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

RE: [gambit-list] Gambit 4.0 under MS Windows
by Dan Hristodorescu 13 Jul '05

13 Jul '05

I used mingw on Windows, the latest beta 13 can be compiled without problems even when using C++ (--enable-cplusplus). I haven't compiled b13 with Visual Studio but I compiled b12 and I had to fix the file reference in the project but tha's fairly easy. Regards, Dan Hristodorescu. -----Original Message----- From: gambit-list-bounces(a)iro.umontreal.ca [mailto:gambit-list-bounces@iro.umontreal.ca] On Behalf Of Mark H. Shirley Sent: Tuesday, July 12, 2005 11:03 AM To: gambit-list(a)iro.umontreal.ca Subject: [gambit-list] Gambit 4.0 under MS Windows Hello, Is anyone compiling the recent Gambit 4.0 betas under MS windows? I'm interested in trying to use it (integrating it with some other software) and I don't want to reinvent the wheel wrt the make files or visual studio project files. Thank you, Mark Shirley _______________________________________________ Gambit-list mailing list Gambit-list(a)iro.umontreal.ca http://mailman.iro.umontreal.ca/mailman/listinfo/gambit-list

1 0

Gambit 4.0 under MS Windows
by Mark H. Shirley 12 Jul '05

12 Jul '05

Hello, Is anyone compiling the recent Gambit 4.0 betas under MS windows? I'm interested in trying to use it (integrating it with some other software) and I don't want to reinvent the wheel wrt the make files or visual studio project files. Thank you, Mark Shirley

1 0

md5 module, bug(?) in read-subu8vector, and some questions
by Christian 11 Jul '05

11 Jul '05

Hello I've put up a new version of my chjmodule on http://scheme.mine.nu/gambit/chjmodule which, besides some other modules and changes, now contains an md5 module compatible with the chicken one. You can also look at the latter (and it's dependencies) at http://scheme.mine.nu/gambit/chjmodule/from-chjmodule-20050711/ There is one serious problem with the md5 module, though: if you feed md5:digest a port which is delivering more than 1024 bytes, it will only digest the first 1024. It seems this is a problem in read-subu8vector, which returns with length 0 after the first 1024 bytes. I'm using gambit 4.0 beta 13 (on Debian ppc). I also have some questions: - the documentation on read-subu8vector doesn't explicitely say how to reliably find out whether eof has been reached (as opposed to a timeout). What's the correct usage? - how can I (pre-)allocate datastructure *per thread*? I'm currently caching buffers to save gc overhead, but that's surely not thread safe. I could try using parameters, but then how do I find out if the current content of a parameter stems from a previous thread or not? (short of saving the current thread id in the parameter as well, which starts looking complicated) md5.scm is also missing better error handling in in ___BEGIN_CFUN_u8vector_to_pointer better: one should add argpos to the error (how?), and use "the" correct error message (___STOC_U8_ERR is clearly not right) or create a new one (hm). I'm also not sure atm whether C checks arguments like "uint8 a[16]" for the correct length. If so, it would be nice if the FFI would allow to define C types with the length - I can't see how to do that currently, given that only one string can be fed to c-define-type. Cheers Christian.

1 0

You should specialize (make-vector n x) when n is known at compile-time
by Bradley Lucier 29 Jui '05

29 Jui '05

Marc: I took the gcbench.sch file from Clinger, changed (let-class (((make-empty-node) (make-vector 4 0)) to (let-class (((make-empty-node) (vector 0 0 0 0)) and the run time went from (time (gc-benchmark)) 4834 ms real time 4679 ms cpu time (4619 user, 60 system) 38 collections accounting for 1519 ms real time (1465 user, 9 system) 1247819792 bytes allocated 22074 minor faults no major faults to (time (gc-benchmark)) 3110 ms real time 2981 ms cpu time (2902 user, 79 system) 57 collections accounting for 1443 ms real time (1374 user, 16 system) 1247692592 bytes allocated 17367 minor faults no major faults on a 64-bit gambc40b13 on Opteron. This should be done by the compiler. Brad

1 0

builtin syntax, namespaces and expansion loops
by Christian 26 Jui '05

26 Jui '05

Hello I've noticed that one cannot refer to builtin syntax using fully qualified notation (something like foo#quasiquote isn't defined; so builtin syntax is present in every namespace but not accessible with any full qualification) and one thus cannot write macros substituting the builtin syntax with something else and then still call the original syntax in the generated code: (define-macro (quasiquote . l) (display "quasiquote called\n") `(quasiquote ,@l)) `1 => leads to infinite loop, and there's no apparent way around that. Additionally, this loop cannot be terminated by hitting ctl-c, which means the whole process has to be terminated. Is there a way to write the above quasiquote wrapper so that it works? Thanks Christian.

