Update draft release notes to 1.8.3
[git/git.git] / Documentation / RelNotes / 1.8.3.txt
1 Git v1.8.3 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 only the current branch to the branch with the same
12 name, and only when the current branch is set to integrate with that
13 remote branch. Use the user preference configuration variable
14 "push.default" to change this. If you are an old-timer who is used
15 to the "matching" semantics, you can set the variable to "matching"
16 to keep the traditional behaviour. If you want to live in the future
17 early, you can set it to "simple" today without waiting for Git 2.0.
19 When "git add -u" (and "git add -A") is run inside a subdirectory and
20 does not specify which paths to add on the command line, it
21 will operate on the entire tree in Git 2.0 for consistency
22 with "git commit -a" and other commands. There will be no
23 mechanism to make plain "git add -u" behave like "git add -u .".
24 Current users of "git add -u" (without a pathspec) should start
25 training their fingers to explicitly say "git add -u ."
26 before Git 2.0 comes. A warning is issued when these commands are
27 run without a pathspec and when you have local changes outside the
28 current directory, because the behaviour in Git 2.0 will be different
29 from today's version in such a situation.
31 In Git 2.0, "git add <path>" will behave as "git add -A <path>", so
32 that "git add dir/" will notice paths you removed from the directory
33 and record the removal. Versions before Git 2.0, including this
34 release, will keep ignoring removals, but the users who rely on this
35 behaviour are encouraged to start using "git add --ignore-removal <path>"
36 now before 2.0 is released.
39 Updates since v1.8.2
40 --------------------
42 Foreign interface
44 * remote-hg and remote-bzr helpers (in contrib/) have been updated.
47 UI, Workflows & Features
49 * The prompt string generator (in contrib/completion/) learned to
50 show how many changes there are in total and how many have been
51 replayed during a "git rebase" session.
53 * "git branch --vv" learned to paint the name of the branch it
54 integrates with in a different color (color.branch.upstream,
55 which defaults to blue).
57 * In a sparsely populated working tree, "git checkout <pathspec>" no
58 longer unmarks paths that match the given pathspec that were
59 originally ignored with "--sparse" (use --ignore-skip-worktree-bits
60 option to resurrect these paths out of the index if you really want
61 to).
63 * "git log --format" specifier learned %C(auto) token that tells Git
64 to use color when interpolating %d (decoration), %h (short commit
65 object name), etc. for terminal output.
67 * "git bisect" leaves the final outcome as a comment in its bisect
68 log file.
70 * "git clone --reference" can now refer to a gitfile "textual symlink"
71 that points at the real location of the repository.
73 * "git count-objects" learned "--human-readable" aka "-H" option to
74 show various large numbers in Ki/Mi/GiB scaled as necessary.
76 * "git cherry-pick $blob" and "git cherry-pick $tree" are nonsense,
77 and a more readable error message e.g. "can't cherry-pick a tree"
78 is given (we used to say "expected exactly one commit").
80 * The "--annotate" option to "git send-email" can be turned on (or
81 off) by default with sendemail.annotate configuration variable (you
82 can use --no-annotate from the command line to override it).
84 * The "--cover-letter" option to "git format-patch" can be turned on
85 (or off) by default with format.coverLetter configuration
86 variable. By setting it to 'auto', you can turn it on only for a
87 series with two or more patches.
89 * The bash completion support (in contrib/) learned that cherry-pick
90 takes a few more options than it already knew about.
92 * "git help" learned "-g" option to show the list of guides just like
93 list of commands are given with "-a".
95 * A triangular "pull from one place, push to another place" workflow
96 is supported better by new remote.pushdefault (overrides the
97 "origin" thing) and branch.*.pushremote (overrides the
98 branch.*.remote) configuration variables.
100 * "git status" learned to report that you are in the middle of a
101 revert session, just like it does for a cherry-pick and a bisect
102 session.
104 * The handling by "git branch --set-upstream-to" against various forms
105 of erroneous inputs was suboptimal and has been improved.
107 * When the interactive access to git-shell is not enabled, it issues
108 a message meant to help the system administrator to enable it.
