dfe9970c3e · 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 (Mar 2021, #04; Sun, 14)

4 -X-master-at: 8775279891fd186ce847b7ddb7f0ac2339504300

5 -X-next-at: c8bc67eb22fa994c4b7386cb001aaa95940877c6

3 +Subject: What's cooking in git.git (Mar 2021, #05; Wed, 17)

4 +X-master-at: a5828ae6b52137b913b978e16cd2334482eb4c1f

5 +X-next-at: 576ba9dcdaf1007243f5a5cb4bf1a1e7b8fcf850

6

7 -What's cooking in git.git (Mar 2021, #04; Sun, 14)

7 +What's cooking in git.git (Mar 2021, #05; Wed, 17)

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

9

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

@@ -12,7 +12,17 @@ only in 'seen' (formerly 'pu'---proposed updates) while commits prefixed

12 with '+' 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 -Git 2.31 final is expected to be tagged on Mar 15, 2021 (monday).

15 +Git 2.31 has been tagged. We will hopefully either (1) not see

16 +brown-paper-bag breakages at all, or (2) will see some but can

17 +immediately deal with them, in coming couple of days. After that,

18 +the 2.32 cycle will gain steam, starting next week. Let's see.

19 +

20 +Many topics marked as "Will cook in 'next'" should be marked for

21 +merging to 'master' now, but that hasn't happened yet. Please

22 +nominate those that should be kicked back from 'next' to 'seen', if

23 +there are any. Please also nominate those that should be among the

24 +first batch of topics to be in 'master' (after a handful of 2.31

25 +regression fixes go through 'next' down to 'master', that is).

26

27 Copies of the source code to Git live in many repositories, and the

28 following is a list of the ones I push into or their mirrors. Some

@@ -43,59 +53,102 @@ Release tarballs are available at:

53 https://www.kernel.org/pub/software/scm/git/

54

55 --------------------------------------------------

46 -[Graduated to 'master']

56 +[New Topics]

57

48 -* jn/mergetool-hideresolved-is-optional (2021-03-13) 2 commits

49 - (merged to 'next' on 2021-03-13 at 23f5a25716)

50 - + doc: describe mergetool configuration in git-mergetool(1)

51 - + mergetool: do not enable hideResolved by default

58 +* jc/calloc-fix (2021-03-15) 1 commit

59 + - xcalloc: use CALLOC_ARRAY() when applicable

60

53 - Disable the recent mergetool's hideresolved feature by default for

54 - backward compatibility and safety.

61 + Code clean-up.

62

63 + Will merge to 'next'.

64

57 -* tb/pack-revindex-on-disk (2021-02-26) 1 commit

58 - (merged to 'next' on 2021-03-12 at 7aa348d3b3)

59 - + pack-revindex.c: don't close unopened file descriptors

65

61 - Fix for a topic in 'master'.

66 +* ab/tree-walk-with-object-type (2021-03-17) 32 commits

67 + - tree-walk.h API: add a tree_entry_extract_type() function

68 + - blame: emit a better error on 'git blame directory'

69 + - tree-walk.h API: add a get_tree_entry_path() function

70 + - tree-walk.h API: add get_tree_entry_all()

71 + - tree-walk.h API: add a tree_entry_extract_all() function

72 + - tree-entry.h API: rename tree_entry_extract() to tree_entry_extract_mode()

73 + - tree-walk.h API: document and format tree_entry_extract()

74 + - tree-walk.h API: add get_tree_entry_type()

75 + - match-trees: use "tmp" for mode in shift_tree_by()

76 + - tree-walk.h API: rename get_tree_entry() to get_tree_entry_mode()

77 + - tree-walk.h API: formatting changes for subsequent commit

78 + - tree-walk.h users: use temporary variable(s) for "mode"

79 + - fsck.c: switch on "object_type" in fsck_walk_tree()

80 + - merge-ort: correct reference to test in 62fdec17a11

81 + - merge-tree tests: test for the mode comparison in same_entry()

82 + - tree-walk.h users: migrate miscellaneous "mode" to "object_type"

