1Git 2.16 Release Notes 2====================== 3 4Backward compatibility notes and other notable changes. 5 6 * Use of an empty string as a pathspec element that is used for 7 'everything matches' is now an error. 8 9 10Updates since v2.15 11------------------- 12 13UI, Workflows & Features 14 15 * An empty string as a pathspec element that means "everything" 16 i.e. 'git add ""', is now illegal. We started this by first 17 deprecating and warning a pathspec that has such an element in 18 2.11 (Nov 2016). 19 20 * A hook script that is set unexecutable is simply ignored. Git 21 notifies when such a file is ignored, unless the message is 22 squelched via advice.ignoredHook configuration. 23 24 * "git pull" has been taught to accept "--[no-]signoff" option and 25 pass it down to "git merge". 26 27 * The "--push-option=<string>" option to "git push" now defaults to a 28 list of strings configured via push.pushOption variable. 29 30 * "gitweb" checks if a directory is searchable with Perl's "-x" 31 operator, which can be enhanced by using "filetest 'access'" 32 pragma, which now we do. 33 34 * "git stash save" has been deprecated in favour of "git stash push". 35 36 * The set of paths output from "git status --ignored" was tied 37 closely with its "--untracked=<mode>" option, but now it can be 38 controlled more flexibly. Most notably, a directory that is 39 ignored because it is listed to be ignored in the ignore/exclude 40 mechanism can be handled differently from a directory that ends up 41 to be ignored only because all files in it are ignored. 42 43 * The remote-helper for talking to MediaWiki has been updated to 44 truncate an overlong pagename so that ".mw" suffix can still be 45 added. 46 47 * The remote-helper for talking to MediaWiki has been updated to 48 work with mediawiki namespaces. 49 50 * The "--format=..." option "git for-each-ref" takes learned to show 51 the name of the 'remote' repository and the ref at the remote side 52 that is affected for 'upstream' and 'push' via "%(push:remotename)" 53 and friends. 54 55 * Doc and message updates to teach users "bisect view" is a synonym 56 for "bisect visualize". 57 58 * "git bisect run" that did not specify any command to run used to go 59 ahead and treated all commits to be tested as 'good'. This has 60 been corrected by making the command error out. 61 62 * The SubmittingPatches document has been converted to produce an 63 HTML version via AsciiDoc/Asciidoctor. 64 65 * We learned to talk to watchman to speed up "git status" and other 66 operations that need to see which paths have been modified. 67 68 * The "diff" family of commands learned to ignore differences in 69 carriage return at the end of line. 70 71 * Places that know about "sendemail.to", like documentation and shell 72 completion (in contrib/) have been taught about "sendemail.tocmd", 73 too. 74 75 * "git add --renormalize ." is a new and safer way to record the fact 76 that you are correcting the end-of-line convention and other 77 "convert_to_git()" glitches in the in-repository data. 78 79 * "git branch" and "git checkout -b" are now forbidden from creating 80 a branch whose name is "HEAD". 81 82 * "git branch --list" learned to show its output through the pager by 83 default when the output is going to a terminal, which is controlled 84 by the pager.branch configuration variable. This is similar to a 85 recent change to "git tag --list". 86 87 * "git grep -W", "git diff -W" and their friends learned a heuristic 88 to extend a pre-context beyond the line that matches the "function 89 pattern" (aka "diff.*.xfuncname") to include a comment block, if 90 exists, that immediately precedes it. 91 92 * "git config --expiry-date gc.reflogexpire" can read "2.weeks" from 93 the configuration and report it as a timestamp, just like "--int" 94 would read "1k" and report 1024, to help consumption by scripts. 95 96 * The shell completion (in contrib/) learned that "git pull" can take 97 the "--autostash" option. 98 99 * The tagnames "git log --decorate" uses to annotate the commits can 100 now be limited to subset of available refs with the two additional 101 options, --decorate-refs[-exclude]=<pattern>. 