1Git 2.5 Release Notes 2===================== 3 4Updates since v2.4 5------------------ 6 7Ports 8 9 10UI, Workflows & Features 11 12 * The bash completion script (in contrib/) learned a few options that 13 "git revert" takes. 14 15 * Whitespace breakages in deleted and context lines can also be 16 painted in the output of "git diff" and friends with the new 17 --ws-error-highlight option. 18 19 * List of commands shown by "git help" are grouped along the workflow 20 elements to help early learners. 21 22 * "git p4" now detects the filetype (e.g. binary) correctly even when 23 the files are opened exclusively. 24 25 * git p4 attempts to better handle branches in Perforce. 26 27 * "git p4" learned "--changes-block-size <n>" to read the changes in 28 chunks from Perforce, instead of making one call to "p4 changes" 29 that may trigger "too many rows scanned" error from Perforce. 30 31 * Unlike "$EDITOR" and "$GIT_EDITOR" that can hold the path to the 32 command and initial options (e.g. "/path/to/emacs -nw"), 'git p4' 33 did not let the shell interpolate the contents of the environment 34 variable that name the editor "$P4EDITOR" (and "$EDITOR", too). 35 This release makes it in line with the rest of Git, as well as with 36 Perforce. 37 38 * A new short-hand <branch>@{push} denotes the remote-tracking branch 39 that tracks the branch at the remote the <branch> would be pushed 40 to. 41 42 * "git show-branch --topics HEAD" (with no other arguments) did not 43 do anything interesting. Instead, contrast the given revision 44 against all the local branches by default. 45 46 * A replacement for contrib/workdir/git-new-workdir that does not 47 rely on symbolic links and make sharing of objects and refs safer 48 by making the borrowee and borrowers aware of each other. 49 50 * Tweak the sample "store" backend of the credential helper to honor 51 XDG configuration file locations when specified. 52 53 * A heuristic to help the "git <cmd> <revs> <pathspec>" command line 54 convention to catch mistyped paths is to make sure all the non-rev 55 parameters in the later part of the command line are names of the 56 files in the working tree, but that means "git grep $str -- \*.c" 57 must always be disambiguated with "--", because nobody sane will 58 create a file whose name literally is asterisk-dot-see. Loosen the 59 heuristic to declare that with a wildcard string the user likely 60 meant to give us a pathspec. 61 62 * "git merge FETCH_HEAD" learned that the previous "git fetch" could 63 be to create an Octopus merge, i.e. recording multiple branches 64 that are not marked as "not-for-merge"; this allows us to lose an 65 old style invocation "git merge <msg> HEAD $commits..." in the 66 implementation of "git pull" script; the old style syntax can now 67 be deprecated. 68 69 * Help us to find broken test script that splits the body part of the 70 test by mistaken use of wrong kind of quotes. 71 (merge d93d5d5 jc/test-prereq-validate later to maint). 72 73 * Developer support to automatically detect broken &&-chain in the 74 test scripts is now turned on by default. 75 (merge 92b269f jk/test-chain-lint later to maint). 76 77 * Filter scripts were run with SIGPIPE disabled on the Git side, 78 expecting that they may not read what Git feeds them to filter. 79 We however treated a filter that does not read its input fully 80 before exiting as an error. 81 82 This changes semantics, but arguably in a good way. If a filter 83 can produce its output without consuming its input using whatever 84 magic, we now let it do so, instead of diagnosing it as a 85 programming error. 86 87 * Instead of dying immediately upon failing to obtain a lock, the 88 locking (of refs etc) retries after a short while with backoff. 89 90 * Introduce http.<url>.SSLCipherList configuration variable to tweak 91 the list of cipher suite to be used with libcURL when talking with 92 https:// sites. 93 94 * "git subtree" script (in contrib/) used "echo -n" to produce 95 progress messages in a non-portable way. 96 97 * "git subtree" script (in contrib/) does not have --squash option 98 when pushing, but the documentation and help text pretended as if 99 it did. 100 101 * The Git subcommand completion (in contrib/) listed credential 102 helpers among candidates, which is not something the end user would 103 invoke interactively. 