1 To: git@vger.kernel.org
3 Subject: What's cooking in git.git (Apr 2014, #06; Fri, 18)
4 X-master-at: cc291953df19aa4a97bee3590e708dc1fc557500
5 X-next-at: 0bd75c508e7c03265459d39a0fa838a04b338cc4
7 What's cooking in git.git (Apr 2014, #06; Fri, 18)
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 The tip of the 'master' branch is at v2.0.0-rc0, an early preview
15 release. There are a handful of topics still in 'next' (and a few
16 that are not in 'next') that I'd like to have in v2.0.0-rc1, but
17 other than that, this hopefully is fairly a close approximation of
20 Also I am still waiting for acks for a few topics before merging
21 them to 'next'. I would feel it would be good if we can merge them
22 early to 'next' for wider and longer exposure and some of them may
23 even deserve to be in -rc1, but as none of them is a regression fix,
24 it is also OK to wait until 2.0 final.
26 You can find the changes described here in the integration branches
27 of the repositories listed at
29 http://git-blame.blogspot.com/p/git-public-repositories.html
31 --------------------------------------------------
32 [Graduated to "master"]
34 * fc/complete-aliased-push (2014-04-09) 1 commit
35 (merged to 'next' on 2014-04-16 at ab798d1)
36 + completion: fix completing args of aliased "push", "fetch", etc.
39 * fc/prompt-zsh-read-from-file (2014-04-14) 1 commit
40 (merged to 'next' on 2014-04-16 at 0e5fec0)
41 + prompt: fix missing file errors in zsh
44 * fc/remote-helper-fixes (2014-04-14) 5 commits
45 (merged to 'next' on 2014-04-16 at 180482e)
46 + remote-bzr: trivial test fix
47 + remote-bzr: include authors field in pushed commits
48 + remote-bzr: add support for older versions
49 + remote-hg: always normalize paths
50 + remote-helpers: allow all tests running from any dir
53 * jk/config-die-bad-number-noreturn (2014-04-16) 1 commit
54 (merged to 'next' on 2014-04-16 at 4d49036)
55 + config.c: mark die_bad_number as NORETURN
57 Squelch a false compiler warning from older gcc.
59 --------------------------------------------------
62 * fc/publish-vs-upstream (2014-04-11) 8 commits
63 - sha1_name: add support for @{publish} marks
64 - sha1_name: simplify track finding
65 - sha1_name: cleanup interpret_branch_name()
66 - branch: display publish branch
67 - push: add --set-publish option
68 - branch: add --set-publish-to option
69 - Add concept of 'publish' branch
70 - t5516 (fetch-push): fix test restoration
72 Add branch@{publish}; it seems that this is somewhat different from
73 Ram and Peff started working on. There are still many discussion
74 messages going back and forth but not updates to the patches.
76 Seems to have some interactions to break t5516 when merged to 'pu'.
79 * tr/merge-recursive-index-only (2014-02-05) 3 commits
80 - merge-recursive: -Xindex-only to leave worktree unchanged
81 - merge-recursive: internal flag to avoid touching the worktree
82 - merge-recursive: remove dead conditional in update_stages()
83 (this branch is used by tr/remerge-diff.)
88 * tr/remerge-diff (2014-02-26) 5 commits
89 . log --remerge-diff: show what the conflict resolution changed
90 . name-hash: allow dir hashing even when !ignore_case
91 . merge-recursive: allow storing conflict hunks in index
92 . revision: fold all merge diff variants into an enum merge_diff_mode
93 . combine-diff: do not pass revs->dense_combined_merges redundantly
94 (this branch uses tr/merge-recursive-index-only.)
96 "log -p" output learns a new way to let users inspect a merge
97 commit by showing the differences between the automerged result
98 with conflicts the person who recorded the merge would have seen
99 and the final conflict resolution that was recorded in the merge.
101 Needs to be rebased, now kb/fast-hashmap topic is in.
104 * bc/blame-crlf-test (2014-02-18) 1 commit
105 - blame: add a failing test for a CRLF issue.
107 I have a feeling that a fix for this should be fairly isolated and
108 trivial (it should be just the matter of paying attention to the
109 crlf settings when synthesizing the fake commit)---perhaps somebody
110 can squash in a fix to this?
