headers
Jonathan Rees | 2 Jan 1995 23:53
Picon
Picon

(unknown)

   From: Brian D. Carlstrom <bdc <at> clark.lcs.mit.edu>
   Date: Fri, 23 Dec 94 12:25:33 -0500

   not really a bug, but shouldnt rts/arch.scm and vm/arch.scm be the same?
   logically, they are, but diff'ing them you wouldn't know it

Fixed in 0.39 (and later versions) by removal of rts/arch.scm.
Permalink | Reply | 0 comments |
headers
Jonathan Rees | 2 Jan 1995 23:54
Picon
Picon

build-usual-image requires . is in one's path.

   Date: Thu, 22 Dec 1994 14:47:43 -0500
   From: "John D. Ramsdell" <ramsdell <at> linus.mitre.org>

   build-usual-image requires . is in one's path.

This will be fixed in 0.39 and later versions.
Permalink | Reply | 0 comments |
headers
Aubrey Jaffer | 4 Jan 1995 06:08
Picon
Picon

ftp-swiss.ai.mit.edu:pub/scm/test.scm

ftp-swiss.ai.mit.edu:pub/scm/test.scm shows several problems with
scheme48 and scsh.  Scsh distinguishes case in symbols.  This is at
odds with its claim to be R4RS compliant.

Scsh takes more than 10 times as long to load as Scheme48.  Is this
expected?

The example of ,build in the documentation creates an image which
causes a segment violation on my system:
 Linux jacal 1.0 #3 Tue Jun 7 11:39:28 EDT 1994 i486

ftp-swiss.ai.mit.edu:pub/users/jaffer/slib.zip now supports Scheme 48 0.36.
Is it intended that scsh work with saved images?  What is the correct
method for invoking such an image?
Permalink | Reply | 4 comments |
headers
Jonathan Rees | 4 Jan 1995 14:29
Picon
Picon

ftp-swiss.ai.mit.edu:pub/scm/test.scm


Please send scsh bugs to scsh-bugs <at> martigny.ai.mit.edu.

Thanks
Jonathan
Permalink | Reply | 0 comments |
headers
Olin Shivers | 4 Jan 1995 21:25
Picon

Re: ftp-swiss.ai.mit.edu:pub/scm/test.scm

   ftp-swiss.ai.mit.edu:pub/scm/test.scm shows several problems with
   scheme48 and scsh.  Scsh distinguishes case in symbols.  This is at
   odds with its claim to be R4RS compliant.

This is documented at length. Please consult pp. 3-4 of the 0.3 manual.
S48's module system allows for modules to be specified relative to
a particular reader, so you can have files written in R4RS syntax,
and modules written in the scsh variant syntax, and they can happily
coexist.

   Scsh takes more than 10 times as long to load as Scheme48.  Is this
   expected?

Scsh has a large number of foreign-function linkages. The top-level is
very conservative about handling this, relinking all the ff links in the
heap at startup time. We can optimise this a lot, but have not had the
time to do so. The emphasis on scsh development to date has been on
design and implementation of new functionality, not optimisation. We
expect that to change as the important pieces fall into place. Remember,
scsh has been publically released for less than six months.

   The example of ,build in the documentation creates an image which
   causes a segment violation on my system:
    Linux jacal 1.0 #3 Tue Jun 7 11:39:28 EDT 1994 i486

I'm sorry, I'm not sure what you mean here. What documentation? What example?

We are very interested in supporting Linux, and if you are unable to build
scsh on your linux box, we'd love to hear about it.
(Continue reading)

Permalink | Reply | 2 comments |
headers
Brian D. Carlstrom | 5 Jan 1995 01:01
Picon
Picon

random package documentation


the random package (make-random) is documented in user-guide.txt but not
big-scheme.txt. seems like it should be in the later if its in the former

-bri
Permalink | Reply | 1 comment |
headers
kelsey | 5 Jan 1995 01:58
Picon

Re: random package documentation


   From: bdc <at> ai.mit.edu (Brian D. Carlstrom)
   Date: Wed, 4 Jan 95 19:01:49 EST

   the random package (make-random) is documented in user-guide.txt but not
   big-scheme.txt. seems like it should be in the later if its in the former

big-scheme.txt documents the structure BIG-SCHEME, which doesn't
export MAKE-RANDOM.
                                      -Richard
Permalink | Reply | 0 comments |
headers
Aubrey Jaffer | 5 Jan 1995 01:55
Picon
Picon

Re: ftp-swiss.ai.mit.edu:pub/scm/test.scm

   From: Olin Shivers <shivers <at> clark.lcs.mit.edu>
   Date: Wed, 4 Jan 95 15:25:57 -0500
   Cc: scheme-48-bugs <at> martigny.ai.mit.edu, bdc <at> ai.mit.edu
   Reply-To: shivers <at> mintaka.lcs.mit.edu

...
      The example of ,build in the documentation creates an image which
      causes a segment violation on my system:
       Linux jacal 1.0 #3 Tue Jun 7 11:39:28 EDT 1994 i486

   I'm sorry, I'm not sure what you mean here. What documentation? What example?

The documentation and example is from the man page for scheme48:
   ================================================================
       The ,dump and ,build commands put heap  images  in  files.
       The  -i  option  causes the initial heap image to be taken
       from file image.  The -a option  causes  a  string  to  be
       passed  as  the  argument  to an image generated using the
       ,build command.

            > ,build (lambda (a) (display a) (newline)) foo.image
            > ,exit
            $ scheme48 -i foo.image -a mumble
            mumble
            $
   ================================================================

Here is a script of trying this example on berne.ai.mit.edu:
   ================================================================
berne>uname -a
(Continue reading)

Permalink | Reply | 1 comment |
headers
Jonathan Rees | 5 Jan 1995 02:19
Picon
Picon

ftp-swiss.ai.mit.edu:pub/scm/test.scm


Thanks for pointing this out.  The man page frequently gets out of
date because we forget about its existence.  The user guide has better
documentation.  Both places should state that the startup procedure is
supposed to return an integer to be interpreted by the operating
system or whatever.  That non-integers on return cause core dumps is
an unfortunate bug.

Jonathan
Permalink | Reply | 0 comments |
headers
Joshua Peck Macdonald | 5 Jan 1995 12:49
Picon
Picon

porting to FreeBSD


in dynload.c, even though my system HAS_DLOPEN, the 
included file <dlfcn.h> does not exist on FreeBSD.
to correct this I included <nlist.h> and <link.h>
and tried again, the only problem with this is that
dlerror is not defined anywhere so I made a dummy
dlerror and continued.  the make then completed
successfully(there are leading tabs missing in the
makefile though, had to correct this, near line 200)
(oh, also failed on make install cause the directory
doesn't exist, might wanna try mkdir /usr/local/lib/scheme48(
before trying to install things there.)  It runs
I guess, except when trying the example in the 
manual page for building a new image, I get a core
dump, here's a transcript

axis.root-s48 % scheme48
Welcome to Scheme 48 0.36 (made by josh on Thu Jan  5 03:28:24 PST 1995).
Copyright (c) 1993, 1994 by Richard Kelsey and Jonathan Rees.
Please report bugs to scheme-48-bugs <at> martigny.ai.mit.edu.
Type ,? (comma question-mark) for help.
> ,build (lambda (a) (display a) (newline)) foo.image
Writing foo.image
> ,exit
axis.root-s48 % scheme48 -i foo.image -a mumble
(mumble)
Segmentation fault (core dumped)

uhhh so now I don't know what to do, is there a better
more informative test to tell me what's wrong here?
(Continue reading)

Permalink | Reply | 0 comments |

Gmane