501045565d · matheusbsbs/matheus-bagliano

@@ -1,10 +1,10 @@

1 To: git@vger.kernel.org

2 Bcc: lwn@lwn.net

3 -Subject: What's cooking in git.git (Apr 2016, #08; Fri, 29)

4 -X-master-at: bbc61680168542cf6fd3ae637bde395c73b76f0f

5 -X-next-at: 5ace31314f460db9aef2f1e2e1bd58016b1541f1

3 +Subject: What's cooking in git.git (May 2016, #01; Tue, 3)

4 +X-master-at: 4bb51aed1e858bfac174255ae4d8994af3a8e07a

5 +X-next-at: d223320ae286a25b183b0959b1a0160871c6684a

6

7 -What's cooking in git.git (Apr 2016, #08; Fri, 29)

7 +What's cooking in git.git (May 2016, #01; Tue, 3)

8 --------------------------------------------------

9

10 Here are the topics that have been cooking. Commits prefixed with

@@ -12,11 +12,11 @@ Here are the topics that have been cooking. Commits prefixed with

12 '+' are in 'next'. The ones marked with '.' do not appear in any of

13 the integration branches, but I am still holding onto them.

14

15 -The 'master' branch now has the eighths batch of topics of this

16 -cycle. On the 'maint' front, 2.8.2 is out.

15 +The 'master' branch now has the ninth batch of topics of this cycle.

16 +On the 'maint' front, 2.8.2 is out and fixes that have been in

17 +'master' accumulates on it for 2.8.3.

18

18 -There are a handful of topics that are stuck; they are marked as

19 -"Needs review", "Needs an Ack", etc. in the following list.

19 +Ones with questionable status has a '?' character in their comments.

20

21 You can find the changes described here in the integration branches

22 of the repositories listed at

@@ -26,275 +26,161 @@ of the repositories listed at

26 --------------------------------------------------

27 [Graduated to "master"]

28

29 -* da/user-useconfigonly (2016-04-01) 2 commits

30 - (merged to 'next' on 2016-04-22 at 26845a5)

31 - + ident: give "please tell me" message upon useConfigOnly error

32 - + ident: check for useConfigOnly before auto-detection of name/email

33 -

34 - The "user.useConfigOnly" configuration variable makes it an error

35 - if users do not explicitly set user.name and user.email. However,

36 - its check was not done early enough and allowed another error to

37 - trigger, reporting that the default value we guessed from the

38 - system setting was unusable. This was a suboptimal end-user

39 - experience as we want the users to set user.name/user.email without

40 - relying on the auto-detection at all.

41 -

42 -

43 -* jc/merge-refuse-new-root (2016-04-21) 2 commits

44 - (merged to 'next' on 2016-04-22 at 74eb957)

45 - + pull: pass --allow-unrelated-histories to "git merge"

46 - + t3033: avoid 'ambiguous refs' warning

47 -

48 - "git pull" has been taught to pass --allow-unrelated-histories

49 - option to underlying "git merge".

50 -

51 -

52 -* jk/push-client-deadlock-fix (2016-04-20) 5 commits

53 - (merged to 'next' on 2016-04-22 at d59a2af)

54 - + t5504: drop sigpipe=ok from push tests

55 - + fetch-pack: isolate sigpipe in demuxer thread

56 - + send-pack: isolate sigpipe in demuxer thread

57 - + run-command: teach async threads to ignore SIGPIPE

58 - + send-pack: close demux pipe before finishing async process

59 -

60 - "git push" from a corrupt repository that attempts to push a large

61 - number of refs deadlocked; the thread to relay rejection notices

62 - for these ref updates blocked on writing them to the main thread,

63 - after the main thread at the receiving end notices that the push

64 - failed and decides not to read these notices and return a failure.

65 -

66 -

67 -* js/replace-edit-use-editor-configuration (2016-04-20) 1 commit

68 - (merged to 'next' on 2016-04-22 at 8df6d30)

69 - + replace --edit: respect core.editor

70 -

71 - "git replace -e" did not honour "core.editor" configuration.

72 -

73 -

74 -* js/win32-mmap (2016-04-22) 3 commits

75 - (merged to 'next' on 2016-04-22 at cd39c60)

76 - + mmap(win32): avoid expensive fstat() call

77 - + mmap(win32): avoid copy-on-write when it is unnecessary

78 - + win32mmap: set errno appropriately

79 -

80 - mmap emulation on Windows has been optimized and work better without

81 - consuming paging store when not needed.