83 + - tree-walk.h users: refactor chained "mode" if/else into switch

84 + - tree-walk.h users: migrate "p->mode &&" pattern

85 + - tree.h API: make read_tree_fn_t take an "enum object_type"

86 + - archive: get rid of 'stage' parameter

87 + - tree.h users: format argument lists in archive.c

88 + - tree.h: format argument lists of read_tree_recursive() users

89 + - tree-walk.h users: switch object_type(...) to new .object_type

90 + - cache.h: have base_name_compare() take "is tree?", not "mode"

91 + - diff tests: test that "mode" is passed when sorting

92 + - mktree tests: test that "mode" is passed when sorting

93 + - fast-import tests: test for sorting dir/file foo v.s. foo.txt

94 + - tree-walk.c: migrate to using new "object_type" field when possible

95 + - tree-walk.h: add object_type member to name_entry

96 + - cache.h: add a comment to object_type()

97 + - notes & match-trees: use name_entry's "pathlen" member

98 + - diff.c: remove redundant canon_mode() call

99 + (this branch uses ab/read-tree.)

100

63 ---------------------------------------------------

64 -[New Topics]

101 + Code clean-up.

102

66 -* en/ort-perf-batch-9 (2021-03-10) 8 commits

67 - - diffcore-rename: avoid doing basename comparisons for irrelevant sources

68 - - merge-ort: skip rename detection entirely if possible

69 - - merge-ort: use relevant_sources to filter possible rename sources

70 - - merge-ort: precompute whether directory rename detection is needed

71 - - merge-ort: introduce wrappers for alternate tree traversal

72 - - merge-ort: add data structures for an alternate tree traversal

73 - - merge-ort: precompute subset of sources for which we need rename detection

74 - - diffcore-rename: enable filtering possible rename sources

75 - (this branch uses en/ort-perf-batch-8.)

103 + I am not exactly sure where this series wants to go, other than

104 + unnecessarily churning the code. Seems to break "diff --no-index"

105 + rather badly, too (e.g. t4050, t4002, among others).

106

107

78 -* js/http-pki-credential-store (2021-03-11) 2 commits

79 - - http: drop the check for an empty proxy password before approving

80 - - http: store credential when PKI auth is used

108 +* jh/fsmonitor-prework (2021-03-17) 1 commit

109 + - fsmonitor: avoid global-buffer-overflow READ when checking trivial response

110

82 - The http codepath learned to let the credential layer to cache the

83 - password used to unlock a certificate that has successfully been

84 - used.

111 + The fsmonitor interface read from its input without making sure

112 + there is something to read from. This bug is new in 2.31

113 + timeframe.

114

115 + Will merge to 'next'.

116

87 -* rs/avoid-null-statement-after-macro-call (2021-03-13) 2 commits

88 - - mem-pool: drop trailing semicolon from macro definition

89 - - block-sha1: drop trailing semicolon from macro definition

117

118 +* jk/bisect-peel-tag-fix (2021-03-17) 1 commit

119 + - bisect: peel annotated tags to commits

120

92 -* rs/calloc-array (2021-03-13) 2 commits

93 - - use CALLOC_ARRAY

94 - - git-compat-util.h: drop trailing semicolon from macro definition

121 + "git bisect" reimplemented more in C during 2.30 timeframe did not

122 + take an annotated tag as a good/bad endpoint well. This regression

123 + has been corrected.

124

125 + Will merge to 'next'.

126

97 -* jk/slimmed-down (2021-03-14) 1 commit

98 - - vcs-svn: remove header files as well

127 +

128 +* js/fsmonitor-unpack-fix (2021-03-17) 2 commits

129 + - fsmonitor: do not forget to release the token in `discard_index()`

130 + - fsmonitor: fix memory corruption in some corner cases

131 +

132 + The data structure used by fsmonitor interface was not properly

133 + duplicated during an in-core merge, leading to use-after-free etc.

134 +

135 + Will merge to 'next'.

136 +

137 +

138 +* jt/clone-unborn-head (2021-03-17) 1 commit

139 + - t5606: run clone branch name test with protocol v2

140 +

141 + Test fix.

142 +

143 + Will merge to 'next'.

