1 To: git@vger.kernel.org
3 Subject: What's cooking in git.git (May 2015, #03; Fri, 8)
4 X-master-at: 8440f74997cf7958c7e8ec853f590828085049b8
5 X-next-at: 38e70713119c25ab5699df6b2fb13e4133d399ab
7 What's cooking in git.git (May 2015, #03; Fri, 8)
8 --------------------------------------------------
10 Here are the topics that have been cooking. Commits prefixed with
11 '-' are only in 'pu' (proposed updates) while commits prefixed with
14 You can find the changes described here in the integration branches
15 of the repositories listed at
17 http://git-blame.blogspot.com/p/git-public-repositories.html
19 --------------------------------------------------
22 * dl/branch-error-message (2015-05-06) 1 commit
23 - branch: do not call a "remote-tracking branch" a "remote branch"
28 * dl/subtree-push-no-squash (2015-05-07) 1 commit
29 - contrib/subtree: there's no push --squash
34 * ld/p4-editor-multi-words (2015-05-07) 2 commits
35 - git-p4: fix handling of multi-word P4EDITOR
36 - git-p4: add failing test for P4EDITOR handling
38 Unlike "$EDITOR" and "$GIT_EDITOR" that can hold the path to the
39 command and initial options (e.g. "/path/to/emacs -nw"), 'git p4'
40 did not let the shell interpolate the contents of the environment
41 variable that name the editor "$P4EDITOR" (and "$EDITOR", too).
42 Make it in line with the rest of Git, as well as with Perforce.
44 Reported to break some tests that assume the non-interpolating
45 behaviour that need to be fixed.
48 * dl/subtree-avoid-tricky-echo (2015-05-08) 1 commit
49 - contrib/subtree: portability fix for string printing
51 "git subtree" script (in contrib/) used "echo -n" to produce
52 progress messages in a non-portable way.
57 * ls/http-ssl-cipher-list (2015-05-08) 1 commit
58 - http: add support for specifying an SSL cipher list
60 Introduce http.<url>.SSLCipherList configuration variable to tweak
61 the list of cipher suite to be used with libcURL when talking with
67 * ps/bundle-verify-arg (2015-05-08) 1 commit
68 - bundle: verify arguments more strictly
73 * sg/help-subcommands (2015-05-08) 1 commit
74 - command-list.txt: fix whitespace inconsistency
76 A preparatory clean-up step.
80 --------------------------------------------------
83 * kk/log-merges-config (2015-04-21) 5 commits
84 - bash-completion: add support for git-log --merges= and log.merges
85 - t4202-log: add tests for --merges=
86 - Documentation: add git-log --merges= option and log.merges config. var
87 - log: honor log.merges= option
88 - revision: add --merges={show|only|hide} option
90 "git log" (but not other commands in the "log" family) learned to
91 pay attention to the log.merges configuration variable that can be
92 set to "show" (the normal behaviour), "only" (hide non-merge
93 commits), or "hide" (hide merge commits). --merges=(show|only|hide)
94 can be used to override the setting from the command line.
96 The documentation may need to be updated once more ($gmane/267250).
100 * mg/httpd-tests-update-for-apache-2.4 (2015-04-08) 2 commits
101 - t/lib-git-svn: check same httpd module dirs as lib-httpd
102 - t/lib-httpd: load mod_unixd
104 This is the first two commits in a three-patch series $gmane/266962
106 with updated log message ($gmane/268061).
109 * mh/numparse (2015-03-19) 14 commits
110 - diff_opt_parse(): use convert_i() when handling --abbrev=<num>
111 - diff_opt_parse(): use convert_i() when handling "-l<num>"
112 - opt_arg(): simplify pointer handling
113 - opt_arg(): report errors parsing option values
114 - opt_arg(): use convert_i() in implementation
115 - opt_arg(): val is always non-NULL
116 - builtin_diff(): detect errors when parsing --unified argument
117 - handle_revision_opt(): use convert_ui() when handling "--abbrev="
118 - strtoul_ui(), strtol_i(): remove functions
119 - handle_revision_opt(): use convert_i() when handling "-<digit>"
120 - handle_revision_opt(): use skip_prefix() in many places
121 - write_subdirectory(): use convert_ui() for parsing mode
122 - cacheinfo_callback(): use convert_ui() when handling "--cacheinfo"
123 - numparse: new module for parsing integral numbers
125 Many codepaths use unchecked use of strtol() and friends (or even
126 worse, atoi()). Introduce a set of wrappers that try to be more
133 * tf/gitweb-project-listing (2015-03-19) 5 commits
134 - gitweb: make category headings into links when they are directories
135 - gitweb: optionally set project category from its pathname
136 - gitweb: add a link under the search box to clear a project filter
137 - gitweb: if the PATH_INFO is incomplete, use it as a project_filter
138 - gitweb: fix typo in man page
140 Update gitweb to make it more pleasant to deal with a hierarchical
141 forest of repositories.
143 Any comments from those who use or have their own code in Gitweb?
146 * jc/a-lone-dash-stands-for-previous-branch (2015-03-16) 1 commit
147 - "-" and "@{-1}" on various programs
149 Lose special case code to make a lone dash "-" mean the previous
150 branch aka "@{-1}" from a handful subcommands, and instead support
151 the notation throughout the system by reimplementing it at the
154 Needs tests, documentation updates, etc. Also does only a half-way
155 job dealing with range notation, which needs to be fixed before the
156 series goes anywhere.
159 * nd/list-files (2015-02-09) 21 commits
160 - t3080: tests for git-list-files
161 - list-files: -M aka diff-cached
162 - list-files -F: show submodules with the new indicator '&'
163 - list-files: add -F/--classify
164 - list-files: show directories as well as files
165 - list-files: do not show duplicate cached entries
166 - list-files: sort output and remove duplicates
167 - list-files: add -t back
168 - list-files: add -1 short for --no-column
169 - list-files: add -R/--recursive short for --max-depth=-1
170 - list-files: -u does not imply showing stages
171 - list-files: make alias 'ls' default to 'list-files'
172 - list-files: a user friendly version of ls-files and more
173 - ls-files: support --max-depth
174 - ls-files: add --column
175 - ls-files: add --color to highlight file names
176 - ls-files: buffer full item in strbuf before printing
177 - ls_colors.c: highlight submodules like directories
178 - ls_colors.c: add a function to color a file name
179 - ls_colors.c: parse color.ls.* from config file
180 - ls_colors.c: add $LS_COLORS parsing code
182 A new "git list-files" Porcelain command, "ls-files" with bells and
185 Reroll to base on wt-status work ($gmane/265142) has seen some
186 positive discussions.
188 Waiting for a further polished reroll ($gmane/265534).
191 * js/fsck-opt (2015-01-21) 19 commits
192 - fsck: support ignoring objects in `git fsck` via fsck.skiplist
193 - fsck: git receive-pack: support excluding objects from fsck'ing
194 - fsck: introduce `git fsck --quick`
195 - fsck: support demoting errors to warnings
196 - fsck: document the new receive.fsck.* options
197 - fsck: allow upgrading fsck warnings to errors
198 - fsck: optionally ignore specific fsck issues completely
199 - fsck: disallow demoting grave fsck errors to warnings
200 - fsck: add a simple test for receive.fsck.*
201 - fsck: make fsck_tag() warn-friendly
202 - fsck: handle multiple authors in commits specially
203 - fsck: make fsck_commit() warn-friendly
204 - fsck: make fsck_ident() warn-friendly
205 - fsck: report the ID of the error/warning
206 - fsck: allow demoting errors to warnings via receive.fsck.warn = <key>
207 - fsck: offer a function to demote fsck errors to warnings
208 - fsck: provide a function to parse fsck message IDs
209 - fsck: introduce identifiers for fsck messages
210 - fsck: introduce fsck options
212 "fsck.warnings = <list of error tokens>" I suggested turned out to
213 be an unpopular choice (sorry Dscho).
218 * nd/untracked-cache (2015-03-12) 24 commits
219 - git-status.txt: advertisement for untracked cache
220 - untracked cache: guard and disable on system changes
221 - mingw32: add uname()
222 - t7063: tests for untracked cache
223 - update-index: test the system before enabling untracked cache
224 - update-index: manually enable or disable untracked cache
225 - status: enable untracked cache
226 - untracked-cache: temporarily disable with $GIT_DISABLE_UNTRACKED_CACHE
227 - untracked cache: mark index dirty if untracked cache is updated
228 - untracked cache: print stats with $GIT_TRACE_UNTRACKED_STATS
229 - untracked cache: avoid racy timestamps
230 - read-cache.c: split racy stat test to a separate function
231 - untracked cache: invalidate at index addition or removal
232 - untracked cache: load from UNTR index extension
233 - untracked cache: save to an index extension
234 - ewah: add convenient wrapper ewah_serialize_strbuf()
235 - untracked cache: don't open non-existent .gitignore
236 - untracked cache: mark what dirs should be recursed/saved
237 - untracked cache: record/validate dir mtime and reuse cached output
238 - untracked cache: make a wrapper around {open,read,close}dir()
