What's cooking (2012/08 #09)
Junio C Hamano committed
Aug 29, 2012 at 15:58 UTC
69cd2dbff443093d98f90a3ce48902f4ecd91fe0
1 file changed
+284
-359
whats-cooking.txt
+284
-359
@@ -1,20 +1,16 @@
1
To: git@vger.kernel.org
2
Bcc: lwn@lwn.net
3
-Subject: What's cooking in git.git (Aug 2012, #08; Mon, 27)
4
-X-master-at: 31e0100e89c3f7c05bd7fdf54e084b2039d398ec
5
-X-next-at: 32a1b722418b8b32eafe3b98d890fa239ec128f8
3
+Subject: What's cooking in git.git (Aug 2012, #09; Wed, 29)
4
+X-master-at: 16d26b168b371b2f4f86b1adb61470c6b08b27b9
5
+X-next-at: 9b7ff7f6c17d9531b4ebeb199a70a759872c7d4f
6
7
-What's cooking in git.git (Aug 2012, #08; Mon, 27)
7
+What's cooking in git.git (Aug 2012, #09; Wed, 29)
8
--------------------------------------------------
9
10
Here are the topics that have been cooking. Commits prefixed with '-' are
11
only in 'pu' (proposed updates) while commits prefixed with '+' are in 'next'.
12
13
-New topics started to flow in at the pace a bit faster than
14
-reviewers give comments on them. I've merged half of the second
15
-batch of topics to 'master', and will do the remainder later this
16
-week. They are a bit more substantial changes than the first batch
17
-we merged earlier.
13
+The latter half of the second batch of topics are now in 'master'.
14
15
I'm planning to keep this cycle reasonably short and aim for tagging
16
the result as 1.8.0 at the end of 9th week, on October 21st, after
@@ -29,108 +25,233 @@ repositories listed at
25
--------------------------------------------------
26
[Graduated to "master"]
27
32
-* ab/diff-write-incomplete-line (2012-08-05) 1 commit
33
- (merged to 'next' on 2012-08-20 at e28b804)
34
- + Fix '\ No newline...' annotation in rewrite diffs
35
-
36
-(Originally merged to 'next' on 2012-08-07)
37
-
38
-The output from "git diff -B" for a file that ends with an
39
-incomplete line did not put "\ No newline..." on a line of its own.
40
-
41
-* bc/receive-pack-stdout-protection (2012-08-06) 2 commits
42
- (merged to 'next' on 2012-08-20 at b1b0d18)
43
- + receive-pack: do not leak output from auto-gc to standard output
44
- + t/t5400: demonstrate breakage caused by informational message from prune
45
-
46
-(Originally merged to 'next' on 2012-08-07)
47
-
48
-When "git push" triggered the automatic gc on the receiving end, a
49
-message from "git prune" that said it was removing cruft leaked to
50
-the standard output, breaking the communication protocol.
51
-
52
-* da/difftool-updates (2012-08-21) 13 commits
53
- (merged to 'next' on 2012-08-21 at 332afe4)
54
- + difftool: silence warning
55
- (merged to 'next' on 2012-08-20 at 2fecd90)
56
- + Add Code Compare v2.80.4 as a merge / diff tool for Windows
57
- + mergetool,difftool: Document --tool-help consistently
58
- + difftool: Disable --symlinks on cygwin
59
- + difftool: Handle compare() returning -1
60
- + difftool: Wrap long lines for readability
61
- + difftool: Check all return codes from compare()
62
- + difftool: Handle finding mergetools/ in a path with spaces
63
- + difftool: Use symlinks when diffing against the worktree
64
- + difftool: Call the temp directory "git-difftool"
65
- + difftool: Move option values into a hash
66
- + difftool: Eliminate global variables
67
- + difftool: Simplify print_tool_help()
28
+* bc/prune-info (2012-08-07) 1 commit
29
+ (merged to 'next' on 2012-08-20 at 1bc9e66)
30
+ + prune.c: only print informational message in show_only or verbose mode
31
69
-(Originally merged to 'next' on 2012-08-10)
32
+Teach "git prune" without "-v" to be silent about leftover temporary files.
33
71
-"git difftool --dir-diff" learned to use symbolic links to prepare
72
-temporary copy of the working tree when available.
34
+* ef/win32-cred-helper (2012-08-16) 1 commit
35
+ (merged to 'next' on 2012-08-20 at bd5c651)
36
+ + contrib: add win32 credential-helper
37
+ (this branch is used by ph/credential-refactor.)
38
74
-* jc/maint-t7406-rev-parse-max-count-huh (2012-07-30) 1 commit
75
- (merged to 'next' on 2012-08-20 at 3e2598a)
76
- + t7406: fix misleading "rev-parse --max-count=1 HEAD"
39
+Credential helper for Win32 to allow access to the keychain of
40
+the logged-in user.
41
78
-(Originally merged to 'next' on 2012-07-30)
42
+* hv/submodule-path-unmatch (2012-08-14) 1 commit
43
+ (merged to 'next' on 2012-08-20 at cbe6483)
44
+ + Let submodule command exit with error status if path does not exist
45
80
-Test clean-up, with no behaviour change.
46
+As the title says.