113 * jk/makefile (2014-02-05) 16 commits
115 - move LESS/LV pager environment to Makefile
116 - Makefile: teach scripts to include make variables
118 - Makefile: auto-build C strings from make variables
119 - Makefile: drop *_SQ variables
121 - Makefile: add c-quote helper function
122 - Makefile: introduce sq function for shell-quoting
123 - Makefile: always create files via make-var
124 - Makefile: store GIT-* sentinel files in MAKE/
125 - Makefile: prefer printf to echo for GIT-*
126 - Makefile: use tempfile/mv strategy for GIT-*
127 - Makefile: introduce make-var helper function
128 - Makefile: fix git-instaweb dependency on gitweb
129 - Makefile: drop USE_GETTEXT_SCHEME from GIT-CFLAGS
131 Simplify the Makefile rules and macros that exist primarily for
132 quoting purposes, and make it easier to robustly express the
138 * po/everyday-doc (2014-01-27) 1 commit
139 - Make 'git help everyday' work
141 This may make the said command to emit something, but the source is
142 not meant to be formatted into a manual pages to begin with, and
143 also its contents are a bit stale. It may be a good first step in
144 the right direction, but needs more work to at least get the
145 mark-up right before public consumption.
150 * jk/branch-at-publish-rebased (2014-01-17) 5 commits
151 . t1507 (rev-parse-upstream): fix typo in test title
152 . implement @{publish} shorthand
153 . branch_get: provide per-branch pushremote pointers
154 . branch_get: return early on error
155 . sha1_name: refactor upstream_mark
157 Give an easier access to the tracking branches from "other" side in
158 a triangular workflow by introducing B@{publish} that works in a
159 similar way to how B@{upstream} does.
161 Meant to be used as a basis for whatever Ram wants to build on.
163 Ejected from 'pu' to make room for fc/publish-vs-upstream topic.
166 * rb/merge-prepare-commit-msg-hook (2014-01-10) 4 commits
167 - merge: drop unused arg from abort_commit method signature
168 - merge: make prepare_to_commit responsible for write_merge_state
169 - t7505: ensure cleanup after hook blocks merge
170 - t7505: add missing &&
172 Expose more merge states (e.g. $GIT_DIR/MERGE_MODE) to hooks that
173 run during "git merge". The log message stresses too much on one
174 hook, prepare-commit-msg, but it would equally apply to other hooks
175 like post-merge, I think.
177 Waiting for a reroll.
180 * jl/submodule-recursive-checkout (2013-12-26) 5 commits
181 - Teach checkout to recursively checkout submodules
182 - submodule: teach unpack_trees() to update submodules
183 - submodule: teach unpack_trees() to repopulate submodules
184 - submodule: teach unpack_trees() to remove submodule contents
185 - submodule: prepare for recursive checkout of submodules
187 An RFCv2 exists ($gmane/241455) with sizable review comments.
191 * jc/graph-post-root-gap (2013-12-30) 3 commits
192 - WIP: document what we want at the end
193 - graph: remove unused code a bit
194 - graph: stuff the current commit into graph->columns[]
196 This was primarily a RFH ($gmane/239580).
199 * np/pack-v4 (2013-09-18) 90 commits
200 . packv4-parse.c: add tree offset caching
201 . t1050: replace one instance of show-index with verify-pack
202 . index-pack, pack-objects: allow creating .idx v2 with .pack v4
203 . unpack-objects: decode v4 trees
204 . unpack-objects: allow to save processed bytes to a buffer
207 Nico and Duy advancing the eternal vaporware pack-v4. This is here
208 primarily for wider distribution of the preview edition.
210 Needs to be rebased, now the pack-bitmap series is in.
213 * tg/perf-lib-test-perf-cleanup (2013-09-19) 2 commits
214 - perf-lib: add test_perf_cleanup target
215 - perf-lib: split starting the test from the execution
217 Add test_perf_cleanup shell function to the perf suite, that allows
218 the script writers to define a test with a clean-up action.
