What's cooking (2017/05 #08)
Junio C Hamano committed
May 29, 2017 at 15:19 UTC
b82b4db0bd796299811dfc8b10a0df30f811377e
1 file changed
+583
-599
whats-cooking.txt
+583
-599
@@ -1,10 +1,10 @@
1
To: git@vger.kernel.org
2
Bcc: lwn@lwn.net
3
-Subject: What's cooking in git.git (May 2017, #07; Tue, 23)
4
-X-master-at: 1eb437020a2c098a7c12da4c05082fbea10d98c9
5
-X-next-at: 93fb34087a286d99fc95933e0bb699d6263fbad7
3
+Subject: What's cooking in git.git (May 2017, #08; Mon, 29)
4
+X-master-at: e83352ef23cca2701953ed3c915f1db49b255a7d
5
+X-next-at: 88ad292c901b5a8436ade109286d30c4942be41f
6
7
-What's cooking in git.git (May 2017, #07; Tue, 23)
7
+What's cooking in git.git (May 2017, #08; Mon, 29)
8
--------------------------------------------------
9
10
Here are the topics that have been cooking. Commits prefixed with
@@ -12,9 +12,12 @@ 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 'maint' branch is now for the v2.13.x maintenance track. A
16
-handful of fixes in 'master' will soon be merged to it to prepare
17
-for v2.13.1 in a week or so.
15
+A bit more topics are now in 'master'. One unfortunate thing is
16
+that the SHA1 breakage in 2.13 for big-endian platforms were lost in
17
+the noise with excitement felt by some subset of contributors with
18
+the possible use of submodules. The first step in the series is
19
+neutral to the excitement, and should be fast-tracked down to
20
+'maint' soonish.
21
22
You can find the changes described here in the integration branches
23
of the repositories listed at
@@ -24,116 +27,405 @@ of the repositories listed at
27
--------------------------------------------------
28
[Graduated to "master"]
29
27
-* ab/doc-replace-gmane-links (2017-05-09) 2 commits
28
- (merged to 'next' on 2017-05-20 at 2c4f96560c)
29
- + doc: replace more gmane links
30
- + doc: replace a couple of broken gmane links
30
+* bc/object-id (2017-05-08) 53 commits
31
+ (merged to 'next' on 2017-05-20 at e7372733fb)
32
+ + object: convert parse_object* to take struct object_id
33
+ + tree: convert parse_tree_indirect to struct object_id
34
+ + sequencer: convert do_recursive_merge to struct object_id
35
+ + diff-lib: convert do_diff_cache to struct object_id
36
+ + builtin/ls-tree: convert to struct object_id
37
+ + merge: convert checkout_fast_forward to struct object_id
38
+ + sequencer: convert fast_forward_to to struct object_id
39
+ + builtin/ls-files: convert overlay_tree_on_cache to object_id
40
+ + builtin/read-tree: convert to struct object_id
41
+ + sha1_name: convert internals of peel_onion to object_id
42
+ + upload-pack: convert remaining parse_object callers to object_id
43
+ + revision: convert remaining parse_object callers to object_id
44
+ + revision: rename add_pending_sha1 to add_pending_oid
45
+ + http-push: convert process_ls_object and descendants to object_id
46
+ + refs/files-backend: convert many internals to struct object_id
47
+ + refs: convert struct ref_update to use struct object_id
48
+ + ref-filter: convert some static functions to struct object_id
49
+ + Convert struct ref_array_item to struct object_id
50
+ + Convert the verify_pack callback to struct object_id
51
+ + Convert lookup_tag to struct object_id
52
+ + log-tree: convert to struct object_id
53
+ + Convert lookup_tree to struct object_id
54
+ + builtin/reflog: convert tree_is_complete to take struct object_id
55
+ + tree: convert read_tree_1 to use struct object_id internally
56
+ + Convert lookup_blob to struct object_id
57
+ + Convert remaining callers of lookup_blob to object_id
58
+ + builtin/unpack-objects: convert to struct object_id
59
+ + pack: convert struct pack_idx_entry to struct object_id
60
+ + Convert lookup_commit* to struct object_id
61
+ + Convert remaining callers of lookup_commit_reference* to object_id
62
+ + builtin/tag: convert to struct object_id
63
+ + sequencer: convert some functions to struct object_id
64
+ + shallow: convert shallow registration functions to object_id
65
+ + revision: convert prepare_show_merge to struct object_id
66
+ + notes-utils: convert internals to struct object_id
67
+ + http-push: convert some static functions to struct object_id
68
+ + tag: convert parse_tag_buffer to struct object_id
69
+ + builtin/verify-commit: convert to struct object_id
70
+ + reflog_expire: convert to struct object_id
71
+ + parse-options-cb: convert to struct object_id
72
+ + notes-cache: convert to struct object_id
73
+ + submodule: convert merge_submodule to use struct object_id
74
+ + fast-import: convert to struct object_id
75
+ + fast-import: convert internal structs to struct object_id
76
+ + builtin/rev-parse: convert to struct object_id
77
+ + builtin/blame: convert static function to struct object_id
78
+ + branch: convert to struct object_id
79
+ + bundle: convert to struct object_id
80
+ + builtin/prune: convert to struct object_id
81
+ + builtin/name-rev: convert to struct object_id
82
+ + Convert struct cache_tree to use struct object_id
83
+ + Clean up outstanding object_id transforms.
84
+ + fetch-pack: convert to struct object_id
85
+ (this branch is used by mh/packed-ref-store-prep.)
86
+
87
+ Conversion from uchar[20] to struct object_id continues.
88
32
- The Web interface to gmane news archive is long gone, even though
33
- the articles are still accessible via NTTP. Replace the links with
34
- ones to public-inbox.org. Because their message identification is
35
- based on the actual message-id, it is likely that it will be easier
36
- to migrate away from it if/when necessary.
89
90
+* bm/interpret-trailers-cut-line-is-eom (2017-05-18) 1 commit
91
+ (merged to 'next' on 2017-05-23 at 3e5ff08517)
92
+ + interpret-trailers: honor the cut line
93
39
-* ab/fix-poison-tests (2017-05-11) 3 commits
40
- (merged to 'next' on 2017-05-20 at 5b13ba86bd)
41
- + travis-ci: add job to run tests with GETTEXT_POISON
42
- + travis-ci: setup "prove cache" in "script" step
43
- + tests: fix tests broken under GETTEXT_POISON=YesPlease
94
+ "git interpret-trailers", when used as GIT_EDITOR for "git commit
95
+ -v", looked for and appended to a trailer block at the very end,
96
+ i.e. at the end of the "diff" output. The command has been
97
+ corrected to pay attention to the cut-mark line "commit -v" adds to
98
+ the buffer---the real trailer block should appear just before it.
99
45
- Update tests to pass under GETTEXT_POISON (a mechanism to ensure
46
- that output strings that should not be translated are not
47
- translated by mistake), and tell TravisCI to run them.
100
101
+* bw/dir-c-stops-relying-on-the-index (2017-05-06) 14 commits
102
+ (merged to 'next' on 2017-05-20 at 1f1b764ec8)
103
+ + dir: convert fill_directory to take an index
104
+ + dir: convert read_directory to take an index
105
+ + dir: convert read_directory_recursive to take an index
106
+ + dir: convert open_cached_dir to take an index
107
+ + dir: convert is_excluded to take an index
108
+ + dir: convert prep_exclude to take an index
109
+ + dir: convert add_excludes to take an index
110
+ + dir: convert is_excluded_from_list to take an index
111
+ + dir: convert last_exclude_matching_from_list to take an index
112
+ + dir: convert dir_add* to take an index
113
+ + dir: convert get_dtype to take index
114
+ + dir: convert directory_exists_in_index to take index
115
+ + dir: convert read_skip_worktree_file_from_index to take an index
116
+ + dir: stop using the index compatibility macros
117
50
-* ah/log-decorate-default-to-auto (2017-05-15) 1 commit
51
- (merged to 'next' on 2017-05-20 at 209614479c)
52
- + builtin/log: honor log.decorate
118
+ API update.
119
54
- Setting "log.decorate=false" in the configuration file did not take
55
- effect in v2.13, which has been corrected.
120
121
+* ja/do-not-ask-needless-questions (2017-05-12) 3 commits
122
+ (merged to 'next' on 2017-05-23 at 6b373cd892)
123
+ + git-filter-branch: be more direct in an error message
124
+ + read-tree -m: make error message for merging 0 trees less smart aleck
125
+ + usability: don't ask questions if no reply is required
126
58
-* bw/submodule-with-bs-path (2017-05-01) 1 commit
59
- (merged to 'next' on 2017-05-20 at b740f784cb)
60
- + t7400: add !CYGWIN prerequisite to 'add with \\ in path'
127
+ Git sometimes gives an advice in a rhetorical question that does
128
+ not require an answer, which can confuse new users and non native
129
+ speakers. Attempt to rephrase them.
