1Git 2.5 Release Notes 2===================== 3 4Updates since v2.4 5------------------ 6 7Ports 8 9 10UI, Workflows & Features 11 12 * List of commands shown by "git help" are grouped along the workflow 13 elements to help early learners. 14 15 * "git p4" now detects the filetype (e.g. binary) correctly even when 16 the files are opened exclusively. 17 18 * git p4 attempts to better handle branches in Perforce. 19 20 * "git p4" learned "--changes-block-size <n>" to read the changes in 21 chunks from Perforce, instead of making one call to "p4 changes" 22 that may trigger "too many rows scanned" error from Perforce. 23 24 * Unlike "$EDITOR" and "$GIT_EDITOR" that can hold the path to the 25 command and initial options (e.g. "/path/to/emacs -nw"), 'git p4' 26 did not let the shell interpolate the contents of the environment 27 variable that name the editor "$P4EDITOR" (and "$EDITOR", too). 28 This release makes it in line with the rest of Git, as well as with 29 Perforce. 30 31 * A new short-hand <branch>@{push} denotes the remote-tracking branch 32 that tracks the branch at the remote the <branch> would be pushed 33 to. 34 35 * "git show-branch --topics HEAD" (with no other arguments) did not 36 do anything interesting. Instead, contrast the given revision 37 against all the local branches by default. 38 39 * A replacement for contrib/workdir/git-new-workdir that does not 40 rely on symbolic links and make sharing of objects and refs safer 41 by making the borrowee and borrowers aware of each other. 42 43 * Tweak the sample "store" backend of the credential helper to honor 44 XDG configuration file locations when specified. 45 46 * A heuristic to help the "git <cmd> <revs> <pathspec>" command line 47 convention to catch mistyped paths is to make sure all the non-rev 48 parameters in the later part of the command line are names of the 49 files in the working tree, but that means "git grep $str -- \*.c" 50 must always be disambiguated with "--", because nobody sane will 51 create a file whose name literally is asterisk-dot-see. Loosen the 52 heuristic to declare that with a wildcard string the user likely 53 meant to give us a pathspec. 54 55 * "git merge FETCH_HEAD" learned that the previous "git fetch" could 56 be to create an Octopus merge, i.e. recording multiple branches 57 that are not marked as "not-for-merge"; this allows us to lose an 58 old style invocation "git merge <msg> HEAD $commits..." in the 59 implementation of "git pull" script; the old style syntax can now 60 be deprecated. 61 62 * Help us to find broken test script that splits the body part of the 63 test by mistaken use of wrong kind of quotes. 64 (merge d93d5d5 jc/test-prereq-validate later to maint). 65 66 * Developer support to automatically detect broken &&-chain in the 67 test scripts is now turned on by default. 68 (merge 92b269f jk/test-chain-lint later to maint). 69 70 * Filter scripts were run with SIGPIPE disabled on the Git side, 71 expecting that they may not read what Git feeds them to filter. 72 We however treated a filter that does not read its input fully 73 before exiting as an error. 74 75 This changes semantics, but arguably in a good way. If a filter 76 can produce its output without consuming its input using whatever 77 magic, we now let it do so, instead of diagnosing it as a 78 programming error. 79 80 * Instead of dying immediately upon failing to obtain a lock, the 81 locking (of refs etc) retries after a short while with backoff. 82 83 * Introduce http.<url>.SSLCipherList configuration variable to tweak 84 the list of cipher suite to be used with libcURL when talking with 85 https:// sites. 86 87 * "git subtree" script (in contrib/) used "echo -n" to produce 88 progress messages in a non-portable way. 89 90 * "git subtree" script (in contrib/) does not have --squash option 91 when pushing, but the documentation and help text pretended as if 92 it did. 93 94 * The Git subcommand completion (in contrib/) listed credential 95 helpers among candidates, which is not something the end user would 96 invoke interactively. 97 98 * The index file can be taught with "update-index --untracked-cache" 99 to optionally remember already seen untracked files, in order to 100 speed up "git status" in a working tree with tons of cruft. 101 102 * "git mergetool" learned to drive WinMerge as a backend. 103 104 * "git upload-pack" that serves "git fetch" can be told to serve 105 commits that are not at the tip of any ref, as long as they are 106 reachable from a ref, with uploadpack.allowReachableSHA1InWant 107 configuration variable. 108 109 * "git cat-file --batch(-check)" learned the "--follow-symlinks" 110 option that follows an in-tree symbolic link when asked about an 111 object via extended SHA-1 syntax, e.g. HEAD:RelNotes that points at 112 Documentation/RelNotes/2.5.0.txt. With the new option, the command 113 behaves as if HEAD:Documentation/RelNotes/2.5.0.txt was given as 114 input instead. 