Gaia/SMS/Scrum/FxOS-S6

From MozillaWiki
< Gaia‎ | SMS‎ | Scrum
Jump to: navigation, search

List of bugs

SMS issues handled by the SMS subteam (blocks the sprint bug 1198259)

Bugzilla link

ID Assigned to Summary Blocking b2g Feature-b2g Whiteboard Resolution
818000 Julien Wajsberg [:julienw] system messages are ignored if the handler pathname is not the same as the foreground application pathname --- [sms-sprint FxOS-S7 p=3][sms-sprint FxOS-S6 p=2][sms-sprint FxOS-S5 p=3]
1180592 Steve Chung [:steveck] [Messages][NG] mozMobileConnections shim Implementation --- [sms-sprint FxOS-S7 p=1][sms-sprint FxOS-S6 p=2] FIXED
1198266 [Messages] Use ConversationService in the application --- [sms-sprint FxOS-S8 p=1][sms-sprint FxOS-S7 p=3][sms-sprint FxOS-S6 p=2] WONTFIX

3 Total; 1 Open (33.33%); 2 Resolved (66.67%); 0 Verified (0%);


Remaining points and burndown chart

google chart api url for Sprint FxOS-S6

Burndown chart
Remaining points
Start 6
Day 2 6
Day 3 6
Day 4 6
Day 5 6
Day 6 6
Day 7 6
Day 8 6
Day 9 6
Day 10
End


SMS issues handled by the SMS subteam outside of the sprint (contains whiteboard "sms-sprint-FxOS-S6")

ID Assigned to Summary Blocking b2g Feature-b2g Whiteboard Resolution
1199593 Julien Wajsberg [:julienw] [Message]The text content lost when forward the MMS. 2.5+ [2.5-aries-test-run-2][sms-sprint-FxOS-S6] FIXED
1199612 Oleg Zasypkin [:azasypkin][⏰UTC+1] [Messages][NG] API shim host didn't initialize correctly and break the MMS retrieval 2.5+ [sms-sprint-FxOS-S6] FIXED
1200116 Oleg Zasypkin [:azasypkin][⏰UTC+1] [Messages][Tests] ThreadGenerator should be able to create threads with MMS --- [sms-sprint-FxOS-S6] FIXED

3 Total; 0 Open (0%); 2 Resolved (66.67%); 1 Verified (33.33%);


All SMS issues tracked for this sprint (target milestone)

Bugzilla link

ID Assigned to Summary Blocking b2g Feature b2g Resolution
1169576 Oleg Zasypkin [:azasypkin][⏰UTC+1] [Messages][NG] Implement Conversation service: method for streaming joined threads and drafts list --- --- FIXED
1197104 Steve Chung [:steveck] [Messages][Drafts] Threadless drafts are broken 2.5+ --- FIXED
1198259 SMS sprint FxOS-S6 --- --- INCOMPLETE

3 Total; 0 Open (0%); 2 Resolved (66.67%); 1 Verified (33.33%);


Sprint planning

Minutes are on a separate page.

Daily meetings

Day 2: 26th August

Day 3: 27th August

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
    • WIP is ready but not fully tested yet. Will request feedback if everything works fine.
  • bug 1198658 - [Flame KK][Message]Press the download button, there is no message indicating that you need to switch data connection
    • Found a dataclone problem with DOMError instance. Patch created for review.
      (Oleg) Wilson told me some time ago, that he should deal with this errors in the bridge (like extract error message or name from Error instance), can we confirm with him why it doesn't work? Or it's not in reject callback?
      (Steve) For the normal Error instance I think he did extract the message of error instance, and in the reject we will only get message instead of error instance. But seems like DOMError is not considered. Do you think bridge should be able to dealing with the DOMError as well?
      (Julien) I think it should !
      (Oleg) Yeah, I think we'll have ton of cases where such error can occur.
      (Steve) So, redirect the bug to Wilson, or create another issue for him?
      (Julien) I would create another issue and still fix our bug with a comment to this issue. Like "We need to clone the DOMError manually until bug XXX is resolved". Unless he can look at it soon, or you can do a PR to the threads.js lib :)
      (Steve) Got it.

Today:

  • Create a bug to disable the message event broadcast
  • Layout for mozConnections shim
  • Confirm with Etienne about the profiling and make the list layout close to master.

