| mwolff | |
| davidedmundson |
| Frameworks | |
| Plasma |
Third part of https://phabricator.kde.org/D10627
According to the documentation at http://doc.qt.io/qt-5/qabstractnativeeventfilter.html
The type of event eventType is specific to the platform plugin chosen
at run-time, and can be used to cast message to the right type.
On X11, eventType is set to "xcb_generic_event_t", and the message can
be casted to a xcb_generic_event_t pointer.
The other eventType are "windows_generic_MSG" and "mac_generic_NSEvent".
No other eventType starts with an 'x'.
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
what? this is totally unsafe - the reinterpret_cast below shouldn't be done on anything that has event names other than the previously used one...
ok, looking at the reasoning in the other commit:
in general, I don't see how such a comparison can be so costly - the real problem would be too many native events, no? did you maybe profile a debug build or something?
Will be done.
- you need to provide a comment in the code the clarifies what's going on here
Will be done.
in general, I don't see how such a comparison can be so costly - the real problem would be too many native events, no? did you maybe profile a debug build or something?
A qstrcmp is more expensive than a single cpu instruction in i586 builds (debug or release), unless it uses SSE4 instruction in the x64 world.
Unfortunately I haven't found what causes so many native events.
ping? I would like to use the same comment and commit message in https://phabricator.kde.org/D10669
This revision was not accepted when it landed; it landed in state Needs Review.
Weird, David accepted it though?
Yes. Shouldn't this be reported as a bug to phabricator? It has happened to me twice.
Phabricator considers reviews approved only when all individual and group reviewers have approved it.
In this instance, nobody approved on behalf of Frameworks and @mwolff has changes requested still.