[Trac #1236] browser crashing on download

[Zarro Boogs per Child]

#1236: browser crashing on download
-------------------------+--------------------------------------------------
 Reporter:  sj           |        Owner:  marco  
     Type:  defect       |       Status:  new    
 Priority:  high         |    Milestone:  Trial-1
Component:  web browser  |   Resolution:         
 Keywords:               |  
-------------------------+--------------------------------------------------
Comment (by jg):

 Replying to [comment:4 marco]:
 > Replying to [comment:3 jg]:
 > > The crash must be fixed.
 >
 > I'll see what I can do, I have not yet started looking into it so I
 don't know what it will involve.
 >

 Yup.

 > > The server should certainly report the right mimetype for an activity.
 >
 > We need to figure a good mime type for .xo.

 Dragging out of my brain, it will probably be something like
 application/sugar, or some such.  If you like, I can go deal with
 the IETF documents and figure it out.  We may need to register it
 with IANA as well.

 >
 > > We also need the general facility to invoke the right helper
 application for a mimetype, if    > there is no recognized plugin that
 supports it (this should not be hardwired into the browser; > otherwise
 people can't experiment with different viewers)..
 >
 > This is not going to happen for Trial1. It's not on my list of
 priorities for BTest-3 either. If someone contribute patches I will take
 them. A mime system is not something you can design and write in a few
 days.

 We weren't expecting a full blown media player supporting N datatypes to
 land on us at this date.  Sometimes plans have to be revisited.

 We'll see how to resource this if at all possible; I'll check with Walter
 on our options. The Real folks said they were contributing patches (might
 already be in trac), but they might hardwire to the Helix activity, which
 is not desirable since it might inhibit others.  I'll investigate their
 patches Monday, and see how they are implemented.

 >
 > >And yes, this is important enough that we'll test a fix (or
 enhancement, if you prefer to call >it that) to be confident we can deploy
 it.  This is essential to getting the browser really >usable (along with
 the back button bug).
 >
 > I'm unclear on what you want to see for Trial1. It might be a good idea
 to just implement .xo installation from the web browser. It should not be
 too difficult and it's a nice feature.

 It would be dynamite if .xo installation is easy and you can do it.

-- 
Ticket URL: <http://dev.laptop.org/ticket/1236#comment:5>
One Laptop Per Child <http://laptop.org/>