1 To: git@vger.kernel.org
3 Subject: What's cooking in git.git (Apr 2015, #04; Mon, 27)
4 X-master-at: fb3e7d5515308c73c54c797356651d6ca41b58a6
5 X-next-at: 800e5753bd3e417587c1acede4a07b3a0f1c5198
7 What's cooking in git.git (Apr 2015, #04; Mon, 27)
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 I'll merge two small documentation follow-up to the new features
15 in 2.4.0 (mg/status-v-v and nd/versioncmp-prereleases topics) to
16 'master' and tag the final Git 2.4 later this week.
18 You can find the changes described here in the integration branches
19 of the repositories listed at
21 http://git-blame.blogspot.com/p/git-public-repositories.html
23 --------------------------------------------------
24 [Graduated to "master"]
26 * jc/push-cert (2015-04-02) 1 commit
27 (merged to 'next' on 2015-04-08 at aecdd43)
28 + push --signed: tighten what the receiving end can ask to sign
30 The "git push --signed" protocol extension did not limit what the
31 "nonce" that is a server-chosen string can contain or how long it
32 can be, which was unnecessarily lax. Limit both the length and the
33 alphabet to a reasonably small space that can still have enough
37 * ma/bash-completion-leaking-x (2015-04-12) 1 commit
38 (merged to 'next' on 2015-04-14 at 3a52a6d)
39 + completion: fix global bash variable leak on __gitcompappend
41 The completion script (in contrib/) contaminated global namespace
42 and clobbered on a shell variable $x.
45 * ps/grep-help-all-callback-arg (2015-04-12) 1 commit
46 (merged to 'next' on 2015-04-14 at e0a8092)
47 + grep: correctly initialize help-all option
52 * tb/connect-ipv6-parse-fix (2015-04-08) 1 commit
53 (merged to 'next' on 2015-04-14 at e720918)
54 + connect.c: ignore extra colon after hostname
56 An earlier update to the parser that disects a URL broke an
57 address, followed by a colon, followed by an empty string (instead
58 of the port number), e.g. ssh://example.com:/path/to/repo.
60 This should go to the maintenance track, too.
63 * va/fix-git-p4-tests (2015-04-12) 3 commits
64 (merged to 'next' on 2015-04-14 at 261bf90)
65 + t9814: guarantee only one source exists in git-p4 copy tests
66 + git-p4: fix copy detection test
67 + t9814: fix broken shell syntax in git-p4 rename test
69 Test fixes for git-p4.
71 --------------------------------------------------
74 * jc/epochtime-wo-tz (2015-04-15) 2 commits
75 (merged to 'next' on 2015-04-21 at b926f72)
76 + parse_date_basic(): let the system handle DST conversion
77 + parse_date_basic(): return early when given a bogus timestamp
79 "git commit --date=now" or anything that relies on approxidate lost
80 the daylight-saving-time offset.
82 Will merge to 'master' in the second batch of post v2.4 cycle.
85 * jc/plug-fmt-merge-msg-leak (2015-04-20) 1 commit
86 (merged to 'next' on 2015-04-21 at 443c728)
87 + fmt-merge-msg: plug small leak of commit buffer
89 Will merge to 'master' in the second batch of post v2.4 cycle.
92 * cn/bom-in-gitignore (2015-04-16) 5 commits
93 (merged to 'next' on 2015-04-21 at f7d56f1)
94 + attr: skip UTF8 BOM at the beginning of the input file
95 + config: use utf8_bom[] from utf.[ch] in git_parse_source()
96 + utf8-bom: introduce skip_utf8_bom() helper
97 + add_excludes_from_file: clarify the bom skipping logic
98 + dir: allow a BOM at the beginning of exclude files
100 Teach the codepaths that read .gitignore and .gitattributes files
101 that these files encoded in UTF-8 may have UTF-8 BOM marker at the
102 beginning; this makes it in line with what we do for configuration
105 Will merge to 'master' in the second batch of post v2.4 cycle.
108 * ee/clean-remove-dirs (2015-04-26) 5 commits
109 - clean: improve performance when removing lots of directories
110 - p7300: add performance tests for clean
111 - t7300: add tests to document behavior of clean and nested git
112 - setup: sanity check file size in read_gitfile_gently
113 - setup: add gentle version of read_gitfile
115 Replace "is this subdirectory a separate repository that should not
116 be touched?" check "git clean" does by checking if it has .git/HEAD
117 using the submodule-related code with a more optimized check.
119 Will merge to 'next'.
122 * ep/fix-test-lib-functions-report (2015-04-16) 1 commit
123 (merged to 'next' on 2015-04-21 at 23e584f)
124 + test-lib-functions.sh: fix the second argument to some helper functions
126 Will merge to 'master' in the second batch of post v2.4 cycle.
129 * jk/still-interesting (2015-04-17) 1 commit
130 (merged to 'next' on 2015-04-21 at 58b54af)
131 + limit_list: avoid quadratic behavior from still_interesting
133 "git rev-list --objects $old --not --all" to see if everything that
134 is reachable from $old is already connected to the existing refs
135 was very inefficient.
