7be9902743 · 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 (Feb 2017, #06; Tue, 21)

3 +Subject: What's cooking in git.git (Feb 2017, #07; Thu, 23)

4 X-master-at: 80ba04ed9b07c34af9cc644f2183b3b80fd81744

5 -X-next-at: 302250072e59acfdf6a5a069d54402b4f1478844

5 +X-next-at: cbe923c8da4a7f6cf4861a317ae620497394d507

6

7 -What's cooking in git.git (Feb 2017, #06; Tue, 21)

7 +What's cooking in git.git (Feb 2017, #07; Thu, 23)

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

9

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

@@ -12,10 +12,6 @@ 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 tip of the 'master' is still at -rc2 but with a git-svn update

16 -from Eric. I'd like to get pull requests for gitk and git-gui

17 -updates soonish, if we are to have one during this cycle.

18 -

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

16 of the repositories listed at

17

@@ -24,127 +20,84 @@ of the repositories listed at

20 --------------------------------------------------

21 [New Topics]

22

27 -* jk/tempfile-ferror-fclose-confusion (2017-02-17) 1 commit

28 - (merged to 'next' on 2017-02-21 at 479ba0131f)

29 - + tempfile: set errno to a known value before calling ferror()

30 -

31 - A caller of tempfile API that uses stdio interface to write to

32 - files may ignore errors while writing, which is detected when

33 - tempfile is closed (with a call to ferror()). By that time, the

34 - original errno that may have told us what went wrong is likely to

35 - be long gone and was overwritten by an irrelevant value.

36 - close_tempfile() now resets errno to EIO to make errno at least

37 - predictable.

38 -

39 - Will cook in 'next'.

40 -

23 +* js/curl-empty-auth-set-by-default (2017-02-22) 1 commit

24 + - http(s): automatically try NTLM authentication first

25

42 -* ah/doc-ls-files-quotepath (2017-02-20) 1 commit

43 - - Documentation: link git-ls-files to core.quotePath variable

26 + Flip "http.emptyAuth" on by default to help OOB experience for

27 + users of HTTP Negotiate authentication scheme.

28

45 - Documentation for "git ls-files" did not refer to core.quotePath

29 + Under discussion.

30 + cf. <20170222233419.q3fxqmrscosumbjm@genre.crustytoothpaste.net>

31

47 - Needs review.

32

49 -

50 -* dr/doc-check-ref-format-normalize (2017-02-21) 1 commit

51 - (merged to 'next' on 2017-02-21 at 5e88b7a93d)

52 - + git-check-ref-format: clarify documentation for --normalize

33 +* bc/blame-doc-fix (2017-02-22) 1 commit

34 + (merged to 'next' on 2017-02-22 at 81c0ff2283)

35 + + Documentation: use brackets for optional arguments

36

37 Doc update.

38

56 - Will cook in 'next'.

39 + Will merge to 'master'.

40

41

59 -* gp/document-dotfiles-in-templates-are-not-copied (2017-02-17) 1 commit

60 - (merged to 'next' on 2017-02-21 at bbfa2bb7d4)

61 - + init: document dotfiles exclusion on template copy

42 +* bc/worktree-doc-fix-detached (2017-02-22) 1 commit

43 + (merged to 'next' on 2017-02-22 at 8257025363)

44 + + Documentation: correctly spell git worktree --detach

45

46 Doc update.

47

65 - Will cook in 'next'.

48 + Will merge to 'master'.

49

50

68 -* jc/config-case-cmdline (2017-02-21) 2 commits

69 - (merged to 'next' on 2017-02-21 at 354f023a3a)

70 - + config: reject invalid VAR in 'git -c VAR=VAL command'

71 - + config: preserve <subsection> case for one-shot config on the command line

72 -

73 - The code to parse "git -c VAR=VAL cmd" and set configuration

74 - variable for the duration of cmd had two small bugs, which have

75 - been fixed.

76 -

77 - Will cook in 'next'.

51 +* rt/align-add-i-help-text (2017-02-22) 1 commit

52 + (merged to 'next' on 2017-02-22 at a8573afb9a)

53 + + git add -i: replace \t with blanks in the help message

54

55 + Doc update.

56

80 -* jh/memihash-opt (2017-02-17) 5 commits

81 - - name-hash: remember previous dir_entry during lazy_init_name_hash

82 - - name-hash: specify initial size for istate.dir_hash table

83 - - name-hash: precompute hash values during preload-index

84 - - hashmap: allow memihash computation to be continued

85 - - name-hash: eliminate duplicate memihash call

57 + Will merge to 'master'.

