from gnome bug #669260
gmain: block child sources when blocking the parent
When blocking a source that has child sources, we need to consider the
children blocked as well. Otherwise they will still trigger repeatedly
in an inner loop started from the parent source's callback.
Fixes download dialog not rendered upon certain mimetypes on midori,
see https://bugs.launchpad.net/midori/+bug/780133
Note that some paths in gsettings have been deprecated
(/apps/, /desktop/ and /system/) which means that glib-compile-schemas
will now complain about them (just a warning).
It is up to the application itself to fix its schemas.
This module updates access and modification timestamps, creating
nonexistent files where necessary. In a fashion similar to touch(1).
from Abel Abraham Camarillo Ojeda
ok sthen merdely gsoares
zap groff while here.
0.64 Sat Dec 19, 2009, joern
Bugfixes:
- Accumulated group member errors were not reported to
the frontend.
- Use Perl's builtin command execution with a subshell
instead of forking.
The 1.0 serie hasn't seen an update for 5 years and this release
contains lots of fixes for new version desktops.
While here, remove the hook stuff; that is what aliases or ~/bin are for.
Params::Classify provides various type-testing functions. These are
intended for functions that, unlike most Perl code, care what type of
data they are operating on. For example, some functions wish to behave
differently depending on the type of their arguments (like overloaded
functions in C++).
MooseX::NonMoose allows for easily subclassing non-Moose classes with
Moose, taking care of the annoying details connected with doing this,
such as setting up proper inheritance from Moose::Object and installing
(and inlining, at make_immutable time) a constructor that makes sure
things like BUILD methods are called.
MooseX::MarkAsMethods allows one to easily mark certain functions as
Moose methods. This will allow other packages such as
namespace::autoclean to operate without blowing away your overloads.
After using MooseX::MarkAsMethods your overloads will be recognized by
Class::MOP as being methods, and class extension as well as composition
from roles with overloads will "just work".
Often you want to write a persistant daemon that has a pid file, and
responds appropriately to Signals. MooseX::Daemonize provides a set of
basic roles as an infrastructure to do that.
"require EXPR" only accepts "Class/Name.pm" style module names, not
"Class::Name". How frustrating! For that, we provide "load_class
'Class::Name'".
It's often useful to test whether a module can be loaded, instead of
throwing an error when it's not available. For that, we provide
"try_load_class 'Class::Name'".
Finally, sometimes we need to know whether a particular class has been
loaded. Asking %INC is an option, but that will miss inner packages and
any class for which the filename does not correspond to the package
name. For that, we provide "is_class_loaded 'Class::Name'".
Updated from 1.04 port in Andreas Voegele's repository.
''Currently, several ports that could be built with
"CONFIGURE_STYLE=modinst" are built with "CONFIGURE_ARGS += --skipdeps"
as they require a more recent version of Module::Install.''
WWW::Mechanize 1.70. Required by the recent www/p5-WWW-Mechanize update
as well as the Catalyst update.
More recent versions of Test::WWW::Mechanize depend on LWP 6.02, which
hasn't hit OpenBSD's ports tree yet.
From Andreas Voegele's repo.
- built from the official upstream release
- needed for upcoming gnome/js update
- this is not an update to lang/spidermonkey, as that new version was
built upon Ffx 4 js engine and probably only works on amd64/i386/ppc.
It builds on sparc64 but is broken at runtime. Not marking BROKEN-* to
allow depending ports to build. Other archs untested.
- special care was taken to ensure it doesnt conflict with
lang/spidermonkey (install versionned binaries/headers..)
- ports wanting to use it should use devel/spidermonkey>=1.8,<1.9 to
ensure the correct version is picked up
Tested in an amd64 bulk build.
ok/prodding ajacoutot@
- rename libxul{,-embedding}.pc files to libxul19{,-embedding}.pc to
cope with an eventual xulrunner built from latest mozilla
- move some WANTLIB-devel from Makefile.inc to 1.9/Makefile
- use BASE_PKGPATH instead of devel/xulrunner/${MOZILLA_VERSION:R:R}, as
there might be a devel/xulrunner/last someday
- sanitize the subst regexps done on the .pc files, half of them were
useless
- bump REVISIONs.
