Electrolysis/Meetings/2016-08-11: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
(Add clarification)
(aklotz)
Line 21: Line 21:
** I landed a patch, and [https://telemetry.mozilla.org/new-pipeline/evo.html#!aggregates=median&cumulative=0&end_date=2016-08-04&keys=!WITHOUT_PREVIEW!__none__!__none__&max_channel_version=nightly%252F51&measure=FX_SESSION_RESTORE_RESTORE_WINDOW_MS&min_channel_version=nightly%252F48&product=Firefox&sanitize=1&sort_keys=submissions&start_date=2016-08-01&trim=1&use_submission_date=0 it appears to be coming down a bit]. The work in {{bug|1290280}} should probably bring this down even more.
** I landed a patch, and [https://telemetry.mozilla.org/new-pipeline/evo.html#!aggregates=median&cumulative=0&end_date=2016-08-04&keys=!WITHOUT_PREVIEW!__none__!__none__&max_channel_version=nightly%252F51&measure=FX_SESSION_RESTORE_RESTORE_WINDOW_MS&min_channel_version=nightly%252F48&product=Firefox&sanitize=1&sort_keys=submissions&start_date=2016-08-01&trim=1&use_submission_date=0 it appears to be coming down a bit]. The work in {{bug|1290280}} should probably bring this down even more.
* The usual slew of reviews and needinfos.
* The usual slew of reviews and needinfos.
== aklotz ==
* {{bug|1273635}} is currently holding up a11y work - tps hangs on Windows 8 x64 e10s. Not clear why this is happening, working on a loaner from releng in an effort to repro.
* Once that clears the remaining patches should be landable, hope to be able to point 51 Nightly to a11y client vendors for testing sometime next week.

Revision as of 15:48, 11 August 2016


« previous week | index | next week »

mconley

  • bug 1291078 - <select> change event is fired before mouse events if e10s is enabled
    • Fixed across the board, and riding in the 48.0.1 point release.
  • Have been dealing with testing fallout from bug 1261842 (make initial browser remote)
    • bug 1290186 - Intermittent test_end for test_windows.py TestBaseWindow.test_open_close logged while not in progress. Logged with data: {"status": "ERROR", "extra": {"class_name": "test_windows.TestBaseWindow", "method_name": “test_open_close"}
      • We have a non-remote window in the background that opens and closes a bunch of windows. This causes many content processes to be created and queued for destruction. Sometimes, a content process is scheduled to be killed very soon after it's first created, which means that it doesn't get a chance to initialize before the shutdown kill timer kills it. This creates a silent shutdown kill paired minidump, which mozcrash interprets as a test failure.
      • Fixed this by having the test flip a pref that disables the shutdown kill timer. We should probably do something better for short-lived content processes.
    • bug 1290567 - "ERROR: GC found live Cell ... at shutdown" errors since bug 1261842
      • Apparently after a shutdown GC, there's still stuff in the heap? Still investigating.
    • bug 1290280 - [e10s] Tab crashes on startup
      • We're doing some silly remoteness flipping for the initial browser tab when restoring windows. I have a patch that fixes this and some new tests.
      • Blocked on review, but will request uplift to Aurora once this goes through.
  • bug 1291860 - Investigate FX_SESSION_RESTORE_RESTORE_WINDOW_MS regression starting on 2016-07-26
  • The usual slew of reviews and needinfos.

aklotz

  • bug 1273635 is currently holding up a11y work - tps hangs on Windows 8 x64 e10s. Not clear why this is happening, working on a loaner from releng in an effort to repro.
  • Once that clears the remaining patches should be landable, hope to be able to point 51 Nightly to a11y client vendors for testing sometime next week.
Retrieved from "https://wiki.mozilla.org/index.php?title=Electrolysis/Meetings/2016-08-11&oldid=1143589"