Documentation / RelNotes / 2.10.0.txton commit Merge branch 'master' of git://github.com/git-l10n/git-po (fe1280d)
   1Git 2.10 Release Notes
   2======================
   3
   4Backward compatibility notes
   5----------------------------
   6
   7Updates since v2.9
   8------------------
   9
  10UI, Workflows & Features
  11
  12 * "git pull --rebase --verify-signature" learned to warn the user
  13   that "--verify-signature" is a no-op when rebasing.
  14
  15 * An upstream project can make a recommendation to shallowly clone
  16   some submodules in the .gitmodules file it ships.
  17
  18 * "git worktree add" learned that '-' can be used as a short-hand for
  19   "@{-1}", the previous branch.
  20
  21 * Update the funcname definition to support css files.
  22
  23 * The completion script (in contrib/) learned to complete "git
  24   status" options.
  25
  26 * Messages that are generated by auto gc during "git push" on the
  27   receiving end are now passed back to the sending end in such a way
  28   that they are shown with "remote: " prefix to avoid confusing the
  29   users.
  30
  31 * "git add -i/-p" learned to honor diff.compactionHeuristic
  32   experimental knob, so that the user can work on the same hunk split
  33   as "git diff" output.
  34
  35 * "upload-pack" allows a custom "git pack-objects" replacement when
  36   responding to "fetch/clone" via the uploadpack.packObjectsHook.
  37   (merge b738396 jk/upload-pack-hook later to maint).
  38
  39 * Teach format-patch and mailsplit (hence "am") how a line that
  40   happens to begin with "From " in the e-mail message is quoted with
  41   ">", so that these lines can be restored to their original shape.
  42   (merge d9925d1 ew/mboxrd-format-am later to maint).
  43
  44 * "git repack" learned the "--keep-unreachable" option, which sends
  45   loose unreachable objects to a pack instead of leaving them loose.
  46   This helps heuristics based on the number of loose objects
  47   (e.g. "gc --auto").
  48   (merge e26a8c4 jk/repack-keep-unreachable later to maint).
  49
  50 * "log --graph --format=" learned that "%>|(N)" specifies the width
  51   relative to the terminal's left edge, not relative to the area to
  52   draw text that is to the right of the ancestry-graph section.  It
  53   also now accepts negative N that means the column limit is relative
  54   to the right border.
  55
  56 * A careless invocation of "git send-email directory/" after editing
  57   0001-change.patch with an editor often ends up sending both
  58   0001-change.patch and its backup file, 0001-change.patch~, causing
  59   embarrassment and a minor confusion.  Detect such an input and
  60   offer to skip the backup files when sending the patches out.
  61   (merge 531220b jc/send-email-skip-backup later to maint).
  62
  63 * "git submodule update" that drives many "git clone" could
  64   eventually hit flaky servers/network conditions on one of the
  65   submodules; the command learned to retry the attempt.
  66
  67 * The output coloring scheme learned two new attributes, italic and
  68   strike, in addition to existing bold, reverse, etc.
  69
  70 * "git log" learns log.showSignature configuration variable, and a
  71   command line option "--no-show-signature" to countermand it.
  72   (merge fce04c3 mj/log-show-signature-conf later to maint).
  73
  74 * More markings of messages for i18n, with updates to various tests
  75   to pass GETTEXT_POISON tests.
  76
  77 * "git archive" learned to handle files that are larger than 8GB and
  78   commits far in the future than expressible by the traditional US-TAR
  79   format.
  80   (merge 5caeeb8 jk/big-and-future-archive-tar later to maint).
  81
  82 * A new configuration variable core.sshCommand has been added to
  83   specify what value for GIT_SSH_COMMAND to use per repository.
  84
  85 * "git worktree prune" protected worktrees that are marked as
  86   "locked" by creating a file in a known location.  "git worktree"
  87   command learned a dedicated command pair to create and remove such
  88   a file, so that the users do not have to do this with editor.
  89
  90 * A handful of "git svn" updates.
  91
  92 * "git push" learned to accept and pass extra options to the
  93   receiving end so that hooks can read and react to them.
  94
  95 * "git status" learned to suggest "merge --abort" during a conflicted
  96   merge, just like it already suggests "rebase --abort" during a
  97   conflicted rebase.
  98
  99 * "git jump" script (in contrib/) has been updated a bit.
