5a6f4274a8 · 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 2016, #01; Mon, 3)

4 -X-master-at: 0cf36115dce7438a0eafad54a81cc57175e8fb54

5 -X-next-at: 9af6bb63e9b9efbcd0064516a7811092873d1ddf

3 +Subject: What's cooking in git.git (Oct 2016, #02; Thu, 6)

4 +X-master-at: a23ca1b8dc42ffd4de2ef30d67ce1e21ded29886

5 +X-next-at: d4ef1078455682bd5fab0e8b62e2351d25268477

6

7 -What's cooking in git.git (Oct 2016, #01; Mon, 3)

7 +What's cooking in git.git (Oct 2016, #02; Thu, 6)

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

9

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

@@ -26,193 +26,6 @@ of the repositories listed at

26 --------------------------------------------------

27 [Graduated to "master"]

28

29 -* dt/mailinfo (2016-09-26) 1 commit

30 - (merged to 'next' on 2016-09-27 at 59e95dbc0e)

31 - + add David Turner's Two Sigma address

32 -

33 -

34 -* dt/tree-fsck (2016-09-27) 2 commits

35 - (merged to 'next' on 2016-09-28 at afdfdbbf37)

36 - + fsck: handle bad trees like other errors

37 - + tree-walk: be more specific about corrupt tree errors

38 -

39 - The codepath in "git fsck" to detect malformed tree objects has

40 - been updated not to die but keep going after detecting them.

41 -

42 -

43 -* ik/gitweb-force-highlight (2016-09-25) 2 commits

44 - (merged to 'next' on 2016-09-27 at cbb8391a76)

45 - + gitweb: use highlight's shebang detection

46 - + gitweb: remove unused guess_file_syntax() parameter

47 -

48 - "gitweb" can spawn "highlight" to show blob contents with

49 - (programming) language-specific syntax highlighting, but only

50 - when the language is known. "highlight" can however be told

51 - to make the guess itself by giving it "--force" option, which

52 - has been enabled.

53 -

54 -

55 -* jc/verify-loose-object-header (2016-09-26) 2 commits

56 - (merged to 'next' on 2016-09-27 at 2947f95f14)

57 - + unpack_sha1_header(): detect malformed object header

58 - + streaming: make sure to notice corrupt object

59 -

60 - Codepaths that read from an on-disk loose object were too loose in

61 - validating what they are reading is a proper object file and

62 - sometimes read past the data they read from the disk, which has

63 - been corrected. H/t to Gustavo Grieco for reporting.

64 -

65 -

66 -* jc/worktree-config (2016-09-27) 1 commit

67 - (merged to 'next' on 2016-09-28 at 0c262f6)

68 - + worktree: honor configuration variables

69 -

70 - "git worktree", even though it used the default_abbrev setting that

71 - ought to be affected by core.abbrev configuration variable, ignored

72 - the variable setting. The command has been taught to read the

73 - default set of configuration variables to correct this.

74 -

75 -

76 -* jk/ident-ai-canonname-could-be-null (2016-09-23) 1 commit

77 - (merged to 'next' on 2016-09-26 at 0eefb29)

78 - + ident: handle NULL ai_canonname

79 -

80 - In the codepath that comes up with the hostname to be used in an

81 - e-mail when the user didn't tell us, we looked at ai_canonname

82 - field in struct addrinfo without making sure it is not NULL first.

83 -

84 -

85 -* jk/verify-packfile-gently (2016-09-22) 1 commit

86 - (merged to 'next' on 2016-09-26 at f5abba5)

87 - + verify_packfile: check pack validity before accessing data

88 -

89 - A low-level function verify_packfile() was meant to show errors

90 - that were detected without dying itself, but under some conditions

91 - it didn't and died instead, which has been fixed.

92 -

93 -

94 -* jt/fetch-pack-in-vain-count-with-stateless (2016-09-23) 1 commit

95 - (merged to 'next' on 2016-09-26 at 9645629)

96 - + fetch-pack: do not reset in_vain on non-novel acks

97 -

98 - When "git fetch" tries to find where the history of the repository

99 - it runs in has diverged from what the other side has, it has a

100 - mechanism to avoid digging too deep into irrelevant side branches.

101 - This however did not work well over the "smart-http" transport due

102 - to a design bug, which has been fixed.

