[Gcl-devel] Solaris basically working (was Re: tests uuencoded)

[Camm Maguire]

Greetings!  Thanks for the solaris gcl/maxima test results, Dan!
Only two differences, right?

1) Could you please send me again the full build output for gcl, as
   you did last time, but with the latest cvs, and also include your
   maxima build output?

2) Can you try running rtest15.max by itself?  Is the error
   reproducible?  Can you narrow it down?  Can you build with -g only
   and optimally figure out what's happening at the C level?

Thanks again!  I'm hoping now that gcl is the only vehicle taking
maxima to sparc architecture.  I'm going to review the differences
between sparc-linux.h and solaris.h to see if I can cure a SIGBUS I'm
getting there.

One thing that will have to change in the solaris.h is the CLEAR_CACHE
macro, as indicated in the other thread on fasloading.  This might be
the culprit in the mistakes you've found.  It would be very helful if
you can figure out for me what the assembly instructions are on your
sparc hardware to flush the data cache at a certain address.

Take care,


>The number of differences found was 2 in problems: (4 3)

>Error summary:
>Error(s) found in rtest15.mac: (4 3)
>real time : 14.000 secs
>run time  : 13.360 secs


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