239 - untracked cache: invalidate dirs recursively if .gitignore changes
240 - untracked cache: initial untracked cache validation
241 - untracked cache: record .gitignore information and dir hierarchy
242 - dir.c: optionally compute sha-1 of a .gitignore file
244 Need extra sets of eyes to review this.
247 * nd/pathspec-strip-fix (2015-04-18) 1 commit
248 - pathspec: adjust prefixlen after striping trailing slash
250 Does not quite fix ($gmane/267614).
253 * jc/diff-b-m (2015-02-23) 5 commits
256 - diffcore-rename: allow easier debugging
257 - diffcore-rename.c: add locate_rename_src()
258 - diffcore-break: allow debugging
260 "git diff -B -M" produced incorrect patch when the postimage of a
261 completely rewritten file is similar to the preimage of a removed
262 file; such a resulting file must not be expressed as a rename from
265 The fix in this patch is broken, unfortunately.
268 * pw/remote-set-url-fetch (2014-11-26) 1 commit
269 - remote: add --fetch and --both options to set-url
274 * tr/remerge-diff (2014-11-10) 9 commits
275 - t4213: avoid "|" in sed regexp
276 - log --remerge-diff: show what the conflict resolution changed
277 - name-hash: allow dir hashing even when !ignore_case
278 - merge-recursive: allow storing conflict hunks in index
279 - merge_diff_mode: fold all merge diff variants into an enum
280 - combine-diff: do not pass revs->dense_combined_merges redundantly
281 - merge-recursive: -Xindex-only to leave worktree unchanged
282 - merge-recursive: internal flag to avoid touching the worktree
283 - merge-recursive: remove dead conditional in update_stages()
285 "log -p" output learns a new way to let users inspect a merge
286 commit by showing the differences between the automerged result
287 with conflicts the person who recorded the merge would have seen
288 and the final conflict resolution that was recorded in the merge.
290 Waiting for a reroll ($gmane/256591).
293 * hv/submodule-config (2014-11-11) 4 commits
294 - do not die on error of parsing fetchrecursesubmodules option
295 - use new config API for worktree configurations of submodules
296 - extract functions for submodule config set and lookup
297 - implement submodule config cache for lookup of submodule names
299 Kicked back to 'pu' per request ($gmane/255610).
302 * tg/perf-lib-test-perf-cleanup (2013-09-19) 2 commits
303 - perf-lib: add test_perf_cleanup target
304 - perf-lib: split starting the test from the execution
306 Add test_perf_cleanup shell function to the perf suite, that allows
307 the script writers to define a test with a clean-up action.
312 * jc/show-branch (2014-03-24) 5 commits
313 - show-branch: use commit slab to represent bitflags of arbitrary width
314 - show-branch.c: remove "all_mask"
315 - show-branch.c: abstract out "flags" operation
316 - show-branch.c: lift all_mask/all_revs to a global static
317 - show-branch.c: update comment style
319 Waiting for the final step to lift the hard-limit.
321 --------------------------------------------------
324 * jc/daemon-no-ipv6-for-2.4.1 (2015-05-05) 1 commit
325 (merged to 'next' on 2015-05-05 at 1eb279f)
326 + daemon: unbreak NO_IPV6 build regression
328 "git daemon" fails to build from the source under NO_IPV6
329 configuration (regression in 2.4).
331 Will merge to 'master' in the second batch.
334 * sb/prefix-path-free-results (2015-05-05) 1 commit
335 (merged to 'next' on 2015-05-07 at 64f15a8)
336 + prefix_path(): unconditionally free results in the callers
338 Code clean-up (not a leak-fix).
340 Will merge to 'master' in the second batch.
343 * sg/completion-omit-credential-helpers (2015-05-06) 1 commit
344 - completion: remove credential helpers from porcelain commands
346 The Git subcommand completion (in contrib/) listed credential
347 helpers among candidates, which is not something the end user would
350 Will merge to 'next'.
353 * jk/skip-http-tests-under-no-curl (2015-05-07) 2 commits
354 - tests: skip dav http-push tests under NO_EXPAT=NoThanks
355 - t/lib-httpd.sh: skip tests if NO_CURL is defined
359 Will merge to 'next'.
362 * mh/ref-directory-file (2015-05-05) 19 commits
364 - reflog_expire(): integrate lock_ref_sha1_basic() errors into ours
365 - ref_transaction_commit(): delete extra "the" from error message
366 - ref_transaction_commit(): provide better error messages
367 - rename_ref(): integrate lock_ref_sha1_basic() errors into ours
368 - lock_ref_sha1_basic(): improve diagnostics for D/F conflicts
369 - lock_ref_sha1_basic(): report errors via a "struct strbuf *err"
370 - verify_refname_available(): report errors via a "struct strbuf *err"
371 - verify_refname_available(): rename function
372 - refs: check for D/F conflicts among refs processed in a transaction
373 - ref_transaction_commit(): use a string_list for detecting duplicates
374 - is_refname_available(): use dirname in first loop
375 - struct nonmatching_ref_data: store a refname instead of a ref_entry
376 - report_refname_conflict(): inline function
377 - entry_matches(): inline function
378 - is_refname_available(): convert local variable "dirname" to strbuf
379 - is_refname_available(): avoid shadowing "dir" variable
380 - is_refname_available(): explain the reason for an early exit
381 - t1404: new tests of D/F conflicts within ref transactions
382 (this branch uses mh/ref-lock-avoid-running-out-of-fds.)