58

87 - Expecting an update for perf?

59

60 +* jh/send-email-one-cc (2017-02-23) 1 commit

61 + - send-email: only allow one address per body tag

62

90 -* km/delete-ref-reflog-message (2017-02-20) 4 commits

91 - (merged to 'next' on 2017-02-21 at 4ee4ce3f64)

92 - + branch: record creation of renamed branch in HEAD's log

93 - + rename_ref: replace empty message in HEAD's log

94 - + update-ref: pass reflog message to delete_ref()

95 - + delete_ref: accept a reflog message argument

63 + "Cc:" on the trailer part does not have to conform to RFC strictly,

64 + unlike in the e-mail header. "git send-email" has been updated to

65 + ignore anything after '>' when picking addresses, to allow non-address

66 + cruft like " # stable 4.4" after the address.

67

97 - "git update-ref -d" and other operations to delete references did

98 - not leave any entry in HEAD's reflog when the reference being

99 - deleted was the current branch. This is not a problem in practice

100 - because you do not want to delete the branch you are currently on,

101 - but caused renaming of the current branch to something else not to

102 - be logged in a useful way.

68 + Will merge to and then cook in 'next'.

69

104 - Will cook in 'next'.

70

71 +* jk/http-auth (2017-02-23) 1 commit

72 + - http: restrict auth methods to what the server advertises

73

107 -* nd/prune-in-worktree (2017-02-19) 15 commits

108 - - rev-list: expose and document --single-worktree

109 - - revision.c: --reflog add HEAD reflog from all worktrees

110 - - files-backend: make reflog iterator go through per-worktree reflog

111 - - refs: add refs_for_each_reflog[_ent]()

112 - - revision.c: --all adds HEAD from all worktrees

113 - - refs: remove dead for_each_*_submodule()

114 - - revision.c: use refs_for_each*() instead of for_each_*_submodule()

115 - - refs: add a refs_for_each_in() and friends

116 - - refs: add refs_for_each_ref()

117 - - refs: add refs_head_ref()

118 - - refs: add refs_read_ref[_full]()

119 - - refs: move submodule slash stripping code to get_submodule_ref_store

120 - - revision.c: --indexed-objects add objects from all worktrees

121 - - revision.c: refactor add_index_objects_to_pending()

122 - - revision.h: new flag in struct rev_info wrt. worktree-related refs

123 - (this branch uses mh/ref-remove-empty-directory, mh/submodule-hash, nd/files-backend-git-dir and nd/worktree-kill-parse-ref.)

74 + Reduce authentication round-trip over HTTP when the server supports

75 + just a single authentication method.

76

125 - "git gc" and friends when multiple worktrees are used off of a

126 - single repository did not consider the index and per-worktree refs

127 - of other worktrees as the root for reachability traversal, making

128 - objects that are in use only in other worktrees to be subject to

129 - garbage collection.

77 + Will merge to and then cook in 'next'.

78 + There is a follow-up that would supersede js/curl-empty-auth-set-by-default

79 + topic; as it expected to be refined further, it is not queued here yet.

80

81

132 -* mm/fetch-show-error-message-on-unadvertised-object (2017-02-21) 1 commit

133 - - fetch: print an error when declining to request an unadvertised object

82 +* jk/ident-empty (2017-02-23) 4 commits

83 + - ident: do not ignore empty config name/email

84 + - ident: reject all-crud ident name

85 + - ident: handle NULL email when complaining of empty name

86 + - ident: mark error messages for translation

87

135 - "git fetch" that requests a commit by object name, when the other

136 - side does not allow such an request, failed without much

137 - explanation.

88 + user.email that consists of only cruft chars should have

89 + consistently errored out, but didn't.

90

139 - Expecting a split series?

91 + Will merge to and then cook in 'next'.

92

93

142 -* rl/remote-allow-missing-branch-name-merge (2017-02-21) 1 commit

143 - - remote: ignore failure to remove missing branch.<name>.merge

94 +* jt/upload-pack-error-report (2017-02-23) 1 commit

95 + - upload-pack: report "not our ref" to client

96

145 - "git remote rm X", when a branch has remote X configured as the

146 - value of its branch.*.remote, tried to remove branch.*.remote and

147 - branch.*.merge and failed if either is unset.

97 + "git upload-pack", which is a counter-part of "git fetch", did not

98 + report a request for a ref that was not advertised as invalid.

