Bug 4249 – SVG JavaScript bindings are busted in latest WebCore+SVG

• Component: SVG
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

REGRESSION: JavaScript support is busted in WebCore+SVG

JavaScript used to work fine in WebCore+SVG.  I believe this was a result of mjs JavaScriptCore 
optimizations.  (http://bugzilla.opendarwin.org/show_bug.cgi?id=4124)

[Eric Seidel (no email)]

Created attachment 3192 [details]
One example of a busted SVG (which used to work)

[Eric Seidel (no email)]

Ben is working on this one.

[Eric Seidel (no email)]

Actually, this is going to require more heavy lifting than ben will be able to give us.  Maciej or I need to re-
work the underlying WebCore javascript bindings to be autogenerated, and then make SVG use those 
autogenerated bindings as well.  With KDOM now removed from WebCore+SVG it's no longer possible for 
the SVG javascript bindings to sit on top of KDOM's bindings, but rather they need to be ported over to 
WebCore's once WebCore's bindings are fixed to be autogenerated.

[Joost de Valk (AlthA)]

Adding Regression keyword.

[Eric Seidel (no email)]

Basic DOM autobinding support has been landed. It will probably be a couple more weeks before it's ready to support SVG however.

[Eric Seidel (no email)]

Marking as P2, and adding to the SVG HitList.  We could never ship SVG w/o some form of SVG JS bindings.

[Eric Seidel (no email)]

Created attachment 8601 [details]
First stab and turning some of these back on

This patch is not quite ready for review.  I'm posting it hear for discussion purposes.

[Eric Seidel (no email)]

Created attachment 8610 [details]
Much better patch, ready for review.

This patch also has some various style cleanups (sorry).

[Anders Carlsson]

Comment on attachment 8610 [details]
Much better patch, ready for review.

Some comments:

-    if (@{$dataNode->parents} > 1) {
-        die "A class can't have more than one parent";
-    }
+    #if (@{$dataNode->parents} > 1) {
+    #    die "A class can't have more than one parent";
+    #}

There should be a comment explaining why this is commented out.

I think the IDL files should be LGPL