4 Backward compatibility notes
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 is now the "simple" semantics,
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
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.
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.
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.
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.
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").
42 "git request-pull" lost a few "heuristics" that often led to mistakes.
45 Updates since v1.9 series
46 -------------------------
48 UI, Workflows & Features
50 * "git commit --cleanup=<mode>" learned a new mode, scissors.
52 * "git tag --list" output can be sorted using "version sort" with
53 "--sort=version:refname".
55 * Discard the accumulated "heuristics" to guess from which branch the
56 result wants to be pulled from and make sure what the end user
57 specified is not second-guessed by "git request-pull", to avoid
58 mistakes. When you pushed out your 'master' branch to your public
59 repository as 'for-linus', use the new "master:for-linus" syntax to
60 denote the branch to be pulled.
62 * "git grep" learned to behave in a way similar to native grep when
63 "-h" (no header) and "-c" (count) options are given.
65 * transport-helper, fast-import and fast-export have been updated to
66 allow the ref mapping and ref deletion in a way similar to the
67 natively supported transports.
69 * The "simple" mode is the default for "git push".
71 * "git add -u" and "git add -A", when run without any pathspec, is a
72 tree-wide operation even when run inside a subdirectory of a
75 * "git add <path> is the same as "git add -A <path>" now.
77 * "core.statinfo" configuration variable, which is a
78 never-advertised synonym to "core.checkstat", has been removed.
80 * The "-q" option to "git diff-files", which does *NOT* mean
81 "quiet", has been removed (it told Git to ignore deletion, which
82 you can do with "git diff-files --diff-filter=d").
84 * Server operators can loosen the "tips of refs only" restriction for
85 the remote archive service with the uploadarchive.allowUnreachable
88 * The progress indicators from various time-consuming commands have
89 been marked for i18n/l10n.
91 * "git notes -C <blob>" diagnoses an attempt to use an object that
92 is not a blob as an error.
94 * "git config" learned to read from the standard input when "-" is
95 given as the value to its "--file" parameter (attempting an
96 operation to update the configuration in the standard input of
99 * Trailing whitespaces in .gitignore files, unless they are quoted
100 for fnmatch(3), e.g. "path\ ", are warned and ignored. Strictly
101 speaking, this is a backward incompatible change, but very unlikely
102 to bite any sane user and adjusting should be obvious and easy.
104 * Many commands that create commits, e.g. "pull", "rebase",
105 learned to take the --gpg-sign option on the command line.
107 * "git commit" can be told to always GPG sign the resulting commit
108 by setting "commit.gpgsign" configuration variable to true (the
109 command line option --no-gpg-sign should override it).
111 * "git pull" can be told to only accept fast-forward by setting the
112 new "pull.ff" configuration.
114 * "git reset" learned "-N" option, which does not reset the index
115 fully for paths the index knows about but the tree-ish the command
116 resets to does not (these paths are kept as intend-to-add entries).
118 * Newly cloned submodule repositories by "git submodule update",
119 when the "checkout" update mode is used, will be on a local
120 branch instead of on a detached HEAD, just like submodules added
121 with "git submodule add".
124 Performance, Internal Implementation, etc.
126 * We started using wildmatch() in place of fnmatch(3) a few releases
127 ago; complete the process and stop using fnmatch(3).
129 * Uses of curl's "multi" interface and "easy" interface do not mix
130 well when we attempt to reuse outgoing connections. Teach the RPC
131 over http code, used in the smart HTTP transport, not to use the
134 * The bitmap-index feature from JGit has been ported, which should
135 significantly improve performance when serving objects form a
136 repository that uses it.
138 * The way "git log --cc" shows a combined diff against multiple
139 parents have been optimized.
141 * The prefixcmp() and suffixcmp() functions are gone. Use
142 starts_with() and ends_with(), and also consider if skip_prefix()
143 suits your needs better when using the former.
146 Also contains various documentation updates and code clean-ups. Many
147 of them came from flurry of activities as GSoC candidate microproject
151 Fixes since v1.9 series
152 -----------------------
154 Unless otherwise noted, all the fixes since v1.9 in the maintenance
155 track are contained in this release (see the maintenance releases'
158 * "git mv" that moves a submodule forgot to adjust the array that
159 uses to keep track of which submodules were to be moved to update
161 (merge fb8a4e8 jk/mv-submodules-fix later to maint).
163 * Length limit for the pathname used when removing a path in a deep
164 subdirectory has been removed to avoid buffer overflows.
165 (merge 2f29e0c mh/remove-subtree-long-pathname-fix later to maint).
167 * The test helper lib-terminal always run an actual test_expect_*
168 when included, which screwed up with the use of skil-all that may
169 have to be done later.
