Apps/StatusMeetings/Engineering/2012-08-08: Difference between revisions

Jump to navigation Jump to search

Apps/StatusMeetings/Engineering/2012-08-08 (view source)

Revision as of 17:15, 8 August 2012

5,366 bytes added ,  8 August 2012
no edit summary
No edit summary
 
Line 1: Line 1:
= Details =
= Details =


* Time: Tuesday, 2012 August 8, 9:00 - 9:45am PT (16:00 - 16:45 UTC)
* Time: Wednesday, 2012 August 8, 9:05 - 9:55am PT (16:05 - 16:55 UTC)
* Backchannel:
* Backchannel:
** irc.mozilla.org #openwebapps
** irc.mozilla.org #openwebapps
Line 26: Line 26:


= Minutes =
= Minutes =
== Significant Updates ==
=== google.com mobile sites ===
perhaps as a result of evangelism efforts, google.com sites are now serving fennec touch-based mobile sites!
=== firefox os ua string change ===
* firefox os ua string change makes most sites serve it their desktop versions
* now we're bikeshedding some more on the string
* current string is what we'd like it to be long-term, but hard to justify the problems it causes
* further changes are likely
==  Checkin/Status On Major Challenges ==
=== payments integration ===
* pmpcat talking with partners about integration of payments
* trying to figure out when they can do end-to-end test
=== state of packaged apps ===
* clouserw was in recent meeting on the subject
* lasted an hour and a half, only got through three questions
* fabrice put together example of packaged app hosted on his people site
* jsmith has figured out how to test the api
* only works on firefox os
* doesn't work on android/desktop, although it gets a little farther on android before it fails
* next meeting about packaged apps tomorrow (Thursday) 9am pt
* very focused small team meeting
* https://etherpad.mozilla.org/packaged-apps
=== startup problems w/apps on android ===
* talked about it in triage
* there's a bug about looking into that
* working on metrics to measure it
* figure out where the slowness is on different devices
* tough problem to solve
* gecko is slow to start up on many devices
* most apps don't use appcache
* appcache doesn't necessarily fix the whole problem
* we're going to look into perception solutions
* would packaged apps help?
* leaving chrome files packaged helped for firefox
* worth looking into
=== api changes for multiple apps per origin ===
* freeze date for gecko 17 is end of august
* sicking would really like to do it before then
* because that's the version we're going to ship in firefox os
* at risk
* https://bugzilla.mozilla.org/show_bug.cgi?id=779982
* sicking to argue for blocking-basecamp, but thinks that doesn't necessarily mean anything
* -> bwalker to talk to pmpcat about tackling coordination of compatibility fixes for multiple apps per origin
== Questions And Concerns ==
=== implementing nsIPrincipal::appStatus ===
* who is implementing nsIPrincipal::appStatus (i.e. identify  apps as "web", "privileged", or "certified"), and what's the status of  that work?
* fabrice working on it, doesn't yet have a good sense of what makes an app certified
* depending on bsmith to implement signing pieces
* last ditch solution, if signing doesn't land, would be to trust manifest as long as manifest is coming from app store
=== packaged apps critical path ===
* jsmith doesn't understand what needs to be done on packaged apps for v1
* sent email about this
* seems in danger of scope creep
* what needs to be done by the various functions to make this happen?
* motivation is to do privileged/certified apps, for which our only current solution is packaged apps
* we need packaged apps for privileged apps
* we need privileged apps for access to various apis (camera, device storage)
* we also need a way to update apps that are bundled with phone
* we're going to have two different storage areas, one on system partition and one for installed apps
* fabrice has a patch for that
* b2g-specific
* does not rely on apps coming from mozilla marketplace
* so doesn't depend on marketplace to update built-in apps
* built-in apps will be updated together with platform
* you won't be able to remove built-in apps
=== marketplace design changes ===
* marketplace designs not final
* but they are mobile-only
* team was waiting for desktop designs
* but is now planning to ship mobile-only designs this week or next
* marketplace will change drastically, might stop working on desktop
=== navigator.id.watch vs. oauth for authentication ===
* we're going to use oath
* reviewed payments and refunds workflow this morning
* tech leads are going to discuss in existing thread
* timeline remains end of august
* kumar and fabrice were part of discussion (along with qa folks tony, krupa, and jason)
* initial requirement was to integrate persona for v1
* but then changed our mind and decided not to integrate persona into firefox os for v1
* instead you'll use your telefonica/bluevia id for various transactions
* oauth is a consequence of decision
* we're also now considering letting users install free apps anonymously, without signing in
* rforbes (security) and clouserw (marketplace engineering) impacted by decision
=== web compatibility and use of navigator.pay ===
* why is this a dom api instead of a hosted service?
* we need something in the dom
* we can't completely implement it in javascript
* because we want user agent interaction
* we could do a two-layered approach with an api in a js library
* and an obscure dom api that the js library uses
* at some point we need to expose something in a web page for people to initiate payment
* there has to be a way for the page to communicate with the user agent
* provided we want the user agent to mediate the payment process, which we do!
* intent is to make it possible for trusted chrome ui to give user assurance that entering confidential payment details is safe


= Actions =
= Actions =
* bwalker to talk to pmpcat about tackling coordination of compatibility fixes for multiple apps per origin
MykMelez
canmove, Confirmed users
2,056

edits

Retrieved from "https://wiki.mozilla.org/Special:MobileDiff/459227"

Navigation menu