https://bugs.kde.org/show_bug.cgi?id=470484
ratijas <m...@ratijas.tk> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |aleix...@kde.org,
| |vlad.zahorod...@kde.org,
| |xaver.h...@gmail.com
Status|REPORTED |CONFIRMED
Ever confirmed|0 |1
--- Comment #3 from ratijas <m...@ratijas.tk> ---
here,
> #5 QtWayland::org_kde_kwin_dpms::set(unsigned int) (mode=<optimized out>,
this=0x10) at /usr/src/debug/libkscreen/build/src/libdpms/qwayland-dpms.cpp:126
`this` is a null pointer, although incremented by 0x10 probably to account for
nested object layout within inheritance hierarchy of C++:
> class Dpms : public QObject, public QtWayland::org_kde_kwin_dpms
Perhaps, this call should be guarded?
> void WaylandDpmsHelper::trigger(KScreen::Dpms::Mode mode, const QList<QScreen
> *> &screens)
> {
> // ...
> for (auto screen : screens) {
> auto dpms = m_dpmsManager->fetch(screen);
> dpms->set(level);
> }
Because having an output per se does not seem to guarantee the existence of a
Dpms object available for fetch:
> Dpms *fetch(QScreen *screen)
> {
> return m_dpmsPerScreen.value(screen);
> }
>
> private:
> void addScreen(QScreen *screen)
> {
> // We can't rely on checking the wl_output being null yet
> // https://codereview.qt-project.org/c/qt/qtwayland/+/464669
> const bool fake = screen->geometry().isEmpty() ||
> screen->name().isEmpty();
> if (fake) {
> return;
> }
>
> QPlatformNativeInterface *native = qGuiApp->platformNativeInterface();
> wl_output *output = reinterpret_cast<wl_output
> *>(native->nativeResourceForScreen(QByteArrayLiteral("output"), screen));
> if (output) {
> m_dpmsPerScreen[screen] = new Dpms(get(output), m_dpms, screen);
> }
> }
As you can see there are several conditions under which `m_dpmsPerScreen` hash
map would be left without a value for some given screen.
--
You are receiving this mail because:
You are watching all bug changes.