Tested in a full bulk build with all xulrunner users..
For the theory of Memoization, please see the Memoize module
documentation. This module implements an expiry policy for Memoize
that follows LRU semantics, that is, the last n results, where n
is specified as the argument to the CACHESIZE parameter, will be
cached.
This module always exports a single function, Dumper, which can be
called with an array of values to dump those values.
It exists, fundamentally, as a convenient way to reproduce a set
of Dumper options that we've found ourselves using across large
numbers of applications, primarily for debugging output.
The principle guiding theme is "all the concision you can get while
still having a useful dump and not doing anything cleverer than
setting Data::Dumper options" - it's been pointed out to us that
Data::Dump::Streamer can produce shorter output with less lines of
code. We know. This is simpler and we've never seen it segfault.
But for complex/weird structures, it generally rocks. You should
use it as well, when Concise is underkill. We do.
Why is deparsing on when the aim is concision? Because you often
want to know what subroutine refs you have when debugging and because
if you were planning to eval this back in you probably wanted to
remove subrefs first and add them back in a custom way anyway. Note
that this -does- force using the pure perl Dumper rather than the
XS one, but I've never in my life seen Data::Dumper show up in a
profile so "who cares?".
These flags types were originally incorrectly handled in glib as being
enums. That bug was fixed, but they're still enums here, leading to
warnings about the mismatch.
Change them to flags.
just set CONFIGURE_ARGS on mips64el, don't provide a way to build with
--english-only on other arch (as the only reason for using this seems to
be to work around compiler/toolchain problems). Avoids oddity with
out-of-date reported by fgsch@.
Also make sure the arch check is done after including bsd.port.arch.mk;
ARCH was in my environment when I tested before, sigh... this problem
reported by kili@.
"it's shorter than the previous version, so it's automagically ok" kili@ ;)
This extension provides a fiber/coroutine implementation for nodejs. It
also ships with a futures implementation that wraps existing nodejs
async functions, allowing synchronous style code that handles exceptions
properly and doesn't block the nodejs event loop.
OK sthen@
building support for other languages). Use it on mips64el by default
as ld(1) has trouble without it. Information from Brian Callahan and
Bryan Irvine.
- while there, prefer http MASTER_SITES
ok landry@
- add MODPY_BADEGGS to fix cleaning as non-root
PyWBEM is a Python library for making CIM operations over HTTP using the
WBEM CIM-XML protocol. It is based on the idea that a good WBEM client
should be easy to use and not necessarily require a large amount of
programming knowledge. PyWBEM is suitable for a large range of tasks
from simply poking around to writing web and GUI applications.
checking if they're defined is wrong as -1 is a valid value to indicate
the feature is not supported. No package bumps as the code sections in
question are not being built at the moment.
from Brad
Iniparse is an INI parser for Python which is API compatible with the
standard library's ConfigParser, preserves structure of INI files (order
of sections & options, indentation, comments, and blank lines are
preserved when data is updated), and is more convenient to use.
Log4cplus is a package for providing a logging framework for C++ based
on log4j ideas.
It provides various filters, run-time re-configurations, Wide
Character (UNICODE) support, log levels, hierarchal loggers, NDC
(Nested Diagnostic Context), and log rotation support. It can log
to the console, files, syslog, Windows events, or via sockets.
Note that this is a devel version but the new stable is expected within a
week or so. Committing now so that I have time to deal with the breakage
and can start the upcoming gnome3 update.
Please contact me _directly_ if you find any issue with it.
simple solution for using all kinds of Input Devices (Keyboards,
Mice, etc) and feedback devices (e.g. force feedback).
Written in C++ using Object Oriented Design pattern
with tweaks from kirby@, mpi@ and Antti Harri.
ok kirby@
the previous commit, and bump. pointed out by landry@, thanks!
