Update draft release notes to 1.8.4
[git/git.git] / Documentation / RelNotes / 1.8.4.txt
1 Git v1.8.4 Release Notes
2 ========================
3
4 Backward compatibility notes (for Git 2.0)
5 ------------------------------------------
6
7 When "git push [$there]" does not say what to push, we have used the
8 traditional "matching" semantics so far (all your branches were sent
9 to the remote as long as there already are branches of the same name
10 over there). In Git 2.0, the default will change to the "simple"
11 semantics that pushes:
12
13 - only the current branch to the branch with the same name, and only
14 when the current branch is set to integrate with that remote
15 branch, if you are pushing to the same remote as you fetch from; or
16
17 - only the current branch to the branch with the same name, if you
18 are pushing to a remote that is not where you usually fetch from.
19
20 Use the user preference configuration variable "push.default" to
21 change this. If you are an old-timer who is used to the "matching"
22 semantics, you can set the variable to "matching" to keep the
23 traditional behaviour. If you want to live in the future early, you
24 can set it to "simple" today without waiting for Git 2.0.
25
26 When "git add -u" (and "git add -A") is run inside a subdirectory and
27 does not specify which paths to add on the command line, it
28 will operate on the entire tree in Git 2.0 for consistency
29 with "git commit -a" and other commands. There will be no
30 mechanism to make plain "git add -u" behave like "git add -u .".
31 Current users of "git add -u" (without a pathspec) should start
32 training their fingers to explicitly say "git add -u ."
33 before Git 2.0 comes. A warning is issued when these commands are
34 run without a pathspec and when you have local changes outside the
35 current directory, because the behaviour in Git 2.0 will be different
36 from today's version in such a situation.
37
38 In Git 2.0, "git add <path>" will behave as "git add -A <path>", so
39 that "git add dir/" will notice paths you removed from the directory
40 and record the removal. Versions before Git 2.0, including this
41 release, will keep ignoring removals, but the users who rely on this
42 behaviour are encouraged to start using "git add --ignore-removal <path>"
43 now before 2.0 is released.
44
45
46 Updates since v1.8.3
47 --------------------
48
49 Foreign interfaces, subsystems and ports.
50
51 * "git rebase -i" now honors --strategy and -X options.
52
53 * Git-gui has been updated to its 0.18.0 version.
54
55 * MediaWiki remote helper (in contrib/) has been updated to use the
56 credential helper interface from Git.pm.
57
58 * Update build for Cygwin 1.[57]. Torsten Bögershausen reports that
59 this is fine with Cygwin 1.7 ($gmane/225824) so let's try moving it
60 ahead.
61
62 * The credential helper to talk to keychain on OS X (in contrib/) has
63 been updated to kick in not just when talking http/https but also
64 imap(s) and smtp.
65
66 * Remote transport helper has been updated to report errors and
67 maintain ref hierarchy used to keep track of its own state better.
68
69 * With "export" remote-helper protocol, (1) a push that tries to
70 update a remote ref whose name is different from the pushing side
71 does not work yet, and (2) the helper may not know how to do
72 --dry-run; these problematic cases are disabled for now.
73
74 * git-remote-hg/bzr (in contrib/) updates.
75
76 * git-remote-mw (in contrib/) hints users to check the certificate,
77 when https:// connection failed.
78
79
80 UI, Workflows & Features
81
82 * The "push.default=simple" mode of "git push" has been updated to
83 behave like "current" without requiring a remote tracking
84 information, when you push to a remote that is different from where
85 you fetch from (i.e. a triangular workflow).
86
87 * Having multiple "fixup!" on a line in the rebase instruction sheet
88 did not work very well with "git rebase -i --autosquash".
89
90 * "git log" learned the "--author-date-order" option, with which the
91 output is topologically sorted and commits in parallel histories
92 are shown intermixed together based on the author timestamp.
93
94 * Various subcommands of "git submodule" refused to run from anywhere
95 other than the top of the working tree of the superproject, but
96 they have been taught to let you run from a subdirectory.
97
98 * "git diff" learned a mode that ignores hunks whose change consists
99 only of additions and removals of blank lines, which is the same as
100 "diff -B" (ignore blank lines) of GNU diff.
101
102 * "git rm" gives a single message followed by list of paths to report
103 multiple paths that cannot be removed.
104
105 * "git rebase" can be told with ":/look for this string" syntax commits
106 to replay the changes onto and where the work to be replayed begins.
