gmane.lisp.scheme.stklos

Re: segfault reading ratio

Erick Gallesio — 2008-08-10T17:56:39

Thanks for signaling it. I'll try to fix it asap.

segfault reading ratio

Bill Schottstaedt — 2008-08-09T20:00:31

* STklos version 0.98 * Copyright (C) 1999-2008 Erick Gallesio - I3S-CNRS/ESSI essi.fr> * * [Linux-2.6.25-14.fc9.i686-i686/pthread] stklos> 0/1234000000 Received a SIGSGV signal. Try to augment stack size (--stack-size option). If the problem persists, send a mail to stklos.net> But, stklos> 0/123400000 0

Re: typo in the link to GTK+ documentation

Erick Gallesio — 2008-06-27T10:12:32

Hello Marijn, This is corrected now. Thanks for signaling it.

typo in the link to GTK+ documentation

Marijn Schouten (hkBst — 2008-06-26T10:42:50

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, http://www.stklos.org/ScmPkg/Doc/stklos-gtk-base/stklos-gtk-base.html contains a typo in the link to GTK+ documentation. It should link to http://www.gtk.org/documentation.html but instead links to that plus a final slash. Marijn - -- Marijn Schouten (hkBst), Gentoo Lisp project, Gentoo ML , #gentoo-{lisp,ml} on FreeNode -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkhjcqoACgkQp/VmCx0OL2xKpACeK63NBF8fNWfs+q5xVxcSJG7r 5zQAoJVdiROQUlMatcOPxTVt2+DhS7Qo =3ZiR -----END PGP SIGNATURE-----

Re: [Announce] STklos-0.98 release

Marijn Schouten (hkBst — 2008-05-18T21:19:57

Re: [Announce] STklos-0.98 release

Erick Gallesio — 2008-05-18T20:27:29

> Erick Gallesio wrote: > | This is an announce for the 0.98 release of STklos. > | > | Changes in this release: > | > | * Replaced the C/invoke FFI library by libffi: Now GTk2 ScmPkg > | packages work on MacOs > > It looks like it isn't possible to use an external libffi. Is that correct? I would really > prefer not to use any bundled libs. > > Hello Marijn, That is true for this release, and I'll change that for the next one, if needed . The change was mainly due to permit the usage of GTK2 ScmPkg on MacOS. However, the situation on libffi is not very clear. The version I use is one which was released on April 3, 2008. It is numbered 3.0.5. On Archlinux, the current version of libffi is 4.2.0 and is not the autonomous one but the one which is extracted from GCC sources. Gentoo seems to have a version 3.4.x which is unknown on libffi site. This is the kind of situation that I generally don't like, since it cans be the source of subtle bugs. Anyway, I will try to use the system lib

Re: [Announce] STklos-0.98 release

Marijn Schouten (hkBst — 2008-05-18T19:26:52

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Erick Gallesio wrote: | This is an announce for the 0.98 release of STklos. Hi Erick, it seems that the config option for ffi interprets --enable-ffi the same as --disable-ffi in both cases disabling ffi. Marijn - -- Marijn Schouten (hkBst), Gentoo Lisp project, Gentoo ML , #gentoo-{lisp,ml} on FreeNode -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkgwgvwACgkQp/VmCx0OL2xpoQCbBDLay0q4l3KOBV2Pq23ota+T frYAoMS9LYHhaH3GRavoYA58/VSiUzB6 =fv3c -----END PGP SIGNATURE-----

Re: [Announce] STklos-0.98 release

Marijn Schouten (hkBst — 2008-05-18T19:44:40

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Erick Gallesio wrote: | This is an announce for the 0.98 release of STklos. | | Changes in this release: | | * Replaced the C/invoke FFI library by libffi: Now GTk2 ScmPkg | packages work on MacOs It looks like it isn't possible to use an external libffi. Is that correct? I would really prefer not to use any bundled libs. Marijn - -- Marijn Schouten (hkBst), Gentoo Lisp project, Gentoo ML , #gentoo-{lisp,ml} on FreeNode -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkgwhygACgkQp/VmCx0OL2wG2wCgj8ni97yN2Uzk7ZjyzK7t7sqi 8o8Ani9nH697KbuKMmKRzyEGXiRrjX92 =qrVF -----END PGP SIGNATURE-----

[Announce] STklos-0.98 release

Erick Gallesio — 2008-05-18T18:21:17

This is an announce for the 0.98 release of STklos. STklos is a free Scheme System mostly conform to R5RS. The aim of this implementation is to be fast as well as light. The implementation is based on an ad-hoc Virtual Machine. STklos can also be compiled as a library, so that one can easily embeds it in an application. The salient points of STklos are: * efficient and powerful object system based on CLOS providing - Multiple Inheritance, - Generic Functions, - Multi-methods - an efficient MOP (Meta Object Protocol) * a simple to use module system * implements the full tower of numbers defined in R5RS * easy connection to the GTK+ toolkit * Perl compatible regular expressions thanks to the PCRE package. * tail recursive calls are properly implemented Changes in this release: * Replaced the C/invoke FFI library by libffi: Now GTk2 ScmPkg packages work on MacOs * Addded option --build-sync-file to ease personal ScmPk

Re: Extending STklos: how to deal with multible returnvalues?