47
82
-* jk/docs-docbook-monospace-display (2012-08-07) 1 commit
83
- (merged to 'next' on 2012-08-20 at 0ed30dc)
84
- + docs: monospace listings in docbook output
48
+* jc/capabilities (2012-08-13) 3 commits
49
+ (merged to 'next' on 2012-08-20 at 1976e47)
50
+ + fetch-pack: mention server version with verbose output
51
+ + parse_feature_request: make it easier to see feature values
52
+ + fetch-pack: do not ask for unadvertised capabilities
53
+ (this branch uses jk/version-string.)
54
86
-(Originally merged to 'next' on 2012-08-18)
55
+Some capabilities were asked by fetch-pack even when upload-pack did
56
+not advertise that they are available. Fix fetch-pack not to do so.
57
88
-The documentation in the TeXinfo format was using indented output
89
-for materials meant to be examples that are better typeset in
90
-monospace.
58
+* jc/maint-protect-sh-from-ifs (2012-08-08) 1 commit
59
+ (merged to 'next' on 2012-08-20 at 8b00562)
60
+ + sh-setup: protect from exported IFS
61
92
-* jk/maint-diff-nul40-as-sentinel (2012-07-29) 1 commit
93
- + diff: do not use null sha1 as a sentinel value
94
- (this branch is used by jk/maint-null-in-trees.)
62
+When the user exports a non-default IFS without HT, scripts that
63
+rely on being able to parse "ls-files -s | while read a b c..."
64
+start to fail. Protect them from such a misconfiguration.
65
96
-"git diff" had a confusion between taking data from a path in the
97
-working tree and taking data from an object that happens to have
98
-name 0{40} recorded in a tree.
66
+* jk/version-string (2012-08-10) 3 commits
67
+ (merged to 'next' on 2012-08-20 at 6d207bb)
68
+ + do not send client agent unless server does first
69
+ + send-pack: fix capability-sending logic
70
+ + include agent identifier in capability string
71
+ (this branch is used by jc/capabilities.)
72
100
-* jk/maint-null-in-trees (2012-07-29) 2 commits
101
- (merged to 'next' on 2012-08-20 at 5bc2c49)
102
- + fsck: detect null sha1 in tree entries
103
- + do not write null sha1s to on-disk index
104
- (this branch uses jk/maint-diff-nul40-as-sentinel.)
73
+(Originally merged to 'next' on 2012-08-10)
74
106
-(Originally merged to 'next' on 2012-07-30)
75
+Leftover bits from the "git version" code restructuring, that give
76
+us an ability to expose the Git version over the wire in the
77
+protocol exchange. But we do so carefully only to the other end that
78
+understands it.
79
108
-We do not want a link to 0{40} object stored anywhere in our objects.
80
+* mg/rebase-i-onto-reflog-in-full (2012-08-10) 1 commit
81
+ (merged to 'next' on 2012-08-20 at f671bd2)
82
+ + rebase -i: use full onto sha1 in reflog
83
110
-* js/grep-patterntype-config (2012-08-03) 1 commit
111
- (merged to 'next' on 2012-08-20 at 39a0431)
112
- + grep: add a grep.patternType configuration setting
84
+The reflog entries left by "git rebase" and "git rebase -i" were
85
+inconsistent.
86
114
-(Originally merged to 'next' on 2012-08-09)
87
+* mh/maint-config-doc-proxy-command (2012-08-18) 2 commits
88
+ (merged to 'next' on 2012-08-20 at a484625)
89
+ + git-config doc: unconfuse an example
90
+ + git-config.txt: fix example
91
116
-"grep" learned to use a non-standard pattern type by default if a
117
-configuration variable tells it to.
92
+A minor documentation update.
93
119
-* mm/push-default-switch-warning (2012-06-24) 1 commit
120
- (merged to 'next' on 2012-08-20 at 553e6c0)
121
- + push: start warning upcoming default change for push.default
94
+* mz/cherry-code-cleanup (2012-07-29) 3 commits
95
+ (merged to 'next' on 2012-08-20 at bfb0457)
96
+ + cherry: remove redundant check for merge commit
97
+ + cherry: don't set ignored rev_info options
98
+ + remove unnecessary parameter from get_patch_ids()
99
123
-(Originally merged to 'next' on 2012-06-26)
100
+Minor code clean-up on the cherry-pick codepath.
101
+
102
+* mz/empty-rebase-test (2012-08-09) 1 commit
103
+ (merged to 'next' on 2012-08-20 at 53e2419)
104
+ + add tests for 'git rebase --keep-empty'
105
+
106
+* rr/precompose-utf8-cleanup (2012-08-20) 2 commits
107
+ (merged to 'next' on 2012-08-20 at a883c2a)
108
+ + precompose-utf8: do not call checks for non-ascii "utf8"
109
+ + cleanup precompose_utf8
110
125
-In the next major release, we will switch "git push [$there]" that
126
-does not say what to push from the traditional "matching" to the
127
-updated "simple" semantics, that pushes the current branch to the
128
-branch with the same name only when the current branch is set to
129
-integrate with that remote branch (all other cases will error out).
111
+A style and code cleanup.
