On 9/3/05, Justin Haygood <[EMAIL PROTECTED]> wrote: > I had this problem when I first started porting the collector. However, the > code I used worked after fixing it. The patch on Bugzilla that i submitted > for it should contain the code that works. The error is probably the stack > base and address aren't the correct ones. Have you debugged the program at > the crash and looked at those values, and see if they make sense? > > If not, try using a different method for determining the base address. If > they do make sense however, I don't know what to tell you
Yes, I'm positive that the stack base is correct and I've verified that the value is on the stack and is being scanned during marking stage of GC. But, as i described, it is not recognized as a valid CollectorCell, because the compiler, before putting the value on the stack, increases the pointer by 4 (I assume the reson for that are implict casts between ValueImp and AllocatedValueImp). -- kjk _______________________________________________ webkit-dev mailing list [email protected] http://www.opendarwin.org/mailman/listinfo/webkit-dev
