It seems like migrating to Qt6 is a win for multiple efforts, and I want us to evaluate switching to it as an alternative to monkey-patching Qt5/qmake to make it work properly.
- First class support for Apple Silicon for #20371 (see: https://bugreports.qt.io/browse/QTBUG-85279)
- Transition from
qmake
tocmake
(see: https://www.qt.io/blog/qt-6-build-system)- We already use
cmake
in depends to build several other tools cmake
is much better maintained thanqmake
- This might make a native/build split possible for
qt
, the last holdout package of native/build mixing afterboost
was recently split - For PRs like #20600, where we’re trying to work around
qmake
’s jankiness, I don’t want people to pour their heart and soul into makingqmake
work just to be completely replaced bycmake
when we bump in the future
- We already use
Supposedly, the switching cost is also minimal, as there’s a Qt5Compat module (see: https://www.qt.io/blog/porting-from-qt-5-to-qt-6-using-qt5compat-library)