7d3989f219 · matheusbsbs/matheus-bagliano

@@ -1,17 +1,18 @@

1 To: git@vger.kernel.org

2 Bcc: lwn@lwn.net

3 -Subject: What's cooking in git.git (Oct 2012, #04; Thu, 11)

4 -X-master-at: cd46259ebf2e624bcee2aaae05c36663d414e1a2

5 -X-next-at: 639036db058fbb54865ae729c26c689777a96550

3 +Subject: What's cooking in git.git (Oct 2012, #05; Tue, 16)

4 +X-master-at: 4a182f6174e1b958777ba6fbc6262e5457151049

5 +X-next-at: 8a75ac8ee22538d824252e6f2aa1c01a120ced77

6

7 -What's cooking in git.git (Oct 2012, #04; Thu, 11)

7 +What's cooking in git.git (Oct 2012, #05; Tue, 16)

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

9

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

11 '-' are only in 'pu' (proposed updates) while commits prefixed with

12 '+' are in 'next'.

13

14 -The tip of 'master' is 1.8.0-rc2.

14 +We are almost there for 1.8.0-rc3. The final is expected to be

15 +tagged this weekend.

16

17 I'm planning to keep this cycle reasonably short and aim for tagging

18 the result as 1.8.0 at the end of 9th week, on October 21st, after

@@ -24,103 +25,113 @@ repositories listed at

25 http://git-blame.blogspot.com/p/git-public-repositories.html

26

27 --------------------------------------------------

27 -[New Topics]

28 +[Graduated to "master"]

29

29 -* js/format-2047 (2012-10-09) 5 commits

30 - - format-patch: tests: check rfc822+rfc2047 in to+cc headers

31 - - format-patch: fix rfc2047 address encoding with respect to rfc822 specials

32 - - format-patch: introduce helper function last_line_length()

33 - - format-patch: do not wrap rfc2047 encoded headers too late

34 - - format-patch: do not wrap non-rfc2047 headers too early

30 +* jk/maint-http-half-auth-push (2012-10-12) 1 commit

31 + (merged to 'next' on 2012-10-12 at 838ab48)

32 + + http: fix segfault in handle_curl_result

33 + (this branch is used by jk/maint-http-init-not-in-result-handler.)

34

36 - Fixes many rfc2047 quoting issues in the output from format-patch.

35 + Fixes a regression in maint-1.7.11 (v1.7.11.7), maint (v1.7.12.1)

36 + and master (v1.8.0-rc0).

37

38 +--------------------------------------------------

39 +[New Topics]

40

39 -* km/send-email-compose-encoding (2012-10-10) 1 commit

40 - (merged to 'next' on 2012-10-11 at d94bd05)

41 - + git-send-email: introduce compose-encoding

41 +* jk/maint-http-init-not-in-result-handler (2012-10-12) 2 commits

42 + (merged to 'next' on 2012-10-16 at cc88829)

43 + + http: do not set up curl auth after a 401

44 + + remote-curl: do not call run_slot repeatedly

45

43 - "git send-email --compose" can let the user create a non-ascii

44 - cover letter message, but there was not a way to mark it with

45 - appropriate content type before sending it out.

46 + Further clean-up to the http codepath that picks up results after

47 + cURL library is done with one request slot.

48

49 + Will merge to 'master' in the second batch after 1.8.0 ships.

50

48 -* so/prompt-command (2012-10-11) 2 commits

49 - (merged to 'next' on 2012-10-11 at 1a14825)

50 - + show color hints based on state of the git tree

51 - + Allow __git_ps1 to be used in PROMPT_COMMAND

51

53 - Updates __git_ps1 so that it can be used as $PROMPT_COMMAND,

54 - instead of being used for command substitution in $PS1, to embed

55 - color escape sequences in its output.

52 +* nd/grep-true-path (2012-10-12) 1 commit

53 + (merged to 'next' on 2012-10-16 at 8a75ac8)

54 + + grep: stop looking at random places for .gitattributes

55

56 + "git grep -e pattern <tree>" asked the attribute system to read

57 + "<tree>:.gitattributes" file in the working tree, which was

58 + nonsense.

