10d07ef269 · 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 (Oct 2017, #01; Wed, 4)

4 -X-master-at: 8fb8a945bc2dea2bb04249213ad8dacffbfc604f

5 -X-next-at: 6663358d3414aff4d7dac0c6f06cbfcc2cb48ddb

3 +Subject: What's cooking in git.git (Oct 2017, #02; Wed, 11)

4 +X-master-at: 111ef79afe185f8731920569450f6a65320f5d5f

5 +X-next-at: 92ea9504509327dacaba929d631d198e57533c14

6

7 -What's cooking in git.git (Oct 2017, #01; Wed, 4)

7 +What's cooking in git.git (Oct 2017, #02; Wed, 11)

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

9

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

@@ -12,9 +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 -I wanted to do v2.15-rc0 today but it seems that this has to slip,

16 -with too many topics still in flight in 'next' among which many are

17 -fixes, one is even a fix for a recent regression.

15 +2.15-rc1 has been tagged. Some of the topic marked "Will merge to

16 +'master'" below might still become part of 2.15 final, while others

17 +may be left to be merged after the final, to become part of the

18 +release after that one (2.16?).

19

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

21 of the repositories listed at

@@ -24,248 +25,145 @@ of the repositories listed at

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

26 [Graduated to "master"]

27

27 -* ad/doc-markup-fix (2017-09-29) 1 commit

28 - (merged to 'next' on 2017-10-02 at fdf600e8f3)

29 - + doc: correct command formatting

30 -

31 - Docfix.

32 -

28 +* ar/request-pull-phrasofix (2017-10-03) 1 commit

29 + (merged to 'next' on 2017-10-03 at 07807bba90)

30 + + request-pull: capitalise "Git" to make it a proper noun

31

34 -* bc/rev-parse-parseopt-fix (2017-09-25) 7 commits

35 - (merged to 'next' on 2017-09-26 at f3e013eaa4)

36 - + parse-options: only insert newline in help text if needed

37 - + parse-options: write blank line to correct output stream

38 - + t0040,t1502: Demonstrate parse_options bugs

39 - (merged to 'next' on 2017-09-24 at e479bce0ff)

40 - + git-rebase: don't ignore unexpected command line arguments

41 - + rev-parse parseopt: interpret any whitespace as start of help text

42 - + rev-parse parseopt: do not search help text for flag chars

43 - + t1502: demonstrate rev-parse --parseopt option mis-parsing

32 + Spell the name of our system as "Git" in the output from

33 + request-pull script.

34

45 - Recent versions of "git rev-parse --parseopt" did not parse the

46 - option specification that does not have the optional flags (*=?!)

47 - correctly, which has been corrected.

35

36 +* ds/avoid-overflow-in-midpoint-computation (2017-10-10) 1 commit

37 + (merged to 'next' on 2017-10-10 at 6279867a5d)

38 + + cleanup: fix possible overflow errors in binary search

39

50 -* bw/git-clang-format (2017-10-01) 1 commit

51 - (merged to 'next' on 2017-10-02 at ee84c89b07)

52 - + clang-format: adjust line break penalties

53 - (this branch is used by sb/git-clang-format.)

40 + Code clean-up.

41

55 - Adjust clang-format penalty parameters.

42

43 +* er/fast-import-dump-refs-on-checkpoint (2017-09-29) 1 commit

44 + (merged to 'next' on 2017-10-03 at 4e7b0e7ec1)

45 + + fast-import: checkpoint: dump branches/tags/marks even if object_count==0

46

58 -* hn/path-ownership-comment (2017-09-27) 2 commits

59 - (merged to 'next' on 2017-09-28 at 7cf8481c29)

60 - + read_gitfile_gently: clarify return value ownership.

61 - + real_path: clarify return value ownership

47 + The checkpoint command "git fast-import" did not flush updates to

48 + refs and marks unless at least one object was created since the

49 + last checkpoint, which has been corrected, as these things can

50 + happen without any new object getting created.

51

63 - Add comment to a few functions that use a short-lived buffer the

64 - caller can peek and copy out of.

52

53 +* hn/string-list-doc (2017-10-06) 1 commit

54 + (merged to 'next' on 2017-10-10 at d214a7c369)

55 + + api-argv-array.txt: remove broken link to string-list API

56

67 -* hn/string-list-doc (2017-09-27) 1 commit

68 - (merged to 'next' on 2017-09-28 at ffc8f65f4a)

69 - + string-list.h: move documentation from Documentation/api/ into header

57 + Docfix.

58

71 - Doc reorg.

59

60 +* jk/refs-df-conflict (2017-10-07) 2 commits

61 + (merged to 'next' on 2017-10-10 at d5953a1ced)

62 + + refs_resolve_ref_unsafe: handle d/f conflicts for writes

63 + + t3308: create a real ref directory/file conflict

64

74 -* hn/submodule-comment (2017-09-26) 1 commit

75 - (merged to 'next' on 2017-09-28 at a071ca3fde)

76 - + submodule.c: describe submodule_to_gitdir() in a new comment

65 + An ancient bug that made Git misbehave with creation/renaming of

66 + refs has been fixed.

