| blob | a60e92e0dd758c26d4fe83dd2e7a72a8bdd7f825 |
1 Git 2.9 Release Notes
2 =====================
4 Backward compatibility note
5 ---------------------------
7 The end-user facing Porcelain level commands in the "git diff" and
8 "git log" by default enables the rename detection; you can still use
9 "diff.renames" configuration variable to disable this.
11 Merging two branches that have no common ancestor with "git merge" is
12 by default forbidden now to prevent creating such an unusual merge by
13 mistake.
15 The output formats of "git log" that indents the commit log message by
16 4 spaces now expands HT in the log message by default. You can use
17 the "--no-expand-tabs" option to disable this.
19 "git commit-tree" plumbing command required the user to always sign
20 its result when the user sets the commit.gpgsign configuration
21 variable, which was an ancient mistake, which this release corrects.
22 A script that drives commit-tree, if it relies on this mistake, now
23 needs to read commit.gpgsign and pass the -S option as necessary.
26 Updates since v2.8
27 ------------------
29 UI, Workflows & Features
31 * Comes with git-multimail 1.3.1 (in contrib/).
33 * The end-user facing Porcelain level commands like "diff" and "log"
34 now enables the rename detection by default.
36 * The credential.helper configuration variable is cumulative and
37 there is no good way to override it from the command line. As
38 a special case, giving an empty string as its value now serves
39 as the signal to clear the values specified in various files.
41 * A new "interactive.diffFilter" configuration can be used to
42 customize the diff shown in "git add -i" session.
44 * "git p4" now allows P4 author names to be mapped to Git author
45 names.
47 * "git rebase -x" can be used without passing "-i" option.
49 * "git -c credential.<var>=<value> submodule" can now be used to
50 propagate configuration variables related to credential helper
51 down to the submodules.
53 * "git tag" can create an annotated tag without explicitly given an
54 "-a" (or "-s") option (i.e. when a tag message is given). A new
55 configuration variable, tag.forceSignAnnotated, can be used to tell
56 the command to create signed tag in such a situation.
58 * "git merge" used to allow merging two branches that have no common
59 base by default, which led to a brand new history of an existing
60 project created and then get pulled by an unsuspecting maintainer,
61 which allowed an unnecessary parallel history merged into the
62 existing project. The command has been taught not to allow this by
63 default, with an escape hatch "--allow-unrelated-histories" option
64 to be used in a rare event that merges histories of two projects
65 that started their lives independently.
67 * "git pull" has been taught to pass --allow-unrelated-histories
68 option to underlying "git merge".
70 * "git apply -v" learned to report paths in the patch that were
71 skipped via --include/--exclude mechanism or being outside the
72 current working directory.
74 * Shell completion (in contrib/) updates.
76 * The commit object name reported when "rebase -i" stops has been
77 shortened.
79 * "git worktree add" can be given "--no-checkout" option to only
80 create an empty worktree without checking out the files.
82 * "git mergetools" learned to drive ExamDiff.
84 * "git pull --rebase" learned "--[no-]autostash" option, so that
85 the rebase.autostash configuration variable set to true can be
86 overridden from the command line.
88 * When "git log" shows the log message indented by 4-spaces, the
89 remainder of a line after a HT does not align in the way the author
90 originally intended. The command now expands tabs by default in
91 such a case, and allows the users to override it with a new option,
92 "--no-expand-tabs".
94 * "git send-email" now uses a more readable timestamps when
95 formulating a message ID.
97 * "git rerere" can encounter two or more files with the same conflict
98 signature that have to be resolved in different ways, but there was
99 no way to record these separate resolutions.
100 (merge d9d501b068 jc/rerere-multi later to maint).
102 * "git p4" learned to record P4 jobs in Git commit that imports from
103 the history in Perforce.