Julien

  • worked more on system messages issue: I have a working solution with (I think) no bug for the current apps; I'm doing a test app with split views + event handling that could be even closer to what we'll need in SMS (esp opening a specific conversation when the user taps the notification) to see if this works with this solution. Will continue today.
  • spent some time on the "low storage condition" spec; I'll file some bugs later this week, there is small work for us.
  • some bug triaging

Today: I want to:

  • do left-over reviews (I left them for some days ;) )
  • work on the system messages issue

If all this moves forward well, I could:

  • continue the prototype caching the thread list to a single db (including contacts/drafts/etc).

Oleg

  • bug 1155534 - [Messages][NG] Extract NewMessage view from Conversation view
    • No updates (in background).
  • bug 1176976 - [Messages][Drafts] Remove the draft saving/replacing action menu
    • Rebased on latest master to re-run integration tests (in review).
  • bug 1169576 - [Messages][NG] Implement Conversation service: method for streaming joined threads and drafts list
    • Finally added "checkin-needed" (checkin-needed).
      (Julien) I think you could add it earlier already :) Sheriffs don't land when the tree is closed ;)
      (Oleg) Ahhhhh, good to know :)
  • bug 1198266 - [Messages] Use ConversationService in the application
    • Analyzing data I've got with WebIDE - will dump first portion to the bug soon (in progress).
      (Julien) you can share profiles from cleopatra too. You can upload a profile to https://people.mozilla.org/~bgirard/cleopatra/ and share from here.
      (Oleg) I thought cleopatra kind of replaced with built-in profiler or not yet? I admit I don't remember what type of data cleopatra provides :)
    • So far l10n.js and gaia-header are the most time consuming things on the startup.
      (Julien) more than in normal master ? l10n I'm not surprised but gaia-header should not :/
      (Oleg) I think the same as in master, just my statement of disappointment :) gaia-header doesn't look like the main thing in the app, but it consumes a lot of time :)
      (Julien) I'll look at it, maybe it regressed. Or maybe it's only JS and not Reflow/Repaint and in that case why not.
      (Oleg) Basically both, initially it spends a lot of time for createShadowRoot (~90ms) and later on it causes huge reflow (~160ms, if I understood it correctly) with sumButtonWidth - I'll comment on the bug.
      (Julien) OK, this should not happen at all :/ in Inbox, I did work (with title-start/title-end) to avoid reflows at startup, and other panels have "skip-font-fit". So I'll look and file a bug if needed.
      (Julien) also you should capture Gecko as well (there is an option, I think)

Other:

  • Started to look at "bug 1160049 - [Messages] Attach menu should not dismiss when user cancel the replace attachment request", left on question at github;
  • Looked through Bevis's proposal, need some time to digest it first.

Today:

  • Will handle review/feedback/need-info requests;
  • Will work on review comments and assigned bugs.

Day 4: 28th August

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
    • Rebase to master (Still in test)
  • bug 1198658 - [Flame KK][Message]Press the download button, there is no message indicating that you need to switch data connection
  • Mail to Etienne about how to utilize raptor profiling with fast-list

Today:

  • Create a bug to disable the message event broadcast
  • Layout for mozConnections shim

Julien

Was a slow day because of some meetings.

  • Still happy we discussed and estimated the NGA plan.
  • reviewed the 2 simple patches from Gandalf, regressions from the l10n changes he did: bug 1171206 and bug 1192744
  • no work on system messages issue

Today: I want to:

  • do left-over reviews
  • work on the system messages issue

If all this moves forward well, I could:

  • continue the prototype caching the thread list to a single db (including contacts/drafts/etc).