67

68

79 -* jk/no-optional-locks (2017-09-27) 1 commit

80 - (merged to 'next' on 2017-09-28 at eaffe9c638)

81 - + git: add --no-optional-locks option

69 +* jk/sha1-loose-object-info-fix (2017-10-06) 1 commit

70 + (merged to 'next' on 2017-10-10 at bcfd9c4e3f)

71 + + sha1_loose_object_info: handle errors from unpack_sha1_rest

72

83 - Some commands (most notably "git status") makes an opportunistic

84 - update when performing a read-only operation to help optimize later

85 - operations in the same repository. The new "--no-optional-locks"

86 - option can be passed to Git to disable them.

73 + Leakfix and futureproofing.

74

75

89 -* jk/read-in-full (2017-09-27) 7 commits

90 - (merged to 'next' on 2017-09-28 at 9d109b9233)

91 - + worktree: check the result of read_in_full()

92 - + worktree: use xsize_t to access file size

93 - + distinguish error versus short read from read_in_full()

94 - + avoid looking at errno for short read_in_full() returns

95 - + prefer "!=" when checking read_in_full() result

96 - + notes-merge: drop dead zero-write code

97 - + files-backend: prefer "0" for write_in_full() error check

76 +* jk/ui-color-always-to-auto (2017-10-04) 3 commits

77 + (merged to 'next' on 2017-10-05 at 792ae936cf)

78 + + Merge branch 'jk/ui-color-always-to-auto-maint' into jk/ui-color-always-to-auto

79 + + t7301: use test_terminal to check color

80 + + t4015: use --color with --color-moved

81 + (this branch uses jk/ui-color-always-to-auto-maint.)

82

99 - Code clean-up to prevent future mistakes by copying and pasting

100 - code that checks the result of read_in_full() function.

83 + Fix regression of "git add -p" for users with "color.ui = always"

84 + in their configuration, by merging the topic below and adjusting it

85 + for the 'master' front.

86

87

103 -* jk/validate-headref-fix (2017-09-27) 3 commits

104 - (merged to 'next' on 2017-09-28 at dcea9d16f9)

105 - + validate_headref: use get_oid_hex for detached HEADs

106 - + validate_headref: use skip_prefix for symref parsing

107 - + validate_headref: NUL-terminate HEAD buffer

88 +* jk/ui-color-always-to-auto-maint (2017-10-04) 10 commits

89 + + color: make "always" the same as "auto" in config

90 + + provide --color option for all ref-filter users

91 + + t3205: use --color instead of color.branch=always

92 + + t3203: drop "always" color test

93 + + t6006: drop "always" color config tests

94 + + t7502: use diff.noprefix for --verbose test

95 + + t7508: use test_terminal for color output

96 + + t3701: use test-terminal to collect color output

97 + + t4015: prefer --color to -c color.diff=always

98 + + test-terminal: set TERM=vt100

99 + (this branch is used by jk/ui-color-always-to-auto.)

100

109 - Code clean-up.

101 + Users with "color.ui = always" in their configuration were broken

102 + by a recent change that made plumbing commands to pay attention to

103 + them as the patch created internally by "git add -p" were colored

104 + (heh) and made unusable. Fix this regression by redefining

105 + 'always' to mean the same thing as 'auto'.

106

107

112 -* js/rebase-i-final (2017-07-27) 10 commits

113 - (merged to 'next' on 2017-09-26 at ea3f8f5e11)

114 - + rebase -i: rearrange fixup/squash lines using the rebase--helper

115 - + t3415: test fixup with wrapped oneline

116 - + rebase -i: skip unnecessary picks using the rebase--helper

117 - + rebase -i: check for missing commits in the rebase--helper

118 - + t3404: relax rebase.missingCommitsCheck tests

119 - + rebase -i: also expand/collapse the SHA-1s via the rebase--helper

120 - + rebase -i: do not invent onelines when expanding/collapsing SHA-1s

121 - + rebase -i: remove useless indentation

122 - + rebase -i: generate the script via rebase--helper

123 - + t3415: verify that an empty instructionFormat is handled as before

124 -

125 - The final batch to "git rebase -i" updates to move more code from

126 - the shell script to C.

127 -

128 -

129 -* mh/mmap-packed-refs (2017-09-25) 21 commits

130 - (merged to 'next' on 2017-09-29 at 3639417666)