82 -

29 +* en/merge-fixes (2016-04-12) 2 commits

30 + (merged to 'next' on 2016-04-27 at 14d7d10)

31 + + merge-recursive: do not check working copy when creating a virtual merge base

32 + + merge-recursive: remove duplicate code

33

84 -* nd/test-helpers (2016-04-15) 2 commits

85 - (merged to 'next' on 2016-04-22 at 55ea5cd)

86 - + test helpers: move test-* to t/helper/ subdirectory

87 - + Makefile: clean *.o files we create

34 + "merge-recursive" strategy incorrectly checked if a path that is

35 + involved in its internal merge exists in the working tree.

36

89 - Sources to many test helper binaries (and the generated helpers)

90 - have been moved to t/helper/ subdirectory to reduce clutter at the

91 - top level of the tree.

37

38 +* jd/p4-jobs-in-commit (2016-04-19) 2 commits

39 + (merged to 'next' on 2016-04-27 at 654d946)

40 + + git-p4: add P4 jobs to git commit message

41 + + git-p4: clean-up code style in tests

42

94 -* sb/mv-submodule-fix (2016-04-19) 1 commit

95 - (merged to 'next' on 2016-04-22 at 089e788)

96 - + mv: allow moving nested submodules

43 + "git p4" learned to record P4 jobs in Git commit that imports from

44 + the history in Perforce.

45

98 - "git mv old new" did not adjust the path for a submodule that lives

99 - as a subdirectory inside old/ directory correctly.

46

47 +* jd/send-email-to-whom (2016-04-25) 1 commit

48 + (merged to 'next' on 2016-04-27 at 47ae363)

49 + + send-email: fix grammo in the prompt that asks e-mail recipients

50

102 -* st/verify-tag (2016-04-22) 6 commits

103 - (merged to 'next' on 2016-04-22 at 98ba239)

104 - + tag -v: verify directly rather than exec-ing verify-tag

105 - + verify-tag: move tag verification code to tag.c

106 - + verify-tag: prepare verify_tag for libification

107 - + verify-tag: update variable name and type

108 - + t7030: test verifying multiple tags

109 - + builtin/verify-tag.c: ignore SIGPIPE in gpg-interface

51 + A question by "git send-email" to ask the identity of the sender

52 + has been updated.

53

111 - Unify internal logic between "git tag -v" and "git verify-tag"

112 - commands by making one directly call into the other.

54

114 ---------------------------------------------------

115 -[New Topics]

55 +* jk/fix-attribute-macro-in-2.5 (2016-04-25) 1 commit

56 + (merged to 'next' on 2016-04-27 at 2e42613)

57 + + remote.c: spell __attribute__ correctly

58

117 -* ab/hooks (2016-04-26) 4 commits

118 - - hooks: allow customizing where the hook directory is

119 - - githooks.txt: minor improvements to the grammar & phrasing

120 - - githooks.txt: amend dangerous advice about 'update' hook ACL

121 - - githooks.txt: improve the intro section

59 + Code fixup.

60

123 - A new configuration variable core.hooksPath allows customizing

124 - where the hook directory is.

61

126 - Almost there.

127 - ($gmane/292635)

62 +* js/name-rev-use-oldest-ref (2016-04-22) 1 commit

63 + (merged to 'next' on 2016-04-27 at 8fdc0ac)

64 + + name-rev: include taggerdate in considering the best name

65

66 + "git describe --contains" often made a hard-to-justify choice of

67 + tag to give name to a given commit, because it tried to come up

68 + with a name with smallest number of hops from a tag, causing an old

69 + commit whose close descendant that is recently tagged were not

70 + described with respect to an old tag but with a newer tag. It did

71 + not help that its computation of "hop" count was further tweaked to

72 + penalize being on a side branch of a merge. The logic has been

73 + updated to favor using the tag with the oldest tagger date, which

74 + is a lot easier to explain to the end users: "We describe a commit

75 + in terms of the (chronologically) oldest tag that contains the

76 + commit."

77

130 -* jc/merge-impossible-no-commit (2016-04-26) 2 commits

131 - - merge: warn --no-commit merge when no new commit is created

132 - - merge: do not contaminate option_commit with --squash

78

134 - "git merge --no-commit" silently succeeded when there is no need to

135 - create any commit, either when you are more recent than the commit

136 - you tried to merge, or you can fast-forward to the commit you tried

137 - to merge. The command gives a warning message in such cases.

79 +* nd/remove-unused (2016-04-22) 2 commits

80 + (merged to 'next' on 2016-04-27 at 7917efa)

81 + + wrapper.c: delete dead function git_mkstemps()

82 + + dir.c: remove dead function fnmatch_icase()

83

139 - Just tying loose ends in a discussion. Unless somebody else

140 - champions this topic, I'll drop it.

84 + Code cleanup.

