10 UI, Workflows & Features
12 * "git p4" now detects the filetype (e.g. binary) correctly even when
13 the files are opened exclusively.
15 * git p4 attempts to better handle branches in Perforce.
17 * "git p4" learned "--changes-block-size <n>" to read the changes in
18 chunks from Perforce, instead of making one call to "p4 changes"
19 that may trigger "too many rows scanned" error from Perforce.
21 * "git show-branch --topics HEAD" (with no other arguments) did not
22 do anything interesting. Instead, contrast the given revision
23 against all the local branches by default.
25 * A replacement for contrib/workdir/git-new-workdir that does not
26 rely on symbolic links and make sharing of objects and refs safer
27 by making the borrowee and borrowers aware of each other.
29 * Tweak the sample "store" backend of the credential helper to honor
30 XDG configuration file locations when specified.
32 * A heuristic to help the "git <cmd> <revs> <pathspec>" command line
33 convention to catch mistyped paths is to make sure all the non-rev
34 parameters in the later part of the command line are names of the
35 files in the working tree, but that means "git grep $str -- \*.c"
36 must always be disambiguated with "--", because nobody sane will
37 create a file whose name literally is asterisk-dot-see. Loosen the
38 heuristic to declare that with a wildcard string the user likely
39 meant to give us a pathspec.
41 * "git merge FETCH_HEAD" learned that the previous "git fetch" could
42 be to create an Octopus merge, i.e. recording multiple branches
43 that are not marked as "not-for-merge"; this allows us to lose an
44 old style invocation "git merge <msg> HEAD $commits..." in the
45 implementation of "git pull" script; the old style syntax can now
48 * Help us to find broken test script that splits the body part of the
49 test by mistaken use of wrong kind of quotes.
50 (merge d93d5d5 jc/test-prereq-validate later to maint).
52 * Developer support to automatically detect broken &&-chain in the
53 test scripts is now turned on by default.
54 (merge 92b269f jk/test-chain-lint later to maint).
56 * Filter scripts were run with SIGPIPE disabled on the Git side,
57 expecting that they may not read what Git feeds them to filter.
58 We however treated a filter that does not read its input fully
59 before exiting as an error.
61 This changes semantics, but arguably in a good way. If a filter
62 can produce its output without consuming its input using whatever
63 magic, we now let it do so, instead of diagnosing it as a
66 * Instead of dying immediately upon failing to obtain a lock, the
67 locking (of refs etc) retries after a short while with backoff.
69 * Introduce http.<url>.SSLCipherList configuration variable to tweak
70 the list of cipher suite to be used with libcURL when talking with
73 * "git subtree" script (in contrib/) used "echo -n" to produce
74 progress messages in a non-portable way.
76 * "git subtree" script (in contrib/) does not have --squash option
77 when pushing, but the documentation and help text pretended as if
80 * The Git subcommand completion (in contrib/) listed credential
81 helpers among candidates, which is not something the end user would
85 Performance, Internal Implementation, Development Support etc.
87 * "unsigned char [20]" used thoughout the code to represent object
88 names are being converted into a semi-opaque "struct object_id".
89 This effort is expected to interfere with other topics in flight,
90 but hopefully will give us one extra level of abstraction in the
93 * Catch a programmer mistake to feed a pointer not an array to
94 ARRAY_SIZE() macro, by using a couple of GCC extensions.
95 (merge 89c855e ep/do-not-feed-a-pointer-to-array-size later to maint).
97 * Some error messages in "git config" were emitted without calling
98 the usual error() facility.
100 * When "add--interactive" splits a hunk into two overlapping hunks
101 and then let the user choose only one, it sometimes feeds an
102 incorrect patch text to "git apply". Add tests to demonstrate
105 I have a slight suspicion that this may be $gmane/87202 coming back
106 and biting us (I seem to have said "let's run with this and see
107 what happens" back then).
109 * More line-ending tests.
111 * An earlier rewrite to use strbuf_getwholeline() instead of fgets(3)
112 to read packed-refs file revealed that the former is unacceptably
115 * The refs API uses ref_lock struct which had its own "int fd", even
116 though the same file descriptor was in the lock struct it contains.
117 Clean-up the code to lose this redundant field.
119 * Add the "--allow-unknown-type" option to "cat-file" to allow
120 inspecting loose objects of an experimental or a broken type.
122 * Many long-running operations show progress eye-candy, even when
123 they are later backgrounded. Hide the eye-candy when the process
124 is sent to the background instead.
125 (merge 9a9a41d lm/squelch-bg-progress later to maint).