131 - + packed-backend.c: rename a bunch of things and update comments

132 - + mmapped_ref_iterator: inline into `packed_ref_iterator`

133 - + ref_cache: remove support for storing peeled values

134 - + packed_ref_store: get rid of the `ref_cache` entirely

135 - + ref_store: implement `refs_peel_ref()` generically

136 - + packed_read_raw_ref(): read the reference from the mmapped buffer

137 - + packed_ref_iterator_begin(): iterate using `mmapped_ref_iterator`

138 - + read_packed_refs(): ensure that references are ordered when read

139 - + packed_ref_cache: keep the `packed-refs` file mmapped if possible

140 - + packed-backend.c: reorder some definitions

141 - + mmapped_ref_iterator_advance(): no peeled value for broken refs

142 - + mmapped_ref_iterator: add iterator over a packed-refs file

143 - + packed_ref_cache: remember the file-wide peeling state

144 - + read_packed_refs(): read references with minimal copying

145 - + read_packed_refs(): make parsing of the header line more robust

146 - + read_packed_refs(): only check for a header at the top of the file

147 - + read_packed_refs(): use mmap to read the `packed-refs` file

148 - + die_unterminated_line(), die_invalid_line(): new functions

149 - + packed_ref_cache: add a backlink to the associated `packed_ref_store`

150 - + prefix_ref_iterator: break when we leave the prefix

151 - + ref_iterator: keep track of whether the iterator output is ordered

152 -

153 - Operations that do not touch (majority of) packed refs have been

154 - optimized by making accesses to packed-refs file lazy; we no longer

155 - pre-parse everything, and an access to a single ref in the

156 - packed-refs does not touch majority of irrelevant refs, either.

157 -

158 -

159 -* mr/doc-negative-pathspec (2017-09-25) 1 commit

160 - (merged to 'next' on 2017-09-29 at 147bb7ef8b)

161 - + docs: improve discoverability of exclude pathspec

162 -

163 - Doc updates.

164 -

165 -

166 -* sb/diff-color-move (2017-09-28) 1 commit

167 - (merged to 'next' on 2017-09-28 at cbdbc741be)

168 - + diff: correct newline in summary for renamed files

169 -

170 - The output from "git diff --summary" was broken in a recent topic

171 - that has been merged to 'master' and lost a LF after reporting of

172 - mode change. This has been fixed.

173 -

174 -

175 -* sb/submodule-diff-header-fix (2017-09-28) 1 commit

176 - (merged to 'next' on 2017-09-28 at 10e0f59303)

177 - + submodule: correct error message for missing commits

108 +* jn/strbuf-doc-re-reuse (2017-10-04) 1 commit

109 + (merged to 'next' on 2017-10-04 at 5940d412d9)

110 + + strbuf doc: reuse after strbuf_release is fine

111

179 - Error message tweak.

112

113 +* jr/hash-migration-plan-doc (2017-09-28) 1 commit

114 + (merged to 'next' on 2017-10-04 at b47b3bb656)

115 + + technical doc: add a design doc for hash function transition

116

182 -* sb/test-submodule-update-config (2017-09-27) 1 commit

183 - (merged to 'next' on 2017-09-28 at 497214603c)

184 - + t7406: submodule.<name>.update command must not be run from .gitmodules

117 + Lay out plans for weaning us off of SHA-1.

118

119

187 -* sd/branch-copy (2017-09-24) 4 commits

188 - (merged to 'next' on 2017-09-28 at a6eceefa02)

189 - + branch: fix "copy" to never touch HEAD

190 - + branch: add a --copy (-c) option to go with --move (-m)

191 - + branch: add test for -m renaming multiple config sections

192 - + config: create a function to format section headers

120 +* jt/oidmap (2017-10-01) 1 commit

121 + (merged to 'next' on 2017-10-05 at e41445fc33)

122 + + oidmap: map with OID as key

123

194 - "git branch" learned "-c/-C" to create a new branch by copying an

195 - existing one.

124 + Introduce a new "oidmap" API and rewrite oidset to use it.

125

197 ---------------------------------------------------

198 -[New Topics]

126

200 -* jc/merge-symlink-ours-theirs (2017-09-26) 1 commit

201 - - merge: teach -Xours/-Xtheirs to symbolic link merge

127 +* ks/branch-tweak-error-message-for-extra-args (2017-10-04) 1 commit

128 + (merged to 'next' on 2017-10-05 at aa0b656bf0)

129 + + branch: change the error messages to be more meaningful

130

203 - "git merge -Xours/-Xtheirs" learned to use our/their version when

204 - resolving a conflicting updates to a symbolic link.

131 + Error message tweak.

132

206 - Needs review.

133

134 +* ks/verify-filename-non-option-error-message-tweak (2017-10-04) 1 commit

135 + (merged to 'next' on 2017-10-05 at 2a7030f6fe)

136 + + setup: update error message to be more meaningful

137

209 -* jt/oidmap (2017-10-01) 1 commit

210 - - oidmap: map with OID as key

138 + Error message tweak.

139

212 - Introduce a new "oidmap" API and rewrite oidset to use it.

140

214 - Will merge to 'next'.

141 +* ls/filter-process-delayed (2017-10-10) 5 commits

142 + (merged to 'next' on 2017-10-10 at 7b26d72991)

143 + + write_entry: untangle symlink and regular-file cases

144 + + write_entry: avoid reading blobs in CE_RETRY case

145 + + write_entry: fix leak when retrying delayed filter

146 + + entry.c: check if file exists after checkout

147 + + entry.c: update cache entry only for existing files

148

216 -

217 -* jt/partial-clone-lazy-fetch (2017-10-02) 18 commits

218 - - fetch-pack: restore save_commit_buffer after use

219 - - unpack-trees: batch fetching of missing blobs

220 - - clone: configure blobmaxbytes in created repos

221 - - clone: support excluding large blobs

222 - - fetch: support excluding large blobs

223 - - fetch: refactor calculation of remote list

224 - - fetch-pack: support excluding large blobs

225 - - pack-objects: support --blob-max-bytes

226 - - pack-objects: rename want_.* to ignore_.*

227 - - gc: do not repack promisor packfiles

228 - - rev-list: support termination at promisor objects

229 - - sha1_file: support lazily fetching missing objects

230 - - introduce fetch-object: fetch one promisor object

231 - - index-pack: refactor writing of .keep files

232 - - fsck: support promisor objects as CLI argument

233 - - fsck: support referenced promisor objects

234 - - fsck: support refs pointing to promisor objects

235 - - fsck: introduce partialclone extension

236 -

237 - A journey for "git clone" and "git fetch" to become "lazier" by

238 - depending more on its remote repository---this is the beginning of

239 - it.

