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.
44 * remote-hg and remote-bzr helpers (in contrib/) have been updated;
45 especially, the latter has been accelerated to help Emacs folks,
46 whose primary SCM seems to be stagnating.
48 UI, Workflows & Features
50 * The prompt string generator (in contrib/completion/) learned to
51 show how many changes there are in total and how many have been
52 replayed during a "git rebase" session.
54 * "git branch --vv" learned to paint the name of the branch it
55 integrates with in a different color (color.branch.upstream,
56 which defaults to blue).
58 * In a sparsely populated working tree, "git checkout <pathspec>" no
59 longer unmarks paths that match the given pathspec that were
60 originally ignored with "--sparse" (use --ignore-skip-worktree-bits
61 option to resurrect these paths out of the index if you really want
64 * "git log --format" specifier learned %C(auto) token that tells Git
65 to use color when interpolating %d (decoration), %h (short commit
66 object name), etc. for terminal output.
68 * "git bisect" leaves the final outcome as a comment in its bisect
71 * "git clone --reference" can now refer to a gitfile "textual symlink"
72 that points at the real location of the repository.
74 * "git count-objects" learned "--human-readable" aka "-H" option to
75 show various large numbers in Ki/Mi/GiB scaled as necessary.
77 * "git cherry-pick $blob" and "git cherry-pick $tree" are nonsense,
78 and a more readable error message e.g. "can't cherry-pick a tree"
79 is given (we used to say "expected exactly one commit").
81 * The "--annotate" option to "git send-email" can be turned on (or
82 off) by default with sendemail.annotate configuration variable (you
83 can use --no-annotate from the command line to override it).
85 * The "--cover-letter" option to "git format-patch" can be turned on
86 (or off) by default with format.coverLetter configuration
87 variable. By setting it to 'auto', you can turn it on only for a
88 series with two or more patches.
90 * The bash completion support (in contrib/) learned that cherry-pick
91 takes a few more options than it already knew about.
93 * "git help" learned "-g" option to show the list of guides just like
94 list of commands are given with "-a".
96 * A triangular "pull from one place, push to another place" workflow
97 is supported better by new remote.pushdefault (overrides the
98 "origin" thing) and branch.*.pushremote (overrides the
99 branch.*.remote) configuration variables.
101 * "git status" learned to report that you are in the middle of a
102 revert session, just like it does for a cherry-pick and a bisect
105 * The handling by "git branch --set-upstream-to" against various forms
106 of erroneous inputs was suboptimal and has been improved.
108 * When the interactive access to git-shell is not enabled, it issues
109 a message meant to help the system administrator to enable it.
110 An explicit way to help the end users who connect to the service by
111 issuing custom messages to refuse such an access has been added.
113 * In addition to the case where the user edits the log message with
114 the "e)dit" option of "am -i", replace the "Applying: this patch"
115 message with the final log message contents after applymsg hook
118 * "git status" suggests users to look into using --untracked=no option
119 when it takes too long.
121 * "git status" shows a bit more information during a
122 rebase/bisect session.
124 * "git fetch" learned to fetch a commit at the tip of an unadvertised
125 ref by specifying a raw object name from the command line when the
126 server side supports this feature.
128 * Output from "git log --graph" works better with submodule log
131 * "git count-objects -v" learned to report leftover temporary
132 packfiles and other garbage in the object store.
134 * A new read-only credential helper (in contrib/) to interact with
135 the .netrc/.authinfo files has been added.
137 * "git send-email" can be used with the credential helper system.
139 * There was no Porcelain way to say "I no longer am interested in
140 this submodule", once you express your interest in a submodule with
141 "submodule init". "submodule deinit" is the way to do so.
143 * "git pull --rebase" learned to pass "-v/-q" options to underlying
146 * The new "--follow-tags" option tells "git push" to push relevant
147 annotated tags when pushing branches out.
149 * "git merge" and "git pull" can optionally be told to inspect and
150 reject when merging a commit that does not carry a trusted GPG
153 * "git mergetool" now feeds files to the "p4merge" backend in the
154 order that matches the p4 convention, where "theirs" is usually
155 shown on the left side, which is the opposite from what other backends
158 * "show/log" now honors gpg.program configuration just like other
159 parts of the code that use GnuPG.
161 * "git log" that shows the difference between the parent and the
162 child has been optimized somewhat.
164 * "git difftool" allows the user to write into the temporary files
165 being shown; if the user makes changes to the working tree at the
166 same time, it now refrains from overwriting the copy in the working
167 tree and leaves the temporary file so that changes can be merged
170 * There was no good way to ask "I have a random string that came from
171 outside world. I want to turn it into a 40-hex object name while
172 making sure such an object exists". A new peeling suffix ^{object}
173 can be used for that purpose, together with "rev-parse --verify".
