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 * In bare repositories, "git shortlog" and other commands now read 43 mailmap files from the tip of the history, to help running these 44 tools in server settings. 45 46 * Color specifiers, e.g. "%C(blue)Hello%C(reset)", used in the 47 "--format=" option of "git log" and friends can be disabled when 48 the output is not sent to a terminal by prefixing them with 49 "auto,", e.g. "%C(auto,blue)Hello%C(auto,reset)". 50 51 * Scripts can ask Git that wildcard patterns in pathspecs they give do 52 not have any significance, i.e. take them as literal strings. 53 54 * The patterns in .gitignore and .gitattributes files can have **/, 55 as a pattern that matches 0 or more levels of subdirectory. 56 E.g. "foo/**/bar" matches "bar" in "foo" itself or in a 57 subdirectory of "foo". 58 59 * When giving arguments without "--" disambiguation, object names 60 that come earlier on the command line must not be interpretable as 61 pathspecs and pathspecs that come later on the command line must 62 not be interpretable as object names. This disambiguation rule has 63 been tweaked so that ":/" (no other string before or after) is 64 always interpreted as a pathspec; "git cmd -- :/" is no longer 65 needed, you can just say "git cmd :/". 66 67 * Various "hint" lines Git gives when it asks the user to edit 68 messages in the editor are commented out with '#' by default. The 69 core.commentchar configuration variable can be used to customize 70 this '#' to a different character. 71 72 * "git add -u" and "git add -A" without pathspec issues warning to 73 make users aware that they are only operating on paths inside the 74 subdirectory they are in. Use ":/" (everything from the top) or 75 "." (everything from the $cwd) to disambiguate. 76 77 * "git blame" (and "git diff") learned the "--no-follow" option. 78 79 * "git check-ignore" command to help debugging .gitignore files has 80 been added. 81 82 * "git cherry-pick" can be used to replay a root commit to an unborn 83 branch. 84 85 * "git commit" can be told to use --cleanup=whitespace by setting the 86 configuration variable commit.cleanup to 'whitespace'. 87 88 * "git fetch --mirror" and fetch that uses other forms of refspec 89 with wildcard used to attempt to update a symbolic ref that match 90 the wildcard on the receiving end, which made little sense (the 91 real ref that is pointed at by the symbolic ref would be updated 92 anyway). Symbolic refs no longer are affected by such a fetch. 93 94 * "git format-patch" now detects more cases in which a whole branch 95 is being exported, and uses the description for the branch, when 96 asked to write a cover letter for the series. 97 98 * "git format-patch" learned "-v $count" option, and prepends a 99 string "v$count-" to the names of its output files, and also 100 automatically sets the subject prefix to "PATCH v$count". This 101 allows patches from rerolled series to be stored under different 102 names and makes it easier to reuse cover letter messsages. 103 104 * "git log" and friends can be told with --use-mailmap option to 105 rewrite the names and email addresses of people using the mailmap 106 mechanism. 107 108 * "git push" now requires "-f" to update a tag, even if it is a 109 fast-forward, as tags are meant to be fixed points. 110 111 * Error messages from "git push" when it stops to prevent remote refs 112 from getting overwritten by mistake have been improved to explain 113 various situations separately. 114 115 * "git push" will stop without doing anything if the new "pre-push" 116 hook exists and exits with a failure. 117 118 * When "git rebase" fails to generate patches to be applied (e.g. due 119 to oom), it failed to detect the failure and instead behaved as if 120 there were nothing to do. A workaround to use a temporary file has 121 been applied, but we probably would want to revisit this later, as 122 it hurts the common case of not failing at all. 123 124 * Input and preconditions to "git reset" has been loosened where 125 appropriate. "git reset $fromtree Makefile" requires $fromtree to 126 be any tree (it used to require it to be a commit), for example. 127 "git reset" (without options or parameters) used to error out when 128 you do not have any commits in your history, but it now gives you 129 an empty index (to match non-existent commit you are not even on). 130 131 * "git submodule" started learning a new mode to integrate with the 132 tip of the remote branch (as opposed to integrating with the commit 133 recorded in the superproject's gitlink). 134 135 136Foreign Interface 137 138 * "git fast-export" has been updated for its use in the context of 139 the remote helper interface. 140 141 * A new remote helper to interact with bzr has been added to contrib/. 142 143 * "git p4" got various bugfixes around its branch handling. It is 144 also made usable with Python 2.4/2.5. In addition, its various 145 portability issues for Cygwin have been addressed. 146 147 * The remote helper to interact with Hg in contrib/ has seen a few 148 fixes. 149 150 151Performance, Internal Implementation, etc. 152 153 * "git fsck" has been taught to be pickier about entries in tree 154 objects that should not be there, e.g. ".", ".git", and "..". 155 156 * Matching paths with common forms of pathspecs that contain wildcard 157 characters has been optimized further. 