1Git 2.10 Release Notes 2====================== 3 4Backward compatibility notes 5---------------------------- 6 7Updates since v2.9 8------------------ 9 10UI, Workflows & Features 11 12 * "git pull --rebase --verify-signature" learned to warn the user 13 that "--verify-signature" is a no-op when rebasing. 14 15 * An upstream project can make a recommendation to shallowly clone 16 some submodules in the .gitmodules file it ships. 17 18 * "git worktree add" learned that '-' can be used as a short-hand for 19 "@{-1}", the previous branch. 20 21 * Update the funcname definition to support css files. 22 23 * The completion script (in contrib/) learned to complete "git 24 status" options. 25 26 * Messages that are generated by auto gc during "git push" on the 27 receiving end are now passed back to the sending end in such a way 28 that they are shown with "remote: " prefix to avoid confusing the 29 users. 30 31 * "git add -i/-p" learned to honor diff.compactionHeuristic 32 experimental knob, so that the user can work on the same hunk split 33 as "git diff" output. 34 35 * "upload-pack" allows a custom "git pack-objects" replacement when 36 responding to "fetch/clone" via the uploadpack.packObjectsHook. 37 (merge b738396 jk/upload-pack-hook later to maint). 38 39 * Teach format-patch and mailsplit (hence "am") how a line that 40 happens to begin with "From " in the e-mail message is quoted with 41 ">", so that these lines can be restored to their original shape. 42 (merge d9925d1 ew/mboxrd-format-am later to maint). 43 44 * "git repack" learned the "--keep-unreachable" option, which sends 45 loose unreachable objects to a pack instead of leaving them loose. 46 This helps heuristics based on the number of loose objects 47 (e.g. "gc --auto"). 48 (merge e26a8c4 jk/repack-keep-unreachable later to maint). 49 50 * "log --graph --format=" learned that "%>|(N)" specifies the width 51 relative to the terminal's left edge, not relative to the area to 52 draw text that is to the right of the ancestry-graph section. It 53 also now accepts negative N that means the column limit is relative 54 to the right border. 55 56 * A careless invocation of "git send-email directory/" after editing 57 0001-change.patch with an editor often ends up sending both 58 0001-change.patch and its backup file, 0001-change.patch~, causing 59 embarrassment and a minor confusion. Detect such an input and 60 offer to skip the backup files when sending the patches out. 61 (merge 531220b jc/send-email-skip-backup later to maint). 62 63 * "git submodule update" that drives many "git clone" could 64 eventually hit flaky servers/network conditions on one of the 65 submodules; the command learned to retry the attempt. 66 67 * The output coloring scheme learned two new attributes, italic and 68 strike, in addition to existing bold, reverse, etc. 69 70 * "git log" learns log.showSignature configuration variable, and a 71 command line option "--no-show-signature" to countermand it. 72 (merge fce04c3 mj/log-show-signature-conf later to maint). 73 74 * More markings of messages for i18n, with updates to various tests 75 to pass GETTEXT_POISON tests. 76 77 * "git archive" learned to handle files that are larger than 8GB and 78 commits far in the future than expressible by the traditional US-TAR 79 format. 80 (merge 5caeeb8 jk/big-and-future-archive-tar later to maint). 81 82 * A new configuration variable core.sshCommand has been added to 83 specify what value for GIT_SSH_COMMAND to use per repository. 84 85 * "git worktree prune" protected worktrees that are marked as 86 "locked" by creating a file in a known location. "git worktree" 87 command learned a dedicated command pair to create and remove such 88 a file, so that the users do not have to do this with editor. 89 90 * A handful of "git svn" updates. 91 92 * "git push" learned to accept and pass extra options to the 93 receiving end so that hooks can read and react to them. 94 95 * "git status" learned to suggest "merge --abort" during a conflicted 96 merge, just like it already suggests "rebase --abort" during a 97 conflicted rebase. 98 99 * "git jump" script (in contrib/) has been updated a bit. 100 (merge a91e692 jk/git-jump later to maint). 101 102 * "git push" and "git clone" learned to give better progress meters 103 to the end user who is waiting on the terminal. 104 105 * An entry "git log --decorate" for the tip of the current branch is 106 shown as "HEAD -> name" (where "name" is the name of the branch); 107 the arrow is now painted in the same color as "HEAD", not in the 108 color for commits. 