59

58 -* jc/test-say-color-avoid-echo-escape (2012-10-11) 1 commit

59 - (merged to 'next' on 2012-10-11 at 639036d)

60 - + test-lib: Fix say_color () not to interpret \a\b\c in the message

60 + Will merge to 'master' in the second batch after 1.8.0 ships.

61

62 - Recent nd/wildmatch series was the first to reveal this ancient bug

63 - in the test scaffolding.

62

63 +* cr/cvsimport-local-zone (2012-10-15) 1 commit

64 + - git-cvsimport: allow author-specific timezones

65

66 -* aw/rebase-am-failure-detection (2012-10-11) 1 commit

67 - - rebase: Handle cases where format-patch fails

66 + Allows "cvsimport" to read per-author timezone from the author info

67 + file.

68

69

70 -* da/mergetools-p4 (2012-10-11) 1 commit

71 - - mergetools/p4merge: Handle "/dev/null"

70 +* fc/completion-send-email-with-format-patch (2012-10-16) 1 commit

71 + - completion: add format-patch options to send-email

72

73 ---------------------------------------------------

74 -[Graduated to "master"]

73

76 -* bw/cp-a-is-gnuism (2012-10-08) 1 commit

77 - (merged to 'next' on 2012-10-08 at 6ddb60d)

78 - + tests: "cp -a" is a GNUism

74 +* fc/zsh-completion (2012-10-15) 3 commits

75 + - completion: add new zsh completion

76 + - tests: use __gitcompadd to simplify completion tests

77 + - completion: add new __gitcompadd helper

78

79

81 -* jc/doc-long-options (2012-10-04) 1 commit

82 - (merged to 'next' on 2012-10-08 at 7183dc0)

83 - + gitcli: parse-options lets you omit tail of long options

80 +* jc/apply-trailing-blank-removal (2012-10-12) 1 commit

81 + - apply.c:update_pre_post_images(): the preimage can be truncated

82

85 - The end-user facing documentation claimed to be an "exhaustive"

86 - list of features of parse-options command line parser, but it

87 - didn't mention anything about abbreviating long option names to

88 - their unique prefixes. Mention it, warn against relying too much on

89 - it (as uniqueness may disappear in future versions of Git), and

90 - drop the claim to be exhaustive.

83 + Fix to update_pre_post_images() that did not take into account the

84 + possibility that whitespace fix could shrink the preimage and

85 + change the number of lines in it.

86

87 + Extra set of eyeballs appreciated.

88

93 -* jc/maint-t1450-fsck-order-fix (2012-10-02) 1 commit

94 - (merged to 'next' on 2012-10-02 at cc1a62e)

95 - + t1450: the order the objects are checked is undefined

89

97 - The fsck test assumed too much on what kind of error it will

98 - detect. The only important thing is the inconsistency is detected

99 - as an error.

90 +* jn/warn-on-inaccessible-loosen (2012-10-14) 4 commits

91 + - config: exit on error accessing any config file

92 + - doc: advertise GIT_CONFIG_NOSYSTEM

93 + - config: treat user and xdg config permission problems as errors

94 + - config, gitignore: failure to access with ENOTDIR is ok

95

96 + An RFC to deal with a situation where .config/git is a file and we

97 + notice .config/git/config is not readable due to ENOTDIR, not

98 + ENOENT; I think a bit more refactored approach to consistently

99 + address permission errors across config, exclude and attrs is

100 + desirable. Don't we also need a check for an opposite situation

101 + where we open .config/git/config or .gitattributes for reading but

102 + they turn out to be directories?

103

102 -* jc/merge-bases-paint-fix (2012-10-04) 1 commit

103 - (merged to 'next' on 2012-10-07 at 9fa34dc)

104 - + paint_down_to_common(): parse commit before relying on its timestamp

104

106 - "git fmt-merge-msg" (actually an internal helper reduce_heads() it

107 - uses) had a severe performance regression; an empty "git pull" took

108 - forever to finish as the result.