112
113
--------------------------------------------------
114
[New Topics]
115
116
+* jk/maint-quiet-is-synonym-to-s-in-log (2012-08-28) 1 commit
117
+ - log: fix --quiet synonym for -s
118
+
119
+We tried to bend backwards to allow "--quiet" to be a synonym as
120
+"-s" when given as e.g. "git show --quiet", but did not quite
121
+succeed.
122
+
123
+Will merge to 'next'.
124
+
125
+* nd/checkout-option-parsing-fix (2012-08-29) 3 commits
126
+ - checkout: reorder option handling
127
+ - checkout: move more parameters to struct checkout_opts
128
+ - checkout: pass "struct checkout_opts *" as const pointer
129
+
130
+The option parsing of "git checkout" had error checking, dwim and
131
+defaulting missing options, all mixed in the code, and issuing an
132
+appropriate error message with useful context was getting harder.
133
+Reorganize the code and allow giving a proper diagnosis when the
134
+user says "git checkout -b -t foo bar" (e.g. "-t" is not a good name
135
+for a branch).
136
+
137
+* jc/maint-name-rev (2012-08-29) 3 commits
138
+ - name-rev: --weight option (WIP)
139
+ - name_rev: clarify when a new tip-name is assigned to a commit
140
+ - name-rev: lose unnecessary typedef
141
+
142
+"git name-rev" names the given revision based on a ref that can be
143
+reached in the smallest number of steps from the rev, but that is
144
+not useful when the caller wants to know which tag is the oldest one
145
+that contains the rev. This teaches a new mode to the command that
146
+uses the oldest ref among those which contain the rev.
147
+
148
+--------------------------------------------------
149
+[Stalled]
150
+
151
+* er/doc-fast-import-done (2012-08-22) 1 commit
152
+ - fast-import: document the --done option
153
+
154
+Parked in 'pu' in case ESR responds with "Sorry, forgot to sign-off".
155
+
156
+* jc/add-delete-default (2012-08-13) 1 commit
157
+ - git add: notice removal of tracked paths by default
158
+
159
+"git add dir/" updated modified files and added new files, but does
160
+not notice removed files, which may be "Huh?" to some users. They
161
+can of course use "git add -A dir/", but why should they?
162
+
163
+Resurrected from graveyard, as I thought it was a worthwhile thing
164
+to do in the longer term; waiting for comments.
165
+
166
+* tx/relative-in-the-future (2012-08-16) 2 commits
167
+ - date: show relative dates in the future
168
+ - date: refactor the relative date logic from presentation
169
+
170
+Not my itch; rewritten an earlier submission by Tom Xue into
171
+somewhat more maintainable form, though it breaks existing i18n.
172
+
173
+Anybody interested in fixing it up? Otherwise may discard.
174
+
175
+* pw/p4-submit-conflicts (2012-08-16) 12 commits
176
+ - git p4: add submit --prepare-p4-only option
177
+ - git p4: add submit --dry-run option
178
+ - git p4: accept -v for --verbose
179
+ - git p4: revert deleted files after submit cancel
180
+ - git p4: rearrange submit template construction
181
+ - git p4: test clean-up after failed submit, fix added files
182
+ - git p4: standardize submit cancel due to unchanged template
183
+ - git p4: move conflict prompt into run, use [c]ontinue and [q]uit
184
+ - git p4: remove submit failure options [a]pply and [w]rite
185
+ - git p4: gracefully fail if some commits could not be applied
186
+ - git p4 test: use p4d -L option to suppress log messages
187
+ - git p4 test: remove bash-ism of combined export/assignment
188
+
189
+Waiting for the dust to settle.
190
+
191
+* mz/rebase-range (2012-07-18) 7 commits
192
+ . rebase (without -p): correctly calculate patches to rebase
193
+ . rebase -p: don't request --left-right only to ignore left side
194
+ . rebase -p: use --cherry-mark for todo file
195
+ . git-rebase--interactive.sh: look up subject in add_pick_line
196
+ . git-rebase--interactive: group all $preserve_merges code
197
+ . git-rebase--interactive.sh: extract function for adding "pick" line
198
+ . git-rebase--am.sh: avoid special-casing --keep-empty
199
+
200
+Expecting a reroll.
201
+
202
+Performance concerns from Windows folks. Also the series lacks
203
+proper sign-offs.
204
+
205
+* mb/remote-default-nn-origin (2012-07-11) 6 commits
206
+ - Teach get_default_remote to respect remote.default.
207
+ - Test that plain "git fetch" uses remote.default when on a detached HEAD.
208
+ - Teach clone to set remote.default.
209
+ - Teach "git remote" about remote.default.
210
+ - Teach remote.c about the remote.default configuration setting.
211
+ - Rename remote.c's default_remote_name static variables.
212
+
213
+When the user does not specify what remote to interact with, we
214
+often attempt to use 'origin'. This can now be customized via a
215
+configuration variable.
216
+
217
+Expecting a reroll.
218
+
219
+"The first remote becomes the default" bit is better done as a
220
+separate step.
