Gambit-list Décembre 2011

gambit-list@iro.umontreal.ca

11 participants
11 discussions

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

04 Avr '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

slime (swank) and gambit
by Attila Lendvai 13 Fév '12

13 Fév '12

dear list, i'm a new hacker on board. i'm planning to use gambit on an embedded device. i have many years of CL experience, but i'm open for anything good. i have quite some background with slime and been using it and hacking on it for long years, so i'd like to use it with gambit. i've seen this mail, but noone answered: https://mercure.iro.umontreal.ca/pipermail/gambit-list/2011-August/005262.h… what is the status of swank-gambit? who is the current maintainer? where is its official repo? i've found two, but neither seems anything like 'official', but rather a random repo on github. https://github.com/jlongster/swank-gambit https://github.com/shortsightedsid/swank-gambit i'm planning to invest time into making it more useful, but i don't want to waste efforts on something that has already been done, just hidden from my google skills. any answers/hints are greatly appreciated, -- attila Notice the erosion of your (digital) freedom, and do something about it! PGP: 2FA1 A9DC 9C1E BA25 A59C 963F 5D5F 45C7 DFCD 0A39 OTR XMPP: 8647EEAC EA30FEEF E1B55146 573E52EE 21B1FF06

5 14

How to ue a module?
by Hendrik Boom 25 Jan '12

25 Jan '12

I just made bootstrap from the 4.6.1 development sources, which I downloaded using git. When I start gsi and, according to the section "Subdivide code into modules" in the tour of Scheme in Gambit, I use "use" it tells me hendrik@notlookedfor:~/dv/lang/gambit/gambit$ gsi/gsi Gambit v4.6.1 > (use ssax-sxml) *** ERROR IN (console)(a)1.2 -- Unbound variable: use 1> So evidently something is still missing, unconfigured, unbuilt, or whatever. It doesn't even get around to looking for the library -- which is another question: how does it find the libraries, assuming it gets that far. And if the X module system is itself a module, how do I bootstrap it into the system? Or should I be using an entirely different mechanism? -- hendrik

8 24

How to pass a Scheme string to a const char*?
by Vok Vojwo 21 Déc '11

21 Déc '11

I have defined wrappers for openlog and syslog. The openlog function requires an ident string. Whatever I pass to openlog does not survive until a call to syslog. I have tried to protect the string with ##still-copy (with and without box/unbox) but it does not work. Here is the example: (define const-ident #f) (define openlog (let ((openlog (c-lambda (char-string int int) void "openlog"))) (lambda (ident option facility) (set! const-ident (box (##still-copy ident))) (openlog (unbox const-ident) option facility)))) (define syslog (let ((syslog (c-lambda (int nonnull-char-string) void "syslog"))) (lambda (priority . args) (syslog priority (apply string-append args))))) This program: (define (main . args) (openlog "gambit" (+ LOG_NDELAY LOG_PID) LOG_USER) (syslog LOG_INFO "hello") (closelog)) will generate random syslog tags: Nov 21 17:58:49 sn-e0692 ��*#002[11621]: hello What is the correct way to protect a string to be able to pass it to a const char* argument?

4 7

Defining signal handlers
by Sascha Ziemann 19 Déc '11

19 Déc '11

I was searching for a way to define signal handlers. I found a description here: https://mercure.iro.umontreal.ca/pipermail/gambit-list/2006-January/000529.… which is also in examples/misc/signals.scm. It works by calling ___EXT(___raise_interrupt) (___INTR_6); in the C signal handler and registering an corresponding function in Scheme: (##interrupt-vector-set! 6 (lambda () (println 'got-SIGUSR1))) Is this still the preferred way to install signal handlers? And if so how are the numbers chosen for the call to ##interrupt-vector-set!? I found the definition in lib/_kernel.scm: (define ##interrupt-vector (##vector #f #f #f #f #f #f #f #f)) (define-prim (##interrupt-vector-set! code handler) (##declare (not interrupts-enabled)) (##vector-set! ##interrupt-vector code handler)) It seems to me that 7 is the last interrupt one can use. If I use 6 and 7 for USR1 and USR2 how can I set a handler for SIGHUP, SIGTERM, SIGQUIT, SIGPWR and SIGIO?

