509301 - Value chosen for callee in args_resolve is nonsense

Status: RESOLVED FIXED

Description

[Jason Orendorff [:jorendorff]]

Symptomless bug, but the reinterpret_cast I just introduced in bug 506721 would actually be kind of dangerous if the value were used. It's not.

Comment 1

[Jason Orendorff [:jorendorff]]

Attachment: v1

Comment 2

[Jason Orendorff [:jorendorff]]

Uh, to clarify, it's not the reinterpret_cast that's dangerous. That INT_TO_JSVAL was already meaningless before my change.

The change in bug 506721 didn't affect the meaning of that code, except I guess it might flunk an assertion now.

Comment 3

[Brendan Eich [:brendan]]

The patch for bug 507573 had a s/OBJECT/INT/ thinko affecting the line in question. Better to use null or undefined, since the getter never reads *vp and only writes it, for sure.

/be

Comment 4

[Jason Orendorff [:jorendorff]]

http://hg.mozilla.org/tracemonkey/rev/3d4074b8928e

dvander reports that bug 506721 causes an assertion in debug browser builds. This might fix it. If not, the assertion is catching a thinko somewhere else...

Comment 5

[Brendan Eich [:brendan]]

(In reply to comment #4)
> http://hg.mozilla.org/tracemonkey/rev/3d4074b8928e
> 
> dvander reports that bug 506721 causes an assertion in debug browser builds.

Isn't the bug dvander filed bug 509557 ? It's different, but there could be more like it.

/be

Comment 6

[Jason Orendorff [:jorendorff]]

Right, my patch in bug 506721 added two assertions, each of which trips on every debug browser startup, due to pre-existing bugs. The first is bug 509557. I don't see the other one on file yet. Maybe I'm searching wrong.

Comment 7

[Robert Sayre]

http://hg.mozilla.org/mozilla-central/rev/3d4074b8928e