103 -

104 -

105 -* jt/mailinfo-fold-in-body-headers (2016-09-21) 3 commits

106 - (merged to 'next' on 2016-09-26 at 4235eb6)

107 - + mailinfo: handle in-body header continuations

108 - + mailinfo: make is_scissors_line take plain char *

109 - + mailinfo: separate in-body header processing

110 -

111 - When "git format-patch --stdout" output is placed as an in-body

112 - header and it uses the RFC2822 header folding, "git am" failed to

113 - put the header line back into a single logical line. The

114 - underlying "git mailinfo" was taught to handle this properly.

115 -

116 -

117 -* kd/mailinfo-quoted-string (2016-09-28) 2 commits

118 - (merged to 'next' on 2016-09-28 at 2aaeb57804)

119 - + mailinfo: unescape quoted-pair in header fields

120 - + t5100-mailinfo: replace common path prefix with variable

121 -

122 - An author name, that spelled a backslash-quoted double quote in the

123 - human readable part "My \"double quoted\" name", was not unquoted

124 - correctly while applying a patch from a piece of e-mail.

125 -

126 -

127 -* mh/diff-indent-heuristic (2016-09-27) 1 commit

128 - (merged to 'next' on 2016-09-27 at 3d6fb6605a)

129 - + xdiff: rename "struct group" to "struct xdlgroup"

130 -

131 - Clean-up for a recently graduated topic.

132 -

133 -

134 -* nd/init-core-worktree-in-multi-worktree-world (2016-09-25) 5 commits

135 - (merged to 'next' on 2016-09-27 at 619f7f3a3b)

136 - + init: kill git_link variable

137 - + init: do not set unnecessary core.worktree

138 - + init: kill set_git_dir_init()

139 - + init: call set_git_dir_init() from within init_db()

140 - + init: correct re-initialization from a linked worktree

141 -

142 - "git init" tried to record core.worktree in the repository's

143 - 'config' file when GIT_WORK_TREE environment variable was set and

144 - it was different from where GIT_DIR appears as ".git" at its top,

145 - but the logic was faulty when .git is a "gitdir:" file that points

146 - at the real place, causing trouble in working trees that are

147 - managed by "git worktree". This has been corrected.

148 -

149 -

150 -* pb/rev-list-reverse-with-count (2016-09-27) 1 commit

151 - (merged to 'next' on 2016-09-28 at 2905d0adbc)

152 - + rev-list-options: clarify the usage of --reverse

153 -

154 - Doc update to clarify what "log -3 --reverse" does.

155 -

156 -

157 -* rs/copy-array (2016-09-25) 2 commits

158 - (merged to 'next' on 2016-09-27 at c92e020669)

159 - + use COPY_ARRAY

160 - + add COPY_ARRAY

161 -

162 - Code cleanup.

163 -

164 -

165 -* rs/git-gui-use-modern-git-merge-syntax (2016-09-26) 2 commits

166 - (merged to 'next' on 2016-09-27 at f55850df7d)

167 - + Merge branch 'rs/use-modern-git-merge-syntax' of git-gui into rs/git-gui-use-modern-git-merge-syntax

168 - + git-gui: stop using deprecated merge syntax

169 -

170 - The original command line syntax for "git merge", which was "git

171 - merge <msg> HEAD <parent>...", has been deprecated for quite some

172 - time, and "git gui" was the last in-tree user of the syntax. This

173 - is finally fixed, so that we can move forward with the deprecation.

174 -

175 -

176 -* va/git-gui-i18n (2016-09-26) 3 commits

177 - (merged to 'next' on 2016-09-27 at ab0f66ff8a)

178 - + Merge branch 'va/i18n' of ../git-gui into va/git-gui-i18n

179 - + git-gui: l10n: add Portuguese translation

180 - + git-gui i18n: mark strings for translation

181 -

182 - "git gui" l10n to Portuguese.

183 -

184 ---------------------------------------------------

185 -[New Topics]

186 -

187 -* rs/cocci (2016-10-03) 4 commits

188 - (merged to 'next' on 2016-10-03 at 758cc6de9c)

189 - + coccicheck: make transformation for strbuf_addf(sb, "...") more precise

190 - (merged to 'next' on 2016-09-28 at 26462645f9)

191 - + use strbuf_add_unique_abbrev() for adding short hashes, part 2

192 - + use strbuf_addstr() instead of strbuf_addf() with "%s", part 2

193 - + gitignore: ignore output files of coccicheck make target

194 -

195 - Code clean-up with help from coccinelle tool continues.