144 +

145 +

146 +* km/config-doc-typofix (2021-03-17) 1 commit

147 + - config.txt: add missing period

148 +

149 + Docfix.

150 +

151 + Will merge to 'next'.

152

153 --------------------------------------------------

154 [Stalled]

@@ -116,47 +169,8 @@ Release tarballs are available at:

169 Looking good.

170

171

119 -* ab/fsck-api-cleanup (2021-02-18) 10 commits

120 - - fsck.h: update FSCK_OPTIONS_* for object_name

121 - - fsck.c: give "FOREACH_MSG_ID" a more specific name

122 - - fsck.c: undefine temporary STR macro after use

123 - - fsck.c: call parse_msg_type() early in fsck_set_msg_type()

124 - - fsck.h: move FSCK_{FATAL,INFO,ERROR,WARN,IGNORE} into an enum

125 - - fsck.c: rename remaining fsck_msg_id "id" to "msg_id"

126 - - fsck.c: move definition of msg_id into append_msg_id()

127 - - fsck.c: rename variables in fsck_set_msg_type() for less confusion

128 - - fsck.h: use "enum object_type" instead of "int"

129 - - fsck.h: indent arguments to of fsck_set_msg_type

130 -

131 - Preliminary fsck API clean-up.

132 -

133 - Expecting a reroll.

134 - cf. <xmqqczwxc8bw.fsf@gitster.g>

135 -

136 -

137 -* jk/symlinked-dotgitx-files (2020-10-23) 9 commits

138 - - docs: document symlink restrictions for .git* files

139 - - fsck: complain when .gitattributes or .gitignore is a symlink

140 - - verify_path(): disallow symlinks in .gitattributes and .gitignore

141 - - t0060: test obscured .gitattributes and .gitignore matching

142 - - t7450: test .gitmodules symlink matching against obscured names

143 - - t7450: test verify_path() handling of gitmodules

144 - - t7415: rename to expand scope

145 - - fsck_tree(): wrap some long lines

146 - - fsck_tree(): fix shadowed variable

147 -

148 - "git fsck" and the corresponding check done during the transport

149 - learned to ensure that in-tree files like `.gitignore` and

150 - `.gitattributes` are not symbolic links.

151 -

152 - It seems that there are real projects with .gitignore recorded as

153 - symlinks, which may need to loosen the fsck setting. Do we need to

154 - introduce a class that is separate from symlinked .gitmodules that

155 - has potential consequences that is more/less grave, so that these

156 - projects can opt out of the new checks?

157 -

158 -

159 -* ag/merge-strategies-in-c (2020-11-24) 13 commits

172 +* ag/merge-strategies-in-c (2021-03-15) 14 commits

173 + - SQUASH??? CALLOC_ARRAY()

174 - sequencer: use the "octopus" merge strategy without forking

175 - sequencer: use the "resolve" strategy without forking

176 - merge: use the "octopus" strategy without forking

@@ -179,6 +193,85 @@ Release tarballs are available at:

193 --------------------------------------------------

194 [Cooking]

195

196 +* ab/fsck-api-cleanup (2021-03-17) 19 commits

197 + - fetch-pack: use new fsck API to printing dangling submodules

198 + - fetch-pack: use file-scope static struct for fsck_options

199 + - fetch-pack: don't needlessly copy fsck_options

200 + - fsck.c: move gitmodules_{found,done} into fsck_options

201 + - fsck.c: add an fsck_set_msg_type() API that takes enums

202 + - fsck.c: pass along the fsck_msg_id in the fsck_error callback

203 + - fsck.[ch]: move FOREACH_FSCK_MSG_ID & fsck_msg_id from *.c to *.h

204 + - fsck.c: give "FOREACH_MSG_ID" a more specific name

205 + - fsck.c: undefine temporary STR macro after use

206 + - fsck.c: call parse_msg_type() early in fsck_set_msg_type()

207 + - fsck.h: re-order and re-assign "enum fsck_msg_type"

208 + - fsck.h: move FSCK_{FATAL,INFO,ERROR,WARN,IGNORE} into an enum

209 + - fsck.c: refactor fsck_msg_type() to limit scope of "int msg_type"

210 + - fsck.c: rename remaining fsck_msg_id "id" to "msg_id"

211 + - fsck.c: move definition of msg_id into append_msg_id()

212 + - fsck.c: rename variables in fsck_set_msg_type() for less confusion

213 + - fsck.h: use "enum object_type" instead of "int"

214 + - fsck.h: use designed initializers for FSCK_OPTIONS_{DEFAULT,STRICT}

215 + - fsck.c: refactor and rename common config callback

216 +

217 + Fsck API clean-up.

