From MozillaWiki
Jump to: navigation, search

« previous week | index | next week »


  • Continuing to work on bug 1262671. Should have complete set of patches to post by end of week.


  • Investigating checksumming IPC messages to detect message corruption
  • bug 1258317 REOPENED:: crash in mozilla::ipc::TransferHandleToProcess (PBackground related)
    • Updated bug 1258317 with new findings from the annotations of new crashes
  • bug 1259480 NEW:: [e10s] FatalError crash in PContentChild::Read from PContentChild::SendReadPrefsArray
    • Investigating dump and tracing code


  • bug 1253956 NEW:: [e10s] Windows debug browser/base/content/test/general application crashed [@ mozilla::LinkedList<js::WeakMapBase>::~LinkedList<js::WeakMapBase>()] (Assertion failure: isEmpty(), at LinkedList.h:332) on shutdown
    • Running crash tests on windows 8


  • bug 1245068 NEW:: [e10s] Ctrl+mouse wheel to change zoom level doesn't work if web page is not focused
    • Found the line that causes the issue, not sure how to fix it yet
  • bug 1190297 NEW:: Urlbar is cleared after invalid url was entered when electrolysis is enabled.
  • bug 1220517 NEW:: Intermittent TEST-UNEXPECTED-FAIL | leakcheck | default process: 84 bytes leaked (Mutex, nsBaseAppShell)
    • Responded to the mail from Andrew Overholt for bug 1220517
  • bug 1257869 NEW:: Window Resize of large page takes seconds to change viewport
    • Checked live resize events for windows and gtk, also ni'd :jimm for whether it is still a P1?


  • bug 1264642 NEW:: Reduce the size of PBrowser::Msg_AsyncMessage IPC messages
    • Discussed StructuredClone with baku
    • Should have working patch by the end of the week.


  • Performed a variety of analysis on 47 and 48 for UI Smooooooooothness release criteria
    • TL;DR: 48 is signed off, with the proviso that I prioritize the last tpaint regression fix in my work queue and fix it ASAP for 49.
    • APZ is responsible for the CART and TART regressions. APZ team says this is the cost of APZ, and difference is not too bad. Trade-off is obvious.
  • bug 1261842 - Make initial browser remote sooner if we're defaulting to using remote tabs
    • Blocked on bug 1267720 - Re-factor nsWindowWatcher::OpenWindowInternal to not be such a rat's nest
      • Hoping to have patches up early next week.
  • M9 bug 1263887 - [e10s] The dropdown selector is not working properly.
    • Landed and bounced a bunch of times, and then finally stuck.
    • Going to wait a few more days and then request uplift.
  • bug 1166351 - XBL parent bindings not properly attached - tabs suddenly do not switch or close properly
  • The usual reviews. Been on a few interview panels.


  • Nothing much. Ran some analysis for figuring out the GA release proportion on bug 1266800. Turns out only ~40% of release users have disqualifying addons (compared to ~60% on beta).


  • bug 1247420 - (m9 1255138 - Resizing windows from JS is not supported in e10s mode) - Being reviewed by Mike.
  • bug 1262661 - (Many large SessionStore:update messages) - Fixed a bug but it got back from inbound because of a test, I'm investigating it. I don't see any other easy fixes. One approach is to try to import some diff-patch mechanism for JS objects. The other is to try to hook in a compressor in message managers sendAsyncMessage function. I have some plans about how it could be done but Bill is not convinced about this approach, I'm waiting for his response.
  • some more reviews...


  • Handled work needed for the experiment before uplift to 47
    • bug 1266213 block gtk 3.20 users on beta
    • bug 1264437 separate disqualified into disqualified-test and disqualified-control cohorts
  • Created the E10s Experiments wiki page
  • Reviews
  • Updated test spreadsheets
  • Helping with the calculations for target audience on release
  • Started looking at the Marionette tests that are failing only on Beta


  • bug 1266484 - crash in libsystem_kernel.dylib@0x16db6 when using Pinboard bookmarklet -- Landed
  • bug 1248683 [e10s] Find bar sometimes gets keypresses in the wrong order when system is under heavy load -- posted a hacky proof-of-concept incomplete patch with questions for masayuki


  • bug 1114647 - content/chrome process consolidation. many, many issues making sandboxing work with just the one process. been working closely with bob owen and jed davis to try and resolve the sandboxing issues. I think I've done all I can reasonably do at this point, and it would probably be best for the sandboxing guys to take over from here.


  • bug 1035125 - On Windows, plugin-container.exe is linked against the sandbox_s library twice - turned into a fight between crt allocation and jemalloc, had to get rid of all the crt static linking to keeping it working when compiled with VS2013 - nearly there.


  • Review data with chutten and felipe for bug 1266800, confirm with Saptarshi (standing in for Jeff G), closed bug. (?)
  • Prioritizing work for after this week.
  • Working on grokking all things Telemetry and using re:dash.
  • Talking to Lonnen about the best way to get some crash-stats data usable via re:dash (per BDS request from bug 1265015).



  • bug 1268151 Circular includes in IPDL headers leading to odd behavior. a11y protocol has to include PFileDescriptorSet even though it doesn't actually reference it!
  • bug 1268544 Refactoring a11y to make room for remoting COM objects
  • bug 1267368 Reviewing the COM apartment used by main content thread.
  • Discussion with tbsaunde about deadlock scenarios, will probably need to do some work to mitigate this


  • bug 1263774 - Send memory report with content process crash reports that were low on memory. I have a working patch that I will be seeking review for today if my Try run passes.
  • Plan to work on bug 1260209 next.



  • OOM work is waiting on new Beta, various session store improvements by billm, kanru and gabor.
  • bug 1265680 - Investigating IPC deserialization crash.
  • bug 1266578 - Missing OOM crash annotation in IPC::Channel::ChannelImpl::ProcessIncomingMessages.
  • bug 1227347 - Remove whitelisting of CompositorChild leaks for XPCOM and LSan leak checkers.
  • bug 1264820 - Fixed up and landed Bill's patch to measure IPC reply size in telemetry.