240 -

241 - Needs review.

149 + Bugfixes to an already graduated series.

150

151

152 * ma/builtin-unleak (2017-10-02) 1 commit

245 - - builtin/: add UNLEAKs

153 + (merged to 'next' on 2017-10-05 at 380cb74320)

154 + + builtin/: add UNLEAKs

155

156 Many variables that points at a region of memory that will live

157 throughout the life of the program have been marked with UNLEAK

158 marker to help the leak checkers concentrate on real leaks..

159

160

252 -* ma/lockfile-fixes (2017-10-02) 11 commits

253 - - read-cache: roll back lock on error with `COMMIT_LOCK`

254 - - read-cache: don't leave dangling pointer in `do_write_index()`

255 - - read-cache: require flags for `write_locked_index()`

256 - - cache.h: document `write_locked_index()`

257 - - apply: remove `newfd` from `struct apply_state`

258 - - apply: move lockfile into `apply_state`

259 - - cache-tree: simplify locking logic

260 - - tempfile: fix documentation on `delete_tempfile()`

261 - - lockfile: fix documentation on `close_lock_file_gently()`

262 - - treewide: prefer lockfiles on the stack

263 - - sha1_file: do not leak `lock_file`

161 +* rb/compat-poll-fix (2017-09-29) 1 commit

162 + (merged to 'next' on 2017-10-05 at c31029202d)

163 + + poll.c: always set revents, even if to zero

164

265 - An earlier update made it possible to use an on-stack in-core

266 - lockfile structure (as opposed to having to deliberately leak an

267 - on-heap one). Many codepaths have been updated to take advantage

268 - of this new facility.

165 + Backports a moral equivalent of 2015 fix to the poll emulation from

166 + the upstream gnulib to fix occasional breakages on HPE NonStop.

167

168

169 * rs/cleanup-strbuf-users (2017-10-02) 3 commits

@@ -276,8 +174,6 @@ of the repositories listed at

174

175 Code clean-up.

176

279 - Will merge to 'master'.

280 -

177

178 * rs/cocci-de-paren-call-params (2017-10-02) 1 commit

179 (merged to 'next' on 2017-10-03 at ee97fe5c24)

@@ -285,7 +181,30 @@ of the repositories listed at

181

182 Code clean-up.

183

288 - Will merge to 'master'.

184 +

185 +* rs/fsck-null-return-from-lookup (2017-10-06) 1 commit

186 + (merged to 'next' on 2017-10-10 at 32dfa424cf)

187 + + fsck: handle NULL return of lookup_blob() and lookup_tree()

188 +

189 + Improve behaviour of "git fsck" upon finding a missing object.

190 +

191 +

192 +* rs/qsort-s (2017-10-04) 1 commit

193 + (merged to 'next' on 2017-10-04 at 5f10a6e263)

194 + + test-stringlist: avoid buffer underrun when sorting nothing

195 +

196 +

197 +* rs/resolve-ref-optional-result (2017-10-01) 2 commits

198 + (merged to 'next' on 2017-10-03 at f129d4b0e9)

199 + + refs: pass NULL to resolve_refdup() if hash is not needed

200 + + refs: pass NULL to refs_resolve_refdup() if hash is not needed

201 +

202 + Code clean-up.

203 +

204 +

205 +* rs/rs-mailmap (2017-10-06) 1 commit

206 + (merged to 'next' on 2017-10-10 at 020d29b68c)

207 + + .mailmap: normalize name for René Scharfe

208

209

210 * rs/run-command-use-alloc-array (2017-10-03) 1 commit

@@ -294,8 +213,6 @@ of the repositories listed at

213

214 Code clean-up.

215

297 - Will merge to 'master'.

298 -

216

217 * rs/tag-null-pointer-arith-fix (2017-10-02) 1 commit

218 (merged to 'next' on 2017-10-03 at e05d92fec9)

@@ -303,8 +220,6 @@ of the repositories listed at

220

221 Code clean-up.

222

306 - Will merge to 'master'.

307 -

223

224 * rs/use-free-and-null (2017-10-02) 1 commit

225 (merged to 'next' on 2017-10-03 at 6340234d67)

@@ -312,7 +227,10 @@ of the repositories listed at

227

228 Code clean-up.

229

315 - Will merge to 'master'.

