I've had a problem with many "500: Connection reset by peer" errors
showing up at seemingly random points in my script.

It always dies at the first call to sysread() in the "request"
subroutine in http.pm

I can perform the same action that my script does in Netscape or Lynx,
without ever getting these errors.

I did put in some code to repeatedly try the request over again until
it works, but sometimes it takes over 15 tries.  I even had it wait a
random number of seconds between 1 and 10 between each try.

I know this is an operating system generated error when the network
connection is dropped.


My question is: is there something with regard to fault tolerance on
socket connections that Lynx implements which LWP does not?

Is there a way to work around this problem (other than trying the
request over and over again)?


If anyone has thoughts or suggestions, they would be greatly appreciated.


-- 
--------------------------------------------------------------
Fair Winds,     Software Engineer and WWW developer
Chris Dunn      MRM, inc.

Email: [EMAIL PROTECTED]
Phone: (919) 544-6500  Ext 228
Pager: (919) 506-0819

http://www4.ncsu.edu/~w_dunn
--------------------------------------------------------------

Reply via email to