What's cooking (2017/03 #03)
Junio C Hamano committed
Mar 8, 2017 at 15:42 UTC
3732b4209945f600b77e6c7757b3665c71ad54e5
1 file changed
+137
-62
whats-cooking.txt
+137
-62
@@ -1,10 +1,10 @@
1
To: git@vger.kernel.org
2
Bcc: lwn@lwn.net
3
-Subject: What's cooking in git.git (Mar 2017, #02; Fri, 3)
4
-X-master-at: 3bc53220cb2dcf709f7a027a3f526befd021d858
5
-X-next-at: a2ecc84866b651ad965003275b748da92fd99ab4
3
+Subject: What's cooking in git.git (Mar 2017, #03; Wed, 8)
4
+X-master-at: e0688e9b28f2c5ff711460ee8b62077be5df2360
5
+X-next-at: 76c07830f9451c898309f3a605eb1f610cf92335
6
7
-What's cooking in git.git (Mar 2017, #02; Fri, 3)
7
+What's cooking in git.git (Mar 2017, #03; Wed, 8)
8
--------------------------------------------------
9
10
Here are the topics that have been cooking. Commits prefixed with
@@ -20,78 +20,81 @@ of the repositories listed at
20
--------------------------------------------------
21
[New Topics]
22
23
-* ew/markdown-url-in-readme (2017-03-01) 1 commit
24
- (merged to 'next' on 2017-03-03 at 3d35e3a991)
25
- + README: create HTTP/HTTPS links from URLs in Markdown
23
+* ew/http-alternates-as-redirects-warning (2017-03-06) 2 commits
24
+ - http: release strbuf on disabled alternates
25
+ - http: inform about alternates-as-redirects behavior
26
27
- Doc update.
27
+ Recent versions of Git treats http alternates (used in dumb http
28
+ transport) just like HTTP redirects and requires the client to
29
+ enable following it, due to security concerns. But we forgot to
30
+ give a warning when we decide not to honor the alternates.
31
29
- Will merge to 'master'.
32
+ Will merge to 'next'.
33
34
32
-* jk/add-i-patch-do-prompt (2017-03-02) 1 commit
33
- - add--interactive: fix missing file prompt for patch mode with "-i"
35
+* jk/ewah-use-right-type-in-sizeof (2017-03-06) 1 commit
36
+ - ewah: fix eword_t/uint64_t confusion
37
35
- The patch subcommand of "git add -i" was meant to have paths
36
- selection prompt just like other subcommand, unlike "git add -p"
37
- directly jumps to hunk selection. Recently, this was broken and
38
- "add -i" lost the paths selection dialog, but it now has been
39
- fixed.
38
+ Code clean-up.
39
40
Will merge to 'next'.
41
42
44
-* ax/line-log-range-merge-fix (2017-03-03) 1 commit
45
- - line-log.c: prevent crash during union of too many ranges
43
+* jk/push-deadlock-regression-fix (2017-03-07) 6 commits
44
+ - send-pack: report signal death of pack-objects
45
+ - send-pack: read "unpack" status even on pack-objects failure
46
+ - send-pack: improve unpack-status error messages
47
+ - send-pack: use skip_prefix for parsing unpack status
48
+ - send-pack: extract parsing of "unpack" response
49
+ - receive-pack: fix deadlock when we cannot create tmpdir
50
47
- The code to parse "git log -L..." command line was buggy when there
48
- are many ranges specified with -L; overrun of the allocated buffer
49
- has been fixed.
51
+ "git push" had a handful of codepaths that could lead to a deadlock
52
+ when unexpected error happened, which has been fixed.
53
54
Will merge to 'next'.
55
56
54
-* js/early-config (2017-03-03) 11 commits
55
- - t1309: test read_early_config()
56
- - read_early_config(): really discover .git/
57
- - read_early_config(): avoid .git/config hack when unneeded
58
- - setup: make read_early_config() reusable
59
- - setup: export the discover_git_directory() function
60
- - SQUASH??? ERROR: trailing statements should be on next line
61
- - setup_git_directory_1(): avoid changing global state
62
- - setup: prepare setup_discovered_git_directory() the root directory
63
- - SQUASH??? ERROR: trailing statements should be on next line
64
- - setup_git_directory(): use is_dir_sep() helper
65
- - t7006: replace dubious test
57
+* vn/line-log-memcpy-size-fix (2017-03-06) 1 commit
58
+ - line-log: use COPY_ARRAY to fix mis-sized memcpy
59
67
- The start-up sequence of "git" needs to figure out some configured
68
- settings before it finds and set itself up in the location of the
69
- repository and was quite messy due to its "chicken-and-egg" nature.