230 +

231 +* sb/branch-avoid-repeated-strbuf-release (2017-10-04) 1 commit

232 + (merged to 'next' on 2017-10-04 at bd86ef77fc)

233 + + branch: reset instead of release a strbuf

234

235

236 * sb/git-clang-format (2017-10-02) 1 commit

@@ -322,102 +240,153 @@ of the repositories listed at

240 Add comment to clarify that the style file is meant to be used with

241 clang-5 and the rules are still work in progress.

242

325 - Will merge to 'master'.

243

244 +* sb/test-cmp-expect-actual (2017-10-07) 1 commit

245 + (merged to 'next' on 2017-10-10 at c47a4dcc72)

246 + + tests: fix diff order arguments in test_cmp

247

328 -* rs/resolve-ref-optional-result (2017-10-01) 2 commits

329 - (merged to 'next' on 2017-10-03 at f129d4b0e9)

330 - + refs: pass NULL to resolve_refdup() if hash is not needed

331 - + refs: pass NULL to refs_resolve_refdup() if hash is not needed

248 + Test tweak.

249

333 - Code clean-up.

250

335 - Will merge to 'master'.

251 +* tb/complete-describe (2017-10-07) 1 commit

252 + (merged to 'next' on 2017-10-10 at 2607253946)

253 + + completion: add --broken and --dirty to describe

254

255 + Docfix.

256

338 -* ar/request-pull-phrasofix (2017-10-03) 1 commit

339 - (merged to 'next' on 2017-10-03 at 07807bba90)

340 - + request-pull: capitalise "Git" to make it a proper noun

257

342 - Spell the name of our system as "Git" in the output from

343 - request-pull script.

258 +* tb/delimit-pretty-trailers-args-with-comma (2017-10-02) 1 commit

259 + (merged to 'next' on 2017-10-04 at de5446a01c)

260 + + pretty.c: delimit "%(trailers)" arguments with ","

261 + (this branch is used by tb/show-trailers-in-ref-filter.)

262

345 - Will merge to 'master'.

263 + The feature that allows --pretty='%(trailers)' to take modifiers

264 + like "fold" and "only" used to separate these modifiers with a

265 + comma, i.e. "%(trailers:fold:only)", but we changed our mind and

266 + use a comma, i.e. "%(trailers:fold,only)". Fast track this change

267 + before this new feature becomes part of any official release.

268

269

348 -* tb/show-trailers-in-ref-filter (2017-10-02) 7 commits

349 - - ref-filter.c: parse trailers arguments with %(contents) atom

350 - - ref-filter.c: use trailer_opts to format trailers

351 - - t6300: refactor %(trailers) tests

352 - - doc: use "`<literal>`"-style quoting for literal strings

353 - - doc: 'trailers' is the preferred way to format trailers

354 - - t4205: unfold across multiple lines

355 - - pretty.c: delimit "%(trailers)" arguments with ","

270 +* tb/ref-filter-empty-modifier (2017-10-05) 1 commit

271 + (merged to 'next' on 2017-10-05 at 62cb67688a)

272 + + ref-filter.c: pass empty-string as NULL to atom parsers

273

357 - "git for-each-ref --format=..." learned a new format element,

358 - %(trailers), to show only the commit log trailer part of the log

359 - message.

274 + In the "--format=..." option of the "git for-each-ref" command (and

275 + its friends, i.e. the listing mode of "git branch/tag"), "%(atom:)"

276 + (e.g. "%(refname:)", "%(body:)" used to error out. Instead, treat

277 + them as if the colon and an empty string that follows it were not

278 + there.

279

361 - Will merge to 'next'.

280

281 +* tb/show-trailers-in-ref-filter (2017-10-02) 6 commits

282 + (merged to 'next' on 2017-10-05 at c7a697f15d)

283 + + ref-filter.c: parse trailers arguments with %(contents) atom

284 + + ref-filter.c: use trailer_opts to format trailers

285 + + t6300: refactor %(trailers) tests

286 + + doc: use "`<literal>`"-style quoting for literal strings

287 + + doc: 'trailers' is the preferred way to format trailers

288 + + t4205: unfold across multiple lines

289

364 -* jk/ui-color-always-to-auto (2017-10-04) 3 commits

365 - - Merge branch 'jk/ui-color-always-to-auto-maint' into jk/ui-color-always-to-auto

366 - - t7301: use test_terminal to check color

367 - - t4015: use --color with --color-moved

368 - (this branch uses jk/ui-color-always-to-auto-maint.)

290 + "git for-each-ref --format=..." learned a new format element,

291 + %(trailers), to show only the commit log trailer part of the log

292 + message.

293

294

371 -* jk/ui-color-always-to-auto-maint (2017-10-04) 10 commits

372 - - color: make "always" the same as "auto" in config

373 - - provide --color option for all ref-filter users

374 - - t3205: use --color instead of color.branch=always

375 - - t3203: drop "always" color test

376 - - t6006: drop "always" color config tests

377 - - t7502: use diff.noprefix for --verbose test

378 - - t7508: use test_terminal for color output

379 - - t3701: use test-terminal to collect color output

380 - - t4015: prefer --color to -c color.diff=always

381 - - test-terminal: set TERM=vt100

382 - (this branch is used by jk/ui-color-always-to-auto.)

