{minutes} HTML WG media telecon 2014-06-03 - EME bug status

http://www.w3.org/2014/06/03-html-media-minutes.html

Joe Steele



HTML Media Task Force Teleconference

03 Jun 2014

Agenda

See also: IRC log

Attendees

Present
+1.781.221.aaaa, paulc, markw, joesteele, cpearce, +1.425.936.aabb, ddorwin, Niels_Thorwirth, [Microsoft], BobLund, +1.425.605.aacc, adrianba
Regrets
Chair
paulc
Scribe
joesteele
Contents

Topics
Agenda
Role Call
EME status and bugs
New EME bugs
[NEW] Bug 25595 - Better definitions needed for session, keys and license
Bug 25866
[NEW] Bug 25896 - Why is EME creating new DOMException subclasses?
[NEW] Bug 25899 - Do not specify messages for DOMExceptions
[NEW] Bug 25900 - [EME] MediaKeys.loadSession(sessionId) should specify what happens if a loaded sessionId is passed
[NEW] Bug 25901 - [EME] HTMLMediaElement.mediaKeys should be nullable
[NEW] Bug 25902 - [EME] Don't use Date for MediaKeySession.expiration
[NEW] Bug 25903 - [EME] Tighten up MediaKey Events definitions
[NEW] Bug 25911 - HTMLSourceElement's IDL keysystem attribute should reflect the keysystem content attribute
[NEW] Bug 25918 - Remove keySystem attribute from MediaKeySession
[NEW] Bug 25920 - Remove extraction of default URL from createSession() algorithm
[NEW] Bug 25923 - isTypeSupported should be asynchronous
Other EME bugs for discussion
Bug 18515: HTMLMedieElement behavior and possibly replacing waitingFor with a promise
Bug 25594: getUsablekeyIds() - synchronous vs promise
EME Use cases Wiki
[EME] Reuse of session
Summary of Action Items
<trackbot> Date: 03 June 2014
<scribe> scribe: joesteele
Agenda

Role Call

paulc: welcome Chris Pearce!
... (discussion of procedure)
EME status and bugs

paulc: editors draft available. Let's speak to email David sent yesterday
<paulc> Recent changes: http://lists.w3.org/Archives/Public/public-html-media/2014Jun/0002.html
paulc: anything in particular we should be looking at?
ddorwin: nothing in particular - just wanted to make folks aware
... (breaking up)
<cpearce> My microphone is muted
ddorwin: ClearKey is no implementable, common key is available, generic PSSH
paulc: PSSH was bug 24027
ddorwin: Mark sent an email about this
<cpearce> paulc: got it.
paulc: remember silence is assent -- please review changes
... closing bugs but new bugs have been filed
New EME bugs

[NEW] Bug 25595 - Better definitions needed for session, keys and license

https://www.w3.org/Bugs/Public/show_bug.cgi?id=25595
ddorwin: dropping off to rejoin
... most of these were filed by Mozilla and are clarifications
... first one does not follow that model
joesteele: think the text is pretty clear here
ddorwin: think this bugs needs to be addressed after use cases
paulc: use cases are at agenda #7 we will pick uyp after that
Bug 25866

https://www.w3.org/Bugs/Public/show_bug.cgi?id=25866
<scribe> [NEW] [NEW] Bug 25866 - \"needkey\" event name is misleading - "needkey" event name is misleading - "needkey" event name is misleading
paulc: what is the status?
ddorwin: need feedback on other names
... needkey is confusing since it is just providing initData
joesteele: I think this is OK for now. Unless we change what the initData actually is -- I would leave this as-is
... I will respond on the bug
[NEW] Bug 25896 - Why is EME creating new DOMException subclasses?

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25896
paulc: believe this may be related to the next bug as well
... Bug 25899
... have we discussed this before?
ddorwin: That is not an EME bug -- it is a WebApps bugs
... indicates we are violating the intent by subclassing
... we are going to lose system code in our exceptions and errors
... still put that in messages if desired
jdsmith: we were talking about DOMException before when we suggested this approach
... David you had a proposal?
ddorwin: that is in the next bug, messages can be customized by the user event. System code could be put in the message somehow
... main concern was that you could not respond or log them unless they were their own integer
... original proposal was to go with DOMException and if experience shows we need it we could add a new member
... now TAG is telling us to get rid of it
... still think if we do need a system code we could consider other options - like a method
... still have an error attribute - just would not be in what is fired
jdsmith: answers my question. we were concerned about being able to log
joesteele: I will add a comment as well
<paulc> jdsmith is going to add a comment to 25896
[NEW] Bug 25899 - Do not specify messages for DOMExceptions

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25899
paulc: was this covered by previous discussion?
ddorwin: no -- need to delete the spec text and the table
paulc: assigned to David and he will withdraw the change
[NEW] Bug 25900 - [EME] MediaKeys.loadSession(sessionId) should specify what happens if a loaded sessionId is passed

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25900
cpearce: might be my 2nd bug. Something I picked up while implementing - something not handled
... when you try to load session already loaded - what happens?
paulc: comment from David in bug
<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25900#c1
paulc: could return NULL
... that was acceptable?
cpearce: yes that is fine -- just needs to be consistent
paulc: assigned to David.
ddorwin: yes I will return NULL
[NEW] Bug 25901 - [EME] HTMLMediaElement.mediaKeys should be nullable

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25901
ddorinw: just a syntax thing I will fix
[NEW] Bug 25902 - [EME] Don't use Date for MediaKeySession.expiration