102 103 * "git grep" compiled with libpcre2 sometimes triggered a segfault, 104 which is being fixed. 105 106 * "git send-email" tries to see if the sendmail program is available 107 in /usr/lib and /usr/sbin; extend the list of locations to be 108 checked to also include directories on $PATH. 109 110 * "git diff" learned, "--anchored", a variant of the "--patience" 111 algorithm, to which the user can specify which 'unique' line to be 112 used as anchoring points. 113 114 * The way "git worktree add" determines what branch to create from 115 where and checkout in the new worktree has been updated a bit. 116 117 * Ancient part of codebase still shows dots after an abbreviated 118 object name just to show that it is not a full object name, but 119 these ellipses are confusing to people who newly discovered Git 120 who are used to seeing abbreviated object names and find them 121 confusing with the range syntax. 122 123 124Performance, Internal Implementation, Development Support etc. 125 126 * An earlier update made it possible to use an on-stack in-core 127 lockfile structure (as opposed to having to deliberately leak an 128 on-heap one). Many codepaths have been updated to take advantage 129 of this new facility. 130 131 * Calling cmd_foo() as if it is a general purpose helper function is 132 a no-no. Correct two instances of such to set an example. 133 134 * We try to see if somebody runs our test suite with a shell that 135 does not support "local" like bash/dash does. 136 137 * An early part of piece-by-piece rewrite of "git bisect" in C. 138 139 * GSoC to piece-by-piece rewrite "git submodule" in C. 140 141 * Optimize the code to find shortest unique prefix of object names. 142 143 * Pathspec-limited revision traversal was taught not to keep finding 144 unneeded differences once it knows two trees are different inside 145 given pathspec. 146 147 * Conversion from uchar[20] to struct object_id continues. 148 149 * Code cleanup. 150 151 * A single-word "unsigned flags" in the diff options is being split 152 into a structure with many bitfields. 153 154 * TravisCI build updates. 155 156 * Parts of a test to drive the long-running content filter interface 157 has been split into its own module, hopefully to eventually become 158 reusable. 159 160 * Drop (perhaps overly cautious) sanity check before using the index 161 read from the filesystem at runtime. 162 163 * The build procedure has been taught to avoid some unnecessary 164 instability in the build products. 165 166 * A new mechanism to upgrade the wire protocol in place is proposed 167 and demonstrated that it works with the older versions of Git 168 without harming them. 169 170 * An infrastructure to define what hash function is used in Git is 171 introduced, and an effort to plumb that throughout various 172 codepaths has been started. 173 174 * The code to iterate over loose object files got optimized. 175 176 * An internal function that was left for backward compatibility has 177 been removed, as there is no remaining callers. 178 179 * Historically, the diff machinery for rename detection had a 180 hardcoded limit of 32k paths; this is being lifted to allow users 181 trade cycles with a (possibly) easier to read result. 182 183 * The tracing infrastructure has been optimized for cases where no 184 tracing is requested. 185 186Also contains various documentation updates and code clean-ups. 187 188 189Fixes since v2.15 190----------------- 191 192 * "auto" as a value for the columnar output configuration ought to 193 judge "is the output consumed by humans?" with the same criteria as 194 "auto" for coloured output configuration, i.e. either the standard 195 output stream is going to tty, or a pager is in use. We forgot the 196 latter, which has been fixed. 197 198 * The experimental "color moved lines differently in diff output" 199 feature was buggy around "ignore whitespace changes" edges, which 200 has been corrected. 201 202 * Instead of using custom line comparison and hashing functions to 203 implement "moved lines" coloring in the diff output, use the pair 204 of these functions from lower-layer xdiff/ code. 205 206 * Some codepaths did not check for errors when asking what branch the 207 HEAD points at, which have been fixed. 