Re: [Chicken-users] tcp-*-timeout in types.db incorrect?

[Felix]

From: Jörg F. Wittenberger <address@hidden>
Subject: [Chicken-users] tcp-*-timeout in types.db incorrect?
Date: 05 Sep 2011 18:53:26 +0200

> I made chicken print out funny logic!
> 
> given tcp-accept-timeout set to #f I've seen the code taking the path
> supposed to be called with a number.  (See my recent message
> http://lists.nongnu.org/archive/html/chicken-users/2011-08/msg00074.html
> where I "fixed" the problem by passing the value through "identity".)
> 
> Now with the latest changes to types.db the trick is broken.
> 
> Which is a good thing, because I know now what to watch for.
> 
> Having gotten the problem back on top I inserted my famous "debug"
> procedure into the code.  I printed the value and a test.  The value
> (by the name "tma" here) was found to be #f and the test
> "(and  (number? tma) (> tma 0))" would print #t !!!
> 
> I circumvent the issue at this time by explicitly testing
> "(not (eq? tma) #f)" instead of just "tma".
> 
> But I guess the correct fix would be in types.db, since tcp-*-timeout
> don't convert a #f argument into anything else, the result would need
> to list the possible boolean value too, shouldn't it?
> 

That's right. Should be fixed now, thanks!


cheers,
felix