1Git v1.8.2 Release Notes 2======================== 3 4Backward compatibility notes 5---------------------------- 6 7In the upcoming major release (tentatively called 1.8.2), we will 8change the behavior of the "git push" command. 9 10When "git push [$there]" does not say what to push, we have used the 11traditional "matching" semantics so far (all your branches were sent 12to the remote as long as there already are branches of the same name 13over there). We will use the "simple" semantics that pushes the 14current branch to the branch with the same name, only when the current 15branch is set to integrate with that remote branch. There is a user 16preference configuration variable "push.default" to change this. 17 18"git push $there tag v1.2.3" used to allow replacing a tag v1.2.3 19that already exists in the repository $there, if the rewritten tag 20you are pushing points at a commit that is a decendant of a commit 21that the old tag v1.2.3 points at. This was found to be error prone 22and starting with this release, any attempt to update an existing 23ref under refs/tags/ hierarchy will fail, without "--force". 24 25 26Updates since v1.8.1 27-------------------- 28 29UI, Workflows & Features 30 31 * Initial ports to QNX and z/OS UNIX System Services have started. 32 33 * Output from the tests is coloured using "green is okay, yellow is 34 questionable, red is bad and blue is informative" scheme. 35 36 * Mention of "GIT/Git/git" in the documentation have been updated to 37 be more uniform and consistent. The name of the system and the 38 concept it embodies is "Git"; the command the users type is "git". 39 All-caps "GIT" was merely a way to imitate "Git" typeset in small 40 caps in our ASCII text only documentation and to be avoided. 41 42 * The completion script (in contrib/completion) used to let the 43 default completer to suggest pathnames, which gave too many 44 irrelevant choices (e.g. "git add" would not want to add an 45 unmodified path). It learnt to use a more git-aware logic to 46 enumerate only relevant ones. 47 48 * In bare repositories, "git shortlog" and other commands now read 49 mailmap files from the tip of the history, to help running these 50 tools in server settings. 51 52 * Color specifiers, e.g. "%C(blue)Hello%C(reset)", used in the 53 "--format=" option of "git log" and friends can be disabled when 54 the output is not sent to a terminal by prefixing them with 55 "auto,", e.g. "%C(auto,blue)Hello%C(auto,reset)". 56 57 * Scripts can ask Git that wildcard patterns in pathspecs they give do 58 not have any significance, i.e. take them as literal strings. 59 60 * The patterns in .gitignore and .gitattributes files can have **/, 61 as a pattern that matches 0 or more levels of subdirectory. 62 E.g. "foo/**/bar" matches "bar" in "foo" itself or in a 63 subdirectory of "foo". 64 65 * When giving arguments without "--" disambiguation, object names 66 that come earlier on the command line must not be interpretable as 67 pathspecs and pathspecs that come later on the command line must 68 not be interpretable as object names. This disambiguation rule has 69 been tweaked so that ":/" (no other string before or after) is 70 always interpreted as a pathspec; "git cmd -- :/" is no longer 71 needed, you can just say "git cmd :/". 72 73 * Various "hint" lines Git gives when it asks the user to edit 74 messages in the editor are commented out with '#' by default. The 75 core.commentchar configuration variable can be used to customize 76 this '#' to a different character. 77 78 * "git add -u" and "git add -A" without pathspec issues warning to 79 make users aware that they are only operating on paths inside the 80 subdirectory they are in. Use ":/" (everything from the top) or 81 "." (everything from the $cwd) to disambiguate. 82 83 * "git blame" (and "git diff") learned the "--no-follow" option. 84 85 * "git branch" now rejects some nonsense combinations of command line 86 arguments (e.g. giving more than one branch name to rename) with 87 more case-specific error messages. 88 89 * "git check-ignore" command to help debugging .gitignore files has 90 been added. 91 92 * "git cherry-pick" can be used to replay a root commit to an unborn 93 branch. 94 95 * "git commit" can be told to use --cleanup=whitespace by setting the 96 configuration variable commit.cleanup to 'whitespace'. 97 98 * "git diff" and other Porcelain commands can be told to use a 99 non-standard algorithm by setting diff.algorithm configuration 100 variable. 101 102 * "git fetch --mirror" and fetch that uses other forms of refspec 103 with wildcard used to attempt to update a symbolic ref that match 104 the wildcard on the receiving end, which made little sense (the 105 real ref that is pointed at by the symbolic ref would be updated 106 anyway). Symbolic refs no longer are affected by such a fetch. 107 108 * "git format-patch" now detects more cases in which a whole branch 109 is being exported, and uses the description for the branch, when 110 asked to write a cover letter for the series. 