130
62
- A hotfix to a topic that is already in v2.13.
131
132
+* jc/read-tree-empty-with-m (2017-05-10) 1 commit
133
+ (merged to 'next' on 2017-05-20 at 7e86815364)
134
+ + read-tree: "read-tree -m --empty" does not make sense
135
65
-* jt/push-options-doc (2017-05-10) 2 commits
66
- (merged to 'next' on 2017-05-20 at ca7f344111)
67
- + receive-pack: verify push options in cert
68
- + docs: correct receive.advertisePushOptions default
136
+ "git read-tree -m" (no tree-ish) gave a nonsense suggestion "use
137
+ --empty if you want to clear the index". With "-m", such a request
138
+ will still fail anyway, as you'd need to name at least one tree-ish
139
+ to be merged.
140
70
- The receive-pack program now makes sure that the push certificate
71
- records the same set of push options used for pushing.
141
142
+* jc/repack-threads (2017-04-27) 1 commit
143
+ (merged to 'next' on 2017-05-20 at c1fd54a2a4)
144
+ + repack: accept --threads=<n> and pass it down to pack-objects
145
74
-* ls/travis-relays-for-windows-ci (2017-05-04) 2 commits
75
- (merged to 'next' on 2017-05-20 at 47f34b78e2)
76
- + travis-ci: retry if Git for Windows CI returns HTTP error 502 or 503
77
- + travis-ci: handle Git for Windows CI status "failed" explicitly
146
+ "git repack" learned to accept the --threads=<n> option and pass it
147
+ to pack-objects.
148
79
- Will merge to 'master'.
149
150
+* jk/alternate-ref-optim (2017-05-18) 1 commit
151
+ (merged to 'next' on 2017-05-23 at 4b4b3ac463)
152
+ + t5400: avoid concurrent writes into a trace file
153
82
-* rs/checkout-am-fix-unborn (2017-05-08) 2 commits
83
- (merged to 'next' on 2017-05-20 at d4f1ee3ed7)
84
- + am: check return value of resolve_refdup before using hash
85
- + checkout: check return value of resolve_refdup before using hash
154
+ A test allowed both "git push" and "git receive-pack" on the other
155
+ end write their traces into the same file. This is OK on platforms
156
+ that allows atomically appending to a file opened with O_APPEND,
157
+ but on other platforms led to a mangled output, causing
158
+ intermittent test failures. This has been fixed by disabling
159
+ traces from "receive-pack" in the test.
160
87
- A few codepaths in "checkout" and "am" working on an unborn branch
88
- tried to access an uninitialized piece of memory.
161
162
+* jk/bug-to-abort (2017-05-22) 4 commits
163
+ (merged to 'next' on 2017-05-22 at 20371eebb6)
164
+ + usage: add NORETURN to BUG() function definitions
165
+ (merged to 'next' on 2017-05-20 at db65acc882)
166
+ + config: complain about --local outside of a git repo
167
+ + setup_git_env: convert die("BUG") to BUG()
168
+ + usage.c: add BUG() function
169
91
-* tb/dedup-crlf-tests (2017-05-10) 1 commit
92
- (merged to 'next' on 2017-05-20 at 0a7ffff154)
93
- + t0027: tests are not expensive; remove t0025
170
+ Introduce the BUG() macro to improve die("BUG: ...").
171
+
172
+
173
+* jk/diff-submodule-diff-inline (2017-05-08) 1 commit
174
+ (merged to 'next' on 2017-05-20 at da7eb2626a)
175
+ + diff: recurse into nested submodules for inline diff
176
+
177
+ "git diff --submodule=diff" now recurses into nested submodules.
178
+
179
+
180
+* jk/disable-pack-reuse-when-broken (2017-05-09) 2 commits
181
+ (merged to 'next' on 2017-05-20 at 684e921273)
182
+ + t5310: fix "; do" style
183
+ + pack-objects: disable pack reuse for object-selection options
184
+
185
+ "pack-objects" can stream a slice of an existing packfile out when
186
+ the pack bitmap can tell that the reachable objects are all needed
187
+ in the output, without inspecting individual objects. This
188
+ strategy however would not work well when "--local" and other
189
+ options are in use, and need to be disabled.
190
+
191
+
192
+* jk/doc-config-include (2017-05-12) 4 commits
193
+ (merged to 'next' on 2017-05-23 at 4f60843265)
194
+ + docs/config: consistify include.path examples
195
+ + docs/config: avoid the term "expand" for includes
196
+ + docs/config: give a relative includeIf example
197
+ + docs/config: clarify include/includeIf relationship
198
+
199
+ Clarify documentation for include.path and includeIf.<condition>.path
200
+ configuration variables.
201
+
202
+
203
+* jk/ignore-broken-tags-when-ignoring-missing-links (2017-05-20) 1 commit
204
+ (merged to 'next' on 2017-05-23 at 977491e52a)
205
+ + revision.c: ignore broken tags with ignore_missing_links
206
+
207
+ Tag objects, which are not reachable from any ref, that point at
208
+ missing objects were mishandled by "git gc" and friends (they
209
+ should silently be ignored instead)
210
+
211
+
212
+* jk/update-links-in-docs (2017-05-15) 1 commit
213
+ (merged to 'next' on 2017-05-23 at 5b2df9fdbe)
214
+ + doc: use https links to Wikipedia to avoid http redirects
215
+
216
+ A few http:// links that are redirected to https:// in the
217
+ documentation have been updated to https:// links.
218
+
219
+
220
+* js/eol-on-ourselves (2017-05-10) 6 commits
221
+ (merged to 'next' on 2017-05-20 at 8023c277d2)
222
+ + t4051: mark supporting files as requiring LF-only line endings
223
+ + Fix the remaining tests that failed with core.autocrlf=true
224
+ + t3901: move supporting files into t/t3901/
225
+ + completion: mark bash script as LF-only
226
+ + git-new-workdir: mark script as LF-only
227
+ + Fix build with core.autocrlf=true
228
+
229
+ Make sure our tests would pass when the sources are checked out
230
+ with "platform native" line ending convention by default on
231
+ Windows. Some "text" files out tests use and the test scripts
232
+ themselves that are meant to be run with /bin/sh, ought to be
233
+ checked out with eol=LF even on Windows.
234
+
235
+
236
+* js/plug-leaks (2017-05-09) 26 commits
237
+ (merged to 'next' on 2017-05-20 at fb136ea2dc)
238
+ + checkout: fix memory leak
239
+ + submodule_uses_worktrees(): plug memory leak
240
+ + show_worktree(): plug memory leak
241
+ + name-rev: avoid leaking memory in the `deref` case
242
+ + remote: plug memory leak in match_explicit()
243
+ + add_reflog_for_walk: avoid memory leak
244
+ + shallow: avoid memory leak
245
+ + line-log: avoid memory leak
246
+ + receive-pack: plug memory leak in update()
247
+ + fast-export: avoid leaking memory in handle_tag()
248
+ + mktree: plug memory leaks reported by Coverity
249
+ + pack-redundant: plug memory leak
250
+ + setup_discovered_git_dir(): plug memory leak
251
+ + setup_bare_git_dir(): help static analysis
252
+ + split_commit_in_progress(): simplify & fix memory leak
253
+ + checkout: fix memory leak
254
+ + cat-file: fix memory leak
255
+ + mailinfo & mailsplit: check for EOF while parsing
256
+ + status: close file descriptor after reading git-rebase-todo
257
+ + difftool: address a couple of resource/memory leaks
258
+ + get_mail_commit_oid(): avoid resource leak
259
+ + git_config_rename_section_in_file(): avoid resource leak
260
+ + add_commit_patch_id(): avoid allocating memory unnecessarily
261
+ + winansi: avoid buffer overrun
262
+ + winansi: avoid use of uninitialized value
263
+ + mingw: avoid memory leak when splitting PATH
264
+
265
+ Fix memory leaks pointed out by Coverity (and people).
266
+
267
+
268
+* km/log-showsignature-doc (2017-05-20) 1 commit
269
+ (merged to 'next' on 2017-05-23 at f43ea11956)
270
+ + config.txt: add an entry for log.showSignature
271
+
272
+ Doc update.
273
+
274
+
275
+* kn/ref-filter-branch-list (2017-05-20) 1 commit
276
+ (merged to 'next' on 2017-05-23 at e0cc721af0)
277
+ + ref-filter: resolve HEAD when parsing %(HEAD) atom
278
+
279
+ "git for-each-ref --format=..." with %(HEAD) in the format used to
280
+ resolve the HEAD symref as many times as it had processed refs,
281
+ which was wasteful, and "git branch" shared the same problem.