115 116 117Performance, Internal Implementation, Development Support etc. 118 119 * "unsigned char [20]" used throughout the code to represent object 120 names are being converted into a semi-opaque "struct object_id". 121 This effort is expected to interfere with other topics in flight, 122 but hopefully will give us one extra level of abstraction in the 123 end, when completed. 124 125 * Catch a programmer mistake to feed a pointer not an array to 126 ARRAY_SIZE() macro, by using a couple of GCC extensions. 127 (merge 89c855e ep/do-not-feed-a-pointer-to-array-size later to maint). 128 129 * Some error messages in "git config" were emitted without calling 130 the usual error() facility. 131 132 * When "add--interactive" splits a hunk into two overlapping hunks 133 and then let the user choose only one, it sometimes feeds an 134 incorrect patch text to "git apply". Add tests to demonstrate 135 this. 136 137 I have a slight suspicion that this may be $gmane/87202 coming back 138 and biting us (I seem to have said "let's run with this and see 139 what happens" back then). 140 141 * More line-ending tests. 142 143 * An earlier rewrite to use strbuf_getwholeline() instead of fgets(3) 144 to read packed-refs file revealed that the former is unacceptably 145 inefficient. 146 147 * The refs API uses ref_lock struct which had its own "int fd", even 148 though the same file descriptor was in the lock struct it contains. 149 Clean-up the code to lose this redundant field. 150 151 * Add the "--allow-unknown-type" option to "cat-file" to allow 152 inspecting loose objects of an experimental or a broken type. 153 154 * Many long-running operations show progress eye-candy, even when 155 they are later backgrounded. Hide the eye-candy when the process 156 is sent to the background instead. 157 (merge 9a9a41d lm/squelch-bg-progress later to maint). 158 159 * There was a dead code that used to handle "git pull --tags" and 160 show special-cased error message, which was made irrelevant when 161 the semantics of the option changed back in Git 1.9 days. 162 (merge 19d122b pt/pull-tags-error-diag later to maint). 163 164 * for_each_ref() callback functions were taught to name the objects 165 not with "unsigned char sha1[20]" but with "struct object_id". 166 167 168Also contains various documentation updates and code clean-ups. 169 170 171Fixes since v2.4 172---------------- 173 174Unless otherwise noted, all the fixes since v2.4 in the maintenance 175track are contained in this release (see the maintenance releases' 176notes for details). 177 178 * Git 2.4 broke setting verbosity and progress levels on "git clone" 179 with native transports. 180 (merge 822f0c4 mh/clone-verbosity-fix later to maint). 181 182 * "git add -e" did not allow the user to abort the operation by 183 killing the editor. 184 (merge cb64800 jk/add-e-kill-editor later to maint). 185 186 * Memory usage of "git index-pack" has been trimmed by tens of 187 per-cent. 188 (merge c6458e6 nd/slim-index-pack-memory-usage later to maint). 189 190 * "git rev-list --objects $old --not --all" to see if everything that 191 is reachable from $old is already connected to the existing refs 192 was very inefficient. 193 (merge b6e8a3b jk/still-interesting later to maint). 194 195 * "hash-object --literally" introduced in v2.2 was not prepared to 196 take a really long object type name. 197 (merge 1427a7f jc/hash-object later to maint). 198 199 * "git rebase --quiet" was not quite quiet when there is nothing to 200 do. 201 (merge 22946a9 jk/rebase-quiet-noop later to maint). 202 203 * The completion for "log --decorate=" parameter value was incorrect. 204 (merge af16bda sg/complete-decorate-full-not-long later to maint). 205 206 * "filter-branch" corrupted commit log message that ends with an 207 incomplete line on platforms with some "sed" implementations that 208 munge such a line. Work it around by avoiding to use "sed". 209 (merge df06201 jk/filter-branch-use-of-sed-on-incomplete-line later to maint). 210 211 * "git daemon" fails to build from the source under NO_IPV6 212 configuration (regression in 2.4). 213 (merge d358f77 jc/daemon-no-ipv6-for-2.4.1 later to maint). 214 215 * Some time ago, "git blame" (incorrectly) lost the convert_to_git() 216 call when synthesizing a fake "tip" commit that represents the 217 state in the working tree, which broke folks who record the history 218 with LF line ending to make their project portable across platforms 219 while terminating lines in their working tree files with CRLF for 220 their platform. 221 (merge 4bf256d tb/blame-resurrect-convert-to-git later to maint). 