1 To: git@vger.kernel.org
3 Subject: What's cooking in git.git (Feb 2015, #07; Thu, 26)
4 X-master-at: 7f4ba4b6e3ba7075ca6b379ba23fd3088cbe69a8
5 X-next-at: 993ae8bdc3df206f62a424eb695a49f1ba8fa118
7 What's cooking in git.git (Feb 2015, #07; Thu, 26)
8 --------------------------------------------------
10 Here are the topics that have been cooking. Commits prefixed with
11 '-' are only in 'pu' (proposed updates) while commits prefixed with
14 Fifth batch of topics have been merged to 'master', including both
15 fixes and enhancements. First maintenance release for v2.3 was cut
16 with many fixes that have already been merged to 'master'.
18 You can find the changes described here in the integration branches
19 of the repositories listed at
21 http://git-blame.blogspot.com/p/git-public-repositories.html
23 --------------------------------------------------
24 [Graduated to "master"]
26 * ak/git-pm-typofix (2015-02-18) 1 commit
27 (merged to 'next' on 2015-02-20 at 0fa233a)
28 + Git.pm: two minor typo fixes
33 * dp/remove-duplicated-header-inclusion (2015-02-13) 1 commit
34 (merged to 'next' on 2015-02-18 at a1bf108)
35 + do not include the same header twice
40 * jc/max-io-size-and-ssize-max (2015-02-12) 1 commit
41 (merged to 'next' on 2015-02-18 at 0c8a4da)
42 + xread/xwrite: clip MAX_IO_SIZE to SSIZE_MAX
44 Our default I/O size (8 MiB) for large files was too large for some
45 platforms with smaller SSIZE_MAX, leading to read(2)/write(2)
49 * jc/send-email-sensible-encoding (2015-02-13) 1 commit
50 (merged to 'next' on 2015-02-18 at 7457655)
51 + send-email: ask confirmation if given encoding name is very short
53 "git send-email" used to accept a mistaken "y" (or "yes") as an
54 answer to "What encoding do you want to use [UTF-8]? " without
55 questioning. Now it asks for confirmation when the answer looks
56 too short to be a valid encoding name.
59 * jk/fast-import-die-nicely-fix (2015-02-10) 1 commit
60 (merged to 'next' on 2015-02-18 at e249425)
61 + fast-import: avoid running end_packfile recursively
63 "git fast-import" used to crash when it could not close and
64 conclude the resulting packfile cleanly.
67 * jk/sanity (2015-02-15) 3 commits
68 (merged to 'next' on 2015-02-18 at 5c54b53)
69 + test-lib.sh: set prerequisite SANITY by testing what we really need
70 + tests: correct misuses of POSIXPERM
71 + t/lib-httpd: switch SANITY check for NOT_ROOT
73 The tests that wanted to see that file becomes unreadable after
74 running "chmod a-r file", and the tests that wanted to make sure it
75 is not run as root, we used "can we write into the / directory?" as
76 a cheap substitute, but on some platforms that is not a good
77 heuristics. The tests and their prerequisites have been updated to
78 check what they really require.
81 * jk/strbuf-doc-to-header (2015-01-16) 7 commits
82 (merged to 'next' on 2015-02-18 at 0482c65)
83 + strbuf.h: group documentation for trim functions
84 + strbuf.h: drop boilerplate descriptions of strbuf_split_*
85 + strbuf.h: reorganize api function grouping headers
86 + strbuf.h: format asciidoc code blocks as 4-space indent
87 + strbuf.h: drop asciidoc list formatting from API docs
88 + strbuf.h: unify documentation comments beginnings
89 + strbuf.h: integrate api-strbuf.txt documentation
91 The strbuf API was explained between the API documentation and in
92 the header file. Move missing bits to strbuf.h so that programmers
93 can check only one place for all necessary information.
96 * jn/doc-api-errors (2014-12-04) 1 commit
97 (merged to 'next' on 2015-02-18 at f60eda6)
98 + doc: document error handling functions and conventions
100 The error handling functions and conventions are now documented in
104 * mh/transport-capabilities (2015-02-13) 2 commits
105 (merged to 'next' on 2015-02-18 at 87e8fcc)
106 + transport-helper: ask the helper to set the same options for import as for fetch
107 + transport-helper: ask the helper to set progress and verbosity options after asking for its capabilities
109 The transport-helper did not give transport options such as
110 verbosity, progress, cloning, etc. to import and export based
111 helpers, like it did for fetch and push based helpers, robbing them
112 the chance to honor the wish of the end-users better.
