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