Bug#905379: Bug explanation
On Sat, Nov 03, 2018 at 01:46:49PM +0100, Roland Hieber wrote: > On Fri, 26 Oct 2018 16:09:19 +0100 Julian Gilbey wrote: > > I've just uploaded a version based on git to unstable > > (2.1.5+dfsg+2.1.6-test.20181026.b4f4e65c-1) so you can test it. Does > > this version work for you? > > Thanks! At least it shows a functional UI without any warnings, and > editing and studying decks works, so I guess that's a yes :-) Super - thanks for letting me know! So I'll close this bug report when 2.1.6 is released. Best wishes, Julian
Bug#905379: Bug explanation
On Fri, 26 Oct 2018 16:09:19 +0100 Julian Gilbey wrote: > I've just uploaded a version based on git to unstable > (2.1.5+dfsg+2.1.6-test.20181026.b4f4e65c-1) so you can test it. Does > this version work for you? Thanks! At least it shows a functional UI without any warnings, and editing and studying decks works, so I guess that's a yes :-) - Roland
Bug#905379: Bug explanation
On Thu, Oct 25, 2018 at 11:20:02PM +0100, Roland Hieber wrote: > On Thu, 20 Sep 2018 08:23:15 +0200 Julien Puydt > wrote: > That particular part of the code was refactored in upstream commit > https://github.com/dae/anki/commit/1af556cb8d62a638f7fa0bcb6289bf932d338691 > and although I was able to backport that single commit onto 2.1.5+dfsg-1 > and that particular error went away, I still didn't get a very > functional Anki window afterwards (only a menu bar with a grey window, > but also three new tabs with data:text/html URLs in my Firefox o_O). So > I guess there is still more to it than meets the eye for full pyqt 5.11 > support. Dear Roland, Thanks so much for letting me know this - I hadn't spotted it. Woo-hoo, it works for me! I wonder whether you needed to rebuild the UI? I've just uploaded a version based on git to unstable (2.1.5+dfsg+2.1.6-test.20181026.b4f4e65c-1) so you can test it. Does this version work for you? I won't close this bug report until the next official release (2.1.6) (assuming that it works for you, of course), as I don't want the package to migrate to testing until then; it may not fully match the packaged upstream version (translations etc.). Best wishes, Julian
Bug#905379: Bug explanation
On Thu, 20 Sep 2018 08:23:15 +0200 Julien Puydt wrote: > Hi, > > I hadn't launched anki since long and hence got stuck. > > Looking at the trace I see: > self.focusProxy().installEventFilter(self) > and an attribute error that NoneType has no installEventFilter, so it's > quite clear what the problem is: > > self.focusProxy() returns None! > > That might mean that some versions of PyQT were having a default focus > proxy and that line worked, but newer versions need an explicit setup. > > Commenting the line makes it possible to go further, with a few error > windows showing up. > > I would suggest asking around. That particular part of the code was refactored in upstream commit https://github.com/dae/anki/commit/1af556cb8d62a638f7fa0bcb6289bf932d338691 and although I was able to backport that single commit onto 2.1.5+dfsg-1 and that particular error went away, I still didn't get a very functional Anki window afterwards (only a menu bar with a grey window, but also three new tabs with data:text/html URLs in my Firefox o_O). So I guess there is still more to it than meets the eye for full pyqt 5.11 support. - Roland
Bug#905379: Bug explanation
Hi, I hadn't launched anki since long and hence got stuck. Looking at the trace I see: self.focusProxy().installEventFilter(self) and an attribute error that NoneType has no installEventFilter, so it's quite clear what the problem is: self.focusProxy() returns None! That might mean that some versions of PyQT were having a default focus proxy and that line worked, but newer versions need an explicit setup. Commenting the line makes it possible to go further, with a few error windows showing up. I would suggest asking around. jpuydt on irc.debian.org