I have 2 working LAMPP boxes, both "identical" on FC4.
In both cases, the install and the webserving and mysql and php and all that LAMPP stuff works fine.
In both cases:
wget http://php.net WORKS
links http://php.net WORKS
telnet php.net 80 \n GET / HTTP/1.0 \n Host: php.net \n\n FAILS
php -r "echo file_get_contents('http://php.net')" FAILS
php with a similar curl script WORKS
ping and traceroute behave as expected
the same above scripts on another box (Windows and cygwin, no less) on the same subnet works, so it's probably not my network topology or DNS, right?
To clarify, definition of FAILS is:
hangs for 2 minutes and then errors out with:
Warning: file_get_contents(http://php.net): failed to open stream: HTTP request failed! in /opt/lampp/htdocs/test/php.net on line 1
and a similar result with different words from 'telnet' software.
I'm pretty certain that on a stock FC4 installation these would all work, as there would be much wailing and gnashing of teeth in my Google results to find this problem if it hadn't... I mean, PHP file_get_contents and telnet get used a LOT, you know.
So did LAMPP re-configure something to break telnet and file_get_contents while not breaking wget and links and php|curl?
I'm very puzzled...
I guess the only other things I should eliminate are:
#1 both boxes are on the same hub, so I guess it COULD be hardware... But how would a hub break telnet/file_get_contents but not break wget, links and curl? Seems very unlikely
#2 there's a 2-minute timeout before the failure, and a 2-minute warning in American football, so with the Bears being 5-0 and me being in Chicago... :-)
I'm mostly posting here because I'm out of ideas of what to test.
As a work-around, I'm using php|curl instead of file_get_contents() so I can move forward, but it's like pulling out a cannon to kill a fly.
Is it just me, or have others experienced this?
I'm not finding it in this forum or Google so far.
My IT/network guys here are mostly Windows, so haven't responded yet...