Instead of checking that PKGSRC_COMPILER is equal to gcc, see if gcc
is contained in it. Thus, if PKGSRC_COMPILER is "ccache gcc" or
"ccache distcc gcc", the hack will still be applied. (As a side
effect, I can confirm that the hack is still needed on recent netbsd-7
i386.)
-----------------------------
News in 11.89
=============
* You can now run all commands needed to compile a document and then
open the viewer with a single command: 'TeX-command-run-all', bound
to 'C-c C-a'.
* Commands such as LaTeX and View can now be executed conveniently on
the current section (or part, chapter, subsection, etc). See
'LaTeX-command-section' and 'LaTeX-command-section-change-level'.
* Forward and backward search with Evince now also work when only a
region of the document is compiled/viewed.
* To open the PDF output file you can now use also PDF Tools, a
document viewer for Emacs. With it, as a plus, forward and
backward search is accurate at word level.
* With new option 'TeX-PDF-via-dvips-ps2pdf' it is possible to
compile a document to DVI and then convert it to PDF using
'dvips'-'ps2pdf' before viewing it.
* New option 'TeX-file-line-error' allows to select file:line:error
style for error messages.
* Indent '\[...\]' math mode as a regular environment by default.
* Now AUCTeX suggests to run 'makeindex' when appropriate.
* 'TeX-view-program-list' can contain, as third optional element of
each item, the name of the executable(s) needed to open the viewer.
* 'TeX-expand-list' variable has been split into 'TeX-expand-list'
and 'TeX-expand-list-builtin'. Only the former is intended to be
customized by the user, the latter contains built-in expanders.
You might want to keep in 'TeX-expand-list' only new expansion
strings.
* Before running commands like TeX and LaTeX, now AUCTeX performs
some checks. If 'TeX-check-TeX' is non-nil, it will test whether a
working TeX distribution is actually present in the system and
available to Emacs. Instead, when 'TeX-check-engine' is non-nil,
before running LaTeX commands AUCTeX will check whether the correct
engine has been set, based upon known restrictions posed by LaTeX
packages.
* Basic support to ConTeXt Mark IV has been added. Users can now
select the Mark version to be used with new option
'ConTeXt-Mark-version', and AUCTeX is able to catch error messages
in the output log of a Mark IV document.
* Support for tons of LaTeX packages has been addedd.
* Numbers of bugs have been fixed, many minor features have been
addedd.
* Remove dbus-glib dependency and add dbus option (from Robert Swindells)
* Fix potential build failure in skia (from Robert Swindells)
Changelog:
Fixed
Fix an audio regression impacting some major websites (bug 1295296)
Fix a top crash in the JavaScript engine (Bug 1290469)
Fix a startup crash issue caused by Websense (Bug 1291738)
Fix a different behavior with e10s / non-e10s on <select> and mouse events (Bug 1291078)
Fix a top crash caused by plugin issues (Bug 1264530)
Fix an unsigned add-ons issue on Windows
Fix a shutdown issue (Bug 1276920)
Fix a crash in WebRTC
lf (as in "list files") is a terminal file manager written in Go. It is heavily
inspired by ranger with some missing and extra features. Some of the missing
features are deliberately ommited since it is better if they are handled by
external tools.
Features
- no external runtime dependencies (except for terminfo database)
- fast startup and low memory footprint (due to native code and static
binaries)
- server/client architecture to share selection between multiple instances
- custom commands as shell scripts (hence any other language as well)
- sync (waiting and skipping) and async commands
- fully customizable keybindings
Non-Features
- tabs or windows (handled by the window manager or the terminal multiplexer)
- built-in pager (handled by your pager of choice)
Termbox is a library that provides a minimalistic API which allows the
programmer to write text-based user interfaces. The library is crossplatform
and has both terminal-based implementations on *nix operating systems and a
winapi console based implementation for windows operating systems. The basic
idea is an abstraction of the greatest common subset of features available on
all major terminals and other terminal-like APIs in a minimalistic fashion.
Small API means it is easy to implement, test, maintain and learn it, that's
what makes the termbox a distinct library in its area.