137 Will merge to 'master' in the second batch of post v2.4 cycle.
140 * jk/type-from-string-gently (2015-04-17) 1 commit
141 (merged to 'next' on 2015-04-20 at a97611f)
142 + type_from_string_gently: make sure length matches
144 "git cat-file bl $blob" failed to barf even though there is no
145 object type that is "bl".
147 Will merge to 'master' in the first batch of post v2.4 cycle.
150 * ls/p4-changes-block-size (2015-04-20) 1 commit
151 (merged to 'next' on 2015-04-21 at 830eeed)
152 + git-p4: use -m when running p4 changes
154 "git p4" learned "--changes-block-size <n>" to read the changes in
155 chunks from Perforce, instead of making one call to "p4 changes"
156 that may trigger "too many rows scanned" error from Perforce.
158 Will merge to 'master' in the second batch of post v2.4 cycle.
161 * mg/show-notes-doc (2015-04-17) 1 commit
162 (merged to 'next' on 2015-04-20 at 2e93969)
163 + rev-list-options.txt: complete sentence about notes matching
167 Will merge to 'master' in the first batch of post v2.4 cycle.
170 * mm/add-p-split-error (2015-04-16) 5 commits
171 (merged to 'next' on 2015-04-23 at b1bd21a)
172 + stash -p: demonstrate failure of split with mixed y/n
173 + t3904-stash-patch: factor PERL prereq at the top of the file
174 + t3904-stash-patch: fix test description
175 + add -p: demonstrate failure when running 'edit' after a split
176 + t3701-add-interactive: simplify code
178 When "add--interactive" splits a hunk into two overlapping hunks
179 and then let the user choose only one, it sometimes feeds an
180 incorrect patch text to "git apply". Add tests to demonstrate
183 I have a slight suspicion that this may be $gmane/87202 coming back
184 and biting us (I seem to have said "let's run with this and see
185 what happens" back then).
187 Will merge to 'master'.
190 * mm/usage-log-l-can-take-regex (2015-04-20) 2 commits
191 (merged to 'next' on 2015-04-21 at 03e951c)
192 + log -L: improve error message on malformed argument
193 + Documentation: change -L:<regex> to -L:<funcname>
195 Will merge to 'master' in the second batch of post v2.4 cycle.
198 * nd/pathspec-strip-fix (2015-04-18) 1 commit
199 - pathspec: adjust prefixlen after striping trailing slash
201 Does not quite fix ($gmane/267614).
204 * nd/t1509-chroot-test (2015-04-18) 1 commit
205 (merged to 'next' on 2015-04-21 at 60d35d3)
206 + t1509: update prepare script to be able to run t1509 in chroot again
210 Will merge to 'master' in the first batch of post v2.4 cycle.
213 * oh/fix-config-default-user-name-section (2015-04-17) 1 commit
214 (merged to 'next' on 2015-04-20 at b74b914)
215 + config: fix settings in default_user_config template
217 The default $HOME/.gitconfig file created upon "git config --global"
218 that edits it had incorrectly spelled user.name and user.email
221 Will merge to 'master' in the first batch of post v2.4 cycle.
224 * tb/t0027-crlf (2015-04-25) 3 commits
225 - t0027: Add repoMIX and LF_nul
226 (merged to 'next' on 2015-04-21 at 142cb99)
227 + t0027: support NATIVE_CRLF platforms
228 + t0027: cleanup: rename functions; avoid non-leading TABs
230 Will merge to 'master' in the first batch of post v2.4 cycle.
233 * jk/prune-mtime (2015-04-20) 3 commits
234 (merged to 'next' on 2015-04-21 at 9990d41)
235 + sha1_file: only freshen packs once per run
236 + sha1_file: freshen pack objects before loose
237 + reachable: only mark local objects as recent
239 Access to objects in repositories that borrow from another one on a
240 slow NFS server unnecessarily got more expensive due to recent code
241 becoming more cautious in a naive way not to lose objects to pruning.
243 Will merge to 'master' in the second batch of post v2.4 cycle.
246 * jc/gitignore-precedence (2015-04-22) 1 commit
247 - ignore: info/exclude should trump core.excludesfile
249 core.excludesfile (defaulting to $XDG_HOME/git/ignore) is supposed
250 to be overridden by repository-specific .git/info/exclude file, but
251 the order was swapped from the beginning. This belatedly fixes it.