170 (merge 7e27173 jk/lib-terminal-lazy later to maint).
172 * "git index-pack" used a wrong variable to name the keep-file in an
173 error message when the file cannot be written or closed.
174 (merge de983a0 nd/index-pack-error-message later to maint).
176 * "rebase -i" produced a broken insn sheet when the title of a commit
177 happened to contain '\n' (or ended with '\c') due to a careless use
179 (merge cb1aefd us/printf-not-echo later to maint).
181 * There were a few instances of 'git-foo' remaining in the
182 documentation that should have been spelled 'git foo'.
183 (merge 3c3e6f5 rr/doc-merge-strategies later to maint).
185 * Serving objects from a shallow repository needs to write a
186 new file to hold the temporary shallow boundaries but it was not
187 cleaned when we exit due to die() or a signal.
188 (merge 7839632 jk/shallow-update-fix later to maint).
190 * When "git stash pop" stops after failing to apply the stash
191 (e.g. due to conflicting changes), the stash is not dropped. State
192 that explicitly in the output to let the users know.
193 (merge 2d4c993 jc/stash-pop-not-popped later to maint).
195 * The labels in "git status" output that describe the nature of
196 conflicts (e.g. "both deleted") were limited to 20 bytes, which was
197 too short for some l10n (e.g. fr).
198 (merge c7cb333 jn/wt-status later to maint).
200 * "git clean -d pathspec" did not use the given pathspec correctly
201 and ended up cleaning too much.
202 (merge 1f2e108 jk/clean-d-pathspec later to maint).
204 * "git difftool" misbehaved when the repository is bound to the
205 working tree with the ".git file" mechanism, where a textual file
206 ".git" tells us where it is.
207 (merge fcfec8b da/difftool-git-files later to maint).
209 * "git push" did not pay attention to branch.*.pushremote if it is
210 defined earlier than remote.pushdefault; the order of these two
211 variables in the configuration file should not matter, but it did
213 (merge 98b406f jk/remote-pushremote-config-reading later to maint).
215 * Codepaths that parse timestamps in commit objects have been
217 (merge 3f419d4 jk/commit-dates-parsing-fix later to maint).
219 * "git diff --external-diff" incorrectly fed the submodule directory
220 in the working tree to the external diff driver when it knew it is
221 the same as one of the versions being compared.
222 (merge aba4727 tr/diff-submodule-no-reuse-worktree later to maint).
224 * "git reset" needs to refresh the index when working in a working
225 tree (it can also be used to match the index to the HEAD in an
226 otherwise bare repository), but it failed to set up the working
227 tree properly, causing GIT_WORK_TREE to be ignored.
228 (merge b7756d4 nd/reset-setup-worktree later to maint).
230 * "git check-attr" when working on a repository with a working tree
231 did not work well when the working tree was specified via the
232 --work-tree (and obviously with --git-dir) option.
233 (merge cdbf623 jc/check-attr-honor-working-tree later to maint).
235 * "merge-recursive" was broken in 1.7.7 era and stopped working in
236 an empty (temporary) working tree, when there are renames
237 involved. This has been corrected.
238 (merge 6e2068a bk/refresh-missing-ok-in-merge-recursive later to maint.)
240 * "git rev-parse" was loose in rejecting command line arguments
241 that do not make sense, e.g. "--default" without the required
242 value for that option.
243 (merge a43219f ds/rev-parse-required-args later to maint.)
245 * include.path variable (or any variable that expects a path that
246 can use ~username expansion) in the configuration file is not a
247 boolean, but the code failed to check it.
248 (merge 67beb60 jk/config-path-include-fix later to maint.)
250 * Commands that take pathspecs on the command line misbehaved when
251 the pathspec is given as an absolute pathname (which is a
252 practice not particularly encouraged) that points at a symbolic
253 link in the working tree.
254 (merge later 655ee9e mw/symlinks to maint.)
256 * "git diff --quiet -- pathspec1 pathspec2" sometimes did not return
257 correct status value.
258 (merge f34b205 nd/diff-quiet-stat-dirty later to maint.)
260 * Attempting to deepen a shallow repository by fetching over smart
261 HTTP transport failed in the protocol exchange, when no-done
262 extension was used. The fetching side waited for the list of
263 shallow boundary commits after the sending end stopped talking to
265 (merge 0232852 nd/http-fetch-shallow-fix later to maint.)
267 * Allow "git cmd path/", when the 'path' is where a submodule is
268 bound to the top-level working tree, to match 'path', despite the
269 extra and unnecessary trailing slash (such a slash is often
270 given by command line completion).
271 (merge 2e70c01 nd/submodule-pathspec-ending-with-slash later to maint.)