223 * jc/format-patch (2013-04-22) 2 commits
224 - format-patch: --inline-single
225 - format-patch: rename "no_inline" field
227 A new option to send a single patch to the standard output to be
228 appended at the bottom of a message. I personally have no need for
229 this, but it was easy enough to cobble together. Tests, docs and
230 stripping out more MIMEy stuff are left as exercises to interested
234 * jc/show-branch (2014-03-24) 5 commits
235 - show-branch: use commit slab to represent bitflags of arbitrary width
236 - show-branch.c: remove "all_mask"
237 - show-branch.c: abstract out "flags" operation
238 - show-branch.c: lift all_mask/all_revs to a global static
239 - show-branch.c: update comment style
241 Waiting for the final step to lift the hard-limit before sending it out.
243 --------------------------------------------------
246 * jl/git-gui-show-added-submodule-changes (2014-04-15) 1 commit
247 - git-gui: show staged submodules regardless of ignore config
249 Tentatively queued what I expect to receive via Pat Thoyts.
252 * jl/gitk-show-added-submodule-changes (2014-04-15) 3 commits
253 - gitk: show staged submodules regardless of ignore config
254 - gitk: Merge branch 'new' of https://github.com/vnwildman/gitk
255 - l10n: Init Vietnamese translation
257 Tentatively queued what I expect to receive via Paul Mackerras.
260 * bg/rebase-off-of-previous-branch (2014-04-16) 1 commit
261 - git-rebase: print name of rev when using shorthand
263 Teach "git rebase -" to report the concrete name of the branch
264 (i.e. the previous one).
266 But it stops short and does not do the same for "git rebase @{-1}".
270 * ef/send-email-absolute-path-to-the-command (2014-04-16) 1 commit
271 - send-email: recognize absolute path on Windows
273 Will merge to 'next'.
276 * ep/shell-command-substitution (2014-04-17) 14 commits
277 (merged to 'next' on 2014-04-18 at ebd996d)
278 + t9362-mw-to-git-utf8.sh: use the $( ... ) construct for command substitution
279 + t9360-mw-to-git-clone.sh: use the $( ... ) construct for command substitution
280 + git-tag.sh: use the $( ... ) construct for command substitution
281 + git-revert.sh: use the $( ... ) construct for command substitution
282 + git-resolve.sh: use the $( ... ) construct for command substitution
283 + git-repack.sh: use the $( ... ) construct for command substitution
284 + git-merge.sh: use the $( ... ) construct for command substitution
285 + git-ls-remote.sh: use the $( ... ) construct for command substitution
286 + git-fetch.sh: use the $( ... ) construct for command substitution
287 + git-commit.sh: use the $( ... ) construct for command substitution
288 + git-clone.sh: use the $( ... ) construct for command substitution
289 + git-checkout.sh: use the $( ... ) construct for command substitution
290 + install-webdoc.sh: use the $( ... ) construct for command substitution
291 + howto-index.sh: use the $( ... ) construct for command substitution
293 Will merge to 'master' by -rc1 if no regressions are reported.
296 * jh/submodule-tests (2014-04-17) 1 commit
297 - t7410: 210 tests for various 'git submodule update' scenarios
300 * jx/i18n (2014-04-18) 4 commits
301 (merged to 'next' on 2014-04-18 at 2ace126)
302 + i18n: mention "TRANSLATORS:" marker in Documentation/CodingGuidelines
303 + i18n: only extract comments marked with "TRANSLATORS:"
304 + i18n: remove obsolete comments for translators in diffstat generation
305 + i18n: fix uncatchable comments for translators in date.c
307 Will merge to 'master' by -rc1 if no regressions are reported.
310 * rs/ref-update-check-errors-early (2014-04-17) 2 commits
311 - commit.c: check for lock error and return early
312 - sequencer.c: check for lock failure and bail early in fast_forward_to
314 Will merge to 'next' and keep it there for the remainder of the cycle.
317 * sk/svn-parse-datestamp (2014-04-17) 1 commit
318 - SVN.pm::parse_svn_date: allow timestamps with a single-digit hour
320 Will merge to 'next' and keep it there for the remainder of the cycle.
323 * fc/transport-helper-sync-error-fix (2014-04-14) 5 commits
324 - transport-helper: fix sync issue on crashes
325 - transport-helper: trivial cleanup
326 - transport-helper: propagate recvline() error pushing
327 - remote-helpers: make recvline return an error
328 - transport-helper: remove barely used xchgline()
330 Make sure the marks are not written out when the transport helper
331 did not finish happily, to avoid marks file that is out of sync
334 As I had to resolve some conflicts with the recent code, will wait
335 until getting an OK from Felipe on the conflict resolution and then
336 hopefully can be merged to 'next' and then later to 'master'.