127 * There was a dead code that used to handle "git pull --tags" and
128 show special-cased error message, which was made irrelevant when
129 the semantics of the option changed back in Git 1.9 days.
130 (merge 19d122b pt/pull-tags-error-diag later to maint).
133 Also contains various documentation updates and code clean-ups.
139 Unless otherwise noted, all the fixes since v2.4 in the maintenance
140 track are contained in this release (see the maintenance releases'
143 * Git 2.4 broke setting verbosity and progress levels on "git clone"
144 with native transports.
145 (merge 822f0c4 mh/clone-verbosity-fix later to maint).
147 * "git add -e" did not allow the user to abort the operation by
149 (merge cb64800 jk/add-e-kill-editor later to maint).
151 * Memory usage of "git index-pack" has been trimmed by tens of
153 (merge c6458e6 nd/slim-index-pack-memory-usage later to maint).
155 * "git rev-list --objects $old --not --all" to see if everything that
156 is reachable from $old is already connected to the existing refs
157 was very inefficient.
158 (merge b6e8a3b jk/still-interesting later to maint).
160 * "hash-object --literally" introduced in v2.2 was not prepared to
161 take a really long object type name.
162 (merge 1427a7f jc/hash-object later to maint).
164 * "git rebase --quiet" was not quite quiet when there is nothing to
166 (merge 22946a9 jk/rebase-quiet-noop later to maint).
168 * The completion for "log --decorate=" parameter value was incorrect.
169 (merge af16bda sg/complete-decorate-full-not-long later to maint).
171 * "filter-branch" corrupted commit log message that ends with an
172 incomplete line on platforms with some "sed" implementations that
173 munge such a line. Work it around by avoiding to use "sed".
174 (merge df06201 jk/filter-branch-use-of-sed-on-incomplete-line later to maint).
176 * "git daemon" fails to build from the source under NO_IPV6
177 configuration (regression in 2.4).
178 (merge d358f77 jc/daemon-no-ipv6-for-2.4.1 later to maint).
180 * Some time ago, "git blame" (incorrectly) lost the convert_to_git()
181 call when synthesizing a fake "tip" commit that represents the
182 state in the working tree, which broke folks who record the history
183 with LF line ending to make their project portabile across
184 platforms while terminating lines in their working tree files with
185 CRLF for their platform.
186 (merge 4bf256d tb/blame-resurrect-convert-to-git later to maint).
188 * We avoid setting core.worktree when the repository location is the
189 ".git" directory directly at the top level of the working tree, but
190 the code misdetected the case in which the working tree is at the
191 root level of the filesystem (which arguably is a silly thing to
192 do, but still valid).
193 (merge 84ccad8 jk/init-core-worktree-at-root later to maint).
195 * "git commit --date=now" or anything that relies on approxidate lost
196 the daylight-saving-time offset.
197 (merge f6e6362 jc/epochtime-wo-tz later to maint).
199 * Access to objects in repositories that borrow from another one on a
200 slow NFS server unnecessarily got more expensive due to recent code
201 becoming more cautious in a naive way not to lose objects to pruning.
202 (merge ee1c6c3 jk/prune-mtime later to maint).
204 * The codepaths that read .gitignore and .gitattributes files have been
205 taught that these files encoded in UTF-8 may have UTF-8 BOM marker at
206 the beginning; this makes it in line with what we do for configuration
208 (merge 27547e5 cn/bom-in-gitignore later to maint).
210 * a few helper scripts in the test suite did not report errors
212 (merge de248e9 ep/fix-test-lib-functions-report later to maint).
214 * The default $HOME/.gitconfig file created upon "git config --global"
215 that edits it had incorrectly spelled user.name and user.email
217 (merge 7e11052 oh/fix-config-default-user-name-section later to maint).
219 * "git cat-file bl $blob" failed to barf even though there is no
220 object type that is "bl".
221 (merge b7994af jk/type-from-string-gently later to maint).
223 * The usual "git diff" when seeing a file turning into a directory
224 showed a patchset to remove the file and create all files in the
225 directory, but "git diff --no-index" simply refused to work. Also,
226 when asked to compare a file and a directory, imitate POSIX "diff"
227 and compare the file with the file with the same name in the
228 directory, instead of refusing to run.
229 (merge 0615173 jc/diff-no-index-d-f later to maint).
231 * "git rebase -i" moved the "current" command from "todo" to "done" a
232 bit too prematurely, losing a step when a "pick" did not even start.
233 (merge 8cbc57c ph/rebase-i-redo later to maint).