115 * nd/attr-optim (2014-12-29) 3 commits
116 (merged to 'next' on 2015-02-18 at 598e68a)
117 + attr: avoid heavy work when we know the specified attr is not defined
118 + attr: do not attempt to expand when we know it's not a macro
119 + attr.c: rename arg name attr_nr to avoid shadowing the global one
121 Optimize attribute look-up, mostly useful in "git grep" on a
122 project that does not use many attributes, by avoiding it when we
123 (should) know that the attributes are not defined in the first
127 * sb/hex-object-name-is-at-most-41-bytes-long (2015-02-13) 1 commit
128 (merged to 'next' on 2015-02-18 at 53d522b)
129 + hex.c: reduce memory footprint of sha1_to_hex static buffers
134 * sb/plug-leak-in-make-cache-entry (2015-02-17) 1 commit
135 (merged to 'next' on 2015-02-18 at e637f65)
136 + read-cache.c: free cache entry when refreshing fails
138 "update-index --refresh" used to leak when an entry cannot be
139 refreshed for whatever reason.
142 * tc/missing-http-proxyauth (2015-02-03) 1 commit
143 (merged to 'next' on 2015-02-18 at 8ff01ad)
144 + http: support curl < 7.10.7
146 We did not check the curl library version before using
147 CURLOPT_PROXYAUTH feature that may not exist.
149 --------------------------------------------------
152 * mg/detached-head-report (2015-02-23) 1 commit
153 - branch: name detached HEAD analogous to status
155 "git branch" on a detached HEAD always said "(detached from xyz)",
156 even when "git status" would report "detached at xyz". The HEAD is
157 actually at xyz and haven't been moved since it was detached in
158 such a case, but the user cannot read what the current value of
159 HEAD is when "detached from" is used.
161 The goal is sound. May want to share code between branch and
162 status in a reroll? Are there other commands that give the same
163 information that share the code, too?
166 * mg/sequencer-commit-messages-always-verbatim (2015-02-23) 1 commit
167 - sequencer: preserve commit messages
169 Always replay the original commit message verbatim, regardless of
170 the commit.cleanup configuration settings by the user.
175 * rj/no-xopen-source-for-cygwin (2015-02-22) 1 commit
176 (merged to 'next' on 2015-02-23 at 323ac75)
177 + git-compat-util.h: remove redundant code
179 Will merge to 'master'.
181 --------------------------------------------------
184 * jk/push-config (2015-02-17) 4 commits
185 - [NEEDSACK] push: allow --follow-tags to be set by config push.followTags
186 - cmd_push: pass "flags" pointer to config callback
187 - cmd_push: set "atomic" bit directly
188 - git_push_config: drop cargo-culted wt_status pointer
190 Waiting for Ack and/or update for the tip one from Dave Olszewski
191 ($gmane/263880, $gmane/263991).
194 * nd/untracked-cache (2015-02-09) 24 commits
195 - git-status.txt: advertisement for untracked cache
196 - untracked cache: guard and disable on system changes
197 - mingw32: add uname()
198 - t7063: tests for untracked cache
199 - update-index: test the system before enabling untracked cache
200 - update-index: manually enable or disable untracked cache
201 - status: enable untracked cache
202 - untracked-cache: temporarily disable with $GIT_DISABLE_UNTRACKED_CACHE
203 - untracked cache: mark index dirty if untracked cache is updated
204 - untracked cache: print stats with $GIT_TRACE_UNTRACKED_STATS
205 - untracked cache: avoid racy timestamps
206 - read-cache.c: split racy stat test to a separate function
207 - untracked cache: invalidate at index addition or removal
208 - untracked cache: load from UNTR index extension
209 - untracked cache: save to an index extension
210 - ewah: add convenient wrapper ewah_serialize_strbuf()
211 - untracked cache: don't open non-existent .gitignore
212 - untracked cache: mark what dirs should be recursed/saved
213 - untracked cache: record/validate dir mtime and reuse cached output
214 - untracked cache: make a wrapper around {open,read,close}dir()
