Re: scm_must_realloc problem with "make check" in current stable tree.

[Marius Vollmer]

Rob Browning <address@hidden> writes:

> Marius Vollmer <address@hidden> writes:
> 
> >> This is with --enable-debug-malloc,
> >
> > I don't see this... but it should be easily analyzed, right?
> 
> Don't know.  I don't know much yet about the garbage collector or the
> debugging allocator, so I don't really know what the message is
> telling me, or how would be appropriate to track it down.  I'd be
> happy to work on this, but where should I start (are there relevant
> docs)?

I would run the program in gdb until it aborts, and then go up the
call stack to find the offending call to scm_must_realloc, and then
start to think hard about what is wrong with it.
 
> The problem doesn't occur if you add --enable-guile-debug, but comes
> back if you also add --enable-debug-freelist.  It also doesn't occur
> if I add a couple of display statements to the test suite's lib.scm,
> so it seems like it may be some kind of memory stomping problem.

Oops, this does make it sound much harder than I thought it would
be...