221
+
222
+* jc/split-blob (2012-04-03) 6 commits
223
+ - chunked-object: streaming checkout
224
+ - chunked-object: fallback checkout codepaths
225
+ - bulk-checkin: support chunked-object encoding
226
+ - bulk-checkin: allow the same data to be multiply hashed
227
+ - new representation types in the packstream
228
+ - packfile: use varint functions
229
+
230
+Not ready.
231
+
232
+I finished the streaming checkout codepath, but as explained in
233
+127b177 (bulk-checkin: support chunked-object encoding, 2011-11-30),
234
+these are still early steps of a long and painful journey. At least
235
+pack-objects and fsck need to learn the new encoding for the series
236
+to be usable locally, and then index-pack/unpack-objects needs to
237
+learn it to be used remotely.
238
+
239
+Given that I heard a lot of noise that people want large files, and
240
+that I was asked by somebody at GitTogether'11 privately for an
241
+advice on how to pay developers (not me) to help adding necessary
242
+support, I am somewhat dissapointed that the original patch series
243
+that was sent long time ago still remains here without much comments
244
+and updates from the developer community. I even made the interface
245
+to the logic that decides where to split chunks easily replaceable,
246
+and I deliberately made the logic in the original patch extremely
247
+stupid to entice others, especially the "bup" fanbois, to come up
248
+with a better logic, thinking that giving people an easy target to
249
+shoot for, they may be encouraged to help out. The plan is not
250
+working :-<.
251
+
252
+--------------------------------------------------
253
+[Cooking]
254
+
255
* cn/branch-set-upstream-to (2012-08-23) 3 commits
256
- branch: deprecate --set-upstream and show help if we detect possible mistaken use
257
- branch: add --unset-upstream option
@@ -143,24 +264,24 @@ the current branch. With a plan to deprecate this option, introduce
264
current branch to integrate with 'origin/master' remote tracking
265
branch.
266
146
-Wainting for a reroll
267
+Waiting for a reroll
268
that tightens its error checking.
269
270
* jc/dotdot-is-parent-directory (2012-08-23) 1 commit
150
- - specifying ranges: we did not mean to make ".." an empty set
271
+ (merged to 'next' on 2012-08-28 at 2f2ae87)
272
+ + specifying ranges: we did not mean to make ".." an empty set
273
274
"git log .." errored out saying it is both rev range and a path when
275
there is no disambiguating "--" is on the command line. Update the
276
command line parser to interpret ".." as a path in such a case.
277
156
-Will merge to 'next'.
278
+Will merge to 'master' as part of the fourth batch.
279
280
* jc/maint-mergetool-style-fix (2012-08-23) 1 commit
159
- - mergetool: style fixes
160
-
161
-As the title says.
281
+ (merged to 'next' on 2012-08-28 at ced5868)
282
+ + mergetool: style fixes
283
163
-Will merge to 'next'.
284
+Will merge to 'master' as part of the third batch.
285
286
* js/compat-itimer (2012-08-24) 1 commit
287
- Support for setitimer() on platforms lacking it
@@ -168,25 +289,31 @@ Will merge to 'next'.
289
Pieces to support compilation on __TANDEM.
290
291
* js/compat-mkdir (2012-08-24) 1 commit
171
- - compat: some mkdir() do not like a slash at the end
292
+ (merged to 'next' on 2012-08-28 at e92bf74)
293
+ + compat: some mkdir() do not like a slash at the end
294
173
-Pieces to support compilation on __TANDEM.
295
+Some mkdir(2) implementations do not want to see trailing slash in
296
+its parameter.
297
175
-Will merge to 'next'.
298
+Will merge to 'master' as part of the third batch.
299
300
* js/no-curl-easy-strerror-on-old-curl (2012-08-23) 1 commit
178
- - http.c: don't use curl_easy_strerror prior to curl-7.12.0
301
+ (merged to 'next' on 2012-08-28 at 25e2e7a)
302
+ + http.c: don't use curl_easy_strerror prior to curl-7.12.0
303
180
-Pieces to support compilation on __TANDEM.
304
+Done to support compilation on __TANDEM, but is independently useful
305
+for people with older version of libcURL.
306
182
-Will merge to 'next'.
307
+Will merge to 'master' as part of the third batch.
308
309
* js/use-sc-open-max (2012-08-24) 1 commit
185
- - sha1_file.c: introduce get_max_fd_limit() helper
310
+ (merged to 'next' on 2012-08-28 at 8b9ac7c)
311
+ + sha1_file.c: introduce get_max_fd_limit() helper
312
187
-Pieces to support compilation on __TANDEM.
313
+Introduce get_max_fd_limit() to absorb platforms that do not have
314
+getrlimit(RLIMIT_NOFILE) and/or sysconf(_SC_OPEN_MAX).
315
189
-Will merge to 'next'.
316
+Will merge to 'master' as part of the third batch.
317
318
* ph/credential-refactor (2012-08-27) 6 commits
319
- wincred: port to generic credential helper (UNTESTED)
@@ -195,7 +322,6 @@ Will merge to 'next'.
322
- gnome-keyring: port to generic helper implementation
323
- contrib: add generic credential helper
324
- contrib: add credential helper for GnomeKeyring
198
- (this branch uses ef/win32-cred-helper.)
