Re: Fix backtrace for interpreted code

[Bryce McKinlay <mckinlay at redhat dot com>]

Andrew Haley wrote:

> Anthony Green writes:
> > On Thu, 2004-10-14 at 07:48, Andrew Haley wrote:
> > > That's right.  The solution I came up with is the only one I know to
> > > be correct.
> > 
> > Does the mingw port support named sections?  If so, we could put the
> > special method in its own section and surround it with
> > _Jv_StartOfInterpreter, etc in the linker script.
>
> I suppose so.  I guess this means we'd have to create a customer
> linker script for mingw.  I don't object, but I can't do it.
>
> Surely it's better to fix MD_FALLBACK_FRAME_STATE_FOR to unwind
> through SEH frames.  That would bring many benefits.
>  

Indeed. If we want proper stack traces on Windows - particularly ones 
that are good enough for the security model, this must be done. The only 
alternative I can think of is to have GCJ to emit both start and end 
pointers for each function as part of our metadata, and use that to 
associate IPs with functions instead of the Unwind library.

Bryce