Documentation / RelNotes / 2.12.0.txton commit Merge branch 'js/difftool-builtin' (b7786bb)
   1Git 2.12 Release Notes
   2======================
   3
   4Backward compatibility notes.
   5
   6 * Use of an empty string that is used for 'everything matches' is
   7   still warned and Git asks users to use a more explicit '.' for that
   8   instead.  The hope is that existing users will not mind this
   9   change, and eventually the warning can be turned into a hard error,
  10   upgrading the deprecation into removal of this (mis)feature.  That
  11   is not scheduled to happen in the upcoming release (yet).
  12
  13 * The historical argument order "git merge <msg> HEAD <commit>..."
  14   has been deprecated for quite some time, and will be removed in the
  15   upcoming release.
  16
  17
  18Updates since v2.11
  19-------------------
  20
  21UI, Workflows & Features
  22
  23 * Various updates to "git p4".
  24
  25 * "git p4" didn't interact with the internal of .git directory
  26   correctly in the modern "git-worktree"-enabled world.
  27
  28 * "git branch --list" and friends learned "--ignore-case" option to
  29   optionally sort branches and tags case insensitively.
  30
  31 * In addition to %(subject), %(body), "log --pretty=format:..."
  32   learned a new placeholder %(trailers).
  33
  34 * "git rebase" learned "--quit" option, which allows a user to
  35   remove the metadata left by an earlier "git rebase" that was
  36   manually aborted without using "git rebase --abort".
  37
  38 * "git clone --reference $there --recurse-submodules $super" has been
  39   taught to guess repositories usable as references for submodules of
  40   $super that are embedded in $there while making a clone of the
  41   superproject borrow objects from $there; extend the mechanism to
  42   also allow submodules of these submodules to borrow repositories
  43   embedded in these clones of the submodules embedded in the clone of
  44   the superproject.
  45
  46 * Porcelain scripts written in Perl are getting internationalized.
  47
  48 * "git merge --continue" has been added as a synonym to "git commit"
  49   to conclude a merge that has stopped due to conflicts.
  50
  51 * Finer-grained control of what protocols are allowed for transports
  52   during clone/fetch/push have been enabled via a new configuration
  53   mechanism.
  54
  55 * "git shortlog" learned "--committer" option to group commits by
  56   committer, instead of author.
  57
  58 * GitLFS integration with "git p4" has been updated.
  59
  60 * The isatty() emulation for Windows has been updated to eradicate
  61   the previous hack that depended on internals of (older) MSVC
  62   runtime.
  63
  64 * Some platforms no longer understand "latin-1" that is still seen in
  65   the wild in e-mail headers; replace them with "iso-8859-1" that is
  66   more widely known when conversion fails from/to it.
  67   (merge df3755888b jc/latin-1 later to maint).
  68
  69 * "git grep" has been taught to optionally recurse into submodules.
  70
  71 * "git rm" used to refuse to remove a submodule when it has its own
  72   git repository embedded in its working tree.  It learned to move
  73   the repository away to $GIT_DIR/modules/ of the superproject
  74   instead, and allow the submodule to be deleted (as long as there
  75   will be no loss of local modifications, that is).
  76
  77 * A recent updates to "git p4" was not usable for older p4 but it
  78   could be made to work with minimum changes.  Do so.
  79
  80 * "git diff" learned diff.interHunkContext configuration variable
  81   that gives the default value for its --inter-hunk-context option.
  82
  83 * The prereleaseSuffix feature of version comparison that is used in
  84   "git tag -l" did not correctly when two or more prereleases for the
  85   same release were present (e.g. when 2.0, 2.0-beta1, and 2.0-beta2
  86   are there and the code needs to compare 2.0-beta1 and 2.0-beta2).
  87
  88
  89Performance, Internal Implementation, Development Support etc.
  90
  91 * Commands that operate on a log message and add lines to the trailer
  92   blocks, such as "format-patch -s", "cherry-pick (-x|-s)", and
  93   "commit -s", have been taught to use the logic of and share the
  94   code with "git interpret-trailer".
  95
  96 * The default Travis-CI configuration specifies newer P4 and GitLFS.
  97
  98 * The "fast hash" that had disastrous performance issues in some
  99   corner cases has been retired from the internal diff.
 100
 101 * The character width table has been updated to match Unicode 9.0
 102
 103 * Update the procedure to generate "tags" for developer support.
 104   (merge 046e4c1c09 jk/make-tags-find-sources-tweak later to maint).
 105
 106 * The codeflow of setting NOATIME and CLOEXEC on file descriptors Git
 107   opens has been simplified.
 108   (merge b4d065df03 jc/git-open-cloexec later to maint).
 109
 110 * "git diff" and its family had two experimental heuristics to shift
 111   the contents of a hunk to make the patch easier to read.  One of
 112   them turns out to be better than the other, so leave only the
 113   "--indent-heuristic" option and remove the other one.
 114   (merge 3cde4e02ee jc/retire-compaction-heuristics later to maint).