 100   (merge a91e692 jk/git-jump later to maint).
 101
 102 * "git push" and "git clone" learned to give better progress meters
 103   to the end user who is waiting on the terminal.
 104
 105 * An entry "git log --decorate" for the tip of the current branch is
 106   shown as "HEAD -> name" (where "name" is the name of the branch);
 107   the arrow is now painted in the same color as "HEAD", not in the
 108   color for commits.
 109
 110 * "git format-patch" learned format.from configuration variable to
 111   specify the default settings for its "--from" option.
 112
 113 * "git am -3" calls "git merge-recursive" when it needs to fall back
 114   to a three-way merge; this call has been turned into an internal
 115   subroutine call instead of spawning a separate subprocess.
 116
 117 * The command line completion scripts (in contrib/) now knows about
 118   "git branch --delete/--move [--remote]".
 119   (merge 2703c22 vs/completion-branch-fully-spelled-d-m-r later to maint).
 120
 121 * "git rev-parse --git-path hooks/<hook>" learned to take
 122   core.hooksPath configuration variable (introduced during 2.9 cycle)
 123   into account.
 124   (merge 9445b49 ab/hooks later to maint).
 125
 126 * "git log --show-signature" and other commands that display the
 127   verification status of PGP signature now shows the longer key-id,
 128   as 32-bit key-id is so last century.
 129
 130
 131Performance, Internal Implementation, Development Support etc.
 132
 133 * "git fast-import" learned the same performance trick to avoid
 134   creating too small a packfile as "git fetch" and "git push" have,
 135   using *.unpackLimit configuration.
 136
 137 * When "git daemon" is run without --[init-]timeout specified, a
 138   connection from a client that silently goes offline can hang around
 139   for a long time, wasting resources.  The socket-level KEEPALIVE has
 140   been enabled to allow the OS to notice such failed connections.
 141
 142 * "git upload-pack" command has been updated to use the parse-options
 143   API.
 144
 145 * The "git apply" standalone program is being libified; the first
 146   step to move many state variables into a structure that can be
 147   explicitly (re)initialized to make the machinery callable more
 148   than once has been merged.
 149
 150 * HTTP transport gained an option to produce more detailed debugging
 151   trace.
 152   (merge 73e57aa ep/http-curl-trace later to maint).
 153
 154 * Instead of taking advantage of the fact that a struct string_list
 155   that is allocated with all NULs happens to be the INIT_NODUP kind,
 156   the users of string_list structures are taught to initialize them
 157   explicitly as such, to document their behaviour better.
 158   (merge 2721ce2 jk/string-list-static-init later to maint).
 159
 160 * HTTPd tests learned to show the server error log to help diagnosing
 161   a failing tests.
 162   (merge 44f243d nd/test-lib-httpd-show-error-log-in-verbose later to maint).
 163
 164 * The ownership rule for the piece of memory that hold references to
 165   be fetched in "git fetch" was screwy, which has been cleaned up.
 166
 167 * "git bisect" makes an internal call to "git diff-tree" when
 168   bisection finds the culprit, but this call did not initialize the
 169   data structure to pass to the diff-tree API correctly.
 170
 171 * Further preparatory clean-up for "worktree" feature continues.
 172   (merge 0409e0b nd/worktree-cleanup-post-head-protection later to maint).
 173
 174 * Formats of the various data (and how to validate them) where we use
 175   GPG signature have been documented.
 176
 177 * A new run-command API function pipe_command() is introduced to
 178   sanely feed data to the standard input while capturing data from
 179   the standard output and the standard error of an external process,
 180   which is cumbersome to hand-roll correctly without deadlocking.
 181
 182 * The codepath to sign data in a prepared buffer with GPG has been
 183   updated to use this API to read from the status-fd to check for
 184   errors (instead of relying on GPG's exit status).
 185   (merge efee955 jk/gpg-interface-cleanup later to maint).
 186
 187 * Allow t/perf framework to use the features from the most recent
 188   version of Git even when testing an older installed version.
 189
 190 * The commands in the "log/diff" family have had an FILE* pointer in the
 191   data structure they pass around for a long time, but some codepaths
 192   used to always write to the standard output.  As a preparatory step
 193   to make "git format-patch" available to the internal callers, these
 194   codepaths have been updated to consistently write into that FILE*
 195   instead.
