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