compat/bswap.h: fix endianness detection
[git/git.git] / Documentation / RelNotes / 2.0.0.txt
1 Git v2.0 Release Notes
2 ======================
3
4 Backward compatibility notes
5 ----------------------------
6
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 is now the "simple" semantics,
11 which pushes:
12
13 - only the current branch to the branch with the same name, and only
14 when the current branch is set to integrate with that remote
15 branch, if you are pushing to the same remote as you fetch from; or
16
17 - only the current branch to the branch with the same name, if you
18 are pushing to a remote that is not where you usually fetch from.
19
20 You can use the configuration variable "push.default" to change
21 this. If you are an old-timer who wants to keep using the
22 "matching" semantics, you can set the variable to "matching", for
23 example. Read the documentation for other possibilities.
24
25 When "git add -u" and "git add -A" are run inside a subdirectory
26 without specifying which paths to add on the command line, they
27 operate on the entire tree for consistency with "git commit -a" and
28 other commands (these commands used to operate only on the current
29 subdirectory). Say "git add -u ." or "git add -A ." if you want to
30 limit the operation to the current directory.
31
32 "git add <path>" is the same as "git add -A <path>" now, so that
33 "git add dir/" will notice paths you removed from the directory and
34 record the removal. In older versions of Git, "git add <path>" used
35 to ignore removals. You can say "git add --ignore-removal <path>" to
36 add only added or modified paths in <path>, if you really want to.
37
38 The "-q" option to "git diff-files", which does *NOT* mean "quiet",
39 has been removed (it told Git to ignore deletion, which you can do
40 with "git diff-files --diff-filter=d").
41
42 "git request-pull" lost a few "heuristics" that often led to mistakes.
43
44 The default prefix for "git svn" has changed in Git 2.0. For a long
45 time, "git svn" created its remote-tracking branches directly under
46 refs/remotes, but it now places them under refs/remotes/origin/ unless
47 it is told otherwise with its --prefix option.
48
49
50 Updates since v1.9 series
51 -------------------------
52
53 UI, Workflows & Features
54
55 * The "multi-mail" post-receive hook (in contrib/) has been updated
56 to a more recent version from the upstream.
57
58 * "git gc --aggressive" learned "--depth" option and
59 "gc.aggressiveDepth" configuration variable to allow use of a less
60 insane depth than the built-in default value of 250.
61
62 * "git log" learned the "--show-linear-break" option to show where a
63 single strand-of-pearls is broken in its output.
64
65 * The "rev-parse --parseopt" mechanism used by scripted Porcelains to
66 parse command line options and to give help text learned to take
67 the argv-help (the placeholder string for an option parameter,
68 e.g. "key-id" in "--gpg-sign=<key-id>").
69
70 * The pattern to find where the function begins in C/C++ used in
71 "diff" and "grep -p" have been updated to help C++ source better.
72
73 * "git rebase" learned to interpret a lone "-" as "@{-1}", the
74 branch that we were previously on.
75
76 * "git commit --cleanup=<mode>" learned a new mode, scissors.
77
78 * "git tag --list" output can be sorted using "version sort" with
79 "--sort=version:refname".
80
81 * Discard the accumulated "heuristics" to guess from which branch the
82 result wants to be pulled from and make sure what the end user
83 specified is not second-guessed by "git request-pull", to avoid
84 mistakes. When you pushed out your 'master' branch to your public
85 repository as 'for-linus', use the new "master:for-linus" syntax to
86 denote the branch to be pulled.
87
88 * "git grep" learned to behave in a way similar to native grep when
89 "-h" (no header) and "-c" (count) options are given.
90
91 * "git push" via transport-helper interface (e.g. remote-hg) has
92 been updated to allow ref deletion in a way similar to the natively
93 supported transports.
94
95 * The "simple" mode is the default for "git push".