208 209 * "git commit", after making a commit, did not check for errors when 210 asking on what branch it made the commit, which has been corrected. 211 212 * "git status --ignored -u" did not stop at a working tree of a 213 separate project that is embedded in an ignored directory and 214 listed files in that other project, instead of just showing the 215 directory itself as ignored. 216 217 * A broken access to object databases in recent update to "git grep 218 --recurse-submodules" has been fixed. 219 220 * A recent regression in "git rebase -i" that broke execution of git 221 commands from subdirectories via "exec" instruction has been fixed. 222 223 * A (possibly flakey) test fix. 224 225 * "git check-ref-format --branch @{-1}" bit a "BUG()" when run 226 outside a repository for obvious reasons; clarify the documentation 227 and make sure we do not even try to expand the at-mark magic in 228 such a case, but still call the validation logic for branch names. 229 230 * "git fetch --recurse-submodules" now knows that submodules can be 231 moved around in the superproject in addition to getting updated, 232 and finds the ones that need to be fetched accordingly. 233 234 * Command line completion (in contrib/) update. 235 236 * Description of blame.{showroot,blankboundary,showemail,date} 237 configuration variables have been added to "git config --help". 238 239 * After an error from lstat(), diff_populate_filespec() function 240 sometimes still went ahead and used invalid data in struct stat, 241 which has been fixed. 242 243 * UNC paths are also relevant in Cygwin builds and they are now 244 tested just like Mingw builds. 245 246 * Correct start-up sequence so that a repository could be placed 247 immediately under the root directory again (which was broken at 248 around Git 2.13). 249 250 * The credential helper for libsecret (in contrib/) has been improved 251 to allow possibly prompting the end user to unlock secrets that are 252 currently locked (otherwise the secrets may not be loaded). 253 254 * MinGW updates. 255 256 * Error checking in "git imap-send" for empty response has been 257 improved. 258 259 * Recent update to the refs infrastructure implementation started 260 rewriting packed-refs file more often than before; this has been 261 optimized again for most trivial cases. 262 263 * Some error messages did not quote filenames shown in it, which have 264 been fixed. 265 266 * "git rebase -i" recently started misbehaving when a submodule that 267 is configured with 'submodule.<name>.ignore' is dirty; this has 268 been corrected. 269 270 * Building with NO_LIBPCRE1_JIT did not disable it, which has been fixed. 271 272 * We used to add an empty alternate object database to the system 273 that does not help anything; it has been corrected. 274 275 * Doc update around use of "format-patch --subject-prefix" etc. 276 277 * A fix for an ancient bug in "git apply --ignore-space-change" codepath. 278 279 * Clarify and enhance documentation for "merge-base --fork-point", as 280 it was clear what it computed but not why/what for. 281 282 * A few scripts (both in production and tests) incorrectly redirected 283 their error output. These have been corrected. 284 285 * "git notes" sent its error message to its standard output stream, 286 which was corrected. 287 288 * The three-way merge performed by "git cherry-pick" was confused 289 when a new submodule was added in the meantime, which has been 290 fixed (or "papered over"). 291 292 * The sequencer machinery (used by "git cherry-pick A..B", and "git 293 rebase -i", among other things) would have lost a commit if stopped 294 due to an unlockable index file, which has been fixed. 295 296 * "git apply --inaccurate-eof" when used with "--ignore-space-change" 297 triggered an internal sanity check, which has been fixed. 298 299 * Command line completion (in contrib/) has been taught about the 300 "--copy" option of "git branch". 301 302 * When "git rebase" prepared an mailbox of changes and fed it to "git 303 am" to replay them, it was confused when a stray "From " happened 304 to be in the log message of one of the replayed changes. This has 305 been corrected. 