> "C. Derek Fields" <derek@gamekeeper.bellcore.com> writes:
>
> >Has anyone written an HTTP daemon in Python?
>
> I haven't, but it sounds like an interesting project.
A "toy" HTTP server is simple -
but I assume the reason for doing one in Python would be to add
some functionality over the existing ones - which might also mean
replicating all of the access-control, cgi-support, proxy support,
etc. of CERN & NCSA's httpd.
However, an alternative is to run two servers in parallel - a "standard"
httpd and a special version that handles redirects from the other.
That way you can experiment with added functionality without needing
to duplicate all of the existing capabilities.
Paul Everitt and some others have been trying to build a server that
serves from an object data-base name-space rather than from a unix
filesystem name space. ( And that is what I wrote server.py for -
as an easy framework to write specialized servers that do a bunch of
one time initialization up front and then fork off processes to
handle requests. Trying to put all that initialization into a CGI
running under HTTPD was too much overhead. )
BTW: I think I've solved (in my head! :-) and reconciled the
redirection problems with the previously posted versions of server.py.
I just haven't had time to write and test it. ( And I forgot about the
problem until this thread reminded me! )
---| Steven D. Majewski (804-982-0831) <sdm7g@Virginia.EDU> |---
---| Computer Systems Engineer University of Virginia |---
---| Department of Molecular Physiology and Biological Physics |---
---| Box 449 Health Science Center Charlottesville,VA 22908 |---