99 + This is generally not a problem (because "git fetch" will stop

100 + before making such a request), but is the right thing to do.

101

102 Will merge to and then cook in 'next'.

103

@@ -283,6 +236,136 @@ of the repositories listed at

236 --------------------------------------------------

237 [Cooking]

238

239 +* jk/tempfile-ferror-fclose-confusion (2017-02-17) 1 commit

240 + (merged to 'next' on 2017-02-21 at 479ba0131f)

241 + + tempfile: set errno to a known value before calling ferror()

242 +

243 + A caller of tempfile API that uses stdio interface to write to

244 + files may ignore errors while writing, which is detected when

245 + tempfile is closed (with a call to ferror()). By that time, the

246 + original errno that may have told us what went wrong is likely to

247 + be long gone and was overwritten by an irrelevant value.

248 + close_tempfile() now resets errno to EIO to make errno at least

249 + predictable.

250 +

251 + Will cook in 'next'.

252 +

253 +

254 +* ah/doc-ls-files-quotepath (2017-02-22) 1 commit

255 + - Documentation: clarify core.quotePath and update git-ls-files doc

256 +

257 + Documentation for "git ls-files" did not refer to core.quotePath

258 +

259 + Looked good, but another reroll is still incoming?

260 + cf. <3c801e54-28c7-52d0-6915-ee7aaf1d89c9@gmail.com>

261 +

262 +

263 +* dr/doc-check-ref-format-normalize (2017-02-21) 1 commit

264 + (merged to 'next' on 2017-02-21 at 5e88b7a93d)

265 + + git-check-ref-format: clarify documentation for --normalize

266 +

267 + Doc update.

268 +

269 + Will merge to 'master'.

270 +

271 +

272 +* gp/document-dotfiles-in-templates-are-not-copied (2017-02-17) 1 commit

273 + (merged to 'next' on 2017-02-21 at bbfa2bb7d4)

274 + + init: document dotfiles exclusion on template copy

275 +

276 + Doc update.

277 +

278 + Will merge to 'master'.

279 +

280 +

281 +* jc/config-case-cmdline (2017-02-21) 3 commits

282 + (merged to 'next' on 2017-02-21 at 63d3652c77)

283 + + config: squelch stupid compiler

284 + (merged to 'next' on 2017-02-21 at 354f023a3a)

285 + + config: reject invalid VAR in 'git -c VAR=VAL command'

286 + + config: preserve <subsection> case for one-shot config on the command line

287 +

288 + The code to parse "git -c VAR=VAL cmd" and set configuration

289 + variable for the duration of cmd had two small bugs, which have

290 + been fixed.

291 +

292 + Will cook in 'next'.

293 +

294 +

295 +* jh/memihash-opt (2017-02-17) 5 commits

296 + - name-hash: remember previous dir_entry during lazy_init_name_hash

297 + - name-hash: specify initial size for istate.dir_hash table

298 + - name-hash: precompute hash values during preload-index

299 + - hashmap: allow memihash computation to be continued

300 + - name-hash: eliminate duplicate memihash call

301 +

302 + Expecting an update for perf?

303 +

304 +

305 +* km/delete-ref-reflog-message (2017-02-20) 4 commits

306 + (merged to 'next' on 2017-02-21 at 4ee4ce3f64)

307 + + branch: record creation of renamed branch in HEAD's log

308 + + rename_ref: replace empty message in HEAD's log

309 + + update-ref: pass reflog message to delete_ref()

310 + + delete_ref: accept a reflog message argument

311 +

312 + "git update-ref -d" and other operations to delete references did

313 + not leave any entry in HEAD's reflog when the reference being

314 + deleted was the current branch. This is not a problem in practice

315 + because you do not want to delete the branch you are currently on,

316 + but caused renaming of the current branch to something else not to

317 + be logged in a useful way.

318 +

319 + Will cook in 'next'.

320 +

321 +

322 +* nd/prune-in-worktree (2017-02-19) 15 commits

323 + . rev-list: expose and document --single-worktree

324 + . revision.c: --reflog add HEAD reflog from all worktrees

325 + . files-backend: make reflog iterator go through per-worktree reflog

326 + . refs: add refs_for_each_reflog[_ent]()

327 + . revision.c: --all adds HEAD from all worktrees

328 + . refs: remove dead for_each_*_submodule()

329 + . revision.c: use refs_for_each*() instead of for_each_*_submodule()

330 + . refs: add a refs_for_each_in() and friends

331 + . refs: add refs_for_each_ref()

332 + . refs: add refs_head_ref()

333 + . refs: add refs_read_ref[_full]()

334 + . refs: move submodule slash stripping code to get_submodule_ref_store

335 + . revision.c: --indexed-objects add objects from all worktrees

336 + . revision.c: refactor add_index_objects_to_pending()

337 + . revision.h: new flag in struct rev_info wrt. worktree-related refs

338 + (this branch uses mh/ref-remove-empty-directory, mh/submodule-hash and nd/worktree-kill-parse-ref; is tangled with nd/files-backend-git-dir.)