109 110 * "git format-patch" learned format.from configuration variable to 111 specify the default settings for its "--from" option. 112 113 * "git am -3" calls "git merge-recursive" when it needs to fall back 114 to a three-way merge; this call has been turned into an internal 115 subroutine call instead of spawning a separate subprocess. 116 117 * The command line completion scripts (in contrib/) now knows about 118 "git branch --delete/--move [--remote]". 119 (merge 2703c22 vs/completion-branch-fully-spelled-d-m-r later to maint). 120 121 122Performance, Internal Implementation, Development Support etc. 123 124 * "git fast-import" learned the same performance trick to avoid 125 creating too small a packfile as "git fetch" and "git push" have, 126 using *.unpackLimit configuration. 127 128 * When "git daemon" is run without --[init-]timeout specified, a 129 connection from a client that silently goes offline can hang around 130 for a long time, wasting resources. The socket-level KEEPALIVE has 131 been enabled to allow the OS to notice such failed connections. 132 133 * "git upload-pack" command has been updated to use the parse-options 134 API. 135 136 * The "git apply" standalone program is being libified; the first 137 step to move many state variables into a structure that can be 138 explicitly (re)initialized to make the machinery callable more 139 than once has been merged. 140 141 * HTTP transport gained an option to produce more detailed debugging 142 trace. 143 (merge 73e57aa ep/http-curl-trace later to maint). 144 145 * Instead of taking advantage of the fact that a struct string_list 146 that is allocated with all NULs happens to be the INIT_NODUP kind, 147 the users of string_list structures are taught to initialize them 148 explicitly as such, to document their behaviour better. 149 (merge 2721ce2 jk/string-list-static-init later to maint). 150 151 * HTTPd tests learned to show the server error log to help diagnosing 152 a failing tests. 153 (merge 44f243d nd/test-lib-httpd-show-error-log-in-verbose later to maint). 154 155 * The ownership rule for the piece of memory that hold references to 156 be fetched in "git fetch" was screwy, which has been cleaned up. 157 158 * "git bisect" makes an internal call to "git diff-tree" when 159 bisection finds the culprit, but this call did not initialize the 160 data structure to pass to the diff-tree API correctly. 161 162 * Further preparatory clean-up for "worktree" feature continues. 163 (merge 0409e0b nd/worktree-cleanup-post-head-protection later to maint). 164 165 * Formats of the various data (and how to validate them) where we use 166 GPG signature have been documented. 167 168 * A new run-command API function pipe_command() is introduced to 169 sanely feed data to the standard input while capturing data from 170 the standard output and the standard error of an external process, 171 which is cumbersome to hand-roll correctly without deadlocking. 172 173 The codepath to sign data in a prepared buffer with GPG has been 174 updated to use this API to read from the status-fd to check for 175 errors (instead of relying on GPG's exit status). 176 (merge efee955 jk/gpg-interface-cleanup later to maint). 177 178 * Allow t/perf framework to use the features from the most recent 179 version of Git even when testing an older installed version. 180 181 * The commands in the "log/diff" family have had an FILE* pointer in the 182 data structure they pass around for a long time, but some codepaths 183 used to always write to the standard output. As a preparatory step 184 to make "git format-patch" available to the internal callers, these 185 codepaths have been updated to consistently write into that FILE* 186 instead. 187 188 * Conversion from unsigned char sha1[20] to struct object_id 189 continues. 190 191 * Improve the look of the way "git fetch" reports what happened to 192 each ref that was fetched. 193 194 * The .c/.h sources are marked as such in our .gitattributes file so 195 that "git diff -W" and friends would work better. 196 197 * Code clean-up to avoid using a variable string that compilers may 198 feel untrustable as printf-style format given to write_file() 199 helper function. 200 201 * "git p4" used a location outside $GIT_DIR/refs/ to place its 202 temporary branches, which has been moved to refs/git-p4-tmp/. 203 204 * Existing autoconf generated test for the need to link with pthread 205 library did not check all the functions from pthread libraries; 206 recent FreeBSD has some functions in libc but not others, and we 207 mistakenly thought linking with libc is enough when it is not. 208 209 * When "git fsck" reports a broken link (e.g. a tree object contains 210 a blob that does not exist), both containing object and the object 211 that is referred to were reported with their 40-hex object names. 