 196
 197 * Conversion from unsigned char sha1[20] to struct object_id
 198   continues.
 199
 200 * Improve the look of the way "git fetch" reports what happened to
 201   each ref that was fetched.
 202
 203 * The .c/.h sources are marked as such in our .gitattributes file so
 204   that "git diff -W" and friends would work better.
 205
 206 * Code clean-up to avoid using a variable string that compilers may
 207   feel untrustable as printf-style format given to write_file()
 208   helper function.
 209
 210 * "git p4" used a location outside $GIT_DIR/refs/ to place its
 211   temporary branches, which has been moved to refs/git-p4-tmp/.
 212
 213 * Existing autoconf generated test for the need to link with pthread
 214   library did not check all the functions from pthread libraries;
 215   recent FreeBSD has some functions in libc but not others, and we
 216   mistakenly thought linking with libc is enough when it is not.
 217
 218 * When "git fsck" reports a broken link (e.g. a tree object contains
 219   a blob that does not exist), both containing object and the object
 220   that is referred to were reported with their 40-hex object names.
 221   The command learned the "--name-objects" option to show the path to
 222   the containing object from existing refs (e.g. "HEAD~24^2:file.txt").
 223
 224 * Allow http daemon tests in Travis CI tests.
 225
 226 * Makefile assumed that -lrt is always available on platforms that
 227   want to use clock_gettime() and CLOCK_MONOTONIC, which is not a
 228   case for recent Mac OS X.  The necessary symbols are often found in
 229   libc on many modern systems and having -lrt on the command line, as
 230   long as the library exists, had no effect, but when the platform
 231   removes librt.a that is a different matter--having -lrt will break
 232   the linkage.
 233
 234   This change could be seen as a regression for those who do need to
 235   specify -lrt, as they now specifically ask for NEEDS_LIBRT when
 236   building. Hopefully they are in the minority these days.
 237
 238 * Further preparatory work on the refs API before the pluggable
 239   backend series can land.
 240
 241 * Error handling in the codepaths that updates refs has been
 242   improved.
 243
 244 * The API to iterate over all the refs (i.e. for_each_ref(), etc.)
 245   has been revamped.
 246
 247 * The handling of the "text=auto" attribute has been corrected.
 248   $ echo "* text=auto eol=crlf" >.gitattributes
 249   used to have the same effect as
 250   $ echo "* text eol=crlf" >.gitattributes
 251   i.e. declaring all files are text (ignoring "auto").  The
 252   combination has been fixed to be equivalent to doing
 253   $ git config core.autocrlf true
 254
 255 * Documentation has been updated to show better example usage
 256   of the updated "text=auto" attribute.
 257
 258 * A few tests that specifically target "git rebase -i" have been
 259   added.
 260
 261 * Dumb http transport on the client side has been optimized.
 262   (merge ecba195 ew/http-walker later to maint).
 263
 264 * Users of the parse_options_concat() API function need to allocate
 265   extra slots in advance and fill them with OPT_END() when they want
 266   to decide the set of supported options dynamically, which makes the
 267   code error-prone and hard to read.  This has been corrected by tweaking
 268   the API to allocate and return a new copy of "struct option" array.
 269
 270 * "git fetch" exchanges batched have/ack messages between the sender
 271   and the receiver, initially doubling every time and then falling
 272   back to enlarge the window size linearly.  The "smart http"
 273   transport, being an half-duplex protocol, outgrows the preset limit
 274   too quickly and becomes inefficient when interacting with a large
 275   repository.  The internal mechanism learned to grow the window size
 276   more aggressively when working with the "smart http" transport.
 277
 278 * Tests for "git svn" have been taught to reuse the lib-httpd test
 279   infrastructure when testing the subversion integration that
 280   interacts with subversion repositories served over the http://
 281   protocol.
 282   (merge a8a5d25 ew/git-svn-http-tests later to maint).
 283
 284 * "git pack-objects" has a few options that tell it not to pack
 285   objects found in certain packfiles, which require it to scan .idx
 286   files of all available packs.  The codepaths involved in these
 287   operations have been optimized for a common case of not having any
 288   non-local pack and/or any .kept pack.
 289
 290 * The t3700 test about "add --chmod=-x" have been made a bit more
 291   robust and generally cleaned up.
 292   (merge 766cdc4 ib/t3700-add-chmod-x-updates later to maint).