85

86

143 -* jk/fix-attribute-macro-in-2.5 (2016-04-25) 1 commit

144 - (merged to 'next' on 2016-04-27 at 2e42613)

145 - + remote.c: spell __attribute__ correctly

87 +* nf/mergetool-prompt (2016-04-25) 1 commit

88 + (merged to 'next' on 2016-04-27 at 1c992df)

89 + + difftool/mergetool: make the form of yes/no questions consistent

90

147 - Code fixup.

91 + UI consistency improvements.

92

149 - Will merge to 'master'.

93

94 +* rt/string-list-lookup-cleanup (2016-04-25) 1 commit

95 + (merged to 'next' on 2016-04-27 at 53514e1)

96 + + string_list: use string-list API in unsorted_string_list_lookup()

97

152 -* js/http-custom-headers (2016-04-27) 1 commit

153 - (merged to 'next' on 2016-04-27 at 0c97a50)

154 - + http: support sending custom HTTP headers

98 + Code cleanup.

99

156 - HTTP transport clients learned to throw extra HTTP headers at the

157 - server, specified via http.extraHeader configuration variable.

100

159 - Will merge to 'master'.

101 +* sg/test-lib-simplify-expr-away (2016-04-22) 1 commit

102 + (merged to 'next' on 2016-04-27 at 8f40952)

103 + + test-lib: simplify '--option=value' parsing

104

105 + Code cleanup.

106

162 -* sb/bisect (2016-04-15) 22 commits

163 - - SQUASH???

164 - - bisect: get back halfway shortcut

165 - - bisect: compute best bisection in compute_relevant_weights()

166 - - bisect: use a bottom-up traversal to find relevant weights

167 - - bisect: prepare for different algorithms based on find_all

168 - - bisect: rename count_distance() to compute_weight()

169 - - bisect: make total number of commits global

170 - - bisect: introduce distance_direction()

171 - - bisect: extract get_distance() function from code duplication

172 - - bisect: use commit instead of commit list as arguments when appropriate

173 - - bisect: replace clear_distance() by unique markers

174 - - bisect: use struct node_data array instead of int array

175 - - bisect: get rid of recursion in count_distance()

176 - - bisect: make algorithm behavior independent of DEBUG_BISECT

177 - - bisect: make bisect compile if DEBUG_BISECT is set

178 - - bisect: plug the biggest memory leak

179 - - bisect: add test for the bisect algorithm

180 - - t6030: generalize test to not rely on current implementation

181 - - t: use test_cmp_rev() where appropriate

182 - - t/test-lib-functions.sh: generalize test_cmp_rev

183 - - bisect: allow 'bisect run' if no good commit is known

184 - - bisect: write about `bisect next` in documentation

107 +--------------------------------------------------

108 +[New Topics]

109

186 - The internal algorithm used in "git bisect" to find the next commit

187 - to check has been optimized greatly.

110 +* ew/normal-to-e (2016-05-02) 1 commit

111 + - .mailmap: update to my shorter email address

112

189 - Expecting a reroll.

190 - ($gmane/291163)

113 + Will merge to 'next'.

114

115

193 -* sb/config-exit-status-list (2016-04-26) 1 commit

194 - (merged to 'next' on 2016-04-27 at 44fe343)

195 - + config doc: improve exit code listing

116 +* jc/diff-compact-always-use-blank-heuristics (2016-04-29) 3 commits

117 + - diff: enable "compaction heuristics" and lose experimentation knob

118 + (merged to 'next' on 2016-04-22 at 0c117ea)

119 + + xdiff: implement empty line chunk heuristic

120 + + xdiff: add recs_match helper function

121 + (this branch is tangled with jk/diff-compact-heuristic.)

122

197 - Doc update.

123 + This should be discarded.

124

199 - Will merge to 'master'.

125

126 +* js/close-packs-before-gc (2016-05-02) 1 commit

127 + - t5510: run auto-gc in the foreground

128

202 -* mh/split-under-lock (2016-04-27) 29 commits

203 - - lock_ref_sha1_basic(): only handle REF_NODEREF mode

204 - - commit_ref_update(): remove the flags parameter

205 - - lock_ref_for_update(): don't resolve symrefs

206 - - lock_ref_for_update(): don't re-read non-symbolic references

207 - - refs: resolve symbolic refs first

208 - - ref_transaction_update(): check refname_is_safe() at a minimum

209 - - unlock_ref(): move definition higher in the file

210 - - lock_ref_for_update(): new function

211 - - add_update(): initialize the whole ref_update

212 - - verify_refname_available(): adjust constness in declaration

213 - - refs: don't dereference on rename

214 - - refs: allow log-only updates

215 - - delete_branches(): use resolve_refdup()