2016-08-19 Richard Russon <rich@flatcap.org>
* Contrib
- Updates to Keybase Support
Joshua Jordi (JakkinStewart)
* Bug Fixes
- Fix data-loss when appending a compressed file
- Don't paint invisible progress bars
- Revert to Mutt keybindings
- Don't de-tag emails after labelling them
- Don't whine if getrandom() fails
Adam Borowski (kilobyte)
- Fix display when 'from' field is invalid
* Config
- Support for $XDG_CONFIG_HOME and $XDG_CONFIG_DIRS
Marco Hinz (mhinz)
* Docs
- Fix DocBook validation
- Document NotMuch queries
* Build
- More Autoconf improvements
Darshit Shah (darnir)
- Create Distribution Tarballs with autogen sources
Darshit Shah (darnir)
2016-08-08 Richard Russon <rich@flatcap.org>
* New Features
- Timeout Hook - Run a command periodically
- Multiple fcc - Save multiple copies of outgoing mail
* Contrib
- Keybase Integration
Joshua Jordi (JakkinStewart)
* Devel
- Attached - Prevent missing attachments
Darshit Shah (darnir)
- Virtual Unmailboxes - Remove unwanted virtual mailboxes
Richard Russon (flatcap)
* Bug Fixes
- Sidebar's inbox occasionally shows zero/wrong value
- Fix crash opening a second compressed mailbox
* Config
- Look for /etc/NeoMuttrc and ~/.neomuttrc
* Docs
- Fix broken links, typos
- Update project link
- Fix version string in the manual
* Build
- Add option to disable fmemopen
- Install all the READMEs and contribs
- Big overhaul of the build
Darshit Shah (darnir)
The purpose of this package is to generate random (plausible) text
sentences and paargraphs based on a dictionary and a sample text.
By default this package will generate Lorem Ipsum style text, but
you can customize the generator to effectively load any dictionary
and any sample text you like.
Perlin noise is ubiquitous in modern CGI. Used for procedural
texturing, animation, and enhancing realism, Perlin noise has been
called the "salt" of procedural content. Perlin noise is a type of
gradient noise, smoothly interpolating across a pseudo-random matrix
of values.
The noise library includes native-code implementations of Perlin
"improved" noise and Perlin simplex noise. It also includes a fast
implementation of Perlin noise in GLSL, for use in OpenGL shaders.
Important note about pre-1.0.4 potential repo corruption
--------------------------------------------------------
Some external errors (like network or disk I/O errors) could lead to
corruption of the backup repository due to issue #1138.
A sign that this happened is if "E" status was reported for a file that can
not be explained by problems with the source file. If you still have logs from
"borg create -v --list", you can check for "E" status.
Here is what could cause corruption and what you can do now:
1) I/O errors (e.g. repo disk errors) while writing data to repo.
This could lead to corrupted segment files.
Fix::
# check for corrupt chunks / segments:
borg check -v --repository-only REPO
# repair the repo:
borg check -v --repository-only --repair REPO
# make sure everything is fixed:
borg check -v --repository-only REPO
2) Unreliable network / unreliable connection to the repo.
This could lead to archive metadata corruption.
Fix::
# check for corrupt archives:
borg check -v --archives-only REPO
# delete the corrupt archives:
borg delete --force REPO::CORRUPT_ARCHIVE
# make sure everything is fixed:
borg check -v --archives-only REPO
3) In case you want to do more intensive checking.
The best check that everything is ok is to run a dry-run extraction::
borg extract -v --dry-run REPO::ARCHIVE
Version 1.0.7 (2016-08-19)
--------------------------
Security fixes:
- borg serve: fix security issue with remote repository access, #1428
If you used e.g. --restrict-to-path /path/client1/ (with or without trailing
slash does not make a difference), it acted like a path prefix match using
/path/client1 (note the missing trailing slash) - the code then also allowed
working in e.g. /path/client13 or /path/client1000.
As this could accidentally lead to major security/privacy issues depending on
the pathes you use, the behaviour was changed to be a strict directory match.
