#8920 NORM 9.1.0: Activity bundle icons in Journal aren't colored correctly

[Zarro Boogs per Child]

#8920: Activity bundle icons in Journal aren't colored correctly
--------------------+-------------------------------------------------------
 Reporter:  Eben    |         Owner:  tomeu        
     Type:  defect  |        Status:  new          
 Priority:  normal  |     Milestone:  9.1.0        
Component:  sugar   |       Version:  not specified
 Keywords:          |   Next_action:  diagnose     
 Verified:  0       |     Blockedby:               
 Blocking:          |  
--------------------+-------------------------------------------------------
 This is a deeper issue than it sounds, on the surface, and I title the bug
 in this manner simply because I don't understand the root of the issue.

 To reproduce the problem, simply look at the icon for any activity bundle
 which has been downloaded to the Journal.  I'll reference the list view
 for the purposes of this ticket, but the problem surfaces in the detail as
 well.  As you can see in the code [1], the icon should appear as a gray
 outline.  This is not the case.  Most frequently, it will look like a
 black outline, but in truth it is rendered in the colors specified in the
 SVG in the bundle, whatever they might be.

 After some experimentation, I've discovered an interesting difference
 between the well-behaved SVGs (on non-bundle entries) and those for
 activity bundles.  By adding this to line 71 [2]:

 {{{
 logging.error("SVG filepath: %s" % self.props.file_name);
 }}}

 It becomes evident that the bundle icons are first copied to /tmp in the
 form:

 {{{
 /tmp/tmp[random_hash][icon_name]
 }}}

 Diffing the icons in /tmp with those in the bundle reveals that they are
 identical.  The only other differences I can spot are:

  1. The permissions are different:  600 in /tmp, but 644 in the bundle
 itself.
  2. The file in /tmp lacks the .svg extension.

 Anyone have a clue if either of these details could be causing the
 problem, and how to fix it?  Since the image appears correctly, I think we
 can rule out (1).

 [1]
 http://dev.laptop.org/git?p=sugar;a=blob;f=src/jarabe/journal/collapsedentry.py;h=c2cc9c8a13209b8a88bc3492ac96f8708b4f45e7;hb=HEAD#l232
 [2]
 http://dev.laptop.org/git?p=sugar;a=blob;f=src/jarabe/journal/collapsedentry.py;h=c2cc9c8a13209b8a88bc3492ac96f8708b4f45e7;hb=HEAD#l71

-- 
Ticket URL: <http://dev.laptop.org/ticket/8920>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system