[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: awg-mod_lisp: mod_lisp for R5RS Scheme

From: R. Mattes
Subject: Re: awg-mod_lisp: mod_lisp for R5RS Scheme
Date: Tue, 20 Sep 2005 16:26:04 +0200

On Tue, 2005-09-20 at 08:06 -0400, Alan Grover wrote:
> R. Mattes wrote:
> > Alan, there seem to be a vew bugs in this code.
> Thank you for finding & reporting them. Fixed in latest release.
> > First, the guile-specific module contains '(use-modules (awg debug))'
> > and a commented-out trivial debugging version. Since you don't
> > distribute your debugging module with the code it might be better to
> > have the 'use-module' part commented out and the trivial debugging
> > function enabled.
> I opted to include my debug module.
> > Second: i don't understand the following:
> > 
> > (let* (
> >          (fn-name (or (and (pair? argv) (car argv) "once")))
> >  ....
> > 
> > Isn't there a second argument to 'or' missing?
> > In the context of your sample code this always evals to "once".
> The 'and was only supposed to have two arguments, making "once" the default.
> > Kind of frustrating since the daemon only handles one request
> > and your suggested test 
> > 
> >  ( sleep 1 ; wget -q -t 1 --post-data "post-data&foryou" ...
> >  
> > fails.
> By "fails", I assume you mean because of the awg-debug module.

No - since your '(or (and' construct always evaluates to "once"
the server only answers _one_ request and the exits. 
> Try with "thread", and the test daemon should keep listening now.

Probably not since fn-name will always be "once". BTW, there seems
to be a bug with the argument handling as well: your code defines
main with a 'rest' argument. The command line args are passed in as
a list already so you end up with a list-o-list.
So 'args' ends up being '(("mod_lisp-for-guile.scm" "thread"))'.
Note also that the first member of args is the program name.
So you need to change 

(define (main .args) .. => (define (main args) ...

and (car args) => (cadr args)

  HTH Ralf Mattes

reply via email to

[Prev in Thread] Current Thread [Next in Thread]