339 * km/avoid-non-function-return-in-rebase (2014-04-17) 2 commits
340 (merged to 'next' on 2014-04-18 at 6982c3e)
341 + Revert "rebase: fix run_specific_rebase's use of "return" on FreeBSD"
342 + rebase: avoid non-function use of "return" on FreeBSD
344 Work around /bin/sh that does not like "return" at the top-level
345 of a file that is dot-sourced from inside a function definition.
347 Will merge to 'master' by -rc1 if no regressions are reported.
350 * db/make-with-curl (2014-04-15) 2 commits
351 (merged to 'next' on 2014-04-16 at b9c8527)
352 + Makefile: allow static linking against libcurl
353 + Makefile: use curl-config to determine curl flags
355 Ask curl-config how to link with the curl library, instead of
356 having only a limited configurability knobs in the Makefile.
358 Will merge to 'master' by -rc1 if no regressions are reported.
361 * nd/index-pack-one-fd-per-thread (2014-04-16) 1 commit
362 (merged to 'next' on 2014-04-16 at b38d5a9)
363 + index-pack: work around thread-unsafe pread()
365 Enable threaded index-pack on platforms without thread-unsafe
368 Will keep in 'next' for the remainder of the cycle.
371 * ym/fix-opportunistic-index-update-race (2014-04-10) 2 commits
372 (merged to 'next' on 2014-04-16 at cb92f4f)
373 + read-cache.c: verify index file before we opportunistically update it
374 + wrapper.c: add xpread() similar to xread()
376 Read-only operations such as "git status" that internally refreshes
377 the index write out the refreshed index to the disk to optimize
378 future accesses to the working tree, but this could race with a
379 "read-write" operation that modify the index while it is running.
380 Detect such a race and avoid overwriting the index.
382 Duy raised a good point that we may need to do the same for the
383 normal writeout codepath, not just the "opportunistic" update
384 codepath. While that is true, nobody sane would be running two
385 simultaneous operations that are clearly write-oriented competing
386 with each other against the same index file. So in that sense that
387 can be done as a less urgent follow-up for this topic.
389 Will keep in 'next' for the remainder of the cycle.
392 * jl/status-added-submodule-is-never-ignored (2014-04-07) 2 commits
393 - commit -m: commit staged submodules regardless of ignore config
394 - status/commit: show staged submodules regardless of ignore config
396 There also are a few patches Ronald Weiss and Jens are working on
397 polishing around this topic, and a patch from Jens each for gitk
400 Waiting for the dust to settle until picking them up all.
403 * mh/lockfile (2014-04-15) 25 commits
404 - trim_last_path_elm(): replace last_path_elm()
405 - resolve_symlink(): take a strbuf parameter
406 - resolve_symlink(): use a strbuf for internal scratch space
407 - change lock_file::filename into a strbuf
408 - commit_lock_file(): use a strbuf to manage temporary space
409 - try_merge_strategy(): use a statically-allocated lock_file object
410 - try_merge_strategy(): remove redundant lock_file allocation
411 - struct lock_file: declare some fields volatile
412 - lockfile: avoid transitory invalid states
413 - commit_lock_file(): die() if called for unlocked lockfile object
414 - commit_lock_file(): inline temporary variable
415 - remove_lock_file(): call rollback_lock_file()
416 - lock_file(): exit early if lockfile cannot be opened
417 - write_packed_entry_fn(): convert cb_data into a (const int *)
418 - prepare_index(): declare return value to be (const char *)
419 - delete_ref_loose(): don't muck around in the lock_file's filename
420 - cache.h: define constants LOCK_SUFFIX and LOCK_SUFFIX_LEN
421 - lockfile.c: document the various states of lock_file objects
422 - lock_file(): always add lock_file object to lock_file_list
423 - hold_lock_file_for_append(): release lock on errors
424 - lockfile: unlock file if lockfile permissions cannot be adjusted
425 - rollback_lock_file(): set fd to -1
426 - rollback_lock_file(): do not clear filename redundantly
427 - api-lockfile: expand the documentation
428 - unable_to_lock_die(): rename function from unable_to_lock_index_die()
430 Refactor and fix corner-case bugs in the lockfile API, all looked
436 * mt/patch-id-stable (2014-03-31) 3 commits
437 (merged to 'next' on 2014-04-08 at 0188d44)
438 + patch-id-test: test --stable and --unstable flags
439 + patch-id: document new behaviour
440 + patch-id: make it stable against hunk reordering
442 Introduce a new way to compute patch-id for a patch that is not
443 affected by the order of the paths that appear in the input.
