403932 - Fx3b1 is unable to load http://ejohn.org/files/ecma-cloud.svg

Status: VERIFIED FIXED

(Core :: SVG, defect, P3)

Description

[Frédéric Buclin]

http://ejohn.org/files/ecma-cloud.svg loads within one second with Fx 2.0.0.9 but still doesn't load after 10 minutes with Fx3b1; I had to kill the application -> regression. Maybe that's a dupe of bug 313314, but as I'm not a SVG expert, I'm filing this bug just in case.

Requesting blocking as it completely freezes my browser and is a regression.

Comment 1

[Damon Sicore (:damons)]

On OS X this is slow, but definitely faster than Linux.  Still, there's something bad happening here.  +'ing and setting to P3.

Comment 2

[Robert Longson [:longsonr]]

Not a regression but a consequence of new functionality in Fx 3 which implements parts of the SVG spec that Fx 2 ignores i.e. filters.

Our implementation of GaussianBlur filtering is very slow. Completing bug 399488 would help.

Comment 3

[Philip Potter]

I'm seeing this on Windows too.

Comment 4

[Robert Longson [:longsonr]]

We end up creating an approximate 1000 x 1000 pixel surface for each filter. That seems awfully large.

Comment 5

[Wesley Johnston (:wesj)]

This should be helped by Bug 416305.

Comment 6

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

Actually bug 414784 is the bug that will fix this.

Comment 7

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

Should be fixed now.

Comment 8

[Frédéric Buclin]

Still doesn't work with Mozilla/5.0 (X11; U; Linux i686; fr; rv:1.9b4pre) Gecko/2008021704 Minefield/3.0b4pre. It probably doesn't has the fix yet. Will try with the next nightly.

Comment 9

[Frédéric Buclin]

Verified fixed with:

Mozilla/5.0 (X11; U; Linux i686; fr; rv:1.9b4pre) Gecko/2008021804 Minefield/3.0b4pre

The loading time is now comparable to Fx2.

Comment 10

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

Except Fx2 didn't display the shadows and now we do :-)