222 223 * We avoid setting core.worktree when the repository location is the 224 ".git" directory directly at the top level of the working tree, but 225 the code misdetected the case in which the working tree is at the 226 root level of the filesystem (which arguably is a silly thing to 227 do, but still valid). 228 (merge 84ccad8 jk/init-core-worktree-at-root later to maint). 229 230 * "git commit --date=now" or anything that relies on approxidate lost 231 the daylight-saving-time offset. 232 (merge f6e6362 jc/epochtime-wo-tz later to maint). 233 234 * Access to objects in repositories that borrow from another one on a 235 slow NFS server unnecessarily got more expensive due to recent code 236 becoming more cautious in a naive way not to lose objects to pruning. 237 (merge ee1c6c3 jk/prune-mtime later to maint). 238 239 * The codepaths that read .gitignore and .gitattributes files have been 240 taught that these files encoded in UTF-8 may have UTF-8 BOM marker at 241 the beginning; this makes it in line with what we do for configuration 242 files already. 243 (merge 27547e5 cn/bom-in-gitignore later to maint). 244 245 * a few helper scripts in the test suite did not report errors 246 correctly. 247 (merge de248e9 ep/fix-test-lib-functions-report later to maint). 248 249 * The default $HOME/.gitconfig file created upon "git config --global" 250 that edits it had incorrectly spelled user.name and user.email 251 entries in it. 252 (merge 7e11052 oh/fix-config-default-user-name-section later to maint). 253 254 * "git cat-file bl $blob" failed to barf even though there is no 255 object type that is "bl". 256 (merge b7994af jk/type-from-string-gently later to maint). 257 258 * The usual "git diff" when seeing a file turning into a directory 259 showed a patchset to remove the file and create all files in the 260 directory, but "git diff --no-index" simply refused to work. Also, 261 when asked to compare a file and a directory, imitate POSIX "diff" 262 and compare the file with the file with the same name in the 263 directory, instead of refusing to run. 264 (merge 0615173 jc/diff-no-index-d-f later to maint). 265 266 * "git rebase -i" moved the "current" command from "todo" to "done" a 267 bit too prematurely, losing a step when a "pick" did not even start. 268 (merge 8cbc57c ph/rebase-i-redo later to maint). 269 270 * The connection initiation code for "ssh" transport tried to absorb 271 differences between the stock "ssh" and Putty-supplied "plink" and 272 its derivatives, but the logic to tell that we are using "plink" 273 variants were too loose and falsely triggered when "plink" appeared 274 anywhere in the path (e.g. "/home/me/bin/uplink/ssh"). 275 (merge baaf233 bc/connect-plink later to maint). 276 277 * "git stash pop/apply" forgot to make sure that not just the working 278 tree is clean but also the index is clean. The latter is important 279 as a stash application can conflict and the index will be used for 280 conflict resolution. 281 (merge ed178ef jk/stash-require-clean-index later to maint). 282 283 * We have prepended $GIT_EXEC_PATH and the path "git" is installed in 284 (typically "/usr/bin") to $PATH when invoking subprograms and hooks 285 for almost eternity, but the original use case the latter tried to 286 support was semi-bogus (i.e. install git to /opt/foo/git and run it 287 without having /opt/foo on $PATH), and more importantly it has 288 become less and less relevant as Git grew more mainstream (i.e. the 289 users would _want_ to have it on their $PATH). Stop prepending the 290 path in which "git" is installed to users' $PATH, as that would 291 interfere the command search order people depend on (e.g. they may 292 not like versions of programs that are unrelated to Git in /usr/bin 293 and want to override them by having different ones in /usr/local/bin 294 and have the latter directory earlier in their $PATH). 295 (merge a0b4507 jk/git-no-more-argv0-path-munging later to maint). 296 297 * core.excludesfile (defaulting to $XDG_HOME/git/ignore) is supposed 298 to be overridden by repository-specific .git/info/exclude file, but 299 the order was swapped from the beginning. This belatedly fixes it. 300 (merge 099d2d8 jc/gitignore-precedence later to maint). 301 302 * After "git add -N", the path appeared in output of "git diff HEAD" 303 and "git diff --cached HEAD", leading "git status" to classify it 304 as "Changes to be committed". Such a path, however, is not yet to 305 be scheduled to be committed. "git diff" showed the change to the 306 path as modification, not as a "new file", in the header of its 307 output. 308 309 Treat such paths as "yet to be added to the index but Git already 310 know about them"; "git diff HEAD" and "git diff --cached HEAD" 311 should not talk about them, and "git diff" should show them as new 312 files yet to be added to the index. 