96
97 * "git add -u" and "git add -A", when run without any pathspec, is a
98 tree-wide operation even when run inside a subdirectory of a
99 working tree.
100
101 * "git add <path> is the same as "git add -A <path>" now.
102
103 * "core.statinfo" configuration variable, which is a
104 never-advertised synonym to "core.checkstat", has been removed.
105
106 * The "-q" option to "git diff-files", which does *NOT* mean
107 "quiet", has been removed (it told Git to ignore deletion, which
108 you can do with "git diff-files --diff-filter=d").
109
110 * Server operators can loosen the "tips of refs only" restriction for
111 the remote archive service with the uploadarchive.allowUnreachable
112 configuration option.
113
114 * The progress indicators from various time-consuming commands have
115 been marked for i18n/l10n.
116
117 * "git notes -C <blob>" diagnoses an attempt to use an object that
118 is not a blob as an error.
119
120 * "git config" learned to read from the standard input when "-" is
121 given as the value to its "--file" parameter (attempting an
122 operation to update the configuration in the standard input of
123 course is rejected).
124
125 * Trailing whitespaces in .gitignore files, unless they are quoted
126 for fnmatch(3), e.g. "path\ ", are warned and ignored. Strictly
127 speaking, this is a backward incompatible change, but very unlikely
128 to bite any sane user and adjusting should be obvious and easy.
129
130 * Many commands that create commits, e.g. "pull", "rebase",
131 learned to take the --gpg-sign option on the command line.
132
133 * "git commit" can be told to always GPG sign the resulting commit
134 by setting "commit.gpgsign" configuration variable to true (the
135 command line option --no-gpg-sign should override it).
136
137 * "git pull" can be told to only accept fast-forward by setting the
138 new "pull.ff" configuration.
139
140 * "git reset" learned "-N" option, which does not reset the index
141 fully for paths the index knows about but the tree-ish the command
142 resets to does not (these paths are kept as intend-to-add entries).
143
144
145 Performance, Internal Implementation, etc.
146
147 * The compilation options to port to AIX and to MSVC have been
148 updated.
149
150 * We started using wildmatch() in place of fnmatch(3) a few releases
151 ago; complete the process and stop using fnmatch(3).
152
153 * Uses of curl's "multi" interface and "easy" interface do not mix
154 well when we attempt to reuse outgoing connections. Teach the RPC
155 over http code, used in the smart HTTP transport, not to use the
156 "easy" interface.
157
158 * The bitmap-index feature from JGit has been ported, which should
159 significantly improve performance when serving objects form a
160 repository that uses it.
161
162 * The way "git log --cc" shows a combined diff against multiple
163 parents have been optimized.
164
165 * The prefixcmp() and suffixcmp() functions are gone. Use
166 starts_with() and ends_with(), and also consider if skip_prefix()
167 suits your needs better when using the former.
168
169
170 Also contains various documentation updates and code clean-ups. Many
171 of them came from flurry of activities as GSoC candidate microproject
172 exercises.
173
174
175 Fixes since v1.9 series
176 -----------------------
177
178 Unless otherwise noted, all the fixes since v1.9 in the maintenance
179 track are contained in this release (see the maintenance releases'
180 notes for details).
181
182 * The remote-helepr interface to fast-import/fast-export via the
183 transport-helper has been tightened to avoid leaving the import
184 marks file from a failed/crashed run, as such a file that is out of
185 sync with the reality confuses a later invocation of itself.
186
187 * "git rebase" used a POSIX shell construct FreeBSD /bin/sh does not
188 work well with.
189 (merge 8cd6596 km/avoid-non-function-return-in-rebase later to maint).
190
191 * zsh prompt (in contrib/) leaked unnecessary error messages.
192
193 * bash completion (in contrib/) did not complete the refs and remotes
194 correctly given "git pu<TAB>" when "pu" is aliased to "push".
195
196 * Some more Unicode codepoints defined in Unicode 6.3 as having zero
197 width have been taught to our display column counting logic.