That means --restrict-to-path /path/client1 (with or without trailing slash
does not make a difference) now uses /path/client1/ internally (note the
trailing slash here!) for matching and allows precisely that path AND any
path below it. So, /path/client1 is allowed, /path/client1/repo1 is allowed,
but not /path/client13 or /path/client1000.
If you willingly used the undocumented (dangerous) previous behaviour, you
may need to rearrange your --restrict-to-path pathes now. We are sorry if
that causes work for you, but we did not want a potentially dangerous
behaviour in the software (not even using a for-backwards-compat option).
Bug fixes:
- fixed repeated LockTimeout exceptions when borg serve tried to write into
a already write-locked repo (e.g. by a borg mount), #502 part b)
This was solved by the fix for #1220 in 1.0.7rc1 already.
- fix cosmetics + file leftover for "not a valid borg repository", #1490
- Cache: release lock if cache is invalid, #1501
- borg extract --strip-components: fix leak of preloaded chunk contents
- Repository, when a InvalidRepository exception happens:
- fix spurious, empty lock.roster
- fix repo not closed cleanly
New features:
- implement borg debug-info, fixes#1122
(just calls already existing code via cli, same output as below tracebacks)
Other changes:
- skip the O_NOATIME test on GNU Hurd, fixes#1315
(this is a very minor issue and the GNU Hurd project knows the bug)
- document using a clean repo to test / build the release
Version 1.0.7rc2 (2016-08-13)
-----------------------------
Bug fixes:
- do not write objects to repository that are bigger than the allowed size,
borg will reject reading them, #1451.
Important: if you created archives with many millions of files or
directories, please verify if you can open them successfully,
e.g. try a "borg list REPO::ARCHIVE".
- lz4 compression: dynamically enlarge the (de)compression buffer, the static
buffer was not big enough for archives with extremely many items, #1453
- larger item metadata stream chunks, raise archive item limit by 8x, #1452
- fix untracked segments made by moved DELETEs, #1442
Impact: Previously (metadata) segments could become untracked when deleting data,
these would never be cleaned up.
- extended attributes (xattrs) related fixes:
- fixed a race condition in xattrs querying that led to the entire file not
being backed up (while logging the error, exit code = 1), #1469
- fixed a race condition in xattrs querying that led to a crash, #1462
- raise OSError including the error message derived from errno, deal with
path being a integer FD
Other changes:
- print active env var override by default, #1467
- xattr module: refactor code, deduplicate, clean up
- repository: split object size check into too small and too big
- add a transaction_id assertion, so borg init on a broken (inconsistent)
filesystem does not look like a coding error in borg, but points to the
real problem.
- explain confusing TypeError caused by compat support for old servers, #1456
- add forgotten usage help file from build_usage
- refactor/unify buffer code into helpers.Buffer class, add tests
- docs:
- document archive limitation, #1452
- improve prune examples
Version 1.0.7rc1 (2016-08-05)
-----------------------------
Bug fixes:
- fix repo lock deadlocks (related to lock upgrade), #1220
- catch unpacker exceptions, resync, #1351
- fix borg break-lock ignoring BORG_REPO env var, #1324
- files cache performance fixes (fixes unneccessary re-reading/chunking/
hashing of unmodified files for some use cases):
- fix unintended file cache eviction, #1430
- implement BORG_FILES_CACHE_TTL, update FAQ, raise default TTL from 10
to 20, #1338
- FUSE:
- cache partially read data chunks (performance), #965, #966
- always create a root dir, #1125
- use an OrderedDict for helptext, making the build reproducible, #1346
- RemoteRepository init: always call close on exceptions, #1370 (cosmetic)
- ignore stdout/stderr broken pipe errors (cosmetic), #1116
New features:
- better borg versions management support (useful esp. for borg servers
wanting to offer multiple borg versions and for clients wanting to choose
a specific server borg version), #1392:
- add BORG_VERSION environment variable before executing "borg serve" via ssh
- add new placeholder {borgversion}
- substitute placeholders in --remote-path
- borg init --append-only option (makes using the more secure append-only mode
more convenient. when used remotely, this requires 1.0.7+ also on the borg
server), #1291.
