0ed03171ca · 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 (Jul 2016, #04; Mon, 11)

3 +Subject: What's cooking in git.git (Jul 2016, #05; Wed, 13)

4 X-master-at: b1ec08fda8e0c3ed19b680404770ca4ec313ca8c

5 X-next-at: aa0ab43525ebd305d89ffde30b5d5837a535491c

6

7 -What's cooking in git.git (Jul 2016, #04; Mon, 11)

7 +What's cooking in git.git (Jul 2016, #05; Wed, 13)

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

9

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

@@ -12,7 +12,10 @@ 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 first maintenance release for 2.9.x series has been issued.

15 +It turns out that v2.9.1 had tests that weren't even meant to pass

16 +on platforms with 32-bit time_t/unsigned long, but did not properly

17 +exclude them. A fix nas been prepared and I expect we'll cut v2.9.2

18 +with only that change relative to v2.9.1 soonish.

19

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

21 of the repositories listed at

@@ -20,139 +23,186 @@ of the repositories listed at

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

24

25 --------------------------------------------------

23 -[Graduated to "master"]

24 -

25 -* jc/send-email-skip-backup (2016-04-12) 1 commit

26 - (merged to 'next' on 2016-06-28 at 67dac59)

27 - + send-email: detect and offer to skip backup files

28 -

29 - A careless invocation of "git send-email directory/" after editing

30 - 0001-change.patch with an editor often ends up sending both

31 - 0001-change.patch and its backup file, 0001-change.patch~, causing

32 - embarrassment and a minor confusion. Detect such an input and

33 - offer to skip the backup files when sending the patches out.

34 - Perhaps people will enhance it more once it gains more visibility.

35 -

36 -

37 -* jk/ansi-color (2016-06-23) 7 commits

38 - (merged to 'next' on 2016-06-28 at 354989c)

39 - + color: support strike-through attribute

40 - + color: support "italic" attribute

41 - + color: allow "no-" for negating attributes

42 - + color: refactor parse_attr

43 - + add skip_prefix_mem helper

44 - + doc: refactor description of color format

45 - + color: fix max-size comment

46 -

47 - The output coloring scheme learned two new attributes, italic and

48 - strike, in addition to existing bold, reverse, etc.

49 -

50 -

51 -* jk/perf-any-version (2016-06-22) 2 commits

52 - (merged to 'next' on 2016-06-28 at 6a41c30)

53 - + p4211: explicitly disable renames in no-rename test

54 - + t/perf: fix regression in testing older versions of git

55 -

56 - Allow t/perf framework to use the features from the most recent

57 - version of Git even when testing an older installed version.

58 -

59 -

60 -* jn/preformatted-doc-url (2016-06-22) 1 commit

61 - (merged to 'next' on 2016-06-28 at 60a99eb)

62 - + doc: git-htmldocs.googlecode.com is no more

26 +[New Topics]

27

64 - The top level documentation "git help git" still pointed at the

65 - documentation set hosted at now-defunct google-code repository.

66 - Update it to point to https://git.github.io/htmldocs/git.html

67 - instead.

28 +* js/t0006-for-v2.9.2 (2016-07-12) 1 commit

29 + - t0006: skip "far in the future" test when unsigned long is not long enough

30 +

31 + A test merged to v2.9.1 forgot that the feature it was testing

32 + would not work on a platform with 32-bit time_t/unsigned long and

33 + reported breakage. Skip the tests that cannot run correctly on

34 + such platforms.

35 +

36 + Waiting for an Ack; will fast-track down to 'maint' to cut v2.9.2.

