Re: [clisp-list] Re: Clisp 2.36 for x86-64 crash on `make check-tests'

[Dr. W. F. <we...@su...>]

On Mon, Dec 12, 2005 at 07:33:58PM +0100, Dr. Werner Fink wrote:
> On Fri, Dec 09, 2005 at 10:46:06AM -0500, Sam Steingold wrote:
> > > * Dr. Werner Fink <jreare@fhfr.qr> [2005-12-08 14:09:29 +0100]:
> > >
> > > Hi,
> > >
> > > this is the result from a compile for Linux on x86-64 with gcc 4.1.0:
> > >
> > > Signal 6 is SIGABRT on x86-64. The crash is always at the same
> > > point regardless of the choosen optimization or SAFETY level.
> > 
> > http://sourceforge.net/tracker/index.php?func=detail&aid=1376646&group_id=1355&atid=101355
> 
> Does this mean that 2.35 also chrash on AMG64?
> Btw: A few printf's for debugging and the addresses
> and the file location found by gdb change a lot.
> In other words: the addresse space is fully broken.

Q: Does the limit of 36 bits for mmap() ind spvw.d makes sense 
in combination of 48 bits choosen for the address handling?
Q: Is it known that --with-debug leads to an error within the
gc_unmarkcheck() function  of spvw_garcol.d forced by -DDEBUG_SPVW.


     Werner

-- 
  "Having a smoking section in a restaurant is like having
          a peeing section in a swimming pool." -- Edward Burr