Other changes:
- Vagrantfile:
- darwin64: upgrade to FUSE for macOS 3.4.1 (aka osxfuse), #1378
- xenial64: use user "ubuntu", not "vagrant" (as usual), #1331
- tests:
- fix fuse tests on OS X, #1433
- docs:
- FAQ: add backup using stable filesystem names recommendation
- FAQ about glibc compatibility added, #491, glibc-check improved
- FAQ: 'A' unchanged file; remove ambiguous entry age sentence.
- OS X: install pkg-config to build with FUSE support, fixes#1400
- add notes about shell/sudo pitfalls with env. vars, #1380
- added platform feature matrix
- implement borg debug-dump-repo-objs
Version 1.0.6 (2016-07-12)
--------------------------
Bug fixes:
- Linux: handle multiple LD_PRELOAD entries correctly, #1314, #1111
- Fix crash with unclear message if the libc is not found, #1314, #1111
Other changes:
- tests:
- Fixed O_NOATIME tests for Solaris and GNU Hurd, #1315
- Fixed sparse file tests for (file) systems not supporting it, #1310
- docs:
- Fixed syntax highlighting, #1313
- misc docs: added data processing overview picture
Version 1.0.6rc1 (2016-07-10)
-----------------------------
New features:
- borg check --repair: heal damaged files if missing chunks re-appear (e.g. if
the previously missing chunk was added again in a later backup archive),
#148. (*) Also improved logging.
Bug fixes:
- sync_dir: silence fsync() failing with EINVAL, #1287
Some network filesystems (like smbfs) don't support this and we use this in
repository code.
- borg mount (FUSE):
- fix directories being shadowed when contained paths were also specified,
#1295
- raise I/O Error (EIO) on damaged files (unless -o allow_damaged_files is
used), #1302. (*)
- borg extract: warn if a damaged file is extracted, #1299. (*)
- Added some missing return code checks (ChunkIndex._add, hashindex_resize).
- borg check: fix/optimize initial hash table size, avoids resize of the table.
Other changes:
- tests:
- add more FUSE tests, #1284
- deduplicate fuse (u)mount code
- fix borg binary test issues, #862
- docs:
- changelog: added release dates to older borg releases
- fix some sphinx (docs generator) warnings, #881
Notes:
(*) Some features depend on information (chunks_healthy list) added to item
metadata when a file with missing chunks was "repaired" using all-zero
replacement chunks. The chunks_healthy list is generated since borg 1.0.4,
thus borg can't recognize such "repaired" (but content-damaged) files if the
repair was done with an older borg version.
Version 1.0.5 (2016-07-07)
--------------------------
Bug fixes:
- borg mount: fix FUSE crash in xattr code on Linux introduced in 1.0.4, #1282
Other changes:
- backport some FAQ entries from master branch
- add release helper scripts
- Vagrantfile:
- centos6: no FUSE, don't build binary
- add xz for redhat-like dists
Version 1.0.4 (2016-07-07)
--------------------------
New features:
- borg serve --append-only, #1168
This was included because it was a simple change (append-only functionality
was already present via repository config file) and makes better security now
practically usable.
- BORG_REMOTE_PATH environment variable, #1258
This was included because it was a simple change (--remote-path cli option
was already present) and makes borg much easier to use if you need it.
- Repository: cleanup incomplete transaction on "no space left" condition.
In many cases, this can avoid a 100% full repo filesystem (which is very
problematic as borg always needs free space - even to delete archives).
Bug fixes:
- Fix wrong handling and reporting of OSErrors in borg create, #1138.
This was a serious issue: in the context of "borg create", errors like
repository I/O errors (e.g. disk I/O errors, ssh repo connection errors)
were handled badly and did not lead to a crash (which would be good for this
case, because the repo transaction would be incomplete and trigger a
transaction rollback to clean up).
Now, error handling for source files is cleanly separated from every other
error handling, so only problematic input files are logged and skipped.
