diff -r -u mozilla/js/rhino/src/org/mozilla/javascript/optimizer/
Codegen.java mozilla-patched/js/rhino/src/org/mozilla/javascript/
optimizer/Codegen.java
--- mozilla/js/rhino/src/org/mozilla/javascript/optimizer/Codegen.java
2008-03-25 07:32:27.000000000 -0700
+++ mozilla-patched/js/rhino/src/org/mozilla/javascript/optimizer/
Codegen.java 2008-04-30 15:53:21.000000000 -0700
@@ -1541,8 +1541,17 @@
}
}
}
-
- if (fnCurrent != null && !inDirectCallFunction
+
+ // This is site of a Jetstream specific modification that is
not Mozilla.
+ // The the reason for the change is that Jetstream requires
static scoping
+ // for functions that do not require activation. This is
because Jetstream
+ // uses a class derived from ScriptableObject as a kind of
package. The
+ // code that is supposed to be in the package is executed
against the object
+ // (which is parented to the global scope) and when it is
done the Package
+ // object contains all of the functions and other variables
defined in the
+ // package source file as properties. Direct call was causing
these functions
+ // to lose their static scoping.
+ if (fnCurrent != null
&& (!compilerEnv.isUseDynamicScope()
|| fnCurrent.fnode.getIgnoreDynamicScope()))
{
FWIW, it seems that the dynamic scoping check is completely vestigial.
If I'm not mistaken (and I easily could be) dynamic scoping is handled
entirely inside of ScriptRuntime now so it's probably safe to get rid
of that check as well.I think it should be safe to always load the
proper static scope, even if a direct call is being made and I can't
imagine that the performance cost is atrocious. That being said, I
don't have a good benchmark to prove it.
It should be noted that without the above patch, Rhino behaves very
differently with respect to certain kinds of scoping between the
compiler and interpreter (the code in the previous post shows how to
reproduce this if anyone is interested). I'm already maintaining a
patch-set for my project so it isn't a big deal, but some change
should probably be made to improve the consistency between the
interpreter and optimizer.
Cheers,
Ben
_______________________________________________
dev-tech-js-engine-rhino mailing list
dev-tech-js-engine-rhino@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-tech-js-engine-rhino
