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 7.0 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 completion script for bash (in contrib/) has been updated to
46 handle aliases that define complex sequence of commands better.
48 * The "core.preloadindex" configuration variable is by default
49 enabled, allowing modern platforms to take advantage of the
50 multiple cores they have.
52 * "git clone" applies the "if cloning from a local disk, physically
53 copy repository using hardlinks, unless otherwise told not to with
54 --no-local" optimization when url.*.insteadOf mechanism rewrites a
55 "git clone $URL" that refers to a repository over the network to a
56 clone from a local disk.
58 * "git commit --date=<date>" option learned to read from more
59 timestamp formats, including "--date=now".
61 * The `core.commentChar` configuration variable is used to specify a
62 custom comment character other than the default "#" to be used in
63 the commit log editor. This can be set to `auto` to attempt to
64 choose a different character that does not conflict with what
65 already starts a line in the message being edited for cases like
68 * "git format-patch" learned --signature-file=<file> to take the mail
71 * "git grep" learned grep.fullname configuration variable to force
72 "--full-name" to be default. This may cause regressions on
73 scripted users that do not expect this new behaviour.
75 * "git imap-send" learned to ask the credential helper for auth
78 * "git log" and friends now understand the value "auto" set to the
79 "log.decorate" configuration variable to enable the "--decorate"
80 option automatically when the output is sent to tty.
82 * "git merge" without argument, even when there is an upstream
83 defined for the current branch, refused to run until
84 merge.defaultToUpstream is set to true. Flip the default of that
85 configuration variable to true.
87 * "git mergetool" learned to drive the vimdiff3 backend.
89 * mergetool.prompt used to default to 'true', always asking "do you
90 really want to run the tool on this path?". Among the two
91 purposes this prompt serves, ignore the use case to confirm that
92 the user wants to view particular path with the named tool, and
93 redefine the meaning of the prompt only to confirm the choice of
94 the tool made by the autodetection (for those who configured the
95 tool explicitly, the prompt shown for the latter purpose is
98 Strictly speaking, this is a backward incompatible change and the
99 users need to explicitly set the variable to 'true' if they want
100 to resurrect the now-ignored use case.
102 * "git replace" learned the "--edit" subcommand to create a
103 replacement by editing an existing object.
105 * "git replace" learned a "--graft" option to rewrite parents of a
108 * "git send-email" learned "--to-cover" and "--cc-cover" options, to
109 tell it to copy To: and Cc: headers found in the first input file
110 when emitting later input files.
112 * "git svn" learned to cope with malformed timestamps with only one
113 digit in the hour part, e.g. 2014-01-07T5:01:02.048176Z, emitted
114 by some broken subversion server implementations.
116 * "git tag" when editing the tag message shows the name of the tag
117 being edited as a comment in the editor.
119 * "git tag" learned to pay attention to "tag.sort" configuration, to
120 be used as the default sort order when no --sort=<value> the option
123 * "git verify-commit" command to check GPG signature in signed
124 commits, in a way similar to "git verify-tag" is used to check
125 signed tags, was added.
128 Performance, Internal Implementation, etc.
130 * Build procedure for 'subtree' (in contrib/) has been cleaned up.
132 * The support for the profile-feedback build, which has been left
133 bit-rotten for quite a while, has been updated.
135 * An experimental format to use two files (the base file and
136 incremental changes relative to it) to represent the index has been
137 introduced; this may reduce I/O cost of rewriting a large index
138 when only small part of the working tree changes.
140 * Effort to shrink the size of patches Windows folks maintain on top
141 by upstreaming them continues.
143 * Patches maintained by msysgit folks for Windows port are being
144 upstreamed here a bit by bit.
146 * The leaf function to check validity of a refname format has been
147 micro-optimized, using SSE2 instructions when available. A few
148 breakages during its development have been caught and fixed already
149 but there might remain some more still; please test and report if
152 * The `core.deltabasecachelimit` used to default to 16 MiB , but this
153 proved to be too small, and has been bumped to 96 MiB.
155 * "git blame" has been optimized greatly by reorganising the data
156 structure that is used to keep track of the work to be done.
158 * "git diff" that compares 3-or-more trees (e.g. parents and the
159 result of a merge) have been optimized.
161 * The API to update/delete references are being converted to handle
162 updates to multiple references in a transactional way. As an
163 example, "update-ref --stdin [-z]" has been updated to use this
166 * skip_prefix() and strip_suffix() API functions are used a lot more
167 widely throughout the codebase now.
169 * Parts of the test scripts can be skipped by using a range notation,
170 e.g. "sh t1234-test.sh --run='1-4 6 8-'" to omit test piece 5 and 7
171 and run everything else.
174 Also contains various documentation updates and code clean-ups.
180 Unless otherwise noted, all the fixes since v2.0 in the maintenance
181 track are contained in this release (see the maintenance releases'
184 * We used to unconditionally disable the pager in the pager process
185 we spawn to feed out output, but that prevented people who want to
186 run "less" within "less" from doing so.
