1052320 - [Tarako][fl] can not download anything when select OMA files

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Gaia::Browser, defect)

Description

[defang.zhang]

DEFECT DESCRIPTION:
  can not download anything  when select OMA files

Steps to reproduce:
----------------------------------------------------
1.open http://imps.tcl-ta.com/cailiang/content_test/download/download.html 
2.select "Sound" "Image" and so on 
3.not trigger download

Actual result:
---------------------------------------
 not response when select test web page OMA file
 

Expected result:
--------------------------------------- 
  normal download


Additional info:
--------------------------------------
Reproduce rate:  5/5

Comment 1

[defang.zhang]

In fact,we found the reason that caused by this problem.Modify bug 1026846 caused by this problem. Normal downlaod if we rollback bug 1026846 submission.
"type": { "value":"application/vnd.oma.dd+xml", "required":true } modify 
"type": "application/vnd.oma.dd+xml"

Comment 2

[David Flanagan [:djf]]

If you back out bug 1026846, then any app that tries to do a view activity request without specifying a type field will fail. It isn't clear to me what a view activity like that would be doing, so maybe backing out is okay.

It would be better, however, to understand what is actually happening here.  It appears to me that the files at http://imps.tcl-ta.com/cailiang/content_test/download/download.html are being served with the correct content-type. I wonder whether the type property is not being passed through to the activity properly. With the required:true code removed, it would be interesting to know what parameters are being passed to the activity when it is invoked...

I'm not sure who assigned this bug to me.  I can take it, but since it is not marked as a blocker of any sort, I don't know what kind of priority to give it. Are there operators selling the Tarako and depending on the DRM downloads? If so, then fixing this is presumably quite urgent.  But if not, then I won't worry about it too much.

Comment 3

[David Flanagan [:djf]]

ying xu: did you land any other DRM-related patches in 1.3T?  I assume that you currently have 1020002 and 1026846 in 1.3T, correct?  

Did you land 971618? That might have affected this, I suppose.

Comment 4

[ying.xu]

(In reply to David Flanagan [:djf] from comment #3)
> ying xu: did you land any other DRM-related patches in 1.3T?  I assume that
> you currently have 1020002 and 1026846 in 1.3T, correct?  
> 
> Did you land 971618? That might have affected this, I suppose.

I can't access the bug 1020002 and 1026846. It says
"You are not authorized to access bug #1026846."
So we don't merge them in 1.3t.

And we don't have bug971618 in 1.3t either.

This is the git log of fl app. On github, I only see you merge it into master. 
I don't know why it's in 1.3t.

ffos@ffos2:~/yingxu/6821/gaia/apps/fl$ git log .
commit b973b8079d07bf05d6cf925f3b4132e238b8d790
Author: David Flanagan <[e-mail]>
Date:   Thu Jun 19 14:39:06 2014 -0700

    Merge pull request #20712 from davidflanagan/bug1026846
    
    Bug 1026846: require the type parameter for FL app view activity r=amac

commit 07315d8a283fae4fb24e9f3cc65532ec933f255b
Author: punamdahiya <[e-mail]>
Date:   Wed Dec 11 16:45:05 2013 -0800

    Merge pull request #14542 from punamdahiya/Bug947028
    
    Bug 947028 - Fix of overlapping checkboxes while selecting ring tones r=davidflanagan
    (cherry picked from commit dea74850442df6aff8405333a973d49a9f1af2ec)

Comment 5

[ying.xu]

(In reply to ying.xu from comment #4)
> (In reply to David Flanagan [:djf] from comment #3)
> > ying xu: did you land any other DRM-related patches in 1.3T?  I assume that
> > you currently have 1020002 and 1026846 in 1.3T, correct? 
> > I can't access the bug 1020002 and 1026846. It says
> "You are not authorized to access bug #1026846."
> So we don't merge them in 1.3t.