339 +

340 + "git gc" and friends when multiple worktrees are used off of a

341 + single repository did not consider the index and per-worktree refs

342 + of other worktrees as the root for reachability traversal, making

343 + objects that are in use only in other worktrees to be subject to

344 + garbage collection.

345 +

346 +

347 +* mm/fetch-show-error-message-on-unadvertised-object (2017-02-22) 4 commits

348 + - fetch-pack: add specific error for fetching an unadvertised object

349 + - fetch_refs_via_pack: call report_unmatched_refs

350 + - squash??? remove unfinished sentence

351 + - fetch-pack: move code to report unmatched refs to a function

352 +

353 + "git fetch" that requests a commit by object name, when the other

354 + side does not allow such an request, failed without much

355 + explanation.

356 +

357 +

358 +* rl/remote-allow-missing-branch-name-merge (2017-02-21) 1 commit

359 + (merged to 'next' on 2017-02-22 at cbe923c8da)

360 + + remote: ignore failure to remove missing branch.<name>.merge

361 +

362 + "git remote rm X", when a branch has remote X configured as the

363 + value of its branch.*.remote, tried to remove branch.*.remote and

364 + branch.*.merge and failed if either is unset.

365 +

366 + Will cook in 'next'.

367 +

368 +

369 * vn/xdiff-func-context (2017-01-15) 1 commit

370 (merged to 'next' on 2017-02-21 at 838eab8d93)

371 + xdiff -W: relax end-of-file function detection

@@ -293,13 +376,30 @@ of the repositories listed at

376 Will cook in 'next'.

377

378

296 -* nd/worktree-kill-parse-ref (2017-02-19) 5 commits

297 - - refs: kill set_worktree_head_symref()

298 - - refs: add refs_create_symref()

299 - - worktree.c: kill parse_ref() in favor of refs_resolve_ref_unsafe()

300 - - refs.c: add refs_resolve_ref_unsafe()

301 - - refs: introduce get_worktree_ref_store()

302 - (this branch is used by nd/prune-in-worktree; uses mh/ref-remove-empty-directory, mh/submodule-hash and nd/files-backend-git-dir.)

379 +* nd/worktree-kill-parse-ref (2017-02-19) 22 commits

380 + . refs: kill set_worktree_head_symref()

381 + . refs: add refs_create_symref()

382 + . worktree.c: kill parse_ref() in favor of refs_resolve_ref_unsafe()

383 + . refs.c: add refs_resolve_ref_unsafe()

384 + . refs: introduce get_worktree_ref_store()

385 + . refs: rename get_ref_store() to get_submodule_ref_store() and make it public

386 + . files-backend: remove submodule_allowed from files_downcast()

387 + . refs: move submodule code out of files-backend.c

388 + . path.c: move some code out of strbuf_git_path_submodule()

389 + . refs.c: make get_main_ref_store() public and use it

390 + . refs.c: kill register_ref_store(), add register_submodule_ref_store()

391 + . refs.c: flatten get_ref_store() a bit

392 + . refs: rename lookup_ref_store() to lookup_submodule_ref_store()

393 + . refs.c: introduce get_main_ref_store()

394 + . files-backend: remove the use of git_path()

395 + . refs.c: share is_per_worktree_ref() to files-backend.c

396 + . files-backend: replace *git_path*() with files_path()

397 + . files-backend: add files_path()

398 + . files-backend: convert git_path() to strbuf_git_path()

399 + . refs-internal.c: make files_log_ref_write() static

400 + . Merge branch 'mh/ref-remove-empty-directory' into nd/files-backend-git-dir

401 + . Merge branch 'mh/submodule-hash' into nd/files-backend-git-dir

402 + (this branch is used by nd/prune-in-worktree; uses mh/ref-remove-empty-directory and mh/submodule-hash; is tangled with nd/files-backend-git-dir.)

403

404 (hopefully) a beginning of safer "git worktree" that is resistant

405 to "gc".