Oleg

  • bug 1155534 - [Messages][NG] Extract NewMessage view from Conversation view
    • No updates (in background).
  • bug 1176976 - [Messages][Drafts] Remove the draft saving/replacing action menu
    • No progress (in review).
  • bug 1169576 - [Messages][NG] Implement Conversation service: method for streaming joined threads and drafts list
    • Landed (landed).
  • bug 1198266 - [Messages] Use ConversationService in the application
    • Added my analysis to the bug and ni? Zibi and Julien for the pieces they know better.
    • Zibi clarified that some bottlenecks in l10n are resolved in l20n - good :) Julien, do you know if plan to have l20n in Messages in 2.5?
    • Noticed twice that "Intl.DateTimeFormat" initialization could take ~80ms, but don't see it with the latest PVT build anymore.
      (Julien) yeah we know this is long, that's why we cache it (don't remember the exact place). If it doesn't happen with current build, it's good ! I'd like to see which bug made this better :)
      (Oleg) Ah, okay, I'll double check and figure out what fixed that.
    • Some preliminary checks: removing gradient photo placeholders didn't improve thread node paint time, didn't notice big difference between "setTimeout(0)" and setImmediate shim via window.postMessage though I expected it to be much faster - will look deeper.
      (Julien) why did you expect to be faster ? postMessage is using the event loop, just like setTimeout.
      (OIeg) I've seen several _synthetic_ jsperf tests, were setTimeout was significantly slower, but it's just _synthetic_. That was the main reason I was thinking so :)
      (Julien) I think setTimeout has the 4ms penalty, but not in all cases (usually nested setTimeout have the penalty). But in our case I don't think this helps because we have a stream of things, we don't wait for the setTimeout before doing the next task. (or I think ?). Not sure the 'nested' rule apply to us, in my previous tests I did see differences between apps and browsing... but don't remember exactly what :p
      (Oleg) Yeah, like mentioned here https://developer.mozilla.org/en-US/docs/Web/API/Window/setImmediate, yeah we don't have nested.
    • Started to profile with Built-In profile, trying to figure out if I can get more details from it. One thing that I don't understand is tiny consequent restyles, do you know what is it? Looks like incremental restyle :)
      (Julien) when you click on the restyle itself, you have the "HINT" on the panel that can tell you where it comes from. It could be an animation for example.
      (Oleg) They almost all "StyleAttribute", didn't find much info what is this exactly.
      (Julien) a "style" attribute change ? :)
      (Oleg) Yeah, but I can't think of cases where we change style attribute that much.
      (Julien) yeah; I'll look at your profile later today anyway :) Right now I'm just guessing :p
      (Oleg) Good!

Other:

  • Reviewed "bug 1160049 - [Messages] Attach menu should not dismiss when user cancel the replace attachment request", just one last question at github.

Today:

  • Will handle review/feedback/need-info requests;
  • Will work on review comments and assigned bugs.

Day 5: 31st August

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
    • Found that we didn't initialize the host correctly in the master. Although I create bug 1199612 to doing the same thing that copy the initHost from inbox startup to shared startup, it seems there's much more task to do to fix the messaging shim issue. Hi Oleg, do you think it's better to address the instance in bug 1199612 or I do it in my patch?
      (Oleg) I have almost ready PR that address both problems (init shim host in main startup + appInstanceId integration into MessagingService/Client), will be ready today so will ask you for review :) Just holding because I want to have integrating test for it - so that it's harder for me to break it again :)
      (Steve) That's fast :) I forgot to mention the app instance issue last week. Thanks!
  • bug 1198658 - [Flame KK][Message]Press the download button, there is no message indicating that you need to switch data connection
    • Reply some wilson's question on github. Patch is in review.

Today:

  • Create a bug to disable the message event broadcast
  • Layout for mozConnections shim update(based on shim host fix)

Julien

Note: PTO 2nd -> 10th September

  • worked on bug 1199593 (forward is broken with complex MMS); it was quick, but then I spent some time to try to do integration tests (which needs bug 1200116)
  • Spent some time on bug 1168363: an integration test (in the System app, but uses the SMS app) fails sometimes when loading the app. But I couldn't reproduce locally...
  • Wrote the mail summarizing our estimation for NGA to David
  • Had a meeting about prerendering to try to understand how we could work to not need it.
    • Filed bug 1199674 about the white flash and with profiles when changing panels.
    • I think we have an issue with gaia-header in the conversation panel because we don't use title-start/end in this panel (because we didn't need it for startup initially, and it has different cases: with/without call button).
  • still no work on system messages issue :(

Today: I want to:

  • do left-over reviews
  • work on the system messages issue

If all this moves forward well, I could:

  • continue the prototype caching the thread list to a single db (including contacts/drafts/etc).

Oleg

PTO from 2nd September to 4th September, inclusive, will be traveling with limited access to internet :)