 293
 294 * The build procedure learned PAGER_ENV knob that lists what default
 295   environment variable settings to export for popular pagers.  This
 296   mechanism is used to tweak the default settings to MORE on FreeBSD.
 297   (merge 995bc22 ew/build-time-pager-tweaks later to maint).
 298
 299 * The http-backend (the server-side component of smart-http
 300   transport) used to trickle the HTTP header one at a time.  Now
 301   these write(2)s are batched.
 302   (merge b36045c ew/http-backend-batch-headers later to maint).
 303
 304 * When "git rebase" tries to compare set of changes on the updated
 305   upstream and our own branch, it computes patch-id for all of these
 306   changes and attempts to find matches. This has been optimized by
 307   lazily computing the full patch-id (which is expensive) to be
 308   compared only for changes that touch the same set of paths.
 309   (merge b3dfeeb kw/patch-ids-optim later to maint).
 310
 311 * A handful of tests that were broken under gettext-poison build have
 312   been fixed.
 313
 314 * The recent i18n patch we added during this cycle did a bit too much
 315   refactoring of the messages to avoid word-legos; the repetition has
 316   been reduced to help translators.
 317
 318
 319Also contains various documentation updates and code clean-ups.
 320
 321
 322Fixes since v2.9
 323----------------
 324
 325Unless otherwise noted, all the fixes since v2.8 in the maintenance
 326track are contained in this release (see the maintenance releases'
 327notes for details).
 328
 329 * The commands in `git log` family take %C(auto) in a custom format
 330   string.  This unconditionally turned the color on, ignoring
 331   --no-color or with --color=auto when the output is not connected to
 332   a tty; this was corrected to make the format truly behave as
 333   "auto".
 334
 335 * "git rev-list --count" whose walk-length is limited with "-n"
 336   option did not work well with the counting optimized to look at the
 337   bitmap index.
 338
 339 * "git show -W" (extend hunks to cover the entire function, delimited
 340   by lines that match the "funcname" pattern) used to show the entire
 341   file when a change added an entire function at the end of the file,
 342   which has been fixed.
 343
 344 * The documentation set has been updated so that literal commands,
 345   configuration variables and environment variables are consistently
 346   typeset in fixed-width font and bold in manpages.
 347
 348 * "git svn propset" subcommand that was added in 2.3 days is
 349   documented now.
 350
 351 * The documentation tries to consistently spell "GPG"; when
 352   referring to the specific program name, "gpg" is used.
 353
 354 * "git reflog" stopped upon seeing an entry that denotes a branch
 355   creation event (aka "unborn"), which made it appear as if the
 356   reflog was truncated.
 357
 358 * The git-prompt scriptlet (in contrib/) was not friendly with those
 359   who uses "set -u", which has been fixed.
 360
 361 * compat/regex code did not cleanly compile.
 362
 363 * A codepath that used alloca(3) to place an unbounded amount of data
 364   on the stack has been updated to avoid doing so.
 365
 366 * "git update-index --add --chmod=+x file" may be usable as an escape
 367   hatch, but not a friendly thing to force for people who do need to
 368   use it regularly.  "git add --chmod=+x file" can be used instead.
 369
 370 * Build improvements for gnome-keyring (in contrib/)
 371
 372 * "git status" used to say "working directory" when it meant "working
 373   tree".
 374
 375 * Comments about misbehaving FreeBSD shells have been clarified with
 376   the version number (9.x and before are broken, newer ones are OK).
 377
 378 * "git cherry-pick A" worked on an unborn branch, but "git
 379   cherry-pick A..B" didn't.
 380
 381 * Fix an unintended regression in v2.9 that breaks "clone --depth"
 382   that recurses down to submodules by forcing the submodules to also
 383   be cloned shallowly, which many server instances that host upstream
 384   of the submodules are not prepared for.
 385
 386 * Fix unnecessarily waste in the idiomatic use of ': ${VAR=default}'
 387   to set the default value, without enclosing it in double quotes.
 388
 389 * Some platform-specific code had non-ANSI strict declarations of C
 390   functions that do not take any parameters, which has been
 391   corrected.
 392
 393 * The internal code used to show local timezone offset is not
 394   prepared to handle timestamps beyond year 2100, and gave a
 395   bogus offset value to the caller.  Use a more benign looking
 396   +0000 instead and let "git log" going in such a case, instead
 397   of aborting.