235 * The connection initiation code for "ssh" transport tried to absorb
236 differences between the stock "ssh" and Putty-supplied "plink" and
237 its derivatives, but the logic to tell that we are using "plink"
238 variants were too loose and falsely triggered when "plink" appeared
239 anywhere in the path (e.g. "/home/me/bin/uplink/ssh").
240 (merge baaf233 bc/connect-plink later to maint).
242 * "git stash pop/apply" forgot to make sure that not just the working
243 tree is clean but also the index is clean. The latter is important
244 as a stash application can conflict and the index will be used for
246 (merge ed178ef jk/stash-require-clean-index later to maint).
248 * We have prepended $GIT_EXEC_PATH and the path "git" is installed in
249 (typically "/usr/bin") to $PATH when invoking subprograms and hooks
250 for almost eternity, but the original use case the latter tried to
251 support was semi-bogus (i.e. install git to /opt/foo/git and run it
252 without having /opt/foo on $PATH), and more importantly it has
253 become less and less relevant as Git grew more mainstream (i.e. the
254 users would _want_ to have it on their $PATH). Stop prepending the
255 path in which "git" is installed to users' $PATH, as that would
256 interfere the command search order people depend on (e.g. they may
257 not like versions of programs that are unrelated to Git in /usr/bin
258 and want to override them by having different ones in /usr/local/bin
259 and have the latter directory earlier in their $PATH).
260 (merge a0b4507 jk/git-no-more-argv0-path-munging later to maint).
262 * core.excludesfile (defaulting to $XDG_HOME/git/ignore) is supposed
263 to be overridden by repository-specific .git/info/exclude file, but
264 the order was swapped from the beginning. This belatedly fixes it.
265 (merge 099d2d8 jc/gitignore-precedence later to maint).
267 * After "git add -N", the path appeared in output of "git diff HEAD"
268 and "git diff --cached HEAD", leading "git status" to classify it
269 as "Changes to be committed". Such a path, however, is not yet to
270 be scheduled to be committed. "git diff" showed the change to the
271 path as modification, not as a "new file", in the header of its
274 Treat such paths as "yet to be added to the index but Git already
275 know about them"; "git diff HEAD" and "git diff --cached HEAD"
276 should not talk about them, and "git diff" should show them as new
277 files yet to be added to the index.
278 (merge d95d728 nd/diff-i-t-a later to maint).
280 * There was a commented-out (instead of being marked to expect
281 failure) test that documented a breakage that was fixed since the
282 test was written; turn it into a proper test.
283 (merge 66d2e04 sb/t1020-cleanup later to maint).
285 * The "log --decorate" enhancement in Git 2.4 that shows the commit
286 at the tip of the current branch e.g. "HEAD -> master", did not
287 work with --decorate=full.
288 (merge 429ad20 mg/log-decorate-HEAD later to maint).
290 * The ref API did not handle cases where 'refs/heads/xyzzy/frotz' is
291 removed at the same time as 'refs/heads/xyzzy' is added (or vice
293 (merge c628edf mh/ref-directory-file later to maint).
295 * Multi-ref transaction support we merged a few releases ago
296 unnecessarily kept many file descriptors open, risking to fail with
297 resource exhaustion. This is for 2.4.x track.
298 (merge 185ce3a mh/write-refs-sooner-2.4 later to maint).
300 * "git bundle verify" did not diagnose extra parameters on the
302 (merge 7886cfa ps/bundle-verify-arg later to maint).
304 * Various documentation mark-up fixes to make the output more
305 consistent in general and also make AsciiDoctor (an alternative
307 (merge d0258b9 jk/asciidoc-markup-fix later to maint).
308 (merge ad3967a jk/stripspace-asciidoctor-fix later to maint).
309 (merge 975e382 ja/tutorial-asciidoctor-fix later to maint).
311 * Code cleanups and documentation updates.
312 (merge 0269f96 mm/usage-log-l-can-take-regex later to maint).
313 (merge 64f2589 nd/t1509-chroot-test later to maint).
314 (merge f86a374 sb/test-bitmap-free-at-end later to maint).
315 (merge 05bfc7d sb/line-log-plug-pairdiff-leak later to maint).
316 (merge 846e5df pt/xdg-config-path later to maint).
317 (merge 1154aa4 jc/plug-fmt-merge-msg-leak later to maint).
318 (merge 319b678 jk/sha1-file-reduce-useless-warnings later to maint).
319 (merge 9a35c14 fg/document-commit-message-stripping later to maint).
320 (merge bbf431c ps/doc-packfile-vs-pack-file later to maint).
321 (merge 309a9e3 jk/skip-http-tests-under-no-curl later to maint).
322 (merge ccd593c dl/branch-error-message later to maint).