- Implement fail-safe error handling for borg extract.
Note that this isn't nearly as critical as the borg create error handling
bug, since nothing is written to the repo. So this was "merely" misleading
error reporting.
- Add missing error handler in directory attr restore loop.
- repo: make sure write data hits disk before the commit tag (#1236) and also
sync the containing directory.
- FUSE: getxattr fail must use errno.ENOATTR, #1126
(fixes Mac OS X Finder malfunction: "zero bytes" file length, access denied)
- borg check --repair: do not lose information about the good/original chunks.
If we do not lose the original chunk IDs list when "repairing" a file
(replacing missing chunks with all-zero chunks), we have a chance to "heal"
the file back into its original state later, in case the chunks re-appear
(e.g. in a fresh backup). Healing is not implemented yet, see #148.
- fixes for --read-special mode:
- ignore known files cache, #1241
- fake regular file mode, #1214
- improve symlinks handling, #1215
- remove passphrase from subprocess environment, #1105
- Ignore empty index file (will trigger index rebuild), #1195
- add missing placeholder support for --prefix, #1027
- improve exception handling for placeholder replacement
- catch and format exceptions in arg parsing
- helpers: fix "undefined name 'e'" in exception handler
- better error handling for missing repo manifest, #1043
- borg delete:
- make it possible to delete a repo without manifest
- borg delete --forced allows to delete corrupted archives, #1139
- borg check:
- make borg check work for empty repo
- fix resync and msgpacked item qualifier, #1135
- rebuild_manifest: fix crash if 'name' or 'time' key were missing.
- better validation of item metadata dicts, #1130
- better validation of archive metadata dicts
- close the repo on exit - even if rollback did not work, #1197.
This is rather cosmetic, it avoids repo closing in the destructor.
- tests:
- fix sparse file test, #1170
- flake8: ignore new F405, #1185
- catch "invalid argument" on cygwin, #257
- fix sparseness assertion in test prep, #1264
Other changes:
- make borg build/work on OpenSSL 1.0 and 1.1, #1187
- docs / help:
- fix / clarify prune help, #1143
- fix "patterns" help formatting
- add missing docs / help about placeholders
- resources: rename atticmatic to borgmatic
- document sshd settings, #545
- more details about checkpoints, add split trick, #1171
- support docs: add freenode web chat link, #1175
- add prune visualization / example, #723
- add note that Fnmatch is default, #1247
- make clear that lzma levels > 6 are a waste of cpu cycles
- add a "do not edit" note to auto-generated files, #1250
- update cygwin installation docs
- repository interoperability with borg master (1.1dev) branch:
- borg check: read item metadata keys from manifest, #1147
- read v2 hints files, #1235
- fix hints file "unknown version" error handling bug
- tests: add tests for format_line
- llfuse: update version requirement for freebsd
- Vagrantfile:
- use openbsd 5.9, #716
- do not install llfuse on netbsd (broken)
- update OSXfuse to version 3.3.3
- use Python 3.5.2 to build the binaries
- glibc compatibility checker: scripts/glibc_check.py
- add .eggs to .gitignore
Version 1.0.3 (2016-05-20)
--------------------------
Bug fixes:
- prune: avoid that checkpoints are kept and completed archives are deleted in
a prune run), #997
- prune: fix commandline argument validation - some valid command lines were
considered invalid (annoying, but harmless), #942
- fix capabilities extraction on Linux (set xattrs last, after chown()), #1069
- repository: fix commit tags being seen in data
- when probing key files, do binary reads. avoids crash when non-borg binary
files are located in borg's key files directory.
- handle SIGTERM and make a clean exit - avoids orphan lock files.
