Electrolysis/Meetings/2015-05-07
From MozillaWiki
< Electrolysis | Meetings
Contents
mconley
- M6 - bug 1146955 - Make GMP plugin crash reporting UI work in e10s
- Fixed and landed
- M6 - bug 1025146 - [e10s] Make it so that we can view document source without hitting the network
- Blocked on https://bugzilla.mozilla.org/show_bug.cgi?id=1156493, which has r+ but some review nits. michal said he'd look at this today, hopefully land soon.
- M6 - bug 1128050 - [e10s] Save page as... doesn't always load from cache
- Reviewed, blocked on the same cache key bug as bug 1025146.
- M7 - bug 1124728 - All talos on OSX fails with running in e10s mode
- Found the culprit, have a solution, but it's throwing off our sessionstore talos numbers. Looking into what we can do about that.
- M7 - bug 1091112 - Print dialog doesn't get focus automatically, if e10s is enabled
- Started - working on this when blocked on other things. Feeling pretty confident that this will be done soon.
- M7 - bug 1116188 - [e10s] Stop using sync messages for Gecko profiler
- Blocked by https://bugzilla.mozilla.org/show_bug.cgi?id=1154115 which is going through final review stages.
- M7 - bug 1136478 - [e10s] pagehide/show events not fired when swapping remote browsers
- Speculative patch written - waiting on feedback from smaug while I write a regression test.
- Wrote a blog post about the big tab spinner: http://mikeconley.ca/blog/2015/05/04/electrolysis-and-the-big-tab-spinner-of-doom/
- Telemetry is coming in, which is nice. Still a bit early, but median tab switch times support the claim that OS X is slightly slower than Linux and Windows. Too early to draw any strong conclusions.
- Enn and I have interns starting on May 19th. What should they work on?
- Page Info dialog
- Flesh out about:tabcrashed with input fields
- Make lightweight theme web installer work
- Clearing out a bunch of unsafe CPOW usage
- Move about: pages to remote
jimm
- bug 1129040 (addon/plugin blocklist + tests) - landed
- bug 1156592 (tab switch telemetry) - landed
- bug 1159327 (enable a11y more broadly) - landed, we should reset the disable pref next time we advertise.
- bug 1153205 (fix for missing annotations in automation crash dumps) - latest wip posted for review.
- bug 1160142 (take browser stacks sooner for plugin hangs) - wip
- bug 1156857 (histogram child data missing) - up for review
gabor
- bug 1148188 (call interposition / addon API) landed
- reviews
felipe
- bug 1156613 - Fix #ifdef NIGHTLY_BUILD blocks - landed and tested on aurora as well
- bug 1161260 - Prompt for e10s on Aurora - almost ready to go
- bug 1129957 - RemoteWebNavigation doesn't accept post headers - finishing test
- Reviews
Mossop
- Tentative patch for bug 1160279 dragging a file into content doesn't load the file
- Trying to get the SDK code uplifted to bring in more e10s fixes
Larissa
- keeping tabs on pre-aurora uplift needs
- starting more detailed tracking of add-on dev outreach in the arewee10s spreadsheet (and thank you so much Tracy for adding the "popularity" sort)
mrbkap
- Wrote and landed patch for m6 bug 1155494 (performance problem with fullscreen).
- Working on patch for last m6 bug 1024437 (<datalist> doesn't work in e10s).
- This led to a few followup bugs to get satchel tests working on e10s.
gw280
- Been looking at a non-hacky solution for bug 1049285 (popups positioned incorrectly if openPopupAtScreen() changes the popup coordinates to stop it from appearing offscreen). I tried some dead ends but I want to talk to Neil to see what he thinks before trying any more.
- Investigated bug 1095781 ("graphics problems with e10s"), or rather, "corruption seen when scrolling pages on windows". I think this can be kicked out of triage because I think they should now be getting WARP instead of D3D9 and the bug only seems to pertain to D3D9 layers (it was kept open because Basic OMTC. In any case this is a gfx bug not e10s.
- Bug 1157941 and bug 1140414 are dupes of each other. Basically, there's an issue with paint suppression and differences between e10s and non-e10s. tn's comment https://bugzilla.mozilla.org/show_bug.cgi?id=1157941#c6 is a pretty good descriptor of the issue.
handyman
- bug 1125325 - Web content is not rescaled when OS X moves browser window from HiDPI laptop display to external display
- Related to bug 1096550, requested by roc/enndeakin. Suspected this patch of breaking a test. Figured out the cause, working on fix.
- bug 1096550 - Dragging tab from one window to another on different displays zooms in
- Took over for enndeakin since other patch solves most of issue
- Will land with bug 1125325 (patch for 1096550 is already done)
- bug 1083365 - [e10s] Tooltips don't always go away
- Landed
- bug 1158315 - Window does not respond to keyboard shortcuts when a plugin has focus
- Investigated and dropped as non-e10s. Handed off to smichaud, who says this issue has been known for years and wont be fixed any time soon. Notably, this is not born of the big Cocoa plugin refactor a few months ago.
- bug 863514 - Electrolysis: Make gesture support work
- Sloppy WIP POC patch posted.
- Cleaning up the patch. This is actually a good deal of work.
- One somewhat big issue — APZ code in the ESM that claims all mouse events over a scrollable region. A simple change fixes this but introduces a bug (swiping past the end of a horiz scroll can trigger a history event — this should only happen if the scrollbar was at end of scroll when gesture was begun). Doing this right will take some fairly heavy ESM engineering. Most recently resolving this issue with Felipe.
Giorgio
- Mailed the "Download Flash And Video" folks again, still no answer, gave up for now
- Worked on outstanding E10s issues in NoScript & FlashGot, deciding to use a progressive approach (E10s-aware modules for new code based on message manager, leaving SDK porting for better times).
Blassey
- bug 1161166 Improve performance of shared memory on OSX - ran into a roadblock, shmget() is failing to allocate memory, which seems broken. I'm going to try fallocate to see if that performs better. If not, I'm stumped.