Re: Garnet, SUBSEQ

[haible (Bruno Haible)]

Douglas Ambrisko <ambrisko@kpc.com> writes:

> | *** - UNIX error 9 (EBADF): Bad file number 
> I just grabbed the clisp versin of garnet and built it.  I hit the same
> problem with the "saveinitmem" on a DEC alpha box .  The problem 
> is that garnet assumes you do a load of all compiled modules including 
> opal/defs which calls "initialize-x11-values".

Ouch. Open file descriptors don't survive a "saveinitmem", of course.

> So a work-around that 
> I'm doing now is to load the "saveinitmem" version then loading opal/defs. 

(opal::initialize-x11-values (opal::get-full-display-name))
should work, too.

Bela Pecsek <P2043742@csdvax.csd.unsw.EDU.AU> writes:

> *** - SUBSEQ: :END should be an integer >=0, not NIL

As I noted, this will be fixed in the next release. Meanwhile, you can
also use the following patch:

(in-package "LISP")
(proclaim '(notinline subseq))
(setf (symbol-function 'old-subseq) (symbol-function 'subseq))
(defun subseq (seq start &optional (end nil))
  (if (null end)
    (old-subseq seq start)
    (old-subseq seq start end)
) )

> I was happier than ever and then the shock came.
> *** - SUBSEQ: :END should be an integer >=0, not NIL
> I really don't know where this error comes from but hopefully somebody
> out there knows about it.

You have to  (proclaim '(notinline subseq))  before compiling Garnet.


                    Bruno Haible
                    haible@ma2s2.mathematik.uni-karlsruhe.de

!! To unsubscribe from the clisp-list mailing list, send mail to     !!
!!           listserv@ma2s2.mathematik.uni-karlsruhe.de              !!
!! including the two words "unsubscribe clisp-list" as message body. !!