445 This changes the generated patch-id unless the users add an extra
446 option to their command lines, but I deliberately queued the series
447 to 'next' without reverting that compatibility breakage to see if
448 people complain. It could be that we do not have to worry about
449 the default flipping at all. We'll see.
451 Will keep in 'next' for the remainder of the cycle.
454 * mh/ref-transaction (2014-04-07) 27 commits
455 (merged to 'next' on 2014-04-16 at a99f84d)
456 + ref_transaction_commit(): work with transaction->updates in place
457 + struct ref_update: add a type field
458 + struct ref_update: add a lock field
459 + ref_transaction_commit(): simplify code using temporary variables
460 + struct ref_update: store refname as a FLEX_ARRAY
461 + struct ref_update: rename field "ref_name" to "refname"
462 + refs: remove API function update_refs()
463 + update-ref --stdin: reimplement using reference transactions
464 + refs: add a concept of a reference transaction
465 + update-ref --stdin: harmonize error messages
466 + update-ref --stdin: improve the error message for unexpected EOF
467 + t1400: test one mistake at a time
468 + update-ref --stdin -z: deprecate interpreting the empty string as zeros
469 + update-ref.c: extract a new function, parse_next_sha1()
470 + t1400: test that stdin -z update treats empty <newvalue> as zeros
471 + update-ref --stdin: simplify error messages for missing oldvalues
472 + update-ref --stdin: make error messages more consistent
473 + update-ref --stdin: improve error messages for invalid values
474 + update-ref.c: extract a new function, parse_refname()
475 + parse_cmd_verify(): copy old_sha1 instead of evaluating <oldvalue> twice
476 + update-ref --stdin: read the whole input at once
477 + update_refs(): fix constness
478 + refs.h: rename the action_on_err constants
479 + t1400: add some more tests involving quoted arguments
480 + parse_arg(): really test that argument is properly terminated
481 + t1400: provide more usual input to the command
482 + t1400: fix name and expected result of one test
484 Update "update-ref --stdin [-z]" and then introduce a transactional
485 support for (multi-)reference updates.
487 Will keep in 'next' for the remainder of the cycle.
490 * jc/apply-ignore-whitespace (2014-03-26) 1 commit
491 (merged to 'next' on 2014-04-04 at 53779a7)
492 + apply --ignore-space-change: lines with and without leading whitespaces do not match
494 "--ignore-space-change" option of "git apply" ignored the
495 spaces at the beginning of line too aggressively, which is
496 inconsistent with the option of the same name "diff" and "git diff"
499 Will keep in 'next' for the remainder of the cycle.
502 * as/grep-fullname-config (2014-03-20) 1 commit
503 (merged to 'next' on 2014-03-28 at 810a076)
504 + grep: add grep.fullName config variable
506 Add a configuration variable to force --full-name to be default for
509 This may cause regressions on scripted users that do not expect
512 Will keep in 'next' for the remainder of the cycle.
515 * nd/multiple-work-trees (2014-03-25) 28 commits
516 - count-objects: report unused files in $GIT_DIR/repos/...
517 - gc: support prune --repos
518 - gc: style change -- no SP before closing bracket
519 - prune: strategies for linked checkouts
520 - checkout: detach if the branch is already checked out elsewhere
521 - checkout: clean up half-prepared directories in --to mode
522 - checkout: support checking out into a new working directory
523 - use new wrapper write_file() for simple file writing
524 - wrapper.c: wrapper to open a file, fprintf then close
525 - setup.c: support multi-checkout repo setup
526 - setup.c: detect $GIT_COMMON_DIR check_repository_format_gently()
527 - setup.c: convert check_repository_format_gently to use strbuf
528 - setup.c: detect $GIT_COMMON_DIR in is_git_directory()
529 - setup.c: convert is_git_directory() to use strbuf
530 - git-stash: avoid hardcoding $GIT_DIR/logs/....