325
326
Adds a new credential helper to support Gnome keyring and attempts
327
to refactor to share code among OSX keychain, Gnome keyring and
@@ -206,7 +332,8 @@ to decide whether the series should be rerolled or reduced to only
332
the first one.
333
334
* jc/apply-binary-p0 (2012-08-24) 1 commit
209
- - apply: compute patch->def_name correctly under -p0
335
+ (merged to 'next' on 2012-08-28 at 6fa690c)
336
+ + apply: compute patch->def_name correctly under -p0
337
338
"git apply -p0" did not parse pathnames on "diff --git" line
339
correctly. This caused patches that had pathnames in no other
@@ -216,10 +343,11 @@ mode changes have preimage and postimage pathnames in different
343
places in a form that can be parsed unambiguously and did not suffer
344
from this problem.
345
219
-Will merge to 'next'.
346
+Will merge to 'master' as part of the fourth batch.
347
348
* jc/maint-doc-checkout-b-always-takes-branch-name (2012-08-26) 1 commit
222
- - doc: "git checkout -b/-B/--orphan" always takes a branch name
349
+ (merged to 'next' on 2012-08-29 at e0d0f5c)
350
+ + doc: "git checkout -b/-B/--orphan" always takes a branch name
351
352
The synopsis said "checkout [-B branch]" to make it clear the branch
353
name is a parameter to the option, but the heading for the option
@@ -227,25 +355,24 @@ description was "-B::", not "-B branch::", making the documentation
355
misleading. There may be room in documentation pages of other
356
commands for similar improvements (hint, hint).
357
230
-Will merge to 'next'.
358
+Will merge to 'master' as part of the fourth batch.
359
360
* jk/maint-http-half-auth-push (2012-08-27) 8 commits
233
- - http: prompt for credentials on failed POST
234
- - http: factor out http error code handling
235
- - t: test http access to "half-auth" repositories
236
- - t: test basic smart-http authentication
237
- - t/lib-httpd: recognize */smart/* repos as smart-http
238
- - t/lib-httpd: only route auth/dumb to dumb repos
239
- - t5550: factor out http auth setup
240
- - t5550: put auth-required repo in auth/dumb
361
+ (merged to 'next' on 2012-08-29 at 0ed32f5)
362
+ + http: prompt for credentials on failed POST
363
+ + http: factor out http error code handling
364
+ + t: test http access to "half-auth" repositories
365
+ + t: test basic smart-http authentication
366
+ + t/lib-httpd: recognize */smart/* repos as smart-http
367
+ + t/lib-httpd: only route auth/dumb to dumb repos
368
+ + t5550: factor out http auth setup
369
+ + t5550: put auth-required repo in auth/dumb
370
371
Pushing to smart HTTP server with recent Git fails without having
372
the username in the URL to force authentication, if the server is
373
configured to allow GET anonymously, while requiring authentication
374
for POST.
375
247
-Will merge to 'next'.
248
-
376
* jx/test-real-path (2012-08-27) 1 commit
377
- test: set the realpath of CWD as TRASH_DIRECTORY
378
@@ -253,7 +380,8 @@ Running tests with the "trash" directory elsewhere with the "--root"
380
option did not work well if the directory was specified by a symbolic
381
link pointing at it.
382
256
-Waiting for a Tested-by.
383
+Seems broken as it makes $(pwd) and TRASH_DIRECTORY inconsistent.
384
+Needs rerolling.
385
386
* mh/fetch-filter-refs (2012-08-26) 17 commits
387
- filter_refs(): simplify logic
@@ -278,134 +406,29 @@ Code simplification and clarification.
406
Waiting for a follow-up patch based on Peff's idea.
407
408
* nd/branch-v-alignment (2012-08-27) 1 commit
281
- - branch -v: align even when branch names are in UTF-8
409
+ (merged to 'next' on 2012-08-29 at 5c0d226)
410
+ + branch -v: align even when branch names are in UTF-8
411
412
Output from "git branch -v" contains "(no branch)" that could be
413
localized, but the code to align it along with the names of branches
414
were counting in bytes, not in display columns.
415
287
-Will merge to 'next'.
416
+Will merge to 'master' as part of the fourth batch.
417
289
-* jc/merge-bases (2012-08-27) 5 commits
418
+* jc/merge-bases (2012-08-28) 5 commits
419
- get_merge_bases_many(): walk from many tips in parallel
420
- in_merge_bases(): omit unnecessary redundant common ancestor reduction
421
- http-push: use in_merge_bases() for fast-forward check
422
- receive-pack: use in_merge_bases() for fast-forward check
423
- in_merge_bases(): support only one "other" commit
424
+ (this branch is used by jc/xprm-merge-bases-many.)
425
426
Optimises the "merge-base" computation a bit, and also updates its
427
users that do not need the full merge-base information to call a
428
cheaper subset. The most important one is the second from the tip
429
done by Thomas.
430
301
-Needs sign-off.
302
-
303
---------------------------------------------------
304
-[Stalled]
305
-
306
-* er/doc-fast-import-done (2012-08-22) 1 commit
307
- - fast-import: document the --done option
308
-
309
-Parked in 'pu' in case ESR responds with "Sorry, forgot to sign-off".
