Re: [Gcl-devel] lcm added back to random tester

[Camm Maguire]

Greetings!

1) Great to hear from you!  I was going through Paul Dietz withdrawal
   :-).

2) I'm trying to reproduce what you describe now, but if you have a
   chance to post a small test case, this would be helpful.

3) I'm assuming you've seen the tests I've put into misc.lsp as you
   requested, and that they are ok.  I'm also assuming that you are
   also seeing the speed improvements in the tester that I'm seeing
   with the compiler change of a few weeks ago. (Not that further
   improvements cannot be made, of course.)  You might also be
   interested in the new time macro.

4) Did you get my earlier message about backporting a minimal number
   of your files in ansi-tests to the stable branch to enable running
   the random tester against this version?  Advice appreciated here.  

5) I'm also assuming that apart from what you report below and the
   known lack of load-time-value implementation, your random tester
   runs are also proceeding essentially indefinitely without showing
   failures.  Please let me know if this is not the case.

Your contributions in both the ansi and random tester arenas are
invaluable, IMHO, and much appreciated.

Take care,

"Paul F. Dietz" <address@hidden> writes:

> I've removed the #-gcl from the occurences of LCM in the random
> form generator.  This reexposes a bug in gcl related to division
> by zero (and also exposes the bug that this condition isn't always being
> caught by the random tester, breaking the pruner.)
> 
>       Paul
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah