1Git 2.7 Release Notes 2===================== 3 4Updates since v2.6 5------------------ 6 7UI, Workflows & Features 8 9 * The appearance of "gitk", particularly on high DPI monitors, have 10 been improved. "gitk" also comes with an undated translation for 11 Swedish and Japanese. 12 13 * "git remote" learned "get-url" subcommand to show the URL for a 14 given remote name used for fetching and pushing. 15 16 * There was no way to defeat a configured rebase.autostash variable 17 from the command line, as "git rebase --no-autostash" was missing. 18 19 * "git log --date=local" used to only show the normal (default) 20 format in the local timezone. The command learned to take 'local' 21 as an instruction to use the local timezone with other formats, 22 23 * The refs used during a "git bisect" session is now per-worktree so 24 that independent bisect sessions can be done in different worktrees 25 created with "git worktree add". 26 27 * Users who are too busy to type three extra keystrokes to ask for 28 "git stash show -p" can now set stash.showPatch configuration 29 variable to true to always see the actual patch, not just the list 30 of paths affected with feel for the extent of damage via diffstat. 31 32 * "quiltimport" allows to specify the series file by honoring the 33 $QUILT_SERIES environment and also --series command line option. 34 35 * The use of 'good/bad' in "git bisect" made it confusing to use when 36 hunting for a state change that is not a regression (e.g. bugfix). 37 The command learned 'old/new' and then allows the end user to 38 say e.g. "bisect start --term-old=fast --term-new=slow" to find a 39 performance regression. 40 41 * "git interpret-trailers" can now run outside of a Git repository. 42 43 * "git p4" learned to reencode the pathname it uses to communicate 44 with the p4 depot with a new option. 45 46 * Give progress meter to "git filter-branch". 47 48 * Allow a later "!/abc/def" to override an earlier "/abc" that 49 appears in the same .gitignore file to make it easier to express 50 "everything in /abc directory is ignored, except for ...". 51 52 * Teach "git p4" to send large blobs outside the repository by 53 talking to Git LFS. 54 55 * Prepare for Git on-disk repository representation to undergo 56 backward incompatible changes by introducing a new repository 57 format version "1", with an extension mechanism. 58 59 * "git worktree" learned a "list" subcommand. 60 61 * "git clone --dissociate" learned that it can be used even when 62 "--reference" was not used at the same time. 63 64 * "git blame" learnt to take "--first-parent" and "--reverse" at the 65 same time when it makes sense. 66 67 * "git checkout" did not follow the usual "--[no-]progress" 68 convention and implemented only "--quiet" that is essentially 69 a superset of "--no-progress". Extend the command to support the 70 usual "--[no-]progress". 71 72 * The semantics of transfer.hideRefs configuration variable have been 73 extended to work better with the ref "namespace" feature that lets 74 you throw unrelated bunches of repositories in a single physical 75 repository and virtually serve them as separate ones. 76 77 * send-email config variables whose values are pathnames now go 78 through the ~username/ expansion. 79 80 * bash completion learnt to TAB-complete recipient addresses given 81 to send-email. 82 83 * The credential-cache daemon can be told to ignore SIGHUP to work 84 around issue when running Git from inside emacs. 85 86 * "git push" learned new configuration for doing "--recurse-submodules" 87 on each push. 88 89 * "format-patch" has learned a new option to zero-out the commit 90 object name on the mbox "From " line. 91 92 93Performance, Internal Implementation, Development Support etc. 94 95 * The infrastructure to rewrite "git submodule" in C is being built 96 incrementally. Let's polish these early parts well enough and make 97 them graduate to 'next' and 'master', so that the more involved 98 follow-up can start cooking on a solid ground. 99 100 * Some features from "git tag -l" and "git branch -l" have been made 101 available to "git for-each-ref" so that eventually the unified 102 implementation can be shared across all three. The version merged 103 to the 'master' branch earlier had a performance regression in "tag 104 --contains", which has since been corrected. 