Platform/2014-Q4-Goals

GFX

Items marked here with release 35 and 36 are part of the Q4 landings.

The current plans is to concentrate on delivering functionality in releases 37 and later, with releases 35 and 36 being used to stage the partial implementations. For example, 2015 Q1 plans call for:

• Firefox OS 2.0/2.1 device support (as not handled by the device team)
• Async pan zoom improvements for Firefox OS (driven by UX or performance requirements, including support for Project Silk)
• WebGL1 performance. Move towards parity with Chrome (targeting Windows as the priority)
• WebGL2 demo MWC/GDC.

DOM

• [ON TRACK] WebComponents: one of :host (bug 992245) or :content (TBD which is more important) selector landed (preffed off) on m-c
• [ON TRACK] respond to Service Worker feedback, adapt to spec changes, and improve performance
• [ON TRACK] have FileSystem API (non-DeviceStorage backend) implementation up for review
• [ON TRACK] bug 701634 land IndexedDB in workers on m-c
• [ON TRACK] bug 942542 land new quota API on PBackground on m-c
• [ON TRACK] bug 874016 ensure all plugin work related to e10s is complete
• Produce a draft for Sharing (annevk)
• Define persistent notifications (notifications associated with a service worker; annevk)
• [ON TRACK] bug 997779 Implement more w3c manifest (marcosc)

Accessibility

• e10s goal - TBD
• [ON TRACK] bug 1069598 Gesture quicknav (ffos a11y)
• [ON TRACK] bug 1069600 Gesture text nav (ffos a11y)
• [ON TRACK] bug 1069574 Screen reader hints (ffos a11y)
• [ON TRACK] GAIA: Fix Productivity P1 a11y bugs (~47 at this time)
• etc/TBD

A*Team

(draft only)

Highlights:

• Develop a multi-binary harness for devtools tests
• Add e10s support to Talos, and either Marionette or Mozmill
• Make end-to-end testing in Autoland possible
• Deploy ReviewBoard for developers to start using (carry over from Q3)
• Add B2G support to the games benchmarking harness (mozbench)
• Add B2G support to mochitest-chrome
• Add support to Bugzilla for the Gmail transition
• Prototype alternate bug views in Bugzilla to improve performance and usability
• Continue improving sheriff and developer workflows in Treeherder

See full list: Auto-tools/Goals/2014Q4

Web Engineering

(draft)

crash-stats

• [ON TRACK] begin migrating subsystems to AWS
  • e.g. migrate hbase to S3, architect cloud infra, build production-quality AMIs
• [ON TRACK] enhance processor flexibility
  • e.g. deploy processor2015, enhance tooling for developing post-crash classifiers, prototype alternate processing pipelines
• [ON TRACK] enhance search functionality
  • e.g. pull users from postgres search so we can shut it off, upgrade ES to 1.x, data dumps of search results

elmo

• [ON TRACK] continue progress towards buildbot retirement
  • e.g. migrating diff/compare tools to a10n, migrate remaining buildbot steps to a10n

airmo

• [ON TRACK] prototype self-service publishing
  • e.g. getUserMedia-based new hire video recording, user uploads of pictures and videos

DXR

• [ON TRACK] Continue improving infrastructure.
  • For example: move file rendering to request time, implement independent tree rendering, or finally replace the config subsystem
• [ON TRACK] Pick newly low-hanging fruit.
  • With the new plugin architecture and Elasticsearch backend, a lot of previously difficult possiblities are opened to us. Examples: index a new language, show result counts, show indexed revision identifiers, do some result mixing, or default the query parser to interpret text runs as phrase matches.

SUMO and Input

(draft only)

• Implement v1.0 of BuddyUp, the live support app for FirefoxOS, and integrate it with existing SUMO web app.
• Update SUMO forums.
• (Input goals still under discussion)

Release Engineering / Release Operations / Developer Services

(draft only)

• Identify, measure, and publish core release engineering KPIs.
• Develop a plan for disaster recovery and business continuation in the event of a disaster for release infrastructure.
• Make try stable and fast.
• Support Windows 64 initiatives as needed.
• Automate merge day processes to free up engineer time and reduce opportunity for error.
• Develop automation and self-serve APIs for developers to remove the need to have one full time engineer on buildduty.
• Automate release processes, again to free engineer time and reduce opportunity for error.