headers
Shiro Kawai | 3 Aug 2004 01:06
Favicon

Gauche-0.8.1

Gauche-0.8.1 is out.

This is a maintainance release.  I tried to fix as many reported 
bugs at this moment as possible, except several 'hard' ones.
If you have reported a problem and didn't get an ack from me
and don't see a fix in this version, let me know.
There's also a few minor feature enhancements.

Some of the important changes:

- Internal defines that are generated as a a result of macro
  are now handled.  Note that this issue is a quite nasty one,
  where even Scheme Macro Gurus debates now and then; the following
  posting summarizes the problem well:

  http://groups.google.com/groups?selm=87660ebzlb.fsf%40radish.petrofsky.org

  I took the 'chez scheme' approach described above, if I understand
  it correctly.

- Macroexpand works for autoloaded macros.

- www.cgi handles file uploads.

For other changes, take a look at the project webpage.

http://www.shiro.dreamhost.com/scheme/gauche/index.html

Enjoy.
(Continue reading)

Permalink | Reply | 5 comments |
headers
Shiro Kawai | 3 Aug 2004 01:11
Favicon

rpath problem (Re: tentative schedule of 0.8.1)

In 0.8.1, if you pass --with-rpath=no to configure, you can get
rid of rpath.  Note that Gauche needs gosh during the build process,
and it should be able to use just-built libgauche.so; if you choose
not to use rpath, you may need LD_LIBRARY_PATH to be set to look
at the build source directory.

Please let me know if this can be a solution of this problem or not.

--shiro

From: hipster <hipster <at> xs4all.nl>
Subject: Re: [Gauche-devel] tentative schedule of 0.8.1
Date: Wed, 21 Jul 2004 10:31:15 +0200

> The embedded rpath needs to be stripped from gosh & libgauche as well.
> 
> This will help doing proper Debian packaging, see
> 
>    http://lintian.debian.org/reports/mHatta_Shuzo.html
> 
> specifically the binary-or-shlib-defines-rpath and 
> sharedobject-in-library-directory-not-actually-a-shlib entries.

-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
Permalink | Reply | 0 comments |
headers
Shiro Kawai | 3 Aug 2004 01:12
Favicon

Re: tentative schedule of 0.8.1

Oops, I didn't do this.  For next release, maybe.

--shiro

From: Shiro Kawai <shiro <at> lava.net>
Subject: Re: [Gauche-devel] tentative schedule of 0.8.1
Date: Tue, 20 Jul 2004 21:42:01 -1000 (HST)

> From: Kimura Fuyuki <fuyuki <at> hadaly.org>
> Subject: Re: [Gauche-devel] tentative schedule of 0.8.1
> Date: Tue, 20 Jul 2004 20:32:33 +0900
> 
> > Perhaps you should put extra care in testing with gcc 3.4.1 while
> > release engineering cycle. It will soon be a mainstream compiler and
> > may bring us a bunch of annoying problems. (Of course this might just
> > be an unreasonable fear of new things. :)
> 
> Thanks for the advice.  I think I can try to compile with gcc 3.4.1.
> 
> --shiro

-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
Permalink | Reply | 1 comment |
headers
Kimura Fuyuki | 3 Aug 2004 12:23

Re: tentative schedule of 0.8.1

At Mon, 02 Aug 2004 13:12:06 -1000 (HST),
Shiro Kawai <shiro <at> lava.net> wrote:
> 
> Oops, I didn't do this.  For next release, maybe.

Don't worry, at least it was cleanly built on my box with gcc 3.4. :)

-- fuyuki

-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
Permalink | Reply | 0 comments |
headers
Mario Domgörgen | 3 Aug 2004 18:57
Picon
Picon

Re: Gauche-0.8.1

Hello!

Could it be that `while' is undocumented? Or shouldn't it there be at all?

Mario

--

-- 
Three things are certain:
Death, taxes, and lost data.
Guess which has occurred.

-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
Permalink | Reply | 4 comments |
headers
Shiro Kawai | 4 Aug 2004 13:45
Favicon

while (Re: Gauche-0.8.1)

From: Mario Domgörgen <kanaldrache <at> gmx.de>
Subject: Re: [Gauche-devel] Gauche-0.8.1
Date: Tue, 03 Aug 2004 18:57:47 +0200

> Could it be that `while' is undocumented? Or shouldn't it there be at all?

It's undocumented since I'm not sure we need it.  I think
once I thought I'd give it a try, and until now I've never
needed it.  (I guess it's partly because 'while' implies
side effects---if (while <expr> <body> ...) ever exits,
it means <expr> returns a different value at some point).

So I doubt it's worth to keep it in the bundled library.
What do you guys think?  If it is popular enough, I'll support
it 'officially'.

