Ciao,
I
corrected this morning a previous bug with HTTP retrieving. It
occurred when persistent connections were on and the max document size
was lower than the actual document size. So the stream wasn't completely
read.
In
the next request, the connection stayed up and the used stream was the
same. So, when read, it contained the previous request body (not already
read). So, here is a <Not header> error ...
I
override this bug, by closing the connection if the max document size has
been reached. Tell me your opinion, please. It's up to you.
The
same mechanism is the one I used for <Not parsable> documents, when
they are asked with a GET method. This way, the body is not read and the
stream is still full ... So, we close the connection.
The
real problem is that if after a read I send a write command on the same
connection, and then I read again, I still have on the stream the first
call result. Or better, I think so because I haven't been able to do the
opposite.
Anyway,
I think that now the mechanism works pretty good.
Well, that's it.
Ciao again
-Gabriele
------------------------------------------------------
Gabriele Bartolini - Prato (PO) - Italy
E-Mail: [EMAIL PROTECTED]
http://space.tin.it/io/gabrbart
http://www.po-net.prato.it
http://www.comune.prato.it/associa/acslucia
------------------------------------------------------
Share what you know. Learn what you don't.
------------------------------------------------------