37 +

38 +

39 +* js/am-3-merge-recursive-direct (2016-07-12) 21 commits

40 + - merge-recursive: flush output buffer even when erroring out

41 + - merge_trees(): ensure that the output buffer is released after calling it

42 + - merge-recursive: offer an option to retain the output in 'obuf'

43 + - merge-recursive: write the commit title in one go

44 + - merge-recursive: flush output buffer before printing error messages

45 + - am -3: use merge_recursive() directly again

46 + - merge-recursive: switch to returning errors instead of dying

47 + - merge-recursive: handle return values indicating errors

48 + - merge-recursive: allow write_tree_from_memory() to error out

49 + - merge-recursive: avoid returning a wholesale struct

50 + - merge_recursive: abort properly upon errors

51 + - prepare the builtins for a libified merge_recursive()

52 + - merge-recursive: clarify code in was_tracked()

53 + - die(_("BUG")): avoid translating bug messages

54 + - die("bug"): report bugs consistently

55 + - t5520: verify that `git pull --rebase` shows the advice when failing

56 + - Merge branch 'jh/clean-smudge-annex' into HEAD

57 + - Merge branch 'bc/cocci' into js/am-3-merge-recursive-direct

58 + - Merge branch 'js/am-call-theirs-theirs-in-fallback-3way' into js/am-3-merge-recursive-direct

59 + - Merge branch 'cc/apply-am' into js/am-3-merge-recursive-direct

60 + - Merge branch 'va/i18n-even-more' into js/am-3-merge-recursive-direct

61 + (this branch uses bc/cocci, cc/apply-am, jh/clean-smudge-annex, js/am-call-theirs-theirs-in-fallback-3way and va/i18n-even-more.)

62 +

63 + "git am -3" calls "git merge-recursive" when it needs to fall back

64 + to a three-way merge; this call has been turned into an internal

65 + subroutine call instead of spawning a separate subprocess.

66 +

67 + Will need to wait for all the topics this depends on graduate to 'master'.

68 +

69 +

70 +* ls/travis-enable-httpd-tests (2016-07-12) 1 commit

71 + - travis-ci: enable web server tests t55xx on Linux

72 +

73 + Allow http daemon tests in Travis CI tests.

74 +

75 +

76 +* nd/cache-tree-ita (2016-07-12) 4 commits

77 + - cache-tree: do not generate empty trees as a result of all i-t-a subentries

78 + - cache-tree.c: fix i-t-a entry skipping directory updates sometimes

79 + - test-lib.sh: introduce and use $_EMPTY_BLOB

80 + - test-lib.sh: introduce and use $_EMPTY_TREE

81 +

82 + "git add -N dir/file && git write-tree" produced an incorrect tree

83 + when there are other paths in the same directory that sorts after

84 + "file".

85 +

86 + Looked mostly OK. Further review comments are welcome.

87 +

88 +

89 +* nd/pack-ofs-4gb-limit (2016-07-13) 7 commits

90 + - fsck: use streaming interface for large blobs in pack

91 + - pack-objects: do not truncate result in-pack object size on 32-bit systems

92 + - index-pack: correct "offset" type in unpack_entry_data()

93 + - index-pack: report correct bad object offsets even if they are large

94 + - index-pack: correct "len" type in unpack_data()

95 + - sha1_file.c: use type off_t* for object_info->disk_sizep

96 + - pack-objects: pass length to check_pack_crc() without truncation

97 +

98 + "git pack-objects" and "git index-pack" mostly operate with off_t

99 + when talking about the offset of objects in a packfile, but there

100 + were a handful of places that used "unsigned long" to hold that

101 + value, leading to an unintended truncation.

102

103 + Will merge to 'next'.

104

70 -* js/color-on-windows-comment (2016-06-28) 1 commit

71 - (merged to 'next' on 2016-06-28 at 38a2ea1)

72 - + color.h: remove obsolete comment about limitations on Windows

105

74 - For a long time, we carried an in-code comment that said our

75 - colored output would work only when we use fprintf/fputs on

76 - Windows, which no longer is the case for the past few years.

106 +* sb/push-options (2016-07-12) 5 commits

107 + - add a test for push options

108 + - push: accept push options

109 + - SQUASH???

110 + - receive-pack: implement advertising and receiving push options

111 + - push options: {pre,post}-receive hook learns about push options

112

113 + "git push" learned to accept and pass extra options to the

114 + receiving end so that hooks can read and react to them.