105 * "git describe --contains" often made a hard-to-justify choice of
106 tag to give name to a given commit, because it tried to come up
107 with a name with smallest number of hops from a tag, causing an old
108 commit whose close descendant that is recently tagged were not
109 described with respect to an old tag but with a newer tag. It did
110 not help that its computation of "hop" count was further tweaked to
111 penalize being on a side branch of a merge. The logic has been
112 updated to favor using the tag with the oldest tagger date, which
113 is a lot easier to explain to the end users: "We describe a commit
114 in terms of the (chronologically) oldest tag that contains the
115 commit."
116 (merge 7550424 js/name-rev-use-oldest-ref later to maint).
118 * "git clone" learned "--shallow-submodules" option.
120 * HTTP transport clients learned to throw extra HTTP headers at the
121 server, specified via http.extraHeader configuration variable.
123 * Patch output from "git diff" and friends has been tweaked to be
124 more readable by using a blank line as a strong hint that the
125 contents before and after it belong to a logically separate unit.
127 * A new configuration variable core.hooksPath allows customizing
128 where the hook directory is.
130 * An earlier addition of "sanitize_submodule_env" with 14111fc4 (git:
131 submodule honor -c credential.* from command line, 2016-02-29)
132 turned out to be a convoluted no-op; implement what it wanted to do
133 correctly, and stop filtering settings given via "git -c var=val".
135 * "git commit --dry-run" reported "No, no, you cannot commit." in one
136 case where "git commit" would have allowed you to commit, and this
137 improves it a little bit ("git commit --dry-run --short" still does
138 not give you the correct answer, for example). This is a stop-gap
139 measure in that "commit --short --dry-run" still gives an incorrect
140 result.
142 * The experimental "multiple worktree" feature gains more safety to
143 forbid operations on a branch that is checked out or being actively
144 worked on elsewhere, by noticing that e.g. it is being rebased.
146 * "git format-patch" learned a new "--base" option to record what
147 (public, well-known) commit the original series was built on in
148 its output.
150 * "git commit" learned to pay attention to "commit.verbose"
151 configuration variable and act as if "--verbose" option was
152 given from the command line.
155 Performance, Internal Implementation, Development Support etc.
157 * The embedded args argv-array in the child process is used to build
158 the command line to run pack-objects instead of using a separate
159 array of strings.
161 * A test for tags has been restructured so that more parts of it can
162 easily be run on a platform without a working GnuPG.
164 * The startup_info data, which records if we are working inside a
165 repository (among other things), are now uniformly available to Git
166 subcommand implementations, and Git avoids attempting to touch
167 references when we are not in a repository.
169 * The command line argument parser for "receive-pack" has been
170 rewritten to use parse-options.
172 * A major part of "git submodule update" has been ported to C to take
173 advantage of the recently added framework to run download tasks in
174 parallel.
176 * Rename bunch of tests on "git clone" for better organization.
178 * The tests that involve running httpd leaked the system-wide
179 configuration in /etc/gitconfig to the tested environment.
181 * Build updates for MSVC.
183 * The repository set-up sequence has been streamlined (the biggest
184 change is that there is no longer git_config_early()), so that we
185 do not attempt to look into refs/* when we know we do not have a
186 Git repository.
188 * Code restructuring around the "refs" area to prepare for pluggable
189 refs backends.
191 * Sources to many test helper binaries (and the generated helpers)
192 have been moved to t/helper/ subdirectory to reduce clutter at the
193 top level of the tree.
195 * Unify internal logic between "git tag -v" and "git verify-tag"
196 commands by making one directly call into the other.
197 (merge bef234b st/verify-tag later to maint).
199 * "merge-recursive" strategy incorrectly checked if a path that is
200 involved in its internal merge exists in the working tree.
202 * The test scripts for "git p4" (but not "git p4" implementation
203 itself) has been updated so that they would work even on a system
204 where the installed version of Python is python 3.
206 * As nobody maintains our in-tree git.spec.in and distros use their
207 own spec file, we stopped pretending that we support "make rpm".
209 * Move from unsigned char[20] to struct object_id continues.
