arder?:
- string-foldcase
- all string/char comparison with more than two parameters
- vector-map and vector-for-each on different length vectors
- digit-value
Thanks for any help, greetings!
Peter
;; simple aliasing "implementation" of r7rs bytevectors
;; just map everything to vec
d as the first argument to
bytevector-length, is not the correct type. (ellipsis by me)
It seems that there is a bit of confusion in http-get as to whether the
body is a bytevector or a string :-/
This nonwithstanding, great to see a move towards R7RS!
Greetings,
Peter
__
(get-trimmed
string
(min the-start the-end)
the-end))
This moves the computation of the-start and the-end above the actual
call to substring, and uses the minimum of start and end (to prevent
start being further along than end).
Would anyone be kind enough to me
As a follow-up, here is a patch including some basic tests (in
tests/runtime/test-string.scm).
Greetings, Peter
diff --git a/src/runtime/ustring.scm b/src/runtime/ustring.scm
index 2cada76b3..adaf7e530 100644
--- a/src/runtime/ustring.scm
+++ b/src/runtime/ustring.scm
@@ -1962,23 +1962,25 @@ USA
ive it a go and send a patch. Is it in fact correct to make tcp
sockets binary ports?
Thanks for any assistance!
Greetings,
Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
redundant-decls -Wextra
-Wno-sign-compare -Wno-unused-parameter -Wold-style-definition
-Wunreachable-code -Wwrite-strings -g -O2 -MT x11-shim.lo -MD -MP -MF
.deps/x11-shim.Tpo -c x11-shim.c -o x11-shim.o >/dev/null 2>&1
mv -f .deps/x11-shim.Tpo .deps/x11-shim.Plo
/bin/sh ./libtool --tag=CC --mode=link gcc -Wall -Wundef -Wpointer-arith
-Winline -Wstrict-prototypes -Wnested-externs -Wredundant-decls -Wextra
-Wno-sign-compare -Wno-unused-parameter -Wold-style-definition
-Wunreachable-code -Wwrite-strings -g -O2 -module -avoid-version -shared -o
x11-shim.la -rpath /usr/local/lib/mit-scheme-x86-64/ x11base.lo x11color.lo
x11graph.lo x11term.lo x11-shim.lo -lX11
libtool: link: gcc -shared -fPIC -DPIC .libs/x11base.o .libs/x11color.o
.libs/x11graph.o .libs/x11term.o .libs/x11-shim.o -lX11 -g -O2 -Wl,-soname
-Wl,x11-shim.so -o .libs/x11-shim.so
libtool: link: ( cd ".libs" && rm -f "x11-shim.la" && ln -s "../x11-shim.la"
"x11-shim.la" )
make x11-const
make[2]: Entering directory '/tmp/mit-scheme-10.1.4/src/x11'
make[2]: 'x11-const' is up to date.
make[2]: Leaving directory '/tmp/mit-scheme-10.1.4/src/x11'
echo "done" > stamp-x11-const
./x11-const
echo '(sf "x11-const")' | mit-scheme-x86-64 --batch-mode
;Generating SCode for file: "x11-const.scm" => "x11-const.bin"... done
./compile.sh
;Loading "compile.scm"...
;The object #t, passed as an argument to merge-pathnames, is not an object
satisfying #[compiled-procedure 13 ("pathnm" #x9) #x1a #x9a8f0a].
;To continue, call RESTART with an option number:
; (RESTART 1) => Return to read-eval-print level 1.
2 error>
End of input stream reached.make[1]: *** [Makefile:1326: stamp-scheme] Error 1
make[1]: Leaving directory '/tmp/mit-scheme-10.1.4/src/x11'
make: *** [Makefile:486: compile-x11] Error 2
Note: I do have an incompatible old mit-scheme 9 git version installed
locally.
Taylor found this commit, which should probably be merged into
10.1.5:
commit f9240424db225dafffa5fb6078d371ce1cd0bad8
Author: Chris Hanson
Date: Thu Nov 15 23:03:45 2018 -0800
Change plugins to use the already-compiled toolchain.
Thanks!
Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
uild the default plugins.
I think the problem is that building the default plugins does not use
the just newly built scheme from the binary release tarball, but instead
finds my system installed one. Thus I suggested this might be a good
thing to fix in one of the next
As an addendum:
The tarball I got from the homepage
(https://www.gnu.org/software/mit-scheme/):
http://ftp.gnu.org/gnu/mit-scheme/stable.pkg/10.1.4/
My local Scheme says:
Image saved on Wednesday November 15, 2017 at 11:01:14 PM
Release 9.2.1 || Microcode 15.3 || Runtime 15.7 || SF 4.41 || LIA
It seems I actually got the source tarball, not the prebuilt binary one,
so that one actually depended on a working installed local binary :-/
I'm sorry about the confusion, I just now tried the x86_64 binary
tarball, that works fine without any problems.
Sorry again,
Greetings,
e called
on it.