105 106 * Because "test_when_finished" in our test framework queues the 107 clean-up tasks to be done in a shell variable, it should not be 108 used inside a subshell. Add a mechanism to allow 'bash' to catch 109 such uses, and fix the ones that were found. 110 111 * The debugging infrastructure for pkt-line based communication has 112 been improved to mark the side-band communication specifically. 113 114 * Update "git branch" that list existing branches, using the 115 ref-filter API that is shared with "git tag" and "git 116 for-each-ref". 117 118 * The test for various line-ending conversions has been enhanced. 119 120 * A few test scripts around "git p4" have been improved for 121 portability. 122 123 * Many allocations that is manually counted (correctly) that are 124 followed by strcpy/sprintf have been replaced with a less error 125 prone constructs such as xstrfmt. 126 127 * The internal stripspace() function has been moved to where it 128 logically belongs to, i.e. strbuf API, and the command line parser 129 of "git stripspace" has been updated to use the parse_options API. 130 131 * "git am" used to spawn "git mailinfo" via run_command() API once 132 per each patch, but learned to make a direct call to mailinfo() 133 instead. 134 135 * The implementation of "git mailinfo" was refactored so that a 136 mailinfo() function can be directly called from inside a process. 137 138 * With a "debug" helper, debugging of a single "git" invocation in 139 our test scripts has become a lot easier. 140 141 * The "configure" script did not test for -lpthread correctly, which 142 upset some linkers. 143 144 * Cross completed task off of subtree project's todo list. 145 146 * Test cleanups for the subtree project. 147 148 * Clean up style in an ancient test t9300. 149 150 * Work around some test flakiness with p4d. 151 152 * Fsck did not correctly detect a NUL-truncated header in a tag. 153 154 * Use a safer behavior when we hit errors verifying remote certificates. 155 156 * Speed up filter-branch for cases where we only care about rewriting 157 commits, not tree data. 158 159 * The parse-options API has been updated to make "-h" command line 160 option work more consistently in all commands. 161 162 * "git svn rebase/mkdirs" got optimized by keeping track of empty 163 directories better. 164 165 * Fix some racy client/server tests by treating SIGPIPE the same as a 166 normal non-zero exit. 167 168 * The necessary infrastructure to build topics using the free Travis 169 CI has been added. Developers forking from this topic (and enabling 170 Travis) can do their own builds, and we can turn on auto-builds for 171 git/git (including build-status for pull requests that people 172 open). 173 174 175Also contains various documentation updates and code clean-ups. 176 177 178Fixes since v2.6 179---------------- 180 181Unless otherwise noted, all the fixes since v2.6 in the maintenance 182track are contained in this release (see the maintenance releases' 183notes for details). 184 185 * Very small number of options take a parameter that is optional 186 (which is not a great UI element as they can only appear at the end 187 of the command line). Add notice to documentation of each and 188 every one of them. 189 190 * "git blame --first-parent v1.0..v2.0" was not rejected but did not 191 limit the blame to commits on the first parent chain. 192 193 * "git subtree" (in contrib/) now can take whitespaces in the 194 pathnames, not only in the in-tree pathname but the name of the 195 directory that the repository is in. 196 197 * The ssh transport, just like any other transport over the network, 198 did not clear GIT_* environment variables, but it is possible to 199 use SendEnv and AcceptEnv to leak them to the remote invocation of 200 Git, which is not a good idea at all. Explicitly clear them just 201 like we do for the local transport. 202 203 * Correct "git p4 --detect-labels" so that it does not fail to create 204 a tag that points at a commit that is also being imported. 205 206 * The Makefile always runs the library archiver with hardcoded "crs" 207 options, which was inconvenient for exotic platforms on which 208 people want to use programs with totally different set of command 209 line options. 210 211 * Customization to change the behaviour with "make -w" and "make -s" 212 in our Makefile was broken when they were used together. 