216 - - ref_transaction_commit(): correctly report close_ref() failure

217 - - ref_transaction_create(): disallow recursive pruning

218 - - refs: make error messages more consistent

219 - - lock_ref_sha1_basic(): remove unneeded local variable

220 - - read_raw_ref(): improve docstring

221 - - read_raw_ref(): rename symref argument to referent

222 - - read_raw_ref(): clear *type at start of function

223 - - read_raw_ref(): rename flags argument to type

224 - - ref_transaction_commit(): remove local variable n

225 - - rename_ref(): remove unneeded local variable

226 - - commit_ref_update(): write error message to *err, not stderr

227 - - refname_is_safe(): insist that the refname already be normalized

228 - - refname_is_safe(): don't allow the empty string

229 - - refname_is_safe(): use skip_prefix()

230 - - remove_dir_recursively(): add docstring

231 - - safe_create_leading_directories(): improve docstring

129 + Will merge to 'next'.

130

233 - Further preparatory work on the refs API before the pluggable

234 - backend series can land.

131

236 - Almost there.

237 - ($gmane/292772)

132 +* ls/travis-submitting-patches (2016-05-02) 1 commit

133 + - Documentation: add setup instructions for Travis CI

134

135 + Will merge to 'next'.

136

240 -* bn/http-cookiefile-config (2016-04-29) 2 commits

241 - - http: expand http.cookieFile as a path

242 - - Documentation: config: improve word ordering for http.cookieFile

243 - (this branch uses jc/config-pathname-type.)

137

245 - "http.cookieFile" configuration variable clearly wants a pathname,

246 - but we forgot to treat it as such by e.g. applying tilde expansion.

138 +* rn/glossary-typofix (2016-05-02) 1 commit

139 + - Documentation: fix typo 'In such these cases'

140

248 - Waiting for an Ack to what's queued with tweaks, or a reroll.

249 - ($gmane/292969)

141 + Will merge to 'next'.

142

143

252 -* ew/doc-split-pack-disables-bitmap (2016-04-28) 1 commit

253 - - pack-objects: warn on split packs disabling bitmaps

144 +* jc/commit-tree-ignore-commit-gpgsign (2016-05-03) 1 commit

145 + - commit-tree: do not pay attention to commit.gpgsign

146

255 - Doc update.

147 + "git commit-tree" plumbing command required the user to always sign

148 + its result when the user sets the commit.gpgsign configuration

149 + variable, which was an ancient mistake. Rework "git rebase" that

150 + relied on this mistake so that it reads commit.gpgsign and pass (or

151 + not pass) the -S option to "git commit-tree" to keep the end-user

152 + expectation the same, while teaching "git commit-tree" to ignore

153 + the configuration variable. This will stop requiring the users to

154 + sign commit objects used internally as an implementation detail of

155 + "git stash".

156

157 Will merge to 'next'.

158

159

260 -* jc/config-pathname-type (2016-04-29) 1 commit

261 - - config: describe 'pathname' value type

262 - (this branch is used by bn/http-cookiefile-config.)

263 -

264 - Consolidate description of tilde-expansion that is done to

265 - configuration variables that take pathname to a single place.

160 +* jk/push-client-deadlock-fix (2016-05-02) 1 commit

161 + - Windows: add pthread_sigmask() that does nothing

162

163 Will merge to 'next'.

164

165

270 -* jk/submodule-config-sanitize-fix (2016-04-28) 5 commits

271 - - submodule: use prepare_submodule_repo_env consistently

272 - - submodule--helper: move config-sanitizing to submodule.c

273 - - submodule: export sanitized GIT_CONFIG_PARAMETERS

274 - - t5550: break submodule config test into multiple sub-tests

275 - - t5550: fix typo in $HTTPD_URL

276 -

277 - An earlier addition of "sanitize_submodule_env" with 14111fc4 (git:

278 - submodule honor -c credential.* from command line, 2016-02-29)

279 - turned out to be a convoluted no-op; implement what it wanted to do

280 - correctly.

166 +* sb/clean-test-fix (2016-05-03) 1 commit

167 + - t7300: mark test with SANITY

168

282 - With a rethink of the merit of "sanitization" going on, we may end

283 - up doing the configuration propagation very differently, though.

169 + Will merge to 'next'.

170

285 - Will hold.

171

172 +* sb/submodule-module-list-pathspec-fix (2016-05-03) 1 commit

173 + - submodule deinit test: fix broken && chain in subshell

174

288 -* mh/connect-leak (2016-04-28) 1 commit

289 - - git_connect(): fix memory leak with CONNECT_DIAG_URL

175 + Will merge to 'next'.

