mozilla :: #fx-team

8 Sep 2017
11:42makmikedeboer: not sure if known, but opening the Downloads view from the Library panel is quite slower than the other views, it just seems to take too much
11:42makProbably cause we show so many
11:43mikedeboermak: yeah, perhaps - I just landed the limit=42 followup
11:44makmikedeboer: is that more or less than what is in Nightly?
11:44mikedeboermak: less... Nightly is unlimited
11:44makI see
11:44makWill check the next nightly then
11:54mikedeboeralright, thanks
12:17makhm, the bookmarks toolbar dropmarker is completely broken...
12:17makit appears in the tabs toolbar...
12:18makdao: any idea which change may have done that?
12:29* mak suspects one of the recent toolbar restyles
12:31daomak: it's on the right toolbar over here
12:32makdao: uh, strange I can reproduce in currently Nightly and a custom build :\ Win10x64
12:33makdao: I mean the dropmarker when reordering bookmarks on the toolbar, to be clear
12:36Aryxdao: hi, failures for 1396205 backout incoming
14:21RyanVMman, customize mode is so much snappier than it used to be
16:26mikedeboernhnt11: ping
16:44mikedeboernhnt11: unoing
16:57mconleygandalf: ping
16:58gandalfmconley: pong
16:59mconleygandalf: hi! I notice you've requested review on bug 1397365. Does that mean you've sorted the test failures all out?
16:59firebot ASSIGNED, Set nodefaultsrc on initialBrowser in tabbrowser.xml
16:59gandalfugh, no, sorry. MozReview is resetting it every time I update the patch
17:00gandalflet me remove that
17:00mconleyprobably because you have r?mconley in the commit message.
17:00gandalfI'll remove it in the next update
17:01gandalfmconley: but if you had a chance to look at any of the failures and suggest what might be going on, it might unblock me! :)
17:01mconleygetting there
17:01mconleygoing through reviews before needinfo's
17:05dolskemikedeboer / mconley: what triggers the "oops we're having trouble restoring your session" warning when E10S is in play? If the content process crashes, do we ignore that for this purpose? (i.e. only show this if the chrome process is crashing). Or do content crashes still cause it?
17:05mconleyI think parent process crashes cause it, because (as I recall), that message means that we didn't successfully persist the last session to disk
17:05mconleyprobably because we crashed before we could finish
17:06mconleyand so we're attempting to restore from a back-up
17:06mconleymikedeboer (or Yoric or ttaubert) can correct me if I'm wrong there
17:06mconleydolske: ^--
17:07mikedeboermconley, dolske That's correct. The content process crashing is not triggering that behavior
17:10dolskecool, thanks. so the good news is that users are presumably seeing this less often, since a lot of crashes are presumably content-only.
17:12mikedeboerperhaps... until I finish my evil plan to restore sessions on startup by default, starting with OSX :)
17:22mconleydolske: yeah, that seems likely.
17:33mconleybobowen: ping
17:33bobowenmconley: pong
17:34mconleybobowen: hey, I'm reviewing dao's patch in bug 1394765, and I have some questions about how it impacts bug 1351358
17:34firebot ASSIGNED, dao+bmo Remote type set by gBrowserInit.onDOMContentLoaded is bogus when restoring a session
17:34firebot FIXED, bobowencode Can't submit form to http(s) URL using POST method from a file:// page
17:34mconleybobowen: was that code just an attempt to avoid creating a useless content process?
17:35mconleythe code being removed
17:35bobowenmconley: yes and generally it would have worked ... if I hadn't screwed up that check
17:36bobowenmconley: although I think he's removed the nsIArray code as well
17:36bobowenmconley: not sure what impact that might have
17:36mconleybobowen: right - so, yeah, I get the string removal one, since I guess that branch of the code is never reached
17:37bobowenmconley: yeah, although my feeling is we should have fixed that, but I'm not clear on the issue he thinks it's causing (or contributing to) with session restore
17:38mconleydao: ping
17:38mconleybobowen: okay, thanks
17:40* mconley re-reads bug
17:59daomconley: pong
17:59mconleydao: hey, n/m - just put it in a bug comment. See
17:59firebotBug 1394765 ASSIGNED, Remote type set by gBrowserInit.onDOMContentLoaded is bogus when restoring a session
18:00mconleydao: To be clear, I see that this string check is busted, but shouldn't we just fix the string check, or am I missing something? I'm having difficulty tying it together with bobowen's work in the other bug.
18:02daomconley: fixing the string check means we'd do the wrong thing when restoring a session, as we won't load the homepage in that case
19:12jawsbesides ntim, who else is working on the janitor and might be able to get me an invite?
19:13johannhjaws: janx
19:13jawsjanx: hi, could i get an invite for the janitor?
19:14johannh(I think he's in my time zone, so might not be available right now)
19:18jawsyeah and on a friday at that :)
19:18mconleyjaws: hey - so I'm trying to figure out what we should do for bug 1397092
19:18firebot ASSIGNED, Severe page loading delays in Firefox 57 32-bit builds compared to Firefox 57 builds from 2017-08-20
19:18mconleyI wonder if we should consider using the APNG if there are many tabs in the loading state?
19:19jawsmconley: i think figuring out that metric is the hardest part. secondly, instead of doing the APNG we could also try to drop the framerate from 60fps to 30fps and see if that is better
19:20jawsdropping the frame rate would be better in the long run instead of supporting two different implementations
19:20mconleyjaws: how hard would it be for you to give shellye a build with a 30fps ping-pong to try?
19:20jawsi'll try to put one together. i also need to get shellye a build with the APNG
19:24mconleyjaws: thanks. Gonna figure this out.
19:26Mardakmy xpcom is rusty. :( for an idl that says `[optional] in unsigned short`. when called without providing an argument, does an implementation with uint16_t get the value 0?
19:27mconleyoh geez, good question
19:28Mardakwell easy fix, we'll pass in 0 to be safe :p
19:28mconleylet's ask one of our new XPCOM peers! John-Galt, do you know?
19:28mconleyer, I guess they're XPConnect peers
19:29mconleybut hey, glue is glue
19:29Mardaki could.. just compile firefox and print out the value.. ;)
19:29mconleythat's true
19:29jawsyes lol
19:30jawsMardak: if it's uninitalized memory, then a debug build will probably print out 0. if it's an opt build you would get garbage
19:30jawsso just take that into mind when trying to understand what gets printed :)
19:30John-GaltMardak: Yes
19:30Mardakursula: ^^ :)
19:32ursulahaha passing in 0 it is!
19:33John-GaltMardak: Although, to be technically accurate, it depends on the function. If it asks for [optional_argc] it gets an explicit arg count, and can use whatever default value it wants.
20:00janxjaws: sure!
20:00jawsjanx: awesome thanks!
20:09MardakJohn-Galt: oh neat good to know. i see the argc comes in between the params and return. thanks!
9 Sep 2017
No messages
Last message: 11 days and 6 hours ago