--shiro

-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
Permalink | Reply | 3 comments |
headers
Kogule, Ryo | 5 Aug 2004 09:33

Patches to 0.8.1

Hi,

libtool (or configure script) of gc seems to have EGREP problem
which is same as SED's.  A patch for this issue is:

--- ltmain.sh.orig      Wed Jul  9 20:45:48 2003
+++ ltmain.sh   Thu Aug  5 16:06:01 2004
 <at>  <at>  -51,6 +51,7  <at>  <at> 
  # [SK: configure generated by older autoconf doesn't create libtool
  # that defines SED.]
  : ${SED=sed}
+: ${EGREP=grep -E}
  # [SK: so as max_cmd_len]
  : ${max_cmd_len=-1}

Additionally a patch for slibcat-in-place target hasn't been
applied yet.  That is:

--- lib/Makefile.in.orig        Fri May 14 21:06:35 2004
+++ lib/Makefile.in     Thu Aug  5 16:07:01 2004
 <at>  <at>  -110,7 +110,7  <at>  <at> 
  # before installing gosh.
  slibcat-in-place :
         if test -f $(SLIB_DIR)/require.scm; then \
-         $(SHELL) ./mklibcat $(top_builddir)/bin/gosh ; \
+         $(SHELL) ./mklibcat $(top_builddir)/src/gosh ; \
         fi

  clean:
(Continue reading)

Permalink | Reply | 1 comment |
headers
Shiro Kawai | 5 Aug 2004 10:01
Favicon

Re: Patches to 0.8.1

Patch applied.  Thanks.

--shiro

From: "Kogule, Ryo" <kogule <at> opendarwin.org>
Subject: [Gauche-devel] Patches to 0.8.1
Date: Thu, 5 Aug 2004 16:33:36 +0900

> Hi,
> 
> libtool (or configure script) of gc seems to have EGREP problem
> which is same as SED's.  A patch for this issue is:
> 
> --- ltmain.sh.orig      Wed Jul  9 20:45:48 2003
> +++ ltmain.sh   Thu Aug  5 16:06:01 2004
>  <at>  <at>  -51,6 +51,7  <at>  <at> 
>   # [SK: configure generated by older autoconf doesn't create libtool
>   # that defines SED.]
>   : ${SED=sed}
> +: ${EGREP=grep -E}
>   # [SK: so as max_cmd_len]
>   : ${max_cmd_len=-1}
> 
> 
> Additionally a patch for slibcat-in-place target hasn't been
> applied yet.  That is:
> 
> --- lib/Makefile.in.orig        Fri May 14 21:06:35 2004
> +++ lib/Makefile.in     Thu Aug  5 16:07:01 2004
>  <at>  <at>  -110,7 +110,7  <at>  <at>
(Continue reading)

Permalink | Reply | 0 comments |
headers
Russell Senior | 8 Aug 2004 02:52

threads, mutex, condition variables and doing what I want ...


Folks

I have never done threads before and I am having a little trouble
finding the solution to what I want to do.  This is the master/slave
parallel processing problem I've been working on.  Basically, I have
slaves that are either already connected or might connect during a
"job".  I "submit" a job in the master via a listener, basically by
setting a variable:

  (define job #f)
  (define job-mutex (make-mutex))
  (define job-condition (make-condition-variable))

  (set! job (build-job-from-file "foo"))

The slaves that have already connected (and have spawned threads
talking to them) are waiting for a job to come along.  When it does, I
want these threads to wake up and feed the job to the slaves.  Also,
if slaves connect during the job, I want them to get the job sent
immediately.  

My talk-to-slave thunk looks roughly like:

   (define (talk-to-slave slave)
     (let ((output (socket-output-port slave))
           (input (socket-input-port slave :buffering #f)))
       (do () (#f)
         (display "OKAY\n" output)
         (flush output)
(Continue reading)

Permalink | Reply | 1 comment |
headers
Mario Domgörgen | 8 Aug 2004 14:20
Picon
Picon

Memory storage of unbound objects

Hallo list!

I'm working on a scheme library to a wm. The wm sends events with
arguments when something happen. It seems naturally to introduce a event
class who holds the arguments. The problem is that the arguments for
various events differs a lot. Some have 25 args, others only two... Now
my question: Would it be a problem if the class provides slots for every
events arguments (would be so around fifty) but does not initialize
them? Would that be a efficency leak?

Thanks in advance!

Mario

--

-- 
Serious error.
All shortcuts have disappeared.
Screen. Mind. Both are blank.

-------------------------------------------------------
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to announce. We are now OSTG- Open Source Technology
Group. Come see the changes on the new OSTG site. www.ostg.com
Permalink | Reply | 3 comments |

Gmane