115

79 -* js/find-commit-subject-ignore-leading-blanks (2016-06-29) 5 commits

80 - (merged to 'next' on 2016-06-29 at 9717e00)

81 - + reset --hard: skip blank lines when reporting the commit subject

82 - + sequencer: use skip_blank_lines() to find the commit subject

83 - + commit -C: skip blank lines at the beginning of the message

84 - (merged to 'next' on 2016-06-28 at 32bbd94)

85 - + commit.c: make find_commit_subject() more robust

86 - + pretty: make the skip_blank_lines() function public

116 + Discussion continues, expecting a further reroll.

117 + ($gmane/299156)

118

88 - A helper function that takes the contents of a commit object and

89 - finds its subject line did not ignore leading blank lines, as is

90 - commonly done by other codepaths. Make it ignore leading blank

91 - lines to match.

119

120 +* ew/http-walker (2016-07-12) 3 commits

121 + - http-walker: reduce O(n) ops with doubly-linked list

122 + - http: avoid disconnecting on 404s for loose objects

123 + - http-walker: remove unused parameter from fetch_object

124

94 -* mj/log-show-signature-conf (2016-06-24) 3 commits

95 - (merged to 'next' on 2016-06-28 at c8b3347)

96 - + log: add log.showSignature configuration variable

97 - + log: add "--no-show-signature" command line option

98 - + t4202: refactor test

125 + Optimize dumb http transport on the client side.

126

100 - "git log" learns log.showSignature configuration variable, and a

101 - command line option "--no-show-signature" to countermand it.

127

128 +* nd/log-decorate-color-head-arrow (2016-07-12) 1 commit

129 + - log: decorate HEAD -> branch with the same color for arrow and HEAD

130

104 -* sb/submodule-clone-retry (2016-06-13) 2 commits

105 - (merged to 'next' on 2016-06-28 at 8a86d54)

106 - + submodule update: continue when a clone fails

107 - + submodule--helper: initial clone learns retry logic

131 + An entry "git log --decorate" for the tip of the current branch is

132 + shown as "HEAD -> name" (where "name" is the name of the branch);

133 + paint the arrow in the same color as "HEAD", not in the color for

134 + commits.

135

109 - "git submodule update" that drives many "git clone" could

110 - eventually hit flaky servers/network conditions on one of the

111 - submodules; the command learned to retry the attempt.

136 + Comments? Personally I find it more-or-less "Meh".

137

113 ---------------------------------------------------

114 -[New Topics]

138

116 -* js/ignore-space-at-eol (2016-07-11) 2 commits

117 - - diff: fix a double off-by-one with --ignore-space-at-eol

118 - - diff: demonstrate a bug with --patience and --ignore-space-at-eol

139 +* rs/rm-strbuf-optim (2016-07-12) 1 commit

140 + - rm: reuse strbuf for all remove_dir_recursively() calls

141

120 - An age old bug that caused "git diff --ignore-space-at-eol"

121 - misbehave has been fixed.

142 + The use of strbuf in "git rm" to build filename to remove was a bit

143 + suboptimal, which has been fixed.

144

145 Will merge to 'next'.

146

147

126 -* js/rebase-i-tests (2016-07-07) 3 commits

127 - - rebase -i: we allow extra spaces after fixup!/squash!

128 - - rebase -i: demonstrate a bug with --autosquash

129 - - t3404: add a test for the --gpg-sign option

148 +* rs/worktree-use-strbuf-absolute-path (2016-07-12) 1 commit

149 + - worktree: use strbuf_add_absolute_path() directly

150

131 - A few tests that specifically target "git rebase -i" have been

132 - added.

151 + Code simplification.

