Re: [Reduce-algebra-developers] FreeBSD patch

[Rainer S. <rai...@gm...>]

Hello Pedro,

On Mon, 2 Apr 2012 at 23:13 -0700, Pedro F. Giffuni wrote:

 > Thank you Rainer!
 > 
 > --- Lun 2/4/12, Rainer Schöpf <rai...@gm...> ha scritto:
 > 
 > >  > 
 > >  > The attached patch just followed the instructions and
 > >  > lets me build the CSL version.
 > > 
 > > Thanks for pointing this out; please try the changes I
 > > checked in today.
 > > 
 > 
 > It built on FreeBSD 9.0-amd64 (actually PC-BSD),
 > Thank You!
 > 
 > Hmm.. it's very weird but typing "gmake install"
 > brought the system to a busy state and eventually I
 > had to coldboot, I had never happened to me!

It's an infinite recursion of "gmake install" invocations (took me a while to 
understand why).  There is no toplevel "install" target, instead it is passed to 
a shell script that calls make again.  Through some error, the "install" target 
is passed to each and every sub-make, leading to an infinite recursion.

An install target needs extra work; I believe the best way is to build packages 
for the various operating system.  I've already created the necessary files to 
build a package for Debian or Ubuntu Linux (in subdir debianbuild).  Someone 
familiar with the FreeBSD package building could easily adapt this. 

 > 
 > >  > The PSL version gets to here:
 > >  > ___
 > >  > 
 > >  > ...
 > >  > configure: Will build this PSL using the unknown
 > > initial binaries
 > >  > cp:
 > > /usr/ports/math/reduce/work/reduce-algebra/psl/psl-unknown/*:
 > > No such file or directory
 > > 
 > > Yes, FreeBSD is not a known system.  By modifying the
 > > script scripts/pslver.sh 
 > > (see below) I was able to run the linux port of PSL (see
 > > 
 > 
 > A yes, the linuxulator is always pretty handy (and fast).

But it can only run 32bit Linux. With a little tweaking, I could compile and run 
the PSL kernel on FreeBSD. This was just a quick and dirty solution; it needs a 
bit of work to integrate it into the build system.

 > Just wondering, you are probably aware of the classic
 > CMU Lisp implementations:
 > 
 > http://www.freshports.org/lang/cmucl/
 > http://www.freshports.org/lang/sbcl/
 > 
 > Is there a chance that REDUCE can work with them?

I'm know it can.  A long time ago, I ported Reduce to MacLisp, one of the 
precursors of Common Lisp.  The Standard Lisp defined as a basis for Reduce is 
not a Common Lisp, being somewhat older than the Common Lisp Standard - see the 
doc/misc directory for a description. Therefore, some interface layer is 
necessary. This is not very difficult, just tedious.

  Rainer