212 The command learned the "--name-objects" option to show the path to 213 the containing object from existing refs (e.g. "HEAD~24^2:file.txt"). 214 215 * Allow http daemon tests in Travis CI tests. 216 217 * Makefile assumed that -lrt is always available on platforms that 218 want to use clock_gettime() and CLOCK_MONOTONIC, which is not a 219 case for recent Mac OS X. The necessary symbols are often found in 220 libc on many modern systems and having -lrt on the command line, as 221 long as the library exists, had no effect, but when the platform 222 removes librt.a that is a different matter--having -lrt will break 223 the linkage. 224 225 This change could be seen as a regression for those who do need to 226 specify -lrt, as they now specifically ask for NEEDS_LIBRT when 227 building. Hopefully they are in the minority these days. 228 229 * Further preparatory work on the refs API before the pluggable 230 backend series can land. 231 232 * Error handling in the codepaths that updates refs has been 233 improved. 234 235 * The API to iterate over all the refs (i.e. for_each_ref(), etc.) 236 has been revamped. 237 238 * The handling of the "text=auto" attribute has been corrected. 239 $ echo "* text=auto eol=crlf" >.gitattributes 240 used to have the same effect as 241 $ echo "* text eol=crlf" >.gitattributes 242 i.e. declaring all files are text (ignoring "auto"). The 243 combination has been fixed to be equivalent to doing 244 $ git config core.autocrlf true 245 246 * A few tests that specifically target "git rebase -i" have been 247 added. 248 249 * Dumb http transport on the client side has been optimized. 250 (merge ecba195 ew/http-walker later to maint). 251 252 * Users of the parse_options_concat() API function need to allocate 253 extra slots in advance and fill them with OPT_END() when they want 254 to decide the set of supported options dynamically, which makes the 255 code error-prone and hard to read. This has been corrected by tweaking 256 the API to allocate and return a new copy of "struct option" array. 257 258 * "git fetch" exchanges batched have/ack messages between the sender 259 and the receiver, initially doubling every time and then falling 260 back to enlarge the window size linearly. The "smart http" 261 transport, being an half-duplex protocol, outgrows the preset limit 262 too quickly and becomes inefficient when interacting with a large 263 repository. The internal mechanism learned to grow the window size 264 more aggressively when working with the "smart http" transport. 265 266 * Tests for "git svn" have been taught to reuse the lib-httpd test 267 infrastructure when testing the subversion integration that 268 interacts with subversion repositories served over the http:// 269 protocol. 270 (merge a8a5d25 ew/git-svn-http-tests later to maint). 271 272 * "git pack-objects" has a few options that tell it not to pack 273 objects found in certain packfiles, which require it to scan .idx 274 files of all available packs. The codepaths involved in these 275 operations have been optimized for a common case of not having any 276 non-local pack and/or any .kept pack. 277 278 * The t3700 test about "add --chmod=-x" have been made a bit more 279 robust and generally cleaned up. 280 (merge 766cdc4 ib/t3700-add-chmod-x-updates later to maint). 281 282 * The build procedure learned PAGER_ENV knob that lists what default 283 environment variable settings to export for popular pagers. This 284 mechanism is used to tweak the default settings to MORE on FreeBSD. 285 (merge 995bc22 ew/build-time-pager-tweaks later to maint). 286 287 * The http-backend (the server-side component of smart-http 288 transport) used to trickle the HTTP header one at a time. Now 289 these write(2)s are batched. 290 (merge b36045c ew/http-backend-batch-headers later to maint). 291 292 * When "git rebase" tries to compare set of changes on the updated 293 upstream and our own branch, it computes patch-id for all of these 294 changes and attempts to find matches. This has been optimized by 295 lazily computing the full patch-id (which is expensive) to be 296 compared only for changes that touch the same set of paths. 297 (merge b3dfeeb kw/patch-ids-optim later to maint). 298 299 * A handful of tests that were broken under gettext-poison build have 300 been fixed. 301 302 303Also contains various documentation updates and code clean-ups. 304 305 306Fixes since v2.9 307---------------- 308 309Unless otherwise noted, all the fixes since v2.8 in the maintenance 310track are contained in this release (see the maintenance releases' 311notes for details). 