105 +* rs/lock-correct-ref-during-delete (2012-10-16) 1 commit

106 + - refs: lock symref that is to be deleted, not its target

107

108 + When "update-ref -d --no-deref SYM" tried to delete a symbolic ref

109 + SYM, it incorrectly locked the underlying reference pointed by SYM,

110 + not the symbolic ref itself.

111

111 -* nd/doc-ignore (2012-10-07) 1 commit

112 - (merged to 'next' on 2012-10-08 at 7b46836)

113 - + gitignore.txt: suggestions how to get literal # or ! at the beginning

112 + Will merge to 'next'.

113

115 - Documentation updates for excluding path whose name begin with # or !.

114 +--------------------------------------------------

115 +[Stalled]

116

117 +* mh/ceiling (2012-09-29) 9 commits

118 + - t1504: stop resolving symlinks in GIT_CEILING_DIRECTORIES

119 + - longest_ancestor_length(): resolve symlinks before comparing paths

120 + - longest_ancestor_length(): use string_list_longest_prefix()

121 + - longest_ancestor_length(): always add a slash to the end of prefixes

122 + - longest_ancestor_length(): explicitly filter list before loop

123 + - longest_ancestor_length(): use string_list_split()

124 + - Introduce new function real_path_if_valid()

125 + - real_path_internal(): add comment explaining use of cwd

126 + - Introduce new static function real_path_internal()

127

118 -* rr/git-uri-doc (2012-10-08) 1 commit

119 - (merged to 'next' on 2012-10-08 at a28ae12)

120 - + Git url doc: mark ftp/ftps as read-only and deprecate them

128 + Elements of GIT_CEILING_DIRECTORIES list may not match the real

129 + pathname we obtain from getcwd(), leading the GIT_DIR discovery

130 + logic to escape the ceilings the user thought to have specified.

131 +

132 + The solution felt a bit unnecessarily convoluted to me.

133 + Expecting a reroll.

134

122 ---------------------------------------------------

123 -[Stalled]

135

136 * rc/maint-complete-git-p4 (2012-09-24) 1 commit

137 (merged to 'next' on 2012-09-25 at 116e58f)

@@ -129,28 +140,6 @@ repositories listed at

140 Comment from Pete will need to be addressed in a follow-up patch.

141

142

132 -* as/check-ignore (2012-09-27) 17 commits

133 - - [SQUASH-FIX] 283d072 (Add git-check-ignore sub-command, 2012-09-20)

134 - - [SQUASH-FIX] 283d072 (Add git-check-ignore sub-command, 2012-09-20)

135 - - [REROLL NEEDED] minimum compilation fix

136 - - Add git-check-ignore sub-command

137 - - dir.c: provide free_directory() for reclaiming dir_struct memory

138 - - pathspec.c: move reusable code from builtin/add.c

139 - - dir.c: refactor treat_gitlinks()

140 - - dir.c: keep track of where patterns came from

141 - - dir.c: refactor is_path_excluded()

142 - - dir.c: refactor is_excluded()

143 - - dir.c: refactor is_excluded_from_list()

144 - - dir.c: rename excluded() to is_excluded()

145 - - dir.c: rename excluded_from_list() to is_excluded_from_list()

146 - - dir.c: rename path_excluded() to is_path_excluded()

147 - - dir.c: rename cryptic 'which' variable to more consistent name

148 - - Improve documentation and comments regarding directory traversal API

149 - - Update directory listing API doc to match code

150 -

151 - Waiting for a reroll.