104 105 * The index file can be taught with "update-index --untracked-cache" 106 to optionally remember already seen untracked files, in order to 107 speed up "git status" in a working tree with tons of cruft. 108 109 * "git mergetool" learned to drive WinMerge as a backend. 110 111 * "git upload-pack" that serves "git fetch" can be told to serve 112 commits that are not at the tip of any ref, as long as they are 113 reachable from a ref, with uploadpack.allowReachableSHA1InWant 114 configuration variable. 115 116 * "git cat-file --batch(-check)" learned the "--follow-symlinks" 117 option that follows an in-tree symbolic link when asked about an 118 object via extended SHA-1 syntax, e.g. HEAD:RelNotes that points at 119 Documentation/RelNotes/2.5.0.txt. With the new option, the command 120 behaves as if HEAD:Documentation/RelNotes/2.5.0.txt was given as 121 input instead. 122 123 * "git send-email" learned the alias file format used by the sendmail 124 program (in an abbreviated form). 125 126 127Performance, Internal Implementation, Development Support etc. 128 129 * "unsigned char [20]" used throughout the code to represent object 130 names are being converted into a semi-opaque "struct object_id". 131 This effort is expected to interfere with other topics in flight, 132 but hopefully will give us one extra level of abstraction in the 133 end, when completed. 134 135 * Catch a programmer mistake to feed a pointer not an array to 136 ARRAY_SIZE() macro, by using a couple of GCC extensions. 137 (merge 89c855e ep/do-not-feed-a-pointer-to-array-size later to maint). 138 139 * Some error messages in "git config" were emitted without calling 140 the usual error() facility. 141 142 * When "add--interactive" splits a hunk into two overlapping hunks 143 and then let the user choose only one, it sometimes feeds an 144 incorrect patch text to "git apply". Add tests to demonstrate 145 this. 146 147 I have a slight suspicion that this may be $gmane/87202 coming back 148 and biting us (I seem to have said "let's run with this and see 149 what happens" back then). 150 151 * More line-ending tests. 152 153 * An earlier rewrite to use strbuf_getwholeline() instead of fgets(3) 154 to read packed-refs file revealed that the former is unacceptably 155 inefficient. 156 157 * The refs API uses ref_lock struct which had its own "int fd", even 158 though the same file descriptor was in the lock struct it contains. 159 Clean-up the code to lose this redundant field. 160 161 * Add the "--allow-unknown-type" option to "cat-file" to allow 162 inspecting loose objects of an experimental or a broken type. 163 164 * Many long-running operations show progress eye-candy, even when 165 they are later backgrounded. Hide the eye-candy when the process 166 is sent to the background instead. 167 (merge 9a9a41d lm/squelch-bg-progress later to maint). 168 169 * There was a dead code that used to handle "git pull --tags" and 170 show special-cased error message, which was made irrelevant when 171 the semantics of the option changed back in Git 1.9 days. 172 (merge 19d122b pt/pull-tags-error-diag later to maint). 173 174 * for_each_ref() callback functions were taught to name the objects 175 not with "unsigned char sha1[20]" but with "struct object_id". 176 177 * Error reporting mechanism used in "refs" API has been made more 178 consistent. 179 180 * "git pull" has more test coverage now. 181 182 183Also contains various documentation updates and code clean-ups. 184 185 186Fixes since v2.4 187---------------- 188 189Unless otherwise noted, all the fixes since v2.4 in the maintenance 190track are contained in this release (see the maintenance releases' 191notes for details). 192 193 * Git 2.4 broke setting verbosity and progress levels on "git clone" 194 with native transports. 195 (merge 822f0c4 mh/clone-verbosity-fix later to maint). 196 197 * "git add -e" did not allow the user to abort the operation by 198 killing the editor. 199 (merge cb64800 jk/add-e-kill-editor later to maint). 