282
+
283
+
284
+* nd/split-index-unshare (2017-05-08) 2 commits
285
+ (merged to 'next' on 2017-05-20 at 9048cdc2ce)
286
+ + p3400: add perf tests for rebasing many changes
287
+ + split-index: add and use unshare_split_index()
288
+
289
+ Plug some leaks and updates internal API used to implement the
290
+ split index feature to make it easier to avoid such a leak in the
291
+ future.
292
+
293
+
294
+* pw/rebase-i-regression-fix (2017-05-20) 3 commits
295
+ (merged to 'next' on 2017-05-23 at 14676d0013)
296
+ + rebase -i: add missing newline to end of message
297
+ + rebase -i: silence stash apply
298
+ + rebase -i: fix reflog message
299
+
300
+ Regression fix to topic recently merged to 'master'.
301
95
- Will merge to 'master'.
302
+
303
+* sb/checkout-recurse-submodules (2017-05-04) 3 commits
304
+ (merged to 'next' on 2017-05-20 at f972b2069f)
305
+ + submodule: properly recurse for read-tree and checkout
306
+ + submodule: avoid auto-discovery in new working tree manipulator code
307
+ + submodule_move_head: reuse child_process structure for futher commands
308
+
309
+ "git checkout --recurse-submodules" did not quite work with a
310
+ submodule that itself has submodules.
311
+
312
+
313
+* sb/reset-recurse-submodules (2017-04-23) 4 commits
314
+ (merged to 'next' on 2017-05-20 at 8adafdf1be)
315
+ + builtin/reset: add --recurse-submodules switch
316
+ + submodule.c: submodule_move_head works with broken submodules
317
+ + submodule.c: uninitialized submodules are ignored in recursive commands
318
+ + entry.c: submodule recursing: respect force flag correctly
319
+ (this branch is used by sb/submodule-blanket-recursive.)
320
+
321
+ "git reset" learned "--recurse-submodules" option.
322
+
323
+
324
+* tg/stash-push-fixup (2017-05-17) 1 commit
325
+ (merged to 'next' on 2017-05-23 at 338eb650f2)
326
+ + completion: add git stash push
327
+
328
+ The shell completion script (in contrib/) learned "git stash" has
329
+ a new "push" subcommand.
330
331
--------------------------------------------------
332
[New Topics]
333
100
-* ah/doc-interpret-trailers-ifexists (2017-05-23) 1 commit
101
- - Documentation: fix reference to ifExists for interpret-trailers
334
+* ab/pcre-v2 (2017-05-26) 7 commits
335
+ - grep: add support for PCRE v2
336
+ - grep: un-break building with PCRE < 8.20
337
+ - grep: un-break building with PCRE < 8.32
338
+ - grep: add support for the PCRE v1 JIT API
339
+ - log: add -P as a synonym for --perl-regexp
340
+ - grep: skip pthreads overhead when using one thread
341
+ - grep: don't redundantly compile throwaway patterns under threading
342
+ (this branch uses ab/grep-preparatory-cleanup.)
343
103
- Documentation fix.
344
+ Update "perl-compatible regular expression" support to enable JIT
345
+ and also allow linking with the newer PCRE v2 library.
346
347
Will merge to 'next'.
348
349
108
-* ah/doc-pretty-format-fix (2017-05-23) 1 commit
109
- - Documentation: fix formatting typo in pretty-formats.txt
350
+* ab/ref-filter-no-contains (2017-05-23) 1 commit
351
+ (merged to 'next' on 2017-05-29 at 5d39fd2961)
352
+ + tag: duplicate mention of --contains should mention --no-contains
353
111
- Documentation fix.
354
+ Doc update to a recent topic.
355
+
356
+ Will merge to 'master'.
357
+
358
+
359
+* rs/mingw-path-lookup-simplify (2017-05-23) 1 commit
360
+ (merged to 'next' on 2017-05-29 at 4365139a7f)
361
+ + mingw: simplify PATH handling
362
+
363
+ Code simplification.
364
+
365
+ Will merge to 'master'.
366
+
367
+
368
+* ab/sha1dc-maint (2017-05-22) 1 commit
369
+ - sha1dc: update from upstream
370
+ (this branch is used by ab/sha1dc.)
371
+
372
+ The "collision detecting" SHA-1 implementation shipped with 2.13
373
+ was quite broken on some big-endian platforms and/or platforms that
374
+ do not like unaligned fetches. Update to the upstream code which
375
+ has already fixed these issues.
376
377
Will merge to 'next'.
378
379
116
-* sb/submodule-blanket-recursive (2017-05-23) 6 commits
117
- . builtin/push.c: respect 'submodule.recurse' option
118
- . builtin/grep.c: respect 'submodule.recurse' option
119
- . builtin/fetch.c: respect 'submodule.recurse' option
120
- . Introduce submodule.recurse option for worktree manipulators
121
- . submodule test invocation: only pass additional arguments
122
- . submodule.c: add has_submodules to check if we have any submodules
123
- (this branch uses sb/reset-recurse-submodules.)
380
+* ab/t3070-test-dedup (2017-05-29) 1 commit
381
+ - wildmatch test: remove redundant duplicate test
382
125
- A new configuration variable "submodule.recurse" can be set to true
126
- to force various commands run at the top-level superproject to
127
- behave as if they were invoked with the "--recurse-submodules"
128
- option.
383
+ Test cleanup.
384
130
- Seems to break t7814 when merged to 'pu'.
385
+ Will merge to 'next'.
386
387
133
-* sb/t5531-update-desc (2017-05-23) 1 commit
134
- - t5531: fix test description
388
+* ah/doc-filter-branch-export-env (2017-05-29) 1 commit
389
+ - doc: filter-branch does not require re-export of vars
390
136
- The description strings for a few tests have been updated.
391
+ Docfix.
392
+
393
+ Will merge to 'next'.
394
+
395
+
396
+* jk/connect-symref-info-leak-fix (2017-05-26) 1 commit
397
+ - connect.c: fix leak in parse_one_symref_info()
398
+
399
+ Leakfix.
400
+
401
+ Will merge to 'next'.
402
+
403
+
404
+* jk/pathspec-magic-disambiguation (2017-05-29) 6 commits
405
+ - verify_filename(): flip order of checks
406
+ - verify_filename(): treat ":(magic)" as a pathspec
407
+ - check_filename(): handle ":^" path magic
408
+ - check_filename(): use skip_prefix
409
+ - check_filename(): refactor ":/" handling
410
+ - t4208: add check for ":/" without matching file
411
+
412
+ The convention for a command line is to follow "git cmdname
413
+ --options" with revisions followed by an optional "--"
414
+ disambiguator and then finally pathspecs. When "--" is not there,
415
+ we make sure early ones are all interpretable as revs (and do not
416
+ look like paths) and later ones are the other way around. A
417
+ pathspec with "magic" (e.g. ":/p/a/t/h" that matches p/a/t/h from
418
+ the top-level of the working tree, no matter what subdirectory you
419
+ are working from) are conservatively judged as "not a path", which
420
+ required disambiguation more often. The command line parser
421
+ learned to say "it's a pathspec" a bit more often when the syntax
422
+ looks like so.
423
+
424
+
425
+* sd/t3200-typofix (2017-05-29) 1 commit
426
+ - branch test: fix invalid config key access
427
+
428
+ Test fix.
429
430
Will merge to 'next'.
431
@@ -158,6 +450,9 @@ of the repositories listed at
450
"git describe --debug --contains" did not add any meaningful
451
information, even though without "--contains" it did.
452
453
+ Will discard for now.
454
+ cf. <c85bc2a1-56d8-8a02-6089-2b8cb3d39e99@grubix.eu>
455
+
456
457
* nd/worktree-move (2017-04-20) 6 commits
458
- worktree remove: new command
@@ -175,22 +470,6 @@ of the repositories listed at
470
cf. <d0e81b1e-5869-299e-f462-4d43dc997bd1@ramsayjones.plus.com>
471
472
178
-* df/dir-iter-remove-subtree (2017-04-19) 5 commits
179
- - remove_subtree(): reimplement using iterators
180
- - dir_iterator: rewrite state machine model
181
- - dir_iterator: refactor dir_iterator_advance
182
- - remove_subtree(): test removing nested directories
183
- - dir_iterator: add tests for dir_iterator API
184
-
185
- Update the dir-iterator API and use it to reimplement
186
- remove_subtree().
187
-
188
- A reroll exists that is based on the updated 'master', but I ran
189
- out of time trying to get it to work with other topics in flight
190
- in 'pu'.
191
- GSoC microproject.