@@ -307,25 +407,34 @@ of the repositories listed at

407 Needs review.

408

409

310 -* nd/files-backend-git-dir (2017-02-19) 17 commits

311 - - refs: rename get_ref_store() to get_submodule_ref_store() and make it public

312 - - files-backend: remove submodule_allowed from files_downcast()

313 - - refs: move submodule code out of files-backend.c

314 - - path.c: move some code out of strbuf_git_path_submodule()

315 - - refs.c: make get_main_ref_store() public and use it

316 - - refs.c: kill register_ref_store(), add register_submodule_ref_store()

317 - - refs.c: flatten get_ref_store() a bit

318 - - refs: rename lookup_ref_store() to lookup_submodule_ref_store()

319 - - refs.c: introduce get_main_ref_store()

320 - - files-backend: remove the use of git_path()

321 - - refs.c: share is_per_worktree_ref() to files-backend.c

322 - - files-backend: replace *git_path*() with files_path()

323 - - files-backend: add files_path()

324 - - files-backend: convert git_path() to strbuf_git_path()

325 - - refs-internal.c: make files_log_ref_write() static

326 - - Merge branch 'mh/ref-remove-empty-directory' into nd/files-backend-git-dir

327 - - Merge branch 'mh/submodule-hash' into nd/files-backend-git-dir

328 - (this branch is used by nd/prune-in-worktree and nd/worktree-kill-parse-ref; uses mh/ref-remove-empty-directory and mh/submodule-hash.)

410 +* nd/files-backend-git-dir (2017-02-22) 26 commits

411 + . t1406: new tests for submodule ref store

412 + . t1405: some basic tests on main ref store

413 + . t/helper: add test-ref-store to test ref-store functions

414 + . refs: delete pack_refs() in favor of refs_pack_refs()

415 + . files-backend: avoid ref api targetting main ref store

416 + . refs: new transaction related ref-store api

417 + . refs: add new ref-store api

418 + . refs: rename get_ref_store() to get_submodule_ref_store() and make it public

419 + . files-backend: replace submodule_allowed check in files_downcast()

420 + . refs: move submodule code out of files-backend.c

421 + . path.c: move some code out of strbuf_git_path_submodule()

422 + . refs.c: make get_main_ref_store() public and use it

423 + . refs.c: kill register_ref_store(), add register_submodule_ref_store()

424 + . refs.c: flatten get_ref_store() a bit

425 + . refs: rename lookup_ref_store() to lookup_submodule_ref_store()

426 + . refs.c: introduce get_main_ref_store()

427 + . files-backend: remove the use of git_path()

428 + . files-backend: add and use files_refname_path()

429 + . files-backend: add and use files_reflog_path()

430 + . files-backend: move "logs/" out of TMP_RENAMED_LOG

431 + . files-backend: convert git_path() to strbuf_git_path()

432 + . files-backend: add and use files_packed_refs_path()

433 + . files-backend: make files_log_ref_write() static

434 + . refs.h: add forward declaration for structs used in this file

435 + . Merge branch 'mh/ref-remove-empty-directory' into nd/files-backend-git-dir

436 + . Merge branch 'mh/submodule-hash' into nd/files-backend-git-dir

437 + (this branch uses mh/ref-remove-empty-directory and mh/submodule-hash; is tangled with nd/prune-in-worktree and nd/worktree-kill-parse-ref.)

438

439 The "submodule" specific field in the ref_store structure is

440 replaced with a more generic "gitdir" that can later be used also

@@ -402,7 +511,7 @@ of the repositories listed at

511 cf. <xmqqmvdfh4az.fsf@gitster.mtv.corp.google.com>

512

513

405 -* bc/object-id (2017-02-20) 19 commits

514 +* bc/object-id (2017-02-22) 19 commits

515 - wt-status: convert to struct object_id

516 - builtin/merge-base: convert to struct object_id

517 - Convert object iteration callbacks to struct object_id

@@ -425,8 +534,8 @@ of the repositories listed at

534

535 "uchar [40]" to "struct object_id" conversion continues.

536

428 - Expecting a reroll.

429 - cf. <20170221011058.cpx7uio6ibkvrtbv@genre.crustytoothpaste.net>

537 + Now at v5; looked alright.

538 + cf. <20170221234737.894681-1-sandals@crustytoothpaste.net>

539

540

541 * jk/grep-no-index-fix (2017-02-14) 7 commits

What's cooking (2017/02 #07)