253 Will merge to 'next'.
256 * jk/git-no-more-argv0-path-munging (2015-04-22) 1 commit
257 - stop putting argv[0] dirname at front of PATH
259 We have prepended $GIT_EXEC_PATH and the path "git" is installed in
260 (typically "/usr/bin") to $PATH when invoking subprograms and hooks
261 for almost eternity, but the original use case the latter tried to
262 support was semi-bogus (i.e. install git to /opt/foo/git and run it
263 without having /opt/foo on $PATH), and more importantly it has
264 become less and less relevant as Git grew more mainstream (i.e. the
265 users would _want_ to have it on their $PATH). Stop prepending the
266 path in which "git" is installed to users' $PATH, as that would
267 interfere the command search order people depend on (e.g. they may
268 not like versions of programs that are unrelated to Git in /usr/bin
269 and want to override them by having different ones in /usr/local/bin
270 and have the latter directory earlier in their $PATH).
272 Will merge to 'next'.
275 * jk/stash-require-clean-index (2015-04-22) 3 commits
276 - stash: require a clean index to apply
277 - t3903: avoid applying onto dirty index
278 - t3903: stop hard-coding commit sha1s
280 "git stash pop/apply" forgot to make sure that not just the working
281 tree is clean but also the index is clean. The latter is important
282 as a stash application can conflict and the index will be used for
285 Will merge to 'next'.
288 * jc/merge (2015-04-25) 14 commits
289 - merge: deprecate 'git merge <message> HEAD <commit>' syntax
290 - merge: handle FETCH_HEAD internally
291 - merge: decide if we auto-generate the message early in collect_parents()
292 - merge: make collect_parents() auto-generate the merge message
293 - merge: extract prepare_merge_message() logic out
294 - merge: narrow scope of merge_names
295 - merge: split reduce_parents() out of collect_parents()
296 - merge: clarify collect_parents() logic
297 - merge: small leakfix and code simplification
298 - merge: do not check argc to determine number of remote heads
299 - merge: clarify "pulling into void" special case
300 - t5520: test pulling an octopus into an unborn branch
302 - merge: simplify code flow
303 (this branch is used by jc/merge-drop-old-syntax.)
305 "git merge FETCH_HEAD" learned that the previous "git fetch" could
306 be to create an Octopus merge, i.e. recording multiple branches
307 that are not marked as "not-for-merge"; this allows us to lose an
308 old style invocation "git merge <msg> HEAD $commits..." in the
309 implementation of "git pull" script; the old style syntax can now
313 * jk/test-chain-lint (2015-04-22) 1 commit
314 - test-lib: turn on GIT_TEST_CHAIN_LINT by default
316 Will merge to 'next'.
319 * mg/status-v-v (2015-04-23) 1 commit
320 (merged to 'next' on 2015-04-25 at cbcaca4)
321 + status: document the -v/--verbose option
323 Will merge to 'master'.
326 * mh/ref-lock-avoid-running-out-of-fds (2015-04-25) 8 commits
327 - ref_transaction_commit(): only keep one lockfile open at a time
328 - ref_transaction_commit(): remove the local flags variables
329 - write_ref_sha1(): inline function at callers
330 - commit_ref_update(): new function, extracted from write_ref_sha1()
331 - write_ref_to_lockfile(): new function, extracted from write_ref_sha1()
332 - refs.c: remove lock_fd from struct ref_lock
333 - t7004: rename ULIMIT test prerequisite to ULIMIT_STACK_SIZE
334 - update-ref: test handling large transactions properly
335 (this branch is tangled with sb/ref-lock-avoid-running-out-of-fds.)
337 "git update-ref --stdin" was converted to be "transactional" some
338 time ago, but the ref transaction machinery was not prepared to
339 handle many refs, primarily because it tried to keep the loose ref
340 files open at the same time, which would cause us to hit the ENFILE
341 limit. Fix the ref_transaction_commit() code not to keep the file
342 descriptors open unnecessarily.
344 We would probably want to apply the fix to older codebase, but this
345 fix is unfortunately done to apply only on v2.4.x track.
348 * nd/versioncmp-prereleases (2015-04-23) 1 commit
349 (merged to 'next' on 2015-04-25 at f824d7f)
350 + git tag: mention versionsort.prereleaseSuffix in manpage
352 Will merge to 'master'.
355 * bc/connect-plink (2015-04-27) 3 commits
356 - connect: improve check for plink to reduce false positives
357 - t5601: fix quotation error leading to skipped tests
358 - connect: simplify SSH connection code path
359 (this branch uses jc/test-prereq-validate.)
361 The connection initiation code for "ssh" transport tried to absorb
362 differences between the stock "ssh" and Putty-supplied "plink" and
363 its derivatives, but the logic to tell that we are using "plink"
364 variants were too loose and falsely triggered when "plink" appeared
365 anywhere in the path (e.g. "/home/me/bin/uplink/ssh").
367 Will merge to 'next'.
370 * jc/test-prereq-validate (2015-04-26) 1 commit
371 - test: validate prerequistes syntax
372 (this branch is used by bc/connect-plink.)
374 Help us to find broken test script that splits the body part of the
375 test by mistaken use of wrong kind of quotes.