211 * Update of "git submodule" to move pieces of logic to C continues.
213 * The code for warning_errno/die_errno has been refactored and a new
214 error_errno() reporting helper is introduced.
215 (merge 1da045f nd/error-errno later to maint).
217 * Running tests with '-x' option to trace the individual command
218 executions is a useful way to debug test scripts, but some tests
219 that capture the standard error stream and check what the command
220 said can be broken with the trace output mixed in. When running
221 our tests under "bash", however, we can redirect the trace output
222 to another file descriptor to keep the standard error of programs
223 being tested intact.
224 (merge d88785e jk/test-send-sh-x-trace-elsewhere later to maint).
226 * t0040 had too many unnecessary repetitions in its test data. Teach
227 test-parse-options program so that a caller can tell what it
228 expects in its output, so that these repetitions can be cleaned up.
230 * Add perf test for "rebase -i"
232 * Common mistakes when writing gitlink: in our documentation are
233 found by "make check-docs".
235 * t9xxx series has been updated primarily for readability, while
236 fixing small bugs in it. A few scripted Porcelains have also been
237 updated to fix possible bugs around their use of "test -z" and
238 "test -n".
240 * CI test was taught to run git-svn tests.
243 Also contains various documentation updates and code clean-ups.
246 Fixes since v2.8
247 ----------------
249 Unless otherwise noted, all the fixes since v2.8 in the maintenance
250 track are contained in this release (see the maintenance releases'
251 notes for details).
253 * "git config --get-urlmatch", unlike other variants of the "git
254 config --get" family, did not signal error with its exit status
255 when there was no matching configuration.
257 * The "--local-env-vars" and "--resolve-git-dir" options of "git
258 rev-parse" failed to work outside a repository when the command's
259 option parsing was rewritten in 1.8.5 era.
261 * "git index-pack --keep[=<msg>] pack-$name.pack" simply did not work.
263 * Fetching of history by naming a commit object name directly didn't
264 work across remote-curl transport.
266 * A small memory leak in an error codepath has been plugged in xdiff
267 code.
269 * strbuf_getwholeline() did not NUL-terminate the buffer on certain
270 corner cases in its error codepath.
272 * "git mergetool" did not work well with conflicts that both sides
273 deleted.
275 * "git send-email" had trouble parsing alias file in mailrc format
276 when lines in it had trailing whitespaces on them.
278 * When "git merge --squash" stopped due to conflict, the concluding
279 "git commit" failed to read in the SQUASH_MSG that shows the log
280 messages from all the squashed commits.
282 * "git merge FETCH_HEAD" dereferenced NULL pointer when merging
283 nothing into an unborn history (which is arguably unusual usage,
284 which perhaps was the reason why nobody noticed it).
286 * When "git worktree" feature is in use, "git branch -d" allowed
287 deletion of a branch that is checked out in another worktree,
288 which was wrong.
290 * When "git worktree" feature is in use, "git branch -m" renamed a
291 branch that is checked out in another worktree without adjusting
292 the HEAD symbolic ref for the worktree.
294 * "git diff -M" used to work better when two originally identical
295 files A and B got renamed to X/A and X/B by pairing A to X/A and B
296 to X/B, but this was broken in the 2.0 timeframe.
298 * "git send-pack --all <there>" was broken when its command line
299 option parsing was written in the 2.6 timeframe.
301 * "git format-patch --help" showed `-s` and `--no-patch` as if these
302 are valid options to the command. We already hide `--patch` option
303 from the documentation, because format-patch is about showing the
304 diff, and the documentation now hides these options as well.
306 * When running "git blame $path" with unnormalized data in the index
307 for the path, the data in the working tree was blamed, even though
308 "git add" would not have changed what is already in the index, due
309 to "safe crlf" that disables the line-end conversion. It has been
310 corrected.