531 - *.sh: avoid hardcoding $GIT_DIR/hooks/...
532 - git-sh-setup.sh: use rev-parse --git-path to get $GIT_DIR/objects
533 - $GIT_COMMON_DIR: a new environment variable
534 - commit: use SEQ_DIR instead of hardcoding "sequencer"
535 - fast-import: use git_path() for accessing .git dir instead of get_git_dir()
536 - reflog: avoid constructing .lock path with git_path
537 - *.sh: respect $GIT_INDEX_FILE
538 - git_path(): be aware of file relocation in $GIT_DIR
539 - path.c: group git_path(), git_pathdup() and strbuf_git_path() together
540 - path.c: rename vsnpath() to do_git_path()
541 - git_snpath(): retire and replace with strbuf_git_path()
542 - path.c: make get_pathname() call sites return const char *
543 - path.c: make get_pathname() return strbuf instead of static buffer
545 A replacement for contrib/workdir/git-new-workdir that does not
546 rely on symbolic links and make sharing of objects and refs safer
547 by making the borrowee and borrowers aware of each other.
552 * ks/tree-diff-nway (2014-04-09) 20 commits
553 (merged to 'next' on 2014-04-09 at c17228e)
554 + mingw: activate alloca
555 (merged to 'next' on 2014-04-08 at 6b74773)
556 + combine-diff: speed it up, by using multiparent diff tree-walker directly
557 + tree-diff: rework diff_tree() to generate diffs for multiparent cases as well
558 + Portable alloca for Git
559 (merged to 'next' on 2014-03-31 at 16a7bd4)
560 + tree-diff: reuse base str(buf) memory on sub-tree recursion
561 + tree-diff: no need to call "full" diff_tree_sha1 from show_path()
562 + tree-diff: rework diff_tree interface to be sha1 based
563 + tree-diff: diff_tree() should now be static
564 + tree-diff: remove special-case diff-emitting code for empty-tree cases
565 (merged to 'next' on 2014-03-25 at cfcbdac)
566 + tree-diff: simplify tree_entry_pathcmp
567 + tree-diff: show_path prototype is not needed anymore
568 + tree-diff: rename compare_tree_entry -> tree_entry_pathcmp
569 + tree-diff: move all action-taking code out of compare_tree_entry()
570 + tree-diff: don't assume compare_tree_entry() returns -1,0,1
571 (merged to 'next' on 2014-03-21 at d872679)
572 + tree-diff: consolidate code for emitting diffs and recursion in one place
573 + tree-diff: show_tree() is not needed
574 + tree-diff: no need to pass match to skip_uninteresting()
575 + tree-diff: no need to manually verify that there is no mode change for a path
576 + combine-diff: move changed-paths scanning logic into its own function
577 + combine-diff: move show_log_first logic/action out of paths scanning
579 Instead of running N pair-wise diff-trees when inspecting a
580 N-parent merge, find the set of paths that were touched by walking
581 N+1 trees in parallel. These set of paths can then be turned into
582 N pair-wise diff-tree results to be processed through rename
583 detections and such. And N=2 case nicely degenerates to the usual
584 2-way diff-tree, which is very nice.
586 Will keep in 'next' for the remainder of the cycle.
589 * cc/interpret-trailers (2014-04-07) 12 commits
590 - trailer: add blank line before the trailers if needed
591 - Documentation: add documentation for 'git interpret-trailers'
592 - trailer: add tests for commands in config file
593 - trailer: execute command from 'trailer.<name>.command'
594 - trailer: add tests for "git interpret-trailers"
595 - trailer: add interpret-trailers command
596 - trailer: put all the processing together and print
597 - trailer: parse trailers from stdin
598 - trailer: process command line trailer arguments
599 - trailer: read and process config information
600 - trailer: process trailers from stdin and arguments
601 - trailer: add data structures and basic functions
603 A new filter to programatically edit the tail end of the commit log
606 I was planning to merge it to 'next' and keep it there for the
607 remainder of the cycle, but it appears that there still will be
608 another round of reroll, at least for the documentation?