mozilla :: #e10s

20 Mar 2017
09:53Enzohey, I'd like to know why I can't check the box to enable e10s on 54.0a2 (2017-03-20) (64-bit)?
13:08pbrogabor: thanks for your email. I'm going with your suggestion for now: I've removed the prefs and pushing to try again, to see if these devtools tests still fail or not
13:09gaborpbro: thanks a lot for your help, I should have done that part already before asking for help, just were busy with the rest of the tests and a bunch of other urgent to fix bugs :(
13:13pbrogabor: sure, no problem. I'll check later what try says, and I'll go from there
13:15gaborpbro: sounds good
14:41pbrogabor: hmm, this try build is so green it's hard for me to believe :) https://treeherder.mozilla.org/#/jobs?repo=try&revision=a400f2339f744a693a5aa7de919ac6179e31b2c6&group_state=expanded Do you mind checking quickly if I ran the right tests? I did run the devtools mochitests, across platforms and e10s/non-e10s. But should I be running something else to make
14:41pbrosure these tests work in e10s-multi?
14:42mstangewhat happens if you have two instances of Firefox open, one of them updates and restarts itself, and in the other one you start a new content process? seems like you'd get an inconsistent session with an updated content process binary and an outdated parent process binary
14:42mstangeoh, I see what happens: MOZ_RELEASE_ASSERT(parentBuildID == childBuildID)
15:00gaborpbro: Awesome! I think because of blassey's patch that removed a bunch of startup sync messaging... or something else fixed it... a bunch of other tests that used to fail seem to be working correctly now so it's entirely possible that these tests just work now. The patch looks correct to me so we might as well land it then :) Could you upload it under bug
15:00gabor1301015 and flag me with r?
15:02pbrogabor: will do, thanks
17:07The_8472mstange, https://bugzilla.mozilla.org/show_bug.cgi?id=1112937
17:07firebotBug 1112937 ASSIGNED, robert.strong.bugs@gmail.com Firefox can update while still running when using multiple profiles concurrently
17:07mstangeThe_8472: thanks!
17:34spohlsmaug: ping
17:35smaugspohl: pong
17:36spohlsmaug: hi there, I was able to get the response from the child regarding whether or not a WidgetKeyboardEvent was handled. however, I now ran into a different problem. if the event wasn't handled by the child, we need to dispatch the native event to the native menu bar.
17:36spohlsmaug: unfortunately, the native event cannot be copied across process boundaries, see https://dxr.mozilla.org/mozilla-central/source/widget/nsGUIEventIPC.h#440
17:37smaugand you need that? You can't regenerated such native event from the WidgetKeyboardEvent ?
17:38spohlsmaug: hmm, regenerate native event that would indeed be the easiest solution I think
17:39spohlsmaug: I'm not sure if we have places where we do this already. I'll take a look
17:39smaugmstange might recall
17:40smaugI mean whether we do that elsewhere in OSX specific code
17:40mstangeI don't believe we do
17:40mstangewe create new events for sendNativeEvent
17:40mstangeand we do something for passing native events to plugins
17:40mstangeI forget how we solved that particular problem
17:42spohlmstange, smaug: I'll take a look. thanks for the heads up
19:29aklotzhandyman: there is a conversation that might be beginning in #accessibility that might be relevant to your interests
19:30handymanaklotz: thanks
20:00katsaklotz: ping
20:00aklotzkats: pong
20:00katsaklotz: hey, e10s was enabled for 32-bit accessibility users in 52, right?
20:01aklotzkats: For touchscreen users, yes, but not true a11y
20:01katsaklotz: ah ok
20:01katsthanks
20:01aklotznp
21 Mar 2017
No messages
   
Last message: 154 days and 2 hours ago