  • bug 1155534 - [Messages][NG] Extract NewMessage view from Conversation view
    • No updates (in background).
  • bug 1176976 - [Messages][Drafts] Remove the draft saving/replacing action menu
    • No progress (in review).
  • bug 1199612 - [Messages][NG] API shim host didn't initialize correctly and break the MMS retrieval
    • Pretty ugly regression that comes from my ConversationService patch - I forgot that MessagingService uses shim host iframe on master now - so we need 2 things here - bootstrap shim iframe in the main startup (in addition to views/inbox/startup) + add support of AppInstanceId to MessagingClient and MessagingService;
    • PR is almost ready, prettifying, testing and finishing integration test (depends on the bug below) (in progress).
      (Julien) wondering if we should backout regressing bug until the fix is ready. I'm actually afraid foxfood users could get this bug and then wait for 1 month for the next update. I don't know if we got a new build last friday ?
      (Oleg) Yeah, foxfooders is my main fear, maybe it's better to backout. Do you have jlorenzo handy to ask?
      (Julien) not right now :) but I can also ask my phone :p
      (Oleg) Ah :) What is the right workflow to ask for backout?
      (Julien) do it yourself: git revert -m1 - <mergecommit>, do a PR, wait for green try, push.
      (Oleg) Okay, thanks! Will do it now.
      (Julien) OK, looks like we got no build this week. phew :)
  • bug 1200116 - [Messages][Tests] ThreadGenerator should be able to create threads with MMS
    • So eventually the time has come :) We need automatically generated MMS in the integration tests, spent some time to figure out how to make things work and seems found acceptable way, created PR, will ask for review once Treeherder is green (almost in review).
  • bug 1198266 - [Messages] Use ConversationService in the application
    • A "bit" slowed down on this one due to the bugs above, but will continue today (in progress).

Today:

  • Will handle review/feedback/need-info requests;
  • Will work on review comments and assigned bugs.

Day 6: 1st September

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
  • bug 1198658 - [Flame KK][Message]Press the download button, there is no message indicating that you need to switch data connection
    • Landed and file the DOMError clone issue. But I can't reproduce the split view issue.

Today:

  • Create a bug to disable the message event broadcast
  • Review the shim host init issue.
  • Layout for mozConnections shim update(based on shim host fix)

Julien

Note: PTO 2nd -> 9th September

  • did nearly all reviews:
    • bug 1200116 (thread generator to generate MMS)
    • bug 1192744 (fix the 12/24h switch in the report panel)
    • bug 1198658 (fix the MMS retrieval when we need to switch the SIM)
    • bug 1197619 (move checkboxes (edit mode) to web components)
    • bug 1180592 (mozMobileConnections (incl switchSim) implementation in the shim)
  • investigated bug 1190980 (caret and copy/paste behavior in recipient panel); Ting-Yu said there are still platform bugs so we need to check after the platform bugs are fixed.
  • had a meeting about the low storage condition. bug 891344 is now part of 2.5 so we'll need to sync with Gabriele (for the Dialer part) to have a common component for the dialog.
  • did a simple not finished patch for bug 1199674 to see if the gaia-header was the cause for the white flash -- but it's not (or at least not the only cause).
  • still no work on system messages issue :(

Today: I want to:

  • will do integration tests for bug 1199593
  • do left-over reviews
  • work on the system messages issue

If all this moves forward well, I could:

  • continue the prototype caching the thread list to a single db (including contacts/drafts/etc).

Oleg

PTO from 2nd September to 4th September inclusive.

  • bug 1176976 - [Messages][Drafts] Remove the draft saving/replacing action menu
    • Rebased on my latest integration patch since it had conflicts (in review).
  • bug 1169576 - [Messages][NG] Implement Conversation service: method for streaming joined threads and drafts list
    • Backed out PR from master and prepared a new one to reland once regression is fixed (waiting for dependency).
  • bug 1199612 - [Messages][NG] API shim host didn't initialize correctly and break the MMS retrieval
    • Finished PR, fixed unit tests and added integration test (in review).
  • bug 1200116 - [Messages][Tests] ThreadGenerator should be able to create threads with MMS
    • Got r+, fixed review comments and landed (landed).
  • bug 1198266 - [Messages] Use ConversationService in the application
    • More measurements and experiments, started to see perfromance penalty of

Intl.DateTimeFormat again, not sure why I stopped seeing it for some time (in progress).

    • Tried to measure MessageChannel performance with console.time/timeEnd, but seems these functions take more time than actual postMessage and skew result a lot (e.g. 20 ping-pongs cost 500ms with console.time (to measure length of ping and pong) and 40ms without it :)). Will check performance.now & friends.
      (Julien) yeah the functions themselves take time but I think they're quite accurate to measure the time between them.
      (Oleg) Yeah I thought of this, but now I'm not sure :) Will see maybe something odd in my test case.
    • A bit stupid idea, but want to try: theoretically in iframe we can reference to the parent "bridge" object to avoid loading it once again in iframe, it may help a bit since we don't have byte code caching yet.
      (Steve): I thought of this actually, but never tried it out since it might be safer to have 2 identical bridge in different context. Not sure how much time we can safe from reusing the bridge, but we can still give it a try to see if it's fine to use same bridge across different context.
      (Oleg) I'll just try so that we know for sure :)

Other:

  • Left feedback for "bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation";
  • Took a look at "bug 1199593 - [Message]The text content lost when forward the MMS.". Looks good, just need to test on device.