313 (merge d95d728 nd/diff-i-t-a later to maint). 314 315 * There was a commented-out (instead of being marked to expect 316 failure) test that documented a breakage that was fixed since the 317 test was written; turn it into a proper test. 318 (merge 66d2e04 sb/t1020-cleanup later to maint). 319 320 * The "log --decorate" enhancement in Git 2.4 that shows the commit 321 at the tip of the current branch e.g. "HEAD -> master", did not 322 work with --decorate=full. 323 (merge 429ad20 mg/log-decorate-HEAD later to maint). 324 325 * The ref API did not handle cases where 'refs/heads/xyzzy/frotz' is 326 removed at the same time as 'refs/heads/xyzzy' is added (or vice 327 versa) very well. 328 (merge c628edf mh/ref-directory-file later to maint). 329 330 * Multi-ref transaction support we merged a few releases ago 331 unnecessarily kept many file descriptors open, risking to fail with 332 resource exhaustion. This is for 2.4.x track. 333 (merge 185ce3a mh/write-refs-sooner-2.4 later to maint). 334 335 * "git bundle verify" did not diagnose extra parameters on the 336 command line. 337 (merge 7886cfa ps/bundle-verify-arg later to maint). 338 339 * Various documentation mark-up fixes to make the output more 340 consistent in general and also make AsciiDoctor (an alternative 341 formatter) happier. 342 (merge d0258b9 jk/asciidoc-markup-fix later to maint). 343 (merge ad3967a jk/stripspace-asciidoctor-fix later to maint). 344 (merge 975e382 ja/tutorial-asciidoctor-fix later to maint). 345 346 * The code to read pack-bitmap wanted to allocate a few hundred 347 pointers to a structure, but by mistake allocated and leaked memory 348 enough to hold that many actual structures. Correct the allocation 349 size and also have it on stack, as it is small enough. 350 (merge 599dc76 rs/plug-leak-in-pack-bitmaps later to maint). 351 352 * The pull.ff configuration was supposed to override the merge.ff 353 configuration, but it didn't. 354 (merge db9bb28 pt/pull-ff-vs-merge-ff later to maint). 355 356 * "git pull --log" and "git pull --no-log" worked as expected, but 357 "git pull --log=20" did not. 358 (merge 5061a44 pt/pull-log-n later to maint). 359 360 * "git rerere forget" in a repository without rerere enabled gave a 361 cryptic error message; it should be a silent no-op instead. 362 (merge 0544574 jk/rerere-forget-check-enabled later to maint). 363 364 * "git rebase -i" fired post-rewrite hook when it shouldn't (namely, 365 when it was told to stop sequencing with 'exec' insn). 366 (merge 141ff8f mm/rebase-i-post-rewrite-exec later to maint). 367 368 * Clarify that "log --raw" and "log --format=raw" are unrelated 369 concepts. 370 (merge 92de921 mm/log-format-raw-doc later to maint). 371 372 * Make "git stash something --help" error out, so that users can 373 safely say "git stash drop --help". 374 (merge 5ba2831 jk/stash-options later to maint). 375 376 * The clean/smudge interface did not work well when filtering an 377 empty contents (failed and then passed the empty input through). 378 It can be argued that a filter that produces anything but empty for 379 an empty input is nonsense, but if the user wants to do strange 380 things, then why not? 381 (merge f6a1e1e jh/filter-empty-contents later to maint). 382 383 * Communication between the HTTP server and http_backend process can 384 lead to a dead-lock when relaying a large ref negotiation request. 385 Diagnose the situation better, and mitigate it by reading such a 386 request first into core (to a reasonable limit). 387 (merge 636614f jk/http-backend-deadlock later to maint). 388 389 * "git clean pathspec..." tried to lstat(2) and complain even for 390 paths outside the given pathspec. 391 (merge 838d6a9 dt/clean-pathspec-filter-then-lstat later to maint). 392 393 * Code cleanups and documentation updates. 394 (merge 0269f96 mm/usage-log-l-can-take-regex later to maint). 395 (merge 64f2589 nd/t1509-chroot-test later to maint). 396 (merge d201a1e sb/test-bitmap-free-at-end later to maint). 397 (merge 05bfc7d sb/line-log-plug-pairdiff-leak later to maint). 398 (merge 846e5df pt/xdg-config-path later to maint). 399 (merge 1154aa4 jc/plug-fmt-merge-msg-leak later to maint). 400 (merge 319b678 jk/sha1-file-reduce-useless-warnings later to maint). 401 (merge 9a35c14 fg/document-commit-message-stripping later to maint). 402 (merge bbf431c ps/doc-packfile-vs-pack-file later to maint). 403 (merge 309a9e3 jk/skip-http-tests-under-no-curl later to maint). 404 (merge ccd593c dl/branch-error-message later to maint). 405 (merge 22570b6 rs/janitorial later to maint). 406 (merge 5c2a581 mc/commit-doc-grammofix later to maint). 407 (merge ce41720 ah/usage-strings later to maint).