176

291 - Is already made obsolete with a patch in flight under discussion.

292 - ($gmane/292962)

177

178 +* sk/gitweb-highlight-encoding (2016-05-03) 1 commit

179 + - gitweb: apply fallback encoding before highlight

180

295 -* sb/misc-cleanups (2016-04-28) 2 commits

296 - - submodule-config: don't shadow `cache`

297 - - config.c: drop local variable

181 + Some multi-byte encoding can have a backslash byte as a later part

182 + of one letter, which would confuse "highlight" filter used in

183 + gitweb.

184

185 Will merge to 'next'.

186

@@ -309,7 +195,7 @@ of the repositories listed at

195 not, causing "git diff $path" to complain that no such path is

196 known to Git. They have been taught to do the normalization.

197

312 - Will be rerolled?

198 + Will this be rerolled?

199 ($gmane/290724)

200

201

@@ -491,53 +377,184 @@ of the repositories listed at

377 --------------------------------------------------

378 [Cooking]

379

494 -* js/name-rev-use-oldest-ref (2016-04-22) 1 commit

495 - (merged to 'next' on 2016-04-27 at 8fdc0ac)

496 - + name-rev: include taggerdate in considering the best name

380 +* ab/hooks (2016-04-26) 4 commits

381 + - hooks: allow customizing where the hook directory is

382 + - githooks.txt: minor improvements to the grammar & phrasing

383 + - githooks.txt: amend dangerous advice about 'update' hook ACL

384 + - githooks.txt: improve the intro section

385

498 - "git describe --contains" often made a hard-to-justify choice of

499 - tag to give name to a given commit, because it tried to come up

500 - with a name with smallest number of hops from a tag, causing an old

501 - commit whose close descendant that is recently tagged were not

502 - described with respect to an old tag but with a newer tag. It did

503 - not help that its computation of "hop" count was further tweaked to

504 - penalize being on a side branch of a merge. The logic has been

505 - updated to favor using the tag with the oldest tagger date, which

506 - is a lot easier to explain to the end users: "We describe a commit

507 - in terms of the (chronologically) oldest tag that contains the

508 - commit."

386 + A new configuration variable core.hooksPath allows customizing

387 + where the hook directory is.

388

510 - Will merge to 'master'.

389 + Almost there.

390 + ($gmane/292635)

391

392

513 -* nd/remove-unused (2016-04-22) 2 commits

514 - (merged to 'next' on 2016-04-27 at 7917efa)

515 - + wrapper.c: delete dead function git_mkstemps()

516 - + dir.c: remove dead function fnmatch_icase()

393 +* jc/merge-impossible-no-commit (2016-04-26) 2 commits

394 + - merge: warn --no-commit merge when no new commit is created

395 + - merge: do not contaminate option_commit with --squash

396

518 - Code cleanup.

397 + "git merge --no-commit" silently succeeded when there is no need to

398 + create any commit, either when you are more recent than the commit

399 + you tried to merge, or you can fast-forward to the commit you tried

400 + to merge. The command gives a warning message in such cases.

401

520 - Will merge to 'master'.

402 + Just tying loose ends in a discussion. Unless somebody else

403 + champions this topic, I'll drop it.

404

405

523 -* rt/string-list-lookup-cleanup (2016-04-25) 1 commit

524 - (merged to 'next' on 2016-04-27 at 53514e1)

525 - + string_list: use string-list API in unsorted_string_list_lookup()

406 +* js/http-custom-headers (2016-04-27) 1 commit

407 + (merged to 'next' on 2016-04-27 at 0c97a50)

408 + + http: support sending custom HTTP headers

409

527 - Code cleanup.

410 + HTTP transport clients learned to throw extra HTTP headers at the

411 + server, specified via http.extraHeader configuration variable.

412

413 Will merge to 'master'.

414

415

532 -* sg/test-lib-simplify-expr-away (2016-04-22) 1 commit

533 - (merged to 'next' on 2016-04-27 at 8f40952)

534 - + test-lib: simplify '--option=value' parsing

416 +* sb/bisect (2016-04-15) 22 commits

417 + - SQUASH???

