| blob | 2226abeffbc034849879ad2d07d52afd0df28057 |
1 Git v1.8.4 Release Notes
2 ========================
4 Backward compatibility notes (for Git 2.0)
5 ------------------------------------------
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:
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
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.
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.
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.
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.
46 Updates since v1.8.3
47 --------------------
49 Foreign interfaces, subsystems and ports.
51 * "git rebase -i" now honors --strategy and -X options.
53 * Git-gui has been updated to its 0.18.0 version.
55 * MediaWiki remote helper (in contrib/) has been updated to use the
56 credential helper interface from Git.pm.
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.
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.
66 * Remote transport helper has been updated to report errors and
67 maintain ref hierarchy used to keep track of its own state better.
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.
74 * git-remote-hg/bzr (in contrib/) updates.
76 * git-remote-mw (in contrib/) hints users to check the certificate,
77 when https:// connection failed.
79 * git-remote-mw (in contrib/) adds a command to allow previewing the
80 contents locally before pushing it out, when working with a
81 MediaWiki remote.
84 UI, Workflows & Features
86 * "git cat-file --batch-check=<format>" is added, primarily to allow
87 on-disk footprint of objects in packfiles (often they are a lot
88 smaller than their true size, when expressed as deltas) to be
89 reported.
91 * "git rebase [-i]" used to leave just "rebase" as its reflog messages
92 for some operations. They have been reworded to be more informative.
94 * In addition to the choice from "rebase, merge, or checkout-detach",
95 "submodule update" can allow a custom command to be used in to
96 update the working tree of submodules via the "submodule.*.update"
97 configuration variable.
99 * "git submodule update" can optionally clone the submodule
100 repositories shallowly.
102 * "git format-patch" learned "--from[=whom]" option, which sets the
103 "From: " header to the specified person (or the person who runs the
104 command, if "=whom" part is missing) and move the original author
105 information to an in-body From: header as necessary.
107 * The configuration variable "merge.ff" was cleary a tri-state to
108 choose one from "favor fast-forward when possible", "always create
109 a merge even when the history could fast-forward" and "do not
110 create any merge, only update when the history fast-forwards", but
111 the command line parser did not implement the usual convention of
112 "last one wins, and command line overrides the configuration"
113 correctly.
115 * "gitweb" learned to optionally place extra links that point at the
116 levels higher than the Gitweb pages themselves in the breadcrumbs,
117 so that it can be used as part of a larger installation.
119 * "git log --format=" now honors i18n.logoutputencoding configuration
120 variable.
122 * The "push.default=simple" mode of "git push" has been updated to
123 behave like "current" without requiring a remote tracking
124 information, when you push to a remote that is different from where
125 you fetch from (i.e. a triangular workflow).
127 * Having multiple "fixup!" on a line in the rebase instruction sheet
128 did not work very well with "git rebase -i --autosquash".
130 * "git log" learned the "--author-date-order" option, with which the
131 output is topologically sorted and commits in parallel histories
132 are shown intermixed together based on the author timestamp.
134 * Various subcommands of "git submodule" refused to run from anywhere
135 other than the top of the working tree of the superproject, but
136 they have been taught to let you run from a subdirectory.
138 * "git diff" learned a mode that ignores hunks whose change consists
139 only of additions and removals of blank lines, which is the same as
140 "diff -B" (ignore blank lines) of GNU diff.
142 * "git rm" gives a single message followed by list of paths to report
143 multiple paths that cannot be removed.
145 * "git rebase" can be told with ":/look for this string" syntax commits
146 to replay the changes onto and where the work to be replayed begins.
148 * Many tutorials teach users to set "color.ui" to "auto" as the first
149 thing after you set "user.name/email" to introduce yourselves to
150 Git. Now the variable defaults to "auto".
152 * On Cygwin, "cygstart" is now recognised as a possible way to start
153 a web browser (used in "help -w" and "instaweb" among others).
155 * "git status" learned status.branch and status.short configuration
156 variables to use --branch and --short options by default (override
157 with --no-branch and --no-short options from the command line).
159 * "git cmd <name>", when <name> happens to be a 40-hex string,
160 directly uses the 40-hex string as an object name, even if a ref
161 "refs/<some hierarchy>/<name>" exists. This disambiguation order
162 is unlikely to change, but we should warn about the ambiguity just
163 like we warn when more than one refs/ hierachies share the same
164 name.
166 * "git rebase" learned "--[no-]autostash" option to save local
167 changes instead of refusing to run (to which people's normal
168 response was to stash them and re-run). This introduced a corner
169 case breakage to "git am --abort" but it has been fixed.
171 * Instead of typing four capital letters "HEAD", you can say "@" now,
172 e.g. "git log @".
174 * "check-ignore" (new feature since 1.8.2) has been updated to work
175 more like "check-attr" over bidi-pipes.
177 * "git describe" learned "--first-parent" option to limit its closest
178 tagged commit search to the first-parent chain.