2 1

[gambit-list] New array code, with tests and documentation (sort of)
by Bradley Lucier 25 Jui '05

25 Jui '05

I've been working with two, three, and four-dimensional arrays of data in image processing applications for several years now. These arrays have contained integer-valued and floating-point-valued grey- scale and color pixels, 32-bit floating-point numbers, pairs of 64- bit floating-point numbers interpreted as complex numbers, ... In the process, I've developed a library of code to manipulate what I call "generalized arrays" in a way that is fast (well, faster than writing loops in Matlab if the function you want isn't built in) and intuitive (if you're used to high-level functions and Scheme). I'm looking for testers. I intend to submit an SRFI eventually, so I've written some preliminary documentation. There is test code that has pretty good coverage. I'm looking for more testers, comments, etc. The code and documentation are at http://www.math.purdue.edu/~lucier/array-code/ Warning: If you build gambc40b13 on MacOS X 10.4.1 with XCode 2.1 compilers and env CC='gcc -mcpu=970 -m64' ./configure enable-single-host and compile pixel-method.scm with % gsc > (compile-file "pixel-method") #t > and then try to (load "pixel-method") then it may crash your machine. It does mine, consistently. (Marc: You should try it on your new dual 1.8 G5, it's rather unbelievable. I'd like to get this fixed, however.) The code works fine on an Opteron machine running RHEL 3.0. Brad

1 0

Need better error message
by Bradley Lucier 25 Jui '05

25 Jui '05

Marc: I tried to declare (not block) in a long list of declarations and was told: > (compile-file "pixel-method") *** ERROR IN "/Volumes/lucier/text/papers/b-one-infinity/programs/ pixel-method5/generic-arrays.scm"@1294.1 -- Declaration can't be negated #f leaving me to guess which of my (not ...) declarations could not be negated. Brad

1 0

integer->char for an integer from interval [128, 255]
by Dmitry Lizorkin 24 Jui '05

24 Jui '05

Hello! I am working with Gambit interpreter, and the following result is unexpected for me: (string (integer->char 127)) ==> "\177" (string (integer->char 128)) ==> "" (string (integer->char 256)) ==> "\x100" I don't understand why the empty string is returned for an integer from the interval [128, 255]. Could you please exlpain this to me? What should I do to obtain the character that corresponds to an integer from the interval [128, 255]? Dmitry P.S. I am using Gambit version 4.0 beta 12.

2 2

create-directory permissions
by ben＠fuhok.net 17 Jui '05

17 Jui '05

I imagine that it is intentional, since it works as the documentation says it should, but why is the default mode for `create-directory' #o666? It seems to me that one would more usually want something like #o755. Thanks, Ben

2 1

Fwd: A module system - update
by Christian 17 Jui '05

17 Jui '05

This mail didn't make it to the list. >Date: Tue, 14 Jun 2005 11:36:43 +0200 >To: Gambit-list(a)iro.umontreal.ca >From: Christian <christian(a)pflanze.mine.nu> >Subject: A module system - update > >Hello > >I've packaged a new version which corrects a dumb bug which I >introduced while creating the first release (the "import" function >gave an error), improved the note about fully qualified names in the >README file, and added 7 more modules: cj-math, cj-u8vector-util, >cj-realrandom, cj-realrandom-formatted, >cj-webserver-saveload-serialize, cj-webserver-saveload-hashtable, >and cj-webserver (the latter being a modified version of Marc's >web-server example, which doesn't send the continuations to the user >but stores them locally in the filesystem -- do "mkdir tmp" before >running it from the repl with "(main 1234)"; by replacing >cj-webserver-saveload-serialize with cj-webserver-saveload-hashtable >in cj-webserver-module.scm you can switch to a version which keeps >them in memory). > >Christian. > >(PS. I've also corrected an error in my email setup; mail being sent >to me yesterday should still reach me if relays waited long enough >to retry now; in case you've got an error, please try again and >accept my apologies.)

1 0