107
108 * Many tutorials teach users to set "color.ui" to "auto" as the first
109 thing after you set "user.name/email" to introduce yourselves to
110 Git. Now the variable defaults to "auto".
111
112 * On Cygwin, "cygstart" is now recognised as a possible way to start
113 a web browser (used in "help -w" and "instaweb" among others).
114
115 * "git status" learned status.branch and status.short configuration
116 variables to use --branch and --short options by default (override
117 with --no-branch and --no-short options from the command line).
118
119 * "git cmd <name>", when <name> happens to be a 40-hex string,
120 directly uses the 40-hex string as an object name, even if a ref
121 "refs/<some hierarchy>/<name>" exists. This disambiguation order
122 is unlikely to change, but we should warn about the ambiguity just
123 like we warn when more than one refs/ hierachies share the same
124 name.
125
126 * "git rebase" learned "--[no-]autostash" option to save local
127 changes instead of refusing to run (to which people's normal
128 response was to stash them and re-run). This introduced a corner
129 case breakage to "git am --abort" but it has been fixed.
130
131 * Instead of typing four capital letters "HEAD", you can say "@" now,
132 e.g. "git log @".
133
134 * "check-ignore" (new feature since 1.8.2) has been updated to work
135 more like "check-attr" over bidi-pipes.
136
137 * "git describe" learned "--first-parent" option to limit its closest
138 tagged commit search to the first-parent chain.
139
140 * "git merge foo" that might have meant "git merge origin/foo" is
141 diagnosed with a more informative error message.
142
143 * "git log -L<line>,<range>:<filename>" has been added. This may
144 still have leaks and rough edges, though.
145
146 * We used the approxidate() parser for "--expire=<timestamp>" options
147 of various commands, but it is better to treat --expire=all and
148 --expire=now a bit more specially than using the current timestamp.
149 "git gc" and "git reflog" have been updated with a new parsing
150 function for expiry dates.
151
152 * Updates to completion (both bash and zsh) helpers.
153
154 * The behaviour of the "--chain-reply-to" option of "git send-email"
155 have changed at 1.7.0, and we added a warning/advice message to
156 help users adjust to the new behaviour back then, but we kept it
157 around for too long. The message has finally been removed.
158
159 * "git fetch origin master" unlike "git fetch origin" or "git fetch"
160 did not update "refs/remotes/origin/master"; this was an early
161 design decision to keep the update of remote tracking branches
162 predictable, but in practice it turns out that people find it more
163 convenient to opportunistically update them whenever we have a
164 chance, and we have been updating them when we run "git push" which
165 already breaks the original "predictability" anyway.
166
167 * The configuration variable core.checkstat was advertised in the
168 documentation but the code expected core.statinfo instead.
169 For now, we accept both core.checkstat and core.statinfo, but the
170 latter will be removed in the longer term.
171
172
173 Performance, Internal Implementation, etc.
174
175 * The original way to specify remote repository using .git/branches/
176 used to have a nifty feature. The code to support the feature was
177 still in a function but the caller was changed not to call it 5
178 years ago, breaking that feature and leaving the supporting code
179 unreachable. The dead code has been removed.
180
181 * "git pack-refs" that races with new ref creation or deletion have
182 been susceptible to lossage of refs under right conditions, which
183 has been tightened up.
184
185 * We read loose and packed rerferences in two steps, but after
186 deciding to read a loose ref but before actually opening it to read
187 it, another process racing with us can unlink it, which would cause
188 us to barf. The codepath has been updated to retry when such a
189 race is detected, instead of outright failing.
190
191 * Uses of the platform fnmatch(3) function (many places in the code,
192 matching pathspec, .gitignore and .gitattributes to name a few)
193 have been replaced with wildmatch, allowing "foo/**/bar" that would
194 match foo/bar, foo/a/bar, foo/a/b/bar, etc.
195
196 * Memory ownership and lifetime rules for what for-each-ref feeds to
197 its callbacks have been clarified (in short, "you do not own it, so
198 make a copy if you want to keep it").
199
200 * The revision traversal logic to improve culling of irrelevant
201 parents while traversing a mergy history has been updated.
202
203 * Some leaks in unpack-trees (used in merge, cherry-pick and other
204 codepaths) have been plugged.
205
206 * The codepath to read from marks files in fast-import/export did not
207 have to accept anything but 40-hex representation of the object
208 name. Further, fast-export did not need full in-core object
209 representation to have parsed wen reading from them. These
210 codepaths have been optimized by taking advantage of these access
211 patterns.