387 * nd/dwim-wildcards-as-pathspecs (2015-05-03) 1 commit
388 (merged to 'next' on 2015-05-07 at 8176de7)
389 + pathspec: avoid the need of "--" when wildcard is used
391 A heuristic to help the "git <cmd> <revs> <pathspec>" command line
392 convention to catch mistyped paths is to make sure all the non-rev
393 parameters in the later part of the command line are names of the
394 files in the working tree, but that means "git grep $str -- \*.c"
395 must always be disambiguated with "--", because nobody sane will
396 create a file whose name literally is asterisk-dot-see. Loosen the
397 heuristic to declare that with a wildcard string the user likely
398 meant to give us a pathspec.
400 Will merge to 'master' in the third batch.
403 * pt/pull-tests (2015-05-03) 7 commits
404 - t5521: test --dry-run does not make any changes
405 - t5520: test --rebase failure on unborn branch with index
406 - t5520: test --rebase with multiple branches
407 - t5520: test work tree fast-forward when fetch updates head
408 - t5520: test for failure if index has unresolved entries
409 - t5520: implement tests for no merge candidates cases
410 - t5520: test pulling multiple branches into an empty repository
412 Add more test coverage to "git pull". This conflicts slightly with
413 a new test jc/merge topic adds, but resolution is fairly trivial.
415 Expecting a reroll ($gmane/268391).
418 * sg/complete-decorate-full-not-long (2015-05-03) 1 commit
419 (merged to 'next' on 2015-05-07 at 2beb429)
420 + completion: fix and update 'git log --decorate=' options
422 The completion for "log --decorate=" parameter value was incorrect.
424 Will merge to 'master' in the second batch.
427 * sg/completion-no-redundant-all-command-list (2015-05-03) 1 commit
428 (merged to 'next' on 2015-05-07 at 00b4bd9)
429 + completion: remove redundant __git_compute_all_commands() call
433 Will merge to 'master' in the second batch.
436 * jc/hash-object (2015-05-05) 4 commits
437 (merged to 'next' on 2015-05-07 at 9b81a06)
438 + write_sha1_file(): do not use a separate sha1[] array
439 + t1007: add hash-object --literally tests
440 + hash-object --literally: fix buffer overrun with extra-long object type
441 + git-hash-object.txt: document --literally option
443 "hash-object --literally" introduced in v2.2 was not prepared to
444 take a really long object type name.
446 Will merge to 'master' in the second batch.
449 * jn/clean-use-error-not-fprintf-on-stderr (2015-05-04) 1 commit
450 (merged to 'next' on 2015-05-05 at 12391f1)
451 + config: use error() instead of fprintf(stderr, ...)
453 Some error messages in "git config" were emitted without calling
454 the usual error() facility.
456 Will merge to 'master' in the second batch.
459 * jc/clone-bundle (2015-04-30) 1 commit
460 - repack: optionally create a clone.bundle
465 * jk/at-push-sha1 (2015-05-03) 12 commits
466 - for-each-ref: accept "%(push)" format
467 - for-each-ref: use skip_prefix instead of starts_with
468 - sha1_name: implement @{push} shorthand
469 - sha1_name: refactor interpret_upstream_mark
470 - sha1_name: refactor upstream_mark
471 - remote.c: add branch_get_push
472 - remote.c: report specific errors from branch_get_upstream
473 - remote.c: introduce branch_get_upstream helper
474 - remote.c: provide per-branch pushremote name
475 - remote.c: hoist branch.*.remote lookup out of remote_get_1
476 - remote.c: drop "remote" pointer from "struct branch"
477 - remote.c: drop default_remote_name variable
479 Introduce <branch>@{push} short-hand to denote the remote-tracking
480 branch that tracks the branch at the remote the <branch> would be
483 Expecting a reroll ($gmane/268419, etc.).
486 * jk/rebase-quiet-noop (2015-04-28) 1 commit
487 (merged to 'next' on 2015-05-05 at 82780b9)
488 + rebase: silence "git checkout" for noop rebase
490 "git rebase --quiet" was not quite quiet when there is nothing to
493 Will merge to 'master' in the second batch.
496 * ld/p4-case-fold (2015-04-28) 1 commit
497 (merged to 'next' on 2015-05-05 at 03ecbd0)
498 + git-p4: add failing tests for case-folding p4d
500 Will merge to 'master' in the second batch.
503 * va/fix-git-p4-tests (2015-04-28) 1 commit
504 (merged to 'next' on 2015-05-05 at 795e858)
505 + git-p4: t9814: prevent --chain-lint failure
507 Will merge to 'master' in the second batch.
510 * jk/filter-branch-use-of-sed-on-incomplete-line (2015-04-29) 1 commit
511 (merged to 'next' on 2015-05-07 at 849a24d)
512 + filter-branch: avoid passing commit message through sed
514 "filter-branch" corrupted commit log message that ends with an
515 incomplete line on platforms with some "sed" implementations that
516 munge such a line. Work it around by avoiding to use "sed".