418 + - bisect: get back halfway shortcut

419 + - bisect: compute best bisection in compute_relevant_weights()

420 + - bisect: use a bottom-up traversal to find relevant weights

421 + - bisect: prepare for different algorithms based on find_all

422 + - bisect: rename count_distance() to compute_weight()

423 + - bisect: make total number of commits global

424 + - bisect: introduce distance_direction()

425 + - bisect: extract get_distance() function from code duplication

426 + - bisect: use commit instead of commit list as arguments when appropriate

427 + - bisect: replace clear_distance() by unique markers

428 + - bisect: use struct node_data array instead of int array

429 + - bisect: get rid of recursion in count_distance()

430 + - bisect: make algorithm behavior independent of DEBUG_BISECT

431 + - bisect: make bisect compile if DEBUG_BISECT is set

432 + - bisect: plug the biggest memory leak

433 + - bisect: add test for the bisect algorithm

434 + - t6030: generalize test to not rely on current implementation

435 + - t: use test_cmp_rev() where appropriate

436 + - t/test-lib-functions.sh: generalize test_cmp_rev

437 + - bisect: allow 'bisect run' if no good commit is known

438 + - bisect: write about `bisect next` in documentation

439

536 - Code cleanup.

440 + The internal algorithm used in "git bisect" to find the next commit

441 + to check has been optimized greatly.

442 +

443 + Expecting a reroll.

444 + ($gmane/291163)

445 +

446 +

447 +* sb/config-exit-status-list (2016-04-26) 1 commit

448 + (merged to 'next' on 2016-04-27 at 44fe343)

449 + + config doc: improve exit code listing

450 +

451 + Doc update.

452

453 Will merge to 'master'.

454

455

456 +* mh/split-under-lock (2016-04-27) 29 commits

457 + - lock_ref_sha1_basic(): only handle REF_NODEREF mode

458 + - commit_ref_update(): remove the flags parameter

459 + - lock_ref_for_update(): don't resolve symrefs

460 + - lock_ref_for_update(): don't re-read non-symbolic references

461 + - refs: resolve symbolic refs first

462 + - ref_transaction_update(): check refname_is_safe() at a minimum

463 + - unlock_ref(): move definition higher in the file

464 + - lock_ref_for_update(): new function

465 + - add_update(): initialize the whole ref_update

466 + - verify_refname_available(): adjust constness in declaration

467 + - refs: don't dereference on rename

468 + - refs: allow log-only updates

469 + - delete_branches(): use resolve_refdup()

470 + - ref_transaction_commit(): correctly report close_ref() failure

471 + - ref_transaction_create(): disallow recursive pruning

472 + - refs: make error messages more consistent

473 + - lock_ref_sha1_basic(): remove unneeded local variable

474 + - read_raw_ref(): improve docstring

475 + - read_raw_ref(): rename symref argument to referent

476 + - read_raw_ref(): clear *type at start of function

477 + - read_raw_ref(): rename flags argument to type

478 + - ref_transaction_commit(): remove local variable n

479 + - rename_ref(): remove unneeded local variable

480 + - commit_ref_update(): write error message to *err, not stderr

481 + - refname_is_safe(): insist that the refname already be normalized

482 + - refname_is_safe(): don't allow the empty string

483 + - refname_is_safe(): use skip_prefix()

484 + - remove_dir_recursively(): add docstring

485 + - safe_create_leading_directories(): improve docstring

486 +

487 + Further preparatory work on the refs API before the pluggable

488 + backend series can land.

489 +

490 + Almost there.

491 + ($gmane/292772)

492 +

493 +

494 +* bn/http-cookiefile-config (2016-04-29) 2 commits

495 + - http: expand http.cookieFile as a path

496 + - Documentation: config: improve word ordering for http.cookieFile

497 + (this branch uses jc/config-pathname-type.)

498 +

499 + "http.cookieFile" configuration variable clearly wants a pathname,

500 + but we forgot to treat it as such by e.g. applying tilde expansion.

501 +

502 + Waiting for an Ack to what's queued with tweaks, or a reroll.

503 + ($gmane/292969)

504 +

505 +

506 +* ew/doc-split-pack-disables-bitmap (2016-04-28) 1 commit

507 + - pack-objects: warn on split packs disabling bitmaps

508 +

509 + Doc update.

510 +

511 + Will merge to 'next'.

512 +

513 +

514 +* jc/config-pathname-type (2016-04-29) 1 commit

515 + - config: describe 'pathname' value type

516 + (this branch is used by bn/http-cookiefile-config.)

517 +

518 + Consolidate description of tilde-expansion that is done to

519 + configuration variables that take pathname to a single place.

520 +

521 + Will merge to 'next'.

522 +

523 +

524 +* jk/submodule-config-sanitize-fix (2016-04-28) 5 commits

525 + - submodule: use prepare_submodule_repo_env consistently

526 + - submodule--helper: move config-sanitizing to submodule.c

527 + - submodule: export sanitized GIT_CONFIG_PARAMETERS

528 + - t5550: break submodule config test into multiple sub-tests

529 + - t5550: fix typo in $HTTPD_URL

530 +

531 + An earlier addition of "sanitize_submodule_env" with 14111fc4 (git:

532 + submodule honor -c credential.* from command line, 2016-02-29)

533 + turned out to be a convoluted no-op; implement what it wanted to do

534 + correctly.