176 Performance, Internal Implementation, etc.
178 * Updates for building under msvc.
180 * A handful of issues in the code that traverses the working tree to find
181 untracked and/or ignored files have been fixed, and the general
182 codepath involved in "status -u" and "clean" have been cleaned up
185 * The stack footprint of some codepaths that access an object from a
186 pack has been shrunk.
188 * The logic to coalesce the same lines removed from the parents in
189 the output from "diff -c/--cc" has been updated, but with O(n^2)
190 complexity, so this might turn out to be undesirable.
192 * The code to enforce permission bits on files in $GIT_DIR/ for
193 shared repositories has been simplified.
195 * A few codepaths know how much data they need to put in the
196 hashtables they use when they start, but still began with small tables
197 and repeatedly grew and rehashed them.
199 * The API to walk reflog entries from the latest to older, which was
200 necessary for operations such as "git checkout -", was cumbersome
201 to use correctly and also inefficient.
203 * Codepaths that inspect log-message-to-be and decide when to add a
204 new Signed-off-by line in various commands have been consolidated.
206 * The pkt-line API, implementation and its callers have been cleaned
207 up to make them more robust.
209 * The Cygwin port has a faster-but-lying lstat(2) emulation whose
210 incorrectness does not matter in practice except for a few
211 codepaths, and setting permission bits on directories is a codepath
212 that needs to use a more correct one.
214 * "git checkout" had repeated pathspec matches on the same paths,
215 which have been consolidated. Also a bug in "git checkout dir/"
216 that is started from an unmerged index has been fixed.
218 * A few bugfixes to "git rerere" working on corner case merge
219 conflicts have been applied.
222 Also contains various documentation updates and code clean-ups.
228 Unless otherwise noted, all the fixes since v1.8.2 in the maintenance
229 track are contained in this release (see release notes to them for
232 * Recent versions of File::Temp (used by "git svn") started blowing
233 up when its tempfile sub is called as a class method; updated the
234 callsite to call it as a plain vanilla function to fix it.
235 (merge eafc2dd hb/git-pm-tempfile later to maint).
237 * Various subcommands of "git remote" simply ignored extraneous
238 command line arguments instead of diagnosing them as errors.
239 (merge b17dd3f tr/remote-tighten-commandline-parsing later to maint).
241 * When receive-pack detects an error in the pack header it received in
242 order to decide which of unpack-objects or index-pack to run, it
243 returned without closing the error stream, which led to a hung
246 * Zsh completion forgot that the '%' character used to signal untracked
247 files needs to be escaped with another '%'.
249 * A commit object whose author or committer ident are malformed
250 crashed some code that trusted that a name, an email and a
251 timestamp can always be found in it.
253 * When "upload-pack" fails while generating a pack in response to
254 "git fetch" (or "git clone"), the receiving side had
255 a programming error that triggered the die handler
258 * "rev-list --stdin" and friends kept bogus pointers into the input
259 buffer around as human readable object names. This was not a huge
260 problem but was exposed by a new change that uses these names in
262 (merge 70d26c6 tr/copy-revisions-from-stdin later to maint).
264 * Smart-capable HTTP servers were not restricted via the
265 GIT_NAMESPACE mechanism when talking with commit-walking clients,
266 like they are when talking with smart HTTP clients.
267 (merge 6130f86 jk/http-dumb-namespaces later to maint).
269 * "git merge-tree" did not omit a merge result that is identical to
270 the "our" side in certain cases.
271 (merge aacecc3 jk/merge-tree-added-identically later to maint).
273 * Perl scripts like "git-svn" closed (instead of redirecting to /dev/null)
274 the standard error stream, which is not a very smart thing to do.
275 A later open may return file descriptor #2 for an unrelated purpose, and
276 error reporting code may write into it.
278 * "git show-branch" was not prepared to show a very long run of
279 ancestor operators e.g. foobar^2~2^2^2^2...^2~4 correctly.
281 * "git diff --diff-algorithm algo" is also understood as "git diff
282 --diff-algorithm=algo".
284 * The new core.commentchar configuration was not applied in a few
287 * "git bundle" erroneously bailed out when parsing a valid bundle
288 containing a prerequisite commit without a commit message.
290 * "git log -S/-G" started paying attention to textconv filter, but
291 there was no way to disable this. Make it honor the --no-textconv
294 * When used with the "-d temporary-directory" option, "git filter-branch"
295 failed to come back to the original working tree to perform the
296 final clean-up procedure.
298 * "git merge $(git rev-parse v1.8.2)" behaved quite differently from
299 "git merge v1.8.2", as if v1.8.2 were written as v1.8.2^0 and did
300 not pay much attention to the annotated tag payload. Make the code
301 notice the type of the tag object, in addition to the dwim_ref()
302 based classification the current code uses (i.e. the name appears
303 in refs/tags/) to decide when to special-case tag merging.
