* The error message from "git push $there :bogo" (and its equivalent
"git push $there --delete bogo") mentioned that we tried and failed
to guess what ref is being deleted based on the LHS of the refspec,
which we don't.
* A handful of files and directories we create had tighter than
necessary permission bits when the user wanted to have group
writability (e.g. by setting "umask 002").
* "commit --amend" used to refuse amending a commit with an empty log
message, with or without "--allow-empty-message".
* "git commit --amend --only --" was meant to allow "Clever" people to
rewrite the commit message without making any change even when they
have already changes for the next commit added to their index, but
it never worked as advertised since it was introduced in 1.3.0 era.
* Even though the index can record pathnames longer than 1<<12 bytes,
in some places we were not comparing them in full, potentially
replacing index entries instead of adding.
* "git show"'s auto-walking behaviour was an unreliable and
unpredictable hack; it now behaves just like "git log" does when it
walks.
* "git diff", "git status" and anything that internally uses the
comparison machinery was utterly broken when the difference
involved a file with "-" as its name. This was due to the way "git
diff --no-index" was incorrectly bolted on to the system, making
any comparison that involves a file "-" at the root level
incorrectly read from the standard input.
* We did not have test to make sure "git rebase" without extra options
filters out an empty commit in the original history.
* "git fast-export" produced an input stream for fast-import without
properly quoting pathnames when they contain SPs in them.
* "git checkout --detach", when you are still on an unborn branch,
should be forbidden, but it wasn't.
* Some implementations of Perl terminates "lines" with CRLF even when
the script is operating on just a sequence of bytes. Make sure to
use "$PERL_PATH", the version of Perl the user told Git to use, in
our tests to avoid unnecessary breakages in tests.
Also contains minor typofixes and documentation updates.
mkc.subprj.mk:
- Support for "virtual" subproject was added. Subprojects listed
in SUBPRJ and SUBPRJS_DFLT are not necessarily associated with
a subdirectory. See examples/hello_superfs for example.
Virtual subproject is a way to group several subprojects into
new one.
New variables were introduced: COPTS_<proj> OBJCOPTS_<proj>
LDADD_<proj> LDFLAGS_<proj> CPPFLAGS_<proj> CXXFLAGS_<proj>.
See the manual page for details.
Fix: OBJDIR_<subdir> variables now always contain full paths.
Fix in mkc_which(1). Now it differs directories and regular files.
Minor fixes in regression tests for EXPORT_SYMBOLS.
mkc.lib.mk:
- SHLIB_MINOR unconditionally defaults to 0
More slides in .pdf presentation:
- cross-compilation
- EXPORT_SYMBOLS
Fixes for Pascal support
Minor fixes in the man page
DragonFly has gethostbyname_r and gethostbyaddr_r functions while NetBSD
does not. The function calls didn't make the prototype though. Macros
were used to provide the correct arguments and return type for DragonFly.
The library has to be specifically and non-trivially ported to each
platform it supports and currently it is not ported to DragonFly.
Moveover, FreeBSD's USB stack, which uses the same API, is intended to
ported to DragonFly in the near future.
The library has to be specifically and non-trivially ported to each
platform it supports and currently it is not ported to DragonFly.
Moveover, FreeBSD's USB stack, which uses the same API, is intended to
ported to DragonFly in the near future.
Fixes PR pkg/46551 (which also does not contain the requisite NEWSy
commit message text.
(Upstream has no news, and no changelog. Upstream author will be sent
a bug report about the lack of NEWS.)
Changelog:
NEW
Google searches now utilize HTTPS
NEW
Full screen support for Mac OS X Lion implemented
NEW
Plugins can now be configured to only load on click (requires an about:config change)
NEW
The Awesome Bar now auto-completes typed URLs
CHANGED
Improved site identity manager, to prevent spoofing of an SSL connection with favicons
DEVELOPER
Pointer Lock API implemented
DEVELOPER
New API to prevent your display from sleeping
DEVELOPER
New text-transform and font-variant CSS improvements for Turkic languages and Greek
FIXED
Various security fixes
FIXED
GIF animation can gets stuck when src and image size are changed (743598)
FIXED
OS X: nsCocoaWindow::ConstrainPosition uses wrong screen in multi-display setup (752149)
FIXED
CSS :hover regression when an element's class name is set by Javascript (758885
* On Cygwin, the platform pread(2) is not thread safe, just like our
own compat/ emulation, and cannot be used in the index-pack
program. Makefile variable NO_THREAD_SAFE_PREAD can be defined to
avoid use of this function in a threaded program.
