 documentation set hosted at now-defunct google-code repository.
304 Update it to point to https://git.github.io/htmldocs/git.html
307 * A helper function that takes the contents of a commit object and
308 finds its subject line did not ignore leading blank lines, as is
309 commonly done by other codepaths. Make it ignore leading blank
311 (merge 054a5ae js/find-commit-subject-ignore-leading-blanks later to maint).
313 * For a long time, we carried an in-code comment that said our
314 colored output would work only when we use fprintf/fputs on
315 Windows, which no longer is the case for the past few years.
316 (merge 3d0a833 js/color-on-windows-comment later to maint).
318 * "gc.autoPackLimit" when set to 1 should not trigger a repacking
319 when there is only one pack, but the code counted poorly and did
321 (merge 5f4e3bf ew/gc-auto-pack-limit-fix later to maint).
323 * Add a test to specify the desired behaviour that currently is not
324 available in "git rebase -Xsubtree=...".
325 (merge 5f35900 dg/subtree-rebase-test later to maint).
327 * More mark-up updates to typeset strings that are expected to
328 literally typed by the end user in fixed-width font.
329 (merge 661c3e9 mm/doc-tt later to maint).
331 * "git commit --amend --allow-empty-message -S" for a commit without
332 any message body could have misidentified where the header of the
334 (merge 3324dd8 js/sign-empty-commit-fix later to maint).
336 * "git rebase -i --autostash" did not restore the auto-stashed change
337 when the operation was aborted.
338 (merge 33ba9c6 ps/rebase-i-auto-unstash-upon-abort later to maint).
340 * Git does not know what the contents in the index should be for a
341 path added with "git add -N" yet, so "git grep --cached" should not
342 show hits (or show lack of hits, with -L) in such a path, but that
343 logic does not apply to "git grep", i.e. searching in the working
344 tree files. But we did so by mistake, which has been corrected.
345 (merge b8e47d1 nd/ita-cleanup later to maint).
347 * "git blame -M" missed a single line that was moved within the file.
348 (merge 17a07e2 dk/blame-move-no-reason-for-1-line-context later to maint).
350 * Fix recently introduced codepaths that are involved in parallel
351 submodule operations, which gave up on reading too early, and
352 could have wasted CPU while attempting to write under a corner
354 (merge d751dd1 sb/submodule-parallel-fetch later to maint).
356 * "git grep -i" has been taught to fold case in non-ascii locales
358 (merge 695f95b nd/icase later to maint).
360 * A test that unconditionally used "mktemp" learned that the command
361 is not necessarily available everywhere.
362 (merge c578a09 ak/lazy-prereq-mktemp later to maint).
364 * There are certain house-keeping tasks that need to be performed at
365 the very beginning of any Git program, and programs that are not
366 built-in commands had to do them exactly the same way as "git"
367 potty does. It was easy to make mistakes in one-off standalone
368 programs (like test helpers). A common "main()" function that
369 calls cmd_main() of individual program has been introduced to
370 make it harder to make mistakes.
371 (merge de61ceb jk/common-main later to maint).
373 * The test framework learned a new helper test_match_signal to
374 check an exit code from getting killed by an expected signal.
375 (merge 03c39b3 jk/test-match-signal later to maint).
377 * General code clean-up around a helper function to write a
378 single-liner to a file.
379 (merge 7eb6e10 jk/write-file later to maint).
381 * One part of "git am" had an oddball helper function that called
382 stuff from outside "his" as opposed to calling what we have "ours",
383 which was not gender-neutral and also inconsistent with the rest of
384 the system where outside stuff is usuall called "theirs" in
386 (merge 715a51b js/am-call-theirs-theirs-in-fallback-3way later to maint).
388 * "git blame file" allowed the lineage of lines in the uncommitted,
389 unadded contents of "file" to be inspected, but it refused when
390 "file" did not appear in the current commit. When "file" was
391 created by renaming an existing file (but the change has not been
392 committed), this restriction was unnecessarily tight.
393 (merge c66b470 mh/blame-worktree later to maint).
395 * "git add -N dir/file && git write-tree" produced an incorrect tree
396 when there are other paths in the same directory that sorts after
398 (merge 6d6a782 nd/cache-tree-ita later to maint).
400 * "git fetch http://user:pass@host/repo..." scrubbed the userinfo
401 part, but "git push" didn't.
402 (merge 68f3c07 jk/push-scrub-url later to maint).
404 * "git merge" with renormalization did not work well with
405 merge-recursive, due to "safer crlf" conversion kicking in when it
407 (merge 1335d76 jc/renormalize-merge-kill-safer-crlf later to maint).
409 * The use of strbuf in "git rm" to build filename to remove was a bit
410 suboptimal, which has been fixed.
411 (merge deb8e15 rs/rm-strbuf-optim later to maint).
413 * An age old bug that caused "git diff --ignore-space-at-eol"
414 misbehave has been fixed.
415 (merge 044fb19 js/ignore-space-at-eol later to maint).
417 * Other minor clean-ups and documentation updates
418 (merge e51b0df pb/commit-editmsg-path later to maint).
419 (merge b333d0d jk/send-pack-stdio later to maint).
420 (merge fcf0fe9 lf/sideband-returns-void later to maint).
421 (merge c2691e2 ah/unpack-trees-advice-messages later to maint).
422 (merge 82f6178 nd/doc-new-command later to maint).
423 (merge fa90ab4 js/t3404-grammo-fix later to maint).
424 (merge c61b2af lf/recv-sideband-cleanup later to maint).
425 (merge 31471ba rs/use-strbuf-addbuf later to maint).
426 (merge 503e224 nd/test-helpers later to maint).
427 (merge 16726cf jc/doc-diff-filter-exclude later to maint).
428 (merge fd2e7da rs/worktree-use-strbuf-absolute-path later to maint).