187 (merge c0459ca je/pager-do-not-recurse later to maint).
189 * Tools that read diagnostic output in our standard error stream do
190 not want to see terminal control sequence (e.g. erase-to-eol).
191 Detect them by checking if the standard error stream is connected
193 (merge 38de156 mn/sideband-no-ansi later to maint).
195 * Mishandling of patterns in .gitignore that has trailing SPs quoted
196 with backslashes (e.g. ones that end with "\ ") have been
198 (merge 97c1364be6b pb/trim-trailing-spaces later to maint).
200 * Reworded the error message given upon a failure to open an existing
201 loose object file due to e.g. permission issues; it was reported as
202 the object being corrupt, but that is not quite true.
203 (merge d6c8a05 jk/report-fail-to-read-objects-better later to maint).
205 * "git log -2master" is a common typo that shows two commits starting
206 from whichever random branch that is not 'master' that happens to
207 be checked out currently.
208 (merge e3fa568 jc/revision-dash-count-parsing later to maint).
210 * Code to avoid adding the same alternate object store twice was
211 subtly broken for a long time, but nobody seems to have noticed.
212 (merge 80b4785 rs/fix-alt-odb-path-comparison later to maint).
213 (merge 539e750 ek/alt-odb-entry-fix later to maint).
215 * The "%<(10,trunc)%s" pretty format specifier in the log family of
216 commands is used to truncate the string to a given length (e.g. 10
217 in the example) with padding to column-align the output, but did
218 not take into account that number of bytes and number of display
219 columns are different.
220 (merge 7d50987 as/pretty-truncate later to maint).
222 * "%G" (nothing after G) is an invalid pretty format specifier, but
223 the parser did not notice it as garbage.
224 (merge 958b2eb jk/pretty-G-format-fixes later to maint).
226 * A handful of code paths had to read the commit object more than
227 once when showing header fields that are usually not parsed. The
228 internal data structure to keep track of the contents of the commit
229 object has been updated to reduce the need for this double-reading,
230 and to allow the caller find the length of the object.
231 (merge 218aa3a jk/commit-buffer-length later to maint).
233 * The "mailmap.file" configuration option did not support the tilde
234 expansion (i.e. ~user/path and ~/path).
235 (merge 9352fd5 ow/config-mailmap-pathname later to maint).
237 * The completion scripts (in contrib/) did not know about quite a few
238 options that are common between "git merge" and "git pull", and a
239 couple of options unique to "git merge".
240 (merge 8fee872 jk/complete-merge-pull later to maint).
242 * The unix-domain socket used by the sample credential cache daemon
243 tried to unlink an existing stale one at a wrong path, if the path
244 to the socket was given as an overlong path that does not fit in
245 sun_path member of the sockaddr_un structure.
246 (merge 2869b3e rs/fix-unlink-unix-socket later to maint).
248 * An ancient rewrite passed a wrong pointer to a curl library
249 function in a rarely used code path.
250 (merge 479eaa8 ah/fix-http-push later to maint).
252 * "--ignore-space-change" option of "git apply" ignored the spaces
253 at the beginning of line too aggressively, which is inconsistent
254 with the option of the same name "diff" and "git diff" have.
255 (merge 14d3bb4 jc/apply-ignore-whitespace later to maint).
257 * "git blame" miscounted number of columns needed to show localized
258 timestamps, resulting in jaggy left-side-edge of the source code
260 (merge dd75553 jx/blame-align-relative-time later to maint).
262 * "git blame" assigned the blame to the copy in the working-tree if
263 the repository is set to core.autocrlf=input and the file used CRLF
265 (merge 4d4813a bc/blame-crlf-test later to maint).
267 * "git clone -b brefs/tags/bar" would have mistakenly thought we were
268 following a single tag, even though it was a name of the branch,
269 because it incorrectly used strstr().
270 (merge 60a5f5f jc/fix-clone-single-starting-at-a-tag later to maint).
272 * "git commit --allow-empty-messag -C $commit" did not work when the
273 commit did not have any log message.
274 (merge 076cbd6 jk/commit-C-pick-empty later to maint).
276 * "git diff --find-copies-harder" sometimes pretended as if the mode
277 bits have changed for paths that are marked with assume-unchanged
279 (merge 5304810 jk/diff-files-assume-unchanged later to maint).
281 * "filter-branch" left an empty single-parent commit that results when
282 all parents of a merge commit gets mapped to the same commit, even
283 under "--prune-empty".
284 (merge 79bc4ef cb/filter-branch-prune-empty-degenerate-merges later to maint).
286 * "git format-patch" did not enforce the rule that the "--follow"
287 option from the log/diff family of commands must be used with
288 exactly one pathspec.
289 (merge dd63f16 jk/diff-follow-must-take-one-pathspec later to maint).