* "git add" allows adding a regular file to the path where a
submodule used to exist, but "git update-index" does not allow an
equivalent operation to Porcelain writers.
* "git archive" incorrectly computed the header checksum; the symptom
was observed only when using pathnames with hi-bit set.
* "git blame" did not try to make sure that the abbreviated commit
object names in its output are unique.
* Running "git bundle verify" on a bundle that records a complete
history said "it requires these 0 commits".
* "git clone --single-branch" to clone a single branch did not limit
the cloning to the specified branch.
* "git diff --no-index" did not correctly handle relative paths and
did not correctly give exit codes when run under "--quiet" option.
* "git diff --no-index" did not work with pagers correctly.
* "git diff COPYING HEAD:COPYING" gave a nonsense error message that
claimed that the treeish HEAD did not have COPYING in it.
* When "git log" gets "--simplify-merges/by-decoration" together with
"--first-parent", the combination of these options makes the
simplification logic to use in-core commit objects that haven't
been examined for relevance, either producing incorrect result or
taking too long to produce any output. Teach the simplification
logic to ignore commits that the first-parent traversal logic
ignored when both are in effect to work around the issue.
* "git ls-files --exclude=t -i" did not consider anything under t/ as
excluded, as it did not pay attention to exclusion of leading paths
while walking the index. Other two users of excluded() are also
updated.
* "git request-pull $url dev" when the tip of "dev" branch was tagged
with "ext4-for-linus" used the contents from the tag in the output
but still asked the "dev" branch to be pulled, not the tag.
Also contains minor typofixes and documentation updates.
2.0402nb1 to 2.0603.
pkgsrc changes:
- adjust build dependencies (META.json informs much better now)
Upstream changes:
2.0603 Thu, Jun 28, 2012
[BUG FIXES]
* Fix test failure in blead. RT #78085.
2.0602 Mon, May 07, 2012
[BUG FIXES]
* Ensure that the Moose::Exporter-generated init_meta returns the same value
that it did previously. This isn't really a bug, since the return value has
never been tested or documented, but since the generated init_meta is
nothing more than a compatibility shim at this point, there's no reason to
not make it as compatible as possible. Reported by Moritz Onken. (doy)
2.0601 Tue, May 01, 2012
[BUG FIXES]
* Fix init_meta order when multiple also packages are specified (this matters
when one of them is being used to actually initalize the metaclass,
typically with also => 'Moose'). Reported by Randy Stauner. (doy)
2.0600 Sun, Apr 29, 2012
[OTHER]
* Releasing 2.0502 as stable.
2.0502-TRIAL Wed, Apr 25, 2012
[OTHER]
* The Test::DependentModules test now covers a much wider range of downstream
dependents (all of them in fact, for some definition of "all"). This should
allow us to track inadvertent backwards compatibility breakages much more
effectively. (doy)
* A few test tweaks to avoid spurious failures. (doy)
2.0501-TRIAL Tue, Apr 03, 2012
[BUG FIXES]
* Avoid syntax errors on pre-5.14. (doy)
2.0500-TRIAL Tue, Apr 03, 2012
[NEW FEATURES]
* Class::MOP::Class now has methods for introspecting and modifying the
overloaded operators for a class. (doy)
[ENHANCEMENTS]
* The cookbook recipes have all been renamed. Instead of numbered recipes
(Basics::Recipe1), we now have descriptive names
(Basics::Point_AttributesAndSubclassing). This makes it easier for us to
add and remove recipes in the future, and makes it a little easier to
converse about them, since the name gives us some clue of what they
contain.
[BUG FIXES]
* Re-declaring a class_type or role_type constraint that has already been
declared now just returns the original type constraint, rather than
replacing the original constraint and ergo losing any coercions that were
on the original constraint. Fixes RT #73289. (t0m)
* Moose::Exporter now calls init_meta methods in the correct order, when
multiple levels of 'also' parameters are specified. Reported by Rocco
Caputo. (doy, perigrin)
* Moose::Exporter no longer generates init_meta methods in order to apply
metaroles, since the metaclass itself isn't guaranteed to exist yet at that
point. Metaroles are now applied at the end of import, after all
user-defined init_meta methods have been called. Fixes RT #51561. (doy)
* Fixed a memory leak. This occurred when creating an anonymous
class. Immutabilizing an anonymous class still leaks memory due to a bug in
Eval::Closure (which should hopefully be fixed soon). Based on code and bug
report from Carlos Lima. RT #74650.