70
- The code has been restructured.
60
+ The command-line parsing of "git log -L" copied internal data
61
+ structures using incorrect size on ILP32 systems.
62
72
- Will merge to 'next' after squashing niggle-fixes in.
63
+ Will merge to 'next'.
64
65
75
-* jt/perf-updates (2017-03-03) 3 commits
76
- - t/perf: add fallback for pre-bin-wrappers versions of git
77
- - t/perf: use $MODERN_GIT for all repo-copying steps
78
- - t/perf: export variable used in other blocks
66
+* js/realpath-pathdup-fix (2017-03-08) 2 commits
67
+ - real_pathdup(): fix callsites that wanted it to die on error
68
+ - t1501: demonstrate NULL pointer access with invalid GIT_WORK_TREE
69
80
- The t/perf performance test suite was not prepared to test not so
81
- old versions of Git, but now it covers versions of Git that are not
82
- so ancient.
70
+ Git v2.12 was shipped with an embarrassing breakage where various
71
+ operations that verify paths given from the user stopped dying when
72
+ seeing an issue, and instead later triggering segfault.
73
84
- Will merge to 'next'.
74
+ Will merge to 'next' and then to 'master', eventually down to 'maint'.
75
76
87
-* ss/remote-bzr-hg-placeholder-wo-python (2017-03-03) 1 commit
88
- - contrib: git-remote-{bzr,hg} placeholders don't need Python
77
+* kn/ref-filter-branch-list (2017-03-08) 1 commit
78
+ - branch: honor --abbrev/--no-abbrev in --list mode
79
90
- There is no need for Python only to give a few messages to the
91
- standard error stream, but we somehow did.
80
+ "git branch --list" takes the "--abbrev" and "--no-abbrev" options
81
+ to control the output of the object name in its "-v"(erbose)
82
+ output, but a recent update started ignoring them; this fixes it
83
+ before the breakage reaches to any released version.
84
85
Will merge to 'next'.
86
87
+
88
+* sb/rev-parse-show-superproject-root (2017-03-08) 1 commit
89
+ - rev-parse: add --show-superproject-working-tree
90
+
91
+ From a working tree of a repository, a new option of "rev-parse"
92
+ lets you ask if the repository is used as a submodule of another
93
+ project, and where the root level of the working tree of that
94
+ project (i.e. your superproject) is.
95
+
96
+ Almost there, but documentation needs a bit more work.
97
+
98
--------------------------------------------------
99
[Stalled]
100
@@ -147,12 +150,6 @@ of the repositories listed at
150
cf. <CAFZEwPNUXcNY9Qdz=_B7q2kQuaecPzJtTMGdv8YMUPEz2vnp8A@mail.gmail.com>
151
152
150
-* ls/filter-process-delayed (2017-01-08) 1 commit
151
- . convert: add "status=delayed" to filter process protocol
152
-
153
- Ejected, as does not build when merged to 'pu'.
154
-
155
-
153
* sh/grep-tree-obj-tweak-output (2017-01-20) 2 commits
154
- grep: use '/' delimiter for paths
155
- grep: only add delimiter if there isn't one already
@@ -192,6 +189,84 @@ of the repositories listed at
189
--------------------------------------------------
190
[Cooking]
191
192
+* ls/filter-process-delayed (2017-03-06) 1 commit
193
+ - convert: add "status=delayed" to filter process protocol
194
+
195
+ cf. <xmqq60jmnmef.fsf@junio-linux.mtv.corp.google.com>
196
+
197
+
198
+* ew/markdown-url-in-readme (2017-03-01) 1 commit
199
+ (merged to 'next' on 2017-03-03 at 3d35e3a991)
200
+ + README: create HTTP/HTTPS links from URLs in Markdown
201
+
202
+ Doc update.
203
+
204
+ Will merge to 'master'.
205
+
206
+
207
+* jk/add-i-patch-do-prompt (2017-03-02) 1 commit
208
+ - add--interactive: fix missing file prompt for patch mode with "-i"
209
+
210
+ The patch subcommand of "git add -i" was meant to have paths
211
+ selection prompt just like other subcommand, unlike "git add -p"
212
+ directly jumps to hunk selection. Recently, this was broken and
213
+ "add -i" lost the paths selection dialog, but it now has been
214
+ fixed.
215
+
216
+ Will merge to 'next'.
