August 2012 - Gambit-list - mailman.iro.umontreal.ca

Edge triggered I/O with libevent
by Adam Langley 04 Apr '13

04 Apr '13

Gambit uses the select call on POSIX systems to multiplex IO operations. This is fine for small numbers of file descriptors, but falls apart quickly with large numbers[1]. libevent[2] is a wrapper library around a number of modern IO multiplexing systems (epoll on Linux, kqueue on the BSDs as well as poll and select). Attached is a patch which converts Gambit to using edge triggered I/O with epoll and libevent. This also requires a patch (attached) of libevent since the standard distribution doesn't include edge triggering support and the upstream appears to be on holiday. (epoll is required because I don't have a BSD box on which to develop edge triggering support for kqueue) This patch (for Gambit) preprocesses away to nothing unless --enable-libevent is passed to the configure script. (Note that, since this patches the configure.ac script, you'll need to run autoconf to regenerate). I hope I've not broken anything, but at this stage I've only tested that the console and tcp ports work. This patch (in the future) should also allow people, who write programs which use socket calls directly, to integrate into the main event loop at some point. Testing is more than welcome and, if you find a form of IO which is broken, please email me. [1] http://monkey.org/~provos/libevent/libevent-benchmark2.jpg [2] http://monkey.org/~provos/libevent/ -- Adam Langley agl(a)imperialviolet.org http://www.imperialviolet.org (+44) (0)7906 332512 PGP: 9113 256A CC0F 71A6 4C84 5087 CDA5 52DF 2CB6 3D60

2 2

Gambit internals documentation - draft of questions now on the wiki.
by Mikael More 31 Aug '12

31 Aug '12

Hi guys! Short summary: Please check section 4 and on on http://dynamo.iro.umontreal.ca/~gambit/wiki/index.php/Design_guide, and please give feedback the next 10 days! Long version: Please find here http://dynamo.iro.umontreal.ca/~gambit/wiki/index.php/Design_guide a draft of the Gambit documentation document spoken about in May. The purpose with the document is to communicate to the reader how Gambit works internally in the core areas (compilation, the stack with continuations, memory and object handling, threads, and so on). The document is a system of questions, whose answers from Marc will communicate this entire picture. Until now there has been no such documentation, and I believe this document provides a suitable format to give the whole picture in an efficient way. It's really important that there is documentation of Gambit's internals, to enable people to understand for themselves different aspects of Gambit's architecture, and make deep debugging and extensions. I hope and believe the document will prove valuable due to the profound yet conceptual level detail it is at. It's profound enough to provide sufficient information for those those who want to dig into and do hard work with Gambit's internals, yet conceptual and brief enough for the user to read the document in its entirety and get how all of Gambit fits together, with lots of associated aha experiences. I do believe that Marc's answers will give rise to some more questions in response, that will extend the document further a bit when they come. If you have any feedback - suggestions for more questions, clarifications, or perhaps removals, then please tell them here during the next ten days! Best regards, Mikael

3 3

Creating scheme modules with functions callable from C main program
by Andrew Suttles 31 Aug '12

31 Aug '12

Folks - I am writing a simple Windows native app in C using Win32 libraries. I'd like to to have some of my Windows callback functions call routines that I've written (or will write) in GambitC (because scheme is more fun to write and maintain than C). I've found material on how to call functions written in C in the scheme interpreter process, but not the other way around. In other words, when I compile a module of scheme functions into an object file, I don't have a header file to include in my main program written in C. How do I call those scheme functions? Thanks again for the help. I'm fairly new to Scheme and I'm *very new* to GambitC, so please excuse the the naivete of my question. Andrew

3 2

Gambit-C macros
by Bob McIsaac 26 Aug '12

26 Aug '12

Hi: 3 questions... I'm looking for a "when" macro and tried one patterned on "unless". It behaves like "when-not". How do I write a proper "when" macro? (define-syntax when-not (syntax-rules () ((when-not test body ...) (if test #t (begin body ...))))) (define (test-when n) (when-not (not (< n 50)) (println "line1 " n) (println "line2 " n) (println "line3 " n))) (test-when 0) (test-when 3) (test-when 55) ;------------------------------------------- Re "define-syntax" the Gambit-C manual says .."Note that this implementation of syntax-case does not support special forms that are specific to Gambit." What are the implications of this statement? Is there a preferred model for writing Gambit-C macros? ;------------------------------------------------------------ Some code has a "##" prefix like.. (##define-macro (err form) `(let () (##declare (safe) (generic)) (err-form (lambda () ,form)))) Why is that? =bob= ||

2 2

How access arbitrary memory addresses (r/w)?
by Mikael 09 Aug '12

09 Aug '12

Dear Marc, Here is something that is really key for some lowlevel things: I wish to do the equivalent of the C expression * ((byte*) x) in the Scheme world, i.e. the same as something like MOV AL,[EAX] . So, something like the C variant: for (int i = 12345; i < 12445; i++) print("%i ",* ((byte*) i)); would, if we call the read operator ##sysmem-byteref , be expressed in Scheme as something like: (let loop ((i 12345)) (if (< i 12445) (begin (print (##sysmem-byteref i) " ") (loop (##fxnum+ i 1))))) We can call the write operator ##sysmem-byte-set! . Also it would be of great benefit to have one for 16bit, one for 32bit, one for 48bit and one for 64bit words. I think it's justified that there is some way to reach this with highest efficiency, i.e. a way that does this inlined in the code thus producing something like * ((byte*) x) / MOV AL,[EAX] . Indeed this would provide of good efficiency on 64bit machines only, where fixnums can go up to ~10^18 (unlike 32bit machines where it's ~0.5 * 10^9, so pointers would generally be bignums). Is this in place already, if not how is it best implemented? (with ##c-codeand some magick around it perhaps; I'd suppose with ##c-lambda there's always some handling around that would make it clearly more than a single-cpu-instruction operation, even if compiled with (declare (not safe)) ). I may not be correct about this one but I got the impression that ##u8vector-ref/set! would *not* work for this, though perhaps a u8vector object could be produced that is immovable and has zero as its offset memory address? Thanks, Mikael