1 0

cgi in Gambit
by Eduardo Costa 14 Déc '11

14 Déc '11

I need to use Scheme or Common Lisp for writing CGI scripts to process the XSL/XML protocols of a law firm. I started with Stalin, since it is easier to compile. The resulting exec files work perfectly well, and have small foot prints. However, the Stalin community does not seem to be very active. Besides this, Stalin is very slow to compile, and does not provide an interpreter. Therefore, I decided to switch to Gambit, Bigloo or SBCL. Gambit has the advantage of providing native infix notation. However, it is not working in most commercial servers that the law firm uses. I am using the following command line to create the exec files: $ ./gsc -exe gambtest.scm Where gambtest.scm contains the following lines: (display "Content-type: text/html")(newline)(newline)(display "<html> <body> <h1>Hello from Gambit</h1> </body></html>" )(newline) It generates an executable file that works perfectly well in my machine, but not in the servers that host the pages of the law firm. $ ./gambtest Content-type: text/html <html> <body> <h1>Hello from Gambit</h1> </body></html> For reasons that I cannot fathom, the law firm uses many hosting providers. I also tried to use Gambit as a scripting language. The result is the same. In the case of Bigloo, the problem lies in the fact that it uses dynamic libraries incompatible with the server. I cannot understand why this happens, since Stalin works even when linked to dynamic libraries. By the way, I prefer the scripting solution, since the static exec files are very large. I installed gambit in the hosting provider, making sure that both gsi and gsi-script have permission to execute. I also tested the scripts with Dorai Sitaram's Scheme in Common Lisp to make sure that everything is working fine. Here is the script (that works in my machine): #! ./gsi-script -:d- (display "Content-type: text/html")(newline)(newline)(display "<html> <body> <h1>Hello from Gambit</h1> </body></html>" )(newline) Here is the message that I get from Hostgator both for the scripting version and the compiled version: Internal Server ErrorThe server encountered an internal error or misconfiguration and was unable to complete your request.Please contact the server administrator, webmaster(a)advogadosmg.org and inform them of the time the error occurred, and anything you might have done that may have caused the error.More information about this error may be available in the server error log.Additionally, a 404 Not Found error was encountered while trying to use an ErrorDocument to handle the request.Apache/2.2.21 (Unix) mod_ssl/2.2.21 OpenSSL/0.9.8r DAV/2 mod_fcgid/2.3.6 FrontPage/5.0.2.2635 mod_bwlimited/1.4 mod_auth_passthrough/2.1 Server at advogadosmg.org Port 80