198 (merge d813ab9 tb/unicode-6.3-zero-width later to maint).
199
200 * Some tests used shell constructs that did not work well on FreeBSD
201 (merge ff7a1c6 km/avoid-bs-in-shell-glob later to maint).
202 (merge 00764ca km/avoid-cp-a later to maint).
203
204 * "git update-ref --stdin" did not fail a request to create a ref
205 when the ref already existed.
206 (merge b9d56b5 mh/update-ref-batch-create-fix later to maint).
207
208 * "git diff --no-index -Mq a b" fell into an infinite loop.
209 (merge ad1c3fb jc/fix-diff-no-index-diff-opt-parse later to maint).
210
211 * "git fetch --prune", when the right-hand-side of multiple fetch
212 refspecs overlap (e.g. storing "refs/heads/*" to
213 "refs/remotes/origin/*", while storing "refs/frotz/*" to
214 "refs/remotes/origin/fr/*"), aggressively thought that lack of
215 "refs/heads/fr/otz" on the origin site meant we should remove
216 "refs/remotes/origin/fr/otz" from us, without checking their
217 "refs/frotz/otz" first.
218
219 Note that such a configuration is inherently unsafe (think what
220 should happen when "refs/heads/fr/otz" does appear on the origin
221 site), but that is not a reason not to be extra careful.
222 (merge e6f6371 cn/fetch-prune-overlapping-destination later to maint).
223
224 * "git status --porcelain --branch" showed its output with labels
225 "ahead/behind/gone" translated to the user's locale.
226 (merge 7a76c28 mm/status-porcelain-format-i18n-fix later to maint).
227
228 * A stray environment variable $prefix could have leaked into and
229 affected the behaviour of the "subtree" script (in contrib/).
230
231 * When it is not necessary to edit a commit log message (e.g. "git
232 commit -m" is given a message without specifying "-e"), we used to
233 disable the spawning of the editor by overriding GIT_EDITOR, but
234 this means all the uses of the editor, other than to edit the
235 commit log message, are also affected.
236 (merge b549be0 bp/commit-p-editor later to maint).
237
238 * "git mv" that moves a submodule forgot to adjust the array that
239 uses to keep track of which submodules were to be moved to update
240 its configuration.
241 (merge fb8a4e8 jk/mv-submodules-fix later to maint).
242
243 * Length limit for the pathname used when removing a path in a deep
244 subdirectory has been removed to avoid buffer overflows.
245 (merge 2f29e0c mh/remove-subtree-long-pathname-fix later to maint).
246
247 * The test helper lib-terminal always run an actual test_expect_*
248 when included, which screwed up with the use of skil-all that may
249 have to be done later.
250 (merge 7e27173 jk/lib-terminal-lazy later to maint).
251
252 * "git index-pack" used a wrong variable to name the keep-file in an
253 error message when the file cannot be written or closed.
254 (merge de983a0 nd/index-pack-error-message later to maint).
255
256 * "rebase -i" produced a broken insn sheet when the title of a commit
257 happened to contain '\n' (or ended with '\c') due to a careless use
258 of 'echo'.
259 (merge cb1aefd us/printf-not-echo later to maint).
260
261 * There were a few instances of 'git-foo' remaining in the
262 documentation that should have been spelled 'git foo'.
263 (merge 3c3e6f5 rr/doc-merge-strategies later to maint).
264
265 * Serving objects from a shallow repository needs to write a
266 new file to hold the temporary shallow boundaries but it was not
267 cleaned when we exit due to die() or a signal.
268 (merge 7839632 jk/shallow-update-fix later to maint).
269
270 * When "git stash pop" stops after failing to apply the stash
271 (e.g. due to conflicting changes), the stash is not dropped. State
272 that explicitly in the output to let the users know.
273 (merge 2d4c993 jc/stash-pop-not-popped later to maint).