Erick Gallesio — 2008-05-01T16:48:52

Hello Heinrich, Excuse me for being so long to answer you. I finally came to the following code for dealing with multiple values. It is different from yours but I expect that it is a little bit more general. Of course, if it doesn't fit your needs, I can integrate the code you provided in your mail. I propose to add the function to transform values to vectors. SCM STk_values2vector(SCM obj, SCM vect) { vm_thread_t *vm = STk_get_current_vm(); SCM src, retval; int len = vm->valc; if (vect) { /* User has provided a vector for storing result */ if (!VECTORP(vect) || VECTOR_SIZE(vect) != len) STk_error("bad vector ~S", vect); retval = vect; } else { /* Allocate a new vector for result */ retval = STk_makevect(len, STk_void); } vm->val = obj; vm->valc = 1; if (len > 1) { /* multiple values */ if (len <= MAX_VALS) { vm->vals[0] = obj;

Re: Extending STklos: how to deal with multible return values?

Heinrich Wolf — 2008-04-22T21:21:48

The implementation of `call-with-values' could certainly be used but to me this seems a little clumsy and might introduce a small performance penalty. When I am a C-programmer, then a Scheme vector seems the most convenient datum to deal with multible Scheme values. This given, my approach would be to pass as consumer to `call-with-values' a procedure that creates a vector and writes its arguments into this vector. I wrote a C procedure that differs from `call-with-values' essentially by by-passing that trivial consumer. Here is what I did: static SCM apply_index_mapping(SCM procedure, SCM args) { vm_thread_t *vm = STk_get_current_vm(); SCM retval; int n; vm->val = STk_C_apply(procedure, -VECTOR_SIZE(args), VECTOR_DATA(args)); n = vm->valc; retval = STk_makevect(n, STk_void); vm->valc = 1; switch (n) { case 0: break; case 1: *VECTOR_DATA(retval) = vm->val; break; default: { SCM *pd, *sentinel, *pdd; if (n <= MAX_VALS) { vm->vals[0] = vm->va

Two new packages

Erick Gallesio — 2008-04-18T15:26:39

Hi, I have added two packages for STklos which permits to make interfaces with GTK2. One is a simple wrapper to the GTK api, the other one permits to dynamically load XML files produced by the Glade interface builder. Files are packaged as ScmPkg and to install them you just have to do $ stklos-pkg -s $ stklos-pkg -i stklos-gtk-base $ stklos-pkg -i stklos-gtk-glade It seems to work on Linux. On MacOs, I had some problems that I cannot understand,but I'm not sure of my GTK installation. Any help on this platform would be greatly appreciated. NOTE: It will only work with the development release of STklos which is available at http://www.stklos.org/hgweb.cgi/archive/tip.tar.gz

STklos Wiki.

Erick Gallesio — 2008-04-18T15:32:20

Hi, There is now some more documentation (mainly for ScmPkg packages). It is in fact a simple Wiki which is available at http://www.stklos.org/Wiki

Re: Extending STklos: how to deal with multible returnvalues?

Erick Gallesio — 2008-03-31T07:58:34

> I recently implemented SRFI-25--- multidimensional arrays--- as an > extension to STklos. (Has passed test suite, will be released under > GPL as I find time.) That is great. I *really* hope that you will find time to share your code. Of course, I would be really pleased to help you if you want to package all your code for being released. BTW, Is your implementation ScmPkg based, or is it just a dynamically loadable file? > > The `share-array' procedure takes as third argument a procedure, which > may return more than one value. > > To implement this I took a good look at Erick's implementation of > `call-with-values' and soon arrived at a solution that seems to > work--- but: I had to include `vm.h', as my code begins with: > > vm_thread_t *vm = STk_get_current_vm(); > > To me it seems an extension of STklos should not access things that are--- > I suppose--- STklos internals. Did I miss something? > Yes you are absolutely right, you should not see this details. I must admit tha

Extending STklos: how to deal with multible return values?