* Fix a segfault when adding a method to a class which was defined in a
package which was deleted. (doy)
2.0403 Tue, Apr 03, 2012
[OTHER]
* No changes, reupload to fix indexing.
0.091014.
pkgsrc changes:
- add newly introduced dependencies
upstream changes:
0.091014 - 2012-07-16
- load overload.pm explicitly for overload::StrVal
0.091013 - 2012-07-15
- useful and detailed errors for coerce in attrib generation
0.091012 - 2012-07-15
- useful and detailed errors for default checker in attrib generation
- throw an error when trying to extend a role
0.091011 - 2012-06-27
- re-add #web-simple as development IRC
- don't assume Scalar::Util is imported into the current package
0.091010 - 2012-06-26
- isa checks on builders
- additional quote_sub docs
- remove multi-populate code to fix exists/defined new() bug
- document move to #moose and include repository metadata
- no Moo and no Moo::Role
- squelch used only once warnings for $Moo::HandleMoose::MOUSE
- MooClass->meta
- subconstructor handling for Moose classes
0.091009 - 2012-06-20
- squelch redefine warnings in the coderef installation code
0.091008 - 2012-06-19
- bump Role::Tiny dependency to get working modifiers under composition
- handle "has '+foo'" for attrs from superclass or consumed role
- document override -> around translation
- use D::GD if installed rather than re-adding it as a requirement
0.091007 - 2012-05-17
- remove stray reference to Devel::GlobalDestruction
0.091006 - 2012-05-16
- drop a couple of dependencies by minor releases we don't strictly need
0.091005 - 2012-05-14
- temporary switch to an inlined in_global_destruction to avoid needing
to fatpack Sub::Exporter for features we don't use
- re-order is documentation to give readonly styles more prominence
- a weakened value should still be returned on set (fixes lazy + weak_ref)
- add an explicit return to all exported subs so people don't accidentally
rely on the return value
0.091004 - 2012-05-07
- also inhale from Mouse
- clarify how isa and coerce interact
- support isa and coerce together for Moose
- guard _accessor_maker_for calls in Moo::Role in case Moo isn't loaded
- reset handlemoose state on mutation in case somebody reified the
metaclass too early
0.091003 - 2012-05-06
- improve attribute option documentation
- update the incompatibilities section since we're less incompatible now
- fix coderef naming to avoid confusing autoclean
0.091002 - 2012-05-05
- exclude union roles and same-role-as-self from metaclass inflation
- inhale Moose roles before checking for composition conflicts
- enable Moo::sification if only Moo::Role is loaded and not Moo
- preserve attribute ordering
- factor out accessor generation code a bit more to enable extension
0.091001 - 2012-05-02
- bump Role::Tiny dependency to require de-strictures-ed version
- fix test failure where Class::XSAccessor is not available
0.091000 - 2012-04-27
- document MX::AttributeShortcuts 009+ support
- documentation for the metaclass inflation code
- better error message for broken BUILDARGS
- provide 'no Moo::sification' to forcibly disable metaclass inflation
- switch to Devel::GlobalDestruction to correctly disarm the
Moo::sification trigger under threads
- make extends after has work
- name subs if Sub::Name is available for better stracktraces
- undefer all subs before creating a concrete Moose metaclass
- fix bug in _load_module where global vars could cause mis-detection
of the module already being loaded
0.009_017 - 2012-04-16
- mangle constructor meta-method on inflation so make_immutable works
- fix possible infinite loop caused by subconstructor code
0.009_016 - 2012-04-12
- don't accidentally load Moo::HandleMoose during global destruction
- better docs for trigger (and initializer's absence)
0.009_015 - 2012-04-11
- Complete support for MooseX::AttributeShortcuts 0.009
- Allow Moo classes to compose Moose roles
- Introduce Moo::HandleMoose, which should allow Moo classes and roles
to be treated as Moose classes/roles. Supported so far:
- Some level of attributes and methods for both classes and roles
- Required methods in roles
- Method modifiers in roles (they're already applied in classes)
- Type constraints
devel/p5-Role-Tiny as requirement for upcoming update for devel/p5-Moo.