192
-
193
-
473
* ls/filter-process-delayed (2017-03-06) 1 commit
474
- convert: add "status=delayed" to filter process protocol
475
@@ -214,6 +493,33 @@ of the repositories listed at
493
--------------------------------------------------
494
[Cooking]
495
496
+* ah/doc-interpret-trailers-ifexists (2017-05-23) 1 commit
497
+ (merged to 'next' on 2017-05-29 at cb353c1d21)
498
+ + Documentation: fix reference to ifExists for interpret-trailers
499
+
500
+ Documentation fix.
501
+
502
+ Will merge to 'master'.
503
+
504
+
505
+* ah/doc-pretty-format-fix (2017-05-23) 1 commit
506
+ (merged to 'next' on 2017-05-29 at 6e3e8fd80d)
507
+ + Documentation: fix formatting typo in pretty-formats.txt
508
+
509
+ Documentation fix.
510
+
511
+ Will merge to 'master'.
512
+
513
+
514
+* sb/t5531-update-desc (2017-05-23) 1 commit
515
+ (merged to 'next' on 2017-05-29 at 67a91db3cf)
516
+ + t5531: fix test description
517
+
518
+ The description strings for a few tests have been updated.
519
+
520
+ Will merge to 'master'.
521
+
522
+
523
* ab/conditional-config-with-symlinks (2017-05-17) 1 commit
524
(merged to 'next' on 2017-05-23 at 6162222c2e)
525
+ config: match both symlink & realpath versions in IncludeIf.gitdir:*
@@ -229,19 +535,6 @@ of the repositories listed at
535
Will merge to 'master'.
536
537
232
-* bm/interpret-trailers-cut-line-is-eom (2017-05-18) 1 commit
233
- (merged to 'next' on 2017-05-23 at 3e5ff08517)
234
- + interpret-trailers: honor the cut line
235
-
236
- "git interpret-trailers", when used as GIT_EDITOR for "git commit
237
- -v", looked for and appended to a trailer block at the very end,
238
- i.e. at the end of the "diff" output. The command has been
239
- corrected to pay attention to the cut-mark line "commit -v" adds to
240
- the buffer---the real trailer block should appear just before it.
241
-
242
- Will merge to 'master'.
243
-
244
-
538
* jc/skip-test-in-the-middle (2017-05-18) 2 commits
539
(merged to 'next' on 2017-05-23 at 6c962b353b)
540
+ t5545: enhance test coverage when no http server is installed
@@ -255,86 +548,61 @@ of the repositories listed at
548
Will merge to 'master'.
549
550
258
-* jk/alternate-ref-optim (2017-05-18) 1 commit
259
- (merged to 'next' on 2017-05-23 at 4b4b3ac463)
260
- + t5400: avoid concurrent writes into a trace file
261
-
262
- A test allowed both "git push" and "git receive-pack" on the other
263
- end write their traces into the same file. This is OK on platforms
264
- that allows atomically appending to a file opened with O_APPEND,
265
- but on other platforms led to a mangled output, causing
266
- intermittent test failures. This has been fixed by disabling
267
- traces from "receive-pack" in the test.
268
-
269
- Will merge to 'master'.
270
-
271
-
272
-* mh/packed-ref-store-prep (2017-05-23) 26 commits
273
- - cache_ref_iterator_begin(): avoid priming unneeded directories
274
- - ref-filter: limit traversal to prefix
275
- - create_ref_entry(): remove `check_name` option
276
- - refs_ref_iterator_begin(): handle `GIT_REF_PARANOIA`
277
- - read_packed_refs(): report unexpected fopen() failures
278
- - read_packed_refs(): do more of the work of reading packed refs
279
- - get_packed_ref_cache(): assume "packed-refs" won't change while locked
280
- - should_pack_ref(): new function, extracted from `files_pack_refs()`
281
- - ref_update_reject_duplicates(): add a sanity check
282
- - ref_update_reject_duplicates(): use `size_t` rather than `int`
283
- - ref_update_reject_duplicates(): expose function to whole refs module
284
- - ref_transaction_prepare(): new optional step for reference updates
285
- - ref_transaction_commit(): check for valid `transaction->state`
286
- - files_transaction_cleanup(): new helper function
287
- - files_ref_store: put the packed files lock directly in this struct
288
- - files-backend: move `lock` member to `files_ref_store`
289
- - lockfile: add a new method, is_lock_file_locked()
290
- - ref_store: take a `msg` parameter when deleting references
291
- - refs: use `size_t` indexes when iterating over ref transaction updates
292
- - refs_ref_iterator_begin(): don't check prefixes redundantly
293
- - prefix_ref_iterator: don't trim too much
294
- - files-backend: use `die("BUG: ...")`, not `die("internal error: ...")`
295
- - ref_iterator_begin_fn(): fix docstring
296
- - refs.h: clarify docstring for the ref_transaction_update()-related fns
297
- - t3600: clean up permissions test properly
298
- - Merge branch 'bc/object-id'
299
- (this branch uses bc/object-id.)
551
+* mh/packed-ref-store-prep (2017-05-24) 26 commits
552
+ (merged to 'next' on 2017-05-29 at e4f5e06b4c)
553
+ + cache_ref_iterator_begin(): avoid priming unneeded directories
554
+ + ref-filter: limit traversal to prefix
555
+ + create_ref_entry(): remove `check_name` option
556
+ + refs_ref_iterator_begin(): handle `GIT_REF_PARANOIA`
557
+ + read_packed_refs(): report unexpected fopen() failures
558
+ + read_packed_refs(): do more of the work of reading packed refs
559
+ + get_packed_ref_cache(): assume "packed-refs" won't change while locked
560
+ + should_pack_ref(): new function, extracted from `files_pack_refs()`
561
+ + ref_update_reject_duplicates(): add a sanity check
562
+ + ref_update_reject_duplicates(): use `size_t` rather than `int`
563
+ + ref_update_reject_duplicates(): expose function to whole refs module
564
+ + ref_transaction_prepare(): new optional step for reference updates
565
+ + ref_transaction_commit(): check for valid `transaction->state`
566
+ + files_transaction_cleanup(): new helper function
567
+ + files_ref_store: put the packed files lock directly in this struct
568
+ + files-backend: move `lock` member to `files_ref_store`
569
+ + lockfile: add a new method, is_lock_file_locked()
570
+ + ref_store: take a `msg` parameter when deleting references
571
+ + refs: use `size_t` indexes when iterating over ref transaction updates
572
+ + refs_ref_iterator_begin(): don't check prefixes redundantly
573
+ + prefix_ref_iterator: don't trim too much
574
+ + files-backend: use `die("BUG: ...")`, not `die("internal error: ...")`
575
+ + ref_iterator_begin_fn(): fix docstring
576
+ + refs.h: clarify docstring for the ref_transaction_update()-related fns
577
+ + t3600: clean up permissions test properly
578
+ + Merge branch 'bc/object-id'
579
580
The implementation of "ref" API around the "packed refs" have been
581
cleaned up, in preparation for further changes.
582
304
- Will merge to 'next'.
583
+ Will merge to 'master'.
584
585
307
-* sl/clean-d-ignored-fix (2017-05-22) 7 commits
308
- - SQUASH????
309
- - clean: teach clean -d to skip dirs containing ignored files
310
- - dir: expose cmp_name() and check_contains()
311
- - dir: hide untracked contents of untracked dirs
312
- - dir: recurse into untracked dirs for ignored files
313
- - t7061: status --ignored should search untracked dirs
314
- - t7300: clean -d should skip dirs with ignored files
586
+* sl/clean-d-ignored-fix (2017-05-24) 6 commits
587
+ (merged to 'next' on 2017-05-29 at 837c255ae8)
588
+ + clean: teach clean -d to preserve ignored paths
589
+ + dir: expose cmp_name() and check_contains()
590
+ + dir: hide untracked contents of untracked dirs
591
+ + dir: recurse into untracked dirs for ignored files
592
+ + t7061: status --ignored should search untracked dirs
593
+ + t7300: clean -d should skip dirs with ignored files
594
595
"git clean -d" used to clean directories that has ignored files,
596
even though the command should not lose ignored ones without "-x".
597
This has been corrected.
598
320
- Expecting a reroll.
321
- cf. <xmqq1srgko52.fsf@gitster.mtv.corp.google.com>
322
-
323
-
324
-* tg/stash-push-fixup (2017-05-17) 1 commit
325
- (merged to 'next' on 2017-05-23 at 338eb650f2)
326
- + completion: add git stash push
327
-
328
- The shell completion script (in contrib/) learned "git stash" has
329
- a new "push" subcommand.
330
-
599
Will merge to 'master'.