535 +

536 + With a rethink of the merit of "sanitization" going on, we may end

537 + up doing the configuration propagation very differently, though.

538 +

539 + Will hold.

540 +

541 +

542 +* mh/connect-leak (2016-04-28) 1 commit

543 + - git_connect(): fix memory leak with CONNECT_DIAG_URL

544 +

545 + Is already made obsolete with a patch in flight under discussion.

546 + ($gmane/292962)

547 +

548 + Will discard.

549 +

550 +

551 +* sb/misc-cleanups (2016-04-28) 2 commits

552 + - submodule-config: don't shadow `cache`

553 + - config.c: drop local variable

554 +

555 + Will merge to 'next'.

556 +

557 +

558 * ew/fast-import-unpack-limit (2016-04-24) 1 commit

559 - fast-import: implement unpack limit

560

@@ -549,16 +566,6 @@ of the repositories listed at

566 ($gmane/292562)

567

568

552 -* jd/send-email-to-whom (2016-04-25) 1 commit

553 - (merged to 'next' on 2016-04-27 at 47ae363)

554 - + send-email: fix grammo in the prompt that asks e-mail recipients

555 -

556 - A question by "git send-email" to ask the identity of the sender

557 - has been updated.

558 -

559 - Will merge to 'master'.

560 -

561 -

569 * ld/p4-test-py3 (2016-04-26) 3 commits

570 (merged to 'next' on 2016-04-27 at d5d5fca)

571 + git-p4 tests: time_in_seconds should use $PYTHON_PATH

@@ -592,17 +599,6 @@ of the repositories listed at

599 Will merge to 'master'.

600

601

595 -* jd/p4-jobs-in-commit (2016-04-19) 2 commits

596 - (merged to 'next' on 2016-04-27 at 654d946)

597 - + git-p4: add P4 jobs to git commit message

598 - + git-p4: clean-up code style in tests

599 -

600 - "git p4" learned to record P4 jobs in Git commit that imports from

601 - the history in Perforce.

602 -

603 - Will merge to 'master'.

604 -

605 -

602 * ls/p4-lfs (2016-04-28) 3 commits

603 - git-p4: fix Git LFS pointer parsing

604 - travis-ci: express Linux/OS X dependency versions more clearly

@@ -636,9 +632,8 @@ of the repositories listed at

632 core.autocrlf, there is no need to, hence we no longer reject the

633 combination.

634

639 - The last step seems to be identical to what I earlier did and

640 - discarded because the approach is fundamentally wrong, but I may be

641 - misreading the patch.

635 + Earlier steps looked alright, but it veers into a wrong direction

636 + in the middle.

637

638

639 * bc/object-id (2016-04-25) 6 commits

@@ -655,16 +650,15 @@ of the repositories listed at

650 Will merge to 'master'.

651

652

658 -* ep/http-curl-trace (2016-04-20) 3 commits

659 - - git.txt: document the new GIT_TRACE_CURL environment variable

660 - - imap-send.c: introduce the GIT_TRACE_CURL enviroment variable

661 - - http.c: implement the GIT_TRACE_CURL environment variable

653 +* ep/http-curl-trace (2016-05-02) 2 commits

654 + . imap-send.c: introduce the GIT_TRACE_CURL environment variable

655 + . http.c: implement the GIT_TRACE_CURL environment variable

656

657 HTTP transport gained an option to produce more detailed debugging

658 trace.

659

660 Still under discussion.

667 - ($gmane/292074)

661 + ($gmane/292074, 293236)

662

663

664 * nd/worktree-various-heads (2016-04-22) 13 commits

@@ -701,7 +695,8 @@ of the repositories listed at

695 Will merge to 'master'.

696

697

704 -* pb/commit-verbose-config (2016-04-19) 6 commits

698 +* pb/commit-verbose-config (2016-05-02) 7 commits

699 + - t/t7507: tests for broken behavior of status

700 - commit: add a commit.verbose config variable

701 - t7507-commit-verbose: improve test coverage by testing number of diffs

702 - parse-options.c: make OPTION_COUNTUP respect "unspecified" values

@@ -713,19 +708,8 @@ of the repositories listed at

708 configuration variable and act as if "--verbose" option was

709 given from the command line.

710

716 - Need a reroll but it will be some time before that happens.

717 - ($gmane/292160).

718 -

719 -

720 -* en/merge-fixes (2016-04-12) 2 commits

721 - (merged to 'next' on 2016-04-27 at 14d7d10)

722 - + merge-recursive: do not check working copy when creating a virtual merge base

723 - + merge-recursive: remove duplicate code

724 -

725 - "merge-recursive" strategy incorrectly checked if a path that is

726 - involved in its internal merge exists in the working tree.

