Re: [Gcl-devel] The state of gcl/maxima

[Mike Thomas]

> If you can *compile* maxima-5.6, and you've undefined SEPARATE_RSYM
> and defined SEPARATE_FASL_FILE to sfaslbfd.c, then bfd is working.  If
> you can define the GET_FAULT_ADDR macro as in x86 linux, and then
> define SGC, and successfully run 'make test' on a maxima built from
> gcl with these settings, then SGC is working.

Thanks.

Unfortunately I am unable to compile Maxima at all now.  eg:

===============================================
;      - Loading binary file "binary-gcl/risch.o"
Loading binary-gcl/risch.o
start address -T 10953000 Finished loading binary-gcl/risch.o
;    - Compiling module "taylor-series"
Error in LET* [or a callee]: Cannot get the truename of #p"".
Fast links are on: do (use-fast-links nil) for debugging
Broken at NAMESTRING.
 1 (Abort) Return to break level 1.
 2 Return to top level.
dbl:>>>
===============================================

Also, the new saved_ansi_gcl executable has problems when things go wrong
eg:

===============================================
$ ./unixport/saved_ansi_gcl.exe
GCL (GNU Common Lisp)  Version(2.5.0) Wed Jun 26 10:50:41  2002
Licensed under GNU Library General Public License
Contains Enhancements by W. Schelter

>(load "barfo.lsp")  ;;; Non-existent file

Unrecoverable error: bind stack overflow.

abnormal program termination

===============================================

The ansi image does however compile and load files which  exist without
problems.

On top of this, I can't get gdb to work properly with gcl!

I'm going to be working on the Windows version of the latest Grass GIS
pre-release for a day or two, so may not be able to get back to you on these
issues until finished there, sorry.

Cheers

Mike Thomas.