Agenda

• followups from last week
  • Brian to find a piece of string to nail down the DVI adapter in GIGO
• SDK 0.4 postmortem
• SDK 0.5 status
• FlightDeck 1.0a2 status
• platform dependencies brainstorming
• E10S in 0.5
• roundtable

Minutes

• followups
  • finding a piece of string is a MISS, Brian will try again for next week
• SDK 0.4 postmortem
  • landed a lot at the last minute, caused lots of regressions ("crunchy")
  • some of that is unavoidable: need deadlines to encourage work
  • planned to land 10 APIs, actually landed 4. This is up from previous releases. We're getting better at this.
  • we need a real buildbot: Brian will talk to RelEng this week about it, see if they're ready for a handoff.
  • dbuc has a "MailPing" example in development, looking for review prior to a blog post. It will evolve over time, as new APIs become available.
• SDK 0.5
  • two weeks to freeze
  • E10S integration (atul): Atul is still trying to get moz-central+E10Spatch to compile.
    • getting any addon/API working on a separate process is a good first proof-of-concept, to make sure the platform gives us what we need
    • also good to learn how E10S impacts our APIs, so we don't need to break too many APIs. In particular, content-manipulating APIs are the most sensitive to E10S effects.
    • how much confidence can we have around not breaking the APIs we ship in 0.5?
    • two separate goals: E10S proof-of-concept, API subsetting (creating an E10S-compliant subset of each API)
    • API subsetting: Brian can help each API owner to change them to be E10S-compliant. Brian is nominal owner.
    • proof-of-concept is a project goal, not a release goal for 0.5
  • marking chrome-auth packages
    • at-risk, atul says pre-E10S world is feasible, needs separate principals, may not be appropriate for 0.5. This week brian will figure out how much work is involved, decide whether to aim for 0.5 or to punt for later
  • pagemods (nickolay): he's been active, not sure how much progress has been made
  • single UI element (dietrich): boriss will have UI designs by this friday. First cut is a statusbar-like thing.
    • dbuc's extension-autodetection thing (JEP102.1) might get into 0.5, implemented by dietrich, but we won't hold up 0.5 for it
  • Tabs API (dietrich): has a WIP patch, depends on atul's tab-tracker module. Still on-track.
  • docs cleanup (noelle): Myk will get status from noelle
  • Panel API (myk): has WIP patch, mostly complete. Needs tests+docs. Needs E10S-compliancing. Should have ready for review by next week (except for E10S compliance). On-track.
  • Request API (paul): myk will ping paul for status
    • has some tests, needs mock-XHR for tests
  • app compatibility handling (atul): Luca submitted a patch, still r?=atul, sits below E10S stuff on atul's queue.
  • lifecycle (adw): on-track. Has WIP patch. Wants comments on the API (use of 'self' module)
  • L10N (gandalf): will have patch up for review today or tomorrow. Sounds like it's on track.
  • Selection API (erik): new patch posted yesterday, waiting for r?=myk
  • SDK search path (?): no progress. Atul can take it as P2. New definition of deliverable is to merely specify the current behavior, not try to change it at all. on-track.
• flightdeck
  • freeze is next week (tuesday june 8th), RC jun09, final jun16
  • piotr is missing, some status second-hand through dbuc
    • oscar is cranking on UX stuff this week, done with switching module view mode, has read-only bespin and view mode done, working on new module creation now
    • working on XPI creation?
  • improvements to flightdeck addon (atul): will figure out offline w/myk
  • step-by-step tutorial (atul): blocked on new version of frontend
• platform dependency brainstorming
  • dependencies quietly crop up
  • want/need involvement from platform team to avoid/fix
• roundtable
  • security testing? what sorts of security review can be done in parallel with development? suggested API review, mostly on the mozilla-labs-jetpack discussion group

Followups