 115
 116 * A new submodule helper "git submodule embedgitdirs" to make it
 117   easier to move embedded .git/ directory for submodules in a
 118   superproject to .git/modules/ (and point the latter with the former
 119   that is turned into a "gitdir:" file) has been added.
 120
 121 * "git push \\server\share\dir" has recently regressed and then
 122   fixed.  A test has retroactively been added for this breakage.
 123
 124 * Build updates for Cygwin.
 125
 126 * The implementation of "real_path()" was to go there with chdir(2)
 127   and call getcwd(3), but this obviously wouldn't be usable in a
 128   threaded environment.  Rewrite it to manually resolve relative
 129   paths including symbolic links in path components.
 130
 131 * Adjust documentation to help AsciiDoctor render better while not
 132   breaking the rendering done by AsciiDoc.
 133
 134
 135Also contains various documentation updates and code clean-ups.
 136
 137Fixes since v2.10
 138-----------------
 139
 140Unless otherwise noted, all the fixes since v2.9 in the maintenance
 141track are contained in this release (see the maintenance releases'
 142notes for details).
 143
 144 * We often decide if a session is interactive by checking if the
 145   standard I/O streams are connected to a TTY, but isatty() that
 146   comes with Windows incorrectly returned true if it is used on NUL
 147   (i.e. an equivalent to /dev/null).  This has been fixed.
 148
 149 * "git svn" did not work well with path components that are "0", and
 150   some configuration variable it uses were not documented.
 151   (merge ea9a93dcc2 ew/svn-fixes later to maint).
 152
 153 * "git rev-parse --symbolic" failed with a more recent notation like
 154   "HEAD^-1" and "HEAD^!".
 155
 156 * An empty directory in a working tree that can simply be nuked used
 157   to interfere while merging or cherry-picking a change to create a
 158   submodule directory there, which has been fixed..
 159
 160 * The code in "git push" to compute if any commit being pushed in the
 161   superproject binds a commit in a submodule that hasn't been pushed
 162   out was overly inefficient, making it unusable even for a small
 163   project that does not have any submodule but have a reasonable
 164   number of refs.
 165
 166 * "git push --dry-run --recurse-submodule=on-demand" wasn't
 167   "--dry-run" in the submodules.
 168
 169 * The output from "git worktree list" was made in readdir() order,
 170   and was unstable.
 171
 172 * mergetool.<tool>.trustExitCode configuration variable did not apply
 173   to built-in tools, but now it does.
 174
 175 * "git p4" LFS support was broken when LFS stores an empty blob.
 176
 177 * A corner case in merge-recursive regression that crept in
 178   during 2.10 development cycle has been fixed.
 179
 180 * Transport with dumb http can be fooled into following foreign URLs
 181   that the end user does not intend to, especially with the server
 182   side redirects and http-alternates mechanism, which can lead to
 183   security issues.  Tighten the redirection and make it more obvious
 184   to the end user when it happens.
 185
 186 * Update the error messages from the dumb-http client when it fails
 187   to obtain loose objects; we used to give sensible error message
 188   only upon 404 but we now forbid unexpected redirects that needs to
 189   be reported with something sensible.
 190
 191 * When diff.renames configuration is on (and with Git 2.9 and later,
 192   it is enabled by default, which made it worse), "git stash"
 193   misbehaved if a file is removed and another file with a very
 194   similar content is added.
 195
 196 * "git diff --no-index" did not take "--no-abbrev" option.
 197
 198 * "git difftool --dir-diff" had a minor regression when started from
 199   a subdirectory, which has been fixed.
 200
 201 * "git commit --allow-empty --only" (no pathspec) with dirty index
 202   ought to be an acceptable way to create a new commit that does not
 203   change any paths, but it was forbidden, perhaps because nobody
 204   needed it so far.
 205
 206 * Git 2.11 had a minor regression in "merge --ff-only" that competed
 207   with another process that simultanously attempted to update the
 208   index. We used to explain what went wrong with an error message,
 209   but the new code silently failed.  The error message has been
 210   resurrected.
 211
 212 * A pathname that begins with "//" or "\\" on Windows is special but
 213   path normalization logic was unaware of it.
 214
 215 * "git pull --rebase", when there is no new commits on our side since
 216   we forked from the upstream, should be able to fast-forward without
 217   invoking "git rebase", but it didn't.
 218
 219 * The way to specify hotkeys to "xxdiff" that is used by "git
 220   mergetool" has been modernized to match recent versions of xxdiff.
 221
 222 * Unlike "git am --abort", "git cherry-pick --abort" moved HEAD back
 223   to where cherry-pick started while picking multiple changes, when
 224   the cherry-pick stopped to ask for help from the user, and the user
 225   did "git reset --hard" to a different commit in order to re-attempt
 226   the operation.
 227
 228 * Code cleanup in shallow boundary computation.
 229
 230 * A recent update to receive-pack to make it easier to drop garbage
 231   objects made it clear that GIT_ALTERNATE_OBJECT_DIRECTORIES cannot
 232   have a pathname with a colon in it (no surprise!), and this in turn
 233   made it impossible to push into a repository at such a path.  This
 234   has been fixed by introducing a quoting mechanism used when
 235   appending such a path to the colon-separated list.