600
601
334
-* ab/sha1dc (2017-05-22) 3 commits
602
+* ab/sha1dc (2017-05-22) 2 commits
603
- sha1collisiondetection: automatically enable when submodule is populated
604
- sha1dc: optionally use sha1collisiondetection as a submodule
337
- - sha1dc: update from upstream
605
+ (this branch uses ab/sha1dc-maint.)
606
607
The "collission-detecting" implementation of SHA-1 hash we borrowed
608
from is replaced by directly binding the upstream project as our
@@ -345,20 +613,14 @@ of the repositories listed at
613
submodule is not yet fully known, but this lets us dip our toes.
614
615
348
-* bp/fsmonitor (2017-05-21) 10 commits
349
- - SQUASH???
350
- - SQUASH???
351
- - SQUASH???
616
+* bp/fsmonitor (2017-05-26) 6 commits
617
- fsmonitor: add a sample query-fsmonitor hook script for Watchman
618
- fsmonitor: add documentation for the fsmonitor extension.
619
- fsmonitor: add test cases for fsmonitor extension
355
- - SQUASH???
620
- fsmonitor: teach git to optionally utilize a file system monitor to speed up detecting new or changed files.
621
- dir: make lookup_untracked() available outside of dir.c
622
- bswap: add 64 bit endianness helper get_be64
623
360
- Expecting a reroll.
361
-
624
625
* dk/send-email-avoid-net-smtp-ssl-when-able (2017-05-20) 1 commit
626
(merged to 'next' on 2017-05-23 at fc59b8a1d4)
@@ -384,44 +646,34 @@ of the repositories listed at
646
Will merge to 'master'.
647
648
387
-* jk/diff-blob (2017-05-20) 15 commits
388
- - diff: use blob path for blob/file diffs
389
- - diff: use pending "path" if it is available
390
- - diff: use the word "path" instead of "name" for blobs
391
- - diff: pass whole pending entry in blobinfo
392
- - handle_revision_arg: record paths for pending objects
393
- - handle_revision_arg: record modes for "a..b" endpoints
394
- - t4063: add tests of direct blob diffs
395
- - get_sha1_with_context: dynamically allocate oc->path
396
- - get_sha1_with_context: always initialize oc->symlink_path
397
- - sha1_name: consistently refer to object_context as "oc"
398
- - handle_revision_arg: add handle_dotdot() helper
399
- - handle_revision_arg: hoist ".." check out of range parsing
400
- - handle_revision_arg: stop using "dotdot" as a generic pointer
401
- - handle_revision_arg: simplify commit reference lookups
402
- - handle_revision_arg: reset "dotdot" consistently
649
+* jk/diff-blob (2017-05-24) 15 commits
650
+ (merged to 'next' on 2017-05-29 at 5ecc979cc7)
651
+ + diff: use blob path for blob/file diffs
652
+ + diff: use pending "path" if it is available
653
+ + diff: use the word "path" instead of "name" for blobs
654
+ + diff: pass whole pending entry in blobinfo
655
+ + handle_revision_arg: record paths for pending objects
656
+ + handle_revision_arg: record modes for "a..b" endpoints
657
+ + t4063: add tests of direct blob diffs
658
+ + get_sha1_with_context: dynamically allocate oc->path
659
+ + get_sha1_with_context: always initialize oc->symlink_path
660
+ + sha1_name: consistently refer to object_context as "oc"
661
+ + handle_revision_arg: add handle_dotdot() helper
662
+ + handle_revision_arg: hoist ".." check out of range parsing
663
+ + handle_revision_arg: stop using "dotdot" as a generic pointer
664
+ + handle_revision_arg: simplify commit reference lookups
665
+ + handle_revision_arg: reset "dotdot" consistently
666
667
The result from "git diff" that compares two blobs, e.g. "git diff
668
$commit1:$path $commit2:$path", used to be shown with the full
669
object name as given on the command line, but it is more natural to
670
use the $path in the output and use it to look up .gitattributes.
671
409
- Needs review.
410
-
411
-
412
-* jk/ignore-broken-tags-when-ignoring-missing-links (2017-05-20) 1 commit
413
- (merged to 'next' on 2017-05-23 at 977491e52a)
414
- + revision.c: ignore broken tags with ignore_missing_links
415
-
416
- Tag objects, which are not reachable from any ref, that point at
417
- missing objects were mishandled by "git gc" and friends (they
418
- should silently be ignored instead)
419
-
672
Will merge to 'master'.
673
674
423
-* js/bs-is-a-dir-sep-on-windows (2017-05-20) 2 commits
424
- (merged to 'next' on 2017-05-23 at e27040ceb0)
675
+* js/bs-is-a-dir-sep-on-windows (2017-05-26) 2 commits
676
+ (merged to 'next' on 2017-05-26 at 450b39f726)
677
+ Windows: do not treat a path with backslashes as a remote's nick name
678
+ mingw.h: permit arguments with side effects for is_dir_sep
679
@@ -441,60 +693,29 @@ of the repositories listed at
693
Will merge to 'master'.
694
695
444
-* km/log-showsignature-doc (2017-05-20) 1 commit
445
- (merged to 'next' on 2017-05-23 at f43ea11956)
446
- + config.txt: add an entry for log.showSignature
447
-
448
- Will merge to 'master'.
449
-
450
-
451
-* kn/ref-filter-branch-list (2017-05-20) 1 commit
452
- (merged to 'next' on 2017-05-23 at e0cc721af0)
453
- + ref-filter: resolve HEAD when parsing %(HEAD) atom
454
-
455
- "git for-each-ref --format=..." with %(HEAD) in the format used to
456
- resolve the HEAD symref as many times as it had processed refs,
457
- which was wasteful, and "git branch" shared the same problem.
458
-
459
- Will merge to 'master'.
460
-
461
-
462
-* pw/rebase-i-regression-fix (2017-05-20) 3 commits
463
- (merged to 'next' on 2017-05-23 at 14676d0013)
464
- + rebase -i: add missing newline to end of message
465
- + rebase -i: silence stash apply
466
- + rebase -i: fix reflog message
467
-
468
- Regression fix to topic recently merged to 'master'.
469
-
470
- Will merge to 'master'.
471
-
472
-
473
-* sb/diff-color-move (2017-05-23) 17 commits
474
- . diff.c: color moved lines differently
475
- . diff: buffer all output if asked to
476
- . diff.c: emit_line includes whitespace highlighting
477
- . diff.c: convert diff_summary to use emit_line_*
478
- . diff.c: convert diff_flush to use emit_line_*
479
- . diff.c: convert word diffing to use emit_line_*
480
- . diff.c: convert show_stats to use emit_line_*
481
- . diff.c: convert emit_binary_diff_body to use emit_line_*
482
- . submodule.c: convert show_submodule_summary to use emit_line_fmt
483
- . diff.c: convert emit_rewrite_lines to use emit_line_*
484
- . diff.c: convert emit_rewrite_diff to use emit_line_*
485
- . diff.c: convert builtin_diff to use emit_line_*
486
- . diff.c: convert fn_out_consume to use emit_line
487
- . diff: introduce more flexible emit function
488
- . diff.c: factor out diff_flush_patch_all_file_pairs
489
- . diff: move line ending check into emit_hunk_header
490
- . diff: readability fix
696
+* sb/diff-color-move (2017-05-25) 17 commits
697
+ - diff.c: color moved lines differently
698
+ - diff: buffer all output if asked to
699
+ - diff.c: emit_line includes whitespace highlighting
700
+ - diff.c: convert diff_summary to use emit_line_*
701
+ - diff.c: convert diff_flush to use emit_line_*
702
+ - diff.c: convert word diffing to use emit_line_*
703
+ - diff.c: convert show_stats to use emit_line_*
704
+ - diff.c: convert emit_binary_diff_body to use emit_line_*
705
+ - submodule.c: convert show_submodule_summary to use emit_line_fmt
706
+ - diff.c: convert emit_rewrite_lines to use emit_line_*
707
+ - diff.c: convert emit_rewrite_diff to use emit_line_*
708
+ - diff.c: convert builtin_diff to use emit_line_*
709
+ - diff.c: convert fn_out_consume to use emit_line
710
+ - diff: introduce more flexible emit function
711
+ - diff.c: factor out diff_flush_patch_all_file_pairs
712
+ - diff: move line ending check into emit_hunk_header
713
+ - diff: readability fix
714
715
"git diff" has been taught to optionally paint new lines that are
716
the same as deleted lines elsewhere differently from genuinely new
717
lines.
718
496
- Seems to break t4060 when merged to 'next'.
497
-
719
720
* xz/send-email-batch-size (2017-05-23) 1 commit
721
- send-email: --batch-size to work around some SMTP server limit
@@ -520,39 +741,19 @@ of the repositories listed at
741
Will merge to 'master'.