109 An explicit way to help the end users who connect to the service by
110 issuing custom messages to refuse such an access has been added.
112 * In addition to the case where the user edits the log message with
113 the "e)dit" option of "am -i", replace the "Applying: this patch"
114 message with the final log message contents after applymsg hook
115 munges it.
117 * "git status" suggests users to look into using --untracked=no option
118 when it takes too long.
120 * "git status" shows a bit more information during a
121 rebase/bisect session.
123 * "git fetch" learned to fetch a commit at the tip of an unadvertised
124 ref by specifying a raw object name from the command line when the
125 server side supports this feature.
127 * Output from "git log --graph" works better with submodule log
128 output now.
130 * "git count-objects -v" learned to report leftover temporary
131 packfiles and other garbage in the object store.
133 * A new read-only credential helper (in contrib/) to interact with
134 the .netrc/.authinfo files has been added.
136 * "git send-email" can be used with the credential helper system.
138 * There was no Porcelain way to say "I no longer am interested in
139 this submodule", once you express your interest in a submodule with
140 "submodule init". "submodule deinit" is the way to do so.
142 * "git pull --rebase" learned to pass "-v/-q" options to underlying
143 "git rebase".
145 * The new "--follow-tags" option tells "git push" to push relevant
146 annotated tags when pushing branches out.
148 * "git merge" and "git pull" can optionally be told to inspect and
149 reject when merging a commit that does not carry a trusted GPG
150 signature.
152 * "git mergetool" now feeds files to the "p4merge" backend in the
153 order that matches the p4 convention, where "theirs" is usually
154 shown on the left side, which is the opposite from what other backends
155 expect.
157 * "show/log" now honors gpg.program configuration just like other
158 parts of the code that use GnuPG.
160 * "git log" that shows the difference between the parent and the
161 child has been optimized somewhat.
163 * "git difftool" allows the user to write into the temporary files
164 being shown; if the user makes changes to the working tree at the
165 same time, it now refrains from overwriting the copy in the working
166 tree and leaves the temporary file so that changes can be merged
167 manually.
169 * There was no good way to ask "I have a random string that came from
170 outside world. I want to turn it into a 40-hex object name while
171 making sure such an object exists". A new peeling suffix ^{object}
172 can be used for that purpose, together with "rev-parse --verify".
175 Performance, Internal Implementation, etc.
177 * Updates for building under msvc.
179 * A handful of issues in the code that traverses the working tree to find
180 untracked and/or ignored files have been fixed, and the general
181 codepath involved in "status -u" and "clean" have been cleaned up
182 and optimized.
184 * The stack footprint of some codepaths that access an object from a
185 pack has been shrunk.
187 * The logic to coalesce the same lines removed from the parents in
188 the output from "diff -c/--cc" has been updated, but with O(n^2)
189 complexity, so this might turn out to be undesirable.
191 * The code to enforce permission bits on files in $GIT_DIR/ for
192 shared repositories has been simplified.
194 * A few codepaths know how much data they need to put in the
195 hashtables they use when they start, but still began with small tables
196 and repeatedly grew and rehashed them.
198 * The API to walk reflog entries from the latest to older, which was
199 necessary for operations such as "git checkout -", was cumbersome
200 to use correctly and also inefficient.
202 * Codepaths that inspect log-message-to-be and decide when to add a
203 new Signed-off-by line in various commands have been consolidated.
205 * The pkt-line API, implementation and its callers have been cleaned
206 up to make them more robust.
208 * The Cygwin port has a faster-but-lying lstat(2) emulation whose
209 incorrectness does not matter in practice except for a few
210 codepaths, and setting permission bits on directories is a codepath
211 that needs to use a more correct one.
213 * "git checkout" had repeated pathspec matches on the same paths,
214 which have been consolidated. Also a bug in "git checkout dir/"
215 that is started from an unmerged index has been fixed.
217 * A few bugfixes to "git rerere" working on corner case merge
218 conflicts have been applied.
221 Also contains various documentation updates and code clean-ups.
224 Fixes since v1.8.2
225 ------------------
227 Unless otherwise noted, all the fixes since v1.8.2 in the maintenance
228 track are contained in this release (see release notes to them for
229 details).
