198435 - [FIX]Setting src on broken image broke in standards mode

Status: VERIFIED FIXED

(Core :: Layout: Images, Video, and HTML Frames, defect, P1)

Description

[Boris Zbarsky [:bzbarsky]]

The old code used to check for an image frame and reframe if there wasn't one...
I guess we need to try to do that here too.  Of course it's not clear how this
affects <object>....

Perhaps a better way is to reframe if the CURRENT_REQUEST has an error status. 
But what happens if the CURRENT_REQUEST is simply null?  That's covered by
http://web.mit.edu/bzbarsky/www/testcases/image-loading/appendNoSrcAndSetSource-standards.html

Comment 1

[Marius Strumyla]

bug 198989 seems related to this one. u dupe maybe?

Comment 2

[Boris Zbarsky [:bzbarsky]]

Attachment: Proposed patch

Comment 3

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 120545 [details] [diff] [review]
Proposed patch

Reviews?  The basic idea here is that if we CantRenderReplacedElement no one is
watching the notifications from the content object anymore, so the image never
gets shown if the src is changed...

I don't like this patch very much, but the real fix of rewriting the way
CantRenderReplacedElement works is not something I can pull off for 1.4...
(because I'm not sure how it _should_ work, truth to tell).

Comment 4

[Robert O'Callahan (:roc) (email my personal email if necessary)]

can't we just always reframe when the SRC changes? It's only one frame so we
probably won't notice the cost. And that would be pretty easy to do, right? The
same thing would work for OBJECT.

Comment 5

[Boris Zbarsky [:bzbarsky]]

> can't we just always reframe when the SRC changes?

That would make image rollovers suck like no tomorrow (I did try this before
posting this patch).  Also, a lot of "dhtml" sites reset src in tight loops for
the fun of it..

Comment 6

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 120545 [details] [diff] [review]
Proposed patch

OK. There must be a better way to do this though. I wonder if we could handle
broken loads by setting something on the content which triggers XBL with the
"broken image".

Comment 7

[Boris Zbarsky [:bzbarsky]]

Maybe..  Like I said, our way of handling of broken loads in general needs a
nice thwacking.  

Comment 8

[Johnny Stenback (:jst)]

Comment on attachment 120545 [details] [diff] [review]
Proposed patch

sr=jst, I guess that's about as good as this'll get for now...

Comment 9

[Boris Zbarsky [:bzbarsky]]

Fixed.

Comment 10

[[e-mail] (gone - use [e-mail] instead)]

I've filed bug 202346 for a crash regression.

Comment 11

[Kathleen :Brade]

*** Bug 202289 has been marked as a duplicate of this bug. ***

Comment 12

[Stephen Donner [:stephend] Not actively reading bugmail]

Verified FIXED using the testcase at
http://web.mit.edu/bzbarsky/www/testcases/image-loading/appendNoSrcAndSetSource-standards.html
on Windows XP with build 2004-11-12-04