310
-
311
-* jc/add-delete-default (2012-08-13) 1 commit
312
- - git add: notice removal of tracked paths by default
313
-
314
-"git add dir/" updated modified files and added new files, but does
315
-not notice removed files, which may be "Huh?" to some users. They
316
-can of course use "git add -A dir/", but why should they?
317
-
318
-Resurrected from graveyard, as I thought it was a worthwhile thing
319
-to do in the longer term; waiting for comments.
320
-
321
-* tx/relative-in-the-future (2012-08-16) 2 commits
322
- - date: show relative dates in the future
323
- - date: refactor the relative date logic from presentation
324
-
325
-Not my itch; rewritten an earlier submission by Tom Xue into
326
-somewhat more maintainable form, though it breaks existing i18n.
327
-
328
-Anybody interested in fixing it up? Otherwise may discard.
329
-
330
-* pw/p4-submit-conflicts (2012-08-16) 12 commits
331
- - git p4: add submit --prepare-p4-only option
332
- - git p4: add submit --dry-run option
333
- - git p4: accept -v for --verbose
334
- - git p4: revert deleted files after submit cancel
335
- - git p4: rearrange submit template construction
336
- - git p4: test clean-up after failed submit, fix added files
337
- - git p4: standardize submit cancel due to unchanged template
338
- - git p4: move conflict prompt into run, use [c]ontinue and [q]uit
339
- - git p4: remove submit failure options [a]pply and [w]rite
340
- - git p4: gracefully fail if some commits could not be applied
341
- - git p4 test: use p4d -L option to suppress log messages
342
- - git p4 test: remove bash-ism of combined export/assignment
343
-
344
-Waiting for the dust to settle.
345
-
346
-* mz/rebase-range (2012-07-18) 7 commits
347
- . rebase (without -p): correctly calculate patches to rebase
348
- . rebase -p: don't request --left-right only to ignore left side
349
- . rebase -p: use --cherry-mark for todo file
350
- . git-rebase--interactive.sh: look up subject in add_pick_line
351
- . git-rebase--interactive: group all $preserve_merges code
352
- . git-rebase--interactive.sh: extract function for adding "pick" line
353
- . git-rebase--am.sh: avoid special-casing --keep-empty
354
-
355
-Expecting a reroll.
356
-
357
-Performance concerns from Windows folks. Also the series lacks
358
-proper sign-offs.
359
-
360
-* mb/remote-default-nn-origin (2012-07-11) 6 commits
361
- - Teach get_default_remote to respect remote.default.
362
- - Test that plain "git fetch" uses remote.default when on a detached HEAD.
363
- - Teach clone to set remote.default.
364
- - Teach "git remote" about remote.default.
365
- - Teach remote.c about the remote.default configuration setting.
366
- - Rename remote.c's default_remote_name static variables.
367
-
368
-When the user does not specify what remote to interact with, we
369
-often attempt to use 'origin'. This can now be customized via a
370
-configuration variable.
371
-
372
-Expecting a reroll.
373
-
374
-"The first remote becomes the default" bit is better done as a
375
-separate step.
376
-
377
-* jc/split-blob (2012-04-03) 6 commits
378
- - chunked-object: streaming checkout
379
- - chunked-object: fallback checkout codepaths
380
- - bulk-checkin: support chunked-object encoding
381
- - bulk-checkin: allow the same data to be multiply hashed
382
- - new representation types in the packstream
383
- - packfile: use varint functions
384
-
385
-Not ready.
386
-
387
-I finished the streaming checkout codepath, but as explained in
388
-127b177 (bulk-checkin: support chunked-object encoding, 2011-11-30),
389
-these are still early steps of a long and painful journey. At least
390
-pack-objects and fsck need to learn the new encoding for the series
391
-to be usable locally, and then index-pack/unpack-objects needs to
392
-learn it to be used remotely.
393
-
394
-Given that I heard a lot of noise that people want large files, and
395
-that I was asked by somebody at GitTogether'11 privately for an
396
-advice on how to pay developers (not me) to help adding necessary
397
-support, I am somewhat dissapointed that the original patch series
398
-that was sent long time ago still remains here without much comments
399
-and updates from the developer community. I even made the interface
400
-to the logic that decides where to split chunks easily replaceable,
401
-and I deliberately made the logic in the original patch extremely
402
-stupid to entice others, especially the "bup" fanbois, to come up
403
-with a better logic, thinking that giving people an easy target to
404
-shoot for, they may be encouraged to help out. The plan is not
405
-working :-<.
406
-
407
---------------------------------------------------
408
-[Cooking]
431
+Will merge to 'next'.
432
433
* jl/submodule-rm (2012-08-27) 1 commit
434
- Teach rm to remove submodules unless they contain a git directory
@@ -447,7 +470,7 @@ option.
470
471
Will merge to 'master' as part of the fourth batch.
472
450
-* fa/remote-svn (2012-08-22) 16 commits
473
+* fa/remote-svn (2012-08-28) 16 commits
474
- Add a test script for remote-svn
475
- remote-svn: add marks-file regeneration
476
- Add a svnrdump-simulator replaying a dump file for testing
@@ -462,23 +485,20 @@ Will merge to 'master' as part of the fourth batch.