518 Will merge to 'master' in the second batch.
521 * ph/rebase-i-redo (2015-04-29) 1 commit
522 (merged to 'next' on 2015-05-07 at 8fc1f52)
523 + rebase -i: redo tasks that die during cherry-pick
525 "git rebase -i" moved the "current" command from "todo" to "done" a
526 bit too prematurely, losing a step when a "pick" did not even start.
528 Will merge to 'master' in the third batch.
531 * ep/do-not-feed-a-pointer-to-array-size (2015-05-05) 1 commit
532 (merged to 'next' on 2015-05-07 at 64d9a20)
533 + git-compat-util.h: implement a different ARRAY_SIZE macro for for safely deriving the size of array
535 Catch a programmer mistake to feed a pointer not an array to
536 ARRAY_SIZE() macro, by using a couple of GCC extensions.
538 Will merge to 'master' in the second batch.
541 * tb/blame-resurrect-convert-to-git (2015-05-03) 1 commit
542 (merged to 'next' on 2015-05-05 at 8e1974e)
543 + blame: CRLF in the working tree and LF in the repo
545 Some time ago, "git blame" (incorrectly) lost the convert_to_git()
546 call when synthesizing a fake "tip" commit that represents the
547 state in the working tree, which broke folks who record the history
548 with LF line ending to make their project portabile across
549 platforms while terminating lines in their working tree files with
550 CRLF for their platform.
552 Will merge to 'master' in the second batch.
555 * jc/plug-fmt-merge-msg-leak (2015-04-20) 1 commit
556 (merged to 'next' on 2015-05-05 at bd94828)
557 + fmt-merge-msg: plug small leak of commit buffer
559 Originally merged to 'next' on 2015-04-21
561 Will merge to 'master' in the second batch.
564 * ee/clean-remove-dirs (2015-04-26) 5 commits
565 - clean: improve performance when removing lots of directories
566 - p7300: add performance tests for clean
567 - t7300: add tests to document behavior of clean and nested git
568 - setup: sanity check file size in read_gitfile_gently
569 - setup: add gentle version of read_gitfile
571 Replace "is this subdirectory a separate repository that should not
572 be touched?" check "git clean" does by checking if it has .git/HEAD
573 using the submodule-related code with a more optimized check.
575 Waiting for a reroll.
578 * jk/still-interesting (2015-04-17) 1 commit
579 (merged to 'next' on 2015-05-05 at 6a5c89c)
580 + limit_list: avoid quadratic behavior from still_interesting
582 Originally merged to 'next' on 2015-04-21
584 "git rev-list --objects $old --not --all" to see if everything that
585 is reachable from $old is already connected to the existing refs
586 was very inefficient.
588 Will merge to 'master' in the second batch.
591 * ls/p4-changes-block-size (2015-04-20) 1 commit
592 (merged to 'next' on 2015-05-05 at 92596d4)
593 + git-p4: use -m when running p4 changes
595 Originally merged to 'next' on 2015-04-21
597 "git p4" learned "--changes-block-size <n>" to read the changes in
598 chunks from Perforce, instead of making one call to "p4 changes"
599 that may trigger "too many rows scanned" error from Perforce.
601 Will merge to 'master' in the second batch.
604 * mm/add-p-split-error (2015-04-16) 5 commits
605 (merged to 'next' on 2015-05-05 at c556011)
606 + stash -p: demonstrate failure of split with mixed y/n
607 + t3904-stash-patch: factor PERL prereq at the top of the file
608 + t3904-stash-patch: fix test description
609 + add -p: demonstrate failure when running 'edit' after a split
610 + t3701-add-interactive: simplify code
612 Originally merged to 'next' on 2015-04-23
614 When "add--interactive" splits a hunk into two overlapping hunks
615 and then let the user choose only one, it sometimes feeds an
616 incorrect patch text to "git apply". Add tests to demonstrate
619 I have a slight suspicion that this may be $gmane/87202 coming back
620 and biting us (I seem to have said "let's run with this and see
621 what happens" back then).
623 Will merge to 'master' in the second batch.
626 * tb/t0027-crlf (2015-04-25) 3 commits
627 (merged to 'next' on 2015-05-05 at 36accbb)
628 + t0027: Add repoMIX and LF_nul
629 + t0027: support NATIVE_CRLF platforms
630 + t0027: cleanup: rename functions; avoid non-leading TABs
632 Originally merged to 'next' on 2015-04-21
634 More line-ending tests.
636 Will merge to 'master' in the second batch.
639 * jc/gitignore-precedence (2015-04-22) 1 commit
640 (merged to 'next' on 2015-05-05 at 6ef85da)
641 + ignore: info/exclude should trump core.excludesfile
643 core.excludesfile (defaulting to $XDG_HOME/git/ignore) is supposed
644 to be overridden by repository-specific .git/info/exclude file, but
645 the order was swapped from the beginning. This belatedly fixes it.