215 - untracked cache: invalidate dirs recursively if .gitignore changes
216 - untracked cache: initial untracked cache validation
217 - untracked cache: record .gitignore information and dir hierarchy
218 - dir.c: optionally compute sha-1 of a .gitignore file
220 Need extra sets of eyes to review this.
223 * ak/stash-store-create-help (2015-01-13) 1 commit
224 - stash: show "create" and "store" subcommands in usage-help
229 * bp/diff-relative-config (2015-01-07) 2 commits
230 - diff: teach diff.relative to give default to --relative=<value>
231 - diff: teach --no-relative to override earlier --relative
233 No comments? No interests?
236 * jc/diff-b-m (2015-02-23) 5 commits
239 - diffcore-rename: allow easier debugging
240 - diffcore-rename.c: add locate_rename_src()
241 - diffcore-break: allow debugging
242 (this branch uses jc/diff-test-updates.)
244 "git diff -B -M" produced incorrect patch when the postimage of a
245 completely rewritten file is similar to the preimage of a removed
246 file; such a resulting file must not be expressed as a rename from
249 The fix in this patch is broken, unfortunately.
252 * nd/slim-index-pack-memory-usage (2015-02-10) 3 commits
253 - index-pack: kill union delta_base to save memory
255 - index-pack: reduce memory footprint a bit
257 Need to replace with the reroll ($gmane/264126).
260 * pw/remote-set-url-fetch (2014-11-26) 1 commit
261 - remote: add --fetch and --both options to set-url
266 * je/quiltimport-no-fuzz (2014-10-21) 2 commits
267 - git-quiltimport: flip the default not to allow fuzz
268 - git-quiltimport.sh: allow declining fuzz with --exact option
270 "quiltimport" drove "git apply" always with -C1 option to reduce
271 context of the patch in order to give more chance to somewhat stale
272 patches to apply. Add an "--exact" option to disable, and also
273 "-C$n" option to customize this behaviour. The top patch
274 optionally flips the default to "--exact".
276 Tired of waiting for an Ack
280 * tr/remerge-diff (2014-11-10) 9 commits
281 - t4213: avoid "|" in sed regexp
282 - log --remerge-diff: show what the conflict resolution changed
283 - name-hash: allow dir hashing even when !ignore_case
284 - merge-recursive: allow storing conflict hunks in index
285 - merge_diff_mode: fold all merge diff variants into an enum
286 - combine-diff: do not pass revs->dense_combined_merges redundantly
287 - merge-recursive: -Xindex-only to leave worktree unchanged
288 - merge-recursive: internal flag to avoid touching the worktree
289 - merge-recursive: remove dead conditional in update_stages()
291 "log -p" output learns a new way to let users inspect a merge
292 commit by showing the differences between the automerged result
293 with conflicts the person who recorded the merge would have seen
294 and the final conflict resolution that was recorded in the merge.
296 Waiting for a reroll ($gmane/256591).
299 * hv/submodule-config (2014-11-11) 4 commits
300 - do not die on error of parsing fetchrecursesubmodules option
301 - use new config API for worktree configurations of submodules
302 - extract functions for submodule config set and lookup
303 - implement submodule config cache for lookup of submodule names
305 Kicked back to 'pu' per request ($gmane/255610).
308 * ab/add-interactive-show-diff-func-name (2014-05-12) 2 commits
309 - SQUASH??? git-add--interactive: Preserve diff heading when splitting hunks
310 - git-add--interactive: Preserve diff heading when splitting hunks
312 Waiting for a reroll.
315 * jn/gitweb-utf8-in-links (2014-05-27) 1 commit
316 - gitweb: Harden UTF-8 handling in generated links
321 * ss/userdiff-update-csharp-java (2014-06-02) 2 commits
322 - userdiff: support Java try keyword
323 - userdiff: support C# async methods and correct C# keywords
325 Reviews sent; waiting for a response.
328 * bg/rebase-off-of-previous-branch (2014-04-16) 1 commit
329 - git-rebase: print name of rev when using shorthand
331 Teach "git rebase -" to report the concrete name of the branch
332 (i.e. the previous one).
334 But it stops short and does not do the same for "git rebase @{-1}".
338 * rb/merge-prepare-commit-msg-hook (2014-01-10) 4 commits
339 - merge: drop unused arg from abort_commit method signature
340 - merge: make prepare_to_commit responsible for write_merge_state
341 - t7505: ensure cleanup after hook blocks merge
342 - t7505: add missing &&
344 Expose more merge states (e.g. $GIT_DIR/MERGE_MODE) to hooks that
345 run during "git merge". The log message stresses too much on one
346 hook, prepare-commit-msg, but it would equally apply to other hooks
347 like post-merge, I think.