742
743
523
-* jk/bug-to-abort (2017-05-22) 4 commits
524
- (merged to 'next' on 2017-05-22 at 20371eebb6)
525
- + usage: add NORETURN to BUG() function definitions
526
- (merged to 'next' on 2017-05-20 at db65acc882)
527
- + config: complain about --local outside of a git repo
528
- + setup_git_env: convert die("BUG") to BUG()
529
- + usage.c: add BUG() function
530
-
531
- Introduce the BUG() macro to improve die("BUG: ...").
532
-
533
- Will merge to 'master'.
534
-
535
-
536
-* jk/update-links-in-docs (2017-05-15) 1 commit
537
- (merged to 'next' on 2017-05-23 at 5b2df9fdbe)
538
- + doc: use https links to Wikipedia to avoid http redirects
539
-
540
- A few http:// links that are redirected to https:// in the
541
- documentation have been updated to https:// links.
542
-
543
- Will merge to 'master'.
544
-
545
-
546
-* js/blame-lib (2017-05-15) 22 commits
547
- - blame: create entry prepend function in libgit
548
- - blame: create scoreboard setup function in libgit
549
- - blame: create scoreboard init function in libgit
744
+* js/blame-lib (2017-05-25) 29 commits
745
+ - blame: move entry prepend to libgit
746
+ - blame: move scoreboard setup to libgit
747
- blame: move scoreboard-related methods to libgit
748
- blame: move fake-commit-related methods to libgit
749
- blame: move origin-related methods to libgit
750
+ - blame: move core structures to header
751
+ - blame: create entry prepend function
752
+ - blame: create scoreboard setup function
753
+ - blame: create scoreboard init function
754
- blame: rework methods that determine 'final' commit
755
- blame: wrap blame_sort and compare_blame_final
555
- - blame: move progess updates to a scoreboard callback
756
+ - blame: move progress updates to a scoreboard callback
757
- blame: make sanity_check use a callback in scoreboard
758
- blame: move no_whole_file_rename flag to scoreboard
759
- blame: move xdl_opts flags to scoreboard
@@ -561,8 +762,12 @@ of the repositories listed at
762
- blame: move contents_from to scoreboard
763
- blame: move copy/move thresholds to scoreboard
764
- blame: move stat counters to scoreboard
564
- - blame: move scoreboard structure to header
565
- - blame: move origin and entry structures to header
765
+ - blame: rename nth_line function
766
+ - blame: rename ent_score function
767
+ - blame: rename coalesce function
768
+ - blame: rename origin-related functions
769
+ - blame: rename scoreboard structure to blame_scoreboard
770
+ - blame: rename origin structure to blame_origin
771
- blame: remove unused parameters
772
- blame: move textconv_object with related functions
773
- blame: remove unneeded dependency on blob.h
@@ -570,12 +775,7 @@ of the repositories listed at
775
The internal logic used in "git blame" has been libified to make it
776
easier to use by cgit.
777
573
- RFC, and is hard to review due to renaming of symbols and moving of
574
- lines across files are intermixed but I think the overall direction
575
- is a good one.
576
-
577
- Expecting a rerolled non-RFC series.
578
- cf. <20170514031513.9042-1-whydoubt@gmail.com>
778
+ Will merge to 'next'.
779
780
781
* jt/send-email-validate-hook (2017-05-16) 1 commit
@@ -614,76 +814,54 @@ of the repositories listed at
814
Will merge to 'master'.
815
816
617
-* jc/read-tree-empty-with-m (2017-05-10) 1 commit
618
- (merged to 'next' on 2017-05-20 at 7e86815364)
619
- + read-tree: "read-tree -m --empty" does not make sense
620
-
621
- "git read-tree -m" (no tree-ish) gave a nonsense suggestion "use
622
- --empty if you want to clear the index". With "-m", such a request
623
- will still fail anyway, as you'd need to name at least one tree-ish
624
- to be merged.
625
-
626
- Will merge to 'master'.
627
-
628
-
629
-* ab/perf-wildmatch (2017-05-12) 2 commits
630
- (merged to 'next' on 2017-05-23 at 0adb7dac31)
631
- + perf: add test showing exponential growth in path globbing
632
- + perf: add function to setup a fresh test repo
633
-
634
- Add perf-test for wildmatch.
635
-
636
- Will merge to 'master'.
637
-
638
-
639
-* jk/doc-config-include (2017-05-12) 4 commits
640
- (merged to 'next' on 2017-05-23 at 4f60843265)
641
- + docs/config: consistify include.path examples
642
- + docs/config: avoid the term "expand" for includes
643
- + docs/config: give a relative includeIf example
644
- + docs/config: clarify include/includeIf relationship
817
+* ab/perf-wildmatch (2017-05-12) 2 commits
818
+ (merged to 'next' on 2017-05-23 at 0adb7dac31)
819
+ + perf: add test showing exponential growth in path globbing
820
+ + perf: add function to setup a fresh test repo
821
646
- Clarify documentation for include.path and includeIf.<condition>.path
647
- configuration variables.
822
+ Add perf-test for wildmatch.
823
824
Will merge to 'master'.
825
826
652
-* ab/grep-preparatory-cleanup (2017-05-21) 30 commits
653
- - grep: assert that threading is enabled when calling grep_{lock,unlock}
654
- - grep: given --threads with NO_PTHREADS=YesPlease, warn
655
- - pack-objects: fix buggy warning about threads
656
- - pack-objects & index-pack: add test for --threads warning
657
- - test-lib: add a PTHREADS prerequisite
658
- - grep: move is_fixed() earlier to avoid forward declaration
659
- - grep: change internal *pcre* variable & function names to be *pcre1*
660
- - grep: change the internal PCRE macro names to be PCRE1
661
- - grep: factor test for \0 in grep patterns into a function
662
- - grep: remove redundant regflags assignments
663
- - grep: catch a missing enum in switch statement
664
- - perf: add a comparison test of log --grep regex engines
665
- - perf: add a comparison test of grep regex engines with -F
666
- - perf: add a comparison test of grep regex engines
667
- - perf: emit progress output when unpacking & building
668
- - perf: add a GIT_PERF_MAKE_COMMAND for when *_MAKE_OPTS won't do
669
- - grep: add tests to fix blind spots with \0 patterns
670
- - grep: prepare for testing binary regexes containing rx metacharacters
671
- - grep: add a test helper function for less verbose -f \0 tests
672
- - grep: add tests for grep pattern types being passed to submodules
673
- - grep: amend submodule recursion test for regex engine testing
674
- - grep: add tests for --threads=N and grep.threads
675
- - grep: change non-ASCII -i test to stop using --debug
676
- - grep: add a test for backreferences in PCRE patterns
677
- - grep: add a test asserting that --perl-regexp dies when !PCRE
678
- - log: make --regexp-ignore-case work with --perl-regexp
679
- - log: add exhaustive tests for pattern style options & config
680
- - test-lib: rename the LIBPCRE prerequisite to PCRE
681
- - grep & rev-list doc: stop promising libpcre for --perl-regexp
682
- - Makefile & configure: reword inaccurate comment about PCRE
827
+* ab/grep-preparatory-cleanup (2017-05-26) 31 commits
828
+ (merged to 'next' on 2017-05-29 at f2cfa89d3e)
829
+ + grep: assert that threading is enabled when calling grep_{lock,unlock}
830
+ + grep: given --threads with NO_PTHREADS=YesPlease, warn
831
+ + pack-objects: fix buggy warning about threads
832
+ + pack-objects & index-pack: add test for --threads warning
833
+ + test-lib: add a PTHREADS prerequisite
834
+ + grep: move is_fixed() earlier to avoid forward declaration
835
+ + grep: change internal *pcre* variable & function names to be *pcre1*
836
+ + grep: change the internal PCRE macro names to be PCRE1
837
+ + grep: factor test for \0 in grep patterns into a function
838
+ + grep: remove redundant regflags assignments
839
+ + grep: catch a missing enum in switch statement
840
+ + perf: add a comparison test of log --grep regex engines with -F
841
+ + perf: add a comparison test of log --grep regex engines
842
+ + perf: add a comparison test of grep regex engines with -F
843
+ + perf: add a comparison test of grep regex engines
844
+ + perf: emit progress output when unpacking & building
845
+ + perf: add a GIT_PERF_MAKE_COMMAND for when *_MAKE_OPTS won't do
846
+ + grep: add tests to fix blind spots with \0 patterns
847
+ + grep: prepare for testing binary regexes containing rx metacharacters
848
+ + grep: add a test helper function for less verbose -f \0 tests
849
+ + grep: add tests for grep pattern types being passed to submodules
850
+ + grep: amend submodule recursion test for regex engine testing
851
+ + grep: add tests for --threads=N and grep.threads
852
+ + grep: change non-ASCII -i test to stop using --debug
853
+ + grep: add a test for backreferences in PCRE patterns
854
+ + grep: add a test asserting that --perl-regexp dies when !PCRE
855
+ + log: make --regexp-ignore-case work with --perl-regexp
856
+ + log: add exhaustive tests for pattern style options & config
857
+ + test-lib: rename the LIBPCRE prerequisite to PCRE
858
+ + grep & rev-list doc: stop promising libpcre for --perl-regexp
859
+ + Makefile & configure: reword inaccurate comment about PCRE
860
+ (this branch is used by ab/pcre-v2.)
