Gambit-list Septembre 2008

gambit-list@iro.umontreal.ca

24 participants
33 discussions

Is long poll for activity on external file descriptors without Gambit reading data currently doable, and if so, how?
by Mikael More 03 Sep '08

03 Sep '08

Hi! Is there any way to make Gambit's scheduler trig an event when there is activity on external file descriptors, for instance as provided by curl_multi_fdset ( http://curl.haxx.se/libcurl/c/curl_multi_fdset.html ) of libCurl multi ( http://curl.haxx.se/libcurl/c/libcurl-multi.html )? The question was raised in 2004 ( https://webmail.iro.umontreal.ca/pipermail/gambit-list/2004-November/000023…), at that time without answer. What is the current answer to this question, is there any elegant way to have Gambit monitor a set of external file descriptors, and trig an event handler on event? ( There has been a quite extensive related discussion over the years, and one interesting instance was when Allan Langley published a libevent patch for Gambit, https://webmail.iro.umontreal.ca/pipermail/gambit-list/2005-August/000382.h…. Allan has personally not continued developing this respect of Gambit. Has someone else? Ought Allan's code work in the current Gambit version? ) Regards Mikael

3 8

parser bug? (gambit-c v4.2.8)
by naruto canada 02 Sep '08

02 Sep '08

hi try this: cd /usr ; rm -rf stalin* WGET_UNPACK_HERE ftp://ftp.ecn.purdue.edu/qobi/stalin.tar.Z rm -f stalin.tar.Z ; mv stalin-0.11 stalin WGET_UNPACK http://www.iro.umontreal.ca/~gambit/download/gambit/v4.2/source/gambc-v4_2_… ./configure --prefix=/usr --enable-single-host --enable-gcc-opts make || INTERVENTION $LINENO make install (don't worry about WGET_UNPACK WGET_UNPACK_HERE INTERVENTION, they simply wget and untar the tarball or check compilation error and does nothing else) cat /usr/stalin/benchmarks/*.sc > /tmp/1.sc cat /tmp/1.sc | mzscheme -e '(require (lib "pretty.ss"))' -f check-parameter.sc # ok no error cat /tmp/1.sc | /usr/v4.2.8/bin/gsi check-parameter.sc *** ERROR IN (stdin)(a)15339.7 -- Invalid token cat check-parameter.sc ;(set! *parser-canonicalize-symbols?* #f) (define (atom? x) (not (pair? x))) (define prev-token 0) (define (transform exp) (cond ((atom? exp) (begin (set! prev-token exp) exp)) ((list? exp) (map transform exp)) (else (begin (if (not (or (eq? 'lambda prev-token) (eq? 'define prev-token))) (begin (display ";;; pair used not as lambda parameter list? ") (pretty-print (list "exp:" exp "prev-token:" prev-token)) exp)) (display ";;; lambda parameter list : ") (pretty-print exp) exp)))) (define (loop) (let ((text (read))) (if (eof-object? text) (exit) (begin (pretty-print (transform text)) (newline) (loop))))) (newline) (loop) have fun :)

1 1

Re: [gambit-list] DrScheme and GambitC
by William Cook 02 Sep '08

02 Sep '08

What you say about Python, Ruby, and Perl is consistent with my understanding of those languages a year or two ago. However, there has been a big push within the dynamic language and Eclipse community to implement visual debuggers for these languages, and the progress is impressive. They have jumped forward and are now almost as good as Java and C# debuggers. Example include Komodo (commercial) and many Eclipse plugins. While many efforts have to date been separate, the Dynamic Languages Toolkit for Eclipse is attempting to bring them together so that they can reuse code (and simplify development). This means that Scheme is now falling pretty far behind the other languages, although they are much younger. I love Eclipse. I would use it just for the CVS and SVN plugins alone. And I'm happy contributing to it, because I feel I'm participating with and leveraging a very large user base. I am good at debugging too.. also from many years experience. Perhaps its bad that I rely on my debugging skills rather than proving everything up front. But, that's the way it goes. I guess I hadn't realized the degree to which Scheme is simply a tiny niche language without a big user base. Perhaps Scheme is plagued by too many implementations and so there is no critical mass on any of them. Sigh. William David Rush wrote: > Cool. I'm all for that. But here's the thing. All your comparisons are > to languages with a much larger user/implementation ratio. So the > Scheme community in general (and Gambit in particular, but not > especially) is resource-starved on the tools side. And by & large all > open-source languages have sucky debuggers: debuggers are the world's > most unrewarding programs to write (believe me I know what I'm talking > about here). There are a fortunate few languages where debuggers have > been built as later add-ons. Frankly, even a fairly popular newcomer > like Ruby has completely sucky debugging facilities. As far as I'm > concerned the VMS debugger ca 1990 was nearly perfect - and it has to > be said that the u$oft debugging tools are not bad either, but once > again these debuggers have corporate sponsorship. > > So I'm still using debugging techniques that date back to what I had > to do when I was cross-assembling Univac assembly to 80286 machines. > printf() and reading and *reasoning* about the code. And people wonder > how I am able to find and fix so many pernicious bugs that baffle the > rest of the engineering team... > > And to bring this full-circle, I'd love a great Scheme debugger. I'd > be happy enough with a great Gambit debugger (and it does have one of > the best inn Scheme - which is pretty damning). But not under eclipse > (which I hate for a myriad of reasons). And the current state of > affairs *is* workable and not much worse than working in Python, Ruby, > Per, or Tcl. > > david rush

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 Septembre 2008