W3C home > Mailing lists > Public > www-lib@w3.org > January to March 2004

Re: NT SP6 and WouldBlock in HTTCP.C - Seemingly broken

From: Girish S <girishpage@yahoo.com>
Date: Tue, 30 Mar 2004 23:18:46 -0800 (PST)
Message-ID: <20040331071846.55726.qmail@web14008.mail.yahoo.com>
To: www-lib@w3.org

There is another serious BUG(?) related to DoConnect()
implementation in libwww.

I am trying to handle the situation explained below

Our application works with a web server. Here the
client application makes several calls to web server
in the same session. (uses cookie to maintain session)

When the web server goes down or the network is broken
with in active session, at this time if client
applicaiton makes a call to libwww to post some data,
libwww gets into an infinite loop in DoConnect() and
it consumes 100% Processor time.

I tried to handle this situation by using HTTimer to
implement application level time out. When http call
does not return before time out, Application Timer
triggers and Kills the request that is hanginig in
DoConnect() loop

But since, DoConnect block is hogging the processor,
The Timer does not trigger and applicaiton again has
no way to kill this request.

This is a serious problem which has to be addressed in
libwww. 

I am using libwww with WWW_WIN_ASYNC on windows xp.

Is there any solution to handle this kind of situation
or should our application release notes have this as a
known problem?

Please help

Thanks





__________________________________
Do you Yahoo!?
Yahoo! Finance Tax Center - File online. File on time.
http://taxes.yahoo.com/filing.html
Received on Wednesday, 31 March 2004 02:21:57 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 18 February 2014 13:20:04 UTC