111 112 * "git format-patch" learned "-v $count" option, and prepends a 113 string "v$count-" to the names of its output files, and also 114 automatically sets the subject prefix to "PATCH v$count". This 115 allows patches from rerolled series to be stored under different 116 names and makes it easier to reuse cover letter messsages. 117 118 * "git log" and friends can be told with --use-mailmap option to 119 rewrite the names and email addresses of people using the mailmap 120 mechanism. 121 122 * "git log --cc --graph" now shows the combined diff output with the 123 ancestry graph. 124 125 * "git log --grep=<pattern>" honors i18n.logoutputencoding to look 126 for the pattern after fixing the log message to the specified 127 encoding. 128 129 * "git mergetool" and "git difftool" learned to list the available 130 tool backends in a more consistent manner. 131 132 * "git mergetool" is aware of TortoiseGitMerge now and uses it over 133 TortoiseMerge when available. 134 135 * "git push" now requires "-f" to update a tag, even if it is a 136 fast-forward, as tags are meant to be fixed points. 137 138 * Error messages from "git push" when it stops to prevent remote refs 139 from getting overwritten by mistake have been improved to explain 140 various situations separately. 141 142 * "git push" will stop without doing anything if the new "pre-push" 143 hook exists and exits with a failure. 144 145 * When "git rebase" fails to generate patches to be applied (e.g. due 146 to oom), it failed to detect the failure and instead behaved as if 147 there were nothing to do. A workaround to use a temporary file has 148 been applied, but we probably would want to revisit this later, as 149 it hurts the common case of not failing at all. 150 151 * Input and preconditions to "git reset" has been loosened where 152 appropriate. "git reset $fromtree Makefile" requires $fromtree to 153 be any tree (it used to require it to be a commit), for example. 154 "git reset" (without options or parameters) used to error out when 155 you do not have any commits in your history, but it now gives you 156 an empty index (to match non-existent commit you are not even on). 157 158 * "git status" says what branch is being bisected or rebased when 159 able, not just "bisecting" or "rebasing". 160 161 * "git submodule" started learning a new mode to integrate with the 162 tip of the remote branch (as opposed to integrating with the commit 163 recorded in the superproject's gitlink). 164 165 * "git upload-pack" which implements the service "ls-remote" and 166 "fetch" talk to can be told to hide ref hierarchies the server 167 side internally uses (and that clients have no business learning 168 about) with transfer.hiderefs configuration. 169 170 171Foreign Interface 172 173 * "git fast-export" has been updated for its use in the context of 174 the remote helper interface. 175 176 * A new remote helper to interact with bzr has been added to contrib/. 177 178 * "git p4" got various bugfixes around its branch handling. It is 179 also made usable with Python 2.4/2.5. In addition, its various 180 portability issues for Cygwin have been addressed. 181 182 * The remote helper to interact with Hg in contrib/ has seen a few 183 fixes. 184 185 186Performance, Internal Implementation, etc. 187 188 * "git fsck" has been taught to be pickier about entries in tree 189 objects that should not be there, e.g. ".", ".git", and "..". 190 191 * Matching paths with common forms of pathspecs that contain wildcard 192 characters has been optimized further. 193 194 * We stopped paying attention to $GIT_CONFIG environment that points 195 at a single configuration file from any command other than "git config" 196 quite a while ago, but "git clone" internally set, exported, and 197 then unexported the variable during its operation unnecessarily. 198 199 * "git reset" internals has been reworked and should be faster in 200 general. We tried to be careful not to break any behaviour but 201 there could be corner cases, especially when running the command 202 from a conflicted state, that we may have missed. 203 204 * The implementation of "imap-send" has been updated to reuse xml 205 quoting code from http-push codepath, and lost a lot of unused 206 code. 207 208 * There is a simple-minded checker for the test scripts in t/ 209 directory to catch most common mistakes (it is not enabled by 210 default). 211 212 * You can build with USE_WILDMATCH=YesPlease to use a replacement 213 implementation of pattern matching logic used for pathname-like 214 things, e.g. refnames and paths in the repository. This new 215 implementation is not expected change the existing behaviour of Git 216 in this release, except for "git for-each-ref" where you can now 217 say "refs/**/master" and match with both refs/heads/master and 218 refs/remotes/origin/master. We plan to use this new implementation 219 in wider places (e.g. "git ls-files '**/Makefile' may find Makefile 220 at the top-level, and "git log '**/t*.sh'" may find commits that 221 touch a shell script whose name begins with "t" at any level) in 222 future versions of Git, but we are not there yet. By building with 223 USE_WILDMATCH, using the resulting Git daily and reporting when you 224 find breakages, you can help us get closer to that goal. 225 226 * Some reimplementations of Git do not write all the stat info back 227 to the index due to their implementation limitations (e.g. jgit). 