87731 - imagelib should associate a loadgroup with its http requests

Status: RESOLVED WONTFIX

(Core :: Graphics: ImageLib, defect, P2)

Description

[Darin Fisher]

imagelib should associate a loadgroup with the http channels it creates.
the loadgroup's default request needs to point to the channel used to
load the toplevel document.

this is needed to fix:

http://bugzilla.mozilla.org/show_bug.cgi?id=83526
http://bugzilla.mozilla.org/show_bug.cgi?id=87388

Comment 1

[Stephen P. Morse]

*** Bug 87388 has been marked as a duplicate of this bug. ***

Comment 2

[Stuart Parmenter]

i'm going to have to create a fake loadgroup to associate with the channels.. 
sigh.

Comment 3

[Jason Kersey]

-> 0.9.4

Comment 4

[Stuart Parmenter]

.

Comment 5

[Stuart Parmenter]

talked to darin about it.. marking wontfix.  the way imglib manipulates necko 
won't allow this to work without large changes to imglib and/or necko.

Comment 6

[J Luh]

Does this mean that Mozilla will not support third-party cookie blocking (bug
87388)?

Comment 7

[Stephen P. Morse]

But this blocks cookie bug 87388.  If we don't fix this, then we will not be 
able to block third-party cookies.  Netscape browsers have always provided this 
facility and people now expect it.

(Oops, I just saw jamesl's comment (I hadn't seen it when I started typing 
this).  He's saying the same thing that I am.

Comment 8

[Stuart Parmenter]

you'll have to find another way to do it.  this way isn't going to be implemented.

Comment 9

[Stuart Parmenter]

imagelib does not associate a single load with the channel load.  There is no 1
to 1 mapping between loadgroups and channels in imagelib.

Comment 10

[Judson Valeski]

this is a prerequisite for browser behavior 101. we need to solve this one way
or another. if this isn't the right place, pav/morse/darin can you hash out a
new model and file a new bug?

Comment 11

[Stuart Parmenter]

i'm going to mark this wontfix again.  I believe we are on the track to fixing 
the 3rd party cookie stuff in bug 107579.