180 * "git merge foo" that might have meant "git merge origin/foo" is
181 diagnosed with a more informative error message.
183 * "git log -L<line>,<range>:<filename>" has been added. This may
184 still have leaks and rough edges, though.
186 * We used the approxidate() parser for "--expire=<timestamp>" options
187 of various commands, but it is better to treat --expire=all and
188 --expire=now a bit more specially than using the current timestamp.
189 "git gc" and "git reflog" have been updated with a new parsing
190 function for expiry dates.
192 * Updates to completion (both bash and zsh) helpers.
194 * The behaviour of the "--chain-reply-to" option of "git send-email"
195 have changed at 1.7.0, and we added a warning/advice message to
196 help users adjust to the new behaviour back then, but we kept it
197 around for too long. The message has finally been removed.
199 * "git fetch origin master" unlike "git fetch origin" or "git fetch"
200 did not update "refs/remotes/origin/master"; this was an early
201 design decision to keep the update of remote tracking branches
202 predictable, but in practice it turns out that people find it more
203 convenient to opportunistically update them whenever we have a
204 chance, and we have been updating them when we run "git push" which
205 already breaks the original "predictability" anyway.
207 * The configuration variable core.checkstat was advertised in the
208 documentation but the code expected core.statinfo instead.
209 For now, we accept both core.checkstat and core.statinfo, but the
210 latter will be removed in the longer term.
213 Performance, Internal Implementation, etc.
215 * Fetching between repositories with many refs employed O(n^2)
216 algorithm to match up the common objects, which has been corrected.
218 * The original way to specify remote repository using .git/branches/
219 used to have a nifty feature. The code to support the feature was
220 still in a function but the caller was changed not to call it 5
221 years ago, breaking that feature and leaving the supporting code
222 unreachable. The dead code has been removed.
224 * "git pack-refs" that races with new ref creation or deletion have
225 been susceptible to lossage of refs under right conditions, which
226 has been tightened up.
228 * We read loose and packed rerferences in two steps, but after
229 deciding to read a loose ref but before actually opening it to read
230 it, another process racing with us can unlink it, which would cause
231 us to barf. The codepath has been updated to retry when such a
232 race is detected, instead of outright failing.
234 * Uses of the platform fnmatch(3) function (many places in the code,
235 matching pathspec, .gitignore and .gitattributes to name a few)
236 have been replaced with wildmatch, allowing "foo/**/bar" that would
237 match foo/bar, foo/a/bar, foo/a/b/bar, etc.
239 * Memory ownership and lifetime rules for what for-each-ref feeds to
240 its callbacks have been clarified (in short, "you do not own it, so
241 make a copy if you want to keep it").
243 * The revision traversal logic to improve culling of irrelevant
244 parents while traversing a mergy history has been updated.
246 * Some leaks in unpack-trees (used in merge, cherry-pick and other
247 codepaths) have been plugged.
249 * The codepath to read from marks files in fast-import/export did not
250 have to accept anything but 40-hex representation of the object
251 name. Further, fast-export did not need full in-core object
252 representation to have parsed wen reading from them. These
253 codepaths have been optimized by taking advantage of these access
254 patterns.
256 * Object lookup logic, when the object hashtable starts to become
257 crowded, has been optimized.
259 * When TEST_OUTPUT_DIRECTORY setting is used, it was handled somewhat
260 inconsistently between the test framework and t/Makefile, and logic
261 to summarize the results looked at a wrong place.
263 * "git clone" uses a lighter-weight implementation when making sure
264 that the history behind refs are complete.
266 * Many warnings from sparse source checker in compat/ area has been
267 squelched.
269 * The code to reading and updating packed-refs file has been updated,
270 correcting corner case bugs.
273 Also contains various documentation updates and code clean-ups.
276 Fixes since v1.8.3
277 ------------------
279 Unless otherwise noted, all the fixes since v1.8.3 in the maintenance
280 track are contained in this release (see release notes to them for
281 details).
283 * Logic to auto-detect character encodings in the commit log message
284 did not reject overlong and invalid UTF-8 characters.
285 (merge 81050ac bc/commit-invalid-utf8 later to maint).
287 * Pass port number as a separate argument when "send-email" initializes
288 Net::SMTP, instead of as a part of the hostname, i.e. host:port.
289 This allows GSSAPI codepath to match with the hostname given.
290 (merge 1a741bf bc/send-email-use-port-as-separate-param later to maint).
292 * "git diff" refused to even show difference when core.safecrlf is
293 set to true (i.e. error out) and there are offending lines in the
294 working tree files.
295 (merge 5430bb2 jc/maint-diff-core-safecrlf later to maint).
297 * A test that should have failed but didn't revealed a bug that needs
298 to be corrected.
299 (merge 94d75d1 jc/t1512-fix later to maint).
301 * An overlong path to a .git directory may have overflown the
302 temporary path buffer used to create a name for lockfiles.
303 (merge 2fbd4f9 mh/maint-lockfile-overflow later to maint).