200 201 * Memory usage of "git index-pack" has been trimmed by tens of 202 per-cent. 203 (merge a78c5b3 nd/slim-index-pack-memory-usage later to maint). 204 205 * "git rev-list --objects $old --not --all" to see if everything that 206 is reachable from $old is already connected to the existing refs 207 was very inefficient. 208 (merge b6e8a3b jk/still-interesting later to maint). 209 210 * "hash-object --literally" introduced in v2.2 was not prepared to 211 take a really long object type name. 212 (merge 1427a7f jc/hash-object later to maint). 213 214 * "git rebase --quiet" was not quite quiet when there is nothing to 215 do. 216 (merge 22946a9 jk/rebase-quiet-noop later to maint). 217 218 * The completion for "log --decorate=" parameter value was incorrect. 219 (merge af16bda sg/complete-decorate-full-not-long later to maint). 220 221 * "filter-branch" corrupted commit log message that ends with an 222 incomplete line on platforms with some "sed" implementations that 223 munge such a line. Work it around by avoiding to use "sed". 224 (merge df06201 jk/filter-branch-use-of-sed-on-incomplete-line later to maint). 225 226 * "git daemon" fails to build from the source under NO_IPV6 227 configuration (regression in 2.4). 228 (merge d358f77 jc/daemon-no-ipv6-for-2.4.1 later to maint). 229 230 * Some time ago, "git blame" (incorrectly) lost the convert_to_git() 231 call when synthesizing a fake "tip" commit that represents the 232 state in the working tree, which broke folks who record the history 233 with LF line ending to make their project portable across platforms 234 while terminating lines in their working tree files with CRLF for 235 their platform. 236 (merge 4bf256d tb/blame-resurrect-convert-to-git later to maint). 237 238 * We avoid setting core.worktree when the repository location is the 239 ".git" directory directly at the top level of the working tree, but 240 the code misdetected the case in which the working tree is at the 241 root level of the filesystem (which arguably is a silly thing to 242 do, but still valid). 243 (merge 84ccad8 jk/init-core-worktree-at-root later to maint). 244 245 * "git commit --date=now" or anything that relies on approxidate lost 246 the daylight-saving-time offset. 247 (merge f6e6362 jc/epochtime-wo-tz later to maint). 248 249 * Access to objects in repositories that borrow from another one on a 250 slow NFS server unnecessarily got more expensive due to recent code 251 becoming more cautious in a naive way not to lose objects to pruning. 252 (merge ee1c6c3 jk/prune-mtime later to maint). 253 254 * The codepaths that read .gitignore and .gitattributes files have been 255 taught that these files encoded in UTF-8 may have UTF-8 BOM marker at 256 the beginning; this makes it in line with what we do for configuration 257 files already. 258 (merge 27547e5 cn/bom-in-gitignore later to maint). 259 260 * a few helper scripts in the test suite did not report errors 261 correctly. 262 (merge de248e9 ep/fix-test-lib-functions-report later to maint). 263 264 * The default $HOME/.gitconfig file created upon "git config --global" 265 that edits it had incorrectly spelled user.name and user.email 266 entries in it. 267 (merge 7e11052 oh/fix-config-default-user-name-section later to maint). 268 269 * "git cat-file bl $blob" failed to barf even though there is no 270 object type that is "bl". 271 (merge b7994af jk/type-from-string-gently later to maint). 272 273 * The usual "git diff" when seeing a file turning into a directory 274 showed a patchset to remove the file and create all files in the 275 directory, but "git diff --no-index" simply refused to work. Also, 276 when asked to compare a file and a directory, imitate POSIX "diff" 277 and compare the file with the file with the same name in the 278 directory, instead of refusing to run. 279 (merge 0615173 jc/diff-no-index-d-f later to maint). 280 281 * "git rebase -i" moved the "current" command from "todo" to "done" a 282 bit too prematurely, losing a step when a "pick" did not even start. 283 (merge 8cbc57c ph/rebase-i-redo later to maint). 