durusmail: quixote-users: Re: Medusa bug or user misunderstanding? [patch files]
Medusa bug or user misunderstanding?
2004-01-21
2004-01-21
2004-01-21
Re: Medusa bug or user misunderstanding?
2004-01-21
2004-01-21
2004-01-21
Medusa bug or user misunderstanding?
2004-01-21
2004-01-22
Re: Medusa bug or user misunderstanding? [patch files]
2004-01-22
Jason E. Sibre (3 parts)
Re: Medusa bug or user misunderstanding? [patchfiles]
2004-02-04
2004-02-05
Medusa bug or user misunderstanding?
2004-01-22
2004-01-21
Re: Medusa bug or user misunderstanding? [patch files]
Jason E. Sibre
2004-01-22
Here are some patches I whipped up based on what I'm thinking at this
point...

These are based on .7a2 files, but I made sure .7a2 = .7a3 for these files.

I've done some light testing, and it appears that (based on the patches):
 http_request:
   - get_server() uses HTTP_HOST if available, and will fall back to
SERVER_NAME and SERVER_PORT
 medusa_http:
   - SERVER_NAME is set from "Host" if available, omitting the port.
   - if Host is not available, it uses
       "self.server.ip or socket.gethostbyaddr(socket.gethostname())[0]",
       to try and set a sane FQDN that matches what my Apache install does.


This seems to achieve what I think the goal is, at least.  But, I'm not sure
if we're all in agreement about the goal, yet.

Works with mozilla (HTTP/1.0 and HTTP/1.1), MSIE (HTTP/1.0 and HTTP/1.1),
links, and with my trusty telnet Host-less http/1.0 simulator.



Jason
reply