485
- Connect fast-import to the remote-helper via pipe, adding 'bidi-import' capability
486
- Add argv_array_detach and argv_array_free_detached
487
- Add svndump_init_fd to allow reading dumps from arbitrary FDs
465
- - Add git-remote-testsvn to Makefile
488
+ - Add git-remote-testsvn to Makefile and .gitignore
489
- Implement a remote helper for svn in C
490
(this branch is used by fa/vcs-svn.)
491
469
-* fa/vcs-svn (2012-08-26) 8 commits
470
- - t9020-*.sh: Skip all tests when python not available
471
- - t9020-*.sh: Fix urls passed to git-remote in test setup
472
- - remote-testsvn.c: avoid use of the getline() function
473
- - vcs-svn: Fix 'fa/remote-svn' and 'fa/vcs-svn' in pu
492
+Waiting for comments from mentors and stakeholders.
493
+
494
+* fa/vcs-svn (2012-08-28) 4 commits
495
- vcs-svn: remove repo_tree
496
- vcs-svn/svndump: rewrite handle_node(), begin|end_revision()
497
- vcs-svn/svndump: restructure node_ctx, rev_ctx handling
498
- svndump: move struct definitions to .h
499
(this branch uses fa/remote-svn.)
500
480
-The tip ones should be squashed in.
481
-Waiting for a reroll.
501
+Waiting for comments from mentors and stakeholders.
502
503
* nd/i18n-parseopt-help (2012-08-22) 66 commits
504
(merged to 'next' on 2012-08-24 at 2a04ce0)
@@ -576,18 +596,19 @@ need to be diagnosed.
596
Will merge to 'master' as part of the fourth batch.
597
598
* nd/i18n-poison-test-updates (2012-08-27) 7 commits
579
- - Fix tests under GETTEXT_POISON on parseopt
580
- - Fix tests under GETTEXT_POISON on git-remote
581
- - Fix tests under GETTEXT_POISON on pack-object
582
- - Fix tests under GETTEXT_POISON on git-apply
583
- - Fix tests under GETTEXT_POISON on diffstat
584
- - Fix tests under GETTEXT_POISON on git-stash
585
- - Fix tests under GETTEXT_POISON on relative dates
599
+ (merged to 'next' on 2012-08-29 at 9b7ff7f)
600
+ + Fix tests under GETTEXT_POISON on parseopt
601
+ + Fix tests under GETTEXT_POISON on git-remote
602
+ + Fix tests under GETTEXT_POISON on pack-object
603
+ + Fix tests under GETTEXT_POISON on git-apply
604
+ + Fix tests under GETTEXT_POISON on diffstat
605
+ + Fix tests under GETTEXT_POISON on git-stash
606
+ + Fix tests under GETTEXT_POISON on relative dates
607
608
Updates to tests that can be broken with gettext-poison builds.
609
Queued with trivial fix-ups.
610
590
-Will merge to 'next'.
611
+Will merge to 'master' as part of the third batch.
612
613
* jc/daemon-access-hook (2012-08-15) 1 commit
614
(merged to 'next' on 2012-08-20 at c8f5b2e)
@@ -608,34 +629,6 @@ prompt is given with a correctly guessed default.
629
630
Will merge to 'master' as part of the third batch.
631
611
-* ef/win32-cred-helper (2012-08-16) 1 commit
612
- (merged to 'next' on 2012-08-20 at bd5c651)
613
- + contrib: add win32 credential-helper
614
- (this branch is used by ph/credential-refactor.)
615
-
616
-Credential helper for Win32 to allow access to the keychain of
617
-the logged-in user.
618
-
619
-Will merge to 'master' as part of the second batch.
620
-
621
-* rr/precompose-utf8-cleanup (2012-08-20) 2 commits
622
- (merged to 'next' on 2012-08-20 at a883c2a)
623
- + precompose-utf8: do not call checks for non-ascii "utf8"
624
- + cleanup precompose_utf8
625
-
626
-A style and code cleanup.
627
-
628
-Will merge to 'master' as part of the second batch.
629
-
630
-* mh/maint-config-doc-proxy-command (2012-08-18) 2 commits
631
- (merged to 'next' on 2012-08-20 at a484625)
632
- + git-config doc: unconfuse an example
633
- + git-config.txt: fix example
634
-
635
-A minor documentation update.
636
-
637
-Will merge to 'master' as part of the second batch.
638
-
632
* ph/stash-rerere (2012-08-17) 2 commits
633
(merged to 'next' on 2012-08-20 at ae37111)
634
+ stash: invoke rerere in case of conflict
@@ -646,59 +639,6 @@ not trigger "rerere" upon conflicts unlike other mergy operations.
639
640
Will merge to 'master' as part of the third batch.
641
649
-* hv/submodule-path-unmatch (2012-08-14) 1 commit
650
- (merged to 'next' on 2012-08-20 at cbe6483)
651
- + Let submodule command exit with error status if path does not exist
652
-
653
-As the title says.
654
-
655
-Will merge to 'master' as part of the second batch.
656
-
657
-* bc/prune-info (2012-08-07) 1 commit
658
- (merged to 'next' on 2012-08-20 at 1bc9e66)
659
- + prune.c: only print informational message in show_only or verbose mode
660
-
661
-Teach "git prune" without "-v" to be silent about leftover temporary files.