 398
 399 * One among four invocations of readlink(1) in our test suite has
 400   been rewritten so that the test can run on systems without the
 401   command (others are in valgrind test framework and t9802).
 402
 403 * t/perf needs /usr/bin/time with GNU extension; the invocation of it
 404   is updated to "gtime" on Darwin.
 405
 406 * A bug, which caused "git p4" while running under verbose mode to
 407   report paths that are omitted due to branch prefix incorrectly, has
 408   been fixed; the command said "Ignoring file outside of prefix" for
 409   paths that are _inside_.
 410
 411 * The top level documentation "git help git" still pointed at the
 412   documentation set hosted at now-defunct google-code repository.
 413   Update it to point to https://git.github.io/htmldocs/git.html
 414   instead.
 415
 416 * A helper function that takes the contents of a commit object and
 417   finds its subject line did not ignore leading blank lines, as is
 418   commonly done by other codepaths.  Make it ignore leading blank
 419   lines to match.
 420
 421 * For a long time, we carried an in-code comment that said our
 422   colored output would work only when we use fprintf/fputs on
 423   Windows, which no longer is the case for the past few years.
 424
 425 * "gc.autoPackLimit" when set to 1 should not trigger a repacking
 426   when there is only one pack, but the code counted poorly and did
 427   so.
 428
 429 * Add a test to specify the desired behaviour that currently is not
 430   available in "git rebase -Xsubtree=...".
 431
 432 * More mark-up updates to typeset strings that are expected to
 433   literally typed by the end user in fixed-width font.
 434
 435 * "git commit --amend --allow-empty-message -S" for a commit without
 436   any message body could have misidentified where the header of the
 437   commit object ends.
 438
 439 * "git rebase -i --autostash" did not restore the auto-stashed change
 440   when the operation was aborted.
 441
 442 * Git does not know what the contents in the index should be for a
 443   path added with "git add -N" yet, so "git grep --cached" should not
 444   show hits (or show lack of hits, with -L) in such a path, but that
 445   logic does not apply to "git grep", i.e. searching in the working
 446   tree files.  But we did so by mistake, which has been corrected.
 447
 448 * "git blame -M" missed a single line that was moved within the file.
 449
 450 * Fix recently introduced codepaths that are involved in parallel
 451   submodule operations, which gave up on reading too early, and
 452   could have wasted CPU while attempting to write under a corner
 453   case condition.
 454
 455 * "git grep -i" has been taught to fold case in non-ascii locales
 456   correctly.
 457
 458 * A test that unconditionally used "mktemp" learned that the command
 459   is not necessarily available everywhere.
 460
 461 * There are certain house-keeping tasks that need to be performed at
 462   the very beginning of any Git program, and programs that are not
 463   built-in commands had to do them exactly the same way as "git"
 464   potty does.  It was easy to make mistakes in one-off standalone
 465   programs (like test helpers).  A common "main()" function that
 466   calls cmd_main() of individual program has been introduced to
 467   make it harder to make mistakes.
 468   (merge de61ceb jk/common-main later to maint).
 469
 470 * The test framework learned a new helper test_match_signal to
 471   check an exit code from getting killed by an expected signal.
 472
 473 * General code clean-up around a helper function to write a
 474   single-liner to a file.
 475   (merge 7eb6e10 jk/write-file later to maint).
 476
 477 * One part of "git am" had an oddball helper function that called
 478   stuff from outside "his" as opposed to calling what we have "ours",
 479   which was not gender-neutral and also inconsistent with the rest of
 480   the system where outside stuff is usuall called "theirs" in
 481   contrast to "ours".
 482
 483 * "git blame file" allowed the lineage of lines in the uncommitted,
 484   unadded contents of "file" to be inspected, but it refused when
 485   "file" did not appear in the current commit.  When "file" was
 486   created by renaming an existing file (but the change has not been
 487   committed), this restriction was unnecessarily tight.
 488
 489 * "git add -N dir/file && git write-tree" produced an incorrect tree
 490   when there are other paths in the same directory that sorts after
 491   "file".
 492
 493 * "git fetch http://user:pass@host/repo..." scrubbed the userinfo
 494   part, but "git push" didn't.
 495
 496 * "git merge" with renormalization did not work well with
 497   merge-recursive, due to "safer crlf" conversion kicking in when it
 498   shouldn't.