284 285 * The connection initiation code for "ssh" transport tried to absorb 286 differences between the stock "ssh" and Putty-supplied "plink" and 287 its derivatives, but the logic to tell that we are using "plink" 288 variants were too loose and falsely triggered when "plink" appeared 289 anywhere in the path (e.g. "/home/me/bin/uplink/ssh"). 290 (merge baaf233 bc/connect-plink later to maint). 291 292 * "git stash pop/apply" forgot to make sure that not just the working 293 tree is clean but also the index is clean. The latter is important 294 as a stash application can conflict and the index will be used for 295 conflict resolution. 296 (merge ed178ef jk/stash-require-clean-index later to maint). 297 298 * We have prepended $GIT_EXEC_PATH and the path "git" is installed in 299 (typically "/usr/bin") to $PATH when invoking subprograms and hooks 300 for almost eternity, but the original use case the latter tried to 301 support was semi-bogus (i.e. install git to /opt/foo/git and run it 302 without having /opt/foo on $PATH), and more importantly it has 303 become less and less relevant as Git grew more mainstream (i.e. the 304 users would _want_ to have it on their $PATH). Stop prepending the 305 path in which "git" is installed to users' $PATH, as that would 306 interfere the command search order people depend on (e.g. they may 307 not like versions of programs that are unrelated to Git in /usr/bin 308 and want to override them by having different ones in /usr/local/bin 309 and have the latter directory earlier in their $PATH). 310 (merge a0b4507 jk/git-no-more-argv0-path-munging later to maint). 311 312 * core.excludesfile (defaulting to $XDG_HOME/git/ignore) is supposed 313 to be overridden by repository-specific .git/info/exclude file, but 314 the order was swapped from the beginning. This belatedly fixes it. 315 (merge 099d2d8 jc/gitignore-precedence later to maint). 316 317 * After "git add -N", the path appeared in output of "git diff HEAD" 318 and "git diff --cached HEAD", leading "git status" to classify it 319 as "Changes to be committed". Such a path, however, is not yet to 320 be scheduled to be committed. "git diff" showed the change to the 321 path as modification, not as a "new file", in the header of its 322 output. 323 324 Treat such paths as "yet to be added to the index but Git already 325 know about them"; "git diff HEAD" and "git diff --cached HEAD" 326 should not talk about them, and "git diff" should show them as new 327 files yet to be added to the index. 328 (merge d95d728 nd/diff-i-t-a later to maint). 329 330 * There was a commented-out (instead of being marked to expect 331 failure) test that documented a breakage that was fixed since the 332 test was written; turn it into a proper test. 333 (merge 66d2e04 sb/t1020-cleanup later to maint). 334 335 * The "log --decorate" enhancement in Git 2.4 that shows the commit 336 at the tip of the current branch e.g. "HEAD -> master", did not 337 work with --decorate=full. 338 (merge 429ad20 mg/log-decorate-HEAD later to maint). 339 340 * The ref API did not handle cases where 'refs/heads/xyzzy/frotz' is 341 removed at the same time as 'refs/heads/xyzzy' is added (or vice 342 versa) very well. 343 (merge c628edf mh/ref-directory-file later to maint). 344 345 * Multi-ref transaction support we merged a few releases ago 346 unnecessarily kept many file descriptors open, risking to fail with 347 resource exhaustion. This is for 2.4.x track. 348 (merge 185ce3a mh/write-refs-sooner-2.4 later to maint). 349 350 * "git bundle verify" did not diagnose extra parameters on the 351 command line. 352 (merge 7886cfa ps/bundle-verify-arg later to maint). 353 354 * Various documentation mark-up fixes to make the output more 355 consistent in general and also make AsciiDoctor (an alternative 356 formatter) happier. 357 (merge d0258b9 jk/asciidoc-markup-fix later to maint). 358 (merge ad3967a jk/stripspace-asciidoctor-fix later to maint). 359 (merge 975e382 ja/tutorial-asciidoctor-fix later to maint). 360 361 * The code to read pack-bitmap wanted to allocate a few hundred 362 pointers to a structure, but by mistake allocated and leaked memory 363 enough to hold that many actual structures. Correct the allocation 364 size and also have it on stack, as it is small enough. 