158 159 * We stopped paying attention to $GIT_CONFIG environment that points 160 at a single configuration file from any command other than "git config" 161 quite a while ago, but "git clone" internally set, exported, and 162 then unexported the variable during its operation unnecessarily. 163 164 * "git reset" internals has been reworked and should be faster in 165 general. We tried to be careful not to break any behaviour but 166 there could be corner cases, especially when running the command 167 from a conflicted state, that we may have missed. 168 169 * The implementation of "imap-send" has been updated to reuse xml 170 quoting code from http-push codepath, and lost a lot of unused 171 code. 172 173 * There is a simple-minded checker for the test scripts in t/ 174 directory to catch most common mistakes (it is not enabled by 175 default). 176 177 * You can build with USE_WILDMATCH=YesPlease to use a replacement 178 implementation of pattern matching logic used for pathname-like 179 things, e.g. refnames and paths in the repository. This new 180 implementation is not expected change the existing behaviour of Git 181 in this release, except for "git for-each-ref" where you can now 182 say "refs/**/master" and match with both refs/heads/master and 183 refs/remotes/origin/master. We plan to use this new implementation 184 in wider places (e.g. "git ls-files '**/Makefile' may find Makefile 185 at the top-level, and "git log '**/t*.sh'" may find commits that 186 touch a shell script whose name begins with "t" at any level) in 187 future versions of Git, but we are not there yet. By building with 188 USE_WILDMATCH, using the resulting Git daily and reporting when you 189 find breakages, you can help us get closer to that goal. 190 191 * Some reimplementations of Git do not write all the stat info back 192 to the index due to their implementation limitations (e.g. jgit). 193 A configuration option can tell Git to ignore changes to most of 194 the stat fields and only pay attention to mtime and size, which 195 these implementations can reliably update. This can be used to 196 avoid excessive revalidation of contents. 197 198 199Also contains minor documentation updates and code clean-ups. 200 201 202Fixes since v1.8.1 203------------------ 204 205Unless otherwise noted, all the fixes since v1.8.1 in the maintenance 206track are contained in this release (see release notes to them for 207details). 208 209 * An element on GIT_CEILING_DIRECTORIES list that does not name the 210 real path to a directory (i.e. a symbolic link) could have caused 211 the GIT_DIR discovery logic to escape the ceiling. 212 213 * When attempting to read the XDG-style $HOME/.config/git/config and 214 finding that $HOME/.config/git is a file, we gave a wrong error 215 message, instead of treating the case as "a custom config file does 216 not exist there" and moving on. 217 218 * The behaviour visible to the end users was confusing, when they 219 attempt to kill a process spawned in the editor that was in turn 220 launched by Git with SIGINT (or SIGQUIT), as Git would catch that 221 signal and die. We ignore these signals now. 222 (merge 0398fc34 pf/editor-ignore-sigint later to maint). 223 224 * A child process that was killed by a signal (e.g. SIGINT) was 225 reported in an inconsistent way depending on how the process was 226 spawned by us, with or without a shell in between. 227 228 * After failing to create a temporary file using mkstemp(), failing 229 pathname was not reported correctly on some platforms. 230 231 * We used to stuff "user@" and then append what we read from 232 /etc/mailname to come up with a default e-mail ident, but a bug 233 lost the "user@" part. 234 235 * The attribute mechanism didn't allow limiting attributes to be 236 applied to only a single directory itself with "path/" like the 237 exclude mechanism does. The initial implementation of this that 238 was merged to 'maint' and 1.8.1.2 was with a severe performance 239 degradations and needs to merge a fix-up topic. 240 241 * "git am" did not parse datestamp correctly from Hg generated patch, 242 when it is run in a locale outside C (or en). 243 244 * "git apply" misbehaved when fixing whitespace breakages by removing 245 excess trailing blank lines. 246 247 * A tar archive created by "git archive" recorded a directory in a 248 way that made NetBSD's implementation of "tar" sometimes unhappy. 249 250 * "git archive" did not record uncompressed size in the header when 251 streaming a zip archive, which confused some implementations of unzip. 252 253 * "git archive" did not parse configuration values in tar.* namespace 254 correctly. 255 (merge b3873c3 jk/config-parsing-cleanup later to maint). 256 257 * Attempt to "branch --edit-description" an existing branch, while 258 being on a detached HEAD, errored out. 259 260 * "git clean" showed what it was going to do, but sometimes end up 261 finding that it was not allowed to do so, which resulted in a 262 confusing output (e.g. after saying that it will remove an 263 untracked directory, it found an embedded git repository there 264 which it is not allowed to remove). It now performs the actions 265 and then reports the outcome more faithfully. 266 (merge f538a91 zk/clean-report-failure later to maint). 