app-misc/sleepyhead depends on dev-qt/qtwebkit:4 which is ancient and will never see any security updates. If there's no upstream fix available, it might be time to remove this package.
Looks like I'll need to update this to the latest version. I'd been avoiding this as they're seemingly in perpetual beta, but I'll tag a commit. It has been updated to qt5 at least.
Ok, the new version is in the tree. I'll leave this open to track the qt4 dependency until the old version is removed, but I'd like to leave it around for a bit in case the new version causes issues.
Upstream hasn't found it necessary to fix stuff in their beta for ~1 year - do you think the old version is ready for cleanup? :)
(In reply to Andreas Sturmlechner from comment #3) > Upstream hasn't found it necessary to fix stuff in their beta for ~1 year - > do you think the old version is ready for cleanup? :) I don't see that we have much choice if you plan to retire qt4. However, I can definitely vouch that the current version is useless on my CPAP. It looks like they reverted a bunch of stuff where they didn't get permission to re-license it, so it doesn't support some less-common models, including mine. I may end up setting up a container somewhere to run the old version and stop maintaining this at some point.
Created attachment 497288 [details, diff] 0001-app-misc-sleepyhead-revision-bump-ports-to-Qt-5.patch 0.9.3 builds against Qt 5 OK, please test.
Should we just add the patch and cleanup?
Sorry about the delay. I just did some testing and I'm getting segfaults after creating a profile with the patched version. Thread 1 "SleepyHead" received signal SIGSEGV, Segmentation fault. 0x00007ffff5788066 in QOpenGLFunctions::hasOpenGLFeature(QOpenGLFunctions::OpenGLFeature) const () from /usr/lib64/libQt5Gui.so.5 (gdb) bt #0 0x00007ffff5788066 in QOpenGLFunctions::hasOpenGLFeature(QOpenGLFunctions::OpenGLFeature) const () from /usr/lib64/libQt5Gui.so.5 #1 0x00007ffff7f7be11 in QGLContextPrivate::bindTexture(QImage const&, unsigned int, int, long long, QFlags<QGLContext::BindOption>) () from /usr/lib64/libQt5OpenGL.so.5 #2 0x00007ffff7f7c5b3 in QGLContextPrivate::bindTexture(QImage const&, unsigned int, int, QFlags<QGLContext::BindOption>) () from /usr/lib64/libQt5OpenGL.so.5 #3 0x00007ffff7f7c65b in QGLContext::bindTexture(QImage const&, unsigned int, int, QFlags<QGLContext::BindOption>) () from /usr/lib64/libQt5OpenGL.so.5 #4 0x00007ffff7f7c6ae in QGLWidget::bindTexture(QImage const&, unsigned int, int) () from /usr/lib64/libQt5OpenGL.so.5 #5 0x00000000005661bd in gGraphView::setCubeImage (this=0x16fe020, img=0x1262880) at Graphs/gGraphView.cpp:3287 #6 0x000000000047ecbe in Daily::Daily (this=0x1357d20, parent=0xb53650, shared=0x0) at daily.cpp:335 #7 0x00000000004c5415 in MainWindow::MainWindow (this=0x7fffffffcdb0, parent=0x0) at mainwindow.cpp:139 #8 0x0000000000427d0c in main (argc=1, argv=0x7fffffffcf68) at main.cpp:282 It loads fine if I switch back to the qt4 version. So, the patch might build but it seems to have issues. You should be able to reproduce this without any cpap data - the segfault happens just after you create a profile. We may be stuck just going to the 1.0 beta, though that has its own issues. That is basically the direction upstream is going in, so I don't really have a problem with this. I don't see a point in releasing the 0.9 patch until it works. However, I'm fine with cleaning out old versions alongside the qt4 cleanup - don't let this hold anything up.
Affected versions depending on Qt4WebKit were masked.
Cleanup done by rich0 in git commit 28d0bdd20453d231a09a8838deacf441ab7a7df5.