365 (merge 599dc76 rs/plug-leak-in-pack-bitmaps later to maint). 366 367 * The pull.ff configuration was supposed to override the merge.ff 368 configuration, but it didn't. 369 (merge db9bb28 pt/pull-ff-vs-merge-ff later to maint). 370 371 * "git pull --log" and "git pull --no-log" worked as expected, but 372 "git pull --log=20" did not. 373 (merge 5061a44 pt/pull-log-n later to maint). 374 375 * "git rerere forget" in a repository without rerere enabled gave a 376 cryptic error message; it should be a silent no-op instead. 377 (merge 0544574 jk/rerere-forget-check-enabled later to maint). 378 379 * "git rebase -i" fired post-rewrite hook when it shouldn't (namely, 380 when it was told to stop sequencing with 'exec' insn). 381 (merge 141ff8f mm/rebase-i-post-rewrite-exec later to maint). 382 383 * Clarify that "log --raw" and "log --format=raw" are unrelated 384 concepts. 385 (merge 92de921 mm/log-format-raw-doc later to maint). 386 387 * Make "git stash something --help" error out, so that users can 388 safely say "git stash drop --help". 389 (merge 5ba2831 jk/stash-options later to maint). 390 391 * The clean/smudge interface did not work well when filtering an 392 empty contents (failed and then passed the empty input through). 393 It can be argued that a filter that produces anything but empty for 394 an empty input is nonsense, but if the user wants to do strange 395 things, then why not? 396 (merge f6a1e1e jh/filter-empty-contents later to maint). 397 398 * Communication between the HTTP server and http_backend process can 399 lead to a dead-lock when relaying a large ref negotiation request. 400 Diagnose the situation better, and mitigate it by reading such a 401 request first into core (to a reasonable limit). 402 (merge 636614f jk/http-backend-deadlock later to maint). 403 404 * "git clean pathspec..." tried to lstat(2) and complain even for 405 paths outside the given pathspec. 406 (merge 838d6a9 dt/clean-pathspec-filter-then-lstat later to maint). 407 408 * Recent "git prune" traverses young unreachable objects to safekeep 409 old objects in the reachability chain from them, which sometimes 410 caused error messages that are unnecessarily alarming. 411 (merge ce4e7b2 jk/squelch-missing-link-warning-for-unreachable later to maint). 412 413 * The configuration reader/writer uses mmap(2) interface to access 414 the files; when we find a directory, it barfed with "Out of memory?". 415 (merge 9ca0aaf jk/diagnose-config-mmap-failure later to maint). 416 417 * "color.diff.plain" was a misnomer; give it 'color.diff.context' as 418 a more logical synonym. 419 (merge 8dbf3eb jk/color-diff-plain-is-context later to maint). 420 421 * The setup code used to die when core.bare and core.worktree are set 422 inconsistently, even for commands that do not need working tree. 423 (merge fada767 jk/die-on-bogus-worktree-late later to maint). 424 425 * Code cleanups and documentation updates. 426 (merge 0269f96 mm/usage-log-l-can-take-regex later to maint). 427 (merge 64f2589 nd/t1509-chroot-test later to maint). 428 (merge d201a1e sb/test-bitmap-free-at-end later to maint). 429 (merge 05bfc7d sb/line-log-plug-pairdiff-leak later to maint). 430 (merge 846e5df pt/xdg-config-path later to maint). 431 (merge 1154aa4 jc/plug-fmt-merge-msg-leak later to maint). 432 (merge 319b678 jk/sha1-file-reduce-useless-warnings later to maint). 433 (merge 9a35c14 fg/document-commit-message-stripping later to maint). 434 (merge bbf431c ps/doc-packfile-vs-pack-file later to maint). 435 (merge 309a9e3 jk/skip-http-tests-under-no-curl later to maint). 436 (merge ccd593c dl/branch-error-message later to maint). 437 (merge 22570b6 rs/janitorial later to maint). 438 (merge 5c2a581 mc/commit-doc-grammofix later to maint). 439 (merge ce41720 ah/usage-strings later to maint). 440 (merge e6a268c sb/glossary-submodule later to maint). 441 (merge ec48a76 sb/submodule-doc-intro later to maint). 442 (merge 14f8b9b jk/clone-dissociate later to maint). 443 (merge 055c7e9 sb/pack-protocol-mention-smart-http later to maint). 444 (merge 7c37a5d jk/make-fix-dependencies later to maint). 445 (merge fc0aa39 sg/merge-summary-config later to maint).