Re: [Chicken-users] Problem with threads

[F. Wittenberger]

Am Freitag, den 19.11.2010, 15:10 -0200 schrieb Jeronimo Pellegrini:
> Hello,
> 
> I'm using Chicken from git/master branch (commit
> ddaacdfdb11a0a1c0d1e6363b1af3af78746575c).
> 
> I recently noticed a change in the behavior of Chicken's
> scheduler. See this program:
> 
> (use srfi-18)
> 
> (define looping-thread
>   (let ((loop (lambda () (do ((i 0 (add1 i)))
>                         (#f)
>                       (print ':: i)))))
>     (thread-start! (make-thread loop))))

I'm NOT at all sure, that my following observation/assumption is correct
and applicable in your case.

But I ran into a similar situation last night and eventually found out
that this might more be related to port handling:

(Let me add: I have a LOT of asynch i/o going on from different threads
and just today managed to overcome some problem with my recent chicken
upgrade - that's where the "last night" comes from.  Everything works
fine even though I ignored my following observation):

Writes on (standard-output-port) from *different* threads appear to
share end-of-line-states in a different way that before.

And...
That's very much just OK until there is a defined behaviour wrt.
atomicity for each operation on the port.  (Which would be strange in
case of a complex thing like "print".)

So in this case I'd add a (flush-output (standard-output-port)) whenever
you want to actually see the current state.  (That's what you've should
have been done before as well, just to be sure.)