861
862
The internal implementation of "git grep" has seen some clean-up.
863
686
- Will merge to 'next'.
864
+ Will merge to 'master'.
865
866
867
* dt/unpack-save-untracked-cache-extension (2017-05-20) 1 commit
@@ -716,140 +894,6 @@ of the repositories listed at
894
cf. <xmqqa864kops.fsf@gitster.mtv.corp.google.com>
895
896
719
-* sb/checkout-recurse-submodules (2017-05-04) 3 commits
720
- (merged to 'next' on 2017-05-20 at f972b2069f)
721
- + submodule: properly recurse for read-tree and checkout
722
- + submodule: avoid auto-discovery in new working tree manipulator code
723
- + submodule_move_head: reuse child_process structure for futher commands
724
-
725
- "git checkout --recurse-submodules" did not quite work with a
726
- submodule that itself has submodules.
727
-
728
- Will merge to 'master'.
729
-
730
-
731
-* ja/do-not-ask-needless-questions (2017-05-12) 3 commits
732
- (merged to 'next' on 2017-05-23 at 6b373cd892)
733
- + git-filter-branch: be more direct in an error message
734
- + read-tree -m: make error message for merging 0 trees less smart aleck
735
- + usability: don't ask questions if no reply is required
736
-
737
- Git sometimes gives an advice in a rhetorical question that does
738
- not require an answer, which can confuse new users and non native
739
- speakers. Attempt to rephrase them.
740
-
741
- Will merge to 'master'.
742
-
743
-
744
-* bw/dir-c-stops-relying-on-the-index (2017-05-06) 14 commits
745
- (merged to 'next' on 2017-05-20 at 1f1b764ec8)
746
- + dir: convert fill_directory to take an index
747
- + dir: convert read_directory to take an index
748
- + dir: convert read_directory_recursive to take an index
749
- + dir: convert open_cached_dir to take an index
750
- + dir: convert is_excluded to take an index
751
- + dir: convert prep_exclude to take an index
752
- + dir: convert add_excludes to take an index
753
- + dir: convert is_excluded_from_list to take an index
754
- + dir: convert last_exclude_matching_from_list to take an index
755
- + dir: convert dir_add* to take an index
756
- + dir: convert get_dtype to take index
757
- + dir: convert directory_exists_in_index to take index
758
- + dir: convert read_skip_worktree_file_from_index to take an index
759
- + dir: stop using the index compatibility macros
760
-
761
- API update.
762
-
763
- Will merge to 'master'.
764
-
765
-
766
-* jk/diff-submodule-diff-inline (2017-05-08) 1 commit
767
- (merged to 'next' on 2017-05-20 at da7eb2626a)
768
- + diff: recurse into nested submodules for inline diff
769
-
770
- "git diff --submodule=diff" now recurses into nested submodules.
771
-
772
- Will merge to 'master'.
773
-
774
-
775
-* jk/disable-pack-reuse-when-broken (2017-05-09) 2 commits
776
- (merged to 'next' on 2017-05-20 at 684e921273)
777
- + t5310: fix "; do" style
778
- + pack-objects: disable pack reuse for object-selection options
779
-
780
- "pack-objects" can stream a slice of an existing packfile out when
781
- the pack bitmap can tell that the reachable objects are all needed
782
- in the output, without inspecting individual objects. This
783
- strategy however would not work well when "--local" and other
784
- options are in use, and need to be disabled.
785
-
786
- Will merge to 'master'.
787
-
788
-
789
-* js/eol-on-ourselves (2017-05-10) 6 commits
790
- (merged to 'next' on 2017-05-20 at 8023c277d2)
791
- + t4051: mark supporting files as requiring LF-only line endings
792
- + Fix the remaining tests that failed with core.autocrlf=true
793
- + t3901: move supporting files into t/t3901/
794
- + completion: mark bash script as LF-only
795
- + git-new-workdir: mark script as LF-only
796
- + Fix build with core.autocrlf=true
797
-
798
- Make sure our tests would pass when the sources are checked out
799
- with "platform native" line ending convention by default on
800
- Windows. Some "text" files out tests use and the test scripts
801
- themselves that are meant to be run with /bin/sh, ought to be
802
- checked out with eol=LF even on Windows.
803
-
804
- Will merge to 'master'.
805
-
806
-
807
-* nd/split-index-unshare (2017-05-08) 2 commits
808
- (merged to 'next' on 2017-05-20 at 9048cdc2ce)
809
- + p3400: add perf tests for rebasing many changes
810
- + split-index: add and use unshare_split_index()
811
-
812
- Plug some leaks and updates internal API used to implement the
813
- split index feature to make it easier to avoid such a leak in the
814
- future.
815
-
816
- Will merge to 'master'.
817
-
818
-
819
-* js/plug-leaks (2017-05-09) 26 commits
820
- (merged to 'next' on 2017-05-20 at fb136ea2dc)
821
- + checkout: fix memory leak
822
- + submodule_uses_worktrees(): plug memory leak
823
- + show_worktree(): plug memory leak
824
- + name-rev: avoid leaking memory in the `deref` case
825
- + remote: plug memory leak in match_explicit()
826
- + add_reflog_for_walk: avoid memory leak
827
- + shallow: avoid memory leak
828
- + line-log: avoid memory leak
829
- + receive-pack: plug memory leak in update()
830
- + fast-export: avoid leaking memory in handle_tag()
831
- + mktree: plug memory leaks reported by Coverity
832
- + pack-redundant: plug memory leak
833
- + setup_discovered_git_dir(): plug memory leak
834
- + setup_bare_git_dir(): help static analysis
835
- + split_commit_in_progress(): simplify & fix memory leak
836
- + checkout: fix memory leak
837
- + cat-file: fix memory leak
838
- + mailinfo & mailsplit: check for EOF while parsing
839
- + status: close file descriptor after reading git-rebase-todo
840
- + difftool: address a couple of resource/memory leaks
841
- + get_mail_commit_oid(): avoid resource leak
842
- + git_config_rename_section_in_file(): avoid resource leak
843
- + add_commit_patch_id(): avoid allocating memory unnecessarily
844
- + winansi: avoid buffer overrun
845
- + winansi: avoid use of uninitialized value
846
- + mingw: avoid memory leak when splitting PATH
847
-
848
- Fix memory leaks pointed out by Coverity (and people).
849
-
850
- Will merge to 'master'.
851
-
852
-
897
* jh/close-index-before-stat (2017-04-28) 1 commit
898
(merged to 'next' on 2017-05-16 at 0c0372eb02)
899
+ read-cache: close index.lock in do_write_index
@@ -865,124 +909,41 @@ of the repositories listed at
909
910
911
* mb/diff-default-to-indent-heuristics (2017-05-09) 4 commits
868
- - add--interactive: drop diff.indentHeuristic handling
869
- - diff: enable indent heuristic by default
870
- - diff: have the diff-* builtins configure diff before initializing revisions
871
- - diff: make the indent heuristic part of diff's basic configuration
912
+ (merged to 'next' on 2017-05-29 at 7645575e21)
913
+ + add--interactive: drop diff.indentHeuristic handling
914
+ + diff: enable indent heuristic by default
915
+ + diff: have the diff-* builtins configure diff before initializing revisions
916
+ + diff: make the indent heuristic part of diff's basic configuration
917
918
Make the "indent" heuristics the default in "diff" and diff.indentHeuristics
919
configuration variable an escape hatch for those who do no want it.
920
876
- Will merge to 'next'.
921
+ Will merge to 'master'.
922
Looks like there may be follow-on work possible, but the topic
923
itself is in good order.
924
cf. <20170509033258.q7auj5p5lxs4tggd@sigill.intra.peff.net>
925
926
882
-* jc/repack-threads (2017-04-27) 1 commit
883
- (merged to 'next' on 2017-05-20 at c1fd54a2a4)
884
- + repack: accept --threads=<n> and pass it down to pack-objects
885
-
886
- "git repack" learned to accept the --threads=<n> option and pass it
887
- to pack-objects.
888
-
889
- Will merge to 'master'.