Heinrich Wolf — 2008-03-30T23:31:45

I recently implemented SRFI-25--- multidimensional arrays--- as an extension to STklos. (Has passed test suite, will be released under GPL as I find time.) The `share-array' procedure takes as third argument a procedure, which may return more than one value. To implement this I took a good look at Erick's implementation of `call-with-values' and soon arrived at a solution that seems to work--- but: I had to include `vm.h', as my code begins with: vm_thread_t *vm = STk_get_current_vm(); To me it seems an extension of STklos should not access things that are--- I suppose--- STklos internals. Did I miss something?

Re: [Announce] STklos-0.97 release: call/cc & dynamic-wind test fails on amd64

Marijn Schouten (hkBst — 2007-12-16T12:32:52

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Erick, the call/cc & dynamic-wind test fails for me on amd64: (cd tests; make test) make[1]: Entering directory `/var/tmp/portage/dev-scheme/stklos-0.97/work/stklos-0.97/tests' ==== Testing Numbers ... passed ==== Testing Keywords ... passed ==== Testing call/cc & dynamic-wind ... failed Errors found in this section: test coroutines on (with-output-to-string start) expected "abcdabcdabcdabcdabcdabcdabcdabcdabcdabcd" but got #(fail) ==== Testing String Ports (SRFI 6) ... passed ==== Testing File Ports ... passed ==== Testing Threads ... passed ==== Testing SRFIs ... passed ==== Testing MD5 checksum ... passed ==== Testing Base64 functions ... passed

[Announce] STklos-0.97 release

Erick Gallesio — 2007-12-16T12:11:30

This is an announce for the 0.97 release of STklos. STklos is a free Scheme System mostly conform to R5RS. The aim of this implementation is to be fast as well as light. The implementation is based on an ad-hoc Virtual Machine. STklos can also be compiled as a library, so that one can easily embeds it in an application. The salient points of STklos are: * efficient and powerful object system based on CLOS providing - Multiple Inheritance, - Generic Functions, - Multi-methods - an efficient MOP (Meta Object Protocol) * a simple to use module system * implements the full tower of numbers defined in R5RS * easy connection to the GTK+ toolkit * Perl compatible regular expressions thanks to the PCRE package. * tail recursive calls are properly implemented Changes in this release: * Fixed support for recent versions of SLIB * Added boxes: they were used by the system but unavailable to the user. * Improved the stklos-pkg c

Re: DESTDIR patch and another

Erick Gallesio — 2007-11-29T21:48:23

Hello Marijn, patches have been applied. Thanks.

DESTDIR patch and another

Marijn Schouten (hkBst — 2007-11-28T18:00:31

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Eric, please consider the following patches. I tested them against 0.96. These first fixes some issue where files are installed to non-optimal directories. Index: stklos-0.95/extensions/Makefile.am =================================================================== - --- stklos-0.95.orig/extensions/Makefile.am +++ stklos-0.95/extensions/Makefile.am < at >< at > -11,7 +11,7 < at >< at > CC = < at >CC< at > CFLAGS = < at >CFLAGS< at > < at >SH_COMP_FLAGS< at > < at >GTK_CONFIG_CFLAGS< at > < at >GCINC< at > $(INCLUDES) - -shobjectdir = $(prefix)/lib/< at >PACKAGE< at >/< at >VERSION< at > +shobjectdir = $(libdir)/< at >PACKAGE< at >/< at >VERSION< at > shobject_SCRIPTS = ldap.< at >SH_SUFFIX< at > #====================================================================== < at >< at > -20,7 +20,7 < at >< at > shobject_SCRIPTS = ldap.< at >SH_SUFFIX< at > scheme_SRCS = ldap.stk scheme_OBJS = ldap.ostk - -schemedir = $(prefix)/share/< at >PACKAGE< at >/< at >VERSION< at > +schemedir = $(datadir)/< at >PACKAGE< at >/< at >VERSION< at > scheme_DATA = $(scheme_SRCS) $(scheme_OBJS) The next one fixes the DESTDIR issue I m

Re: OSX 10.5 Leopard

Rembrandt Wolpert — 2007-11-03T19:22:48

_______________________________________________ STklos mailing list STklos< at >email.essi.fr http://email.essi.fr/mailman/listinfo/stklos

Re: OSX 10.5 Leopard

Erick Gallesio — 2007-11-03T10:41:11

Hi Rembrandt, Thanks a lot for the information. Perhaps it is time to make dmg files for STklos. If I understand what you said, a dmg file made on 10.4 should run on 10.5. However, I have no idea of the way to make a dmg file. If someone has some idea or link on the subject, it would be great.