How do I close down that network connection? I've tried (apropos
'close), but haven't found anything that looks useful here.
Thanks for any help with this,
Greetings,
Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-dev
> close-binary-port
> close-binary-input-port
> close-binary-output-port
Thanks, these work. However, should not the "normal" procedures do the
same things these ones do? Is that a bug, or is that intentional?
Thanks for the hel
work, only close-output-port does
not (which is probably because the corresponding *input* port is still
open).
Sorry for the noise, everything seems to be working fine ;)
Thanks, Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://list
or making MIT-Scheme such a great system.
Peter#| -*-Scheme-*-
FFI for kyotocabinet (database)
|#
(declare (usual-integrations))
(load-option 'format)
(load-option 'ffi)
(C-include "kyotocabinet")
(define KCOREADER (C-enum "KCOREADER"))
(define KCOWRITER (C-enum
tk interface, for example, implements options
> --without-gtk and --with-gtk, and auto-selection via pkg-config.
If I want it to be automatically built and available,
INSTALLED_SUBDIRS is the only option, right? I've managed to avoid
autotools so far, seems now is the time to learn ;)
Thanks
file environment)) load-files)
(for-each (lambda (file) (compile-file file '() environment)) load-files)))
This loads all your files into the same environment that is then used
to compile your files.
Thanks to Taylor Campbell for explaining to me how this works.
Greetings, Peter
pg
I get bounces from naver.com for netscou...@naver.com in Korean. Google
translate indicates that this e-mail has blocked mail from the list
address. Would it be possible to remove the recipient from the list, in
order to avoid the bounces?
Greetings,
Peter
--- Begin Message ---
NAVER - http
f 2 hours ago. They do not
completely implement everything, but are an improvement over the
existing swank.scm (which does not work at all for me).
Please don't hesitate to tell me if they still are not in an acceptable
format (git-wise or code-wise), so I can further improve them.
Greetings,
h buffer)))
(if (= start end)
buffer
(let ((read-chars (read-substring! buffer start end in)))
(loop
(+ start read-chars)
end))
Would anyone apply this to runtime/swank.scm?
Thank you :)
Peter
_
!
Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
0 to 10.0?
There is a primitive C-POKE-DOUBLE, but I haven't yet found a way to
use this correctly :-/
I'd appreciate any help!
Greetings,
Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
for
struct parameters or return values. Is there any easy way to fix this?
Unfortunately quite a lot of C APIs use structs as parameters or return
values :(
Thanks for any help!
Peter
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://lis
The files cdecls.scm, ctypes.scm and syntax.scm have an incorrect
package comment ;;; package: (ffi syntax). According to ffi.pkg, they
are loaded into (ffi). This makes swank report an error when opening and
editing these files.
---
src/ffi/cdecls.scm | 2 +-
src/ffi/ctypes.scm | 2 +-
src/ffi/sy
If a name is actually a typedef for a struct, as in
typedef struct { int x; } a;
then sizeof(struct a) does not compile (because a does not strictly name
a struct).
This patch removes the "struct" from the declaration of the struct in
grovel_struct_ and from the sizeof() in the -const.c file
genera
struct? Then it should be possible to
support structs as return and parameter types. Are aliens
garbage-collected automatically? Can I just malloc some memory for an
alien, then cons_alien it, then expect it to be garbage-collected later?
Thanks for helping me with my questions here!
Peter
This patch adds support for struct parameter and return types.
Functions that return a struct take an additional first alien
parameter (like functions returning pointers), which must point to a
memory area large enough to hold the struct data. This can for example
be allocated by MALLOC inside Sche
f there are low hanging fruits regarding MIT-SCHEME hacking, I'd
happy in trying to help.
Thanks in advance.
Regards,
Peter.
___
MIT-Scheme-devel mailing list
MIT-Scheme-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
, 16 Feb 2015 11:45:04 +0100
>From: Peter BARABAS
>
>I couldn't find a way to send e-mails however, even after digging
> through
>the sources. Am I right in supposing that the functionality isn't there
>yet? Are there plans to implement it? (I think one
ext:*ARGV*
don't work. I get the error "Unbound variable" (when I type "scheme
--load myprog.scm").
Can MIT/GNU Scheme compile at all? With "--load"?
Greetings
Peter
___
MIT-Scheme-devel mailing list
MIT-Schem
Am 15.05.19 um 14:22 schrieb Aaron S. Hawley:
On Wed, May 15, 2019 at 7:14 AM Peter Wiehe wrote:
How do I access the commandline arguments? argv, *ARGV* and ext:*ARGV*
don't work. I get the error "Unbound variable" (when I type "scheme
--load myprog.scm").
The procedu
e tutorial.
Kind regards
Peter
Chris Hanson schrieb am Do., 16. Mai 2019 00:04:
> Please read the manual before asking this kind of question. Both your
> question about the command line, and the other about compiling files, are
> answered in the User's Manual.
>
> The advice to use
30 matches
Mail list logo