Role::Tiny is a minimalist role composition tool.
ROLE COMPOSITION
Role composition can be thought of as much more clever and meaningful
multiple inheritance. The basics of this implementation of roles is:
* If a method is already defined on a class, that method will not be
composed in from the role.
* If a method that the role "requires" to be implemented is not
implemented, role application will fail loudly.
Unlike Class::C3, where the last class inherited from "wins," role
composition is the other way around, where the class wins. If multiple
roles are applied in a single call (single with statement), then if any
of their provided methods clash, an exception is raised unless the class
provides a method since this conflict indicates a potential problem.
Changelog:
What's new in 1.447.2 (2012/06/11)
Guice injector failure can cause failure of whole Jenkins (issue 13448)
Jenkins runs out of file descriptors (winstone problem) (issue 9882)
Parsing of POM happens before SNAPSHOT-Parents are updated (issue 8663)
Loading All Build History Fails (issue 13238)
Changes in version 2.0.19-stable (3 May 2012)
* Refactor event_persist_closure: raise and extract some common logic
* If time has jumped so we'd reschedule a periodic event in the past, schedule
it for the future instead
* If a higher-priority event becomes active, don't continue running events
of the current priority.
* Fixed potential double-readcb execution with openssl bufferevents.
* Cancel a probe request when the server is freed, and ignore cancelled probe
callbacks
* Remove redundant DNS_ERR_CANCEL check, move comment
* When retransmitting a timed-out DNS request, pick a fresh nameserver.
* Backport: provide EVENT_LOG_* names, and deprecate _EVENT_LOG_*
Changes in version 2.0.18-stable (22 Mar 2012)
* Make uses of open() close-on-exec safe by introducing an internal
evutil_open_closeonexec.
* Properly zero the kevent in kq_setup_kevent()
* Stop crashing in evdns when nameserver probes give a weird error
Changes in version 2.0.17-stable (10 Feb 2012)
* Be absolutely sure to clear pncalls before leaving event_signal_closure
* check for sysctl before we use it
* Remove bogus casts of socket to int before calling ev_callback
* Make evconnlistener work around bug in older Linux when getting nmapped
* Fix a list corruption bug when using event_reinit() with signals present
* Fix a fd leak in event_reinit()
* Do a memberwise comparison of threading function tables
* Use C-style comments in C source files (for compatibility with compilers
such as xlc on AIX).
* Avoid crash when freeing event_iocp and using event_set_mem_functions
* In the kqueue backend, do not report EBADF as an EV_READ
* Fix behavior of evbuffer_peek(buf,-1,NULL,NULL,0)
* Loop on filtering SSL reads until we are blocked or exhausted.
* Force strict validation of HTTP version in response.
* evdns: fix a bug in circular-queue implementation
* Backport evhttp_connection_get_bufferevent to Libevent 2.0
Changes in version 2.0.16-stable (18 Nov 2011)
* More detailed message in case of libevent self-debugging failure.
* epoll: close fd on alloc fail at initialization
* Fix compile warning from saying event2/*.h inside a comment
* Warn when unable to construct base because of failing make_base_notifiable
* Don't try to make notifiable event_base when no threading fns are configured
* unit test for remove_buffer bug
* Fix an evbuffer crash in evbuffer_remove_buffer()
* Refactor amount-to-read calculations in buffervent_ssl consider_reading()
* Move SSL rate-limit enforcement into bytes_to_read()
* Avoid spinning on OpenSSL reads
* Empty DNS reply with OK status is another way to say NODATA.
Changes in version 2.0.15-stable (12 Oct 2011)
* DNS: add ttl for negative answers using RFC 2308 idea.
* Add DNS_ERR_NODATA error code to handle empty replies.
* Make evbuffer callbacks get the right n_added value after evbuffer_add
* Prefer mmap to sendfile unless a DRAINS_TO_FD flag is set. Allows add_file
to work with SSL.
* When a signal callback is activated to run multiple times, allow
event_base_loopbreak to work even before they all have run.
== 1.2.11.4 / 2012-07-03
- Embed PropertySet meta data GUIDs and field lists, to avoid hitting the
filesystem and remove dependency on YAML.
- Update Rakefile to avoid warnings about both deprecated tasks and space
before parentheses.
- Remove Dirent#children=.
PLIST.cs and PLIST.ci for case sensitive/insensitive switch
was not honoured (Mac OS X issue).
The related lines are really revert to revision 1.28.
