| blob | 257f1e7d290b7c614811ae0cc8dba822266c26f2 |
1 Git 2.14 Release Notes
2 ======================
4 Backward compatibility notes.
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 not scheduled to happen in the upcoming
12 release (yet).
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
19 greatly appreciated.
21 * The experiment to improve the hunk-boundary selection of textual
22 diff output has finished, and the "indent heuristics" has now
23 become the default.
26 Updates since v2.13
27 -------------------
29 UI, Workflows & Features
31 * The colors in which "git status --short --branch" showed the names
32 of the current branch and its remote-tracking branch are now
33 configurable.
35 * "git clone" learned the "--no-tags" option not to fetch all tags
36 initially, and also set up the tagopt not to follow any tags in
37 subsequent fetches.
39 * "git archive --format=zip" learned to use zip64 extension when
40 necessary to go beyond the 4GB limit.
41 (merge 867e40ff3a rs/large-zip later to maint).
43 * "git reset" learned "--recurse-submodules" option.
45 * "git diff --submodule=diff" now recurses into nested submodules.
46 (merge 5a5221427c jk/diff-submodule-diff-inline later to maint).
48 * "git repack" learned to accept the --threads=<n> option and pass it
49 to pack-objects.
51 * "git send-email" learned to run sendemail-validate hook to inspect
52 and reject a message before sending it out.
53 (merge 177409e589 jt/send-email-validate-hook later to maint).
55 * There is no good reason why "git fetch $there $sha1" should fail
56 when the $sha1 names an object at the tip of an advertised ref,
57 even when the other side hasn't enabled allowTipSHA1InWant.
59 * The recently introduced "[includeIf "gitdir:$dir"] path=..."
60 mechansim has further been taught to take symlinks into account.
61 The directory "$dir" specified in "gitdir:$dir" may be a symlink to
62 a real location, not something that $(getcwd) may return. In such
63 a case, a realpath of "$dir" is compared with the real path of the
64 current repository to determine if the contents from the named path
65 should be included.
67 * Make the "indent" heuristics the default in "diff" and diff.indentHeuristics
68 configuration variable an escape hatch for those who do no want it.
70 * Many commands learned to pay attention to submodule.recurse
71 configuration.
74 Performance, Internal Implementation, Development Support etc.
76 * The default packed-git limit value has been raised on larger
77 platforms to save "git fetch" from a (recoverable) failure while
78 "gc" is running in parallel.
80 * Code to update the cache-tree has been tightened so that we won't
81 accidentally write out any 0{40} entry in the tree object.
82 (merge a96d3cc3f6 jk/no-null-sha1-in-cache-tree later to maint).
84 * Attempt to allow us notice "fishy" situation where we fail to
85 remove the temporary directory used during the test.
87 * Travis CI gained a task to format the documentation with both
88 AsciiDoc and AsciiDoctor.
90 * Some platforms have ulong that is smaller than time_t, and our
91 historical use of ulong for timestamp would mean they cannot
92 represent some timestamp that the platform allows. Invent a
93 separate and dedicated timestamp_t (so that we can distingiuish
94 timestamps and a vanilla ulongs, which along is already a good
95 move), and then declare uintmax_t is the type to be used as the
96 timestamp_t.
98 * We can trigger Windows auto-build tester (credits: Dscho &
99 Microsoft) from our existing Travis CI tester now.
101 * Conversion from uchar[20] to struct object_id continues.
103 * Simplify parse_pathspec() codepath and stop it from looking at the
104 default in-core index.
105 (merge 08de9151a8 bw/pathspec-sans-the-index later to maint).
107 * Add perf-test for wildmatch.
108 (merge 62ca75a6b9 ab/perf-wildmatch later to maint).
110 * Code from "conversion using external process" codepath has been
111 extracted to a separate sub-process.[ch] module.
112 (merge 4f2a2e9f0e bp/sub-process-convert-filter later to maint).