152 -

153 -

143 * as/test-tweaks (2012-09-20) 7 commits

144 - tests: paint unexpectedly fixed known breakages in bold red

145 - tests: test the test framework more thoroughly

@@ -353,26 +342,111 @@ repositories listed at

342 --------------------------------------------------

343 [Cooking]

344

356 -* nd/wildmatch (2012-10-11) 12 commits

357 - - t3070: disable two fnmatch tests that have different results on different libc

345 +* as/check-ignore (2012-10-15) 12 commits

346 + - Add git-check-ignore sub-command

347 + - dir.c: provide free_directory() for reclaiming dir_struct memory

348 + - pathspec.c: move reusable code from builtin/add.c

349 + - dir.c: refactor treat_gitlinks()

350 + - dir.c: keep track of where patterns came from

351 + - dir.c: refactor is_path_excluded()

352 + - dir.c: refactor is_excluded()

353 + - dir.c: refactor is_excluded_from_list()

354 + - dir.c: rename excluded() to is_excluded()

355 + - dir.c: rename excluded_from_list() to is_excluded_from_list()

356 + - dir.c: rename path_excluded() to is_path_excluded()

357 + - dir.c: rename cryptic 'which' variable to more consistent name

358 + (this branch uses nd/attr-match-optim and nd/attr-match-optim-more; is tangled with nd/wildmatch.)

359 +

360 + Duy helped to reroll this. Needs a bit of documentation updates,

361 + at least.

362 +

363 +

364 +* js/format-2047 (2012-10-09) 5 commits

365 + - format-patch: tests: check rfc822+rfc2047 in to+cc headers

366 + - format-patch: fix rfc2047 address encoding with respect to rfc822 specials

367 + - format-patch: introduce helper function last_line_length()

368 + - format-patch: do not wrap rfc2047 encoded headers too late

369 + - format-patch: do not wrap non-rfc2047 headers too early

370 +

371 + Fixes many rfc2047 quoting issues in the output from format-patch.

372 + The last patch attempts to sneak new policies in without concensus,

373 + which I'd wish to see at least described as such i.e. "this fails

374 + with the current codebase but we would eventually want to do X and

375 + Y at which point it will pass". Then we can discuss if doing X and

376 + Y is a good thing. Alternatively, the bits that adds expect-failure

377 + should be removed from the patch, leaving the tests that cover what

378 + this series is supposed to have fixed.

379 +

380 + Otherwise, I think this is fit for 'next'.

381 +

382 +

383 +* km/send-email-compose-encoding (2012-10-10) 1 commit

384 + (merged to 'next' on 2012-10-11 at d94bd05)

385 + + git-send-email: introduce compose-encoding

386 +

387 + "git send-email --compose" can let the user create a non-ascii

388 + cover letter message, but there was not a way to mark it with

389 + appropriate content type before sending it out.

390 +

391 + Will merge to 'master' in the second batch after 1.8.0 ships.

392 +

393 +

394 +* so/prompt-command (2012-10-11) 2 commits

395 + (merged to 'next' on 2012-10-11 at 1a14825)

396 + + show color hints based on state of the git tree

397 + + Allow __git_ps1 to be used in PROMPT_COMMAND

398 +

399 + Updates __git_ps1 so that it can be used as $PROMPT_COMMAND,

400 + instead of being used for command substitution in $PS1, to embed

401 + color escape sequences in its output.

402 +

403 +

404 +* jc/test-say-color-avoid-echo-escape (2012-10-11) 1 commit

405 + (merged to 'next' on 2012-10-11 at 639036d)

406 + + test-lib: Fix say_color () not to interpret \a\b\c in the message

407 +

408 + Recent nd/wildmatch series was the first to reveal this ancient bug

409 + in the test scaffolding.

410 +

411 + Will merge to 'master' in the first batch after 1.8.0 ships.

412 +

413 +

414 +* aw/rebase-am-failure-detection (2012-10-11) 1 commit

415 + - rebase: Handle cases where format-patch fails

416 +

417 + I am unhappy a bit about the possible performance implications of

418 + having to store the output in a temporary file only for a rare case

419 + of format-patch aborting.