218 + cf. <20210317182054.5986-1-avarab@gmail.com>

219 +

220 +

221 +* en/ort-perf-batch-9 (2021-03-10) 8 commits

222 + - diffcore-rename: avoid doing basename comparisons for irrelevant sources

223 + - merge-ort: skip rename detection entirely if possible

224 + - merge-ort: use relevant_sources to filter possible rename sources

225 + - merge-ort: precompute whether directory rename detection is needed

226 + - merge-ort: introduce wrappers for alternate tree traversal

227 + - merge-ort: add data structures for an alternate tree traversal

228 + - merge-ort: precompute subset of sources for which we need rename detection

229 + - diffcore-rename: enable filtering possible rename sources

230 + (this branch uses en/ort-perf-batch-8.)

231 +

232 + More ort.

233 +

234 + Ready?

235 +

236 +

237 +* js/http-pki-credential-store (2021-03-11) 2 commits

238 + - http: drop the check for an empty proxy password before approving

239 + - http: store credential when PKI auth is used

240 +

241 + The http codepath learned to let the credential layer to cache the

242 + password used to unlock a certificate that has successfully been

243 + used.

244 +

245 + Will merge to 'next'.

246 +

247 +

248 +* rs/avoid-null-statement-after-macro-call (2021-03-17) 2 commits

249 + - mem-pool: drop trailing semicolon from macro definition

250 + - block-sha1: drop trailing semicolon from macro definition

251 +

252 + Fix macros that can silently inject unintended null-statements.

253 +

254 + Will merge to 'next'.

255 +

256 +

257 +* rs/calloc-array (2021-03-15) 3 commits

258 + - cocci: allow xcalloc(1, size)

259 + - use CALLOC_ARRAY

260 + - git-compat-util.h: drop trailing semicolon from macro definition

261 +

262 + CALLOC_ARRAY() macro replaces many uses of xcalloc().

263 +

264 + Will merge to 'next'.

265 +

266 +

267 +* jk/slimmed-down (2021-03-14) 1 commit

268 + - vcs-svn: remove header files as well

269 +

270 + Unused code removal.

271 +

272 + Will merge to 'next'.

273 +

274 +

275 * dl/stash-show-untracked (2021-03-05) 2 commits

276 (merged to 'next' on 2021-03-08 at 510d4cfa5f)

277 + stash show: learn stash.showIncludeUntracked

@@ -247,6 +340,7 @@ Release tarballs are available at:

340

341 Coccicheck update.

342

343 + Expecting a reroll.

344 IIUC, the "speed up" claimed in the intermediate steps are all gone

345 when correctness is brought back in with the last step. The series

346 needs a restructuring to avoid confuing readers with "it produces

@@ -255,17 +349,23 @@ Release tarballs are available at:

349 that fast" progression.

350

351

258 -* ab/read-tree (2021-03-11) 6 commits

259 - . tree.h API: remove "stage" parameter from read_tree_recursive()

260 - . tree.h API: remove support for starting at prefix != ""

261 - . ls-files: refactor away read_tree()

262 - . ls-files: don't needlessly pass around stage variable

263 - . tree.c API: move read_tree() into builtin/ls-files.c

264 - . ls-files tests: add meaningful --with-tree tests

352 +* ab/read-tree (2021-03-17) 9 commits

353 + - tree.h API: expose read_tree_1() as read_tree_at()

354 + - show tests: add test for "git show <tree>"

355 + - tree.h API: rename read_tree_recursive() to read_tree()

356 + - tree.h API: remove "stage" parameter from read_tree_recursive()

357 + - tree.h API: remove support for starting at prefix != ""

358 + - ls-files: refactor away read_tree()

359 + - ls-files: don't needlessly pass around stage variable

360 + - tree.c API: move read_tree() into builtin/ls-files.c

361 + - ls-files tests: add meaningful --with-tree tests

362 + (this branch is used by ab/tree-walk-with-object-type.)

