I agree that the problem most likely does not reside within
curl. But the fact remains that curl cannot retrieve a URL
while Netscape Navigator is able to do so.

I did some packet captures to see what was going on when the
following tools connect to www.cbs.com:
  - Navigator
  - GET (the script bundled with the LWP perl library)
  - curl
  - manual telnet session

The first two successfully retrieved the URL, while the second
two failed to do so.

The only obvious difference I can see so far is that the first
two tools sent hex 0x0D 0x0A 0x0D 0x0A to end their requests,
and the second two only sent 0x0D 0x0A.

This is strange, since I definitely hit "return" twice within
my telnet session. And I would think that curl would send the
CR-LF combination twice.

Any ideas?

Jason Priebe
WRAL OnLine
http://www.wral-tv.com/

> -----Original Message-----
> From: Jochen Roderburg [mailto:Roderburg_at_rrz.Uni-Koeln.DE]
> Sent: Wednesday, September 20, 2000 8:35 a.m.
> To: curl_at_contactor.se
> Subject: Re: hanging URL
>
>
> Hello,
>
> Some times ago I observed similar problems with other hosts
> and one of our
> network experts had found an explanation for them.
> Unfortunately he is on
> vacation in the moment, so I cannot ask him for the exact details.
> My own vague remembrance is that it had something to do with
> a misconfigured
> router or firewall in the neighbourhood of the problematic
> host. With the
> current example www.cbs.com you can see that this host cannot
> be reached with
> ping or traceroute, so obviously some low-level network
> filtering (ICMP Packets
> ?) is used somewhere. The side effect of this was that some automatic
> adjustment of other TCP/IP parameters in connection with
> fragmented packets did
> not work correctly and the end result was that the replies to
> the HTTP-Requests
> never came back. I don't know if there is a workaround for
> this, may be someone
> with deeper TCP/IP knowledge has further ideas.
> Sorry, that I cannot be more precise, but nevertheless I
> thought I share this
> experience with the list, before people do too much research
> at the curl end of
> the problem.
>
> Best regards,
>
> Jochen Roderburg
> ZAIK/RRZK
> University of Cologne
> Robert-Koch-Str. 10 Tel.: +49-221/478-7024
> D-50931 Koeln E-Mail:
> Roderburg_at_rrz.Uni-Koeln.DE
> Germany
>
Received on 2000-09-20