 499   (merge 1335d76 jc/renormalize-merge-kill-safer-crlf later to maint).
 500
 501 * The use of strbuf in "git rm" to build filename to remove was a bit
 502   suboptimal, which has been fixed.
 503
 504 * An age old bug that caused "git diff --ignore-space-at-eol"
 505   misbehave has been fixed.
 506
 507 * "git notes merge" had a code to see if a path exists (and fails if
 508   it does) and then open the path for writing (when it doesn't).
 509   Replace it with open with O_EXCL.
 510
 511 * "git pack-objects" and "git index-pack" mostly operate with off_t
 512   when talking about the offset of objects in a packfile, but there
 513   were a handful of places that used "unsigned long" to hold that
 514   value, leading to an unintended truncation.
 515
 516 * Recent update to "git daemon" tries to enable the socket-level
 517   KEEPALIVE, but when it is spawned via inetd, the standard input
 518   file descriptor may not necessarily be connected to a socket.
 519   Suppress an ENOTSOCK error from setsockopt().
 520
 521 * Recent FreeBSD stopped making perl available at /usr/bin/perl;
 522   switch the default the built-in path to /usr/local/bin/perl on not
 523   too ancient FreeBSD releases.
 524
 525 * "git commit --help" said "--no-verify" is only about skipping the
 526   pre-commit hook, and failed to say that it also skipped the
 527   commit-msg hook.
 528
 529 * "git merge" in Git v2.9 was taught to forbid merging an unrelated
 530   lines of history by default, but that is exactly the kind of thing
 531   the "--rejoin" mode of "git subtree" (in contrib/) wants to do.
 532   "git subtree" has been taught to use the "--allow-unrelated-histories"
 533   option to override the default.
 534
 535 * The build procedure for "git persistent-https" helper (in contrib/)
 536   has been updated so that it can be built with more recent versions
 537   of Go.
 538
 539 * There is an optimization used in "git diff $treeA $treeB" to borrow
 540   an already checked-out copy in the working tree when it is known to
 541   be the same as the blob being compared, expecting that open/mmap of
 542   such a file is faster than reading it from the object store, which
 543   involves inflating and applying delta.  This however kicked in even
 544   when the checked-out copy needs to go through the convert-to-git
 545   conversion (including the clean filter), which defeats the whole
 546   point of the optimization.  The optimization has been disabled when
 547   the conversion is necessary.
 548
 549 * "git -c grep.patternType=extended log --basic-regexp" misbehaved
 550   because the internal API to access the grep machinery was not
 551   designed well.
 552
 553 * Windows port was failing some tests in t4130, due to the lack of
 554   inum in the returned values by its lstat(2) emulation.
 555
 556 * The reflog output format is documented better, and a new format
 557   --date=unix to report the seconds-since-epoch (without timezone)
 558   has been added.
 559   (merge 442f6fd jk/reflog-date later to maint).
 560
 561 * "git difftool <paths>..." started in a subdirectory failed to
 562   interpret the paths relative to that directory, which has been
 563   fixed.
 564
 565 * The characters in the label shown for tags/refs for commits in
 566   "gitweb" output are now properly escaped for proper HTML output.
 567
 568 * FreeBSD can lie when asked mtime of a directory, which made the
 569   untracked cache code to fall back to a slow-path, which in turn
 570   caused tests in t7063 to fail because it wanted to verify the
 571   behaviour of the fast-path.
 572
 573 * Squelch compiler warnings for nedmalloc (in compat/) library.
 574
 575 * A small memory leak in the command line parsing of "git blame"
 576   has been plugged.
 577
 578 * The API documentation for hashmap was unclear if hashmap_entry
 579   can be safely discarded without any other consideration.  State
 580   that it is safe to do so.
 581
 582 * Not-so-recent rewrite of "git am" that started making internal
 583   calls into the commit machinery had an unintended regression, in
 584   that no matter how many seconds it took to apply many patches, the
 585   resulting committer timestamp for the resulting commits were all
 586   the same.
 587
 588 * "git push --force-with-lease" already had enough logic to allow
 589   ensuring that such a push results in creation of a ref (i.e. the
 590   receiving end did not have another push from sideways that would be
 591   discarded by our force-pushing), but didn't expose this possibility
 592   to the users.  It does so now.