No PKGREVISION bump required, already packaged one is not affected.
On a 32-bit OpenIndiana 151a machine, glib2 failed with the error:
"large files are not supported by libelf".
The glib2 makefile doesn't include libelf from pkgsrc, it uses the version
with the header at /usr/include/libelf.h. Libelf will only support large
files (64 bits) with 64-bit versions of SunOS.
This change adds additional checks. When the Makefile determines that the
package is building on a 32-bit SunOS, it will force the configure script
to define _FILE_OFFSET_BITS as 32 which allows glib2 to build completely.
No need to bump revision; it wouldn't have built on 32-bit SunOS before.
Packaged by Jaap Boender.
Extension to OCaml for deriving functions from type declarations. Includes
derivers for pretty-printing, type-safe marshalling with structure-sharing,
dynamic typing, equality, and more.
Overview of changes in Glib 1.260 (stable)
==========================================
* Tell CPAN to ignore POD fragments in Makefile.PL
* Document that SOURCE_CONTINUE and _REMOVE can be exported
Overview of changes in Glib 1.253
=================================
* Export the constants SOURCE_CONTINUE and SOURCE_REMOVE on request.
(RT #48070)
* Create and register a GType for GConnectFlags.
Overview of changes in Glib 1.252
=================================
* Make the recent thread-safety improvements work for non-threaded perls too.
Overview of changes in Glib 1.251
=================================
* Make signal marshalling more thread-safe. When a signal handler is invoked
from a foreign thread without associated Perl interpreter, hand the
marshalling over to the main loop which in turn later wakes up the main
thread and lets it handle the request. Since this approach is experimental,
there is debug print for now whenever it is used:
*** GPerl asked to invoke callback from a foreign thread;
handing it over to the main loop
Overview of changes in Glib 1.250
=================================
* Add a fallback implementation of SvMAGIC_set; Bugzilla bug #665266
* Avoid a syntax error on older perls
* Glib::Object: make ref-counting compatible with perl >= 5.16 (RT#73650)
* Fix some typos in POD (RT#73616)
Experimental version released on July 10th, 2012.
This is the first public release of the kyua-atf-compat package.
This first release is paired with the 0.16 release of ATF, which
disables the build of atf-run and atf-report by default. The goal
of this package is, then, to provide compatibility utilities for
these disabled tools but using Kyua as the backend for the execution
of the tests.
Experimental version released on July 10th, 2012.
* Issue 15: Added automatic stacktrace gathering of crashing test cases.
This relies on GDB and is a best-effort operation.
* Issue 32: Added the '--build-root' option to the debug, list and test
commands. This allows executing test programs from a different
directory than where the Kyuafile scripts live. See the 'Build roots'
section in the manual for more details.
* Issue 33: Removed the kyuaify.sh script. This has been renamed to
atf2kyua and moved to the kyua-atf-compat module, where it ships as a
first-class utility (with a manual page and tests).
* Issue 34: Changed the HTML reports to include the stdout and stderr of
every test case.
* Fixed the build when using a "build directory" and a clean source tree
from the repository.
Experimental version released on July 10th, 2012.
* Added a --enable-tools flag to configure to request the build of the
deprecated ATF tools, whose build is now disabled by default. In order
to continue running tests, you should migrate to Kyua instead of enabling
the build of the deprecated tools. The kyua-atf-compat package provides
transitional compatibility versions of atf-run and atf-report built on
top of Kyua.
* Tweaked the ATF_TEST_CASE macro of atf-c++ so that the compiler can
detect defined but unused test cases.
* PR bin/45859: Fixed some XSLT bugs that resulted in the tc-time and
tp-time XML tags leaking into the generated HTML file. Also improved
the CSS file slightly to correct alignment and color issues with the
timestamps column.
* Optimized atf-c++/macros.hpp so that GNU G++ consumes less memory during
compilation with GNU G++.
* Flipped the default to building shared libraries for atf-c and atf-c++,
and started versioning them. As a side-effect, this removes the
--enable-unstable-shared flag from configure that appears to not work any
more (under NetBSD). Additionally, some distributions require the use of
shared libraries for proper dependency tracking (e.g. Fedora), so it is
better if we do the right versioning upstream.
* Project hosting moved from an adhoc solution (custom web site and
Monotone repository) to Google Code (standard wiki and Git). ATF now
lives in a subcomponent of the Kyua project.