I check the git log , we do have 1020002 and 1026846 in 1.3t
And we don't have bug971618 in 1.3t. So we may miss this bug.
I will try it today.


ffos@ffos2:~/yingxu/6821/gecko$ git log|grep -10 -i 1020002 
commit 550bd5051bc3206fce3f4da35f4bac1d374931c6
Author: B2G Bumper Bot <[e-mail]>
Date:   Wed Jun 11 12:56:27 2014 -0700

    Bumping gaia.json for 2 gaia revision(s) a=gaia-bump
    
    ========
    
    https://hg.mozilla.org/integration/gaia-1_3/rev/88318d7821be
    Author: David Flanagan <[e-mail]>
    Desc: Merge pull request #20093 from davidflanagan/bug1020002-v1.3
    
    Bug1020002 v1.3
    
    ========
    
    https://hg.mozilla.org/integration/gaia-1_3/rev/73f8e881f085
    Author: David Flanagan <[e-mail]>
    Desc: Bug 1020002: don't allow an activity choice that could subvert forward-lock


ffos@ffos2:~/yingxu/6821/gecko$ git log|grep -10 -i 1026846 
commit 7485bfb141102f46f0aa3d90eef525957e37b060
Author: B2G Bumper Bot <[e-mail]>
Date:   Mon Jun 23 13:00:22 2014 -0700

    Bumping gaia.json for 1 gaia revision(s) a=gaia-bump
    
    ========
    
    https://hg.mozilla.org/integration/gaia-1_3/rev/15b1d1141db1
    Author: David Flanagan <[e-mail]>
    Desc: Merge pull request #20712 from davidflanagan/bug1026846
    
    Bug 1026846: require the type parameter for FL app view activity r=amac

Comment 6

[defang.zhang]

(In reply to David Flanagan [:djf] from comment #2)
> If you back out bug 1026846, then any app that tries to do a view activity
> request without specifying a type field will fail. It isn't clear to me what
> a view activity like that would be doing, so maybe backing out is okay.
> 
> It would be better, however, to understand what is actually happening here. 
> It appears to me that the files at
> http://imps.tcl-ta.com/cailiang/content_test/download/download.html are
> being served with the correct content-type. I wonder whether the type
> property is not being passed through to the activity properly. With the
> required:true code removed, it would be interesting to know what parameters
> are being passed to the activity when it is invoked...
> 
> I'm not sure who assigned this bug to me.  I can take it, but since it is
> not marked as a blocker of any sort, I don't know what kind of priority to
> give it. Are there operators selling the Tarako and depending on the DRM
> downloads? If so, then fixing this is presumably quite urgent.  But if not,
> then I won't worry about it too much.

Dear David Flanagan:

after "type": { "value":"application/vnd.oma.dd+xml", "required":true } modify 
"type": "application/vnd.oma.dd+xml"
1.add log in download.js:
function view(activity) {
  debug('view() invoked for activity', activity.source.name,
        'with data', JSON.stringify(activity.source.data));
    console.log('view() invoked for activity', activity.source.name,
        'with data', JSON.stringify(activity.source.data));

  var descriptor = {};  // The parsed form of the download descriptor
  var mimeType;         // Set by checkType() to the primary media type
  var isImage = false;  // These are also set by checkType()
  var isAudio = false;
  var installType;      // Wallpaper, ringtone or song.
...
}
2.click "sound" in test web page,the log as follow:
E/GeckoConsole(  937): Content JS LOG at app://fl.gaiamobile.org/js/utils.js:7 in debug: [Locked Content] 0 view() invoked for activity v    iew with data {"type":"application/vnd.oma.dd+xml","url":"http://imps.tcl-ta.com/cailiang/content_test/download/sound.dd"} 
3.if you not modify,not the log when you click "sound" in test web page,I doubt that it should not trigger function view(activity).
4.I will validate bug 971618 today

Comment 7

[defang.zhang]

Dear David Flanagan:

After bug 971618 cherry-pick in 1.3t,the download not response.the log not found.

Comment 8

[David Flanagan [:djf]]

I can't reproduce this bug on my Flame, and don't have my Tarako with me right now to try it on that device. I can't see why adding the required:true parameter to the manifest would break anything. If you have to, just reverting bug 1026846 is probably okay, but it would be better to figure out why this is failiing.

Perhaps bug 1020002 is causing problems. That patch uses Array.prototype.findIndex() in the system app. That is supposed to be in Gecko 25, so it ought to be in 1.3t, but maybe that is breaking or something?

I have added Ying Xu and Defang Zhang to the cc list on bug 1020002 and 1026846, so you should both be able to read those bugs and see the patches now.  Maybe with access to those bugs you can figure out what is happening.

I'm going to unassign myself from this bug because I don't know that I can help any more than this.  If you have questions feel free to use needinfo. Or escalate the bug though the usual channels to get it assigned back to me.

Comment 9

[BMO Automation]

Firefox OS is not being worked on