349 Waiting for a reroll.
352 * jc/graph-post-root-gap (2013-12-30) 3 commits
353 - WIP: document what we want at the end
354 - graph: remove unused code a bit
355 - graph: stuff the current commit into graph->columns[]
357 This was primarily a RFH ($gmane/239580).
360 * tg/perf-lib-test-perf-cleanup (2013-09-19) 2 commits
361 - perf-lib: add test_perf_cleanup target
362 - perf-lib: split starting the test from the execution
364 Add test_perf_cleanup shell function to the perf suite, that allows
365 the script writers to define a test with a clean-up action.
370 * jc/show-branch (2014-03-24) 5 commits
371 - show-branch: use commit slab to represent bitflags of arbitrary width
372 - show-branch.c: remove "all_mask"
373 - show-branch.c: abstract out "flags" operation
374 - show-branch.c: lift all_mask/all_revs to a global static
375 - show-branch.c: update comment style
377 Waiting for the final step to lift the hard-limit before sending it out.
379 --------------------------------------------------
382 * nd/list-files (2015-02-09) 21 commits
383 - t3080: tests for git-list-files
384 - list-files: -M aka diff-cached
385 - list-files -F: show submodules with the new indicator '&'
386 - list-files: add -F/--classify
387 - list-files: show directories as well as files
388 - list-files: do not show duplicate cached entries
389 - list-files: sort output and remove duplicates
390 - list-files: add -t back
391 - list-files: add -1 short for --no-column
392 - list-files: add -R/--recursive short for --max-depth=-1
393 - list-files: -u does not imply showing stages
394 - list-files: make alias 'ls' default to 'list-files'
395 - list-files: a user friendly version of ls-files and more
396 - ls-files: support --max-depth
397 - ls-files: add --column
398 - ls-files: add --color to highlight file names
399 - ls-files: buffer full item in strbuf before printing
400 - ls_colors.c: highlight submodules like directories
401 - ls_colors.c: add a function to color a file name
402 - ls_colors.c: parse color.ls.* from config file
403 - ls_colors.c: add $LS_COLORS parsing code
405 A new "git list-files" Porcelain command, "ls-files" with bells and
409 * mm/am-c-doc (2015-02-20) 2 commits
410 (merged to 'next' on 2015-02-23 at 1cd9b0f)
411 + Documentation/git-am.txt: mention mailinfo.scissors config variable
412 + Documentation/config.txt: document mailinfo.scissors
414 The configuration variable 'mailinfo.scissors' was hard to
415 discover in the documentation.
417 Will merge to 'master'.
420 * rs/simple-cleanups (2015-02-22) 4 commits
421 (merged to 'next' on 2015-02-23 at a89f034)
422 + sha1_name: use strlcpy() to copy strings
423 + pretty: use starts_with() to check for a prefix
424 + for-each-ref: use skip_prefix() to avoid duplicate string comparison
425 + connect: use strcmp() for string comparison
427 Will merge to 'master'.
430 * tb/connect-ipv6-parse-fix (2015-02-22) 3 commits
431 (merged to 'next' on 2015-02-24 at 6658564)
432 + t5500: show user name and host in diag-url
433 + t5601: add more test cases for IPV6
434 + connect.c: allow ssh://user@[2001:db8::1]/repo.git
436 We did not parse username followed by literal IPv6 address in SSH
437 transport URLs, e.g. ssh://user@[2001:db8::1]:22/repo.git
440 Will merge to 'master'.
443 * jc/decorate-leaky-separator-color (2015-02-20) 2 commits
444 - config.txt: spell out how certain typed values are written
445 - log --decorate: do not leak "commit" color into the next item
447 "git log --decorate" did not reset colors correctly around the
450 Waiting for discussion on the tip one.
453 * ms/submodule-update-config-doc (2015-02-23) 1 commit
454 - submodule: improve documentation of update subcommand
456 The interaction between "git submodule update" and the
457 submodule.*.update configuration was not clearly documented.