377 Will merge to 'next'.
380 * fg/document-commit-message-stripping (2015-04-27) 1 commit
381 - Documentation: clarify how "git commit" cleans up the edited log message
383 Will merge to 'next'.
385 --------------------------------------------------
388 * mh/fdopen-with-retry (2015-03-06) 6 commits
389 - buffer_fdinit(): use fdopen_with_retry()
390 - update_info_file(): use fdopen_with_retry()
391 - copy_to_log(): use fdopen_with_retry()
392 - fdopen_lock_file(): use fdopen_with_retry()
393 - SQUASH??? $gmane/264889
394 - xfdopen(): if first attempt fails, free memory and try again
396 Various parts of the code where they call fdopen() can fail when
397 they run out of memory; attempt to proceed by retrying the
398 operation after freeing some resource.
400 Waiting for further comments.
403 * nd/untracked-cache (2015-03-12) 24 commits
404 - git-status.txt: advertisement for untracked cache
405 - untracked cache: guard and disable on system changes
406 - mingw32: add uname()
407 - t7063: tests for untracked cache
408 - update-index: test the system before enabling untracked cache
409 - update-index: manually enable or disable untracked cache
410 - status: enable untracked cache
411 - untracked-cache: temporarily disable with $GIT_DISABLE_UNTRACKED_CACHE
412 - untracked cache: mark index dirty if untracked cache is updated
413 - untracked cache: print stats with $GIT_TRACE_UNTRACKED_STATS
414 - untracked cache: avoid racy timestamps
415 - read-cache.c: split racy stat test to a separate function
416 - untracked cache: invalidate at index addition or removal
417 - untracked cache: load from UNTR index extension
418 - untracked cache: save to an index extension
419 - ewah: add convenient wrapper ewah_serialize_strbuf()
420 - untracked cache: don't open non-existent .gitignore
421 - untracked cache: mark what dirs should be recursed/saved
422 - untracked cache: record/validate dir mtime and reuse cached output
423 - untracked cache: make a wrapper around {open,read,close}dir()
424 - untracked cache: invalidate dirs recursively if .gitignore changes
425 - untracked cache: initial untracked cache validation
426 - untracked cache: record .gitignore information and dir hierarchy
427 - dir.c: optionally compute sha-1 of a .gitignore file
429 Need extra sets of eyes to review this.
432 * jc/diff-b-m (2015-02-23) 5 commits
435 - diffcore-rename: allow easier debugging
436 - diffcore-rename.c: add locate_rename_src()
437 - diffcore-break: allow debugging
439 "git diff -B -M" produced incorrect patch when the postimage of a
440 completely rewritten file is similar to the preimage of a removed
441 file; such a resulting file must not be expressed as a rename from
444 The fix in this patch is broken, unfortunately.
447 * pw/remote-set-url-fetch (2014-11-26) 1 commit
448 - remote: add --fetch and --both options to set-url
453 * tr/remerge-diff (2014-11-10) 9 commits
454 - t4213: avoid "|" in sed regexp
455 - log --remerge-diff: show what the conflict resolution changed
456 - name-hash: allow dir hashing even when !ignore_case
457 - merge-recursive: allow storing conflict hunks in index
458 - merge_diff_mode: fold all merge diff variants into an enum
459 - combine-diff: do not pass revs->dense_combined_merges redundantly
460 - merge-recursive: -Xindex-only to leave worktree unchanged
461 - merge-recursive: internal flag to avoid touching the worktree
462 - merge-recursive: remove dead conditional in update_stages()
464 "log -p" output learns a new way to let users inspect a merge
465 commit by showing the differences between the automerged result
466 with conflicts the person who recorded the merge would have seen
467 and the final conflict resolution that was recorded in the merge.
469 Waiting for a reroll ($gmane/256591).
472 * hv/submodule-config (2014-11-11) 4 commits
473 - do not die on error of parsing fetchrecursesubmodules option
474 - use new config API for worktree configurations of submodules
475 - extract functions for submodule config set and lookup
476 - implement submodule config cache for lookup of submodule names
478 Kicked back to 'pu' per request ($gmane/255610).
481 * tg/perf-lib-test-perf-cleanup (2013-09-19) 2 commits
482 - perf-lib: add test_perf_cleanup target
483 - perf-lib: split starting the test from the execution
485 Add test_perf_cleanup shell function to the perf suite, that allows
486 the script writers to define a test with a clean-up action.