363

364 Code simplification by removing support for a caller that is long gone.

365

268 - It turns out that a new caller is just about to happen.

366 + Expecting a reroll.

367 + Mostly good, but the progression shows "oops we screwed up" earlier

368 + in the series that is then fixed later.

369

370

371 * ab/unexpected-object-type (2021-03-08) 7 commits

@@ -279,6 +379,7 @@ Release tarballs are available at:

379

380 Error reporting upon object type mismatch has been improved

381

382 + Expecting a reroll.

383 Looked good except for some rewrites.

384

385

@@ -305,6 +406,8 @@ Release tarballs are available at:

406 Plug or annotate remaining leaks that trigger while running the

407 very basic set of tests.

408

409 + Ready?

410 +

411

412 * bc/clone-bare-with-conflicting-config (2021-03-10) 1 commit

413 (merged to 'next' on 2021-03-14 at 963a86199c)

@@ -557,7 +660,7 @@ Release tarballs are available at:

660 available.

661

662

560 -* cm/rebase-i-fixup-amend-reword (2021-03-13) 6 commits

663 +* cm/rebase-i-fixup-amend-reword (2021-03-15) 6 commits

664 - doc/git-commit: add documentation for fixup=[amend|reword] options

665 - t3437: use --fixup with options to create amend! commit

666 - t7500: add tests for --fixup=[amend|reword] options

@@ -572,9 +675,6 @@ Release tarballs are available at:

675 tweak both the message and the contents, and only the message,

676 respectively.

677

575 - Expecting a reroll.

576 - cf. <CAPSFM5c7ULGxThtFjPam33wMzK3kprds97zwJ63stc_Ct24nmQ@mail.gmail.com>

577 -

678

679 * jk/open-dotgitx-with-nofollow (2021-02-16) 6 commits

680 (merged to 'next' on 2021-02-25 at a784bf7be6)

@@ -630,7 +730,8 @@ Release tarballs are available at:

730 Will cook in 'next', together with "cm/rebase-i".

731

732

633 -* jh/simple-ipc (2021-03-09) 12 commits

733 +* jh/simple-ipc (2021-03-15) 13 commits

734 + - SQUASH??? CALLOC_ARRAY()

735 - t0052: add simple-ipc tests and t/helper/test-simple-ipc tool

736 - simple-ipc: add Unix domain socket implementation

737 - unix-stream-server: create unix domain socket under lock

@@ -785,3 +886,28 @@ Release tarballs are available at:

886 . doc: propose hooks managed by the config

887

888 Now replaced by es/config-hooks topic.

889 +

890 +

891 +

892 +* jk/symlinked-dotgitx-files (2020-10-23) 9 commits

893 + . docs: document symlink restrictions for .git* files

894 + . fsck: complain when .gitattributes or .gitignore is a symlink

895 + . verify_path(): disallow symlinks in .gitattributes and .gitignore

896 + . t0060: test obscured .gitattributes and .gitignore matching

897 + . t7450: test .gitmodules symlink matching against obscured names

898 + . t7450: test verify_path() handling of gitmodules

899 + . t7415: rename to expand scope

900 + . fsck_tree(): wrap some long lines

901 + . fsck_tree(): fix shadowed variable

902 +

903 + "git fsck" and the corresponding check done during the transport

904 + learned to ensure that in-tree files like `.gitignore` and

905 + `.gitattributes` are not symbolic links.

906 +

907 + It seems that there are real projects with .gitignore recorded as

908 + symlinks, which may need to loosen the fsck setting. Do we need to

909 + introduce a class that is separate from symlinked .gitmodules that

910 + has potential consequences that is more/less grave, so that these

911 + projects can opt out of the new checks?

912 +

913 +

What's cooking (2021/03 #05)