152

153 Will merge to 'next'.

154

136 -

155 --------------------------------------------------

156 [Stalled]

157

140 -* jh/clean-smudge-annex (2016-07-08) 9 commits

141 - - SQUASH??? compilation fix

142 - - use smudgeToFile filter in recursive merge

143 - - use smudgeToFile filter in git am

144 - - better recovery from failure of smudgeToFile filter

145 - - warn on unusable smudgeToFile/cleanFromFile config

146 - - use smudgeToFile in git checkout etc

147 - - use cleanFromFile in git add

148 - - add smudgeToFile and cleanFromFile filter configs

149 - - clarify %f documentation

158 +* cc/apply-am (2016-06-28) 41 commits

159 + - apply: use error_errno() where possible

160 + - builtin/am: use apply api in run_apply()

161 + - apply: change error_routine when be_silent is set

162 + - usage: add get_error_routine() and get_warn_routine()

163 + - usage: add set_warn_routine()

164 + - apply: don't print on stdout when be_silent is set

165 + - apply: make 'be_silent' incompatible with 'apply_verbosely'

166 + - apply: add 'be_silent' variable to 'struct apply_state'

167 + - write_or_die: use warning() instead of fprintf(stderr, ...)

168 + - environment: add set_index_file()

169 + - apply: make some parsing functions static again

170 + - apply: move libified code from builtin/apply.c to apply.{c,h}

171 + - apply: rename and move opt constants to apply.h

172 + - builtin/apply: rename option parsing functions

173 + - builtin/apply: make create_one_file() return -1 on error

174 + - builtin/apply: make try_create_file() return -1 on error

175 + - builtin/apply: make write_out_results() return -1 on error

176 + - builtin/apply: make write_out_one_result() return -1 on error

177 + - builtin/apply: make create_file() return -1 on error

178 + - builtin/apply: make add_index_file() return -1 on error

179 + - builtin/apply: make add_conflicted_stages_file() return -1 on error

180 + - builtin/apply: make remove_file() return -1 on error

181 + - builtin/apply: make build_fake_ancestor() return -1 on error

182 + - builtin/apply: change die_on_unsafe_path() to check_unsafe_path()

183 + - builtin/apply: make gitdiff_*() return -1 on error

184 + - builtin/apply: make gitdiff_*() return 1 at end of header

185 + - builtin/apply: make parse_traditional_patch() return -1 on error

186 + - builtin/apply: make apply_all_patches() return 128 or 1 on error

187 + - builtin/apply: move check_apply_state() to apply.c

188 + - builtin/apply: make check_apply_state() return -1 instead of die()ing

189 + - apply: make init_apply_state() return -1 instead of exit()ing

190 + - builtin/apply: move init_apply_state() to apply.c

191 + - builtin/apply: make parse_ignorewhitespace_option() return -1 instead of die()ing

192 + - builtin/apply: make parse_whitespace_option() return -1 instead of die()ing

193 + - builtin/apply: make parse_single_patch() return -1 on error

194 + - builtin/apply: make parse_chunk() return a negative integer on error

195 + - builtin/apply: make find_header() return -128 instead of die()ing

196 + - builtin/apply: read_patch_file() return -1 instead of die()ing

197 + - builtin/apply: make apply_patch() return -1 or -128 instead of die()ing

198 + - apply: move 'struct apply_state' to apply.h

199 + - apply: make some names more specific

200 + (this branch is used by jh/clean-smudge-annex and js/am-3-merge-recursive-direct.)

201

151 - The interface to "clean/smudge" filters require Git to feed the

152 - whole contents via pipe, which is suboptimal for some applications.

153 - "cleanFromFile/smudgeToFile" commands are the moral equilvalents

154 - for these filters but they interact with the files on the

155 - filesystem directly.

202 + "git am" has been taught to make an internal call to "git apply"'s

203 + innards without spawning the latter as a separate process.

204 +

205 + Needs review.