212
213 * Object lookup logic, when the object hashtable starts to become
214 crowded, has been optimized.
215
216 * When TEST_OUTPUT_DIRECTORY setting is used, it was handled somewhat
217 inconsistently between the test framework and t/Makefile, and logic
218 to summarize the results looked at a wrong place.
219
220 * "git clone" uses a lighter-weight implementation when making sure
221 that the history behind refs are complete.
222
223 * Many warnings from sparse source checker in compat/ area has been
224 squelched.
225
226 * The code to reading and updating packed-refs file has been updated,
227 correcting corner case bugs.
228
229
230 Also contains various documentation updates and code clean-ups.
231
232
233 Fixes since v1.8.3
234 ------------------
235
236 Unless otherwise noted, all the fixes since v1.8.3 in the maintenance
237 track are contained in this release (see release notes to them for
238 details).
239
240 * "git diff" refused to even show difference when core.safecrlf is
241 set to true (i.e. error out) and there are offending lines in the
242 working tree files.
243 (merge 5430bb2 jc/maint-diff-core-safecrlf later to maint).
244
245 * A test that should have failed but didn't revealed a bug that needs
246 to be corrected.
247 (merge 94d75d1 jc/t1512-fix later to maint).
248
249 * An overlong path to a .git directory may have overflown the
250 temporary path buffer used to create a name for lockfiles.
251 (merge 2fbd4f9 mh/maint-lockfile-overflow later to maint).
252
253 * Invocations of "git checkout" used internally by "git rebase" were
254 counted as "checkout", and affected later "git checkout -" to the
255 the user to an unexpected place.
256 (merge 3bed291 rr/rebase-checkout-reflog later to maint).
257
258 * "git stash save", when your local change turns a tracked file into
259 a directory, has to remove files in that directory in order to
260 revert your working tree to a pristine state. This will lose
261 untracked files in such a directory, and the command now requires
262 you to "--force" it.
263
264 * The configuration variable column.ui was poorly documented.
265 (merge 5e62cc1 rr/column-doc later to maint).
266
267 * "git name-rev --refs=tags/v*" were forbidden, which was a bit
268 inconvenient (you had to give a pattern to match refs fully, like
269 --refs=refs/tags/v*).
270 (merge 98c5c4a nk/name-rev-abbreviated-refs later to maint).
271
272 * "git apply" parsed patches that add new files, generated by
273 programs other than Git, incorrectly. This is an old breakage in
274 v1.7.11 and will need to be merged down to the maintanance tracks.
275 (merge 212eb96 tr/maint-apply-non-git-patch-parsefix later to maint).
276
277 * Older cURL wanted piece of memory we call it with to be stable, but
278 we updated the auth material after handing it to a call.
279 (merge a94cf2c bc/http-keep-memory-given-to-curl later to maint).
280
281 * "git pull" into nothing trashed "local changes" that were in the
282 index, and this avoids it.
283 (merge b4dc085 jk/pull-into-dirty-unborn later to maint).
284
285 * Many "git submodule" operations do not work on a submodule at a
286 path whose name is not in ASCII.
287 (merge bed9470 fg/submodule-non-ascii-path later to maint).
288
289 * "cherry-pick" had a small leak in an error codepath.
290 (merge 706728a fc/sequencer-plug-leak later to maint).
291
292 * Logic used by git-send-email to suppress cc mishandled names like
293 "A U. Thor" <author@example.xz>, where the human readable part
294 needs to be quoted (the user input may not have the double quotes
295 around the name, and comparison was done between quoted and
296 unquoted strings). It also mishandled names that need RFC2047
297 quoting.
298 (merge 1495266 mt/send-email-cc-match-fix later to maint).
299
300 * Call to discard_cache/discard_index (used when we use different
301 contents of the index in-core, in many operations like commit,
302 apply, and merge) used to leak memory that held the array of index
303 entries, which has been plugged.
304 (merge a0fc4db rs/discard-index-discard-array later to maint).
305
306 * "gitweb" forgot to clear a global variable $search_regexp upon each
307 request, mistakenly carrying over the previous search to a new one
308 when used as a persistent CGI.
309 (merge ca7a5dc cm/gitweb-project-list-persistent-cgi-fix later to maint).
310
311 * The wildmatch engine did not honor WM_CASEFOLD option correctly.
312 (merge b79c0c3 ar/wildmatch-foldcase later to maint).