291 * "git gc --auto" was recently changed to run in the background to
292 give control back early to the end-user sitting in front of the
293 terminal, but it forgot that housekeeping involving reflogs should
294 be done without other processes competing for accesses to the refs.
295 (merge 62aad18 nd/daemonize-gc later to maint).
297 * "git grep -O" to show the lines that hit in the pager did not work
298 well with case insensitive search. We now spawn "less" with its
299 "-I" option when it is used as the pager (which is the default).
300 (merge f7febbe sk/spawn-less-case-insensitively-from-grep-O-i later to maint).
302 * We used to disable threaded "git index-pack" on platforms without
303 thread-safe pread(); use a different workaround for such
304 platforms to allow threaded "git index-pack".
305 (merge 3953949 nd/index-pack-one-fd-per-thread later to maint).
307 * The error reporting from "git index-pack" has been improved to
308 distinguish missing objects from type errors.
309 (merge 77583e7 jk/index-pack-report-missing later to maint).
311 * "log --show-signature" incorrectly decided the color to paint a
312 mergetag that was and was not correctly validated.
313 (merge 42c55ce mg/fix-log-mergetag-color later to maint).
315 * "log --show-signature" did not pay attention to "--graph" option.
316 (merge cf3983d zk/log-graph-showsig later to maint).
318 * "git mailinfo" used to read beyond the end of header string while
319 parsing an incoming e-mail message to extract the patch.
320 (merge b1a013d rs/mailinfo-header-cmp later to maint).
322 * On a case insensitive filesystem, merge-recursive incorrectly
323 deleted the file that is to be renamed to a name that is the same
324 except for case differences.
325 (merge baa37bf dt/merge-recursive-case-insensitive later to maint).
327 * Merging changes into a file that ends in an incomplete line made the
328 last line into a complete one, even when the other branch did not
329 change anything around the end of file.
330 (merge ba31180 mk/merge-incomplete-files later to maint).
332 * "git pack-objects" unnecessarily copied the previous contents when
333 extending the hashtable, even though it will populate the table
335 (merge fb79947 rs/pack-objects-no-unnecessary-realloc later to maint).
337 * Recent updates to "git repack" started to duplicate objects that
338 are in packfiles marked with .keep flag into the new packfile by
340 (merge d078d85 jk/repack-pack-keep-objects later to maint).
342 * "git rerere forget" did not work well when merge.conflictstyle
343 was set to a non-default value.
344 (merge de3d8bb fc/rerere-conflict-style later to maint).
346 * "git remote rm" and "git remote prune" can involve removing many
347 refs at once, which is not a very efficient thing to do when very
348 many refs exist in the packed-refs file.
349 (merge e6bea66 jl/remote-rm-prune later to maint).
351 * "git log --exclude=<glob> --all | git shortlog" worked as expected,
352 but "git shortlog --exclude=<glob> --all", which is supposed to be
353 identical to the above pipeline, was not accepted at the command
354 line argument parser level.
355 (merge eb07774 jc/shortlog-ref-exclude later to maint).
357 * The autostash mode of "git rebase -i" did not restore the dirty
358 working tree state if the user aborted the interactive rebase by
359 emptying the insn sheet.
360 (merge ddb5432 rr/rebase-autostash-fix later to maint).
362 * "git rebase --fork-point" did not filter out patch-identical
365 * During "git rebase --merge", a conflicted patch could not be
366 skipped with "--skip" if the next one also conflicted.
367 (merge 95104c7 bc/fix-rebase-merge-skip later to maint).
369 * "git show -s" (i.e. show log message only) used to incorrectly emit
370 an extra blank line after a merge commit.
371 (merge ad2f725 mk/show-s-no-extra-blank-line-for-merges later to maint).
373 * "git status", even though it is a read-only operation, tries to
374 update the index with refreshed lstat(2) info to optimize future
375 accesses to the working tree opportunistically, but this could
376 race with a "read-write" operation that modify the index while it
377 is running. Detect such a race and avoid overwriting the index.
378 (merge 426ddee ym/fix-opportunistic-index-update-race later to maint).
380 * "git status" (and "git commit") behaved as if changes in a modified
381 submodule are not there if submodule.*.ignore configuration is set,
382 which was misleading. The configuration is only to unclutter diff
383 output during the course of development, and should not to hide
384 changes in the "status" output to cause the users forget to commit
386 (merge c215d3d jl/status-added-submodule-is-never-ignored later to maint).
388 * Documentation for "git submodule sync" forgot to say that the subcommand
389 can take the "--recursive" option.
390 (merge 9393ae7 mc/doc-submodule-sync-recurse later to maint).
392 * "git update-index --cacheinfo" in 2.0 release crashed on a
393 malformed command line.
394 (merge c8e1ee4 jc/rev-parse-argh-dashed-multi-words later to maint).
396 * The mode to run tests with HTTP server tests disabled was broken.
397 (merge afa53fe na/no-http-test-in-the-middle later to maint).
projects / git / blob