206

207

208 * tb/convert-peek-in-index (2016-07-07) 3 commits

@@ -262,7 +312,49 @@ of the repositories listed at

312 --------------------------------------------------

313 [Cooking]

314

265 -* jc/renormalize-merge-kill-safer-crlf (2016-07-08) 2 commits

315 +* jh/clean-smudge-annex (2016-07-12) 9 commits

316 + - use smudgeToFile filter in recursive merge

317 + - use smudgeToFile filter in git am

318 + - better recovery from failure of smudgeToFile filter

319 + - warn on unusable smudgeToFile/cleanFromFile config

320 + - use smudgeToFile in git checkout etc

321 + - use cleanFromFile in git add

322 + - add smudgeToFile and cleanFromFile filter configs

323 + - clarify %f documentation

324 + - Merge branch 'cc/apply-am' into jh/clean-smudge-annex

325 + (this branch is used by js/am-3-merge-recursive-direct; uses cc/apply-am.)

326 +

327 + The interface to "clean/smudge" filters require Git to feed the

328 + whole contents via pipe, which is suboptimal for some applications.

329 + "cleanFromFile/smudgeToFile" commands are the moral equilvalents

330 + for these filters but they interact with the files on the

331 + filesystem directly.

332 +

333 + Will need to wait for all the topics this depends on graduate to 'master'.

334 +

335 +

336 +* js/ignore-space-at-eol (2016-07-11) 2 commits

337 + - diff: fix a double off-by-one with --ignore-space-at-eol

338 + - diff: demonstrate a bug with --patience and --ignore-space-at-eol

339 +

340 + An age old bug that caused "git diff --ignore-space-at-eol"

341 + misbehave has been fixed.

342 +

343 + Will merge to 'next'.

344 +

345 +

346 +* js/rebase-i-tests (2016-07-07) 3 commits

347 + - rebase -i: we allow extra spaces after fixup!/squash!

348 + - rebase -i: demonstrate a bug with --autosquash

349 + - t3404: add a test for the --gpg-sign option

350 +

351 + A few tests that specifically target "git rebase -i" have been

352 + added.

353 +

354 + Will merge to 'next'.

355 +

356 +

357 +* jc/renormalize-merge-kill-safer-crlf (2016-07-12) 2 commits

358 - merge: avoid "safer crlf" during recording of merge results

359 - convert: unify the "auto" handling of CRLF

360 (this branch is tangled with tb/convert-peek-in-index.)

@@ -271,7 +363,7 @@ of the repositories listed at

363 merge-recursive, due to "safer crlf" conversion kicking in when it

364 shouldn't.

365

274 - Waiting for further discussion; hopefully can move this forward.

366 + Will merge to 'next', but extra set of eyes are appreciated.

367

368

369 * rs/help-c-source-with-gitattributes (2016-07-07) 1 commit

@@ -350,6 +442,7 @@ of the repositories listed at

442 * js/am-call-theirs-theirs-in-fallback-3way (2016-07-08) 1 commit

443 (merged to 'next' on 2016-07-11 at abae4bc)

444 + am: counteract gender bias

445 + (this branch is used by js/am-3-merge-recursive-direct.)

446

447 One part of "git am" had an oddball helper function that called

448 stuff from outside "his" as opposed to calling what we have "ours",

@@ -377,7 +470,8 @@ of the repositories listed at

470 interpret the paths relative to that directory, which has been

471 fixed.

472

380 - Waiting for an ack.

473 + Favourably reviewed with some questions and a request for new tests.

474 + ($gmane/299319)

475

476

477 * nd/test-helpers (2016-07-11) 3 commits

@@ -441,55 +535,6 @@ of the repositories listed at

535 Will merge to 'master'.

536

537

444 -* cc/apply-am (2016-06-28) 41 commits

445 - - apply: use error_errno() where possible

446 - - builtin/am: use apply api in run_apply()

447 - - apply: change error_routine when be_silent is set

448 - - usage: add get_error_routine() and get_warn_routine()

449 - - usage: add set_warn_routine()

450 - - apply: don't print on stdout when be_silent is set

451 - - apply: make 'be_silent' incompatible with 'apply_verbosely'

452 - - apply: add 'be_silent' variable to 'struct apply_state'

453 - - write_or_die: use warning() instead of fprintf(stderr, ...)