5 4

Re: [gambit-list] cgi in Gambit
by Marc Feeley 14 Déc '11

14 Déc '11

On 2011-12-14, at 1:08 AM, Eduardo Costa wrote: > Hi, Marc. > > I still cannot understand what is going on. When I compile for Bigloo, I get the same error if I compile with the following options: > > bigloo -static-bigloo whatever.scm -o whatever.cgi > > However, if I pass the static option to the gcc compiler, everything works well. > > bigloo -static-bigloo -copt -static whatever.scm -o whatever.cgi > > I cannot see why Gambit cannot have the same options. I mean, an all static linking. BTW, I tried to pass the static option to gcc through cc-options, but I get an error. It is possible to statically link the C runtime by specifying this to the "configure" script. For example: % make clean % CC="gcc -static" ./configure --enable-single-host % make % strip gsi/gsi % ls -lh gsi/gsi -rwxr-x--- 1 feeley parallelisme 4.8M 2011-12-14 09:18 gsi/gsi % ldd gsi/gsi not a dynamic executable % make clean % CC="gcc" ./configure --enable-single-host % make % strip gsi/gsi % ls -lh gsi/gsi -rwxr-x--- 1 feeley parallelisme 4.5M 2011-12-14 09:20 gsi/gsi % ldd gsi/gsi linux-vdso.so.1 => (0x00007fff96d70000) libutil.so.1 => /lib64/libutil.so.1 (0x0000003f6fe00000) libdl.so.2 => /lib64/libdl.so.2 (0x0000003f62600000) libm.so.6 => /lib64/libm.so.6 (0x0000003f62200000) libc.so.6 => /lib64/libc.so.6 (0x0000003f61e00000) /lib64/ld-linux-x86-64.so.2 (0x0000003f60c00000) [Note that for some reason I don't understand, I had to remove the -mieee-fp flag detection logic from the configure script because that was causing a multiple symbol definition at link time on the Linux machine I tried this on.] > I have a completely unrelated question. I noticed that Gambit has a much smaller compiler than SBCL. Depends. On my Mac OS X machine, the Gambit compiler (stripped) is 5.5 MB and the sbcl compiler's sbcl.core file is 41 MB. Don't forget however that the Gambit-C compiler relies on the presence of a C compiler (typically gcc) which also takes quite a bit of space, probably many times more that sbcl.core when you include precompiled libraries and header files. On the other hand it allows you to mix Scheme and C code and easily interface with foreign code. > Do you think that this could be a great advantage for scripting? I mean, since Gambit has a smaller compiler, it would have a faster start-up. Does it make sense? If you only care about using the interpreter (which would make sense for scripting), then Gambit will be advantageous. Gambit does have a fast startup. On my Mac OS X machine which has a SSD, it takes 7 ms to start gsi, and 13 ms to start gsc. So using gsi or gsc for scripting is efficient. Note that on the same machine, sbcl starts up about as quickly as gsc (14 ms). Marc > --- On Tue, 12/13/11, Marc Feeley <feeley(a)iro.umontreal.ca> wrote: > > From: Marc Feeley <feeley(a)iro.umontreal.ca> > Subject: Re: [gambit-list] cgi in Gambit > To: "Eduardo Costa" <edu500ac(a)yahoo.com> > Date: Tuesday, December 13, 2011, 4:14 PM > > > On 2011-12-13, at 4:35 PM, Eduardo Costa wrote: > > > Well, people at hostgator, believe me or not, don't know what a path is... They use something called fakepath. In any case, I created a page so you can make the tests you feel fit. In order to enter that page, type: > > I tested your setup quickly and found that you linked Gambit with GLIBC 2.11, but only GLIBC 2.5 is installed. The message I get when executing gsi is > > ./gsi: /lib/libc.so.6: version `GLIBC_2.11' not found (required by ./gsi) > > You can get this information by visiting the page: > > http://advogadosmg.org/cgi-bin/env.cgi > > You might want to change the script env.cgi to poke around the server to see how it is configured. > > I suggest you recompile Gambit on a system which has GLIBC 2.5 . That should solve your problem. > > Marc >

