Electrolysis/Meetings/2015-02-05

From MozillaWiki
Jump to: navigation, search


gw280

  • bug 1036682 - upped the number of fds allowed by our code and resolved the bug. no reports of any crashes since afaik.
  • bug 1121791 - spent a lot of time tracing through nsWindowWatcher here and found the root cause of the issue, and mconley and I realised that his patch for bug 1047603 would fix it as a result.
  • now working on profiling and finding the issues with the spinner showing (ie - paint times taking too long)

== ejpbruel == (Devtools)

  • Will join the meeting halfway in due to devtools team meeting
  • The following bugs are now fixed:
 * Bug 1067145 - Get Responsive design mode working in E10S
  • I need somebody to look into platform issues for the following bugs:
 * Bug 1107949 - The framerate actor stops recording after a tab navigation
   * Needs confirmation + STR from jsantell
 * Bug 1093535 - nsIRequest.suspend throws NS_ERROR_NOT_AVAILABLE
   * This is blocking us from enabling some debugger tests
 * Bug 1029451 - Synthesized mouseover events are blocked
   * This is blocking us from enabling the layout view tests
  • How important is bug 1011663 - CPOWs should be inspectable?
    • mconley: imo, not that important right now - though having the console / debugger highlight CPOWs (as opposed to calling isCrossProcessWrapper manually) would be super-nice. Probably for add-on authors too.

jimm

  • bug 1116884 (m5, crash in WaitForSingleObjectEx)- this is now a meta, taken off m5 list
    • various child abort fixes have landed, reducing child aborts by about 40%
    • bug 1129575 (70% of remaining aborts, PPluginWidget's RecvCreate) - patch up for review
    • bug 1128457 (35% of remaining aborts, RecvGetCookieString) -> over to blake
    • bug 1128454 (15% of remaining aborts, RecvLoadPlugin) - open
    • there's a short tail of rare aborts people can investigate, see bug 1116884
  • bug 1103177 (native plugin windows torn out of tabs) - wip

Mossop

  • about:tabcrashed changes landed
  • Landed bug 1040158: Allow updating multiprocessCompatible from update.rdf
  • Landed bug 1083281 (Using flags to control where about: and chrome: pages open)
  • Bug 1068186: Fix search engine installation
  • Bug 1129003: Fixing CPOW jank in page-mod

mconley

  • M5 - bug 1077168 - Installing an app on marketplace fails with e10s
    • Have a patch up, still waiting for review from felipe and myk
  • M5 - bug 1112243 - [e10s] <select> dropdowns are blank, if there are many options (e.g. at http://www.bls.gov/data/inflation_calculator.htm )
    • Fixed, landed
  • M5 - bug 1047603 - [e10s] Non-remote tabs in e10s windows do not handle target="_blank" or window.open links properly
    • Landed yesterday, bounced due to unexpected test failures
    • Have a solution for the Marionette failures that require a change in some Marionette code - waiting for cmanchester to get in to consult
    • Trying to sort out the new unexpected sessionstore failure.
    • Happily, it turns out this will also fix the bug that gw280 was working on wrt opened window sizes
  • M5 - bug 1090448 - Make e10s printing work on Linux
    • I have a WIP up that gets printing working on Linux
    • There are still a bunch of special settings that I need to serialize / deserialize, but I think it's all grunt work from here on out
    • Got briefly distracted by a memory leak that I thought I caused - turns out this leak has been around since at least Firefox 4. Filed bug 1129557 about that.
  • M5 - bug 1066812 - e10s: child process continues running if parent process crashes
    • On deck once Linux printing has been sorted out - I've started initial research on this, and tentatively I think this should be straight-forward.
  • Becoming the expert on unsafe CPOW usage in Firefox starts next week.

mrbkap

  • Working on bug, doing reviews, nothing too exciting

gabor

  • Worked on Bug 1090592 mainly and learned more about cpows

« previous week | index | next week »