647 Will merge to 'master' in the third batch.
650 * jk/git-no-more-argv0-path-munging (2015-04-22) 1 commit
651 (merged to 'next' on 2015-05-05 at b3f9a45)
652 + stop putting argv[0] dirname at front of PATH
654 We have prepended $GIT_EXEC_PATH and the path "git" is installed in
655 (typically "/usr/bin") to $PATH when invoking subprograms and hooks
656 for almost eternity, but the original use case the latter tried to
657 support was semi-bogus (i.e. install git to /opt/foo/git and run it
658 without having /opt/foo on $PATH), and more importantly it has
659 become less and less relevant as Git grew more mainstream (i.e. the
660 users would _want_ to have it on their $PATH). Stop prepending the
661 path in which "git" is installed to users' $PATH, as that would
662 interfere the command search order people depend on (e.g. they may
663 not like versions of programs that are unrelated to Git in /usr/bin
664 and want to override them by having different ones in /usr/local/bin
665 and have the latter directory earlier in their $PATH).
667 Will merge to 'master' in the third batch.
670 * jk/stash-require-clean-index (2015-04-22) 3 commits
671 (merged to 'next' on 2015-05-05 at b5f6c32)
672 + stash: require a clean index to apply
673 + t3903: avoid applying onto dirty index
674 + t3903: stop hard-coding commit sha1s
676 "git stash pop/apply" forgot to make sure that not just the working
677 tree is clean but also the index is clean. The latter is important
678 as a stash application can conflict and the index will be used for
681 Will merge to 'master' in the third batch.
684 * jc/merge (2015-04-29) 15 commits
685 (merged to 'next' on 2015-05-07 at 1c56512)
686 + merge: deprecate 'git merge <message> HEAD <commit>' syntax
687 + merge: handle FETCH_HEAD internally
688 + merge: decide if we auto-generate the message early in collect_parents()
689 + merge: make collect_parents() auto-generate the merge message
690 + merge: extract prepare_merge_message() logic out
691 + merge: narrow scope of merge_names
692 + merge: split reduce_parents() out of collect_parents()
693 + merge: clarify collect_parents() logic
694 + merge: small leakfix and code simplification
695 + merge: do not check argc to determine number of remote heads
696 + merge: clarify "pulling into void" special case
697 + t5520: test pulling an octopus into an unborn branch
699 + merge: simplify code flow
700 + merge: test the top-level merge driver
701 (this branch is used by jc/merge-drop-old-syntax.)
703 "git merge FETCH_HEAD" learned that the previous "git fetch" could
704 be to create an Octopus merge, i.e. recording multiple branches
705 that are not marked as "not-for-merge"; this allows us to lose an
706 old style invocation "git merge <msg> HEAD $commits..." in the
707 implementation of "git pull" script; the old style syntax can now
710 Will merge to 'master' in the third batch.
713 * jc/merge-drop-old-syntax (2015-04-29) 1 commit
714 - merge: drop 'git merge <message> HEAD <commit>' syntax
715 (this branch uses jc/merge.)
717 Stop supporting "git merge <messsage> HEAD <commit>" syntax that
718 has been deprecated since October 2007.
720 Will merge to 'next' and keep there during the 2.5 cycle.
723 * jk/test-chain-lint (2015-04-28) 2 commits
724 (merged to 'next' on 2015-05-05 at e6f0290)
725 + test-lib: turn on GIT_TEST_CHAIN_LINT by default
726 + t7502-commit.sh: fix a broken and-chain
728 Developer support to automatically detect broken &&-chain in the
729 test scripts is now turned on by default.
731 Will merge to 'master' in the third batch.
734 * mh/ref-lock-avoid-running-out-of-fds (2015-04-25) 8 commits
735 - ref_transaction_commit(): only keep one lockfile open at a time
736 - ref_transaction_commit(): remove the local flags variables
737 - write_ref_sha1(): inline function at callers
738 - commit_ref_update(): new function, extracted from write_ref_sha1()
739 - write_ref_to_lockfile(): new function, extracted from write_ref_sha1()
740 - refs.c: remove lock_fd from struct ref_lock
741 - t7004: rename ULIMIT test prerequisite to ULIMIT_STACK_SIZE
742 - update-ref: test handling large transactions properly
743 (this branch is used by mh/ref-directory-file.)
745 "git update-ref --stdin" was converted to be "transactional" some
746 time ago, but the ref transaction machinery was not prepared to
747 handle many refs, primarily because it tried to keep the loose ref
748 files open at the same time, which would cause us to hit the ENFILE
749 limit. Fix the ref_transaction_commit() code not to keep the file
750 descriptors open unnecessarily.