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25902
cpearce: our DOM experts picked up on this
... they say that Date is bad and should use unrestricted double
paulc: Boris agreed
ddorwin: agreed we should change
... I will change
[NEW] Bug 25903 - [EME] Tighten up MediaKey Events definitions

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25903
cpearce: we noticied ambiguity in the spec when initializing MediaKey message. Think we need to define the initData and message attrobutes are copied into the message keys event from the original struct. Otherwise if a reference is copied the array may change which is bad
paulc: 3 things in here - David have you looked?
ddorwin: not in detail yet -- makes sense but not sure what the text should be
cpearce: I will add recommended text to the bug
ddorwin: for array copies we did that for the params, not sure we need spec text since theses are coming from the browser
cpearce: we have IDL interfaces in Javascript - could be defined by application
[NEW] Bug 25911 - HTMLSourceElement's IDL keysystem attribute should reflect the keysystem content attribute

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25911
paulc: no responses - but exact proposal from Chris
ddorwin: he is adding the last sentence
<paulc> Proposing to add: The keySystem IDL attribute must reflect the keysystem content attribute."
paulc: any comments?
ddorwin: I will take this
[NEW] Bug 25918 - Remove keySystem attribute from MediaKeySession

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25918
ddorwin: this is a dup of what is on MediaKeys -- sugest we remove
paulc: any comments?
[NEW] Bug 25920 - Remove extraction of default URL from createSession() algorithm

+q
<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25920
<paulc> See Joe's response: https://www.w3.org/Bugs/Public/show_bug.cgi?id=25920#c1
joesteele: my response was probably a bit strong - but would like more information
ddorwin: putting information in the PSSH and relying on it is one of the interop issues we have discussed here and elsewhere
... folks that do not want to rely on the PSSH will need to know the URL anyway
... I am also concerned about the media data telling the application which servers to ping
paulc: what is the interop bug you are thinking of?
<ddorwin> https://www.w3.org/Bugs/Public/show_bug.cgi?id=20944#c44
joesteele: don't think that removing the URL adds anything to the security and privacy since origin rules apply
jdsmith: think that MS actually proposed this in the beginning -- may actually need it as well
[NEW] Bug 25923 - isTypeSupported should be asynchronous

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25923
paulc: This comes from Ana
... fair amount of discussion
ddorwin: I am looking for feedback from app developers
... major change in how this works
paulc: filed on May 30th - already 6 comments
... encourage folks to look at this and will be on next weeks agenda
Other EME bugs for discussion

Bug 18515: HTMLMedieElement behavior and possibly replacing waitingFor with a promise

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18515#c28
paulc: David, this is your comment
ddorwin: that is mostly how MediaElement should work
... looking for feedback (from Jerry and others)
... #29 is another option but not a clear win
jdsmith: I will give feedback
Bug 25594: getUsablekeyIds() - synchronous vs promise

<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25594
ddorwin: we discussed this briefly before
... need a Promise or a method
... I proposed to use Array, but that is going away
... now proposing an Object
... don't think there is much debate except for whether a method or a Promise
... I have no real opinion
... pros and cons
paulc: one of the editors must assign this
... any comments on this?
... will someone volunteer to look at this?
EME Use cases Wiki

<paulc> See Joe's promise: http://lists.w3.org/Archives/Public/public-html-media/2014May/0044.html
paulc: joe promised to look at this but has not
<paulc> Not done yet.
joesteele: will provide some initial input this week but will not be able to make next week most likely
paulc: any objections to convening in 2 weeks?
... ok next meeting will be on the June 17th
... how should we discuss this?
ddorwin: hoping folks would add detailed application models - has only been discussed in the abstract so far
... want to know what we need to do and then discuss the how
... then we could discuss new APIs etc.
... that is one of the reasons I changed the format to be more specifc
joesteele: would like to see these use cases referred to specifically in the bugs
ddorwin: Niels sent an email which referenced to CDM -user agent interaction
... don't think we need to cover that
niels_t: It does relate to some of the functionality that CDMs have
... does not directly impact the specifiction we are talking about
... but will affect it -- that is why I put it in
ddorwin: we may need another section of informative stuff
<paulc> Niels email: http://lists.w3.org/Archives/Public/public-html-media/2014May/0040.html
niels_t: did not see a single hierarchy with one view of the system -- we have multiple views of the system
... none of these cover all of the cases
paulc: almost out of time
... more work on wiki coming
... Joe if you make the changes please send a note to the list to let people know
... you can track changes via your W3C ID
[EME] Reuse of session

<paulc> http://lists.w3.org/Archives/Public/public-html-media/2014May/0070.html
paulc: is there a cluster of bugs here? or just related to one bug?
<paulc> https://www.w3.org/Bugs/Public/show_bug.cgi?id=25268
paulc: Shinya pointed to a bug not being resolved appropriately
ddorwin: this thread is a month old and touched on a bunch of issues
... mostly talking about needkey causing session creation
... I think key rotation is the issue
... need to recall this
paulc: think we are through the agenda now -- no time for more bugs
... REMINDER -- we will meet in 2 weeks June 17th
... Thanks all and welcome again to Chris
... appreciate you getting up at this early hour ;-)
... bye
Summary of Action Items

[End of minutes]
Minutes formatted by David Booth's scribe.perl version 1.138 (CVS log)
$Date: 2014-06-03 16:07:11 $