Documentation / RelNotes / 2.10.0.txton commit Sync with 2.9.3 (2376d31)
   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 * A couple of "git svn" updates.
  75
  76 * More markings of messages for i18n, with updates to various tests
  77   to pass GETTEXT_POISON tests.
  78
  79 * "git archive" learned to handle files that are larger than 8GB and
  80   commits far in the future than expressible by the traditional US-TAR
  81   format.
  82   (merge 5caeeb8 jk/big-and-future-archive-tar later to maint).
  83
  84 * A new configuration variable core.sshCommand has been added to
  85   specify what value for GIT_SSH_COMMAND to use per repository.
  86
  87 * "git worktree prune" protected worktrees that are marked as
  88   "locked" by creating a file in a known location.  "git worktree"
  89   command learned a dedicated command pair to create and remove such
  90   a file, so that the users do not have to do this with editor.
  91
  92 * A handful of "git svn" updates.
  93
  94 * "git push" learned to accept and pass extra options to the
  95   receiving end so that hooks can read and react to them.
  96
  97 * "git status" learned to suggest "merge --abort" during a conflicted
  98   merge, just like it already suggests "rebase --abort" during a
  99   conflicted rebase.
 100
 101 * "git jump" script (in contrib/) has been updated a bit.
 102   (merge a91e692 jk/git-jump later to maint).
 103
 104 * "git push" and "git clone" learned to give better progress meters
 105   to the end user who is waiting on the terminal.
 106
 107 * An entry "git log --decorate" for the tip of the current branch is
 108   shown as "HEAD -> name" (where "name" is the name of the branch);
 109   paint the arrow in the same color as "HEAD", not in the color for
 110   commits.
 111
 112 * "git format-patch" learned format.from configuration variable to
 113   specify the default settings for its "--from" option.
 114
 115 * "git am -3" calls "git merge-recursive" when it needs to fall back
 116   to a three-way merge; this call has been turned into an internal
 117   subroutine call instead of spawning a separate subprocess.
 118
 119 * The command line completion scripts (in contrib/) now knows about
 120   "git branch --delete/--move [--remote]".
 121   (merge 2703c22 vs/completion-branch-fully-spelled-d-m-r later to maint).
 122
 123
 124Performance, Internal Implementation, Development Support etc.
 125
 126 * "git fast-import" learned the same performance trick to avoid
 127   creating too small a packfile as "git fetch" and "git push" have,
 128   using *.unpackLimit configuration.
 129
 130 * When "git daemon" is run without --[init-]timeout specified, a
 131   connection from a client that silently goes offline can hang around
 132   for a long time, wasting resources.  The socket-level KEEPALIVE has
 133   been enabled to allow the OS to notice such failed connections.
 134
 135 * "git upload-pack" command has been updated to use the parse-options
 136   API.
 137
 138 * The "git apply" standalone program is being libified; this is the
 139   first step to move many state variables into a structure that can
 140   be explicitly (re)initialized to make the machinery callable more
 141   than once.
 142
 143 * HTTP transport gained an option to produce more detailed debugging
 144   trace.
 145   (merge 73e57aa ep/http-curl-trace later to maint).
 146
 147 * Instead of taking advantage of a struct string_list that is
 148   allocated with all NULs happens to be STRING_LIST_INIT_NODUP kind,
 149   initialize them explicitly as such, to document their behaviour
 150   better.
 151   (merge 2721ce2 jk/string-list-static-init later to maint).
 152
 153 * HTTPd tests learned to show the server error log to help diagnosing
 154   a failing tests.
 155   (merge 44f243d nd/test-lib-httpd-show-error-log-in-verbose later to maint).
 156
 157 * The ownership rule for the piece of memory that hold references to
 158   be fetched in "git fetch" was screwy, which has been cleaned up.
 159
 160 * "git bisect" makes an internal call to "git diff-tree" when
 161   bisection finds the culprit, but this call did not initialize the
 162   data structure to pass to the diff-tree API correctly.
 163
 164 * Further preparatory clean-up for "worktree" feature continues.
 165   (merge 0409e0b nd/worktree-cleanup-post-head-protection later to maint).
 166
 167 * Formats of the various data (and how to validate them) where we use
 168   GPG signature have been documented.
 169
 170 * A new run-command API function pipe_command() is introduced to
 171   sanely feed data to the standard input while capturing data from
 172   the standard output and the standard error of an external process,
 173   which is cumbersome to hand-roll correctly without deadlocking.
 174
 175   The codepath to sign data in a prepared buffer with GPG has been
 176   updated to use this API to read from the status-fd to check for
 177   errors (instead of relying on GPG's exit status).