890
-
891
-
892
-* nd/fopen-errors (2017-05-09) 23 commits
893
- - t1308: add a test case on open a config directory
894
- - config.c: handle error on failing to fopen()
895
- - xdiff-interface.c: report errno on failure to stat() or fopen()
896
- - wt-status.c: report error on failure to fopen()
897
- - server-info: report error on failure to fopen()
898
- - sequencer.c: report error on failure to fopen()
899
- - rerere.c: report correct errno
900
- - rerere.c: report error on failure to fopen()
901
- - remote.c: report error on failure to fopen()
902
- - commit.c: report error on failure to fopen() the graft file
927
+* nd/fopen-errors (2017-05-26) 13 commits
928
+ - is_missing_file_error(): work around EINVAL on Windows
929
+ - wrapper: factor out is_missing_file_error()
930
- log: fix memory leak in open_next_file()
904
- - log: report errno on failure to fopen() a file
905
- - blame: report error on open if graft_file is a directory
906
- - bisect: report on fopen() error
907
- - ident.c: use fopen_or_warn()
908
- - attr.c: use fopen_or_warn()
909
- - wrapper.c: add fopen_or_warn()
910
- - wrapper.c: add warn_on_fopen_errors()
931
+ - rerere.c: move error_errno() closer to the source system call
932
+ - print errno when reporting a system call error
933
+ - wrapper.c: make warn_on_inaccessible() static
934
+ - wrapper.c: add and use fopen_or_warn()
935
+ - wrapper.c: add and use warn_on_fopen_errors()
936
- config.mak.uname: set FREAD_READS_DIRECTORIES for Darwin, too
937
- config.mak.uname: set FREAD_READS_DIRECTORIES for Linux and FreeBSD
938
- clone: use xfopen() instead of fopen()
914
- - Use xfopen() in more places
939
+ - use xfopen() in more places
940
- git_fopen: fix a sparse 'not declared' warning
941
942
We often try to open a file for reading whose existence is
943
optional, and silently ignore errors from open/fopen; report such
944
errors if they are not due to missing files.
945
921
- Will merge to 'next'.
922
- Looks like this is now in a good shape.
923
- cf. <20170503101706.9223-1-pclouds@gmail.com>
924
-
925
-
926
-* bc/object-id (2017-05-08) 53 commits
927
- (merged to 'next' on 2017-05-20 at e7372733fb)
928
- + object: convert parse_object* to take struct object_id
929
- + tree: convert parse_tree_indirect to struct object_id
930
- + sequencer: convert do_recursive_merge to struct object_id
931
- + diff-lib: convert do_diff_cache to struct object_id
932
- + builtin/ls-tree: convert to struct object_id
933
- + merge: convert checkout_fast_forward to struct object_id
934
- + sequencer: convert fast_forward_to to struct object_id
935
- + builtin/ls-files: convert overlay_tree_on_cache to object_id
936
- + builtin/read-tree: convert to struct object_id
937
- + sha1_name: convert internals of peel_onion to object_id
938
- + upload-pack: convert remaining parse_object callers to object_id
939
- + revision: convert remaining parse_object callers to object_id
940
- + revision: rename add_pending_sha1 to add_pending_oid
941
- + http-push: convert process_ls_object and descendants to object_id
942
- + refs/files-backend: convert many internals to struct object_id
943
- + refs: convert struct ref_update to use struct object_id
944
- + ref-filter: convert some static functions to struct object_id
945
- + Convert struct ref_array_item to struct object_id
946
- + Convert the verify_pack callback to struct object_id
947
- + Convert lookup_tag to struct object_id
948
- + log-tree: convert to struct object_id
949
- + Convert lookup_tree to struct object_id
950
- + builtin/reflog: convert tree_is_complete to take struct object_id
951
- + tree: convert read_tree_1 to use struct object_id internally
952
- + Convert lookup_blob to struct object_id
953
- + Convert remaining callers of lookup_blob to object_id
954
- + builtin/unpack-objects: convert to struct object_id
955
- + pack: convert struct pack_idx_entry to struct object_id
956
- + Convert lookup_commit* to struct object_id
957
- + Convert remaining callers of lookup_commit_reference* to object_id
958
- + builtin/tag: convert to struct object_id
959
- + sequencer: convert some functions to struct object_id
960
- + shallow: convert shallow registration functions to object_id
961
- + revision: convert prepare_show_merge to struct object_id
962
- + notes-utils: convert internals to struct object_id
963
- + http-push: convert some static functions to struct object_id
964
- + tag: convert parse_tag_buffer to struct object_id
965
- + builtin/verify-commit: convert to struct object_id
966
- + reflog_expire: convert to struct object_id
967
- + parse-options-cb: convert to struct object_id
968
- + notes-cache: convert to struct object_id
969
- + submodule: convert merge_submodule to use struct object_id
970
- + fast-import: convert to struct object_id
971
- + fast-import: convert internal structs to struct object_id
972
- + builtin/rev-parse: convert to struct object_id
973
- + builtin/blame: convert static function to struct object_id
974
- + branch: convert to struct object_id
975
- + bundle: convert to struct object_id
976
- + builtin/prune: convert to struct object_id
977
- + builtin/name-rev: convert to struct object_id
978
- + Convert struct cache_tree to use struct object_id
979
- + Clean up outstanding object_id transforms.
980
- + fetch-pack: convert to struct object_id
981
- (this branch is used by mh/packed-ref-store-prep.)
982
-
983
- Conversion from uchar[20] to struct object_id continues.
984
-
985
- Will merge to 'master'.
946
+ Waiting for comments on the two changes at the tip from Windows folks.
947
948
949
* js/rebase-i-final (2017-05-02) 10 commits
@@ -1000,7 +961,9 @@ of the repositories listed at
961
The final batch to "git rebase -i" updates to move more code from
962
the shell script to C.
963
1003
- Needs review.
964
+ Undecided; expecting a reroll.
965
+ This is at its v4.
966
+ cf. <cover.1493414945.git.johannes.schindelin@gmx.de>
967
968
969
* bw/forking-and-threading (2017-05-15) 14 commits
@@ -1026,19 +989,6 @@ of the repositories listed at
989
Will merge to 'master'.
990
991
1029
-* sb/reset-recurse-submodules (2017-04-23) 4 commits
1030
- (merged to 'next' on 2017-05-20 at 8adafdf1be)
1031
- + builtin/reset: add --recurse-submodules switch
1032
- + submodule.c: submodule_move_head works with broken submodules
1033
- + submodule.c: uninitialized submodules are ignored in recursive commands
1034
- + entry.c: submodule recursing: respect force flag correctly
1035
- (this branch is used by sb/submodule-blanket-recursive.)
1036
-
1037
- "git reset" learned "--recurse-submodules" option.
1038
-
1039
- Will merge to 'master'.
1040
-
1041
-
992
* bp/sub-process-convert-filter (2017-05-15) 11 commits
993
(merged to 'next' on 2017-05-23 at 89f5420a82)
994
+ convert: update subprocess_read_status() to not die on EOF
@@ -1137,3 +1087,37 @@ of the repositories listed at
1087
Will discard.
1088
This was more of "the world could have been like this" illustration
1089
rather than a "let's make this change" proposal.
1090
+
1091
+
1092
+* sb/submodule-blanket-recursive (2017-05-23) 6 commits
1093
+ . builtin/push.c: respect 'submodule.recurse' option
1094
+ . builtin/grep.c: respect 'submodule.recurse' option
1095
+ . builtin/fetch.c: respect 'submodule.recurse' option
1096
+ . Introduce submodule.recurse option for worktree manipulators
1097
+ . submodule test invocation: only pass additional arguments
1098
+ . submodule.c: add has_submodules to check if we have any submodules
1099
+
1100
+ A new configuration variable "submodule.recurse" can be set to true
1101
+ to force various commands run at the top-level superproject to
1102
+ behave as if they were invoked with the "--recurse-submodules"
1103
+ option.
1104
+
1105
+ Retracted for now.
1106
+ cf. <CAGZ79kZexcwh=E6_ks83=pJh=ZvKnLvJ54eLsn+HURsTZOpvqg@mail.gmail.com>
1107
+
1108
+
1109
+* df/dir-iter-remove-subtree (2017-05-29) 5 commits
1110
+ . remove_subtree(): reimplement using iterators
1111
+ . dir_iterator: rewrite state machine model
1112
+ . dir_iterator: refactor dir_iterator_advance
1113
+ . remove_subtree(): test removing nested directories
1114
+ . dir_iterator: add tests for dir_iterator API
1115
+
1116
+ Update the dir-iterator API and use it to reimplement
1117
+ remove_subtree().
1118
+
1119
+ GSoC microproject.
1120
+ Ejected as it conflicts with other topics in flight in a
1121
+ non-trivial way.
1122
+
1123
+