mozilla :: #apz

9 Jan 2017
18:29botondkats: ping
18:29katsbotond: pong
18:30botondkats: so I upgraded my machine to Windows 10, but unfortunately, I still cannot reproduce the scrollbar dragging behaviour that you and cpeterson are reporting
18:30katshm, interesting
18:30botondkats: can you think of any relevant difference between our configurations?
18:30botondkats: i'm on a Thinkpad X1 Carbon, dragging using the touchpad
18:30katsbotond: not really. i also tried with touch events turned off on mine and i could still repro it
18:30katsi'm using an external mouse
18:31katsmaybe that's relevant
18:34botondkats: works fine for me with a mouse, too :/
18:35katsbotond: i can investigate more then
18:35botondkats: ok, thanks!
18:35botondkats: sorry i can't be of more help
18:36katsno worries
18:36botondkats: if at some point during the investigation you discover something that allows me to reproduce, my offer to take over the investigation stands
18:36katsbotond: sounds good
18:37botondkats: on the bright side, i now have Windows 10, so I should be able to repro other Win10-specific issues in the future
18:37katsalthough i hope there aren't too many :)
18:37botondlikewise :)
22:35rbarkerkats: ping? So I think I need to fix Bug 1202020 before I can turn on the GPU process. I'm sure you've paged that out long ago but do you have any thoughts before I start?
22:35firebot ASSIGNED, Remove dependencies on async pan/zoom information from Java code
22:35rbarkerI think my biggest concern is the paint ID.
22:37katsrbarker: it's all paged out and i'm heading out shortly but from what i remember the paint ID is only exposed to GeckoLayerClient
22:38katsit's not exposed to the rest of the java code
22:38katsso it shouldn't matter
22:38rbarkerkats: okay, it's just that the paint id is hitching a ride with the frame metric in SyncFrameMetric.
22:39katsrbarker: do you need to get rid of SyncFrameMetrics?
22:39rbarkerkats: I thinks so. I looks like it runs at frame rate?
22:39rbarkerI would need to send it over IPC so probably not a good idea.
22:40katsah true
22:40rbarkerI could probably use shared memory but I think killing it would be easier.
22:40katsrbarker: so then you want to do more than what's covered by bug 1202020
22:41katsrbarker: that bug excluded code in mobile/android/.../gfx/*
22:42katsrbarker: fixing bug 1202020 should be fairly easy, but removing *all* the dependencies on compositor state from java code will be harder
22:42katsprobably still doable
22:42rbarkerkats: Yeah.
22:42katsthe dynamic toolbar is probably the most problematic piece
22:42katsfeel free to rewrite that :)
22:42rbarkerkats: Yeah, but the GPU process will break it anyway I think?
22:43rbarkerSince the paint id will have to get back asynchronously?
22:44katsprobably yeah
22:44rbarkerkats: okay, thanks. Probably need to do a little more thinking on this.
22:45katsrbarker: ok
10 Jan 2017
No messages
Last message: 7 days and 19 hours ago