This is https://bugreports.qt.io/browse/QTBUG-101649
I don't see why qdoc shouldn't repeat a failed lookup for \relates T \fn foo(x) as if \fn T::foo(x) were given, but I guess we all need to first learn to work around it, and then change back once it's fixed. But qdoc being weird should _not_ be a reason to make it non-hidden. We want hidden friends very much, because the alternative is // foo.cpp #include <QtCore> class Foo {}; void f() { QDataStream ds; Foo foo; ds << foo; $ ninja foo | wc -l 518 HTH, Marc On 25.07.22 21:27, Thiago Macieira wrote: In https://codereview.qt-project.org/c/qt/qtbase/+/422690 Build: https://testresults.qt.io/coin/integration/qt/qtbase/tasks/1658767511 /home/qt/work/qt/qtbase/src/corelib/kernel/qmetatype.cpp:3032: (qdoc) warning: clang couldn't find function when parsing \fn int qRegisterMetaType(QMetaType meta) The function exists because it's unit-tested and compiled. Why can't qdoc see it? It's a hidden friend, but that's valid C++ and clang should see it.
_______________________________________________ Development mailing list Development@qt-project.org https://lists.qt-project.org/listinfo/development