312 313 * The commands in `git log` family take %C(auto) in a custom format 314 string. This unconditionally turned the color on, ignoring 315 --no-color or with --color=auto when the output is not connected to 316 a tty; this was corrected to make the format truly behave as 317 "auto". 318 319 * "git rev-list --count" whose walk-length is limited with "-n" 320 option did not work well with the counting optimized to look at the 321 bitmap index. 322 323 * "git show -W" (extend hunks to cover the entire function, delimited 324 by lines that match the "funcname" pattern) used to show the entire 325 file when a change added an entire function at the end of the file, 326 which has been fixed. 327 328 * The documentation set has been updated so that literal commands, 329 configuration variables and environment variables are consistently 330 typeset in fixed-width font and bold in manpages. 331 332 * "git svn propset" subcommand that was added in 2.3 days is 333 documented now. 334 335 * The documentation tries to consistently spell "GPG"; when 336 referring to the specific program name, "gpg" is used. 337 338 * "git reflog" stopped upon seeing an entry that denotes a branch 339 creation event (aka "unborn"), which made it appear as if the 340 reflog was truncated. 341 342 * The git-prompt scriptlet (in contrib/) was not friendly with those 343 who uses "set -u", which has been fixed. 344 345 * compat/regex code did not cleanly compile. 346 347 * A codepath that used alloca(3) to place an unbounded amount of data 348 on the stack has been updated to avoid doing so. 349 350 * "git update-index --add --chmod=+x file" may be usable as an escape 351 hatch, but not a friendly thing to force for people who do need to 352 use it regularly. "git add --chmod=+x file" can be used instead. 353 354 * Build improvements for gnome-keyring (in contrib/) 355 356 * "git status" used to say "working directory" when it meant "working 357 tree". 358 359 * Comments about misbehaving FreeBSD shells have been clarified with 360 the version number (9.x and before are broken, newer ones are OK). 361 362 * "git cherry-pick A" worked on an unborn branch, but "git 363 cherry-pick A..B" didn't. 364 365 * Fix an unintended regression in v2.9 that breaks "clone --depth" 366 that recurses down to submodules by forcing the submodules to also 367 be cloned shallowly, which many server instances that host upstream 368 of the submodules are not prepared for. 369 370 * Fix unnecessarily waste in the idiomatic use of ': ${VAR=default}' 371 to set the default value, without enclosing it in double quotes. 372 373 * Some platform-specific code had non-ANSI strict declarations of C 374 functions that do not take any parameters, which has been 375 corrected. 376 377 * The internal code used to show local timezone offset is not 378 prepared to handle timestamps beyond year 2100, and gave a 379 bogus offset value to the caller. Use a more benign looking 380 +0000 instead and let "git log" going in such a case, instead 381 of aborting. 382 383 * One among four invocations of readlink(1) in our test suite has 384 been rewritten so that the test can run on systems without the 385 command (others are in valgrind test framework and t9802). 386 387 * t/perf needs /usr/bin/time with GNU extension; the invocation of it 388 is updated to "gtime" on Darwin. 389 390 * A bug, which caused "git p4" while running under verbose mode to 391 report paths that are omitted due to branch prefix incorrectly, has 392 been fixed; the command said "Ignoring file outside of prefix" for 393 paths that are _inside_. 394 395 * The top level documentation "git help git" still pointed at the 396 documentation set hosted at now-defunct google-code repository. 397 Update it to point to https://git.github.io/htmldocs/git.html 398 instead. 399 400 * A helper function that takes the contents of a commit object and 401 finds its subject line did not ignore leading blank lines, as is 402 commonly done by other codepaths. Make it ignore leading blank 403 lines to match. 404 405 * For a long time, we carried an in-code comment that said our 406 colored output would work only when we use fprintf/fputs on 407 Windows, which no longer is the case for the past few years. 408 409 * "gc.autoPackLimit" when set to 1 should not trigger a repacking 410 when there is only one pack, but the code counted poorly and did 411 so. 412 413 * Add a test to specify the desired behaviour that currently is not 414 available in "git rebase -Xsubtree=...". 