- repository cache: don't cache large objects (avoid using lots of temp. disk
space), #1063
Other changes:
- Vagrantfile: OS X: update osxfuse / install lzma package, #933
- setup.py: add check for platform_darwin.c
- setup.py: on freebsd, use a llfuse release that builds ok
- docs / help:
- update readthedocs URLs, #991
- add missing docs for "borg break-lock", #992
- borg create help: add some words to about the archive name
- borg create help: document format tags, #894
Ranger is a fast implementation of random forest (Breiman 2001) or
recursive partitioning, particularly suited for high dimensional
data. Classification, regression, probability estimation and survival
forests are supported. Classification and regression forests are
implemented as in the original Random Forest (Breiman 2001), survival
forests as in Random Survival Forests (Ishwaran et al. 2008). For
probability estimation forests see Malley et al. (2012).
New Features
Custom columns: Add a new 'Short text' column type that behaves like the builtin title column
More generally, you can now customize Comments-like columns, controlling how their headings are displayed in the Book Details panel and also specifying what kind of data they contain (plain text, HTML or even Markdown)
Edit book: Function mode: Allow appending text to the end of the marked text when running function mode search/replace on marked text
PDF Output: Add a new variable _TOP_LEVEL_SECTION_ for use in header/footer templates that resolves to the current top-most section in the Table of Contents
Linux/OSX: Shutdown gracefully on receiving the interrupt or terminate Unix signals
Bug Fixes
PDF Output: Fix a regression that caused the last page of a chapter to sometimes be dropped from the output
Do not show an error if the user tries to quit calibre while editing a cell in the book list. Instead, just silently abort the edit.
Closes tickets: 1610539
E-book viewer: Fix 0px margins and percentage margins not being respected when the option to override book margins is turned off.
Closes tickets: 1608458
Book Details: Do not show the 'Copy link' context menu item unless there is an actual link at the cursor location.
Closes tickets: 1605594
Get Books: Update store plugin for nexto
New news sources
Data News by oCkz7bJ_
Improved news sources
Private Eye
Prospect Magazine
De Standaard
Jot Down
Weblogs SL
Hola
El Correo
Fixes since v2.9.2
------------------
* A helper function that takes the contents of a commit object and
finds its subject line did not ignore leading blank lines, as is
commonly done by other codepaths. Make it ignore leading blank
lines to match.
* Git does not know what the contents in the index should be for a
path added with "git add -N" yet, so "git grep --cached" should not
show hits (or show lack of hits, with -L) in such a path, but that
logic does not apply to "git grep", i.e. searching in the working
tree files. But we did so by mistake, which has been corrected.
* "git rebase -i --autostash" did not restore the auto-stashed change
when the operation was aborted.
* "git commit --amend --allow-empty-message -S" for a commit without
any message body could have misidentified where the header of the
commit object ends.
* More mark-up updates to typeset strings that are expected to
literally typed by the end user in fixed-width font.
* For a long time, we carried an in-code comment that said our
colored output would work only when we use fprintf/fputs on
Windows, which no longer is the case for the past few years.
* "gc.autoPackLimit" when set to 1 should not trigger a repacking
when there is only one pack, but the code counted poorly and did
so.
* One part of "git am" had an oddball helper function that called
stuff from outside "his" as opposed to calling what we have "ours",
which was not gender-neutral and also inconsistent with the rest of
the system where outside stuff is usuall called "theirs" in
contrast to "ours".
* The test framework learned a new helper test_match_signal to
check an exit code from getting killed by an expected signal.
* "git blame -M" missed a single line that was moved within the file.
* Fix recently introduced codepaths that are involved in parallel
submodule operations, which gave up on reading too early, and
could have wasted CPU while attempting to write under a corner
case condition.
* "git grep -i" has been taught to fold case in non-ascii locales
correctly.
* A test that unconditionally used "mktemp" learned that the command
is not necessarily available everywhere.
* "git blame file" allowed the lineage of lines in the uncommitted,
unadded contents of "file" to be inspected, but it refused when
"file" did not appear in the current commit. When "file" was
created by renaming an existing file (but the change has not been
committed), this restriction was unnecessarily tight.
* "git add -N dir/file && git write-tree" produced an incorrect tree
when there are other paths in the same directory that sorts after
"file".
* "git fetch http://user:pass [at] hos/repo..." scrubbed the userinfo
part, but "git push" didn't.