Today:

  • Will handle review/feedback/need-info requests;
  • Will work on review comments and assigned bugs.

Day 7: 2nd September

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
  • bug 891344 - [SMS/MMS] When storage is low with "Device space low" message displayed on status bar, device can't send nor receive sms/mms, we need to warn the user.
    • Reply the possible efforts and had some discussion about the possible platform efforts.
  • bug 1201016 - [Messages][NG] Migrate the current Message manager event handling to NGA.
    • Instead of create a bug to disable the message event from shim, maybe it's time to replace the event handling part in message manager.

Today:

  • Layout for mozConnections shim update(based on shim host fix)

Julien

  • In PTO

Oleg

  • In PTO

Day 8: 3rd September

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
    • Patch rebased, working on unit test
  • bug 1201016 - [Messages][NG] Migrate the current Message manager event handling to NGA.
    • Left some thought about the messaging client for views.

Today:

  • Layout for mozConnections shim update

Julien

  • In PTO

Oleg

  • In PTO

Day 9: 4th September

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
    • Patch rebased, working on unit test
  • bug 1201016 - [Messages][NG] Migrate the current Message manager event handling to NGA.
    • Left some thought about the messaging client for views. No progress
  • Reply some bugs that related to message app.

Today:

  • Layout for mozConnections shim update(based on shim host fix)

Julien

  • In PTO

Oleg

  • In PTO

Day 10: 7th September

Steve

  • bug 1179628 - [Messages][NG] Lay out Settings service structure
    • Start mozMobileConnections.
  • bug 1180592 - [Messages][NG] mozMobileConnections shim Implementation
    • unit test almost ready, will submit the patch for review today.
  • bug 1201016 - [Messages][NG] Migrate the current Message manager event handling to NGA.
    • Waiting for more feedback. No progress yet.
  • Reply some bugs that related to message app.

Today:

  • Create layout for mozConnections shim patch for review.

Oleg

  • bug 1176976 - [Messages][Drafts] Remove the draft saving/replacing action menu
    • Rebased once again due to the latest changes in integration tests (in review).
  • bug 1169576 - [Messages][NG] Implement Conversation service: method for streaming joined threads and drafts list
    • Dependency has been landed, so relanded patch again (landed).
  • bug 1199612 - [Messages][NG] API shim host didn't initialize correctly and break the MMS retrieval
    • Got r+, fixed review comments and landed (landed).
  • bug 1198266 - [Messages] Use ConversationService in the application
    • Not much updates since September 1st, Wilson replied - reading (in progress).
    • Still need to measure MessageChannel performance with performance.now & friends.

Other:

  • Reviewed "bug 1199593 - [Message]The text content lost when forward the MMS." + new integration test.

Today:

  • Will read my bug-mail;
  • Will handle review/feedback/need-info requests;
  • Will work on review comments and assigned bugs.

Julien

  • In PTO

Demos

Retrospective

What was good in the last sprint

  • We're using MessagingService in master!
  • I've got more experience on different profilers :)
    • Any other profiler then cleopatra?
    • Mostly Gecko (via Cleopatra) and WebIDE (I like it, but it seems a bit buggy, filed bug for that), and some raptor.
    • Is the use of WebIDE similar to cleopatra?
    • Yeah, they are quite similar, the only thing is that you can't profile application start with WebIDE - there are two tricks I used: 1. from Julien - start app -> start profiler -> fire location.reload(). 2. I've created fake empty start page with link to the actual inbox :)

What was bad in the last sprint

  • Not much progress for NGA at all :p
  • Regressions that had to be fixed.

Any questions

  • Should we plan "Low Device Storage" for this sprint? I feel a bit out of loop on this :)
    • That's what I asked from last week because the spec julien found seems not for 2.5, we can only fulfil part of the spec but I don't know the exact spec for 2.5.
    • Mmm, it feels like it is not really actionable.... Do you think we should do at least something in this sprint (like disabling controls) or just wait for the clear picture first? I think more about the latter, but you know better :)
    • Julien said the spec is defined by Katie(?) Maybe we can clarify the spec for 2.5 first. I can tried to ping the correct person.
    • Okay, let's find out who is this Katie and ni? her :) Yeah, it would be great if you can ping relevant person.
    • Do you think maybe we can create bugs after spec finalized? (in this sprint)
    • I think so

Actions for next sprint