[Gcl-devel] Re: gcl-2.4.3

[Camm Maguire]

Greetings!

1) In CVS, I've just implemented a heuristic to shutoff readline
        a) if stdin is not a tty
        b) if the TERM environment variable is set to "dumb"

2) THe COMMON-LISP package is our work in progress toward making an
   ansi compliant lisp package.  The LISP package is the traditional
   one.  You should only get a (populated) common-lisp if you use CVS
   and configure with --enable-ansi.

3) It does seem that certain new features not strictly necessary for
   the building of maxima have found there way into 2.4.3.  I hope to
   do a 2.4.4 to remove the -I. from cmpmain.lsp, and at this time
   might try to clean things up briefly.  There is quite a bit in CVS,
   though, which is not in 2.4.3 -- full bfd support, support for many
   more platforms, and preliminary ansi lisp image support.

Take care,

Michael Bogomolny <address@hidden> writes:

> Camm,
> 
> Matt Kaufmann (the co-author of ACL2 along with J Moore) found a simple fix
> to the extra prompts, and to the fact that 2 successive ctrl-d's would
> cause a fatal error instead of exiting you from gcl.
> 
> The installation of gcl created the following script to launch gcl:
> 
> #!/bin/sh
> exec /p/lib/gcl-2.5.0/unixport/saved_gcl \
>    -dir /p/lib/gcl-2.5.0/unixport/ \
>    -libdir /p/lib/gcl-2.5.0/ \
>    -eval '(setq si::*allow-gzipped-file* t)' \
>    -eval '(setq si::*tk-library* "/usr/local/lib/tk8.3")' \
>    -eval '(si::init-readline)' \
>      "$@"
> # other options: -load /tmp/foo.o -load jo.lsp -eval "(joe 3)"
> 
> Matt noticed that if you remove the line
>    -eval '(si::init-readline)' \
> from the script, the behavior is as we expect it to be.
> 
> That pretty much solves all our issues (on the Debian side! Still have 
> serious 
> problems on the Solaris machines . . .) except for still wanting to know 
> 1) what the differences are between the LISP and COMMON-LISP packages
> 2) if this is really version 2.5.0 instead of 2.4.3 
>    (are there important differences?)
> 
> thanks in advance,
> bogo
> 
> 
> On Jun 20, 2002 12:38, Michael Bogomolny wrote:
> | Camm,
> | 
> | thanks for your help. The C_INCLUDE_PATH job did the trick. I now have
> | gcl-2.4.3 compiled and installed on some Debian Linux machines.
> | 
> | Three questions:
> | 
> | 1) What changes were made to the packages? The theorem prover ACL2 needs to
> | look at packages because it "disables" functions with side effects and
> | makes sure to complete functions on undefined domains. We can certainly do
> | with packages what we've always done, but are not sure why both "LISP" and
> | "COMMON-LISP" packages exists in gcl nor how they differ.
> | 
> | For example:
> | >(equal 'lisp::car 'common-lisp::car)
> | 
> | NIL
> | 
> | 2) everything other than the name of the tarball seems to indicate that
> | this is release 2.5.0. Is there a difference?
> | 
> | 3) gcl now seems to insert a prompt after every carriage return, whether it
> | is within the middle of a statement or not.
> | 
> | Example:
> | svartalfheim.cs.utexas.edu% /p/bin/gcl-2.4.3 
> | GCL (GNU Common Lisp)  Version(2.5.0) Wed Jun 19 15:41:50 CDT 2002
> | Licensed under GNU Library General Public License
> | Contains Enhancements by W. Schelter
> | 
> | >(+ 2 3)
> | 
> | 5
> | 
> | >(defun double (x)
> | >(* 2 x))
> | 
> | DOUBLE
> | 
> | that extra caret is causing some problems for us. Is that a new "feature"?
> | If so, is there a way to disable it?
> | 
> | thanks,
> | bogo
> 
> 

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