okay espie@ (MAINTAINER), went through a couple of KDE4 builds.
This should been committed a while ago. Even more, I was thinking that this
patch *was* committed at Oct 22... It's probably better late than never.
No point in trying to build it for days before seeing it failing horribly
because hppa/atomic ops/webkit... there might be an easy fix/workaround;
but right now hope is lost.
context-dependent attackers to cause a denial of service (memory
consumption) via an XML Entity Expansion (XEE) attack
From upstream via Arch Linux
OK espie@ (maintainer)
The change which attempted to fix QTBUG-15319 broke keyboard shortcuts
for non latin keyboard layouts.
This patch reverts QTBUG-15319 (f45cdeda8) since it caused a
regression.
From upstream via Arch Linux.
OK espie@ (maintainer)
privilege-separated X and caused broken displays for applications using
qt frames, diff from dcoppa, okay espie.
(An alternative fix would be to disable shm support, i.e.
"CONFIGURE_ARGS += -no-mitshm", I'd prefer this but it would need further
testing).
via Arch Linux).
Feedback from zhuk@, espie@, and Amit Kulkarni.
ABI/API compatibility analysis done with abi-compliance-checker.
ok espie@ (MAINTAINER)
a new libmysqlclient non-blocking API which utilizes co-routines. The X86
specific GCC ASM co-routine support hid the fact that there was an issue.
The only fallback code so far is POSIX user contexts which OpenBSD does not
support.
Input from and Ok sthen@ jasper@
was present at configure time, so the build is non-deterministic.
(notably audio/musique build fails unless qt4 was built with gtk
theme support). Add gtk+2 as a build dep to avoid this. Sync WANTLIB
while there. ok espie@
using it crashes right away at startup, since it tries to preallocate
1Gb of memory.
From webkit's patch-Source_JavaScriptCore_jit_ExecutableAllocatorFixedVMPool_cpp
tested with mnemosyne & arora.
load a specific version of the library to avoid ABI breakages, however it uses
upstream shared lib versions which don't match those in our ports. This has
already been patched where it tries to open gtk+2/cups by removing the version
check. See ports@ thread "UPDATE: kdiff3-0.9.97" where this was discussed with
fgsch, dcoppa, espie and myself, upshot is this isn't totally clean but isn't
really worse than what we had before and it fixes use of icu4c from Qt4.
This diff from fgsch, testing/ok by me, ok espie.
- since we no longer have GCC3_ARCHES, simplify, no way it's gonna compile
on a GCC2_ARCH anyways (size matters)
- so, we have only SHARED_ONLY arches, remove those frags as well.
- fold back qt4-sqlite since sqlite is part of core now.
I had a diff adding proper LIB_DEPENDS-* and WANTLIB- entries,
but espie@ prefers to just disable pulseaudio support for now,
because he's working on a qt4 update.
ok espie@
Unbreak the no_examples PSEUDO_FLAVOR.
Use the recommended dist site to provide for faster and more localized
downloads (from Brad).
Tested in a bulk build by landry@
OK landry@, sthen@