491 * jc/show-branch (2014-03-24) 5 commits
492 - show-branch: use commit slab to represent bitflags of arbitrary width
493 - show-branch.c: remove "all_mask"
494 - show-branch.c: abstract out "flags" operation
495 - show-branch.c: lift all_mask/all_revs to a global static
496 - show-branch.c: update comment style
498 Waiting for the final step to lift the hard-limit.
500 --------------------------------------------------
503 * ah/usage-strings (2015-04-02) 2 commits
504 - branch: fix funny-sounding error message
505 - standardize usage strings that were missed the first time
507 A few usage string updates.
509 Will merge to 'next'.
512 * jk/reading-packed-refs (2015-04-16) 9 commits
513 (merged to 'next' on 2015-04-21 at c9bce73)
514 + t1430: add another refs-escape test
515 + read_packed_refs: avoid double-checking sane refs
516 + strbuf_getwholeline: use getdelim if it is available
517 + strbuf_getwholeline: avoid calling strbuf_grow
518 + strbuf_addch: avoid calling strbuf_grow
519 + config: use getc_unlocked when reading from file
520 + strbuf_getwholeline: use getc_unlocked
521 + git-compat-util: add fallbacks for unlocked stdio
522 + strbuf_getwholeline: use getc macro
524 An earlier rewrite to use strbuf_getwholeline() instead of fgets(3)
525 to read packed-refs file revealed that the former is unacceptably
528 Will merge to 'master' in the second batch of post v2.4 cycle.
531 * kk/log-merges-config (2015-04-21) 5 commits
532 - bash-completion: add support for git-log --merges= and log.merges
533 - t4202-log: add tests for --merges=
534 - Documentation: add git-log --merges= option and log.merges config. var
535 - log: honor log.merges= option
536 - revision: add --merges={show|only|hide} option
538 "git log" (but not other commands in the "log" family) learned to
539 pay attention to the log.merges configuration variable that can be
540 set to "show" (the normal behaviour), "only" (hide non-merge
541 commits), or "hide" (hide merge commits). --merges=(show|only|hide)
542 can be used to override the setting from the command line.
544 The documentation may need to be updated once more.
547 * kn/cat-file-literally (2015-04-15) 5 commits
549 - t1006: add tests for git cat-file --literally
550 - cat-file: add documentation for '--literally' option.
551 - cat-file: teach cat-file a '--literally' option
552 - sha1_file.c: support reading from a loose object of unknown type
554 Add the "--literally" option to "cat-file" to allow inspecting
555 loose objects of an experimental or a broken type.
557 This is v8 $gmane/267213; getting closer to the final, I would say.
560 * ld/p4-filetype-detection (2015-04-04) 3 commits
561 (merged to 'next' on 2015-04-08 at da735b4)
562 + git-p4: fix filetype detection on files opened exclusively
563 + git-p4: small fix for locked-file-move-test
564 + git-p4: fix small bug in locked test scripts
566 Will merge to 'master' in the first batch of post v2.4 cycle.
569 * mg/httpd-tests-update-for-apache-2.4 (2015-04-08) 2 commits
570 - t/lib-git-svn: check same httpd module dirs as lib-httpd
571 - t/lib-httpd: load mod_unixd
573 This is the first two commits in a three-patch series $gmane/266962
576 * lm/squelch-bg-progress (2015-04-15) 2 commits
577 (merged to 'next' on 2015-04-21 at 9a8b1e9)
578 + compat/mingw: stubs for getpgid() and tcgetpgrp()
579 + progress: no progress in background
581 Many long-running operations show progress eye-candy, even when
582 they are later backgrounded. Hide the eye-candy when the process
583 is sent to the background instead.
585 Will merge to 'master' in the second batch of post v2.4 cycle.
588 * sb/test-bitmap-free-at-end (2015-04-12) 1 commit
589 (merged to 'next' on 2015-04-14 at 0ae4759)
590 + pack-bitmap.c: fix a memleak
592 Will merge to 'master' in the first batch of post v2.4 cycle.
595 * va/p4-client-path (2015-04-23) 2 commits
596 (merged to 'next' on 2015-04-23 at e0d2065)
597 + git-p4: improve client path detection when branches are used
598 + t9801: check git-p4's branch detection with client spec enabled
600 git p4 attempts to better handle branches in Perforce.
602 Will merge to 'master'.
605 * jk/at-push-sha1 (2015-03-31) 6 commits
606 - sha1_name: implement @{push} shorthand
607 - sha1_name: refactor upstream_mark
608 - remote.c: provide per-branch pushremote name
609 - remote.c: hoist branch.*.remote lookup out of remote_get_1
610 - remote.c: drop "remote" pointer from "struct branch"
611 - remote.c: drop default_remote_name variable
613 Introduce <branch>@{push} short-hand to denote the remote-tracking
614 branch that tracks the branch at the remote the <branch> would be
617 Waiting for a reroll ($gmane/266573).
620 * jk/sha1-file-reduce-useless-warnings (2015-03-30) 1 commit
621 (merged to 'next' on 2015-04-21 at a8bd995)
622 + sha1_file: squelch "packfile cannot be accessed" warnings
624 Will merge to 'master' in the second batch of post v2.4 cycle.
627 * mh/show-branch-topic (2015-03-31) 1 commit
628 (merged to 'next' on 2015-04-08 at 628ada5)
629 + show-branch: show all local heads when only giving one rev along --topics
631 "git show-branch --topics HEAD" (with no other arguments) did not
632 do anything interesting. Instead, contrast the given revision
633 against all the local branches by default.