196 -

197 - Will merge to 'master'.

198 -

199 -

200 -* nd/ita-empty-commit (2016-09-28) 3 commits

201 - - commit: don't be fooled by ita entries when creating initial commit

202 - - diff-lib.c: enable --shift-ita in index_differs_from()

203 - - Resurrect "diff-lib.c: adjust position of i-t-a entries in diff"

204 -

205 - When new paths were added by "git add -N" to the index, it was

206 - enough to circumvent the check by "git commit" to refrain from

207 - making an empty commit without "--allow-empty". The same logic

208 - prevented "git status" to show such a path as "new file" in the

209 - "Changes not staged for commit" section.

210 -

211 - Expecting a reroll.

212 - cf. <xmqqzimrj03j.fsf@gitster.mtv.corp.google.com>

213 - cf. <xmqq8tubkgg5.fsf@gitster.mtv.corp.google.com>

214 -

215 -

29 * jc/blame-abbrev (2016-09-28) 1 commit

30 (merged to 'next' on 2016-10-03 at 8ec86ff1e1)

31 + blame: use DEFAULT_ABBREV macro

@@ -221,37 +34,29 @@ of the repositories listed at

34 setting for the abbreviation, but "git blame" peeked into

35 underlying variable bypassing the macro for no good reason.

36

224 - Will merge to 'master'.

225 -

226 -

227 -* lt/abbrev-auto (2016-10-03) 3 commits

228 - (merged to 'next' on 2016-10-03 at bb188d00f7)

229 - + abbrev: auto size the default abbreviation

230 - + abbrev: prepare for new world order

231 - + abbrev: add FALLBACK_DEFAULT_ABBREV to prepare for auto sizing

232 - (this branch uses jk/ambiguous-short-object-names.)

233 -

234 - Allow the default abbreviation length, which has historically been

235 - 7, to scale as the repository grows. The logic suggests to use 12

236 - hexdigits for the Linux kernel, and 9 to 10 for Git itself.

237 -

238 - Will merge to 'master'.

239 -

240 -

241 -* dt/http-empty-auth (2016-10-03) 1 commit

242 - - http: http.emptyauth should allow empty (not just NULL) usernames

243 -

244 - http.emptyauth configuration is a way to allow an empty username to

245 - pass when attempting to authenticate using mechanisms like

246 - Kerberos. We took an unspecified (NULL) username and sent ":"

247 - (i.e. no username, no password) to CURLOPT_USERPWD, but did not do

248 - the same when the username is explicitly set to an empty string.

249 -

250 - cf. <20161003210100.t5nqknwfotag3lmj@sigill.intra.peff.net>

37

38 +* jk/ambiguous-short-object-names (2016-09-27) 11 commits

39 + (merged to 'next' on 2016-09-28 at 1b85295323)

40 + + get_short_sha1: make default disambiguation configurable

41 + + get_short_sha1: list ambiguous objects on error

42 + + for_each_abbrev: drop duplicate objects

43 + + sha1_array: let callbacks interrupt iteration

44 + + get_short_sha1: mark ambiguity error for translation

45 + + get_short_sha1: NUL-terminate hex prefix

46 + + get_short_sha1: refactor init of disambiguation code

47 + + get_short_sha1: parse tags when looking for treeish

48 + + get_sha1: propagate flags to child functions

49 + + get_sha1: avoid repeating ourselves via ONLY_TO_DIE

50 + + get_sha1: detect buggy calls with multiple disambiguators

51 + (this branch is used by jk/abbrev-auto and lt/abbrev-auto.)

52

253 -* jc/diff-unique-abbrev-comments (2016-09-30) 1 commit

254 - - diff_unique_abbrev(): document its assumption and limitation

53 + When given an abbreviated object name that is not (or more

54 + realistically, "no longer") unique, we gave a fatal error

55 + "ambiguous argument". This error is now accompanied by hints that

56 + lists the objects that begins with the given prefix. During the

57 + course of development of this new feature, numerous minor bugs were

58 + uncovered and corrected, the most notable one of which is that we

59 + gave "short SHA1 xxxx is ambiguous." twice without good reason.