305 * Invocations of "git checkout" used internally by "git rebase" were
306 counted as "checkout", and affected later "git checkout -" to the
307 the user to an unexpected place.
308 (merge 3bed291 rr/rebase-checkout-reflog later to maint).
310 * "git stash save", when your local change turns a tracked file into
311 a directory, has to remove files in that directory in order to
312 revert your working tree to a pristine state. This will lose
313 untracked files in such a directory, and the command now requires
314 you to "--force" it.
316 * The configuration variable column.ui was poorly documented.
317 (merge 5e62cc1 rr/column-doc later to maint).
319 * "git name-rev --refs=tags/v*" were forbidden, which was a bit
320 inconvenient (you had to give a pattern to match refs fully, like
321 --refs=refs/tags/v*).
322 (merge 98c5c4a nk/name-rev-abbreviated-refs later to maint).
324 * "git apply" parsed patches that add new files, generated by
325 programs other than Git, incorrectly. This is an old breakage in
326 v1.7.11 and will need to be merged down to the maintanance tracks.
328 * Older cURL wanted piece of memory we call it with to be stable, but
329 we updated the auth material after handing it to a call.
331 * "git pull" into nothing trashed "local changes" that were in the
332 index, and this avoids it.
334 * Many "git submodule" operations do not work on a submodule at a
335 path whose name is not in ASCII.
337 * "cherry-pick" had a small leak in an error codepath.
339 * Logic used by git-send-email to suppress cc mishandled names like
340 "A U. Thor" <author@example.xz>, where the human readable part
341 needs to be quoted (the user input may not have the double quotes
342 around the name, and comparison was done between quoted and
343 unquoted strings). It also mishandled names that need RFC2047
344 quoting.
346 * Call to discard_cache/discard_index (used when we use different
347 contents of the index in-core, in many operations like commit,
348 apply, and merge) used to leak memory that held the array of index
349 entries, which has been plugged.
350 (merge a0fc4db rs/discard-index-discard-array later to maint).
352 * "gitweb" forgot to clear a global variable $search_regexp upon each
353 request, mistakenly carrying over the previous search to a new one
354 when used as a persistent CGI.
356 * The wildmatch engine did not honor WM_CASEFOLD option correctly.
358 * "git log -c --follow $path" segfaulted upon hitting the commit that
359 renamed the $path being followed.
361 * When a reflog notation is used for implicit "current branch", we
362 did not say which branch and worse said "branch ''".
364 * "difftool --dir-diff" did not copy back changes made by the
365 end-user in the diff tool backend to the working tree in some
366 cases.
368 * "git push $there HEAD:branch" did not resolve HEAD early enough, so
369 it was easy to flip it around while push is still going on and push
370 out a branch that the user did not originally intended when the
371 command was started.
373 * The bash prompt code (in contrib/) displayed the name of the branch
374 being rebased when "rebase -i/-m/-p" modes are in use, but not the
375 plain vanilla "rebase".
377 * Handling of negative exclude pattern for directories "!dir" was
378 broken in the update to v1.8.3.
380 * zsh prompt script that borrowed from bash prompt script did not
381 work due to slight differences in array variable notation between
382 these two shells.
384 * An entry for "file://" scheme in the enumeration of URL types Git
385 can take in the HTML documentation was made into a clickable link
386 by mistake.
388 * "git push --[no-]verify" was not documented.
390 * Stop installing the git-remote-testpy script that is only used for
391 testing.
393 * "git commit --allow-empty-message -m ''" should not start an
394 editor.
396 * "git merge @{-1}~22" was rewritten to "git merge frotz@{1}~22"
397 incorrectly when your previous branch was "frotz" (it should be
398 rewritten to "git merge frotz~22" instead).
400 * "git diff -c -p" was not showing a deleted line from a hunk when
401 another hunk immediately begins where the earlier one ends.
403 * "git log --ancestry-path A...B" did not work as expected, as it did
404 not pay attention to the fact that the merge base between A and B
405 was the bottom of the range being specified.
407 * Mac OS X does not like to write(2) more than INT_MAX number of
408 bytes; work it around by chopping write(2) into smaller pieces.
410 * Newer MacOS X encourages the programs to compile and link with
411 their CommonCrypto, not with OpenSSL.
413 * "git clone foo/bar:baz" cannot be a request to clone from a remote
414 over git-over-ssh specified in the scp style. This case is now
415 detected and clones from a local repository at "foo/bar:baz".
417 * When $HOME is misconfigured to point at an unreadable directory, we
418 used to complain and die. Loosen the check.
420 * "git subtree" (in contrib/) had one codepath with loose error
421 checks to lose data at the remote side.
423 * "git fetch" into a shallow repository from a repository that does
424 not know about the shallow boundary commits (e.g. a different fork
425 from the repository the current shallow repository was cloned from)
426 did not work correctly.
428 * "git checkout foo" DWIMs the intended "upstream" and turns it into
429 "git checkout -t -b foo remotes/origin/foo". This codepath has been
430 updated to correctly take existing remote definitions into account.