Labs/Jetpack/Weekly Meeting/2010-10-26
From MozillaWiki
< Labs | Jetpack | Weekly Meeting
Agenda
- followups from last meeting
- Brian to find out the deal with the buildbot project
- myk to ping individuals re: 0.10 plan, then finalize it
- SDK 0.9 review
- FlightDeck 1.0a6 status
- SDK 0.10 status
- roundtable
Minutes
- followups
- brian on buildbot stuff: linux builders are dying early (platform/nightly issue) but at least they're starting. windows is dying even earlier, atul is trying to get access to the buildslave to diagnose
- myk will focus on jetpack-vs-FFtrunk this week, which should help
- 0.10 plan is finalized
- brian on buildbot stuff: linux builders are dying early (platform/nightly issue) but at least they're starting. windows is dying even earlier, atul is trying to get access to the buildslave to diagnose
- SDK 0.9
- thawed late, but froze on time, shipped one day late, hit most goals
- perhaps we could have made the go-for-b6-not-trunk decision earlier, but it might not have helped
- FF4b7 will block on jetpack working: platform team is keenly interested in fixing jetpack problems
- changing freeze time (from midnight up to 11am) didn't seem to strongly affect anybody, but is easier for europe
- Flightdeck 1.0a6
- code standardization: one item left to agree on
- look-and-feel is deferred until next quarter
- multiple SDK versions: piotr has details in the bug
- current code has a hack, won't tolerate too many subdirectories in $package/lib, but this can be fixed later
- docs are on track
- brian will figure out a bug# for splitting API parser code out of SDK, which affects wbamberg's work
- freeze still planned for 11/15
- SDK 0.10
- E10S integration: coming along, got one unit test to pass under --e10s
- plan is to land it, maybe this week, with --e10s turned off, and then get it working later
- CommonJS fixes: probably touches manifest/loader stuff that E10S touches. at risk, brian has presentation to give next week
- switch from add/remove to construct/destroy: drew says it should be easy, but not started yet. Concerned about colliding with other people changing the same code: devs should remain aware that he's in that space, coordinate with him.
- E10S-compat for Windows API: felipe will talk to myk about what needs to happen. probably small. myk will get back to felipe with details.
- E10S-compat for Tabs API: at risk, Myk is full
- "add-on SDK": on track, patch has bitrotted but should be easy to update
- docs: on track, patches in review.
- E10S integration: coming along, got one unit test to pass under --e10s
- roundtable
- dietrich's concern that we're putting energy into API changes now instead of things like fixing E10S. General consensus is that we need to make these API changes now since we'll be supporting the 1.0 APIs indefinitely.
- it seems like we're not seeking out add-on developers and getting feedback on APIs
- add/remove vs construct/destroy is a result of feedback from Resig. Some developers of specific well-known add-ons were pinged for feedback.
- but we could do more. dbuc will look into it.
- git: folks are figuring things out, finding a workflow
- dietrich's concern that we're putting energy into API changes now instead of things like fixing E10S. General consensus is that we need to make these API changes now since we'll be supporting the 1.0 APIs indefinitely.
Followups
- atul to find out if he really needs bug 603403 to be fixed in order to get access to the windows buildslaves to diagnose windows issues
- myk to bump priority of bug 603403 if necessary
- brian to figure out bug# for splitting API parser code out of SDK, which affects wbamberg's work
- myk to figure out what felipe needs to do to make Windows API e10s compatible
- dbuc to look into outreach to developers