213 214 * Allocation related functions and stdio are unsafe things to call 215 inside a signal handler, and indeed killing the pager can cause 216 glibc to deadlock waiting on allocation mutex as our signal handler 217 tries to free() some data structures in wait_for_pager(). Reduce 218 these unsafe calls. 219 220 * The way how --ref/--notes to specify the notes tree reference are 221 DWIMmed was not clearly documented. 222 223 * "git gc" used to barf when a symbolic ref has gone dangling 224 (e.g. the branch that used to be your upstream's default when you 225 cloned from it is now gone, and you did "fetch --prune"). 226 227 * "git clone --dissociate" runs a big "git repack" process at the 228 end, and it helps to close file descriptors that are open on the 229 packs and their idx files before doing so on filesystems that 230 cannot remove a file that is still open. 231 232 * Description of the "log.follow" configuration variable in "git log" 233 documentation is now also copied to "git config" documentation. 234 235 * "git rebase -i" had a minor regression recently, which stopped 236 considering a line that begins with an indented '#' in its insn 237 sheet not a comment. Further, the code was still too picky on 238 Windows where CRLF left by the editor is turned into a trailing CR 239 on the line read via the "read" built-in command of bash. Both of 240 these issues are now fixed. 241 242 * After "git checkout --detach", "git status" reported a fairly 243 useless "HEAD detached at HEAD", instead of saying at which exact 244 commit. 245 246 * When "git send-email" wanted to talk over Net::SMTP::SSL, 247 Net::Cmd::datasend() did not like to be fed too many bytes at the 248 same time and failed to send messages. Send the payload one line 249 at a time to work around the problem. 250 251 * When "git am" was rewritten as a built-in, it stopped paying 252 attention to user.signingkey, which was fixed. 253 254 * It was not possible to use a repository-lookalike created by "git 255 worktree add" as a local source of "git clone". 256 257 * On a case insensitive filesystems, setting GIT_WORK_TREE variable 258 using a random cases that does not agree with what the filesystem 259 thinks confused Git that it wasn't inside the working tree. 260 261 * Performance-measurement tests did not work without an installed Git. 262 263 * A test script for the HTTP service had a timing dependent bug, 264 which was fixed. 265 266 * There were some classes of errors that "git fsck" diagnosed to its 267 standard error that did not cause it to exit with non-zero status. 268 269 * Work around "git p4" failing when the P4 depot records the contents 270 in UTF-16 without UTF-16 BOM. 271 272 * When "git gc --auto" is backgrounded, its diagnosis message is 273 lost. Save it to a file in $GIT_DIR and show it next time the "gc 274 --auto" is run. 275 276 * The submodule code has been taught to work better with separate 277 work trees created via "git worktree add". 278 279 * "git gc" is safe to run anytime only because it has the built-in 280 grace period to protect young objects. In order to run with no 281 grace period, the user must make sure that the repository is 282 quiescent. 283 284 * A recent "filter-branch --msg-filter" broke skipping of the commit 285 object header, which is fixed. 286 287 * The normalize_ceiling_entry() function does not muck with the end 288 of the path it accepts, and the real world callers do rely on that, 289 but a test insisted that the function drops a trailing slash. 290 291 * A test for interaction between untracked cache and sparse checkout 292 added in Git 2.5 days were flaky. 293 294 * A couple of commands still showed "[options]" in their usage string 295 to note where options should come on their command line, but we 296 spell that "[<options>]" in most places these days. 297 298 * The synopsis text and the usage string of subcommands that read 299 list of things from the standard input are often shown as if they 300 only take input from a file on a filesystem, which was misleading. 301 302 * "git am -3" had a small regression where it is aborted in its error 303 handling codepath when underlying merge-recursive failed in certain 304 ways, as it assumed that the internal call to merge-recursive will 305 never die, which is not the case (yet). 306 307 * The linkage order of libraries was wrong in places around libcurl. 