459 Waiting for a final reroll.
462 * jc/diff-test-updates (2015-02-23) 7 commits
463 (merged to 'next' on 2015-02-23 at 9e863a7)
464 + test_ln_s_add: refresh stat info of fake symbolic links
465 (merged to 'next' on 2015-02-18 at f152c9f)
466 + t4008: modernise style
467 + t/diff-lib: check exact object names in compare_diff_raw
468 + tests: do not borrow from COPYING and README from the real source
469 + t4010: correct expected object names
470 + t9300: correct expected object names
471 + t4008: correct stale comments
472 (this branch is used by jc/diff-b-m.)
476 Will merge to 'master'.
479 * km/send-email-getopt-long-workarounds (2015-02-16) 1 commit
480 (merged to 'next' on 2015-02-18 at 84c1597)
481 + git-send-email.perl: support no- prefix with older GetOptions
483 Even though we officially haven't dropped Perl 5.8 support, the
484 Getopt::Long package that came with it does not support "--no-"
485 prefix to negate a boolean option; manually add support to help
486 people with older Getopt::Long package.
488 Will merge to 'master'.
491 * jk/daemon-interpolate (2015-02-17) 3 commits
492 (merged to 'next' on 2015-02-20 at 21118a7)
493 + daemon: sanitize incoming virtual hostname
494 + t5570: test git-daemon's --interpolated-path option
495 + git_connect: let user override virtual-host we send to daemon
497 The "interpolated-path" option of "git daemon" inserted any string
498 client declared on the "host=" capability request without checking.
499 Sanitize and limit %H and %CH to a saner and a valid DNS name.
501 Will merge to 'master'.
504 * rs/daemon-interpolate (2015-02-17) 2 commits
505 (merged to 'next' on 2015-02-18 at c0f5750)
506 + daemon: use callback to build interpolated path
507 + daemon: look up client-supplied hostname lazily
509 "git daemon" looked up the hostname even when "%CH" and "%IP"
510 interpolations are not requested, which was unnecessary.
512 Will merge to 'master'.
515 * mh/expire-updateref-fixes (2015-02-09) 10 commits
516 - reflog_expire(): lock symbolic refs themselves, not their referent
517 - reflog_expire(): never update a reference to null_sha1
518 - reflog_expire(): ignore --updateref for symbolic references
519 - reflog: rearrange the manpage
520 - reflog: fix documentation
521 - lock_ref_sha1_basic(): do not set force_write for missing references
522 - write_ref_sha1(): Move write elision test to callers
523 - write_ref_sha1(): remove check for lock == NULL
524 - Merge branch 'sb/atomic-push' into mh/ref-trans-value-check
525 - Merge branch 'mh/reflog-expire' into mh/ref-trans-value-check
527 Under discussion ($gmane/263552)
530 * mh/refs-have-new (2015-02-17) 13 commits
531 (merged to 'next' on 2015-02-23 at b8929db)
532 + refs.h: remove duplication in function docstrings
533 + update_ref(): improve documentation
534 + ref_transaction_verify(): new function to check a reference's value
535 + ref_transaction_delete(): check that old_sha1 is not null_sha1
536 + ref_transaction_create(): check that new_sha1 is valid
537 + commit: avoid race when creating orphan commits
538 + commit: add tests of commit races
539 + ref_transaction_delete(): remove "have_old" parameter
540 + ref_transaction_update(): remove "have_old" parameter
541 + struct ref_update: move "have_old" into "flags"
542 + refs.c: change some "flags" to "unsigned int"
543 + refs: remove the gap in the REF_* constant values
544 + refs: move REF_DELETING to refs.c
546 Simplify the ref transaction API around how "the ref should be
547 pointing at this object" is specified.
549 Will merge to 'master'.
552 * jc/apply-beyond-symlink (2015-02-10) 4 commits
553 (merged to 'next' on 2015-02-18 at 4109de1)
554 + apply: do not touch a file beyond a symbolic link
555 + apply: do not read from beyond a symbolic link
556 + apply: do not read from the filesystem under --index
557 + apply: reject input that touches outside the working area
559 "git apply" was not very careful about reading from, removing,
560 updating and creating paths outside the working tree (under
561 --index/--cached) or the current directory (when used as a
562 replacement for GNU patch).
564 Will merge to 'master'.
567 * jc/apply-ws-fix-expands-report (2015-01-22) 1 commit
568 (merged to 'next' on 2015-02-18 at b571668)
569 + apply: detect and mark whitespace errors in context lines when fixing
571 "git apply --whitespace=fix" fixed whitespace errors in the common
572 context lines but did so without reporting.