114 * When "git checkout", "git merge", etc. manipulates the in-core
115 index, various pieces of information in the index extensions are
116 discarded from the original state, as it is usually not the case
117 that they are kept up-to-date and in-sync with the operation on the
118 main index. The untracked cache extension is copied across these
119 operations now, which would speed up "git status" (as long as the
120 cache is properly invalidated).
122 * The internal implementation of "git grep" has seen some clean-up.
123 (merge 8df4c2953f ab/grep-preparatory-cleanup later to maint).
125 * Update the C style recommendation for notes for translators, as
126 recent versions of gettext tools can work with our style of
127 multi-line comments.
129 * The implementation of "ref" API around the "packed refs" have been
130 cleaned up, in preparation for further changes.
132 * The internal logic used in "git blame" has been libified to make it
133 easier to use by cgit.
135 * Our code often opens a path to an optional file, to work on its
136 contents when we can successfully open it. We can ignore a failure
137 to open if such an optional file does not exist, but we do want to
138 report a failure in opening for other reasons (e.g. we got an I/O
139 error, or the file is there, but we lack the permission to open).
141 The exact errors we need to ignore are ENOENT (obviously) and
142 ENOTDIR (less obvious). Instead of repeating comparison of errno
143 with these two constants, introduce a helper function to do so.
145 * We often try to open a file for reading whose existence is
146 optional, and silently ignore errors from open/fopen; report such
147 errors if they are not due to missing files.
150 Also contains various documentation updates and code clean-ups.
153 Fixes since v2.13
154 -----------------
156 Unless otherwise noted, all the fixes since v2.13 in the maintenance
157 track are contained in this release (see the maintenance releases'
158 notes for details).
160 * "git gc" did not interact well with "git worktree"-managed
161 per-worktree refs.
163 * "git cherry-pick" and other uses of the sequencer machinery
164 mishandled a trailer block whose last line is an incomplete line.
165 This has been fixed so that an additional sign-off etc. are added
166 after completing the existing incomplete line.
168 * The codepath in "git am" that is used when running "git rebase"
169 leaked memory held for the log message of the commits being rebased.
171 * "git clone --config var=val" is a way to populate the
172 per-repository configuration file of the new repository, but it did
173 not work well when val is an empty string. This has been fixed.
175 * Setting "log.decorate=false" in the configuration file did not take
176 effect in v2.13, which has been corrected.
178 * A few codepaths in "checkout" and "am" working on an unborn branch
179 tried to access an uninitialized piece of memory.
181 * The Web interface to gmane news archive is long gone, even though
182 the articles are still accessible via NTTP. Replace the links with
183 ones to public-inbox.org. Because their message identification is
184 based on the actual message-id, it is likely that it will be easier
185 to migrate away from it if/when necessary.
187 * The receive-pack program now makes sure that the push certificate
188 records the same set of push options used for pushing.
190 * Tests have been updated to pass under GETTEXT_POISON (a mechanism
191 to ensure that output strings that should not be translated are
192 not translated by mistake), and TravisCI is told to run them.
194 * "git checkout --recurse-submodules" did not quite work with a
195 submodule that itself has submodules.
197 * Plug some leaks and updates internal API used to implement the
198 split index feature to make it easier to avoid such a leak in the
199 future.
200 (merge de950c5773 nd/split-index-unshare later to maint).
202 * "pack-objects" can stream a slice of an existing packfile out when
203 the pack bitmap can tell that the reachable objects are all needed
204 in the output, without inspecting individual objects. This
205 strategy however would not work well when "--local" and other
206 options are in use, and need to be disabled.
208 * Fix memory leaks pointed out by Coverity (and people).
210 * "git read-tree -m" (no tree-ish) gave a nonsense suggestion "use
211 --empty if you want to clear the index". With "-m", such a request
212 will still fail anyway, as you'd need to name at least one tree-ish
213 to be merged.
215 * Make sure our tests would pass when the sources are checked out
216 with "platform native" line ending convention by default on
217 Windows. Some "text" files out tests use and the test scripts
218 themselves that are meant to be run with /bin/sh, ought to be
219 checked out with eol=LF even on Windows.