312 * A change back in version 2.7 to "git branch" broke display of a
313 symbolic ref in a non-standard place in the refs/ hierarchy (we
314 expect symbolic refs to appear in refs/remotes/*/HEAD to point at
315 the primary branch the remote has, and as .git/HEAD to point at the
316 branch we locally checked out).
318 * A partial rewrite of "git submodule" in the 2.7 timeframe changed
319 the way the gitdir: pointer in the submodules point at the real
320 repository location to use absolute paths by accident. This has
321 been corrected.
323 * "git commit" misbehaved in a few minor ways when an empty message
324 is given via -m '', all of which has been corrected.
326 * Support for CRAM-MD5 authentication method in "git imap-send" did
327 not work well.
329 * Upcoming OpenSSL 1.1.0 will break compilation b updating a few APIs
330 we use in imap-send, which has been adjusted for the change.
332 * The socks5:// proxy support added back in 2.6.4 days was not aware
333 that socks5h:// proxies behave differently.
335 * "git config" had a codepath that tried to pass a NULL to
336 printf("%s"), which nobody seems to have noticed.
338 * On Cygwin, object creation uses the "create a temporary and then
339 rename it to the final name" pattern, not "create a temporary,
340 hardlink it to the final name and then unlink the temporary"
341 pattern.
343 This is necessary to use Git on Windows shared directories, and is
344 already enabled for the MinGW and plain Windows builds. It also
345 has been used in Cygwin packaged versions of Git for quite a while.
346 See http://thread.gmane.org/gmane.comp.version-control.git/291853
348 * "merge-octopus" strategy did not ensure that the index is clean
349 when merge begins.
351 * When "git merge" notices that the merge can be resolved purely at
352 the tree level (without having to merge blobs) and the resulting
353 tree happens to already exist in the object store, it forgot to
354 update the index, which lead to an inconsistent state for later
355 operations.
357 * "git submodule" reports the paths of submodules the command
358 recurses into, but this was incorrect when the command was not run
359 from the root level of the superproject.
361 * The "user.useConfigOnly" configuration variable makes it an error
362 if users do not explicitly set user.name and user.email. However,
363 its check was not done early enough and allowed another error to
364 trigger, reporting that the default value we guessed from the
365 system setting was unusable. This was a suboptimal end-user
366 experience as we want the users to set user.name/user.email without
367 relying on the auto-detection at all.
369 * "git mv old new" did not adjust the path for a submodule that lives
370 as a subdirectory inside old/ directory correctly.
372 * "git replace -e" did not honour "core.editor" configuration.
374 * "git push" from a corrupt repository that attempts to push a large
375 number of refs deadlocked; the thread to relay rejection notices
376 for these ref updates blocked on writing them to the main thread,
377 after the main thread at the receiving end notices that the push
378 failed and decides not to read these notices and return a failure.
380 * mmap emulation on Windows has been optimized and work better without
381 consuming paging store when not needed.
383 * A question by "git send-email" to ask the identity of the sender
384 has been updated.
386 * UI consistency improvements for "git mergetool".
388 * "git rebase -m" could be asked to rebase an entire branch starting
389 from the root, but failed by assuming that there always is a parent
390 commit to the first commit on the branch.
391 (merge 79f4344 bw/rebase-merge-entire-branch later to maint).
393 * Fix a broken "p4 lfs" test.
395 * Recent update to Git LFS broke "git p4" by changing the output from
396 its "lfs pointer" subcommand.
398 * "git fetch" test t5510 was flaky while running a (forced) automagic
399 garbage collection.
401 * Documentation updates to help contributors setting up Travis CI
402 test for their patches.
404 * Some multi-byte encoding can have a backslash byte as a later part
405 of one letter, which would confuse "highlight" filter used in
406 gitweb.