 178   (merge efee955 jk/gpg-interface-cleanup later to maint).
 179
 180 * Allow t/perf framework to use the features from the most recent
 181   version of Git even when testing an older installed version.
 182
 183 * The commands in the "log/diff" family have had an FILE* pointer in the
 184   data structure they pass around for a long time, but some codepaths
 185   used to always write to the standard output.  As a preparatory step
 186   to make "git format-patch" available to the internal callers, these
 187   codepaths have been updated to consistently write into that FILE*
 188   instead.
 189
 190 * Conversion from unsigned char sha1[20] to struct object_id
 191   continues.
 192
 193 * Improve the look of the way "git fetch" reports what happened to
 194   each ref that was fetched.
 195
 196 * The .c/.h sources are marked as such in our .gitattributes file so
 197   that "git diff -W" and friends would work better.
 198
 199 * Code clean-up to avoid using a variable string that compilers may
 200   feel untrustable as printf-style format given to write_file()
 201   helper function.
 202
 203 * "git p4" used a location outside $GIT_DIR/refs/ to place its
 204   temporary branches, which has been moved to refs/git-p4-tmp/.
 205
 206 * Existing autoconf generated test for the need to link with pthread
 207   library did not check all the functions from pthread libraries;
 208   recent FreeBSD has some functions in libc but not others, and we
 209   mistakenly thought linking with libc is enough when it is not.
 210
 211 * When "git fsck" reports a broken link (e.g. a tree object contains
 212   a blob that does not exist), both containing object and the object
 213   that is referred to were reported with their 40-hex object names.
 214   The command learned the "--name-objects" option to show the path to
 215   the containing object from existing refs (e.g. "HEAD~24^2:file.txt").
 216
 217 * Allow http daemon tests in Travis CI tests.
 218
 219 * Makefile assumed that -lrt is always available on platforms that
 220   want to use clock_gettime() and CLOCK_MONOTONIC, which is not a
 221   case for recent Mac OS X.  The necessary symbols are often found in
 222   libc on many modern systems and having -lrt on the command line, as
 223   long as the library exists, had no effect, but when the platform
 224   removes librt.a that is a different matter--having -lrt will break
 225   the linkage.
 226
 227   This change could be seen as a regression for those who do need to
 228   specify -lrt, as they now specifically ask for NEEDS_LIBRT when
 229   building. Hopefully they are in the minority these days.
 230
 231 * Further preparatory work on the refs API before the pluggable
 232   backend series can land.
 233
 234 * Error handling in the codepaths that updates refs has been
 235   improved.
 236
 237 * The API to iterate over all the refs (i.e. for_each_ref(), etc.)
 238   has been revamped.
 239
 240 * A few tests that specifically target "git rebase -i" have been
 241   added.
 242
 243 * Dumb http transport on the client side has been optimized.
 244   (merge ecba195 ew/http-walker later to maint).
 245
 246 * Users of the parse_options_concat() API function need to allocate
 247   extra slots in advance and fill them with OPT_END() when they want
 248   to decide the set of supported options dynamically, which makes the
 249   code error-prone and hard to read.  This has been corrected by tweaking
 250   the API to allocate and return a new copy of "struct option" array.
 251
 252 * "git fetch" exchanges batched have/ack messages between the sender
 253   and the receiver, initially doubling every time and then falling
 254   back to enlarge the window size linearly.  The "smart http"
 255   transport, being an half-duplex protocol, outgrows the preset limit
 256   too quickly and becomes inefficient when interacting with a large
 257   repository.  The internal mechanism learned to grow the window size
 258   more aggressively when working with the "smart http" transport.
 259
 260 * Tests for "git svn" have been taught to reuse the lib-httpd test
 261   infrastructure when testing the subversion integration that
 262   interacts with subversion repositories served over the http://
 263   protocol.
 264   (merge a8a5d25 ew/git-svn-http-tests later to maint).
 265
 266 * "git pack-objects" has a few options that tell it not to pack
 267   objects found in certain packfiles, which require it to scan .idx
 268   files of all available packs.  The codepaths involved in these
 269   operations have been optimized for a common case of not having any
 270   non-local pack and/or any .kept pack.
 271
 272 * The t3700 test about "add --chmod=-x" have been made a bit more
 273   robust and generally cleaned up.
 274   (merge 766cdc4 ib/t3700-add-chmod-x-updates later to maint).
 275
 276 * The build procedure learned PAGER_ENV knob that lists what default
 277   environment variable settings to export for popular pagers.  This
 278   mechanism is used to tweak the default settings to MORE on FreeBSD.