 236
 237 * The function usage_msg_opt() has been updated to say "fatal:"
 238   before the custom message programs give, when they want to die
 239   with a message about wrong command line options followed by the
 240   standard usage string.
 241
 242 * "git index-pack --stdin" needs an access to an existing repository,
 243   but "git index-pack file.pack" to generate an .idx file that
 244   corresponds to a packfile does not.
 245
 246 * Fix for NDEBUG builds.
 247
 248 * A lazy "git push" without refspec did not internally use a fully
 249   specified refspec to perform 'current', 'simple', or 'upstream'
 250   push, causing unnecessary "ambiguous ref" errors.
 251
 252 * "git p4" misbehaved when swapping a directory and a symbolic link.
 253
 254 * Even though an fix was attempted in Git 2.9.3 days, but running
 255   "git difftool --dir-diff" from a subdirectory never worked. This
 256   has been fixed.
 257
 258 * "git p4" that tracks multile p4 paths imported a single changelist
 259   that touches files in these multiple paths as one commit, followed
 260   by many empty commits.  This has been fixed.
 261
 262 * A potential but unlikely buffer overflow in Windows port has been
 263   fixed.
 264
 265 * When the http server gives an incomplete response to a smart-http
 266   rpc call, it could lead to client waiting for a full response that
 267   will never come.  Teach the client side to notice this condition
 268   and abort the transfer.
 269
 270 * Compression setting for producing packfiles were spread across
 271   three codepaths, one of which did not honor any configuration.
 272   Unify these so that all of them honor core.compression and
 273   pack.compression variables the same way.
 274   (merge 8de7eeb54b jc/compression-config later to maint).
 275
 276 * "git fast-import" sometimes mishandled while rebalancing notes
 277   tree, which has been fixed.
 278   (merge 405d7f4af6 mh/fast-import-notes-fix-new later to maint).
 279
 280 * Recent update to the default abbreviation length that auto-scales
 281   lacked documentation update, which has been corrected.
 282   (merge 48d5014dd4 jc/abbrev-autoscale-config later to maint).
 283
 284 * Leakage of lockfiles in the config subsystem has been fixed.
 285   (merge c06fa62dfc nd/config-misc-fixes later to maint).
 286
 287 * It is natural that "git gc --auto" may not attempt to pack
 288   everything into a single pack, and there is no point in warning
 289   when the user has configured the system to use the pack bitmap,
 290   leading to disabling further "gc".
 291   (merge 1c409a705c dt/disable-bitmap-in-auto-gc later to maint).
 292
 293 * "git archive" did not read the standard configuration files, and
 294   failed to notice a file that is marked as binary via the userdiff
 295   driver configuration.
 296   (merge 965cba2e7e jk/archive-zip-userdiff-config later to maint).
 297
 298 * "git blame --porcelain" misidentified the "previous" <commit, path>
 299   pair (aka "source") when contents came from two or more files.
 300   (merge 4e76832984 jk/blame-fixes later to maint).
 301
 302 * "git rebase -i" with a recent update started showing an incorrect
 303   count when squashing more than 10 commits.
 304   (merge 356b8ecff1 jk/rebase-i-squash-count-fix later to maint).
 305
 306 * "git <cmd> @{push}" on a detached HEAD used to segfault; it has
 307   been corrected to error out with a message.
 308   (merge b10731f43d km/branch-get-push-while-detached later to maint).
 309
 310 * Running "git add a/b" when "a" is a submodule correctly errored
 311   out, but without a meaningful error message.
 312   (merge 2d81c48fa7 sb/pathspec-errors later to maint).
 313
 314 * Typing ^C to pager, which usually does not kill it, killed Git and
 315   took the pager down as a collateral damage in certain process-tree
 316   structure.  This has been fixed.
 317   (merge 46df6906f3 jk/execv-dashed-external later to maint).
 318
 319 * "git mergetool" without any pathspec on the command line that is
 320   run from a subdirectory became no-op in Git v2.11 by mistake, which
 321   has been fixed.
 322
 323 * Retire long unused/unmaintained gitview from the contrib/ area.
 324   (merge 3120925c25 sb/remove-gitview later to maint).
 325
 326 * Tighten a test to avoid mistaking an extended ERE regexp engine as
 327   a PRE regexp engine.
 328   (merge 7675c7bd01 jk/grep-e-could-be-extended-beyond-posix later to maint).
 329
 330 * Other minor doc, test and build updates and code cleanups.
 331   (merge f2627d9b19 sb/submodule-config-cleanup later to maint).
 332   (merge 384f1a167b sb/unpack-trees-cleanup later to maint).
 333   (merge 3f05402ac0 ad/bisect-terms later to maint).
 334   (merge 874444b704 rh/diff-orderfile-doc later to maint).
 335   (merge c68d2d7c2b ws/request-pull-code-cleanup later to maint).