574 Will merge to 'master'.
577 * js/fsck-opt (2015-01-21) 19 commits
578 - fsck: support ignoring objects in `git fsck` via fsck.skiplist
579 - fsck: git receive-pack: support excluding objects from fsck'ing
580 - fsck: introduce `git fsck --quick`
581 - fsck: support demoting errors to warnings
582 - fsck: document the new receive.fsck.* options
583 - fsck: allow upgrading fsck warnings to errors
584 - fsck: optionally ignore specific fsck issues completely
585 - fsck: disallow demoting grave fsck errors to warnings
586 - fsck: add a simple test for receive.fsck.*
587 - fsck: make fsck_tag() warn-friendly
588 - fsck: handle multiple authors in commits specially
589 - fsck: make fsck_commit() warn-friendly
590 - fsck: make fsck_ident() warn-friendly
591 - fsck: report the ID of the error/warning
592 - fsck: allow demoting errors to warnings via receive.fsck.warn = <key>
593 - fsck: offer a function to demote fsck errors to warnings
594 - fsck: provide a function to parse fsck message IDs
595 - fsck: introduce identifiers for fsck messages
596 - fsck: introduce fsck options
598 "fsck.warnings = <list of error tokens>" I suggested turned out to
599 be an unpopular choice (sorry Dscho).
604 * nd/multiple-work-trees (2015-01-27) 38 commits
605 (merged to 'next' on 2015-02-18 at b51f696)
606 + t2026 needs procondition SANITY
607 + git-checkout.txt: a note about multiple checkout support for submodules
608 + checkout: add --ignore-other-wortrees
609 + checkout: pass whole struct to parse_branchname_arg instead of individual flags
610 + git-common-dir: make "modules/" per-working-directory directory
611 + checkout: do not fail if target is an empty directory
612 + t2025: add a test to make sure grafts is working from a linked checkout
613 + checkout: don't require a work tree when checking out into a new one
614 + git_path(): keep "info/sparse-checkout" per work-tree
615 + count-objects: report unused files in $GIT_DIR/worktrees/...
616 + gc: support prune --worktrees
617 + gc: factor out gc.pruneexpire parsing code
618 + gc: style change -- no SP before closing parenthesis
619 + checkout: clean up half-prepared directories in --to mode
620 + checkout: reject if the branch is already checked out elsewhere
621 + prune: strategies for linked checkouts
622 + checkout: support checking out into a new working directory
623 + use new wrapper write_file() for simple file writing
624 + wrapper.c: wrapper to open a file, fprintf then close
625 + setup.c: support multi-checkout repo setup
626 + setup.c: detect $GIT_COMMON_DIR check_repository_format_gently()
627 + setup.c: convert check_repository_format_gently to use strbuf
628 + setup.c: detect $GIT_COMMON_DIR in is_git_directory()
629 + setup.c: convert is_git_directory() to use strbuf
630 + git-stash: avoid hardcoding $GIT_DIR/logs/....
631 + *.sh: avoid hardcoding $GIT_DIR/hooks/...
632 + git-sh-setup.sh: use rev-parse --git-path to get $GIT_DIR/objects
633 + $GIT_COMMON_DIR: a new environment variable
634 + commit: use SEQ_DIR instead of hardcoding "sequencer"
635 + fast-import: use git_path() for accessing .git dir instead of get_git_dir()
636 + reflog: avoid constructing .lock path with git_path
637 + *.sh: respect $GIT_INDEX_FILE
638 + git_path(): be aware of file relocation in $GIT_DIR
639 + path.c: group git_path(), git_pathdup() and strbuf_git_path() together
640 + path.c: rename vsnpath() to do_git_path()
641 + git_snpath(): retire and replace with strbuf_git_path()
642 + path.c: make get_pathname() call sites return const char *
643 + path.c: make get_pathname() return strbuf instead of static buffer
645 A replacement for contrib/workdir/git-new-workdir that does not
646 rely on symbolic links and make sharing of objects and refs safer
647 by making the borrowee and borrowers aware of each other.
651 --------------------------------------------------
654 * tc/t9001-noxmailer (2015-01-30) 1 commit
655 . t9001: use older Getopt::Long boolean prefix '--no' rather than '--no-'
657 Kyle's update to explicitly support --no-option with older
658 Getopt::Long ($gmane/263203) makes this unnecessary.