415 416 * More mark-up updates to typeset strings that are expected to 417 literally typed by the end user in fixed-width font. 418 419 * "git commit --amend --allow-empty-message -S" for a commit without 420 any message body could have misidentified where the header of the 421 commit object ends. 422 423 * "git rebase -i --autostash" did not restore the auto-stashed change 424 when the operation was aborted. 425 426 * Git does not know what the contents in the index should be for a 427 path added with "git add -N" yet, so "git grep --cached" should not 428 show hits (or show lack of hits, with -L) in such a path, but that 429 logic does not apply to "git grep", i.e. searching in the working 430 tree files. But we did so by mistake, which has been corrected. 431 432 * "git blame -M" missed a single line that was moved within the file. 433 434 * Fix recently introduced codepaths that are involved in parallel 435 submodule operations, which gave up on reading too early, and 436 could have wasted CPU while attempting to write under a corner 437 case condition. 438 439 * "git grep -i" has been taught to fold case in non-ascii locales 440 correctly. 441 442 * A test that unconditionally used "mktemp" learned that the command 443 is not necessarily available everywhere. 444 445 * There are certain house-keeping tasks that need to be performed at 446 the very beginning of any Git program, and programs that are not 447 built-in commands had to do them exactly the same way as "git" 448 potty does. It was easy to make mistakes in one-off standalone 449 programs (like test helpers). A common "main()" function that 450 calls cmd_main() of individual program has been introduced to 451 make it harder to make mistakes. 452 (merge de61ceb jk/common-main later to maint). 453 454 * The test framework learned a new helper test_match_signal to 455 check an exit code from getting killed by an expected signal. 456 457 * General code clean-up around a helper function to write a 458 single-liner to a file. 459 (merge 7eb6e10 jk/write-file later to maint). 460 461 * One part of "git am" had an oddball helper function that called 462 stuff from outside "his" as opposed to calling what we have "ours", 463 which was not gender-neutral and also inconsistent with the rest of 464 the system where outside stuff is usuall called "theirs" in 465 contrast to "ours". 466 467 * "git blame file" allowed the lineage of lines in the uncommitted, 468 unadded contents of "file" to be inspected, but it refused when 469 "file" did not appear in the current commit. When "file" was 470 created by renaming an existing file (but the change has not been 471 committed), this restriction was unnecessarily tight. 472 473 * "git add -N dir/file && git write-tree" produced an incorrect tree 474 when there are other paths in the same directory that sorts after 475 "file". 476 477 * "git fetch http://user:pass@host/repo..." scrubbed the userinfo 478 part, but "git push" didn't. 479 480 * "git merge" with renormalization did not work well with 481 merge-recursive, due to "safer crlf" conversion kicking in when it 482 shouldn't. 483 (merge 1335d76 jc/renormalize-merge-kill-safer-crlf later to maint). 484 485 * The use of strbuf in "git rm" to build filename to remove was a bit 486 suboptimal, which has been fixed. 487 488 * An age old bug that caused "git diff --ignore-space-at-eol" 489 misbehave has been fixed. 490 491 * "git notes merge" had a code to see if a path exists (and fails if 492 it does) and then open the path for writing (when it doesn't). 493 Replace it with open with O_EXCL. 494 495 * "git pack-objects" and "git index-pack" mostly operate with off_t 496 when talking about the offset of objects in a packfile, but there 497 were a handful of places that used "unsigned long" to hold that 498 value, leading to an unintended truncation. 499 500 * Recent update to "git daemon" tries to enable the socket-level 501 KEEPALIVE, but when it is spawned via inetd, the standard input 502 file descriptor may not necessarily be connected to a socket. 503 Suppress an ENOTSOCK error from setsockopt(). 504 505 * Recent FreeBSD stopped making perl available at /usr/bin/perl; 506 switch the default the built-in path to /usr/local/bin/perl on not 507 too ancient FreeBSD releases. 508 509 * "git commit --help" said "--no-verify" is only about skipping the 510 pre-commit hook, and failed to say that it also skipped the 511 commit-msg hook. 512 513 * "git merge" in Git v2.9 was taught to forbid merging an unrelated 514 lines of history by default, but that is exactly the kind of thing 515 the "--rejoin" mode of "git subtree" (in contrib/) wants to do. 516 "git subtree" has been taught to use the "--allow-unrelated-histories" 517 option to override the default. 