305 * Fix a 1.8.1.x regression that stopped matching "dir" (without a
306 trailing slash) to a directory "dir".
307 (merge efa5f82 jc/directory-attrs-regression-fix later to maint-1.8.1).
309 * "git apply --whitespace=fix" was not prepared to see a line getting
310 longer after fixing whitespaces (e.g. tab-in-indent aka Python).
311 (merge 329b26e jc/apply-ws-fix-tab-in-indent later to maint-1.8.1).
313 * The prompt string generator (in contrib/completion/) did not notice
314 when we are in a middle of a "git revert" session.
316 * "submodule summary --summary-limit" option did not support the
317 "--option=value" form.
319 * "index-pack --fix-thin" used an uninitialized value to compute
320 the delta depths of objects it appends to the resulting pack.
322 * "index-pack --verify-stat" used a few counters outside the protection
323 of a mutex, possibly showing incorrect numbers.
325 * The code to keep track of what directory names are known to Git on
326 platforms with case insensitive filesystems could get confused upon a
327 hash collision between these pathnames and would loop forever.
329 * Annotated tags outside the refs/tags/ hierarchy were not advertised
330 correctly to ls-remote and fetch with recent versions of Git.
332 * Recent optimizations broke shallow clones.
334 * "git cmd -- ':(top'" was not diagnosed as an invalid syntax, and
335 instead the parser kept reading beyond the end of the string.
337 * "git tag -f <tag>" always said "Updated tag '<tag>'" even when
338 creating a new tag (i.e. neither overwriting nor updating).
340 * "git p4" did not behave well when the path to the root of the P4
341 client was not its real path.
342 (merge bbd8486 pw/p4-symlinked-root later to maint).
344 * "git archive" reported a failure when asked to create an archive out
345 of an empty tree. It is more intuitive to give an empty
346 archive back in such a case.
348 * When "format-patch" quoted a non-ascii string in header files,
349 it incorrectly applied rfc2047 and chopped a single character in
350 the middle of the string.
352 * An aliased command spawned from a bare repository that does not say
353 it is bare with "core.bare = yes" was treated as non-bare by mistake.
355 * In "git reflog expire", the REACHABLE bit was not cleared from the
358 * The logic used by "git diff -M --stat" to shorten the names of
359 files before and after a rename did not work correctly when the
360 common prefix and suffix between the two filenames overlapped.
362 * The "--match=<pattern>" option of "git describe", when used with
363 "--all" to allow refs that are not annotated tags to be a
364 base of description, did not restrict the output from the command
365 to those refs that match the given pattern.
367 * Clarify in the documentation "what" gets pushed to "where" when the
368 command line to "git push" does not say these explicitly.
370 * The "--color=<when>" argument to the commands in the diff family
371 was described poorly.
373 * The arguments given to the pre-rebase hook were not documented.
375 * The v4 index format was not documented.
377 * The "--match=<pattern>" argument "git describe" takes uses glob
378 pattern but it wasn't obvious from the documentation.
380 * Some sources failed to compile on systems that lack NI_MAXHOST in
381 their system header (e.g. z/OS).
383 * Add an example use of "--env-filter" in "filter-branch"
386 * "git bundle verify" did not say "records a complete history" for a
387 bundle that does not have any prerequisites.
389 * In the v1.8.0 era, we changed symbols that do not have to be global
390 to file scope static, but a few functions in graph.c were used by
391 CGit sideways, bypassing the entry points of the API the
394 * "git update-index -h" did not do the usual "-h(elp)" thing.
396 * "git index-pack" had a buffer-overflow while preparing an
397 informational message when the translated version of it was too
400 * 'git commit -m "$msg"' used to add an extra newline even when
401 $msg already ended with one.
403 * The SSL peer verification done by "git imap-send" did not ask for
404 Server Name Indication (RFC 4366), failing to connect to SSL/TLS
405 sites that serve multiple hostnames on a single IP.
407 * perl/Git.pm::cat_blob slurped everything in core only to write it
408 out to a file descriptor, which was not a very smart thing to do.
410 * "git branch" did not bother to check nonsense command line
411 parameters. It now issues errors in many cases.
413 * Verification of signed tags was not done correctly when not in C
416 * Some platforms and users spell UTF-8 differently; retry with the
417 most official "UTF-8" when the system does not understand the
418 user-supplied encoding name that is a common alternative
421 * When export-subst is used, "zip" output recorded an incorrect
424 * "git am $maildir/" applied messages in an unexpected order; sort
425 filenames read from the maildir/ in a way that is more likely to
426 sort the messages in the order the writing MUA meant to, by sorting
427 numeric segments in numeric order and non-numeric segments in
430 * "git submodule update", when recursed into sub-submodules, did not
431 accumulate the prefix paths.
projects / thirdparty / git.git / blob