231 * Recent versions of File::Temp (used by "git svn") started blowing
232 up when its tempfile sub is called as a class method; updated the
233 callsite to call it as a plain vanilla function to fix it.
234 (merge eafc2dd hb/git-pm-tempfile later to maint).
236 * Various subcommands of "git remote" simply ignored extraneous
237 command line arguments instead of diagnosing them as errors.
238 (merge b17dd3f tr/remote-tighten-commandline-parsing later to maint).
240 * When receive-pack detects an error in the pack header it received in
241 order to decide which of unpack-objects or index-pack to run, it
242 returned without closing the error stream, which led to a hung
243 sideband thread.
245 * Zsh completion forgot that the '%' character used to signal untracked
246 files needs to be escaped with another '%'.
248 * A commit object whose author or committer ident are malformed
249 crashed some code that trusted that a name, an email and a
250 timestamp can always be found in it.
252 * When "upload-pack" fails while generating a pack in response to
253 "git fetch" (or "git clone"), the receiving side had
254 a programming error that triggered the die handler
255 recursively.
257 * "rev-list --stdin" and friends kept bogus pointers into the input
258 buffer around as human readable object names. This was not a huge
259 problem but was exposed by a new change that uses these names in
260 error output.
261 (merge 70d26c6 tr/copy-revisions-from-stdin later to maint).
263 * Smart-capable HTTP servers were not restricted via the
264 GIT_NAMESPACE mechanism when talking with commit-walking clients,
265 like they are when talking with smart HTTP clients.
266 (merge 6130f86 jk/http-dumb-namespaces later to maint).
268 * "git merge-tree" did not omit a merge result that is identical to
269 the "our" side in certain cases.
270 (merge aacecc3 jk/merge-tree-added-identically later to maint).
272 * Perl scripts like "git-svn" closed (instead of redirecting to /dev/null)
273 the standard error stream, which is not a very smart thing to do.
274 A later open may return file descriptor #2 for an unrelated purpose, and
275 error reporting code may write into it.
277 * "git show-branch" was not prepared to show a very long run of
278 ancestor operators e.g. foobar^2~2^2^2^2...^2~4 correctly.
280 * "git diff --diff-algorithm algo" is also understood as "git diff
281 --diff-algorithm=algo".
283 * The new core.commentchar configuration was not applied in a few
284 places.
286 * "git bundle" erroneously bailed out when parsing a valid bundle
287 containing a prerequisite commit without a commit message.
289 * "git log -S/-G" started paying attention to textconv filter, but
290 there was no way to disable this. Make it honor the --no-textconv
291 option.
293 * When used with the "-d temporary-directory" option, "git filter-branch"
294 failed to come back to the original working tree to perform the
295 final clean-up procedure.
297 * "git merge $(git rev-parse v1.8.2)" behaved quite differently from
298 "git merge v1.8.2", as if v1.8.2 were written as v1.8.2^0 and did
299 not pay much attention to the annotated tag payload. Make the code
300 notice the type of the tag object, in addition to the dwim_ref()
301 based classification the current code uses (i.e. the name appears
302 in refs/tags/) to decide when to special-case tag merging.
304 * Fix a 1.8.1.x regression that stopped matching "dir" (without a
305 trailing slash) to a directory "dir".
306 (merge efa5f82 jc/directory-attrs-regression-fix later to maint-1.8.1).
308 * "git apply --whitespace=fix" was not prepared to see a line getting
309 longer after fixing whitespaces (e.g. tab-in-indent aka Python).
310 (merge 329b26e jc/apply-ws-fix-tab-in-indent later to maint-1.8.1).
312 * The prompt string generator (in contrib/completion/) did not notice
313 when we are in a middle of a "git revert" session.
315 * "submodule summary --summary-limit" option did not support the
316 "--option=value" form.
318 * "index-pack --fix-thin" used an uninitialized value to compute
319 the delta depths of objects it appends to the resulting pack.
321 * "index-pack --verify-stat" used a few counters outside the protection
322 of a mutex, possibly showing incorrect numbers.