1 0

Re: [gambit-list] Questions about Web pages in Scheme
by Marc Feeley 13 Déc '11

13 Déc '11

For the benefit of the readers of this list, I am copying my reply to a question on comp.lang.scheme. Marc On Dec 12, 4:55 pm, Eduardo Costa <edu50...(a)gmail.com> wrote: - Hide quoted text - > I have posted similar questions in the comp.lang.lisp list, but I > would like to know the opinion of the Scheme community as well. I am > writing dynamic web pages in commercial hosting services. Most hosting > services won't let me install my own server, or libraries. Therefore, > I am forced into CGI. > Right now, the programs are working in Stalin. Scheme code includes a > lot of text processing stuff, like Dorai Sitaram's pregexpr, htmltags, > etc. Even so, Stalin has a very small footprint (219 kbytes). There > is no question that Stalin is linking to dynamic libraries. On the > other hand, Bigloo and Gambit code require static linking. The > command line to compile with Stalin is given below: > $ stalin -On -t cross.sc -copt -O3 > The gambit command line takes the form: > $ gsc-script -exe -o gambit.cgi gambtest.scm > The gambit program is one order of magnitute larger than the Stalin > application. What is worse, the gambit program does not work. When I > try to call gambit.cgi from the web, I get: > The server encountered an internal error or misconfiguration and was > unable to complete your request. > My first guess is that I need to compile the program with static C > libraries. Therefore, I modified the command line thus: > $ gsc-script -exe -ld-options -static gambtest.scm > collect2: ld returned 1 exit status > *** ERROR IN ##main -- C link failed while linking "/home/lisp/gambc- > v4_6_2-devel/gsc/gambtest.o" "/home/lisp/gambc-v4_6_2-devel/gsc/ > gambtest_.o" > lisp@LispMachine:~/gambc-v4_6_2-devel/gsc$ gsc-script -exe -ld-options > -static gambtest.scm > Why Stalin linked to dynamic libraries works everywhere, but Gambit > does not? How to fix the Gambit program? First of all it is possible to dynamically link the Gambit-C runtime. When you run the configure script, add the --enable-shared command line option. Static linking is the default because on most desktops it isn't a big issue to copy the runtime system to each program (it is about 5 MB), and it avoids problems with setting up the correct path for loading dynamic libraries. A hello-world program is a few KB when using a dynamically linked Gambit runtime. Moreover, have you considered running your CGI scripts as Scheme scripts? Then the size of your script is really just the size of your source code. This can be achieved by starting the first line of your CGI script like this: #! /usr/bin/gsi-script (display "hello world!\n") Adjust the path if Gambit is not installed at the usual place. Now back to the problem you are encountering. It is probably due to a misuse of the controlling terminal. By default, REPL interaction goes to the controlling TTY of the process (this includes output that is generated with the pp procedure, or any tracing, warning and error messages). This is probably not what you want for a CGI script. You can redirect the interaction channel to the standard input/output by adding the flag -:d- when the program is launched. This can also be achieved by adding that option to the first line of your script, i.e.: #! /usr/bin/gsi-script -:d- (display "hello world!\n") This will work whether the script is interpreted, or compiled. For example, if the above script is called foo.scm, you can compile it to foo.cgi with the command gsc -exe foo.cgi foo.scm If your problems persist or if you have other questions, please direct them to the Gambit mailing list. Marc

1 0

Announcement: ELS 2012 in Zadar/Croatia
by Dr. Damir Cavar 07 Déc '11

07 Déc '11

Hi there, sorry for multiple posting and misusing this list for such announcements, but I would like to point you to the European Lisp Symposium in Zadar, Croatia, in 2012. As you know, Scheme and Gambit topics are definitely welcome there. Please have a look at the conference web page for details: http://european-lisp-symposium.org/ Maybe I'll see some of you there? Sincerely DC -- Dr. Damir Cavar http://cavar.me/damir/ phone: +1 (734) 709-6039 FaceTime: dcavar(a)me.com

1 0

Need help on understanding GVM bytecode
by Meng Zhang 07 Déc '11

07 Déc '11

Hi everyone, I'm trying to understand the GVM bytecode, while I got some problem on understanding its handling of stack frames. In the paper <A Parallel Virtual Machine for Efficient Scheme Compilation: .... JUMP and COND instructions cause the stack pointer to be recalculated, while LABEL instructions recalculated the frame pointer... While I can hardly understand these description, with the ".gvm" I can hardly understand how backend should do with given instruction. I do read the the c backend implementation, it uses one "fp" pointer and calculate stackpointer based on that. But the other operations the it uses is too difficult to understand without any documents, so I'll be grateful if anyone could give me a concrete example on gambit's stack handling. -- Meng Zhang Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

2 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Gambit-list Décembre 2011