295 +* tg/memfixes (2017-10-04) 3 commits

296 + (merged to 'next' on 2017-10-04 at fb338a5531)

297 + + sub-process: use child_process.args instead of child_process.argv

298 + + http-push: fix construction of hex value from path

299 + + path.c: fix uninitialized memory access

300

301 + Fixes for a handful memory access issues identified by valgrind.

302

385 -* jn/strbuf-doc-re-reuse (2017-10-04) 1 commit

386 - - strbuf doc: reuse after strbuf_release is fine

303 +--------------------------------------------------

304 +[New Topics]

305

388 - Will merge to 'next'.

306 +* dm/run-command-ignored-hook-advise (2017-10-10) 1 commit

307 + (merged to 'next' on 2017-10-10 at 0827814922)

308 + + run-command: add hint when a hook is ignored

309

310 + A hook script that is set unexecutable is simply ignored. Git

311 + notifies when such a file is ignored, unless the message is

312 + squelched via advice.ignoredHook configuration.

313

391 -* ks/branch-tweak-error-message-for-extra-args (2017-10-04) 1 commit

392 - - branch: change the error messages to be more meaningful

314 + Will cook in 'next'.

315

394 - Error message tweak.

316

317 +* hv/fetch-moved-submodules-on-demand (2017-10-08) 5 commits

318 + - DONTMERGE: mark failing tests for now

319 + - submodule: simplify decision tree whether to or not to fetch

320 + - implement fetching of moved submodules

321 + - change submodule push test to use proper repository setup

322 + - fetch: add test to make sure we stay backwards compatible

323

397 -* ks/verify-filename-non-option-error-message-tweak (2017-10-04) 1 commit

398 - - setup: update error message to be more meaningful

324 + WIP.

325

400 - Error message tweak.

326

327 +* js/for-each-ref-remote-name-and-ref (2017-10-11) 3 commits

328 + - for-each-ref: test :remotename and :remoteref

329 + - for-each-ref: let upstream/push optionally report the remote ref name

330 + - for-each-ref: let upstream/push optionally report the remote name

331

403 -* rs/qsort-s (2017-10-04) 1 commit

404 - - test-stringlist: avoid buffer underrun when sorting nothing

332 + The "--format=..." option "git for-each-ref" takes learned to show

333 + the name of the 'remote' repository and the ref at the remote side

334 + that is affected for 'upstream' and 'push' via "%(push:remotename)"

335 + and friends.

336

406 - Will merge to 'next'.

337

338 +* pb/bisect-helper (2017-10-06) 6 commits

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

340 + - t6030: explicitly test for bisection cleanup

341 + - bisect--helper: `bisect_clean_state` shell function in C

342 + - bisect--helper: `write_terms` shell function in C

343 + - bisect--helper: rewrite `check_term_format` shell function in C

344 + - bisect--helper: use OPT_CMDMODE instead of OPT_BOOL

345

409 -* sb/branch-avoid-repeated-strbuf-release (2017-10-04) 1 commit

410 - - branch: reset instead of release a strbuf

346 + An early part of piece-by-piece rewrite of "git bisect".

347

348 Will merge to 'next'.

349

350

415 -* tg/memfixes (2017-10-04) 3 commits

416 - - sub-process: use child_process.args instead of child_process.argv

417 - - http-push: fix construction of hex value from path

418 - - path.c: fix uninitialized memory access

419 -

420 - Fixes for a handful memory access issues identified by valgrind.

351 +* bc/object-id (2017-10-09) 24 commits

352 + - refs/files-backend: convert static functions to object_id

353 + - refs: convert read_raw_ref backends to struct object_id

354 + - refs: convert peel_object to struct object_id

355 + - refs: convert resolve_ref_unsafe to struct object_id

356 + - worktree: convert struct worktree to object_id

357 + - refs: convert resolve_gitlink_ref to struct object_id

358 + - resolve_gitlink_ref(): convert remaining callers to object_id

359 + - sha1_file: convert index_path and index_fd to struct object_id

360 + - refs: convert reflog_expire parameter to struct object_id

361 + - refs: convert read_ref_at to struct object_id

362 + - refs: convert peel_ref to struct object_id

363 + - builtin/pack-objects: convert to struct object_id

364 + - pack-bitmap: convert traverse_bitmap_commit_list to object_id

365 + - refs: convert dwim_log to struct object_id

366 + - builtin/reflog: convert remaining unsigned char uses to object_id

367 + - refs: convert dwim_ref and expand_ref to struct object_id

368 + - refs: convert read_ref and read_ref_full to object_id

369 + - refs: convert resolve_refdup and refs_resolve_refdup to struct object_id

370 + - check_connected(): use struct object_id

371 + - refs: update ref transactions to use struct object_id

372 + - refs: convert update_ref and refs_update_ref to use struct object_id

373 + - refs: convert delete_ref and refs_delete_ref to struct object_id

374 + - refs/files-backend: convert struct ref_to_prune to object_id

375 + - walker: convert to struct object_id

376 +

377 +

378 +* jc/no-cmd-as-subroutine (2017-10-11) 2 commits

379 + - merge-ours: do not use cmd_*() as a subroutine

380 + - describe: do not use cmd_*() as a subroutine

381 +

382 + Calling cmd_foo() as if it is a general purpose helper function is

383 + a no-no. Correct two instances of such to set an example.