324 * The code to keep track of what directory names are known to Git on
325 platforms with case insensitive filesystems could get confused upon a
326 hash collision between these pathnames and would loop forever.
328 * Annotated tags outside the refs/tags/ hierarchy were not advertised
329 correctly to ls-remote and fetch with recent versions of Git.
331 * Recent optimizations broke shallow clones.
333 * "git cmd -- ':(top'" was not diagnosed as an invalid syntax, and
334 instead the parser kept reading beyond the end of the string.
336 * "git tag -f <tag>" always said "Updated tag '<tag>'" even when
337 creating a new tag (i.e. neither overwriting nor updating).
339 * "git p4" did not behave well when the path to the root of the P4
340 client was not its real path.
341 (merge bbd8486 pw/p4-symlinked-root later to maint).
343 * "git archive" reported a failure when asked to create an archive out
344 of an empty tree. It is more intuitive to give an empty
345 archive back in such a case.
347 * When "format-patch" quoted a non-ascii string in header files,
348 it incorrectly applied rfc2047 and chopped a single character in
349 the middle of the string.
351 * An aliased command spawned from a bare repository that does not say
352 it is bare with "core.bare = yes" was treated as non-bare by mistake.
354 * In "git reflog expire", the REACHABLE bit was not cleared from the
355 correct objects.
357 * The logic used by "git diff -M --stat" to shorten the names of
358 files before and after a rename did not work correctly when the
359 common prefix and suffix between the two filenames overlapped.
361 * The "--match=<pattern>" option of "git describe", when used with
362 "--all" to allow refs that are not annotated tags to be a
363 base of description, did not restrict the output from the command
364 to those refs that match the given pattern.
366 * Clarify in the documentation "what" gets pushed to "where" when the
367 command line to "git push" does not say these explicitly.
369 * The "--color=<when>" argument to the commands in the diff family
370 was described poorly.
372 * The arguments given to the pre-rebase hook were not documented.
374 * The v4 index format was not documented.
376 * The "--match=<pattern>" argument "git describe" takes uses glob
377 pattern but it wasn't obvious from the documentation.
379 * Some sources failed to compile on systems that lack NI_MAXHOST in
380 their system header (e.g. z/OS).
382 * Add an example use of "--env-filter" in "filter-branch"
383 documentation.
385 * "git bundle verify" did not say "records a complete history" for a
386 bundle that does not have any prerequisites.
388 * In the v1.8.0 era, we changed symbols that do not have to be global
389 to file scope static, but a few functions in graph.c were used by
390 CGit sideways, bypassing the entry points of the API the
391 in-tree users use.
393 * "git update-index -h" did not do the usual "-h(elp)" thing.
395 * "git index-pack" had a buffer-overflow while preparing an
396 informational message when the translated version of it was too
397 long.
399 * 'git commit -m "$msg"' used to add an extra newline even when
400 $msg already ended with one.
402 * The SSL peer verification done by "git imap-send" did not ask for
403 Server Name Indication (RFC 4366), failing to connect to SSL/TLS
404 sites that serve multiple hostnames on a single IP.
406 * perl/Git.pm::cat_blob slurped everything in core only to write it
407 out to a file descriptor, which was not a very smart thing to do.
409 * "git branch" did not bother to check nonsense command line
410 parameters. It now issues errors in many cases.
412 * Verification of signed tags was not done correctly when not in C
413 or en/US locale.
415 * Some platforms and users spell UTF-8 differently; retry with the
416 most official "UTF-8" when the system does not understand the
417 user-supplied encoding name that is a common alternative
418 spelling of UTF-8.
420 * When export-subst is used, "zip" output recorded an incorrect
421 size of the file.
423 * "git am $maildir/" applied messages in an unexpected order; sort
424 filenames read from the maildir/ in a way that is more likely to
425 sort the messages in the order the writing MUA meant to, by sorting
426 numeric segments in numeric order and non-numeric segments in
427 alphabetical order.
429 * "git submodule update", when recursed into sub-submodules, did not
430 accumulate the prefix paths.