752 We would want to apply the fix to older codebase, but this fix is
753 unfortunately done to apply only on v2.4.x track.
755 Expecting a rebase to older codebase.
758 * bc/connect-plink (2015-04-28) 3 commits
759 (merged to 'next' on 2015-05-05 at 9def2e1)
760 + connect: improve check for plink to reduce false positives
761 + t5601: fix quotation error leading to skipped tests
762 + connect: simplify SSH connection code path
764 The connection initiation code for "ssh" transport tried to absorb
765 differences between the stock "ssh" and Putty-supplied "plink" and
766 its derivatives, but the logic to tell that we are using "plink"
767 variants were too loose and falsely triggered when "plink" appeared
768 anywhere in the path (e.g. "/home/me/bin/uplink/ssh").
770 Will merge to 'master' in the third batch.
773 * jc/test-prereq-validate (2015-04-28) 1 commit
774 (merged to 'next' on 2015-05-05 at a30464c)
775 + test: validate prerequistes syntax
777 Help us to find broken test script that splits the body part of the
778 test by mistaken use of wrong kind of quotes.
780 Will merge to 'master' in the third batch.
783 * fg/document-commit-message-stripping (2015-04-27) 1 commit
784 (merged to 'next' on 2015-05-05 at 1892a99)
785 + Documentation: clarify how "git commit" cleans up the edited log message
787 Will merge to 'master' in the third batch.
790 * ah/usage-strings (2015-05-03) 2 commits
791 - branch: fix funny-sounding error message
792 - blame, log: format usage strings similarly to those in documentation
794 A few usage string updates. The tip one still needs work.
797 * jk/reading-packed-refs (2015-04-16) 9 commits
798 (merged to 'next' on 2015-05-05 at 89b5694)
799 + t1430: add another refs-escape test
800 + read_packed_refs: avoid double-checking sane refs
801 + strbuf_getwholeline: use getdelim if it is available
802 + strbuf_getwholeline: avoid calling strbuf_grow
803 + strbuf_addch: avoid calling strbuf_grow
804 + config: use getc_unlocked when reading from file
805 + strbuf_getwholeline: use getc_unlocked
806 + git-compat-util: add fallbacks for unlocked stdio
807 + strbuf_getwholeline: use getc macro
809 Originally merged to 'next' on 2015-04-21
811 An earlier rewrite to use strbuf_getwholeline() instead of fgets(3)
812 to read packed-refs file revealed that the former is unacceptably
815 Will merge to 'master' in the second batch.
818 * kn/cat-file-literally (2015-05-06) 4 commits
819 - t1006: add tests for git cat-file --allow-unknown-type
820 - cat-file: teach cat-file a '--allow-unknown-type' option
821 - cat-file: make the options mutually exclusive
822 - sha1_file: support reading from a loose object of unknown type
824 Add the "--allow-unknown-type" option to "cat-file" to allow
825 inspecting loose objects of an experimental or a broken type.
827 This is v10 $gmane/267960 with fixup! squashed in.
830 * lm/squelch-bg-progress (2015-04-15) 2 commits
831 (merged to 'next' on 2015-05-05 at a2fe74d)
832 + compat/mingw: stubs for getpgid() and tcgetpgrp()
833 + progress: no progress in background
835 Originally merged to 'next' on 2015-04-21
837 Many long-running operations show progress eye-candy, even when
838 they are later backgrounded. Hide the eye-candy when the process
839 is sent to the background instead.
841 Will merge to 'master' in the second batch.
844 * va/p4-client-path (2015-04-23) 2 commits
845 (merged to 'next' on 2015-05-05 at 852facc)
846 + git-p4: improve client path detection when branches are used
847 + t9801: check git-p4's branch detection with client spec enabled
849 Originally merged to 'next' on 2015-04-23
851 git p4 attempts to better handle branches in Perforce.
853 Will merge to 'master' in the second batch.
856 * jk/sha1-file-reduce-useless-warnings (2015-03-30) 1 commit
857 (merged to 'next' on 2015-05-05 at a8de68e)
858 + sha1_file: squelch "packfile cannot be accessed" warnings
860 Originally merged to 'next' on 2015-04-21
862 Will merge to 'master' in the second batch.
865 * pt/credential-xdg (2015-03-25) 4 commits
866 (merged to 'next' on 2015-05-05 at 0d6711f)
867 + t0302: "unreadable" test needs POSIXPERM
868 + t0302: test credential-store support for XDG_CONFIG_HOME
869 + git-credential-store: support XDG_CONFIG_HOME
870 + git-credential-store: support multiple credential files
871 (this branch is used by pt/xdg-config-path.)
