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