 279   (merge 995bc22 ew/build-time-pager-tweaks later to maint).
 280
 281 * The http-backend (the server-side component of smart-http
 282   transport) used to trickle the HTTP header one at a time.  Now
 283   these write(2)s are batched.
 284   (merge b36045c ew/http-backend-batch-headers later to maint).
 285
 286 * When "git rebase" tries to compare set of changes on the updated
 287   upstream and our own branch, it computes patch-id for all of these
 288   changes and attempts to find matches. This has been optimized by
 289   lazily computing the full patch-id (which is expensive) to be
 290   compared only for changes that touch the same set of paths.
 291   (merge b3dfeeb kw/patch-ids-optim later to maint).
 292
 293
 294Also contains various documentation updates and code clean-ups.
 295
 296
 297Fixes since v2.9
 298----------------
 299
 300Unless otherwise noted, all the fixes since v2.8 in the maintenance
 301track are contained in this release (see the maintenance releases'
 302notes for details).
 303
 304 * The commands in `git log` family take %C(auto) in a custom format
 305   string.  This unconditionally turned the color on, ignoring
 306   --no-color or with --color=auto when the output is not connected to
 307   a tty; this was corrected to make the format truly behave as
 308   "auto".
 309
 310 * "git rev-list --count" whose walk-length is limited with "-n"
 311   option did not work well with the counting optimized to look at the
 312   bitmap index.
 313
 314 * "git show -W" (extend hunks to cover the entire function, delimited
 315   by lines that match the "funcname" pattern) used to show the entire
 316   file when a change added an entire function at the end of the file,
 317   which has been fixed.
 318
 319 * The documentation set has been updated so that literal commands,
 320   configuration variables and environment variables are consistently
 321   typeset in fixed-width font and bold in manpages.
 322
 323 * "git svn propset" subcommand that was added in 2.3 days is
 324   documented now.
 325
 326 * The documentation tries to consistently spell "GPG"; when
 327   referring to the specific program name, "gpg" is used.
 328
 329 * "git reflog" stopped upon seeing an entry that denotes a branch
 330   creation event (aka "unborn"), which made it appear as if the
 331   reflog was truncated.
 332
 333 * The git-prompt scriptlet (in contrib/) was not friendly with those
 334   who uses "set -u", which has been fixed.
 335
 336 * compat/regex code did not cleanly compile.
 337
 338 * A codepath that used alloca(3) to place an unbounded amount of data
 339   on the stack has been updated to avoid doing so.
 340
 341 * "git update-index --add --chmod=+x file" may be usable as an escape
 342   hatch, but not a friendly thing to force for people who do need to
 343   use it regularly.  "git add --chmod=+x file" can be used instead.
 344
 345 * Build improvements for gnome-keyring (in contrib/)
 346
 347 * "git status" used to say "working directory" when it meant "working
 348   tree".
 349
 350 * Comments about misbehaving FreeBSD shells have been clarified with
 351   the version number (9.x and before are broken, newer ones are OK).
 352
 353 * "git cherry-pick A" worked on an unborn branch, but "git
 354   cherry-pick A..B" didn't.
 355
 356 * Fix an unintended regression in v2.9 that breaks "clone --depth"
 357   that recurses down to submodules by forcing the submodules to also
 358   be cloned shallowly, which many server instances that host upstream
 359   of the submodules are not prepared for.
 360
 361 * Fix unnecessarily waste in the idiomatic use of ': ${VAR=default}'
 362   to set the default value, without enclosing it in double quotes.
 363
 364 * Some platform-specific code had non-ANSI strict declarations of C
 365   functions that do not take any parameters, which has been
 366   corrected.
 367
 368 * The internal code used to show local timezone offset is not
 369   prepared to handle timestamps beyond year 2100, and gave a
 370   bogus offset value to the caller.  Use a more benign looking
 371   +0000 instead and let "git log" going in such a case, instead
 372   of aborting.
 373
 374 * One among four invocations of readlink(1) in our test suite has
 375   been rewritten so that the test can run on systems without the
 376   command (others are in valgrind test framework and t9802).
 377
 378 * t/perf needs /usr/bin/time with GNU extension; the invocation of it
 379   is updated to "gtime" on Darwin.
 380
 381 * A bug, which caused "git p4" while running under verbose mode to
 382   report paths that are omitted due to branch prefix incorrectly, has
 383   been fixed; the command said "Ignoring file outside of prefix" for
 384   paths that are _inside_.
 385
 386 * The top level documentation "git help git" still pointed at the
 387   documentation set hosted at now-defunct google-code repository.
