yet more -v stuff

Stephen Weeks MLton@sourcelight.com
Sat, 6 Jan 2001 13:54:48 -0800 (PST)


> That's a reasonable solution.  I guess it depends whether not the info
> is really useful.  Once you recognize which GC_gc call is being invoked,
> is it really reasonable to trace it back to some particular part of the
> source program?  I suspect this was difficult with the C backend; it would
> be more so with the x86 backend.

It wasn't difficult with the C backend.  The line number was the line of a GC_gc
in the generated C file.  This was useful for finding backend and gc bugs,
especially when doing things like GC_EVERY_CHECK, because I could see what code
was being executed just before the GC.