 593   (merge 9eed4f3 jk/push-force-with-lease-creation later to maint).
 594
 595 * The mechanism to limit the pack window memory size, when packing is
 596   done using multiple threads (which is the default), is per-thread,
 597   but this was not documented clearly.
 598   (merge 954176c ms/document-pack-window-memory-is-per-thread later to maint).
 599
 600 * "import-tars" fast-import script (in contrib/) used to ignore a
 601   hardlink target and replaced it with an empty file, which has been
 602   corrected to record the same blob as the other file the hardlink is
 603   shared with.
 604   (merge 04e0869 js/import-tars-hardlinks later to maint).
 605
 606 * "git mv dir non-existing-dir/" did not work in some environments
 607   the same way as existing mainstream platforms.  The code now moves
 608   "dir" to "non-existing-dir", without relying on rename("A", "B/")
 609   that strips the trailing slash of '/'.
 610   (merge 189d035 js/mv-dir-to-new-directory later to maint).
 611
 612 * The "t/" hierarchy is prone to get an unusual pathname; "make test"
 613   has been taught to make sure they do not contain paths that cannot
 614   be checked out on Windows (and the mechanism can be reusable to
 615   catch pathnames that are not portable to other platforms as need
 616   arises).
 617   (merge c2cafd3 js/test-lint-pathname later to maint).
 618
 619 * When "git merge-recursive" works on history with many criss-cross
 620   merges in "verbose" mode, the names the command assigns to the
 621   virtual merge bases could have overwritten each other by unintended
 622   reuse of the same piece of memory.
 623   (merge 5447a76 rs/pull-signed-tag later to maint).
 624
 625 * "git checkout --detach <branch>" used to give the same advice
 626   message as that is issued when "git checkout <tag>" (or anything
 627   that is not a branch name) is given, but asking with "--detach" is
 628   an explicit enough sign that the user knows what is going on.  The
 629   advice message has been squelched in this case.
 630   (merge 779b88a sb/checkout-explit-detach-no-advice later to maint).
 631
 632 * "git difftool" by default ignores the error exit from the backend
 633   commands it spawns, because often they signal that they found
 634   differences by exiting with a non-zero status code just like "diff"
 635   does; the exit status codes 126 and above however are special in
 636   that they are used to signal that the command is not executable,
 637   does not exist, or killed by a signal.  "git difftool" has been
 638   taught to notice these exit status codes.
 639   (merge 45a4f5d jk/difftool-command-not-found later to maint).
 640
 641 * On Windows, help.browser configuration variable used to be ignored,
 642   which has been corrected.
 643   (merge 6db5967 js/no-html-bypass-on-windows later to maint).
 644
 645 * The "git -c var[=val] cmd" facility to append a configuration
 646   variable definition at the end of the search order was described in
 647   git(1) manual page, but not in git-config(1), which was more likely
 648   place for people to look for when they ask "can I make a one-shot
 649   override, and if so how?"
 650   (merge ae1f709 dg/document-git-c-in-git-config-doc later to maint).
 651
 652 * The tempfile (hence its user lockfile) API lets the caller to open
 653   a file descriptor to a temporary file, write into it and then
 654   finalize it by first closing the filehandle and then either
 655   removing or renaming the temporary file.  When the process spawns a
 656   subprocess after obtaining the file descriptor, and if the
 657   subprocess has not exited when the attempt to remove or rename is
 658   made, the last step fails on Windows, because the subprocess has
 659   the file descriptor still open.  Open tempfile with O_CLOEXEC flag
 660   to avoid this (on Windows, this is mapped to O_NOINHERIT).
 661   (merge 05d1ed6 bw/mingw-avoid-inheriting-fd-to-lockfile later to maint).
 662
 663 * Other minor clean-ups and documentation updates
 664   (merge 02a8cfa rs/merge-add-strategies-simplification later to maint).
 665   (merge af4941d rs/merge-recursive-string-list-init later to maint).
 666   (merge 1eb47f1 rs/use-strbuf-add-unique-abbrev later to maint).
 667   (merge ddd0bfa jk/tighten-alloc later to maint).
 668   (merge ecf30b2 rs/mailinfo-lib later to maint).
 669   (merge 0eb75ce sg/reflog-past-root later to maint).
 670   (merge 175d38c hv/doc-commit-reference-style later to maint).