60

61

62 * jk/graph-padding-fix (2016-09-29) 1 commit

@@ -263,28 +68,6 @@ of the repositories listed at

68 has been fixed; this did not affect any existing code as nobody

69 tried to write anything after the padding on such a line, though.

70

266 - Will merge to 'master'.

267 -

268 -

269 -* jk/quarantine-received-objects (2016-10-03) 5 commits

270 - - tmp-objdir: do not migrate files starting with '.'

271 - - tmp-objdir: put quarantine information in the environment

272 - - receive-pack: quarantine objects until pre-receive accepts

273 - - tmp-objdir: introduce API for temporary object directories

274 - - check_connected: accept an env argument

275 - (this branch uses jk/alt-odb-cleanup.)

276 -

277 - In order for the receiving end of "git push" to inspect the

278 - received history and decide to reject the push, the objects sent

279 - from the sending end need to be made available to the hook and

280 - the mechanism for the connectivity check, and this was done

281 - traditionally by storing the objects in the receiving repository

282 - and letting "git gc" to expire it. Instead, store the newly

283 - received objects in a temporary area, and make them available by

284 - reusing the alternate object store mechanism to them only while we

285 - decide if we accept the check, and once we decide, either migrate

286 - them to the repository or purge them immediately.

287 -

71

72 * ps/http-gssapi-cred-delegation (2016-09-29) 1 commit

73 (merged to 'next' on 2016-10-03 at 310fbe8f24)

@@ -294,8 +77,6 @@ of the repositories listed at

77 disabled by default due to CVE-2011-2192; introduce a configuration

78 to selectively allow enabling this.

79

297 - Will merge to 'master'.

298 -

80

81 * rs/c-auto-resets-attributes (2016-09-29) 1 commit

82 (merged to 'next' on 2016-10-03 at 6a0b946a79)

@@ -303,24 +84,18 @@ of the repositories listed at

84

85 When "%C(auto)" appears at the very beginning of the pretty format

86 string, it did not need to issue the reset sequence, but it did.

306 -

307 - Will merge to 'master'.

87 This is a small optimization to already graduated topic.

88

89

311 -* rs/qsort (2016-10-03) 6 commits

312 - - show-branch: use QSORT

313 - - use QSORT, part 2

314 - - coccicheck: use --all-includes by default

315 - - remove unnecessary check before QSORT

316 - - use QSORT

317 - - add QSORT

318 -

319 - We call "qsort(array, nelem, sizeof(array[0]), fn)", and most of

320 - the time third parameter is redundant. A new QSORT() macro lets us

321 - omit it.

90 +* rs/cocci (2016-10-03) 4 commits

91 + (merged to 'next' on 2016-10-03 at 758cc6de9c)

92 + + coccicheck: make transformation for strbuf_addf(sb, "...") more precise

93 + (merged to 'next' on 2016-09-28 at 26462645f9)

94 + + use strbuf_add_unique_abbrev() for adding short hashes, part 2

95 + + use strbuf_addstr() instead of strbuf_addf() with "%s", part 2

96 + + gitignore: ignore output files of coccicheck make target

97

323 - Will merge to 'next'.

98 + Code clean-up with help from coccinelle tool continues.

99

100

101 * sg/ref-filter-parse-optim (2016-10-03) 1 commit

@@ -330,29 +105,62 @@ of the repositories listed at

105 The code that parses the format parameter of for-each-ref command

106 has seen a micro-optimization.

107

333 - Will merge to 'master'.

108

109 +* vn/revision-shorthand-for-side-branch-log (2016-09-27) 1 commit

110 + (merged to 'next' on 2016-09-28 at c1237b24f6)

111 + + revision: new rev^-n shorthand for rev^n..rev

112

336 -* jk/alt-odb-cleanup (2016-10-03) 18 commits

337 - - alternates: use fspathcmp to detect duplicates

338 - - sha1_file: always allow relative paths to alternates

339 - - count-objects: report alternates via verbose mode

340 - - fill_sha1_file: write into a strbuf

341 - - alternates: store scratch buffer as strbuf

342 - - fill_sha1_file: write "boring" characters

343 - - alternates: use a separate scratch space

344 - - alternates: encapsulate alt->base munging

345 - - alternates: provide helper for allocating alternate

346 - - alternates: provide helper for adding to alternates list

347 - - link_alt_odb_entry: refactor string handling

348 - - link_alt_odb_entry: handle normalize_path errors

349 - - t5613: clarify "too deep" recursion tests

350 - - t5613: do not chdir in main process

351 - - t5613: whitespace/style cleanups

352 - - t5613: use test_must_fail

353 - - t5613: drop test_valid_repo function

354 - - t5613: drop reachable_via function

355 - (this branch is used by jk/quarantine-received-objects.)