313
314 * "git log -c --follow $path" segfaulted upon hitting the commit that
315 renamed the $path being followed.
316 (merge 46ec510 cb/log-follow-with-combined later to maint).
317
318 * When a reflog notation is used for implicit "current branch", we
319 did not say which branch and worse said "branch ''".
320 (merge 305ebea rr/die-on-missing-upstream later to maint).
321
322 * "difftool --dir-diff" did not copy back changes made by the
323 end-user in the diff tool backend to the working tree in some
324 cases.
325 (merge 32eaf1d ks/difftool-dir-diff-copy-fix later to maint).
326
327 * "git push $there HEAD:branch" did not resolve HEAD early enough, so
328 it was easy to flip it around while push is still going on and push
329 out a branch that the user did not originally intended when the
330 command was started.
331 (merge 0f075b2 rr/push-head later to maint).
332
333 * The bash prompt code (in contrib/) displayed the name of the branch
334 being rebased when "rebase -i/-m/-p" modes are in use, but not the
335 plain vanilla "rebase".
336 (merge 1306321 fc/show-branch-in-rebase-am later to maint).
337
338 * Handling of negative exclude pattern for directories "!dir" was
339 broken in the update to v1.8.3.
340 (merge c3c327d kb/status-ignored-optim-2 later to maint).
341
342 * zsh prompt script that borrowed from bash prompt script did not
343 work due to slight differences in array variable notation between
344 these two shells.
345 (merge d0583da tg/maint-zsh-svn-remote-prompt later to maint).
346
347 * An entry for "file://" scheme in the enumeration of URL types Git
348 can take in the HTML documentation was made into a clickable link
349 by mistake.
350 (merge 4c32e36 nd/urls-doc-no-file-hyperlink-fix later to maint).
351
352 * "git push --[no-]verify" was not documented.
353 (merge 90d32d1 tr/push-no-verify-doc later to maint).
354
355 * Stop installing the git-remote-testpy script that is only used for
356 testing.
357 (merge 416fda6 fc/makefile later to maint).
358
359 * "git commit --allow-empty-message -m ''" should not start an
360 editor.
361 (merge 2520677 rs/commit-m-no-edit later to maint).
362
363 * "git merge @{-1}~22" was rewritten to "git merge frotz@{1}~22"
364 incorrectly when your previous branch was "frotz" (it should be
365 rewritten to "git merge frotz~22" instead).
366 (merge 84cf246 jc/strbuf-branchname-fix later to maint).
367
368 * "git diff -c -p" was not showing a deleted line from a hunk when
369 another hunk immediately begins where the earlier one ends.
370 (merge aac3857 mk/combine-diff-context-horizon-fix later to maint).
371
372 * "git log --ancestry-path A...B" did not work as expected, as it did
373 not pay attention to the fact that the merge base between A and B
374 was the bottom of the range being specified.
375 (merge a765499 kb/ancestry-path-threedots later to maint).
376
377 * Mac OS X does not like to write(2) more than INT_MAX number of
378 bytes; work it around by chopping write(2) into smaller pieces.
379 (merge 6c642a8 fc/macos-x-clipped-write later to maint).
380
381 * Newer MacOS X encourages the programs to compile and link with
382 their CommonCrypto, not with OpenSSL.
383 (merge be4c828 da/darwin later to maint).
384
385 * "git clone foo/bar:baz" cannot be a request to clone from a remote
386 over git-over-ssh specified in the scp style. This case is now
387 detected and clones from a local repository at "foo/bar:baz".
388 (merge 6000334 nd/clone-local-with-colon later to maint).
389
390 * When $HOME is misconfigured to point at an unreadable directory, we
391 used to complain and die. Loosen the check.
392 (merge 4698c8f jn/config-ignore-inaccessible later to maint).
393
394 * "git subtree" (in contrib/) had one codepath with loose error
395 checks to lose data at the remote side.
396 (merge 3212d56 jk/subtree-do-not-push-if-split-fails later to maint).
397
398 * "git fetch" into a shallow repository from a repository that does
399 not know about the shallow boundary commits (e.g. a different fork
400 from the repository the current shallow repository was cloned from)
401 did not work correctly.
402 (merge 71d5f93 mh/fetch-into-shallow later to maint).
403
404 * "git checkout foo" DWIMs the intended "upstream" and turns it into
405 "git checkout -t -b foo remotes/origin/foo". This codepath has been
406 updated to correctly take existing remote definitions into account.
407 (merge 229177a jh/checkout-auto-tracking later to maint).