635 Will merge to 'master' in the first batch of post v2.4 cycle.
638 * sb/line-log-plug-pairdiff-leak (2015-03-30) 1 commit
639 (merged to 'next' on 2015-04-08 at 6ea8b71)
640 + line-log.c: fix a memleak
642 Will merge to 'master' in the first batch of post v2.4 cycle.
645 * jk/init-core-worktree-at-root (2015-04-02) 1 commit
646 (merged to 'next' on 2015-04-08 at 92a942b)
647 + init: don't set core.worktree when initializing /.git
649 We avoid setting core.worktree when the repository location is the
650 ".git" directory directly at the top level of the working tree, but
651 the code misdetected the case in which the working tree is at the
652 root level of the filesystem (which arguably is a silly thing to
653 do, but still valid).
655 Will merge to 'master' in the first batch of post v2.4 cycle.
658 * ts/checkout-advice-plural (2015-04-02) 1 commit
659 (merged to 'next' on 2015-04-08 at dd5c436)
660 + checkout: call a single commit "it" intead of "them"
662 Will merge to 'master' in the first batch of post v2.4 cycle.
665 * pt/credential-xdg (2015-03-25) 4 commits
666 (merged to 'next' on 2015-03-25 at 765128e)
667 + t0302: "unreadable" test needs POSIXPERM
668 (merged to 'next' on 2015-03-24 at 9a3706e)
669 + t0302: test credential-store support for XDG_CONFIG_HOME
670 + git-credential-store: support XDG_CONFIG_HOME
671 + git-credential-store: support multiple credential files
672 (this branch is used by pt/xdg-config-path.)
674 Tweak the sample "store" backend of the credential helper to honor
675 XDG configuration file locations when specified.
677 As this may see further updates, and also this is not an urgent
678 issue anyway, I'll give it a bit more time for it to simmer.
683 * pt/xdg-config-path (2015-04-12) 7 commits
684 - path.c: remove home_config_paths()
685 - git-config: replace use of home_config_paths()
686 - git-commit: replace use of home_config_paths()
687 - credential-store.c: replace home_config_paths() with xdg_config_home()
688 - dir.c: replace home_config_paths() with xdg_config_home()
689 - attr.c: replace home_config_paths() with xdg_config_home()
690 - path.c: implement xdg_config_home()
691 (this branch uses pt/credential-xdg.)
693 Seen some discussions.
694 Waiting for a reroll ($gmane/267518).
697 * jc/diff-no-index-d-f (2015-03-26) 2 commits
698 (merged to 'next' on 2015-03-27 at 1f270f9)
699 + diff-no-index: align D/F handling with that of normal Git
700 + diff-no-index: DWIM "diff D F" into "diff D/F F"
702 The usual "git diff" when seeing a file turning into a directory
703 showed a patchset to remove the file and create all files in the
704 directory, but "git diff --no-index" simply refused to work. Also,
705 when asked to compare a file and a directory, imitate POSIX "diff"
706 and compare the file with the file with the same name in the
707 directory, instead of refusing to run.
709 Will merge to 'master' in the first batch of post v2.4 cycle.
712 * nd/diff-i-t-a (2015-03-23) 1 commit
713 (merged to 'next' on 2015-03-24 at a6be89f)
714 + diff-lib.c: adjust position of i-t-a entries in diff
716 After "git add -N", the path appeared in output of "git diff HEAD"
717 and "git diff --cached HEAD", leading "git status" to classify it
718 as "Changes to be committed". Such a path, however, is not yet to
719 be scheduled to be committed. "git diff" showed the change to the
720 path as modification, not as a "new file", in the header of its
723 Treat such paths as "yet to be added to the index but Git already
724 know about them"; "git diff HEAD" and "git diff --cached HEAD"
725 should not talk about them, and "git diff" should show them as new
726 files yet to be added to the index.
728 Will cook in 'next', as this brings in a new world order.
731 * mh/numparse (2015-03-19) 14 commits
732 - diff_opt_parse(): use convert_i() when handling --abbrev=<num>
733 - diff_opt_parse(): use convert_i() when handling "-l<num>"
734 - opt_arg(): simplify pointer handling
735 - opt_arg(): report errors parsing option values
736 - opt_arg(): use convert_i() in implementation
737 - opt_arg(): val is always non-NULL
738 - builtin_diff(): detect errors when parsing --unified argument
739 - handle_revision_opt(): use convert_ui() when handling "--abbrev="
740 - strtoul_ui(), strtol_i(): remove functions
741 - handle_revision_opt(): use convert_i() when handling "-<digit>"
742 - handle_revision_opt(): use skip_prefix() in many places
743 - write_subdirectory(): use convert_ui() for parsing mode
744 - cacheinfo_callback(): use convert_ui() when handling "--cacheinfo"