727 -

728 - Will merge to 'master'.

711 + Almost there.

712 + ($gmane/293410).

713

714

715 * jc/fsck-nul-in-commit (2016-04-14) 2 commits

@@ -735,28 +719,41 @@ of the repositories listed at

719 "git fsck" learned to catch NUL byte in a commit object as

720 potential error and warn.

721

722 + What was the status of this one? Ready to proceed?

723 +

724

725 * jc/ll-merge-internal (2016-04-27) 3 commits

726 - t6036: remove pointless test that expects failure

727 - ll-merge: use a longer conflict marker for internal merge

728 - ll-merge: fix typo in comment

729

744 - RFC.

730 + "git rerere" can get confused by conflict markers deliberately left

731 + by the inner merge step, because they are indistinguishable from

732 + the real conflict markers left by the outermost merge which are

733 + what the end user and "rerere" need to look at. This was fixed by

734 + making the conflict markers left by the inner merges a bit longer.

735 +

736 + Will rebase to remove the comment after three-dash line and then merge.

737

738

747 -* jk/diff-compact-heuristic (2016-04-19) 2 commits

739 +* jk/diff-compact-heuristic (2016-05-02) 3 commits

740 + (merged to 'next' on 2016-05-02 at 2a74763)

741 + + diff: undocument the compaction heuristic knobs for experimentation

742 (merged to 'next' on 2016-04-22 at 0c117ea)

743 + xdiff: implement empty line chunk heuristic

744 + xdiff: add recs_match helper function

745 + (this branch is tangled with jc/diff-compact-always-use-blank-heuristics.)

746

747 Patch output from "git diff" and friends has been tweaked to be

748 more readable by using a blank line as a strong hint that the

749 contents before and after it belong to a logically separate unit.

750

756 - Will merge to 'master' after removing the experimentation knob.

751 + Will merge to 'master'.

752

753

759 -* sb/submodule-init (2016-04-29) 6 commits

754 +* sb/submodule-init (2016-05-03) 7 commits

755 + (merged to 'next' on 2016-05-03 at 8a5fce4)

756 + + submodule init: redirect stdout to stderr

757 (merged to 'next' on 2016-04-29 at 3e81ee88)

758 + submodule--helper update-clone: abort gracefully on missing .gitmodules

759 + submodule init: fail gracefully with a missing .gitmodules file

@@ -783,15 +780,6 @@ of the repositories listed at

780 Needs review.

781

782

786 -* nf/mergetool-prompt (2016-04-25) 1 commit

787 - (merged to 'next' on 2016-04-27 at 1c992df)

788 - + difftool/mergetool: make the form of yes/no questions consistent

789 -

790 - UI consistency improvements.

791 -

792 - Will merge to 'master'.

793 -

794 -

783 * va/i18n-misc-updates (2016-04-19) 9 commits

784 - i18n: builtin/pull.c: split strings marked for translation

785 - i18n: builtin/pull.c: mark placeholders for translation

@@ -806,6 +794,7 @@ of the repositories listed at

794 Mark several messages for translation.

795

796 Comments? They looked all sensible to me.

797 + Does the lack of response mean lack of interest and support?

798

799

800 * jc/drop-git-spec-in (2016-04-27) 2 commits

@@ -859,7 +848,7 @@ of the repositories listed at

848 ($gmane/292622).

849

850

862 -* dt/index-helper (2016-04-28) 19 commits

851 +* dt/index-helper (2016-05-02) 19 commits

852 . untracked-cache: config option

853 . Add tracing to measure where most of the time is spent

854 . index-helper: optionally automatically run

@@ -870,7 +859,7 @@ of the repositories listed at

859 . unpack-trees: preserve index extensions

860 . update-index: enable/disable watchman support

861 . index-helper: use watchman to avoid refreshing index with lstat()

873 - . Add watchman support to reduce index refresh cost

862 + . watchman: add support to watchman to reduce refresh cost

863 . read-cache: add watchman 'WAMA' extension

864 . index-helper: add --detach

865 . daemonize(): set a flag before exiting the main process

@@ -880,8 +869,8 @@ of the repositories listed at

869 . read-cache: allow to keep mmap'd memory after reading

870 . read-cache.c: fix constness of verify_hdr()

871

883 - Needs review.

884 - t7900#5 seems to fail for me.

872 + Needs review. Reported to break its own tests.

873 + ($gmane/293461).

874

875

876 * jc/bundle (2016-03-03) 6 commits

What's cooking (2016/05 #01)