662
-
663
-Will merge to 'master' as part of the second batch.
664
-
665
-* jc/maint-protect-sh-from-ifs (2012-08-08) 1 commit
666
- (merged to 'next' on 2012-08-20 at 8b00562)
667
- + sh-setup: protect from exported IFS
668
-
669
-When the user exports a non-default IFS without HT, scripts that
670
-rely on being able to parse "ls-files -s | while read a b c..."
671
-start to fail. Protect them from such a misconfiguration.
672
-
673
-Will merge to 'master' as part of the second batch.
674
-
675
-* mg/rebase-i-onto-reflog-in-full (2012-08-10) 1 commit
676
- (merged to 'next' on 2012-08-20 at f671bd2)
677
- + rebase -i: use full onto sha1 in reflog
678
-
679
-The reflog entries left by "git rebase" and "git rebase -i" were
680
-inconsistent.
681
-
682
-Will merge to 'master' as part of the second batch.
683
-
684
-* mz/empty-rebase-test (2012-08-09) 1 commit
685
- (merged to 'next' on 2012-08-20 at 53e2419)
686
- + add tests for 'git rebase --keep-empty'
687
-
688
-Will merge to 'master' as part of the second batch.
689
-
690
-* jc/capabilities (2012-08-13) 3 commits
691
- (merged to 'next' on 2012-08-20 at 1976e47)
692
- + fetch-pack: mention server version with verbose output
693
- + parse_feature_request: make it easier to see feature values
694
- + fetch-pack: do not ask for unadvertised capabilities
695
- (this branch uses jk/version-string.)
696
-
697
-Some capabilities were asked by fetch-pack even when upload-pack did
698
-not advertise that they are available. Fix fetch-pack not to do so.
699
-
700
-Will merge to 'master' as part of the second batch.
701
-
642
* pw/p4-use-client-spec-branch-detection (2012-08-20) 5 commits
643
(merged to 'next' on 2012-08-24 at 08abc61)
644
+ git p4: make branch detection work with --use-client-spec
@@ -712,30 +652,12 @@ used together (the command used to misdetect branches).
652
653
Will merge to 'master' as part of the fourth batch.
654
715
-* jc/maint-push-refs-all (2012-08-07) 2 commits
655
+* jc/maint-push-refs-all (2012-08-27) 2 commits
656
- get_fetch_map(): tighten checks on dest refs
717
- - fetch/push: allow refs/*:refs/*
657
+ - [BROKEN] fetch/push: allow refs/*:refs/*
658
659
Allows pushing and fetching refs/stash.
660
721
-Will merge to 'next'.
722
-
723
-* jk/version-string (2012-08-10) 3 commits
724
- (merged to 'next' on 2012-08-20 at 6d207bb)
725
- + do not send client agent unless server does first
726
- + send-pack: fix capability-sending logic
727
- + include agent identifier in capability string
728
- (this branch is used by jc/capabilities.)
729
-
730
-(Originally merged to 'next' on 2012-08-10)
731
-
732
-Leftover bits from the "git version" code restructuring, that give
733
-us an ability to expose the Git version over the wire in the
734
-protocol exchange. But we do so carefully only to the other end that
735
-understands it.
736
-
737
-Will merge to 'master' as part of the third batch.
738
-
661
* tg/index-v5 (2012-08-17) 13 commits
662
- p0002-index.sh: add perf test for the index formats
663
- update-index.c: rewrite index when index-version is given
@@ -777,16 +699,6 @@ silently proceed.
699
700
Will merge to 'master' as part of the third batch.
701
780
-* mz/cherry-code-cleanup (2012-07-29) 3 commits
781
- (merged to 'next' on 2012-08-20 at bfb0457)
782
- + cherry: remove redundant check for merge commit
783
- + cherry: don't set ignored rev_info options
784
- + remove unnecessary parameter from get_patch_ids()
785
-
786
-Minor code clean-up on the cherry-pick codepath.
787
-
788
-Will merge to 'master' as part of the second batch.
789
-
702
* jc/maint-config-exit-status (2012-07-30) 1 commit
703
(merged to 'next' on 2012-08-20 at e18ddd8)
704
+ config: "git config baa" should exit with status 1
@@ -828,3 +740,16 @@ Teach the code that works around NKD/NKC gotcha on MacOS to call
740
iconv_open() only when it is necessary, in the hope of avoiding
741
set-up overhead. It turns out that there was no noticeable
742
improvements.
743
+
744
+* nd/checkout-branch-name-check (2012-08-27) 1 commit
745
+ . checkout: verify new branch name's validity early
746
+
747
+"git checkout -b --opt y" errors out saying that creating a new
748
+branch to check it out and grabbing contents for paths out of a
749
+commit are incompatible operations. While it is technically correct
750
+(the command line wants to create a new branch whose name is "--opt"
751
+and check it out, and there shouldn't be anything else left on the
752
+command line, but there is "y"), "--opt" is not a valid name of the
753
+branch to begin with, so even without "y", the command will not
754
+succeed. Treat this case specially to complain that "--opt" is not
755
+a valid branch name.