420 +

421 +

422 +* da/mergetools-p4 (2012-10-11) 1 commit

423 + (merged to 'next' on 2012-10-12 at 16f5c06)

424 + + mergetools/p4merge: Handle "/dev/null"

425 +

426 + Will merge to 'master' in the first batch after 1.8.0 ships.

427 +

428 +

429 +* nd/wildmatch (2012-10-15) 13 commits

430 - Support "**" wildcard in .gitignore and .gitattributes

431 - wildmatch: make /**/ match zero or more directories

432 - wildmatch: adjust "**" behavior

433 - wildmatch: fix case-insensitive matching

434 - wildmatch: remove static variable force_lower_case

435 - wildmatch: make wildmatch's return value compatible with fnmatch

436 + - t3070: disable unreliable fnmatch tests

437 - Integrate wildmatch to git

438 + - wildmatch: follow Git's coding convention

439 - wildmatch: remove unnecessary functions

440 - Import wildmatch from rsync

441 - ctype: support iscntrl, ispunct, isxdigit and isprint

442 - ctype: make sane_ctype[] const array

443 + (this branch uses nd/attr-match-optim and nd/attr-match-optim-more; is tangled with as/check-ignore.)

444

445 Allows pathname patterns in .gitignore and .gitattributes files

446 with double-asterisks "foo/**/bar" to match any number of directory

447 hiearchies.

448

374 - This might be missing some off-by-one fix posted on the list, and

375 - needs to be checked.

449 + Will merge to 'next'.

450

451

452 * jk/lua-hackery (2012-10-07) 6 commits

@@ -392,21 +466,29 @@ repositories listed at

466 (merged to 'next' on 2012-10-08 at bfbdd8a)

467 + attr: avoid searching for basename on every match

468 + attr: avoid strlen() on every match

395 - (this branch is used by nd/attr-match-optim-more.)

469 + (this branch is used by as/check-ignore, nd/attr-match-optim-more and nd/wildmatch.)

470

471 Trivial and obvious optimization for finding attributes that match

472 a given path.

473

474 + Will merge to 'master' in the first batch after 1.8.0 ships.

475 +

476

401 -* nd/attr-match-optim-more (2012-10-11) 3 commits

477 +* nd/attr-match-optim-more (2012-10-15) 7 commits

478 - attr: more matching optimizations from .gitignore

479 - gitignore: make pattern parsing code a separate function

404 - - Merge commit 'f9f6e2c' into nd/attr-match-optim-more

405 - (this branch uses nd/attr-match-optim.)

480 + - exclude: split pathname matching code into a separate function

481 + - exclude: fix a bug in prefix compare optimization

482 + - exclude: split basename matching code into a separate function

483 + - exclude: stricten a length check in EXC_FLAG_ENDSWITH case

484 + + Merge commit 'f9f6e2c' into nd/attr-match-optim-more

485 + (this branch is used by as/check-ignore and nd/wildmatch; uses nd/attr-match-optim.)

486

487 Start laying the foundation to build the "wildmatch" after we can

488 agree on its desired semantics.

489

490 + Will merge to 'next'.

491 +

492

493 * jc/grep-pcre-loose-ends (2012-10-09) 7 commits

494 (merged to 'next' on 2012-10-11 at fec8530)

@@ -422,6 +504,8 @@ repositories listed at

504 pattern as extended regular expression, and instead looked for the

505 string literally. The early part of this series is a fix for it.

506

507 + Will merge to 'master' in the second batch after 1.8.0 ships.

508 +

509

510 * jk/peel-ref (2012-10-04) 4 commits

511 (merged to 'next' on 2012-10-08 at 4adfa2f)

@@ -434,6 +518,8 @@ repositories listed at

518 other side of the connection) by reducing the cost to advertise the

519 branches and tags that are available in the repository.

520

521 + Will merge to 'master' in the first batch after 1.8.0 ships.

522 +

523

524 * fa/remote-svn (2012-10-07) 16 commits

525 (merged to 'next' on 2012-10-07 at 7b90cf4)

@@ -456,6 +542,7 @@ repositories listed at

542 (this branch is used by fa/vcs-svn.)