221 * Introduce the BUG() macro to improve die("BUG: ...").
223 * Clarify documentation for include.path and includeIf.<condition>.path
224 configuration variables.
226 * Git sometimes gives an advice in a rhetorical question that does
227 not require an answer, which can confuse new users and non native
228 speakers. Attempt to rephrase them.
230 * A few http:// links that are redirected to https:// in the
231 documentation have been updated to https:// links.
233 * "git for-each-ref --format=..." with %(HEAD) in the format used to
234 resolve the HEAD symref as many times as it had processed refs,
235 which was wasteful, and "git branch" shared the same problem.
237 * Regression fix to topic recently merged to 'master'.
239 * The shell completion script (in contrib/) learned "git stash" has
240 a new "push" subcommand.
242 * "git interpret-trailers", when used as GIT_EDITOR for "git commit
243 -v", looked for and appended to a trailer block at the very end,
244 i.e. at the end of the "diff" output. The command has been
245 corrected to pay attention to the cut-mark line "commit -v" adds to
246 the buffer---the real trailer block should appear just before it.
248 * A test allowed both "git push" and "git receive-pack" on the other
249 end write their traces into the same file. This is OK on platforms
250 that allows atomically appending to a file opened with O_APPEND,
251 but on other platforms led to a mangled output, causing
252 intermittent test failures. This has been fixed by disabling
253 traces from "receive-pack" in the test.
255 * Tag objects, which are not reachable from any ref, that point at
256 missing objects were mishandled by "git gc" and friends (they
257 should silently be ignored instead)
259 * "git describe --contains" penalized light-weight tags so much that
260 they were almost never considered. Instead, give them about the
261 same chance to be considered as an annotated tag that is the same
262 age as the underlying commit would.
263 (merge ef1e74065c jc/name-rev-lw-tag later to maint).
265 * The "run-command" API implementation has been made more robust
266 against dead-locking in a threaded environment.
268 * A recent update to t5545-push-options.sh started skipping all the
269 tests in the script when a web server testing is disabled or
270 unavailable, not just the ones that require a web server. Non HTTP
271 tests have been salvaged to always run in this script.
273 * "git send-email" now uses Net::SMTP::SSL, which is obsolete, only
274 when needed. Recent versions of Net::SMTP can do TLS natively.
276 * "foo\bar\baz" in "git fetch foo\bar\baz", even though there is no
277 slashes in it, cannot be a nickname for a remote on Windows, as
278 that is likely to be a pathname on a local filesystem.
280 * "git clean -d" used to clean directories that has ignored files,
281 even though the command should not lose ignored ones without "-x".
282 "git status --ignored" did not list ignored and untracked files
283 without "-uall". These have been corrected.
285 * The result from "git diff" that compares two blobs, e.g. "git diff
286 $commit1:$path $commit2:$path", used to be shown with the full
287 object name as given on the command line, but it is more natural to
288 use the $path in the output and use it to look up .gitattributes.
289 (merge 30d005c020 jk/diff-blob later to maint).
291 * The "collision detecting" SHA-1 implementation shipped with 2.13
292 was quite broken on some big-endian platforms and/or platforms that
293 do not like unaligned fetches. Update to the upstream code which
294 has already fixed these issues.
296 * "git am -h" triggered a BUG().
298 * The interaction of "url.*.insteadOf" and custom URL scheme's
299 whitelisting is now documented better.
301 * The timestamp of the index file is now taken after the file is
302 closed, to help Windows, on which a stale timestamp is reported by
303 fstat() on a file that is opened for writing and data was written
304 but not yet closed.
306 * "git pull --rebase --autostash" didn't auto-stash when the local history
307 fast-forwards to the upstream.
309 * A flaky test has been corrected.
310 (merge 7c2115aa07 jk/pack-idx-corruption-safety later to maint).
312 * Other minor doc, test and build updates and code cleanups.
313 (merge 8ba74bfd7c jc/diff-tree-stale-comment later to maint).