113 + "git log rev^..rev" is an often-used revision range specification

114 + to show what was done on a side branch merged at rev. This has

115 + gained a short-hand "rev^-1". In general "rev^-$n" is the same as

116 + "^rev^$n rev", i.e. what has happened on other branches while the

117 + history leading to nth parent was looking the other way.

118 +

119 +--------------------------------------------------

120 +[New Topics]

121 +

122 +* jc/ws-error-highlight (2016-10-04) 4 commits

123 + - diff: introduce diff.wsErrorHighlight option

124 + - diff.c: move ws-error-highlight parsing helpers up

125 + - diff.c: refactor parse_ws_error_highlight()

126 + - t4015: split out the "setup" part of ws-error-highlight test

127 +

128 + "git diff/log --ws-error-highlight=<kind>" lacked the corresponding

129 + configuration variable to set it by default.

130 +

131 + Undecided.

132 +

133 +

134 +* jk/abbrev-auto (2016-10-03) 1 commit

135 + - find_unique_abbrev: move logic out of get_short_sha1()

136 + (this branch uses lt/abbrev-auto.)

137 +

138 + Updates the way approximate count of total objects is computed

139 + while attempting to come up with a unique abbreviated object name,

140 + which in turn needs to estimate how many hexdigits are necessary to

141 + ensure uniqueness.

142 +

143 + Undecided.

144 +

145 +

146 +* jk/clone-copy-alternates-fix (2016-10-05) 1 commit

147 + - clone: detect errors in normalize_path_copy

148 +

149 + "git clone" of a local repository can be done at the filesystem

150 + level, but the codepath did not check errors while copying and

151 + adjusting the file that lists alternate object stores.

152 +

153 + Will merge to 'next'.

154 +

155 +

156 +* nd/commit-p-doc (2016-10-05) 1 commit

157 + - git-commit.txt: clarify --patch mode with pathspec

158 +

159 + Documentation for "git commit" was updated to clarify that "commit

160 + -p <paths>" adds to the current contents of the index to come up

161 + with what to commit.

162 +

163 + Will merge to 'next'.

164

165 --------------------------------------------------

166 [Stalled]

@@ -655,32 +463,115 @@ of the repositories listed at

463 --------------------------------------------------

464 [Cooking]

465

658 -* jk/ambiguous-short-object-names (2016-09-27) 11 commits

659 - (merged to 'next' on 2016-09-28 at 1b85295323)

660 - + get_short_sha1: make default disambiguation configurable

661 - + get_short_sha1: list ambiguous objects on error

662 - + for_each_abbrev: drop duplicate objects

663 - + sha1_array: let callbacks interrupt iteration

664 - + get_short_sha1: mark ambiguity error for translation

665 - + get_short_sha1: NUL-terminate hex prefix

666 - + get_short_sha1: refactor init of disambiguation code

667 - + get_short_sha1: parse tags when looking for treeish

668 - + get_sha1: propagate flags to child functions

669 - + get_sha1: avoid repeating ourselves via ONLY_TO_DIE

670 - + get_sha1: detect buggy calls with multiple disambiguators

671 - (this branch is used by lt/abbrev-auto.)

466 +* nd/ita-empty-commit (2016-09-28) 3 commits

467 + - commit: don't be fooled by ita entries when creating initial commit

468 + - diff-lib.c: enable --shift-ita in index_differs_from()

469 + - Resurrect "diff-lib.c: adjust position of i-t-a entries in diff"

470 +

471 + When new paths were added by "git add -N" to the index, it was

472 + enough to circumvent the check by "git commit" to refrain from

473 + making an empty commit without "--allow-empty". The same logic

474 + prevented "git status" to show such a path as "new file" in the

475 + "Changes not staged for commit" section.