217
+
218
+
219
+* ax/line-log-range-merge-fix (2017-03-03) 1 commit
220
+ - line-log.c: prevent crash during union of too many ranges
221
+
222
+ The code to parse "git log -L..." command line was buggy when there
223
+ are many ranges specified with -L; overrun of the allocated buffer
224
+ has been fixed.
225
+
226
+ Will merge to 'next'.
227
+
228
+
229
+* js/early-config (2017-03-07) 10 commits
230
+ - setup_git_directory_gently_1(): avoid die()ing
231
+ - t1309: test read_early_config()
232
+ - read_early_config(): really discover .git/
233
+ - read_early_config(): avoid .git/config hack when unneeded
234
+ - setup: make read_early_config() reusable
235
+ - setup: introduce the discover_git_directory() function
236
+ - setup_git_directory_1(): avoid changing global state
237
+ - setup: prepare setup_discovered_git_directory() the root directory
238
+ - setup_git_directory(): use is_dir_sep() helper
239
+ - t7006: replace dubious test
240
+
241
+ The start-up sequence of "git" needs to figure out some configured
242
+ settings before it finds and set itself up in the location of the
243
+ repository and was quite messy due to its "chicken-and-egg" nature.
244
+ The code has been restructured.
245
+
246
+ Will merge to 'next' after waiting for a few days.
247
+
248
+
249
+* jt/perf-updates (2017-03-03) 3 commits
250
+ - t/perf: add fallback for pre-bin-wrappers versions of git
251
+ - t/perf: use $MODERN_GIT for all repo-copying steps
252
+ - t/perf: export variable used in other blocks
253
+
254
+ The t/perf performance test suite was not prepared to test not so
255
+ old versions of Git, but now it covers versions of Git that are not
256
+ so ancient.
257
+
258
+ Will merge to 'next'.
259
+
260
+
261
+* ss/remote-bzr-hg-placeholder-wo-python (2017-03-03) 1 commit
262
+ - contrib: git-remote-{bzr,hg} placeholders don't need Python
263
+
264
+ There is no need for Python only to give a few messages to the
265
+ standard error stream, but we somehow did.
266
+
267
+ Will merge to 'next'.
268
+
269
+
270
* jk/interpret-branch-name (2017-03-02) 9 commits
271
- checkout: restrict @-expansions when finding branch
272
- strbuf_check_ref_format(): expand only local branches
@@ -226,7 +301,7 @@ of the repositories listed at
301
cf. <CA+55aFxTWqsTTiDKo4DBZT-8Z9t80bGMD3uijzKONa_bYEZABQ@mail.gmail.com>
302
303
229
-* js/travis-32bit-linux (2017-03-03) 1 commit
304
+* js/travis-32bit-linux (2017-03-06) 1 commit
305
- Travis: also test on 32-bit Linux
306
307
Add 32-bit Linux variant to the set of platforms to be tested with
@@ -328,7 +403,7 @@ of the repositories listed at
403
Will merge to 'master'.
404
405
331
-* cc/split-index-config (2017-03-01) 22 commits
406
+* cc/split-index-config (2017-03-06) 22 commits
407
- Documentation/git-update-index: explain splitIndex.*
408
- Documentation/config: add splitIndex.sharedIndexExpire
409
- read-cache: use freshen_shared_index() in read_index_from()
@@ -660,7 +735,8 @@ of the repositories listed at
735
from the other worktrees.
736
737
663
-* sb/checkout-recurse-submodules (2017-02-23) 15 commits
738
+* sb/checkout-recurse-submodules (2017-03-07) 18 commits
739
+ - builtin/read-tree: add --recurse-submodules switch
740
- builtin/checkout: add --recurse-submodules switch
741
- entry.c: update submodules when interesting
742
- read-cache, remove_marked_cache_entries: wipe selected submodules.
@@ -674,14 +750,13 @@ of the repositories listed at
750
- connect_work_tree_and_git_dir: safely create leading directories
751
- make is_submodule_populated gently
752
- lib-submodule-update.sh: define tests for recursing into submodules
753
+ - lib-submodule-update.sh: replace sha1 by hash
754
+ - lib-submodule-update: teach test_submodule_content the -C <dir> flag
755
- lib-submodule-update.sh: do not use ./. as submodule remote
756
- lib-submodule-update.sh: reorder create_lib_submodule_repo
757
758
"git checkout" is taught --recurse-submodules option.
759
682
- Reroll exists, but it appears it still needs minor work on top.
683
- cf. <20170302004759.27852-1-sbeller@google.com>
684
-
760
761
* tg/stash-push (2017-02-28) 6 commits
762
(merged to 'next' on 2017-03-03 at b50fda0389)