Février 2009 - Gambit-list - mailman.iro.umontreal.ca

gcc char* warnings
by lowly coder 18 Fév '09

18 Fév '09

i'm getting a bunch of warnings of the form: warning: deprecated conversion from string constant to 'char*' is this something I should: 1) fix via a gcc option to not give me these warnings or 2) fix something I'm doing wrong with c-ffi that's resulting in these warnings

3 2

Re: [gambit-list] Gambit archeology
by Ken Dickey 18 Fév '09

18 Fév '09

>> Whee, thanks. I wonder if I can get these to run on my 4mb Atari ST :D >Ask Ken Dickey... the first Gambit user... he ported Gambit to his Amiga almost 20 years ago today... Gambit ran just fine in a 1 MB Amiga (once the library changes were made), so it should do well with four times the memory! I probably should refrain from repeating Perlis' quip about some things being an improvement over their successors. 8^}. [NB: with reference to the Amiga which booted from a 256KB EPROM with a real-time kernel and window system -- NOT to Gambit's newer and improving releases!]. Cheers, -KenD

3 2

Re: [gambit-list] Gambit archeology
by Marc Feeley 17 Fév '09

17 Fév '09

On 16-Feb-09, at 10:39 PM, Oisín Mac Fhearaí wrote: > 2009/2/17 Marc Feeley <feeley(a)iro.umontreal.ca>: >> For those interested in studying Gambit archeology, I have added some >> really old versions of Gambit on the distribution page. There is the >> original SIS 0.1 compiler (the Simple Implementation of Scheme in >> Prolog) which generated m68k code and was where it all started. >> >> Read the sources if you want to see some real cool m68k code >> generation hacks! > > Whee, thanks. I wonder if I can get these to run on my 4mb Atari ST :D Ask Ken Dickey... the first Gambit user... he ported Gambit to his Amiga almost 20 years ago today... Marc

1 0

Gambit archeology
by Marc Feeley 17 Fév '09

17 Fév '09

For those interested in studying Gambit archeology, I have added some really old versions of Gambit on the distribution page. There is the original SIS 0.1 compiler (the Simple Implementation of Scheme in Prolog) which generated m68k code and was where it all started. Read the sources if you want to see some real cool m68k code generation hacks! Have fun! Marc

1 0

more fun with c-ffi
by lowly coder 17 Fév '09

17 Fév '09

looking at the glut ffi at the dumping grounds, we have lines like: (define GL_NORMAL_ARRAY #x8075) (define GL_COLOR_ARRAY #x8076) (define GL_INDEX_ARRAY #x8077) (define GL_TEXTURE_COORD_ARRAY #x8078) this is really cool because ... now, in scheme land, I also know the values of these defines suppose however, I didn't care about the values of these globals in scheme land ... and suppose, i'm writing the ffi for naother library and I don't want to manually (or write a script) to get the value of each of these constants, is there anyway I can do something like: (c-define-foreign-varaible gl_normal_array "GL_NORMAL_ARRAY") and have gl_normal_array (only in compiled mode, I recognize this will fail in interpted mode) just get expanded to "GL_NORMAL_ARRAY" during compilation? thanks! (I also dislike the eralier method in that ... if API changes causes the global vars to have different values ... i'll have lots of fun with silent var value changes

5 5

thanks for the help procedure
by Pavel Dudrenov 17 Fév '09

17 Fév '09

I love the help procedure. It is extremely useful for a novice like me. Thank you very much much for introducing it. Pavel

3 5

Etags
by Vincent St-Amour 17 Fév '09

17 Fév '09

Hi everyone, I was using etags the other day (a program that generates emacs TAGS files, which can be used to easily jump to function definitions), and was rather disappointed that it didn't handle nested defines. I made a little script that does pretty much the same thing as etags, but handles nested definitions. I put it on the dumping grounds, in case some of you might be interested. You will need pregexp to use it. It's obviously not perfect, though. For example, having several functions with the same name defined inside different functions confuses definition lookup. Hope this can come in handy, Vincent

1 0

best practices with gambit / compile-to-c / and git
by lowly coder 16 Fév '09

16 Fév '09

foo.scm -> foo.c -> foo.o1 this is cool, except I keep track of stuff with git, and I like "git status" to be clean The foo.o1 files don't bother me (I have a "*.o*" in .gitignore -- but i'm a bit hesistant to put a *.c into my .gitignore) thanks!

2 1

address of var
by lowly coder 16 Fév '09

16 Fév '09

suppose I have: int foo; bar(&foo); <-- I want to duplicate this function call suppose now, I have foo and bar in scheme land; how do I make the above function call? grepping for "\&" in gambit-c.txt brings up the #& box syntax, but I don't think that's what I want thanks!

2 1

Compiler bug with the new container hooks
by Guillaume Cartier 15 Fév '09

15 Fév '09

The following code will crash the compiler instead of reporting the problem of duplicate c-define-type : - - - - - - - - - - - - - - - File: setup.scm (define (some-path->container-hook path) 'foo) (set! ##path->container-hook some-path->container-hook) - - - - - - - - - - - - - - - File: crash.scm (c-define-type jazz.platform.windows.WinTypes.LONG long) (c-define-type jazz.platform.windows.WinTypes.LONG long) - - - - - - - - - - - - - - - % gsc Gambit v4.4.0 > (load "setup") "c:\\setup.scm" > (compile-file "crash") *** ERROR IN EXPRESSION 983041 EXPRESSION 0 EXPRESSION 1975872 EXPRESSION 1978145 EXPRESSION 1988504 - - - - - - - - - - - - - - - Problem seems to be in locat-show in _source.scm when it's doing (if source (locat-show " " (source-locat source)))

1 0