Revert py3.15 support, given that dev-python/matplotlib doesn't have it.
I wish pkgcheck would have reported that...
Signed-off-by: Michał Górny <mgorny@gentoo.org>
There's an option to link directly to pipewire now, which we have
no reason not to use on a source-based distro.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
The option name kind of make it sound like we're gimping the package
by disabling a performance option, but it's a quality vs performance
deal rather than both at once. Even if qtcanvaspainter gets packaged,
at best it'd be a off-by-default flag to match what upstream currently
does (that's subject to change though).
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
Starting with Qt 6.12, qtwebengine is being split from the main Qt
packaging and may or may not receive (security) releases not in sync
with the main Qt version.
It is now being versioned as 6.<chromium-version>.<patch> and a 6.140
branch exists (6.12 branch seems to have stopped receiving updates).
This will complicate a few things given eclass/ebuilds/scripts been
assuming Qt versions all match, quite possible overlooked something.
At least, for ebuilds with lower bounds, e.g. 6.140.0 is still >=6.11
and so it shouldn't conflict, the check just kind of loses meaning.
qt6-build.eclass will also need to fix the path SRC_URI tarballs, but
leaving this alone for now to see how Qt will handle this for the -rc.
qtwebengine itself has a find_package for Qt >= 6.8 and so in theory
doesn't need to depend on Qt >=6.12 or be in sync. Unfortunately need
a binding qtbase:= for rebuilds every Qt bumps due to private symbols.
As much as possible will still try to keep qtwebengine bumps and
stabilizations in sync with Qt to minimize rebuilds. Keeping Qt6's
stabilization-group as-is because of that (subject to change if the
release schedule becomes completely disconnected).
..
Note that 6.140 currently fails to build due to seemingly incomplete
security backports, won't bother patching at this stage and see if
Qt upstream handles it soon (lot or all(?) of the issues are already
fixed in 6.9999).
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
In Qt 6.12, qtwebengine will use version 6.<chromium-version>.<patch>
and no longer needs to be in sync with the rest of Qt. So it should
makes no sense to try to block it here esp. using qtbase's PV.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
In Qt 6.12, qtwebengine is being split from the main Qt packaging
and will be versioned as 6.<chromium-version>.<patch> which makes
a version check based on Qt's main version irrelevant. Will revisit
adding a lower bound only if pyqt6-webengine requires it or if it
starts using the same type of versioning to match (sounds likely).
No need for a revbump given e.g. qtwebengine-6.140.0 is still >=6.11.
The dependency on 6.11 is also not technicaly gone given pyqt6-6.11.x
requires it which we still depend on.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
While looking at 6.12 changes, noticed that Qt added a check for
ShaderTools -- but that wasn't due to new usage, and had always
been needed.
Had low chances to be missing because qtdeclarative has it in BDEPEND,
and then (if USE=qml, default on desktop profiles) qtquick3d further
has it in RDEPEND even for binpkgs.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>