re: Workaround for LWP/Solaris missing EOF bug
John Gerrish (John_Gerrish@dot.state.ak.us)
Wed, 31 Jan 2001 14:57:58 -0900
<-- copy start from "libwww Digest 31 Jan 2001 20:02:20 -0000 Issue 195" ->
Subject: RE: Workaround for LWP/Solaris missing EOF bug
Date: Tue, 30 Jan 2001 16:00:44 -0600
From: Steve Livingston <slivings@DigitalRiver.com>
To: 'Phil Mitchell' <philip_mitchell@harvard.edu>,
libwww@perl.org
Just sun/solaris.
The command line output _appears_ to have a <cr> at the end.
Dunno if we can rule out Apache ... maybe the trailing <cr> is a (PC) ctl^m
-----Original Message-----
From: Phil Mitchell [mailto:philip_mitchell@harvard.edu]
Sent: Tuesday, January 30, 2001 3:38 PM
To: libwww@perl.org
Subject: RE: Workaround for LWP/Solaris missing EOF bug
So in your case, it has nothing to do with Apache... the cgi script is not
terminating its output the way LWP expects? I have not seen this problem w/
any apache servers... Is everything running on solaris? I'm curious to know
whether it would work if LWP was running on linux...
<-- copy start from "libwww Digest 31 Jan 2001 20:02:20 -0000 Issue 195" ->
begin
I had posted "Request Post error useragent.pm line 227" and in reviewing the
digest read your dialogue and realized it's the same problem. I was using LWP
to POST to a secure sever to authorize credit card transactions. It worked
fine from my Linux test machine and when I moved it to solaris it would 1 out
of 3-6 times produce the mentioned error in UserAgent.pm However the
transaction is received by the processor (who is running MS IIS) and
authorized/recorded, but I don't see the confirmation. I get different error
(the must handle timeout different) depending on whether using HTTPS or HTTP
protcol in
use HTTP::Request::Common qw(POST).
my $req = POST "https://...
Vern Hart <v-libwww@vern.com> in message
Subject: Can't call method "request" on an undefined value at ...
Date: Mon, 29 Jan 2001 00:48:23 -0700 (MST)
suggests adding some code to catch symptoms of problem in UserAgent.pm
So is MS IIS really not sending the expected EOF? The telnet test you mention
hangs right after </html> without any sign of CR LF or whatever. So is Sun
dropping what is sent or MS not sending it?
Does this problem persist to LWP 5.50?
Comments?
-Johnnie
--
John Gerrish RMS3 Application Manager 907-465-8861 fax:2024
Alaska Marine Highway System, DOT/PF, State of Alaska
3132 Channel Dr, Juneau, AK 99801
Blessed are the flexible for they shall not be bent out of shape.