Agenda

• followups from last meeting
  • Brian to talk with Will and Piotr about the new docs system
  • Brian to discuss the key pair id in the SDK package.json file
  • Daniel to look into creating bugs from pull requests
• FlightDeck 1.0a7 status
• SDK 1.0b2 status
• roundtable
  • cancel meeting next week?

Minutes

• followups
  • Brian talked to Will about patch to change SDK's behavior when you run cfx docs; Piotr and FlightDeck folks not focusing on docs at the moment; plan of record is to put HTML files into generated sdocs tarball; code to import tarball into FlightDeck can use HTML docs for display while continuing to use JSON for code completion
  • Brian talked to Piotr about adding command to SDK that moves addon from local machine to FlightDeck, uploading both code and keypair;

• 1.0a7 status
  • package sidebar redesign: progress made
  • ability to edit attached files: almost finished
  • ability to upload locally built packages: ZIP is done, XPI not started, unclear whether it'll make the release
  • implement more generic docs system: not started, probably won't make release
  • mechanism for displaying SDK error output: Daniel's thing, and he's not here, but probably won't happen this release
  • papercut bugs: some done, some on track

• SDK 1.0b2 status
  • no plan yet
  • e10s is top priority, but other things can happen too
  • Will: is documenting e10s high priority? Myk: probably not, as it won't affect addon developers in this release (unless it goes much faster than expected and we can flip the switch in this cycle, which seems unlikely)
  • proposed timeline in discussion group with release at end of January
  • expect plan by next week

• roundtable items
  • should we meet next week? Myk proposes no, as some people won't be around, and others will have taken time off such that there isn't enough work time in between meetings to justify getting status; no dissent; resolved, we won't meet next week
  • Piotr: when there is a review of a XPI, the XPI will be secure because of the security of the SDK-provided modules it wraps; but what if an author changes the implementation of a module? Brian: the envisioned review tool will check the integrity of bundled modules and warn reviewers when their content has changed.

Followups

• Daniel to look into creating bugs from pull requests
• Brian to confirm FlightDeck XPI generation disrespects Python plugins