408 * "git commit-tree" plumbing command required the user to always sign
409 its result when the user sets the commit.gpgsign configuration
410 variable, which was an ancient mistake. Rework "git rebase" that
411 relied on this mistake so that it reads commit.gpgsign and pass (or
412 not pass) the -S option to "git commit-tree" to keep the end-user
413 expectation the same, while teaching "git commit-tree" to ignore
414 the configuration variable. This will stop requiring the users to
415 sign commit objects used internally as an implementation detail of
416 "git stash".
417 (merge 6694856 jc/commit-tree-ignore-commit-gpgsign later to maint).
419 * "http.cookieFile" configuration variable clearly wants a pathname,
420 but we forgot to treat it as such by e.g. applying tilde expansion.
421 (merge e5a39ad bn/http-cookiefile-config later to maint).
423 * Consolidate description of tilde-expansion that is done to
424 configuration variables that take pathname to a single place.
425 (merge dca83ab jc/config-pathname-type later to maint).
427 * Correct faulty recommendation to use "git submodule deinit ." when
428 de-initialising all submodules, which would result in a strange
429 error message in a pathological corner case.
430 (merge f6a5279 sb/submodule-deinit-all later to maint).
432 * Many 'linkgit:<git documentation page>' references were broken,
433 which are all fixed with this.
434 (merge 1cca17d jc/linkgit-fix later to maint).
436 * "git rerere" can get confused by conflict markers deliberately left
437 by the inner merge step, because they are indistinguishable from
438 the real conflict markers left by the outermost merge which are
439 what the end user and "rerere" need to look at. This was fixed by
440 making the conflict markers left by the inner merges a bit longer.
441 (merge 0f9fd5c jc/ll-merge-internal later to maint).
443 * CI test was taught to build documentation pages.
444 (merge b98712b ls/travis-build-doc later to maint).
446 * "git fsck" learned to catch NUL byte in a commit object as
447 potential error and warn.
448 (merge 6d2d780 jc/fsck-nul-in-commit later to maint).
450 * Portability enhancement for "rebase -i" to help platforms whose
451 shell does not like "for i in <empty>" (which is not POSIX-kosher).
452 (merge 8e98b35 jk/rebase-interative-eval-fix later to maint).
454 * On Windows, .git and optionally any files whose name starts with a
455 dot are now marked as hidden, with a core.hideDotFiles knob to
456 customize this behaviour.
457 (merge ebf31e7 js/windows-dotgit later to maint).
459 * Documentation for "git merge --verify-signatures" has been updated
460 to clarify that the signature of only the commit at the tip is
461 verified. Also the phrasing used for signature and key validity is
462 adjusted to align with that used by OpenPGP.
463 (merge 05a5869 kf/gpg-sig-verification-doc later to maint).
465 * A couple of bugs around core.autocrlf have been fixed.
466 (merge caa47ad tb/core-eol-fix later to maint).
468 * Many commands normalize command line arguments from NFD to NFC
469 variant of UTF-8 on OSX, but commands in the "diff" family did
470 not, causing "git diff $path" to complain that no such path is
471 known to Git. They have been taught to do the normalization.
472 (merge 90a78b8 ar/diff-args-osx-precompose later to maint).
474 * "git difftool" learned to handle unmerged paths correctly in
475 dir-diff mode.
476 (merge 366f9ce da/difftool later to maint).
478 * Other minor clean-ups and documentation updates
479 (merge 832c0e5 lp/typofixes later to maint).
480 (merge f5ee54a sb/z-is-gnutar-ism later to maint).
481 (merge 2e3926b va/i18n-misc-updates later to maint).
482 (merge f212dcc bn/config-doc-tt-varnames later to maint).
483 (merge f54bea4 nd/remote-plural-ours-plus-theirs later to maint).
484 (merge 2bb0518 ak/t4151-ls-files-could-be-empty later to maint).
485 (merge 4df4313 jc/test-seq later to maint).
486 (merge a75a308 tb/t5601-sed-fix later to maint).
487 (merge 6c1fbe1 va/i18n-remote-comment-to-align later to maint).
488 (merge dee2303 va/mailinfo-doc-typofix later to maint).