384 +

385 +

386 +* jc/doc-checkout (2017-10-11) 1 commit

387 + - checkout doc: clarify command line args for "checkout paths" mode

388 +

389 + Doc update.

390

391 Will merge to 'next'.

392

@@ -478,43 +447,72 @@ of the repositories listed at

447 --------------------------------------------------

448 [Cooking]

449

481 -* er/fast-import-dump-refs-on-checkpoint (2017-09-29) 1 commit

482 - (merged to 'next' on 2017-10-03 at 4e7b0e7ec1)

483 - + fast-import: checkpoint: dump branches/tags/marks even if object_count==0

450 +* jc/merge-symlink-ours-theirs (2017-09-26) 1 commit

451 + - merge: teach -Xours/-Xtheirs to symbolic link merge

452

485 - The checkpoint command "git fast-import" did not flush updates to

486 - refs and marks unless at least one object was created since the

487 - last checkpoint, which has been corrected, as these things can

488 - happen without any new object getting created.

453 + "git merge -Xours/-Xtheirs" learned to use our/their version when

454 + resolving a conflicting updates to a symbolic link.

455

490 - Will merge to 'master'.

456 + Needs review.

457

458

493 -* jr/hash-migration-plan-doc (2017-09-28) 1 commit

494 - - technical doc: add a design doc for hash function transition

459 +* jt/partial-clone-lazy-fetch (2017-10-02) 18 commits

460 + - fetch-pack: restore save_commit_buffer after use

461 + - unpack-trees: batch fetching of missing blobs

462 + - clone: configure blobmaxbytes in created repos

463 + - clone: support excluding large blobs

464 + - fetch: support excluding large blobs

465 + - fetch: refactor calculation of remote list

466 + - fetch-pack: support excluding large blobs

467 + - pack-objects: support --blob-max-bytes

468 + - pack-objects: rename want_.* to ignore_.*

469 + - gc: do not repack promisor packfiles

470 + - rev-list: support termination at promisor objects

471 + - sha1_file: support lazily fetching missing objects

472 + - introduce fetch-object: fetch one promisor object

473 + - index-pack: refactor writing of .keep files

474 + - fsck: support promisor objects as CLI argument

475 + - fsck: support referenced promisor objects

476 + - fsck: support refs pointing to promisor objects

477 + - fsck: introduce partialclone extension

478

496 - Lay out plans for weaning us off of SHA-1.

479 + A journey for "git clone" and "git fetch" to become "lazier" by

480 + depending more on its remote repository---this is the beginning of

481 + it.

482

498 - Will merge to 'next'.

483 + Needs review.

484

485

501 -* ot/mru-on-list (2017-10-01) 1 commit

502 - - mru: use double-linked list from list.h

486 +* ma/lockfile-fixes (2017-10-07) 12 commits

487 + - read_cache: roll back lock in `update_index_if_able()`

488 + - read-cache: leave lock in right state in `write_locked_index()`

489 + - read-cache: drop explicit `CLOSE_LOCK`-flag

490 + - cache.h: document `write_locked_index()`

491 + - apply: remove `newfd` from `struct apply_state`

492 + - apply: move lockfile into `apply_state`

493 + - cache-tree: simplify locking logic

494 + - checkout-index: simplify locking logic

495 + - tempfile: fix documentation on `delete_tempfile()`

496 + - lockfile: fix documentation on `close_lock_file_gently()`

497 + - treewide: prefer lockfiles on the stack

498 + - sha1_file: do not leak `lock_file`

499

504 - The first step to getting rid of mru API and using the

505 - doubly-linked list API directly instead.

500 + An earlier update made it possible to use an on-stack in-core

501 + lockfile structure (as opposed to having to deliberately leak an

502 + on-heap one). Many codepaths have been updated to take advantage

503 + of this new facility.

504

507 - Needs review.

508 - Outreachy.

505 + Looking good. Is this ready for 'next'?

506

507

511 -* rb/compat-poll-fix (2017-09-29) 1 commit

512 - - poll.c: always set revents, even if to zero

508 +* ot/mru-on-list (2017-10-01) 1 commit

509 + (merged to 'next' on 2017-10-07 at e6ad4a16a2)

510 + + mru: use double-linked list from list.h

511

514 - Backports a moral equivalent of 2015 fix to the poll emulation from

515 - the upstream gnulib to fix occasional breakages on HPE NonStop.

512 + The first step to getting rid of mru API and using the

513 + doubly-linked list API directly instead.

514

517 - Needs review.

515 + Will cook in 'next'.

516

517

518 * jm/svn-pushmergeinfo-fix (2017-09-17) 1 commit

@@ -555,12 +553,16 @@ of the repositories listed at

553 What's the doneness of this one?

554

555

558 -* ds/find-unique-abbrev-optim (2017-10-03) 5 commits

556 +* ds/find-unique-abbrev-optim (2017-10-09) 4 commits

557 - sha1_name: minimize OID comparisons during disambiguation

558 - sha1_name: parse less while finding common prefix

559 - sha1_name: unroll len loop in find_unique_abbrev_r()

562 - - p0008-abbrev.sh: test find_unique_abbrev() perf

563 - - test-list-objects: list a subset of object ids

560 + - p4211-line-log.sh: add log --online --raw --parents perf test

561 +

562 + Optimize the code to find shortest unique prefix of object names.