745 - numparse: new module for parsing integral numbers
747 Many codepaths use unchecked use of strtol() and friends (or even
748 worse, atoi()). Introduce a set of wrappers that try to be more
751 Waiting for a reroll ($gmane/266209).
754 * tf/gitweb-project-listing (2015-03-19) 5 commits
755 - gitweb: make category headings into links when they are directories
756 - gitweb: optionally set project category from its pathname
757 - gitweb: add a link under the search box to clear a project filter
758 - gitweb: if the PATH_INFO is incomplete, use it as a project_filter
759 - gitweb: fix typo in man page
761 Update gitweb to make it more pleasant to deal with a hierarchical
762 forest of repositories.
764 Any comments from those who use or have their own code in Gitweb?
767 * jc/a-lone-dash-stands-for-previous-branch (2015-03-16) 1 commit
768 - "-" and "@{-1}" on various programs
770 Lose special case code to make a lone dash "-" mean the previous
771 branch aka "@{-1}" from a handful subcommands, and instead support
772 the notation throughout the system by reimplementing it at the
775 Needs tests, documentation updates, etc. Also does only a half-way
776 job dealing with range notation, which needs to be fixed before the
777 series goes anywhere.
780 * bc/object-id (2015-03-13) 10 commits
781 (merged to 'next' on 2015-03-24 at 3ec4f83)
782 + apply: convert threeway_stage to object_id
783 + patch-id: convert to use struct object_id
784 + commit: convert parts to struct object_id
785 + diff: convert struct combine_diff_path to object_id
786 + bulk-checkin.c: convert to use struct object_id
787 + zip: use GIT_SHA1_HEXSZ for trailers
788 + archive.c: convert to use struct object_id
789 + bisect.c: convert leaf functions to use struct object_id
790 + define utility functions for object IDs
791 + define a structure for object IDs
793 Identify parts of the code that knows that we use SHA-1 hash to
794 name our objects too much, and use (1) symbolic constants instead
795 of hardcoded 20 as byte count and/or (2) use struct object_id
796 instead of unsigned char [20] for object names.
799 Also waiting for reviews on the follow-up series ($gmane/265939)
800 that seems to have seen no interest X-<.
803 * nd/slim-index-pack-memory-usage (2015-04-18) 2 commits
804 (merged to 'next' on 2015-04-21 at 919b80d)
805 + index-pack: kill union delta_base to save memory
806 + index-pack: reduce object_entry size to save memory
808 Memory usage of "git index-pack" has been trimmed by tens of
811 Will merge to 'master' in the second batch of post v2.4 cycle.
814 * nd/list-files (2015-02-09) 21 commits
815 - t3080: tests for git-list-files
816 - list-files: -M aka diff-cached
817 - list-files -F: show submodules with the new indicator '&'
818 - list-files: add -F/--classify
819 - list-files: show directories as well as files
820 - list-files: do not show duplicate cached entries
821 - list-files: sort output and remove duplicates
822 - list-files: add -t back
823 - list-files: add -1 short for --no-column
824 - list-files: add -R/--recursive short for --max-depth=-1
825 - list-files: -u does not imply showing stages
826 - list-files: make alias 'ls' default to 'list-files'
827 - list-files: a user friendly version of ls-files and more
828 - ls-files: support --max-depth
829 - ls-files: add --column
830 - ls-files: add --color to highlight file names
831 - ls-files: buffer full item in strbuf before printing
832 - ls_colors.c: highlight submodules like directories
833 - ls_colors.c: add a function to color a file name
834 - ls_colors.c: parse color.ls.* from config file
835 - ls_colors.c: add $LS_COLORS parsing code
837 A new "git list-files" Porcelain command, "ls-files" with bells and
840 Reroll to base on wt-status work ($gmane/265142) has seen some
841 positive discussions.
843 Waiting for a further polished reroll ($gmane/265534).
846 * js/fsck-opt (2015-01-21) 19 commits
847 - fsck: support ignoring objects in `git fsck` via fsck.skiplist
848 - fsck: git receive-pack: support excluding objects from fsck'ing
849 - fsck: introduce `git fsck --quick`
850 - fsck: support demoting errors to warnings
851 - fsck: document the new receive.fsck.* options
852 - fsck: allow upgrading fsck warnings to errors
853 - fsck: optionally ignore specific fsck issues completely
854 - fsck: disallow demoting grave fsck errors to warnings
855 - fsck: add a simple test for receive.fsck.*
856 - fsck: make fsck_tag() warn-friendly
857 - fsck: handle multiple authors in commits specially
858 - fsck: make fsck_commit() warn-friendly
859 - fsck: make fsck_ident() warn-friendly
860 - fsck: report the ID of the error/warning
861 - fsck: allow demoting errors to warnings via receive.fsck.warn = <key>
862 - fsck: offer a function to demote fsck errors to warnings
863 - fsck: provide a function to parse fsck message IDs
864 - fsck: introduce identifiers for fsck messages
865 - fsck: introduce fsck options
867 "fsck.warnings = <list of error tokens>" I suggested turned out to
868 be an unpopular choice (sorry Dscho).
