2017-08-17
NECKO WEEKLY MEETING NOTES
Thurs August 17th, 2017
STICKY
- Triage duty: nwgh
- new untriaged bugs: https://tinyurl.com/zhk9yjf
- NECKO-ACTIVE bugs: https://tinyurl.com/j8pj35u
- NECKO NEXT bugs: https://tinyurl.com/zjwjfee
- NECKO-BACKLOG bugs: https://tinyurl.com/hpkcxx3
- Triage Notes
- Android crashes on infra are currently not symbolicated: https://bugzilla.mozilla.org/show_bug.cgi?id=1389805 (makes triaging/diagnosing android crash bugs hard/impossible)
Round Table:
---PROJECT UPDATES--- - rust-url meeting notes: https://public.etherpad-mozilla.org/p/mozloha-rust-url - rcwn https://public.etherpad-mozilla.org/p/mozaloha-rcwn
Nick
Honza
Michal
- 1369051 Assertion failure: !entry || !entry->IsFresh()
- 1388684 RCWN breaks HTTP_CHANNEL_DISPOSITION telemetry
- 1390683 HTTP_*_COMPLETE_LOAD probes are used in nsLoadGroup::TelemetryReportChannel when using new cache
- 1382688 Do not allow to enable old disk cache backend
Patrick
last week: mpendingq, 256fd telemetry, 0.9 telemetry, namespaces, neat tls, alt-svc, immutable bug, some non-necko stuff this week: neat tls, alt-svc, immutable bug, some non-necko stuff. more reviews.
Daniel
- NEAT project work
- IPv6 network identification
Valentin
Bug 1381816 - Pref study: Race Cache with Network iteration #2
Bug 1385883 - Cannot delete history with IDN
Bug 1386195 - Cannot delete history with "www." + IDN
Bug 1383831 - require newer version of rust-url
RCWN - telemetry should be available in the next few days, so we can see if it made any improvement
Looked into 1367810 - Crash in mozilla::net::nsHttpChannel::CallOnStartRequest
Dragana
TFO - working, nspr patch
Better TFO telemetry
rel=preload and no-cache
JAR debugging
Gary
CDP Testing
- Background download
- Cynthia works on Sikuli for user interaction tests Bug 1373708 - Convert Jar channels to stop using main thread I/O
Junior
Bug 870460 - Lazy load of cookie service blocks main thread while cookie database loads
- first round feedback, init db in profile-do-change, seems work Bug 1388671 - Crash in nsCookieService::AsyncReadComplete
Amy
Bug 1331680 - Consider not doing sync IPC for document.cookie getter/setter
-
Landed
-
Testing report
-
https://drive.google.com/a/mozilla.com/file/d/0B1DNWGRO_VTuVnUtd1BVeThhV00/view?usp=sharing Bug 1366578 - cookies are not transmitted when handling custom protocol
-
Content type of Load Info is different between e10s and non-e10s in the testing environment of user.
-
User provided testing webpage to me, but I can't reproduce.
-
ni user. Bug 1376009 - Crash in mozalloc_abort | NS_DebugBreak | mozilla::ipc::LogicError | mozilla::net::PCookieServiceChild::SendGetCookieString
Bug 1374939 - Crash in mozalloc_abort | NS_DebugBreak | mozilla::ipc::LogicError | mozilla::net::PCookieService::Transition
-
r+ Measured time of set cookie string.
-
CNN - loaded to homepage.
-
https://perfht.ml/2viFIsB
-
Facebook - logged in to my account.
-
https://perfht.ml/2viz7OQ
-
The longest time is 1ms.
Kershaw
CDP Test
- Create a web site to present test results
- http://10.247.28.241:8888/
Bug 1385576 - Tracking Protection broken in Nightly.
- investigating
Bug 1388247 - Label PCookieServiceChild
- working
SC
[PBg-Http]
-
performance analysis
- quantum_pageload is now measuring first paint (https://treeherder.mozilla.org/perf.html#/compare?originalProject=try&originalRevision=3491ddfe3180171a78febefd3b29551c5cd06604&newProject=try&newRevision=36c892dbea102bbda865cf665c99212d5caec9ba&framework=1&showOnlyImportant=0)
- amazon/youtube shows 10% improvements (40-70ms)
- google shows no noticeable difference
- facebook shows 10% degradation (40ms)
- quantum_pageload is now measuring first paint (https://treeherder.mozilla.org/perf.html#/compare?originalProject=try&originalRevision=3491ddfe3180171a78febefd3b29551c5cd06604&newProject=try&newRevision=36c892dbea102bbda865cf665c99212d5caec9ba&framework=1&showOnlyImportant=0)
-
fix regressions
- Bug 1387285
- Bug 1353829 - 8% Are we slim yet regression found on autoland march 28th from revision e2a697abd5d3
- before/after OMT imagelib holds different set of decoded image, which causes the memory difference
- the regression is reduced to 1.71% [CDP]
-
Bug 1382567 - boost load priority of preloaded image for without height/width attribute
-
under review
-
necko triage overhead