--/--
update to p5-YAML-XS (YAML-LibYAML) 0.38 and add a SECURITY patch to fix
format string problems; https://rt.cpan.org/Public/Bug/Display.html?id=46507
--/--
from CPAN. It provides its own runtests() method for the case where
you need to create an archive of test results generated by running
multiple harnesses sequentially.
ok landry@
exactly the same way except for one detail. In addition to outputting
a running progress of the tests and an ending summary it can also
capture all of the raw TAP from the individual test files or streams
into an archive file (".tar" or ".tar.gz").
ok landry@
- using /dev/arandom instead of /dev/urandom is not needed anymore
- #undef _POSIX_THREAD_PRIORITY_SCHEDULING is not needed since the
switch to rthreads.
from brad.
Tested on alpha amd64 hppa i386 powerpc sparc64, no change in regression
output (well, dtoa still fails, but thats not surprising given
corresponding code is heavily patched)
parts of gnome want vte3, and yet some other stuff (py-vte) still wants vte2.
if we let vte2->vte3 because of gnome, then py-vte will complain, and the
update won't happen.
considering we've set things up so vte2 and vte3 can coexist, the thing to do
is have vte2 update to vte2-newer, let gnome3 pick and install vte3 as well,
keeping py-vte happy and ending with a fully updated machine.
On the way, need to tell quirks AND vte about vte-pyhelper, which vanished
into vte2/vte3 (and again, pick one).
(triggers a bug in pkg_add caching which requires -current, btw ;( )
Solves update issue reported by Mikolaj.
Ports STILL fully locked, this making it in as a vital update thing,
(no don't ask, your silly little update won't make it)
okayed by naddy@
SEGFAULT when using g_closure() on at least amd64; unclear whether this
is a gcc/binutils/glib... bug yet but we want to be on the safe side for
the release.
Hopefully someone will be able to figure this out, I haven't...
Thanks to Stuart for teaching me PROPERTIES.
ok sthen@ naddy@
adaptation of the two-continuation implementation found in the paper
"Backtracking, Interleaving, and Terminating Monad Transformers"
available here: http://okmij.org/ftp/papers/LogicT.pdf.
ok ajacoutot@
of a set together with an associative binary operation. A semigroup
generalizes a monoid in that there might not exist an identity
element. It also (originally) generalized a group (a monoid with
all inverses) to a type where every element did not have to have
an inverse, thus the name semigroup.
ok ajacoutot@
optimized for performance critical use, both in terms of large data
quantities and high speed.
The declared cost of each operation is either worst-case or amortized,
but remains valid even if structures are shared.
ok ajacoutot@
It can load a MANIFEST file that comes in a Perl distribution tarball,
examine the contents, and perform some simple tasks. It can also load
the MANIFEST.SKIP file and check that.
from LEVAI Daniel with tweeks from me. feedback and ok jasper@
This module simply interpolates option file hints in @ARGV by the
contents of the pointed files. This enables option reading from files
instead of or additional to the usual reading from the command line.
ok sthen@
This is a requirement for an update of hs-snap I'm working on.
While here, add myself as maintainer (in addition to jim@, who is
too busy with private stuff for now).
There's a newer version available, but 0.7.1 is much less intrusive
to other depending ports (i.e. it doesn't break anything).
"go for it" jim@ (for co-maintaining and working on snap and related ports)
converted to a hash value. This class exists for the benefit of
hashing-based data structures. The package provides instances for
basic types and a way to combine hash values.
ok dcoppa@
fixes for Modules/FindRuby.cmake
fix wrong use of IS_ABSOLUTE into Modules/GenerateExportHeader.cmake
fix Source/cmLocalUnixMakefileGenerator3.cxx: exit the loop when
we have determined the language
As a consequence, explicitely disable the FAM backend.
This bring a huge improvement compared to what we've been using before
(gio-fam+gamin): file / directory monitoring and notifications are next
to instantaneous without any cpu overhead like it was before.
This means we can also drop the bootstrap FLAVOR of gamin that was
needed because of the cyclic dependency (gamin->glib2->gamin).