873 * nd/multiple-work-trees (2015-03-31) 41 commits
874 (merged to 'next' on 2015-04-02 at 7ea51b4)
875 + prune --worktrees: fix expire vs worktree existence condition
876 (merged to 'next' on 2015-03-24 at 58b5a60)
877 + t1501: fix test with split index
878 (merged to 'next' on 2015-03-20 at cc98ed0)
879 + t2026: fix broken &&-chain
880 (merged to 'next' on 2015-02-18 at b51f696)
881 + t2026 needs procondition SANITY
882 + git-checkout.txt: a note about multiple checkout support for submodules
883 + checkout: add --ignore-other-wortrees
884 + checkout: pass whole struct to parse_branchname_arg instead of individual flags
885 + git-common-dir: make "modules/" per-working-directory directory
886 + checkout: do not fail if target is an empty directory
887 + t2025: add a test to make sure grafts is working from a linked checkout
888 + checkout: don't require a work tree when checking out into a new one
889 + git_path(): keep "info/sparse-checkout" per work-tree
890 + count-objects: report unused files in $GIT_DIR/worktrees/...
891 + gc: support prune --worktrees
892 + gc: factor out gc.pruneexpire parsing code
893 + gc: style change -- no SP before closing parenthesis
894 + checkout: clean up half-prepared directories in --to mode
895 + checkout: reject if the branch is already checked out elsewhere
896 + prune: strategies for linked checkouts
897 + checkout: support checking out into a new working directory
898 + use new wrapper write_file() for simple file writing
899 + wrapper.c: wrapper to open a file, fprintf then close
900 + setup.c: support multi-checkout repo setup
901 + setup.c: detect $GIT_COMMON_DIR check_repository_format_gently()
902 + setup.c: convert check_repository_format_gently to use strbuf
903 + setup.c: detect $GIT_COMMON_DIR in is_git_directory()
904 + setup.c: convert is_git_directory() to use strbuf
905 + git-stash: avoid hardcoding $GIT_DIR/logs/....
906 + *.sh: avoid hardcoding $GIT_DIR/hooks/...
907 + git-sh-setup.sh: use rev-parse --git-path to get $GIT_DIR/objects
908 + $GIT_COMMON_DIR: a new environment variable
909 + commit: use SEQ_DIR instead of hardcoding "sequencer"
910 + fast-import: use git_path() for accessing .git dir instead of get_git_dir()
911 + reflog: avoid constructing .lock path with git_path
912 + *.sh: respect $GIT_INDEX_FILE
913 + git_path(): be aware of file relocation in $GIT_DIR
914 + path.c: group git_path(), git_pathdup() and strbuf_git_path() together
915 + path.c: rename vsnpath() to do_git_path()
916 + git_snpath(): retire and replace with strbuf_git_path()
917 + path.c: make get_pathname() call sites return const char *
918 + path.c: make get_pathname() return strbuf instead of static buffer
920 A replacement for contrib/workdir/git-new-workdir that does not
921 rely on symbolic links and make sharing of objects and refs safer
922 by making the borrowee and borrowers aware of each other.
926 --------------------------------------------------
929 * as/userdiff-sh (2015-03-13) 1 commit
930 . userdiff: funcname and word patterns for sh
932 Add a built-in "userdiff" patterns to word-split and identify
933 notable lines in shell scripts to help presentation of diff and
937 * sb/ref-lock-avoid-running-out-of-fds (2015-04-23) 5 commits
938 . refs.c: enable large transactions
939 . sha1_file.c: move get_max_fd_limit(void) to wrapper.c
940 - refs.c: remove lock_fd from struct ref_lock
941 - t7004: rename ULIMIT test prerequisite to ULIMIT_STACK_SIZE
942 - update-ref: test handling large transactions properly
943 (this branch is tangled with mh/ref-lock-avoid-running-out-of-fds.)
946 * jc/merge-deprecate-old-syntax (2015-03-26) 1 commit
947 . merge: deprecate 'git merge <message> HEAD <commit>' syntax
949 The first step to really start the process of removing the ancient
950 syntax to invoke a two-way merge, which has been deprecated since
953 This regresses "git pull" in various ways; do not merge ($gmane/267432).
956 * jc/merge-drop-old-syntax (2015-04-25) 1 commit
957 - merge: drop 'git merge <message> HEAD <commit>' syntax
958 (this branch uses jc/merge.)
960 Stop supporting "git merge <messsage> HEAD <commit>" syntax that
961 has been deprecated since October 2007.
963 This regresses "git pull" in various ways; do not merge ($gmane/267432).