W3C home > Mailing lists > Public > www-lib@w3.org > July to September 1998

Re: Problem with repeated

From: Henrik Frystyk Nielsen <frystyk@w3.org>
Date: Tue, 18 Aug 1998 15:59:42 -0400
Message-Id: <>
To: "Tom Friedel" <tom@readyink.com>, <www-lib@w3.org>
At 21:48 8/13/98 -0000, Tom Friedel wrote:
>I am trying to write a program that initializes the libwww library and
>repeatedly gets pages from the Internet.  However after about 10 iterations
>the requests fail.
>I added a while loop to w3c-libwww-5.1m/Examples/chunk.c to show this
>behaviour.  That program, BTW, calls HTRequest_new() before initializing the
>library, which appears wrong but probably harmless.  I am short of time
>right now and doubt I will be able to debug this problem, and am hoping for
>some advice.

I have written a slightly updated version of your program and I can't find
any problems. Could you please try the version based on the CVS checkout:


You can find the example at


Linked from


>Also, I noticed that running
>Robot/.libs/webbot -depth 1 -prefix http http://www.microsoft.com
>I get, after a lot of output,
>HTTPReq.c:65 Why is this?
>Aborted (core dumped)
>It doesn't seem like I should be on line 65 to begin with.  Can someone
>check these things, to see if I have a bad libwww.a build.  I happen to be
>running on a Linux system.

This is because IIS 4.0 thinks that it should close the connection on a 404
(Not Found) response and not only does it close the connection, it closes
it in both direction so when libwww does pipelining then it receives a TCP
RST which it tries to recover from.

However, the recover mechanism is quite difficult and at some point after
having recovered many times due to premature TCP closes it looses track.

I have reported this to the MS guys.

Henrik Frystyk Nielsen,
World Wide Web Consortium
Received on Tuesday, 18 August 1998 15:59:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:33:48 UTC