4 Backward compatibility notes
5 ----------------------------
7 * The default value we give to the environment variable LESS has been
8 changed from "FRSX" to "FRX", losing "S" (chop long lines instead
9 of wrapping). Existing users who prefer not to see line-wrapped
10 output may want to set
12 $ git config core.pager "less -S"
14 to restore the traditional behaviour. It is expected that people
15 find output from the most subcommands easier to read with the new
16 default, except for "blame" which tends to produce really long
17 lines. To override the new default only for "git blame", you can
20 $ git config pager.blame "less -S"
22 * A few disused directories in contrib/ have been retired.
28 UI, Workflows & Features
30 * Since the very beginning of Git, we gave the LESS environment a
31 default value "FRSX" when we spawn "less" as the pager. "S" (chop
32 long lines instead of wrapping) has been removed from this default
33 set of options, because it is more or less a personal taste thing,
34 as opposed to others that have good justifications (i.e. "R" is
35 very much justified because many kinds of output we produce are
36 colored and "FX" is justified because output we produce is often
39 * The logic and data used to compute the display width needed for
40 UTF-8 strings have been updated to match Unicode 6.3 better.
42 * HTTP-based transports learned to propagate the error messages from
43 the webserver better to the client coming over the HTTP transport.
45 * The "core.preloadindex" configuration variable is by default
46 enabled, allowing modern platforms to take advantage of the
47 multiple cores they have.
49 * "git commit --date=<date>" option learned to read from more
50 timestamp formats, including "--date=now".
52 * The `core.commentChar` configuration variable is used to specify a
53 custom comment character other than the default "#" to be used in
54 the commit log editor. This can be set to `auto` to attempt to
55 choose a different character that does not conflict with what
56 already starts a line in the message being edited for cases like
59 * "git format-patch" learned --signature-file=<file> to take the mail
62 * "git grep" learned grep.fullname configuration variable to force
63 "--full-name" to be default. This may cause regressions on
64 scripted users that do not expect this new behaviour.
66 * "git imap-send" learned to ask the credential helper for auth
69 * "git log" and friends now understand the value "auto" set to the
70 "log.decorate" configuration variable to enable the "--decorate"
71 option automatically when the output is sent to tty.
73 * "git merge" without argument, even when there is an upstream
74 defined for the current branch, refused to run until
75 merge.defaultToUpstream is set to true. Flip the default of that
76 configuration variable to true.
78 * "git mergetool" learned to drive the vimdiff3 backend.
80 * mergetool.prompt used to default to 'true', always asking "do you
81 really want to run the tool on this path?". Among the two
82 purposes this prompt serves, ignore the use case to confirm that
83 the user wants to view particular path with the named tool, and
84 redefine the meaning of the prompt only to confirm the choice of
85 the tool made by the autodetection (for those who configured the
86 tool explicitly, the prompt shown for the latter purpose is
89 Strictly speaking, this is a backward incompatible change and the
90 users need to explicitly set the variable to 'true' if they want
91 to resurrect the now-ignored use case.
93 * "git replace" learned the "--edit" subcommand.
95 * "git svn" learned to cope with malformed timestamps with only one
96 digit in the hour part, e.g. 2014-01-07T5:01:02.048176Z, emitted
97 by some broken subversion server implementations.
99 * "git tag" when editing the tag message shows the name of the tag
100 being edited as a comment in the editor.
103 Performance, Internal Implementation, etc.
105 * Build procedure for 'subtree' (in contrib/) has been cleaned up.
107 * The `core.deltabasecachelimit` used to default to 16 MiB , but this
108 proved to be too small, and has been bumped to 96 MiB.
110 * "git blame" has been optimized greatly by reorganising the data
111 structure that is used to keep track of the work to be done.
113 * "git diff" that compares 3-or-more trees (e.g. parents and the
114 result of a merge) have been optimized.
116 * The API to update/delete references are being converted to handle
117 updates to multiple references in a transactional way. As an
118 example, "update-ref --stdin [-z]" has been updated to use this
121 * Parts of the test scripts can be skipped by using a range notation,
122 e.g. "sh t1234-test.sh --run='1-4 6 8-'" to omit test piece 5 and 7
123 and run everything else.
126 Also contains various documentation updates and code clean-ups.
132 Unless otherwise noted, all the fixes since v2.0 in the maintenance
133 track are contained in this release (see the maintenance releases'
136 * We used to unconditionally disable the pager in the pager process
137 we spawn to feed out output, but that prevented people who want to
138 run "less" within "less" from doing so.
139 (merge c0459ca je/pager-do-not-recurse later to maint).
141 * Tools that read diagnostic output in our standard error stream do
142 not want to see terminal control sequence (e.g. erase-to-eol).
143 Detect them by checking if the standard error stream is connected
145 (merge 38de156 mn/sideband-no-ansi later to maint).
147 * Mishandling of patterns in .gitignore that has trailing SPs quoted
148 with backslashes (e.g. ones that end with "\ ") have been
150 (merge e61a6c1 pb/trim-trailing-spaces later to maint).
152 * Reworded the error message given upon a failure to open an existing
153 loose object file due to e.g. permission issues; it was reported as
154 the object being corrupt, but that is not quite true.
155 (merge d6c8a05 jk/report-fail-to-read-objects-better later to maint).
157 * The "%<(10,trunc)%s" pretty format specifier in the log family of
158 commands is used to truncate the string to a given length (e.g. 10
159 in the example) with padding to column-align the output, but did
160 not take into account that number of bytes and number of display
161 columns are different.