* An age old bug that caused "git diff --ignore-space-at-eol"
misbehave has been fixed.
* "git notes merge" had a code to see if a path exists (and fails if
it does) and then open the path for writing (when it doesn't).
Replace it with open with O_EXCL.
* "git pack-objects" and "git index-pack" mostly operate with off_t
when talking about the offset of objects in a packfile, but there
were a handful of places that used "unsigned long" to hold that
value, leading to an unintended truncation.
* Recent update to "git daemon" tries to enable the socket-level
KEEPALIVE, but when it is spawned via inetd, the standard input
file descriptor may not necessarily be connected to a socket.
Suppress an ENOTSOCK error from setsockopt().
* Recent FreeBSD stopped making perl available at /usr/bin/perl;
switch the default the built-in path to /usr/local/bin/perl on not
too ancient FreeBSD releases.
* "git status" learned to suggest "merge --abort" during a conflicted
merge, just like it already suggests "rebase --abort" during a
conflicted rebase.
* The .c/.h sources are marked as such in our .gitattributes file so
that "git diff -W" and friends would work better.
* Existing autoconf generated test for the need to link with pthread
library did not check all the functions from pthread libraries;
recent FreeBSD has some functions in libc but not others, and we
mistakenly thought linking with libc is enough when it is not.
* Allow http daemon tests in Travis CI tests.
* Users of the parse_options_concat() API function need to allocate
extra slots in advance and fill them with OPT_END() when they want
to decide the set of supported options dynamically, which makes the
code error-prone and hard to read. This has been corrected by tweaking
the API to allocate and return a new copy of "struct option" array.
* The use of strbuf in "git rm" to build filename to remove was a bit
suboptimal, which has been fixed.
* "git commit --help" said "--no-verify" is only about skipping the
pre-commit hook, and failed to say that it also skipped the
commit-msg hook.
* "git merge" in Git v2.9 was taught to forbid merging an unrelated
lines of history by default, but that is exactly the kind of thing
the "--rejoin" mode of "git subtree" (in contrib/) wants to do.
"git subtree" has been taught to use the "--allow-unrelated-histories"
option to override the default.
* The build procedure for "git persistent-https" helper (in contrib/)
has been updated so that it can be built with more recent versions
of Go.
* There is an optimization used in "git diff $treeA $treeB" to borrow
an already checked-out copy in the working tree when it is known to
be the same as the blob being compared, expecting that open/mmap of
such a file is faster than reading it from the object store, which
involves inflating and applying delta. This however kicked in even
when the checked-out copy needs to go through the convert-to-git
conversion (including the clean filter), which defeats the whole
point of the optimization. The optimization has been disabled when
the conversion is necessary.
* "git -c grep.patternType=extended log --basic-regexp" misbehaved
because the internal API to access the grep machinery was not
designed well.
* Windows port was failing some tests in t4130, due to the lack of
inum in the returned values by its lstat(2) emulation.
* The characters in the label shown for tags/refs for commits in
"gitweb" output are now properly escaped for proper HTML output.
* FreeBSD can lie when asked mtime of a directory, which made the
untracked cache code to fall back to a slow-path, which in turn
caused tests in t7063 to fail because it wanted to verify the
behaviour of the fast-path.
* Squelch compiler warnings for netmalloc (in compat/) library.
* The API documentation for hashmap was unclear if hashmap_entry
can be safely discarded without any other consideration. State
that it is safe to do so.
* Not-so-recent rewrite of "git am" that started making internal
calls into the commit machinery had an unintended regression, in
that no matter how many seconds it took to apply many patches, the
resulting committer timestamp for the resulting commits were all
the same.
* "git difftool <paths>..." started in a subdirectory failed to
interpret the paths relative to that directory, which has been
fixed.
Also contains minor documentation updates and code clean-ups.
15.0.3 (2016-08-05)
Test for given python path actually being an executable file, #939
Only search for copy actual existing Tcl/Tk directories (PR #937)
Generically search for correct Tcl/Tk version (PR #926, PR #933)
Upgrade setuptools to 22.0.5