 388   Update it to point to https://git.github.io/htmldocs/git.html
 389   instead.
 390
 391 * A helper function that takes the contents of a commit object and
 392   finds its subject line did not ignore leading blank lines, as is
 393   commonly done by other codepaths.  Make it ignore leading blank
 394   lines to match.
 395
 396 * For a long time, we carried an in-code comment that said our
 397   colored output would work only when we use fprintf/fputs on
 398   Windows, which no longer is the case for the past few years.
 399
 400 * "gc.autoPackLimit" when set to 1 should not trigger a repacking
 401   when there is only one pack, but the code counted poorly and did
 402   so.
 403
 404 * Add a test to specify the desired behaviour that currently is not
 405   available in "git rebase -Xsubtree=...".
 406
 407 * More mark-up updates to typeset strings that are expected to
 408   literally typed by the end user in fixed-width font.
 409
 410 * "git commit --amend --allow-empty-message -S" for a commit without
 411   any message body could have misidentified where the header of the
 412   commit object ends.
 413
 414 * "git rebase -i --autostash" did not restore the auto-stashed change
 415   when the operation was aborted.
 416
 417 * Git does not know what the contents in the index should be for a
 418   path added with "git add -N" yet, so "git grep --cached" should not
 419   show hits (or show lack of hits, with -L) in such a path, but that
 420   logic does not apply to "git grep", i.e. searching in the working
 421   tree files.  But we did so by mistake, which has been corrected.
 422
 423 * "git blame -M" missed a single line that was moved within the file.
 424
 425 * Fix recently introduced codepaths that are involved in parallel
 426   submodule operations, which gave up on reading too early, and
 427   could have wasted CPU while attempting to write under a corner
 428   case condition.
 429
 430 * "git grep -i" has been taught to fold case in non-ascii locales
 431   correctly.
 432
 433 * A test that unconditionally used "mktemp" learned that the command
 434   is not necessarily available everywhere.
 435
 436 * There are certain house-keeping tasks that need to be performed at
 437   the very beginning of any Git program, and programs that are not
 438   built-in commands had to do them exactly the same way as "git"
 439   potty does.  It was easy to make mistakes in one-off standalone
 440   programs (like test helpers).  A common "main()" function that
 441   calls cmd_main() of individual program has been introduced to
 442   make it harder to make mistakes.
 443   (merge de61ceb jk/common-main later to maint).
 444
 445 * The test framework learned a new helper test_match_signal to
 446   check an exit code from getting killed by an expected signal.
 447
 448 * General code clean-up around a helper function to write a
 449   single-liner to a file.
 450   (merge 7eb6e10 jk/write-file later to maint).
 451
 452 * One part of "git am" had an oddball helper function that called
 453   stuff from outside "his" as opposed to calling what we have "ours",
 454   which was not gender-neutral and also inconsistent with the rest of
 455   the system where outside stuff is usuall called "theirs" in
 456   contrast to "ours".
 457
 458 * "git blame file" allowed the lineage of lines in the uncommitted,
 459   unadded contents of "file" to be inspected, but it refused when
 460   "file" did not appear in the current commit.  When "file" was
 461   created by renaming an existing file (but the change has not been
 462   committed), this restriction was unnecessarily tight.
 463
 464 * "git add -N dir/file && git write-tree" produced an incorrect tree
 465   when there are other paths in the same directory that sorts after
 466   "file".
 467
 468 * "git fetch http://user:pass@host/repo..." scrubbed the userinfo
 469   part, but "git push" didn't.
 470
 471 * "git merge" with renormalization did not work well with
 472   merge-recursive, due to "safer crlf" conversion kicking in when it
 473   shouldn't.
 474   (merge 1335d76 jc/renormalize-merge-kill-safer-crlf later to maint).
 475
 476 * The use of strbuf in "git rm" to build filename to remove was a bit
 477   suboptimal, which has been fixed.
 478
 479 * An age old bug that caused "git diff --ignore-space-at-eol"
 480   misbehave has been fixed.
 481
 482 * "git notes merge" had a code to see if a path exists (and fails if
 483   it does) and then open the path for writing (when it doesn't).
 484   Replace it with open with O_EXCL.
 485
 486 * "git pack-objects" and "git index-pack" mostly operate with off_t
 487   when talking about the offset of objects in a packfile, but there
 488   were a handful of places that used "unsigned long" to hold that
 489   value, leading to an unintended truncation.
 490
 491 * Recent update to "git daemon" tries to enable the socket-level
 492   KEEPALIVE, but when it is spawned via inetd, the standard input
 493   file descriptor may not necessarily be connected to a socket.