308 309 * The name-hash subsystem that is used to cope with case insensitive 310 filesystems keeps track of directories and their on-filesystem 311 cases for all the paths in the index by holding a pointer to a 312 randomly chosen cache entry that is inside the directory (for its 313 ce->ce_name component). This pointer was not updated even when the 314 cache entry was removed from the index, leading to use after free. 315 This was fixed by recording the path for each directory instead of 316 borrowing cache entries and restructuring the API somewhat. 317 318 * "git merge-file" tried to signal how many conflicts it found, which 319 obviously would not work well when there are too many of them. 320 321 * The error message from "git blame --contents --reverse" incorrectly 322 talked about "--contents --children". 323 324 * "git imap-send" did not compile well with older version of cURL library. 325 326 * Merging a branch that removes a path and another that changes the 327 mode bits on the same path should have conflicted at the path, but 328 it didn't and silently favoured the removal. 329 330 * "git --literal-pathspecs add -u/-A" without any command line 331 argument misbehaved ever since Git 2.0. 332 333 * "git daemon" uses "run_command()" without "finish_command()", so it 334 needs to release resources itself, which it forgot to do. 335 336 * "git status --branch --short" accessed beyond the constant string 337 "HEAD", which has been corrected. 338 339 * We peek objects from submodule's object store by linking it to the 340 list of alternate object databases, but the code to do so forgot to 341 correctly initialize the list. 342 343 * The code to prepare the working tree side of temporary directory 344 for the "dir-diff" feature forgot that symbolic links need not be 345 copied (or symlinked) to the temporary area, as the code already 346 special cases and overwrites them. Besides, it was wrong to try 347 computing the object name of the target of symbolic link, which may 348 not even exist or may be a directory. 349 350 * A Range: request can be responded with a full response and when 351 asked properly libcurl knows how to strip the result down to the 352 requested range. However, we were hand-crafting a range request 353 and it did not kick in. 354 355 * Having a leftover .idx file without corresponding .pack file in 356 the repository hurts performance; "git gc" learned to prune them. 357 358 * Apple's common crypto implementation of SHA1_Update() does not take 359 more than 4GB at a time, and we now have a compile-time workaround 360 for it. 361 362 * Produce correct "dirty" marker for shell prompts, even when we 363 are on an orphan or an unborn branch. 364 365 * A build without NO_IPv6 used to use gethostbyname() when guessing 366 user's hostname, instead of getaddrinfo() that is used in other 367 codepaths in such a build. 368 369 * The exit code of git-fsck did not reflect some types of errors 370 found in packed objects, which has been corrected. 371 372 * The helper used to iterate over loose object directories to prune 373 stale objects did not closedir() immediately when it is done with a 374 directory--a callback such as the one used for "git prune" may want 375 to do rmdir(), but it would fail on open directory on platforms 376 such as WinXP. 377 378 * "git p4" used to import Perforce CLs that touch only paths outside 379 the client spec as empty commits. It has been corrected to ignore 380 them instead, with a new configuration git-p4.keepEmptyCommits as a 381 backward compatibility knob. 382 383 * The completion script (in contrib/) used to list "git column" 384 (which is not an end-user facing command) as one of the choices 385 (merge 160fcdb sg/completion-no-column later to maint). 386 387 * The error reporting from "git send-email", when SMTP TLS fails, has 388 been improved. 389 (merge 9d60524 jk/send-email-ssl-errors later to maint). 390 391 * When getpwuid() on the system returned NULL (e.g. the user is not 392 in the /etc/passwd file or other uid-to-name mappings), the 393 codepath to find who the user is to record it in the reflog barfed 394 and died. Loosen the check in this codepath, which already accepts 395 questionable ident string (e.g. host part of the e-mail address is 396 obviously bogus), and in general when we operate fmt_ident() function 397 in non-strict mode. 398 (merge 92bcbb9 jk/ident-loosen-getpwuid later to maint). 399 400 * Code clean-up, minor fixes etc.