518 519 * The build procedure for "git persistent-https" helper (in contrib/) 520 has been updated so that it can be built with more recent versions 521 of Go. 522 523 * There is an optimization used in "git diff $treeA $treeB" to borrow 524 an already checked-out copy in the working tree when it is known to 525 be the same as the blob being compared, expecting that open/mmap of 526 such a file is faster than reading it from the object store, which 527 involves inflating and applying delta. This however kicked in even 528 when the checked-out copy needs to go through the convert-to-git 529 conversion (including the clean filter), which defeats the whole 530 point of the optimization. The optimization has been disabled when 531 the conversion is necessary. 532 533 * "git -c grep.patternType=extended log --basic-regexp" misbehaved 534 because the internal API to access the grep machinery was not 535 designed well. 536 537 * Windows port was failing some tests in t4130, due to the lack of 538 inum in the returned values by its lstat(2) emulation. 539 540 * The reflog output format is documented better, and a new format 541 --date=unix to report the seconds-since-epoch (without timezone) 542 has been added. 543 (merge 442f6fd jk/reflog-date later to maint). 544 545 * "git difftool <paths>..." started in a subdirectory failed to 546 interpret the paths relative to that directory, which has been 547 fixed. 548 549 * The characters in the label shown for tags/refs for commits in 550 "gitweb" output are now properly escaped for proper HTML output. 551 552 * FreeBSD can lie when asked mtime of a directory, which made the 553 untracked cache code to fall back to a slow-path, which in turn 554 caused tests in t7063 to fail because it wanted to verify the 555 behaviour of the fast-path. 556 557 * Squelch compiler warnings for netmalloc (in compat/) library. 558 559 * A small memory leak in the command line parsing of "git blame" 560 has been plugged. 561 562 * The API documentation for hashmap was unclear if hashmap_entry 563 can be safely discarded without any other consideration. State 564 that it is safe to do so. 565 566 * Not-so-recent rewrite of "git am" that started making internal 567 calls into the commit machinery had an unintended regression, in 568 that no matter how many seconds it took to apply many patches, the 569 resulting committer timestamp for the resulting commits were all 570 the same. 571 572 * "git push --force-with-lease" already had enough logic to allow 573 ensuring that such a push results in creation of a ref (i.e. the 574 receiving end did not have another push from sideways that would be 575 discarded by our force-pushing), but didn't expose this possibility 576 to the users. It does so now. 577 (merge 9eed4f3 jk/push-force-with-lease-creation later to maint). 578 579 * The mechanism to limit the pack window memory size, when packing is 580 done using multiple threads (which is the default), is per-thread, 581 but this was not documented clearly. 582 (merge 954176c ms/document-pack-window-memory-is-per-thread later to maint). 583 584 * "import-tars" fast-import script (in contrib/) used to ignore a 585 hardlink target and replaced it with an empty file, which has been 586 corrected to record the same blob as the other file the hardlink is 587 shared with. 588 (merge 04e0869 js/import-tars-hardlinks later to maint). 589 590 * "git mv dir non-existing-dir/" did not work in some environments 591 the same way as existing mainstream platforms. The code now moves 592 "dir" to "non-existing-dir", without relying on rename("A", "B/") 593 that strips the trailing slash of '/'. 594 (merge 189d035 js/mv-dir-to-new-directory later to maint). 595 596 * The "t/" hierarchy is prone to get an unusual pathname; "make test" 597 has been taught to make sure they do not contain paths that cannot 598 be checked out on Windows (and the mechanism can be reusable to 599 catch pathnames that are not portable to other platforms as need 600 arises). 601 (merge c2cafd3 js/test-lint-pathname later to maint). 602 603 * Other minor clean-ups and documentation updates 604 (merge 02a8cfa rs/merge-add-strategies-simplification later to maint). 605 (merge af4941d rs/merge-recursive-string-list-init later to maint). 606 (merge 1eb47f1 rs/use-strbuf-add-unique-abbrev later to maint). 607 (merge ddd0bfa jk/tighten-alloc later to maint). 608 (merge ecf30b2 rs/mailinfo-lib later to maint). 609 (merge 0eb75ce sg/reflog-past-root later to maint).