306 307 * There was a recent semantic mismerge in the codepath to write out a 308 section of a configuration section, which has been corrected. 309 310 * Mentions of "git-rebase" and "git-am" (dashed form) still remained 311 in end-user visible strings emitted by the "git rebase" command; 312 they have been corrected. 313 314 * Contrary to the documentation, "git pull -4/-6 other-args" did not 315 ask the underlying "git fetch" to go over IPv4/IPv6, which has been 316 corrected. 317 318 * "git checkout --recursive" may overwrite and rewind the history of 319 the branch that happens to be checked out in submodule 320 repositories, which might not be desirable. Detach the HEAD but 321 still allow the recursive checkout to succeed in such a case. 322 (merge 57f22bf997 sb/submodule-recursive-checkout-detach-head later to maint). 323 324 * "git branch --set-upstream" has been deprecated and (sort of) 325 removed, as "--set-upstream-to" is the preferred one these days. 326 The documentation still had "--set-upstream" listed on its 327 synopsis section, which has been corrected. 328 (merge a060f3d3d8 tz/branch-doc-remove-set-upstream later to maint). 329 330 * Internally we use 0{40} as a placeholder object name to signal the 331 codepath that there is no such object (e.g. the fast-forward check 332 while "git fetch" stores a new remote-tracking ref says "we know 333 there is no 'old' thing pointed at by the ref, as we are creating 334 it anew" by passing 0{40} for the 'old' side), and expect that a 335 codepath to locate an in-core object to return NULL as a sign that 336 the object does not exist. A look-up for an object that does not 337 exist however is quite costly with a repository with large number 338 of packfiles. This access pattern has been optimized. 339 (merge 87b5e236a1 jk/fewer-pack-rescan later to maint). 340 341 * In addition to "git stash -m message", the command learned to 342 accept "git stash -mmessage" form. 343 (merge 5675473fcb ph/stash-save-m-option-fix later to maint). 344 345 * @{-N} in "git checkout @{-N}" may refer to a detached HEAD state, 346 but the documentation was not clear about it, which has been fixed. 347 (merge 75ce149575 ks/doc-checkout-previous later to maint). 348 349 * A regression in the progress eye-candy was fixed. 350 (merge 9c5951cacf jk/progress-delay-fix later to maint). 351 352 * The code internal to the recursive merge strategy was not fully 353 prepared to see a path that is renamed to try overwriting another 354 path that is only different in case on case insensitive systems. 355 This does not matter in the current code, but will start to matter 356 once the rename detection logic starts taking hints from nearby 357 paths moving to some directory and moves a new path along with them. 358 (merge 4cba2b0108 en/merge-recursive-icase-removal later to maint). 359 360 * An v2.12-era regression in pathspec match logic, which made it look 361 into submodule tree even when it is not desired, has been fixed. 362 (merge eef3df5a93 bw/pathspec-match-submodule-boundary later to maint). 363 364 * Amending commits in git-gui broke the author name that is non-ascii 365 due to incorrect enconding conversion. 366 367 * Recent update to the submodule configuration code broke "diff-tree" 368 by accidentally stopping to read from the index upfront. 369 (merge fd66bcc31f bw/submodule-config-cleanup later to maint). 370 371 * Git shows a message to tell the user that it is waiting for the 372 user to finish editing when spawning an editor, in case the editor 373 opens to a hidden window or somewhere obscure and the user gets 374 lost. 375 (merge abfb04d0c7 ls/editor-waiting-message later to maint). 376 377 * Other minor doc, test and build updates and code cleanups. 378 (merge 1a1fc2d5b5 rd/man-prune-progress later to maint). 379 (merge 0ba014035a rd/man-reflog-add-n later to maint). 380 (merge e54b63359f rd/doc-notes-prune-fix later to maint). 381 (merge ff4c9b413a sp/doc-info-attributes later to maint). 382 (merge 7db2cbf4f1 jc/receive-pack-hook-doc later to maint). 383 (merge 5a0526264b tg/t-readme-updates later to maint). 384 (merge 5e83cca0b8 jk/no-optional-locks later to maint). 385 (merge 826c778f7c js/hashmap-update-sample later to maint). 386 (merge 176b2d328c sg/setup-doc-update later to maint).