2017-11-27

NECKO WEEKLY MEETING NOTES

Thurs November 15th, 2017


TRIAGE

  • Note: use these 2 queries when you're on triage duty: all bugs should be [necko-triaged] and have a priority

    • UNTRIAGED bugs: https://mzl.la/2wAv7XZ
    • MALFORMED bugs (have necko-triaged but no priority): https://mzl.la/2xXf2jc
  • P1 bugs https://mzl.la/2he1C85

  • P2 bugs https://mzl.la/2yaD1aO

  • P3 bugs (backlog) https://mzl.la/2xcYzEL

  • P5 bugs (would take): https://mzl.la/2hdnQXy

  • 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)

Nick

Gary

Honza

Junior <<<<<===== Nov 24 -> 30

Michal

Amy

Daniel

Kershaw

Valentin

SC

Dragana


Round Table:


Nick

Honza

Michal

  • RCWN
  • Bug 1313934 - Intermittent PROCESS-CRASH
  • Bug 1394564 - Version 55 breaks sending large amounts of data over web sockets.

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

  • TRR (Trusted Recursive Resolver over DOH)
  • bug 1416048 use a 32 bit field for TTL data

Valentin

Threadsafe URI project - https://docs.google.com/document/d/1wzbVcxZ2X_3ab6_VRUgE18KBcnZMekVm1gwKS3zIw0M/edit#

Overview of URI classes and interfaces - https://gist.github.com/valenting/1f1550d756c01c9e732d8fcdfe9240b8

Bug 1416791 - Reduce the mutability of nsHostObjectURI

Bug 1415205 - Add add mutate() to nsIURI

Found bugs in URI implementations: 1416193

Bug 1416343 - Mark nsIURI and interfaces extending it as [builtinclass]

Dragana

103 Early Hint

security bug

NEAT

TFO

Gary

Bug 1415115 - PERMA-FAIL | accessible/tests/mochitest/events/docload/test_docload_root.html | Test timed out.

Junior

Bug 1415120 - Crash in mozilla::detail::RunnableFunction<nsCookieService::RebuildCorruptDB(DBState*)::<lambda()>::<lambda()> >::Run

  • r?

  • Root cause: cookie requests during rebuilding and we close the all the connections. The request is trying to connect a db which is renamed and not loaded.

  • solution: access the cookies from hashtable. drop all the db operations because we will backport them to new db anyway. Bug 1413839 - Add telemetry to probe how close we’re going to block main thread

  • r? Bug 1416985 - MOZ_SQLITE_COOKIES_OLD_SCHEMA seems bogus for unreasonable many people using cookies.txt

  • WIP

Amy

Bug 1357676 - Increase cookie count limits to match Chrome’s

  • Got r- from Josh

  • Have to discuss the new batch eviction algorithm to Josh and Martin. Bug 1413512 - Get telemetry on how often we see HTTP(S) -> FTP redirects

  • Start to implement this bug. Bug 1414206 - doesnt logout properly in firefox57

  • Reporter used the wrong setting.

Kershaw

Bug 1413999 - Add support for Server-Timing HTTP trailing header

*- Almost finish the parser and test case

Bug 1416623 - Skeletal code for socket process

*- WIP uploaded *- Basic setup for new process, such as new process type, init XPCOM, etc. *- TODO:

  • Try to create UDPSocket in socket process
  • sandboxing

SC

[Socket Process]

  • study "Endpoint" mechanism for IPC between child process (reference: bug wwwwww)

  • Prototyping

  • design interface and flow for UDPSocket

  • https://docs.google.com/presentation/d/12Jg2eXT7nne2G8MtS3E2p7yRGaQJrvXg8iiQOhYggFU/edit#slide=id.g286761667d_0_0

  • create project wiki page https://wiki.mozilla.org/Necko:_Sandboxing_TCP/UDP_socket_in_a_separate_process

  • discuss sandboxing design with @pauljt

  • put protocol code in separate process is better

  • try not have plain TCP/UDP IPC interface, to have better control on the network usage [Others]

  • Bug 1390708 - .hqx files are all corrupted when downloaded with FF unless going through SAVE AS...

  • remove BinHex stream converter support, landed on m-c (firefox 59)

  • Bug 1416879 - Crash in mozilla::net::HttpChannelParentListener::OnDataAvailable

  • parent ODA called during channel suspension for diversion, might be timing issue caused by intercept stream listener