476 +

477 + Expecting a reroll.

478 + cf. <xmqqzimrj03j.fsf@gitster.mtv.corp.google.com>

479 + cf. <xmqq8tubkgg5.fsf@gitster.mtv.corp.google.com>

480

673 - When given an abbreviated object name that is not (or more

674 - realistically, "no longer") unique, we gave a fatal error

675 - "ambiguous argument". This error is now accompanied by hints that

676 - lists the objects that begins with the given prefix. During the

677 - course of development of this new feature, numerous minor bugs were

678 - uncovered and corrected, the most notable one of which is that we

679 - gave "short SHA1 xxxx is ambiguous." twice without good reason.

481 +

482 +* lt/abbrev-auto (2016-10-03) 3 commits

483 + (merged to 'next' on 2016-10-03 at bb188d00f7)

484 + + abbrev: auto size the default abbreviation

485 + + abbrev: prepare for new world order

486 + + abbrev: add FALLBACK_DEFAULT_ABBREV to prepare for auto sizing

487 + (this branch is used by jk/abbrev-auto.)

488 +

489 + Allow the default abbreviation length, which has historically been

490 + 7, to scale as the repository grows. The logic suggests to use 12

491 + hexdigits for the Linux kernel, and 9 to 10 for Git itself.

492 +

493 + Will hold to see if people scream.

494 +

495 +

496 +* dt/http-empty-auth (2016-10-04) 1 commit

497 + - http: http.emptyauth should allow empty (not just NULL) usernames

498 +

499 + http.emptyauth configuration is a way to allow an empty username to

500 + pass when attempting to authenticate using mechanisms like

501 + Kerberos. We took an unspecified (NULL) username and sent ":"

502 + (i.e. no username, no password) to CURLOPT_USERPWD, but did not do

503 + the same when the username is explicitly set to an empty string.

504 +

505 + Will merge to 'next'.

506 +

507 +

508 +* jc/diff-unique-abbrev-comments (2016-09-30) 1 commit

509 + - diff_unique_abbrev(): document its assumption and limitation

510 +

511 +

512 +* jk/quarantine-received-objects (2016-10-05) 5 commits

513 + - tmp-objdir: do not migrate files starting with '.'

514 + - tmp-objdir: put quarantine information in the environment

515 + - receive-pack: quarantine objects until pre-receive accepts

516 + - tmp-objdir: introduce API for temporary object directories

517 + - check_connected: accept an env argument

518 + (this branch uses jk/alt-odb-cleanup.)

519 +

520 + In order for the receiving end of "git push" to inspect the

521 + received history and decide to reject the push, the objects sent

522 + from the sending end need to be made available to the hook and

523 + the mechanism for the connectivity check, and this was done

524 + traditionally by storing the objects in the receiving repository

525 + and letting "git gc" to expire it. Instead, store the newly

526 + received objects in a temporary area, and make them available by

527 + reusing the alternate object store mechanism to them only while we

528 + decide if we accept the check, and once we decide, either migrate

529 + them to the repository or purge them immediately.

530 +

531 + Will merge to 'next' after whipping jk/alt-odb-cleanup into shape.

532 +

533 +

534 +* rs/qsort (2016-10-03) 6 commits

535 + (merged to 'next' on 2016-10-06 at 32a5bd3c88)

536 + + show-branch: use QSORT

537 + + use QSORT, part 2

538 + + coccicheck: use --all-includes by default

539 + + remove unnecessary check before QSORT

540 + + use QSORT

541 + + add QSORT

542 +

543 + We call "qsort(array, nelem, sizeof(array[0]), fn)", and most of

544 + the time third parameter is redundant. A new QSORT() macro lets us

545 + omit it.

546

547 Will merge to 'master'.

548

549

550 +* jk/alt-odb-cleanup (2016-10-05) 19 commits

551 + - alternates: use fspathcmp to detect duplicates

552 + - sha1_file: always allow relative paths to alternates

553 + - count-objects: report alternates via verbose mode

554 + - fill_sha1_file: write into a strbuf

555 + - alternates: store scratch buffer as strbuf

556 + - fill_sha1_file: write "boring" characters

557 + - alternates: use a separate scratch space

558 + - alternates: encapsulate alt->base munging

559 + - alternates: provide helper for allocating alternate

560 + - alternates: provide helper for adding to alternates list

561 + - link_alt_odb_entry: refactor string handling

562 + - link_alt_odb_entry: handle normalize_path errors

563 + - fixup! t5613: clarify "too deep" recursion tests

564 + - t5613: clarify "too deep" recursion tests

565 + - t5613: do not chdir in main process

566 + - t5613: whitespace/style cleanups

567 + - t5613: use test_must_fail

568 + - t5613: drop test_valid_repo function

569 + - t5613: drop reachable_via function

570 + (this branch is used by jk/quarantine-received-objects.)

