Re: New stack trace code

[Bryce McKinlay <mckinlay at redhat dot com>]

Andrew Haley wrote:

> I've found a few problems.
>
> We're not marking the line table in boehm.cc:
>  
Doh, manual marking bites again!

> Index: boehm.cc
> ===================================================================
> RCS file: /cvs/gcc/gcc/libjava/boehm.cc,v
> retrieving revision 1.45
> diff -u -r1.45 boehm.cc
> --- boehm.cc    25 Nov 2004 03:46:53 -0000      1.45
> +++ boehm.cc    11 Mar 2005 18:52:31 -0000
> @@ -289,6 +289,8 @@
>                    = (_Jv_InterpMethod *) ic->interpreted_methods[i];
>                  if (im)
>                    {
> +                     p = (GC_PTR) im->line_table;
> +                     MAYBE_MARK (p, mark_stack_ptr, mark_stack_limit, ic);
>                      p = (GC_PTR) im->prepared;
>                      MAYBE_MARK (p, mark_stack_ptr, mark_stack_limit, ic);
>                    }
>
> Also, the calls to _Jv_AllocBytes in stacktrace.cc are potentially
> problematic: unless they're rechable from somewhere we scan, the
> memory may be freed before it gets used.
>  
The pointers are always kept on the stack, I think, so should be marked 
fine - but I'll check this and maybe add a comment.

Bryce