The constructor of QMouseEvent actually expects windowPos and/or screenPos but never globalPos.
By only supplying a localPos we have Qt figure it out correctly on its own.
Differential Revision: https://phabricator.kde.org/D22851
- removed useless "virtual" specifications
- removed useless destructors
- cleanup variable initializations
- moved protected methods as private when possible for better encapsulation
Summary:
In Qt 4, the compiler detection macros have no value, so they can't be compared
against. So just use the compiler-specific ones.
It also makes use of the fact that undefined tokens evaluate to 0 in
comparisons.
Test Plan: jriddell built it with GCC 5 and 7 against Qt 4 and 5.
Reviewers: #breeze, davidedmundson
Reviewed By: davidedmundson
Subscribers: plasma-devel
Tags: #plasma
Differential Revision: https://phabricator.kde.org/D12277
Summary:
Qt has its own macro to tell the compiler that the lack of the break
statement is intentional. One of the advantages of using Q_FALLTHROUGH
is that it explicitly says other developers that the break statement is
missing on purpose. Another advantage is that we can use directives,
see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77817
Test Plan:
Compiled Breeze with GCC 7 and Clang, no "this statement may fall
through" warnings whatsoever.
Reviewers: #breeze, davidedmundson
Reviewed By: davidedmundson
Subscribers: plasma-devel
Tags: #plasma
Differential Revision: https://phabricator.kde.org/D11860
- only resize proxy widget on mouse press
- do not clear splitter in leave event if mouseGrabber is still set
- do not disable/enable updates while resizeing the splitter handle
Kai Uwe Broulik
Hugo Pereira Da Costa
Fabian Vogt
Vlad Zagorodniy
Albert Astals Cid
Kevin Funk
Hugo Pereira Da Costa