454 - - environment: add set_index_file()

455 - - apply: make some parsing functions static again

456 - - apply: move libified code from builtin/apply.c to apply.{c,h}

457 - - apply: rename and move opt constants to apply.h

458 - - builtin/apply: rename option parsing functions

459 - - builtin/apply: make create_one_file() return -1 on error

460 - - builtin/apply: make try_create_file() return -1 on error

461 - - builtin/apply: make write_out_results() return -1 on error

462 - - builtin/apply: make write_out_one_result() return -1 on error

463 - - builtin/apply: make create_file() return -1 on error

464 - - builtin/apply: make add_index_file() return -1 on error

465 - - builtin/apply: make add_conflicted_stages_file() return -1 on error

466 - - builtin/apply: make remove_file() return -1 on error

467 - - builtin/apply: make build_fake_ancestor() return -1 on error

468 - - builtin/apply: change die_on_unsafe_path() to check_unsafe_path()

469 - - builtin/apply: make gitdiff_*() return -1 on error

470 - - builtin/apply: make gitdiff_*() return 1 at end of header

471 - - builtin/apply: make parse_traditional_patch() return -1 on error

472 - - builtin/apply: make apply_all_patches() return 128 or 1 on error

473 - - builtin/apply: move check_apply_state() to apply.c

474 - - builtin/apply: make check_apply_state() return -1 instead of die()ing

475 - - apply: make init_apply_state() return -1 instead of exit()ing

476 - - builtin/apply: move init_apply_state() to apply.c

477 - - builtin/apply: make parse_ignorewhitespace_option() return -1 instead of die()ing

478 - - builtin/apply: make parse_whitespace_option() return -1 instead of die()ing

479 - - builtin/apply: make parse_single_patch() return -1 on error

480 - - builtin/apply: make parse_chunk() return a negative integer on error

481 - - builtin/apply: make find_header() return -128 instead of die()ing

482 - - builtin/apply: read_patch_file() return -1 instead of die()ing

483 - - builtin/apply: make apply_patch() return -1 or -128 instead of die()ing

484 - - apply: move 'struct apply_state' to apply.h

485 - - apply: make some names more specific

486 -

487 - "git am" has been taught to make an internal call to "git apply"'s

488 - innards without spawning the latter as a separate process.

489 -

490 - Needs review.

491 -

492 -

538 * dp/autoconf-curl-ssl (2016-06-28) 1 commit

539 - ./configure.ac: detect SSL in libcurl using curl-config

540

@@ -810,11 +855,12 @@ of the repositories listed at

855 + coccinelle: convert hashcpy() with null_sha1 to hashclr()

856 + contrib/coccinelle: add basic Coccinelle transforms

857 + hex: add oid_to_hex_r()

858 + (this branch is used by js/am-3-merge-recursive-direct.)

859

860 Conversion from unsigned char sha1[20] to struct object_id

861 continues.

862

817 - Will cook a bit in 'next' and then merge to 'master'.

863 + Will merge to 'master'.

864

865

866 * jk/big-and-future-archive-tar (2016-07-01) 5 commits

@@ -996,6 +1042,7 @@ of the repositories listed at

1042 + i18n: advice: mark string about detached head for translation

1043 + i18n: builtin/remote.c: fix mark for translation

1044 + Merge branch 'jc/t2300-setup' into HEAD

1045 + (this branch is used by js/am-3-merge-recursive-direct.)

1046

1047 More markings of messages for i18n, with updates to various tests

1048 to pass GETTEXT_POISON tests.

@@ -1065,7 +1112,7 @@ of the repositories listed at

1112 back-burner the threading enhancement to another day.

1113

1114

1068 -* pb/bisect (2016-07-11) 9 commits

1115 +* pb/bisect (2016-07-12) 9 commits

1116 - bisect--helper: `bisect_write` shell function in C

1117 - bisect--helper: `is_expected_rev` & `check_expected_revs` shell function in C

1118 - bisect--helper: `bisect_reset` shell function in C

What's cooking (2016/07 #05)