Gambit-list Novembre 2010

gambit-list@iro.umontreal.ca

23 participants
22 discussions

A register is not initialized in gvm code?
by Oleg Parashchenko 25 Nov '10

25 Nov '10

Hello, I compiled to GVM the following test program: (declare (standard-bindings) (fixnum) (not safe)) (define mkadder (lambda (n) (lambda (x) (+ x n)))) (define add1 (mkadder 1)) (add1 1) In the produced GVM code (below, slightly edited for mail posting), I see that the code of the lambda uses a closure in the register 4, but nowhere in the code I see an assigment to this register. What do I miss? **** #<primitive | test2.o7|> = #line 2 "............" #1 0 entry-point 0 () ~#mkadder = '#<procedure ~#mkadder> #line 6 -1 = +0 +1 = '1 +0 = #3 -2 = . -3 = . -4 = . jump* 4 #2 #2 4 jump 4 ~#mkadder 1 #3 4 return-point ~#add1 = +1 #line 7 +1 = '1 +0 = -1 jump* 4 #4 #4 4 jump 0 ~#add1 1 **** #<procedure ~#mkadder> = #line 2 "................" #1 0 entry-point 1 () close -1 = (#2 +1) +1 = -1 jump 0 +0 #2 0 closure-entry-point 1 () +4 = +4(1) +1 = (##fx+ +1 +4) jump 0 +0 -- Oleg Parashchenko olpa@ http://uucode.com/ http://uucode.com/blog/ XML, TeX, Python, Mac, Chess

4 14

c++ example
by David Dreisigmeyer 24 Nov '10

24 Nov '10

Does anyone have a simple example of calling a C++ function from Gambit? Thanks, -Dave

2 1

C-FFI pointers in c function arguments
by Sid H 19 Nov '10

19 Nov '10

Hi, I'm just starting with the FFI. I'm trying to call a C function that updates a pointer argument. However, that value does not seem to change. My understanding is that all arguments passed to C are treated as immutable. So, if we need to return any value from C then, it cannot be via an argument only. Is that correct? Please see my example below. (c-declare #<<c-declare-end #include <stdio.h> int get_value(int* value) { if(NULL != value) { *value = 42; } else { printf("Invalid pointer"); } return 21; } c-declare-end ) (c-define-type int* (pointer int)) (define get-value (c-lambda (int*) int "get_value")) ;; Define a Wrapper in C (define %get-value (c-lambda () int #<<C-END int tmp; int retval = get_value(&tmp); /* TODO: Handle retval inside the C Wrapper code */ ___result = tmp; C-END )) ;; Can we update a pointer via a c function call arg? (define value #f) (let ((retval (get-value value))) (if (eq? retval 21) (begin (display value) ; Prints #f (newline)) (begin (display "Did not get 21 as return value") (newline)))) ; Now use the wrapper API instead (set! value (%get-value)) (display value) ; No surprise, prints 42 (newline) Thanks, Sid

2 1

Gambit Scheme Inside Out talk at ILC2010
by Marc Feeley 16 Nov '10

16 Nov '10

Hello. I gave a talk on Gambit at ILC2010 yesterday. It presents highlights of both the Gambit Scheme language and its implementation. You can find it at: http://www.iro.umontreal.ca/~gambit/Gambit-inside-out.pdf The .pdf is 24MB so it might be a longish download... Many thanks to Guillaume Cartier, James Long and Robert Lizee for providing some nice demos. For the record, some participants said that now that they know Gambit they see less of a reason to use Common Lisp (whatever that means!). Marc

3 3

Issues with define-syntax compilation
by Evan Hanson 15 Nov '10

15 Nov '10

I'm having issues compiling with the syntax-case extension. I have no issues using only define-syntax, but using define-syntax and introducing the -:s switch seems to cause some, but not all, of the macros I have defined using define-macro to be ignored. ... *** WARNING -- "maybe" is not defined, *** referenced in: ("/home/evanhanson/proj/gm/gm.c") *** WARNING -- "when" is not defined, *** referenced in: ("/home/evanhanson/proj/gm/gm.c") ... Is it the case that syntax-case can't be used alongside lisp-style macros? Redefining the problem macros with define-syntax didn't help, though there remained other definitions using define-macro (I didn't try removing them all...). Thanks, Evan

2 2

terminal behaviour in Gambit
by Adrien Piérard 10 Nov '10

10 Nov '10

2010/11/7 Michele Zaffalon <michele.zaffalon(a)gmail.com>: > Hi all, Hi Michele > I am using Gambit from the terminal: is it it possible to change the > terminal behaviour so that if you type a partial word and hit up > arrow, the terminal tries to complete it with a command from the past > command history that matches the partial word, not just with the last > command? This behaviour is, as far as I know, not very common. Most pervasive completion systems are based on the readline library, which Gambit reimplemented if I remember correctly. So, <up> will always give you the previous command in history (though some of us believe that the previous S-exp would be better). In your case, what you want is to use <tab>, form completion. And this works well. I do not know however if this can be restrained to symbols that are in the history, or if the suggestion order may consider history first. Cheers, P! -- Mail resent to the list, for I initially sent it with a wrong account…

3 3

terminal behaviour in Gambit
by Michele Zaffalon 07 Nov '10

07 Nov '10

Hi all, I am using Gambit from the terminal: is it it possible to change the terminal behaviour so that if you type a partial word and hit up arrow, the terminal tries to complete it with a command from the past command history that matches the partial word, not just with the last command? I sometimes use Ipython and find this feature very practical. Thanks, michele

1 0

Linking
by Brady McCary 04 Nov '10

04 Nov '10

gambit-list, I had composed a long-winded email about some trouble I was having linking against ATLAS http://math-atlas.sourceforge.net/ but I figured it out. Because I am using external C++ dependencies I compiled gambit using C++. It turns out that I want to use (c-declare #<<c-declare-end extern "C" { #include <cblas.h> }; c-declare-end ) and not (c-declare "#include <cblas.h>") This is obvious in retrospect. However I wasted about 90 minutes on this. Hopefully someone will search for gambit, linker, link, linking, atlas, blas, lapack, undefined reference and find this message here. Brady

2 1

Compiling: Invalid characters in file name
by Taylor Venable 03 Nov '10

03 Nov '10

Hi, I'm curious why when compiling to a dynamically loadable object file (.o1, .o2, ...) I can use a filename such as 1138.scm, but when compiling directly to a executable program that is not allowed: 8333 [ taylor @ zareason ] : ~ > gsc -exe 1138.scm *** ERROR -- Invalid characters in file name (must be a symbol with no "#") Thanks, -- Taylor C. Venable http://metasyntax.net/

1 0

Re: [gambit-list] Gambit on Android
by sean chapel 02 Nov '10

02 Nov '10

I believe that the android NDK defines ANDROID so it should be fairly easy to make a patch. TCP/IP networking should work however I haven't tested it. The only thing that I had to comment out that would make much of a difference was in open_pseudo_terminal_slave due to * I_PUSH *not being defined. The only thing I might need help with is the makefile. I was using the latest source snapshot from here http://www.iro.umontreal.ca/~gambit/download/gambit/v4.6/source/gambc-v4_6_… I can check it out from git and create a patch that way if you prefer. Let me know. -Sean Chapel

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Gambit-list Novembre 2010