571 +

572 + Will merge to 'next' after squashing the fixup! in.

573 +

574 +

575 * va/i18n-perl-scripts (2016-09-25) 11 commits

576 - i18n: difftool: mark warnings for translation

577 - i18n: send-email: mark string with interpolation for translation

@@ -696,21 +587,7 @@ of the repositories listed at

587

588 Porcelain scripts written in Perl are getting internationalized.

589

699 - Waiting for a reroll.

700 - cf. <1474913721.1035.9.camel@sapo.pt>

701 -

702 -

703 -* vn/revision-shorthand-for-side-branch-log (2016-09-27) 1 commit

704 - (merged to 'next' on 2016-09-28 at c1237b24f6)

705 - + revision: new rev^-n shorthand for rev^n..rev

706 -

707 - "git log rev^..rev" is an often-used revision range specification

708 - to show what was done on a side branch merged at rev. This has

709 - gained a short-hand "rev^-1". In general "rev^-$n" is the same as

710 - "^rev^$n rev", i.e. what has happened on other branches while the

711 - history leading to nth parent was looking the other way.

712 -

713 - Will merge to 'master'.

590 + Waiting for comments.

591

592

593 * jc/latin-1 (2016-09-26) 2 commits

@@ -725,7 +602,8 @@ of the repositories listed at

602 Will hold to see if people scream.

603

604

728 -* mg/gpg-richer-status (2016-09-28) 1 commit

605 +* mg/gpg-richer-status (2016-10-06) 2 commits

606 + - SQUASH: gpg-interface: use more status letters

607 - gpg-interface: use more status letters

608

609 The GPG verification status shown in "%G?" pretty format specifier

@@ -733,6 +611,8 @@ of the repositories listed at

611 key, a signature made by a revoked key, etc. New output letters

612 have been assigned to express them.

613

614 + Waiting for a response to ping/squash.

615 +

616

617 * jc/blame-reverse (2016-06-14) 2 commits

618 (merged to 'next' on 2016-09-22 at d1a8e9ce99)

@@ -744,13 +624,14 @@ of the repositories listed at

624 in path in an old revision OLD have survived up to the current

625 commit.

626

747 - Will hold to see if it is broken.

627 + Will merge to 'master'.

628

629

750 -* js/libify-require-clean-work-tree (2016-09-12) 5 commits

630 +* js/libify-require-clean-work-tree (2016-10-05) 6 commits

631 + - wt-status: begin error messages with lower-case

632 - wt-status: teach has_{unstaged,uncommitted}_changes() about submodules

752 - - Export also the has_un{staged,committed}_changed() functions

753 - - Make the require_clean_work_tree() function truly reusable

633 + - wt-status: export also the has_un{staged,committed}_changed()

634 + - wt-status: make the require_clean_work_tree() function reusable

635 - pull: make code more similar to the shell script again

636 - pull: drop confusing prefix parameter of die_on_unclean_work_tree()

637

@@ -758,10 +639,7 @@ of the repositories listed at

639 pull" was rewritten from shell; the helper is now made available to

640 other callers in preparation for upcoming "rebase -i" work.

641

761 - Waiting for comments.

762 - Modulo a few minor nits, this looked almost ready.

763 - cf. <xmqqtwdl2bhm.fsf@gitster.mtv.corp.google.com>

764 - cf. <xmqqpoo92bdr.fsf@gitster.mtv.corp.google.com>

642 + Will merge to 'next'.