873 Originally merged to 'next' on 2015-03-25
875 Tweak the sample "store" backend of the credential helper to honor
876 XDG configuration file locations when specified.
878 Will merge to 'master' in the second batch.
881 * pt/xdg-config-path (2015-05-06) 7 commits
882 (merged to 'next' on 2015-05-07 at 38e7071)
883 + path.c: remove home_config_paths()
884 + git-config: replace use of home_config_paths()
885 + git-commit: replace use of home_config_paths()
886 + credential-store.c: replace home_config_paths() with xdg_config_home()
887 + dir.c: replace home_config_paths() with xdg_config_home()
888 + attr.c: replace home_config_paths() with xdg_config_home()
889 + path.c: implement xdg_config_home()
890 (this branch uses pt/credential-xdg.)
892 Code clean-up for xdg configuration path support.
894 Will merge to 'master' in the second batch.
897 * nd/diff-i-t-a (2015-03-23) 1 commit
898 (merged to 'next' on 2015-05-05 at cba9cd9)
899 + diff-lib.c: adjust position of i-t-a entries in diff
901 Originally merged to 'next' on 2015-03-24
903 After "git add -N", the path appeared in output of "git diff HEAD"
904 and "git diff --cached HEAD", leading "git status" to classify it
905 as "Changes to be committed". Such a path, however, is not yet to
906 be scheduled to be committed. "git diff" showed the change to the
907 path as modification, not as a "new file", in the header of its
910 Treat such paths as "yet to be added to the index but Git already
911 know about them"; "git diff HEAD" and "git diff --cached HEAD"
912 should not talk about them, and "git diff" should show them as new
913 files yet to be added to the index.
915 Will merge to 'master' in the third batch.
918 * nd/slim-index-pack-memory-usage (2015-04-18) 2 commits
919 (merged to 'next' on 2015-05-05 at 42bd845)
920 + index-pack: kill union delta_base to save memory
921 + index-pack: reduce object_entry size to save memory
923 Originally merged to 'next' on 2015-04-21
925 Memory usage of "git index-pack" has been trimmed by tens of
928 Will merge to 'master' in the second batch.
931 * nd/multiple-work-trees (2015-03-31) 41 commits
932 (merged to 'next' on 2015-05-05 at 0f04a1c)
933 + prune --worktrees: fix expire vs worktree existence condition
934 + t1501: fix test with split index
935 + t2026: fix broken &&-chain
936 + t2026 needs procondition SANITY
937 + git-checkout.txt: a note about multiple checkout support for submodules
938 + checkout: add --ignore-other-wortrees
939 + checkout: pass whole struct to parse_branchname_arg instead of individual flags
940 + git-common-dir: make "modules/" per-working-directory directory
941 + checkout: do not fail if target is an empty directory
942 + t2025: add a test to make sure grafts is working from a linked checkout
943 + checkout: don't require a work tree when checking out into a new one
944 + git_path(): keep "info/sparse-checkout" per work-tree
945 + count-objects: report unused files in $GIT_DIR/worktrees/...
946 + gc: support prune --worktrees
947 + gc: factor out gc.pruneexpire parsing code
948 + gc: style change -- no SP before closing parenthesis
949 + checkout: clean up half-prepared directories in --to mode
950 + checkout: reject if the branch is already checked out elsewhere
951 + prune: strategies for linked checkouts
952 + checkout: support checking out into a new working directory
953 + use new wrapper write_file() for simple file writing
954 + wrapper.c: wrapper to open a file, fprintf then close
955 + setup.c: support multi-checkout repo setup
956 + setup.c: detect $GIT_COMMON_DIR check_repository_format_gently()
957 + setup.c: convert check_repository_format_gently to use strbuf
958 + setup.c: detect $GIT_COMMON_DIR in is_git_directory()
959 + setup.c: convert is_git_directory() to use strbuf
960 + git-stash: avoid hardcoding $GIT_DIR/logs/....
961 + *.sh: avoid hardcoding $GIT_DIR/hooks/...
962 + git-sh-setup.sh: use rev-parse --git-path to get $GIT_DIR/objects
963 + $GIT_COMMON_DIR: a new environment variable
964 + commit: use SEQ_DIR instead of hardcoding "sequencer"
965 + fast-import: use git_path() for accessing .git dir instead of get_git_dir()
966 + reflog: avoid constructing .lock path with git_path
967 + *.sh: respect $GIT_INDEX_FILE
968 + git_path(): be aware of file relocation in $GIT_DIR
969 + path.c: group git_path(), git_pathdup() and strbuf_git_path() together
970 + path.c: rename vsnpath() to do_git_path()
971 + git_snpath(): retire and replace with strbuf_git_path()
972 + path.c: make get_pathname() call sites return const char *
973 + path.c: make get_pathname() return strbuf instead of static buffer
975 Originally merged to 'next' on 2015-04-02
977 A replacement for contrib/workdir/git-new-workdir that does not
978 rely on symbolic links and make sharing of objects and refs safer
979 by making the borrowee and borrowers aware of each other.
981 Will merge to 'master' in the second batch.