563 +

564 + Almost there.

565 + cf. <20171009134933.vmba67adelqbkx4y@sigill.intra.peff.net>

566

567

568 * cc/perf-run-config (2017-09-24) 9 commits

@@ -589,8 +591,11 @@ of the repositories listed at

591 Waiting for review to conclude.

592

593

592 -* bp/fsmonitor (2017-10-04) 13 commits

593 - - fsmonitor: MINGW support for watchman integration

594 +* bp/fsmonitor (2017-10-05) 15 commits

595 + (merged to 'next' on 2017-10-05 at 964a029d13)

596 + + fsmonitor: preserve utf8 filenames in fsmonitor-watchman log

597 + + fsmonitor: read entirety of watchman output

598 + + fsmonitor: MINGW support for watchman integration

599 (merged to 'next' on 2017-10-02 at cf0c67979c)

600 + fsmonitor: add a performance test

601 + fsmonitor: add a sample integration script for Watchman

@@ -608,6 +613,8 @@ of the repositories listed at

613 We learned to talk to watchman to speed up "git status" and other

614 operations that need to see which paths have been modified.

615

616 + Will cook in 'next'.

617 +

618

619 * bw/protocol-v1 (2017-10-04) 10 commits

620 - ssh: introduce a 'simple' ssh variant

@@ -625,8 +632,14 @@ of the repositories listed at

632 and demonstrated that it works with the older versions of Git

633 without harming them.

634

635 + Expecting a reroll.

636 + cf. <20171003214206.GY19555@aiede.mtv.corp.google.com>

637 + The tip one needs to clarify that it is not just '-o' but also '-p'

638 + option in the documentation. Other than that, the approach taken

639 + by the series basically seems sane and well received.

640 +

641

629 -* pc/submodule-helper (2017-10-02) 3 commits

642 +* pc/submodule-helper (2017-10-07) 3 commits

643 - submodule: port submodule subcommand 'status' from shell to C

644 - submodule--helper: introduce for_each_listed_submodule()

645 - submodule--helper: introduce get_submodule_displaypath()

@@ -635,7 +648,8 @@ of the repositories listed at

648 Is this ready for 'next'?

649

650

638 -* bc/hash-algo (2017-08-20) 5 commits

651 +* bc/hash-algo (2017-10-04) 6 commits

652 + - fixup! hash-algo: integrate hash algorithm support with repo setup

653 - hash-algo: switch empty tree and blob lookups to use hash abstraction

654 - hash-algo: integrate hash algorithm support with repo setup

655 - hash-algo: add structure representing hash algorithm

@@ -690,23 +704,6 @@ of the repositories listed at

704 cf. <f59c8256-716b-9305-2a4f-d4fe49f666ff@web.de>

705

706

693 -* jc/allow-lazy-cas (2017-07-06) 1 commit

694 - - push: disable lazy --force-with-lease by default

695 -

696 - Because "git push --force-with-lease[=<ref>]" that relies on the

697 - stability of remote-tracking branches is unsafe when something

698 - fetches into the repository behind user's back, it is now disabled

699 - by default. A new configuration variable can be used to enable it

700 - by users who know what they are doing. This would pave the way to

701 - possibly turn `--force` into `--force-with-lease`.

702 -

703 - Will discard.

704 - Independent from disabling the feature by default, this stirred up

705 - a discussion to replace the DWIM heuristics with a better one, which

706 - deserves to be its own topic.

707 - cf. <alpine.DEB.2.21.1.1707131435220.4193@virtualbox>

708 -

709 -

707 * ex/deprecate-empty-pathspec-as-match-all (2017-06-23) 2 commits

708 (merged to 'next' on 2017-08-14 at 24db08a6e8)

709 + pathspec: die on empty strings as pathspec

@@ -724,7 +721,6 @@ of the repositories listed at

721 As the draft RelNotes says, let's merge this to 'master' after

722 this release cycle.

723

727 -

724 --------------------------------------------------

725 [Discarded]

726

@@ -742,3 +738,18 @@ of the repositories listed at

738 one) until we come up with a different fix to the problem.

739

740

741 +* jc/allow-lazy-cas (2017-07-06) 1 commit

742 + . push: disable lazy --force-with-lease by default

743 +

744 + Because "git push --force-with-lease[=<ref>]" that relies on the

745 + stability of remote-tracking branches is unsafe when something

746 + fetches into the repository behind user's back, it is now disabled

747 + by default. A new configuration variable can be used to enable it

748 + by users who know what they are doing. This would pave the way to

749 + possibly turn `--force` into `--force-with-lease`.

750 +

751 + Will discard.

752 + Independent from disabling the feature by default, this stirred up

753 + a discussion to replace the DWIM heuristics with a better one, which

754 + deserves to be its own topic.

755 + cf. <alpine.DEB.2.21.1.1707131435220.4193@virtualbox>

What's cooking (2017/10 #02)