643

644

645 * bw/ls-files-recurse-submodules (2016-10-03) 4 commits

@@ -780,18 +658,21 @@ of the repositories listed at

658 Looking good. Is this ready for 'next'?

659

660

783 -* ls/filter-process (2016-09-23) 11 commits

661 +* ls/filter-process (2016-10-04) 14 commits

662 + - contrib/long-running-filter: add long running filter example

663 - convert: add filter.<driver>.process option

664 + - convert: prepare filter.<driver>.process option

665 - convert: make apply_filter() adhere to standard Git error handling

786 - - convert: modernize tests

787 - - convert: quote filter names in error messages

666 - pkt-line: add functions to read/write flush terminated packet streams

667 - pkt-line: add packet_write_gently()

668 - pkt-line: add packet_flush_gently()

669 - pkt-line: add packet_write_fmt_gently()

792 - - run-command: move check_pipe() from write_or_die to run_command

670 - pkt-line: extract set_packet_header()

671 - pkt-line: rename packet_write() to packet_write_fmt()

672 + - run-command: add wait_on_exit

673 + - run-command: move check_pipe() from write_or_die to run_command

674 + - convert: modernize tests

675 + - convert: quote filter names in error messages

676

677 The smudge/clean filter API expect an external process is spawned

678 to filter the contents for each path that has a filter defined. A

@@ -800,9 +681,6 @@ of the repositories listed at

681 all filtering need is served by this single process for multiple

682 paths, reducing the process creation overhead.

683

803 - Somehow I thought this was getting ready for 'next' but it seems

804 - at least another round of reroll is coming?

805 -

684

685 * hv/submodule-not-yet-pushed-fix (2016-09-14) 2 commits

686 - serialize collection of refs that contain submodule changes

@@ -844,20 +722,18 @@ of the repositories listed at

722 complete "git cmd ^mas<HT>" to complete the negative end of

723 reference to "git cmd ^master".

724

847 - Will hold to see if it is broken.

725 + Will merge to 'master'.

726

727

850 -* sb/push-make-submodule-check-the-default (2016-08-24) 1 commit

851 - - push: change submodule default to check

728 +* sb/push-make-submodule-check-the-default (2016-10-06) 2 commits

729 + - push: change submodule default to check when submodules exist

730 + - submodule add: extend force flag to add existing repos

731

853 - Turn the default of "push.recurseSubmodules" to "check".

732 + Turn the default of "push.recurseSubmodules" to "check" when

733 + submodules seem to be in use.

734

735 Will hold to wait for hv/submodule-not-yet-pushed-fix

736

857 - This reveals that the "check" mode is too inefficient to use in

858 - real projects, even in ones as small as git itself.

859 - cf. <xmqqh9aaot49.fsf@gitster.mtv.corp.google.com>

860 -

737

738 * ak/curl-imap-send-explicit-scheme (2016-08-17) 1 commit

739 (merged to 'next' on 2016-09-22 at 4449584c26)

@@ -868,7 +744,7 @@ of the repositories listed at

744 imap(s):// before the destination. To some folks, that didn't work

745 and the library tried to make HTTP(s) requests instead.

746

871 - Will hold to see if it is broken.

747 + Will merge to 'master'.

748

749

750 * jk/pack-objects-optim-mru (2016-08-11) 4 commits

@@ -885,7 +761,7 @@ of the repositories listed at

761 the packfiles are now optimized by checking the most-recently-used

762 packfile first.

763

888 - Will hold to see if people scream.

764 + Will merge to 'master'.

765

766

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

@@ -895,7 +771,7 @@ of the repositories listed at

771 The ./configure script generated from configure.ac was taught how

772 to detect support of SSL by libcurl better.

773

898 - Will hold to see if it is broken.

774 + Will merge to 'master'.

775

776

777 * jc/pull-rebase-ff (2016-07-28) 1 commit

@@ -966,7 +842,7 @@ of the repositories listed at

842 "I am interested only in the recent N months worth of history" and

843 "Give me only the history since that version".

844

969 - Will hold to see if it is broken.

845 + Will merge to 'master'.

846

847

848 * jc/merge-drop-old-syntax (2015-04-29) 1 commit

@@ -980,7 +856,7 @@ of the repositories listed at

856 which needs to be fixed before this final step can proceed.

857 cf. <5671DB28.8020901@kdbg.org>

858

983 - Will hold to wait for rs/git-gui-use-modern-git-merge-syntax

859 + Will merge to 'next'.

860

861 --------------------------------------------------

862 [Discarded]

What's cooking (2016/10 #02)