274
275 * The labels in "git status" output that describe the nature of
276 conflicts (e.g. "both deleted") were limited to 20 bytes, which was
277 too short for some l10n (e.g. fr).
278 (merge c7cb333 jn/wt-status later to maint).
279
280 * "git clean -d pathspec" did not use the given pathspec correctly
281 and ended up cleaning too much.
282 (merge 1f2e108 jk/clean-d-pathspec later to maint).
283
284 * "git difftool" misbehaved when the repository is bound to the
285 working tree with the ".git file" mechanism, where a textual file
286 ".git" tells us where it is.
287 (merge fcfec8b da/difftool-git-files later to maint).
288
289 * "git push" did not pay attention to branch.*.pushremote if it is
290 defined earlier than remote.pushdefault; the order of these two
291 variables in the configuration file should not matter, but it did
292 by mistake.
293 (merge 98b406f jk/remote-pushremote-config-reading later to maint).
294
295 * Codepaths that parse timestamps in commit objects have been
296 tightened.
297 (merge f80d1f9 jk/commit-dates-parsing-fix later to maint).
298
299 * "git diff --external-diff" incorrectly fed the submodule directory
300 in the working tree to the external diff driver when it knew it is
301 the same as one of the versions being compared.
302 (merge aba4727 tr/diff-submodule-no-reuse-worktree later to maint).
303
304 * "git reset" needs to refresh the index when working in a working
305 tree (it can also be used to match the index to the HEAD in an
306 otherwise bare repository), but it failed to set up the working
307 tree properly, causing GIT_WORK_TREE to be ignored.
308 (merge b7756d4 nd/reset-setup-worktree later to maint).
309
310 * "git check-attr" when working on a repository with a working tree
311 did not work well when the working tree was specified via the
312 --work-tree (and obviously with --git-dir) option.
313 (merge cdbf623 jc/check-attr-honor-working-tree later to maint).
314
315 * "merge-recursive" was broken in 1.7.7 era and stopped working in
316 an empty (temporary) working tree, when there are renames
317 involved. This has been corrected.
318 (merge 6e2068a bk/refresh-missing-ok-in-merge-recursive later to maint.)
319
320 * "git rev-parse" was loose in rejecting command line arguments
321 that do not make sense, e.g. "--default" without the required
322 value for that option.
323 (merge a43219f ds/rev-parse-required-args later to maint.)
324
325 * include.path variable (or any variable that expects a path that
326 can use ~username expansion) in the configuration file is not a
327 boolean, but the code failed to check it.
328 (merge 67beb60 jk/config-path-include-fix later to maint.)
329
330 * Commands that take pathspecs on the command line misbehaved when
331 the pathspec is given as an absolute pathname (which is a
332 practice not particularly encouraged) that points at a symbolic
333 link in the working tree.
334 (merge later 655ee9e mw/symlinks to maint.)
335
336 * "git diff --quiet -- pathspec1 pathspec2" sometimes did not return
337 correct status value.
338 (merge f34b205 nd/diff-quiet-stat-dirty later to maint.)
339
340 * Attempting to deepen a shallow repository by fetching over smart
341 HTTP transport failed in the protocol exchange, when no-done
342 extension was used. The fetching side waited for the list of
343 shallow boundary commits after the sending end stopped talking to
344 it.
345 (merge 0232852 nd/http-fetch-shallow-fix later to maint.)
346
347 * Allow "git cmd path/", when the 'path' is where a submodule is
348 bound to the top-level working tree, to match 'path', despite the
349 extra and unnecessary trailing slash (such a slash is often
350 given by command line completion).
351 (merge 2e70c01 nd/submodule-pathspec-ending-with-slash later to maint.)
352
353 * Documentation and in-code comments had many instances of mistaken
354 use of "nor", which have been corrected.
355 (merge 235e8d5 jl/nor-or-nand-and later to maint).