267 268 * When "git clone --separate-git-dir=$over_there" is interrupted, it 269 failed to remove the real location of the $GIT_DIR it created. 270 This was most visible when interrupting a submodule update. 271 272 * "git fetch --depth" was broken in at least three ways. The 273 resulting history was deeper than specified by one commit, it was 274 unclear how to wipe the shallowness of the repository with the 275 command, and documentation was misleading. 276 (merge cfb70e1 nd/fetch-depth-is-broken later to maint). 277 278 * "git log --all -p" that walked refs/notes/textconv/ ref can later 279 try to use the textconv data incorrectly after it gets freed. 280 (merge be5c9fb jk/read-commit-buffer-data-after-free later to maint). 281 282 * We forgot to close the file descriptor reading from "gpg" output, 283 killing "git log --show-signature" on a long history. 284 (merge 7dac3f8 sb/gpg-plug-fd-leak later to maint). 285 286 * The way "git svn" asked for password using SSH_ASKPASS and 287 GIT_ASKPASS was not in line with the rest of the system. 288 289 * The --graph code fell into infinite loop when asked to do what the 290 code did not expect. 291 292 * http transport was wrong to ask for the username when the 293 authentication is done by certificate identity. 294 295 * "git pack-refs" that ran in parallel to another process that 296 created new refs had a nasty race. 297 298 * Rebasing the history of superproject with change in the submodule 299 has been broken since v1.7.12. 300 (merge e28efb1 jc/fake-ancestor-with-non-blobs later to maint). 301 302 * After "git add -N" and then writing a tree object out of the 303 index, the cache-tree data structure got corrupted. 304 305 * "git clone" used to allow --bare and --separate-git-dir=$there 306 options at the same time, which was nonsensical. 307 (merge 95b63f1 nd/clone-no-separate-git-dir-with-bare later to maint). 308 309 * "git rebase --preserve-merges" lost empty merges in recent versions 310 of Git. 311 312 * "git merge --no-edit" computed who were involved in the work done 313 on the side branch, even though that information is to be discarded 314 without getting seen in the editor. 315 316 * "git merge" started calling prepare-commit-msg hook like "git 317 commit" does some time ago, but forgot to pay attention to the exit 318 status of the hook. 319 320 * A failure to push due to non-ff while on an unborn branch 321 dereferenced a NULL pointer when showing an error message. 322 (merge 1d2c14d ft/transport-report-segv later to maint). 323 324 * When users spell "cc:" in lowercase in the fake "header" in the 325 trailer part, "git send-email" failed to pick up the addresses from 326 there. As e-mail headers field names are case insensitive, this 327 script should follow suit and treat "cc:" and "Cc:" the same way. 328 329 * Output from "git status --ignored" showed an unexpected interaction 330 with "--untracked". 331 332 * "gitweb", when sorting by age to show repositories with new 333 activities first, used to sort repositories with absolutely 334 nothing in it early, which was not very useful. 335 336 * "gitweb"'s code to sanitize control characters before passing it to 337 "highlight" filter lost known-to-be-safe control characters by 338 mistake. 339 340 * When a line to be wrapped has a solid run of non space characters 341 whose length exactly is the wrap width, "git shortlog -w" failed 342 to add a newline after such a line. 343 344 * Command line completion leaked an unnecessary error message while 345 looking for possible matches with paths in <tree-ish>. 346 347 * Command line completion for "tcsh" emitted an unwanted space 348 after completing a single directory name. 349 350 * Command line completion code was inadvertently made incompatible with 351 older versions of bash by using a newer array notation. 352 353 * Some shells do not behave correctly when IFS is unset; work it 354 around by explicitly setting it to the default value. 355 356 * Some scripted programs written in Python did not get updated when 357 PYTHON_PATH changed. 358 (cherry-pick 96a4647fca54031974cd6ad1 later to maint). 359 360 * When autoconf is used, any build on a different commit always ran 361 "config.status --recheck" even when unnecessary. 362 363 * A fix was added to the build procedure to work around buggy 364 versions of ccache broke the auto-generation of dependencies, which 365 unfortunately is still relevant because some people use ancient 366 distros. 367 (merge 6978934 jn/auto-depend-workaround-buggy-ccache later to maint). 368 369 * We have been carrying a translated and long-unmaintained copy of an 370 old version of the tutorial; removed. 371 372 * t0050 had tests expecting failures from a bug that was fixed some 373 time ago. 374 375 * t4014, t9502 and t0200 tests had various portability issues that 376 broke on OpenBSD. 377 378 * t9020 and t3600 tests had various portability issues. 379 380 * t9200 runs "cvs init" on a directory that already exists, but a 381 platform can configure this fail for the current user (e.g. you 382 need to be in the cvsadmin group on NetBSD 6.0). 383 384 * t9020 and t9810 had a few non-portable shell script construct. 385 386 * Scripts to test bash completion was inherently flaky as it was 387 affected by whatever random things the user may have on $PATH. 388 (merge 5047822 jc/do-not-let-random-file-interfere-with-completion-tests later to maint).