4 5

In Scheme, what's the benefits and disadvantages with using a Scheme or FFI library for a given purpose?
by Mikael 06 Aug '12

06 Aug '12

(repost w more suitable title) Below, I will understand that the use of SXML is made with the highly qualitative SSAX-SXML library, which has several parts (serialization, sxpath etc.) and is distributed for lots of Scheme implementations on a public domain basis. Essential parts of it are available as Black Hole modules for anyone for whom that might be of interest. Now: What is advantageous of using SXML or LIBXML depends on the application. Motivations for using a C library for a particular mechanism that is already implemented in very very well in Scheme would be, that you are using some code that requires you to use that C library anyhow, or, that there would happen to be some routines implemented in that C library that you need, are not in the Scheme library, and by some reason would be very time-consuming to implement. Also of course, like Cyrille points out, there can be occasions when you want or need a specific library choice due to reasons relating to how you handle the code in your application, or the specific computer environment in which your application will execute. >From a very wide perspective, I relate to such occasions as very rare. SXML is a data format standard that expresses the entire XML format in S-expression form. For several purposes this is highly usable. As a general rule, processing Scheme objects rather than FFI objects carries many benefits with it: the language and routines for processing and manipulation available is extremely rich already (sporadic examples: iteration and serialization), your need of custom code for many operations is zero, and also as a sidenote the memory handling is implicit in all steps, including safety of memory fragmentation. SXML is, generally speaking, an excellent example of when this applies. The only thing that SXML does not include, which a typical XML library does include, is a parent element attribute on (SXML/XML) elements. For the one for whom that would show to be relevant though, it's easy enough to implement, didn't hear anyone ask for that up to now though. If you have further thoughts or questions please feel free to share, Mikael 2012/7/17 Klaus Schilling <schilling.klaus(a)web.de> > What are the major advantages of using libxml2 xmlreader with ffi over > using sxml? > > Klaus Schilling >

2 1

problem with ffi and libxml2
by Cyrille Duret 05 Aug '12

05 Aug '12

hello, I have problem to parse xml string with ffi and libxml2. My code is in the test file libxml-raw.scm : (c-declare #<<end #include <stdlib.h> #include <libxml/xmlreader.h> xmlTextReaderPtr create_reader(const char* buffer, size_t len) { return xmlReaderForMemory(buffer, len, "_.xml", NULL, 0); } int first_type_prime(xmlTextReaderPtr reader) { int type; if( reader != NULL ) { int cont = xmlTextReaderRead(reader); if( cont ) { type = xmlTextReaderNodeType(reader); printf("read %d\n", type); xmlFreeTextReader(reader); return type; } xmlFreeTextReader(reader); return -1; } } int first_type(const char* s, size_t len) { return first_type_prime(create_reader(s, len)); } end ) (c-define-type xmlTextReader "xmlTextReader") (c-define-type xmlTextReaderPtr (pointer xmlTextReader)) (define create-reader (c-lambda (nonnull-char-string unsigned-int) xmlTextReaderPtr "create_reader")) (define first-type-prime (c-lambda (xmlTextReaderPtr) int "first_type_prime")) (define first-type (c-lambda (nonnull-char-string unsigned-int) int "first_type")) Two test functions are defined here : The first-type function take xml body and create the parser within the c body. The first-type-prime uses a pointer on xmlTextReader structure stored and the scheme-side. The first-type-prime function fails and I don't know what's going on : @ gsc-gambit Gambit v4.6.4 > (compile-file "libxml-raw.scm" 'cc-options: "-Wno-write-strings -U___SINGLE_HOST -I/usr/local/include/libxml2" 'ld-options: "-L/usr/local/lib/ -lxml2") "/usr/users/cydu/programming/scheme/modules/xml/libxml-raw.o1" > (load "libxml-raw") "/usr/users/cydu/programming/scheme/modules/xml/libxml-raw.o1" > (define r (create-reader "<xml>toto</xml>" 15)) > r #<xmlTextReader* #2 0x801781300> > (first-type-prime r) _.xml:1: parser error : Char 0x0 out of allowed range <xml ^ _.xml:1: parser error : Couldn't find end of Start Tag xml <xml ^ read 0 0 > while this call works : > (first-type "<xml>toto</xml>" 15) read 1 1 Any idea what's happened ? thanks a lot cyrille

4 5

Running the benchmarks
by Alex Queiroz 03 Aug '12

03 Aug '12

Hallo, With the advent of clang 4.0 I would like to run the Gambit benchmarks[1] so as to compare with GCC 4.7. Is there an easy/convenient way to download and run the benchmarks? [1] - http://www.iro.umontreal.ca/~gambit/bench.html Cheers, -- -alex http://www.artisancoder.com/

3 6