5.9.0rc3 build failure in numerical/slatec/zbesy.lisp, gcl 2.4.3, gcc 3.2

Hi Valerij,

Thanks for the tip, compiling gcl 2.4.3 with gcc 2.96 gives a gcl which
is able to compile maxima.

Compiling gcl 2.4.3 with either gcc 2.96 or gcc 3.2 gives me some mild
problems: i have to manually adjust h/object.h to deal with `bool',
which is in conflict with gcc's ideas about bool, and i have to
manually adjust makedefs and/or makedefc to force the info files and
emacs files to go where i want them to.  On the other hand, compiling
maxima itself is very simple, using gcc 2.96-based gcl 2.4.3 per your
suggestion.  In particular, i didn't have to set C_INCLUDE_PATH as i did,
again per your suggestion, for the gcc 3.2-based gcl 2.4.3.

Thanks also for the advice about using cvs, but i have problems
dealing with the huge influx of files on my isp.  (I know of two
different ways to get the cvs files, depending on where i issue the
commands, but both of them have bottlenecks that make it impractical.)
Anyhow, i can avoid this for now.

Thanks again, and thanks also to Jim and RJF for their earlier remarks.

dan