felix winkelmann wrote:
On Sun, Apr 5, 2009 at 11:29 PM, Matt Jones <[email protected]> wrote:
Kon Lovett wrote:
Versions before 3.0 are unsupported AFAIK. Packaged sources are
available at the Chicken website:
http://call-with-current-continuation.org as are instructions for
accessing the source repository.
Thanks for the tip. I have upgraded to version 3.4. The problem persists,
though. I looked at the source and the relevant part is the same. The
function 'fail appears to close the port -- (##net#close s) -- but somehow
these ports end up open and accumulate.
Is it possible that the condition-case handling somehow skips the
closing of the port? (A lame question, I know - but I don't
know what to do besides trying the code myself, for which I currently haven't
got the time)
cheers,
felix
Hm ... my quesiton is, What is the intended behavior of tcp-connect when
it signals an exception after failing to connect? Is it supposed to
leave a port open? That is what it appears to do.
I can't close the port when I handle the exception because I'm not in
possession of it. (The exception is signalled during the call to
tcp-connect, so it never returns a valid port.) That's why it seems
more likely that it shouldn't be open when the exception is signalled.
(BTW, this part of my message above
>> The function 'fail appears to close the port -- (##net#close s) --
>> but somehow these ports end up open and accumulate.
refers to the code in Chicken's tcp.scm, in the definition of
tcp-connect, not my code. My code isn't trying to close the port. My
point there was that tcp-connect appears to close the port before
signalling an exception, yet the port remains open. I realize that was
probably unclear.)
Best
Matt
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