162 (merge 7d50987 as/pretty-truncate later to maint).
164 * The "mailmap.file" configuration option did not support the tilde
165 expansion (i.e. ~user/path and ~/path).
166 (merge 9352fd5 ow/config-mailmap-pathname later to maint).
168 * The completion scripts (in contrib/) did not know about quite a few
169 options that are common between "git merge" and "git pull", and a
170 couple of options unique to "git merge".
171 (merge 8fee872 jk/complete-merge-pull later to maint).
173 * "--ignore-space-change" option of "git apply" ignored the spaces
174 at the beginning of line too aggressively, which is inconsistent
175 with the option of the same name "diff" and "git diff" have.
176 (merge 14d3bb4 jc/apply-ignore-whitespace later to maint).
178 * "git blame" miscounted number of columns needed to show localized
179 timestamps, resulting in jaggy left-side-edge of the source code
181 (merge dd75553 jx/blame-align-relative-time later to maint).
183 * "git blame" assigned the blame to the copy in the working-tree if
184 the repository is set to core.autocrlf=input and the file used CRLF
186 (merge 4d4813a bc/blame-crlf-test later to maint).
188 * "git commit --allow-empty-messag -C $commit" did not work when the
189 commit did not have any log message.
190 (merge 076cbd6 jk/commit-C-pick-empty later to maint).
192 * "git diff --find-copies-harder" sometimes pretended as if the mode
193 bits have changed for paths that are marked with assume-unchanged
195 (merge 5304810 jk/diff-files-assume-unchanged later to maint).
197 * "git format-patch" did not enforce the rule that the "--follow"
198 option from the log/diff family of commands must be used with
199 exactly one pathspec.
200 (merge dd63f16 jk/diff-follow-must-take-one-pathspec later to maint).
202 * "git gc --auto" was recently changed to run in the background to
203 give control back early to the end-user sitting in front of the
204 terminal, but it forgot that housekeeping involving reflogs should
205 be done without other processes competing for accesses to the refs.
206 (merge 62aad18 nd/daemonize-gc later to maint).
208 * "git grep -O" to show the lines that hit in the pager did not work
209 well with case insensitive search. We now spawn "less" with its
210 "-I" option when it is used as the pager (which is the default).
211 (merge f7febbe sk/spawn-less-case-insensitively-from-grep-O-i later to maint).
213 * We used to disable threaded "git index-pack" on platforms without
214 thread-safe pread(); use a different workaround for such
215 platforms to allow threaded "git index-pack".
216 (merge 3953949 nd/index-pack-one-fd-per-thread later to maint).
218 * The error reporting from "git index-pack" has been improved to
219 distinguish missing objects from type errors.
220 (merge 77583e7 jk/index-pack-report-missing later to maint).
222 * "git mailinfo" used to read beyond the end of header string while
223 parsing an incoming e-mail message to extract the patch.
224 (merge b1a013d rs/mailinfo-header-cmp later to maint).
226 * On a case insensitive filesystem, merge-recursive incorrectly
227 deleted the file that is to be renamed to a name that is the same
228 except for case differences.
229 (merge baa37bf dt/merge-recursive-case-insensitive later to maint).
231 * "git pack-objects" unnecessarily copied the previous contents when
232 extending the hashtable, even though it will populate the table
234 (merge fb79947 rs/pack-objects-no-unnecessary-realloc later to maint).
236 * "git rerere forget" did not work well when merge.conflictstyle
237 was set to a non-default value.
238 (merge de3d8bb fc/rerere-conflict-style later to maint).
240 * "git remote rm" and "git remote prune" can involve removing many
241 refs at once, which is not a very efficient thing to do when very
242 many refs exist in the packed-refs file.
243 (merge e6bea66 jl/remote-rm-prune later to maint).
245 * "git log --exclude=<glob> --all | git shortlog" worked as expected,
246 but "git shortlog --exclude=<glob> --all", which is supposed to be
247 identical to the above pipeline, was not accepted at the command
248 line argument parser level.
249 (merge eb07774 jc/shortlog-ref-exclude later to maint).
251 * The autostash mode of "git rebase -i" did not restore the dirty
252 working tree state if the user aborted the interactive rebase by
253 emptying the insn sheet.
254 (merge ddb5432 rr/rebase-autostash-fix later to maint).
256 * "git show -s" (i.e. show log message only) used to incorrectly emit
257 an extra blank line after a merge commit.
258 (merge ad2f725 mk/show-s-no-extra-blank-line-for-merges later to maint).
260 * "git status", even though it is a read-only operation, tries to
261 update the index with refreshed lstat(2) info to optimize future
262 accesses to the working tree opportunistically, but this could
263 race with a "read-write" operation that modify the index while it
264 is running. Detect such a race and avoid overwriting the index.
265 (merge 426ddee ym/fix-opportunistic-index-update-race later to maint).
267 * "git status" (and "git commit") behaved as if changes in a modified
268 submodule are not there if submodule.*.ignore configuration is set,
269 which was misleading. The configuration is only to unclutter diff
270 output during the course of development, and should not to hide
271 changes in the "status" output to cause the users forget to commit
273 (merge c215d3d jl/status-added-submodule-is-never-ignored later to maint).
275 * "git update-index --cacheinfo" in 2.0 release crashed on a
276 malformed command line.
277 (merge c8e1ee4 jc/rev-parse-argh-dashed-multi-words later to maint).
279 * The mode to run tests with HTTP server tests disabled was broken.
280 (merge afa53fe na/no-http-test-in-the-middle later to maint).