228 A configuration option can tell Git to ignore changes to most of 229 the stat fields and only pay attention to mtime and size, which 230 these implementations can reliably update. This can be used to 231 avoid excessive revalidation of contents. 232 233 * Some platforms ship with old version of expat where xmlparse.h 234 needs to be included instead of expat.h; the build procedure has 235 been taught about this. 236 237 * "make clean" on platforms that cannot compute header dependencies 238 on the fly did not work with implementations of "rm" that do not 239 like an empty argument list. 240 241Also contains minor documentation updates and code clean-ups. 242 243 244Fixes since v1.8.1 245------------------ 246 247Unless otherwise noted, all the fixes since v1.8.1 in the maintenance 248track are contained in this release (see release notes to them for 249details). 250 251 * An element on GIT_CEILING_DIRECTORIES list that does not name the 252 real path to a directory (i.e. a symbolic link) could have caused 253 the GIT_DIR discovery logic to escape the ceiling. 254 255 * When attempting to read the XDG-style $HOME/.config/git/config and 256 finding that $HOME/.config/git is a file, we gave a wrong error 257 message, instead of treating the case as "a custom config file does 258 not exist there" and moving on. 259 260 * The behaviour visible to the end users was confusing, when they 261 attempt to kill a process spawned in the editor that was in turn 262 launched by Git with SIGINT (or SIGQUIT), as Git would catch that 263 signal and die. We ignore these signals now. 264 (merge 0398fc34 pf/editor-ignore-sigint later to maint). 265 266 * A child process that was killed by a signal (e.g. SIGINT) was 267 reported in an inconsistent way depending on how the process was 268 spawned by us, with or without a shell in between. 269 270 * After failing to create a temporary file using mkstemp(), failing 271 pathname was not reported correctly on some platforms. 272 273 * We used to stuff "user@" and then append what we read from 274 /etc/mailname to come up with a default e-mail ident, but a bug 275 lost the "user@" part. 276 277 * The attribute mechanism didn't allow limiting attributes to be 278 applied to only a single directory itself with "path/" like the 279 exclude mechanism does. The initial implementation of this that 280 was merged to 'maint' and 1.8.1.2 was with a severe performance 281 degradations and needs to merge a fix-up topic. 282 283 * The smart HTTP clients forgot to verify the content-type that comes 284 back from the server side to make sure that the request is being 285 handled properly. 286 (merge 3443db5 sp/smart-http-content-type-check later to maint). 287 288 * "git am" did not parse datestamp correctly from Hg generated patch, 289 when it is run in a locale outside C (or en). 290 291 * "git apply" misbehaved when fixing whitespace breakages by removing 292 excess trailing blank lines. 293 294 * "git apply --summary" has been taught to make sure the similarity 295 value shown in its output is sensible, even when the input had a 296 bogus value. 297 (merge afcb6ac jk/apply-similaritly-parsing later to maint). 298 299 * A tar archive created by "git archive" recorded a directory in a 300 way that made NetBSD's implementation of "tar" sometimes unhappy. 301 302 * "git archive" did not record uncompressed size in the header when 303 streaming a zip archive, which confused some implementations of unzip. 304 305 * "git archive" did not parse configuration values in tar.* namespace 306 correctly. 307 (merge b3873c3 jk/config-parsing-cleanup later to maint). 308 309 * Attempt to "branch --edit-description" an existing branch, while 310 being on a detached HEAD, errored out. 311 312 * "git clean" showed what it was going to do, but sometimes end up 313 finding that it was not allowed to do so, which resulted in a 314 confusing output (e.g. after saying that it will remove an 315 untracked directory, it found an embedded git repository there 316 which it is not allowed to remove). It now performs the actions 317 and then reports the outcome more faithfully. 318 (merge f538a91 zk/clean-report-failure later to maint). 319 320 * When "git clone --separate-git-dir=$over_there" is interrupted, it 321 failed to remove the real location of the $GIT_DIR it created. 322 This was most visible when interrupting a submodule update. 323 324 * "git cvsimport" mishandled timestamps at DST boundary. 325 (merge 48c9162 bw/get-tz-offset-perl later to maint). 326 327 * We used to have an arbitrary 32 limit for combined diff input, 328 resulting in incorrect number of leading colons shown when showing 329 the "--raw --cc" output. 330 (merge edbc00e jc/combine-diff-many-parents later to maint). 331 332 * "git fetch --depth" was broken in at least three ways. The 333 resulting history was deeper than specified by one commit, it was 334 unclear how to wipe the shallowness of the repository with the 335 command, and documentation was misleading. 336 (merge cfb70e1 nd/fetch-depth-is-broken later to maint). 