543

544 A GSoC project.

545 + Will merge to 'master' in the first batch after 1.8.0 ships.

546

547

548 * bw/config-lift-variable-name-length-limit (2012-10-01) 1 commit

@@ -465,6 +552,20 @@ repositories listed at

552 The configuration parser had an unnecessary hardcoded limit on

553 variable names that was not checked consistently. Lift the limit.

554

555 + Will merge to 'master' in the first batch after 1.8.0 ships.

556 +

557 +

558 +* jm/diff-context-config (2012-10-02) 2 commits

559 + (merged to 'next' on 2012-10-02 at e57700a)

560 + + t4055: avoid use of sed 'a' command

561 + (merged to 'next' on 2012-10-01 at 509a558)

562 + + diff: diff.context configuration gives default to -U

563 +

564 + Teaches a new configuration variable to "git diff" Porcelain and

565 + its friends.

566 +

567 + Will merge to 'master' in the first batch after 1.8.0 ships.

568 +

569

570 * jl/submodule-add-by-name (2012-09-30) 2 commits

571 (merged to 'next' on 2012-10-08 at 9408d8d)

@@ -481,36 +582,7 @@ repositories listed at

582 place in $GIT_DIR of the superproject that does not conflict with

583 the original submodule.

584

484 -

485 -* jm/diff-context-config (2012-10-02) 2 commits

486 - (merged to 'next' on 2012-10-02 at e57700a)

487 - + t4055: avoid use of sed 'a' command

488 - (merged to 'next' on 2012-10-01 at 509a558)

489 - + diff: diff.context configuration gives default to -U

490 -

491 - Teaches a new configuration variable to "git diff" Porcelain and

492 - its friends.

493 -

494 - Will defer to the next cycle.

495 -

496 -

497 -* mh/ceiling (2012-09-29) 9 commits

498 - - t1504: stop resolving symlinks in GIT_CEILING_DIRECTORIES

499 - - longest_ancestor_length(): resolve symlinks before comparing paths

500 - - longest_ancestor_length(): use string_list_longest_prefix()

501 - - longest_ancestor_length(): always add a slash to the end of prefixes

502 - - longest_ancestor_length(): explicitly filter list before loop

503 - - longest_ancestor_length(): use string_list_split()

504 - - Introduce new function real_path_if_valid()

505 - - real_path_internal(): add comment explaining use of cwd

506 - - Introduce new static function real_path_internal()

507 -

508 - Elements of GIT_CEILING_DIRECTORIES list may not match the real

509 - pathname we obtain from getcwd(), leading the GIT_DIR discovery

510 - logic to escape the ceilings the user thought to have specified.

511 -

512 - The solution felt a bit unnecessarily convoluted to me.

513 - Expecting a reroll.

585 + Will merge to 'master' in the second batch after 1.8.0 ships.

586

587

588 * jl/submodule-rm (2012-09-29) 1 commit

@@ -525,7 +597,7 @@ repositories listed at

597 is not embedded in its working tree but is elsewhere and uses the

598 gitfile mechanism to point at it.

599

528 - Will defer to the next cycle.

600 + Will merge to 'master' in the second batch after 1.8.0 ships.

601

602

603 * nd/pretty-placeholder-with-color-option (2012-09-30) 9 commits

@@ -541,8 +613,11 @@ repositories listed at

613

614

615 * jk/no-more-pre-exec-callback (2012-06-05) 1 commit

544 - - pager: drop "wait for output to run less" hack

616 + (merged to 'next' on 2012-10-12 at 69fed45)

617 + + pager: drop "wait for output to run less" hack

618

619 (Originally merged to 'next' on 2012-07-23)

620 + Removes a workaround for buggy version of less older than version

621 + 406.

622

548 - Will defer to the next cycle.

623 + Will merge to 'master' in the first batch after 1.8.0 ships.

What's cooking (2012/10 #05)