 494   Suppress an ENOTSOCK error from setsockopt().
 495
 496 * Recent FreeBSD stopped making perl available at /usr/bin/perl;
 497   switch the default the built-in path to /usr/local/bin/perl on not
 498   too ancient FreeBSD releases.
 499
 500 * "git commit --help" said "--no-verify" is only about skipping the
 501   pre-commit hook, and failed to say that it also skipped the
 502   commit-msg hook.
 503
 504 * "git merge" in Git v2.9 was taught to forbid merging an unrelated
 505   lines of history by default, but that is exactly the kind of thing
 506   the "--rejoin" mode of "git subtree" (in contrib/) wants to do.
 507   "git subtree" has been taught to use the "--allow-unrelated-histories"
 508   option to override the default.
 509
 510 * The build procedure for "git persistent-https" helper (in contrib/)
 511   has been updated so that it can be built with more recent versions
 512   of Go.
 513
 514 * There is an optimization used in "git diff $treeA $treeB" to borrow
 515   an already checked-out copy in the working tree when it is known to
 516   be the same as the blob being compared, expecting that open/mmap of
 517   such a file is faster than reading it from the object store, which
 518   involves inflating and applying delta.  This however kicked in even
 519   when the checked-out copy needs to go through the convert-to-git
 520   conversion (including the clean filter), which defeats the whole
 521   point of the optimization.  The optimization has been disabled when
 522   the conversion is necessary.
 523
 524 * "git -c grep.patternType=extended log --basic-regexp" misbehaved
 525   because the internal API to access the grep machinery was not
 526   designed well.
 527
 528 * Windows port was failing some tests in t4130, due to the lack of
 529   inum in the returned values by its lstat(2) emulation.
 530
 531 * The reflog output format is documented better, and a new format
 532   --date=unix to report the seconds-since-epoch (without timezone)
 533   has been added.
 534   (merge 442f6fd jk/reflog-date later to maint).
 535
 536 * "git difftool <paths>..." started in a subdirectory failed to
 537   interpret the paths relative to that directory, which has been
 538   fixed.
 539
 540 * The characters in the label shown for tags/refs for commits in
 541   "gitweb" output are now properly escaped for proper HTML output.
 542
 543 * FreeBSD can lie when asked mtime of a directory, which made the
 544   untracked cache code to fall back to a slow-path, which in turn
 545   caused tests in t7063 to fail because it wanted to verify the
 546   behaviour of the fast-path.
 547
 548 * Squelch compiler warnings for netmalloc (in compat/) library.
 549
 550 * A small memory leak in the command line parsing of "git blame"
 551   has been plugged.
 552
 553 * The API documentation for hashmap was unclear if hashmap_entry
 554   can be safely discarded without any other consideration.  State
 555   that it is safe to do so.
 556
 557 * Not-so-recent rewrite of "git am" that started making internal
 558   calls into the commit machinery had an unintended regression, in
 559   that no matter how many seconds it took to apply many patches, the
 560   resulting committer timestamp for the resulting commits were all
 561   the same.
 562
 563 * "git push --force-with-lease" already had enough logic to allow
 564   ensuring that such a push results in creation of a ref (i.e. the
 565   receiving end did not have another push from sideways that would be
 566   discarded by our force-pushing), but didn't expose this possibility
 567   to the users.  It does so now.
 568   (merge 9eed4f3 jk/push-force-with-lease-creation later to maint).
 569
 570 * The mechanism to limit the pack window memory size, when packing is
 571   done using multiple threads (which is the default), is per-thread,
 572   but this was not documented clearly.
 573   (merge 954176c ms/document-pack-window-memory-is-per-thread later to maint).
 574
 575 * "import-tars" fast-import script (in contrib/) used to ignore a
 576   hardlink target and replaced it with an empty file, which has been
 577   corrected to record the same blob as the other file the hardlink is
 578   shared with.
 579   (merge 04e0869 js/import-tars-hardlinks later to maint).
 580
 581 * "git mv dir non-existing-dir/" did not work in some environments
 582   the same way as existing mainstream platforms.  The code now moves
 583   "dir" to "non-existing-dir", without relying on rename("A", "B/")
 584   that strips the trailing slash of '/'.
 585   (merge 189d035 js/mv-dir-to-new-directory later to maint).
 586
 587 * Other minor clean-ups and documentation updates
 588   (merge 02a8cfa rs/merge-add-strategies-simplification later to maint).
 589   (merge af4941d rs/merge-recursive-string-list-init later to maint).
 590   (merge 1eb47f1 rs/use-strbuf-add-unique-abbrev later to maint).