337 338 * "git log --all -p" that walked refs/notes/textconv/ ref can later 339 try to use the textconv data incorrectly after it gets freed. 340 (merge be5c9fb jk/read-commit-buffer-data-after-free later to maint). 341 342 * We forgot to close the file descriptor reading from "gpg" output, 343 killing "git log --show-signature" on a long history. 344 345 * The way "git svn" asked for password using SSH_ASKPASS and 346 GIT_ASKPASS was not in line with the rest of the system. 347 348 * The --graph code fell into infinite loop when asked to do what the 349 code did not expect. 350 351 * http transport was wrong to ask for the username when the 352 authentication is done by certificate identity. 353 354 * "git pack-refs" that ran in parallel to another process that 355 created new refs had a nasty race. 356 357 * Rebasing the history of superproject with change in the submodule 358 has been broken since v1.7.12. 359 360 * After "git add -N" and then writing a tree object out of the 361 index, the cache-tree data structure got corrupted. 362 363 * "git clone" used to allow --bare and --separate-git-dir=$there 364 options at the same time, which was nonsensical. 365 (merge 95b63f1 nd/clone-no-separate-git-dir-with-bare later to maint). 366 367 * "git rebase --preserve-merges" lost empty merges in recent versions 368 of Git. 369 370 * "git merge --no-edit" computed who were involved in the work done 371 on the side branch, even though that information is to be discarded 372 without getting seen in the editor. 373 374 * "git merge" started calling prepare-commit-msg hook like "git 375 commit" does some time ago, but forgot to pay attention to the exit 376 status of the hook. 377 378 * A failure to push due to non-ff while on an unborn branch 379 dereferenced a NULL pointer when showing an error message. 380 381 * When users spell "cc:" in lowercase in the fake "header" in the 382 trailer part, "git send-email" failed to pick up the addresses from 383 there. As e-mail headers field names are case insensitive, this 384 script should follow suit and treat "cc:" and "Cc:" the same way. 385 386 * Output from "git status --ignored" showed an unexpected interaction 387 with "--untracked". 388 389 * "gitweb", when sorting by age to show repositories with new 390 activities first, used to sort repositories with absolutely 391 nothing in it early, which was not very useful. 392 393 * "gitweb"'s code to sanitize control characters before passing it to 394 "highlight" filter lost known-to-be-safe control characters by 395 mistake. 396 397 * "gitweb" pages served over HTTPS, when configured to show picon or 398 gravatar, referred to these external resources to be fetched via 399 HTTP, resulting in mixed contents warning in browsers. 400 (merge 5748558 ab/gitweb-use-same-scheme later to maint). 401 402 * When a line to be wrapped has a solid run of non space characters 403 whose length exactly is the wrap width, "git shortlog -w" failed 404 to add a newline after such a line. 405 406 * Command line completion leaked an unnecessary error message while 407 looking for possible matches with paths in <tree-ish>. 408 409 * Command line completion for "tcsh" emitted an unwanted space 410 after completing a single directory name. 411 412 * Command line completion code was inadvertently made incompatible with 413 older versions of bash by using a newer array notation. 414 415 * "git push" was taught to refuse updating the branch that is 416 currently checked out long time ago, but the user manual was left 417 stale. 418 (merge d9be248 wk/man-deny-current-branch-is-default-these-days later to maint). 419 420 * Some shells do not behave correctly when IFS is unset; work it 421 around by explicitly setting it to the default value. 422 423 * Some scripted programs written in Python did not get updated when 424 PYTHON_PATH changed. 425 (cherry-pick 96a4647fca54031974cd6ad1 later to maint). 426 427 * When autoconf is used, any build on a different commit always ran 428 "config.status --recheck" even when unnecessary. 429 430 * A fix was added to the build procedure to work around buggy 431 versions of ccache broke the auto-generation of dependencies, which 432 unfortunately is still relevant because some people use ancient 433 distros. 434 435 * The autoconf subsystem passed --mandir down to generated 436 config.mak.autogen but forgot to do the same for --htmldir. 437 (merge fc1c541 ct/autoconf-htmldir later to maint). 438 439 * We have been carrying a translated and long-unmaintained copy of an 440 old version of the tutorial; removed. 441 442 * t0050 had tests expecting failures from a bug that was fixed some 443 time ago. 444 445 * t4014, t9502 and t0200 tests had various portability issues that 446 broke on OpenBSD. 447 448 * t9020 and t3600 tests had various portability issues. 449 450 * t9200 runs "cvs init" on a directory that already exists, but a 451 platform can configure this fail for the current user (e.g. you 452 need to be in the cvsadmin group on NetBSD 6.0). 453 454 * t9020 and t9810 had a few non-portable shell script construct. 455 456 * Scripts to test bash completion was inherently flaky as it was 457 affected by whatever random things the user may have on $PATH.