durusmail: quixote-users: Re: R: [Quixote-users] Serving quixote through iis
Serving quixote through iis
2005-10-06
Serving quixote through iis (while fixing a Quixote-WSGI bug)
Re: R: [Quixote-users] Serving quixote through iis (while fixing a Quixote-WSGI bug)
2005-10-06
2005-10-06
2005-10-06
2005-10-06
2005-10-07
Re: R: [Quixote-users] Serving quixote through iis
2005-10-07
Re: R: [Quixote-users] Serving quixote through iis
2005-10-07
Re: R: [Quixote-users] Serving quixote through iis
David M. Cooke
2005-10-07
Daniele Varrazzo  writes:

>> > Did anybody face the same problem and has some tip to share, or
>> some scar to show, about that?
>>
>> Scar indeed.
>>
>> I have used SCGI-CGI (http://arbutus.physics.mcmaster.ca/dmc/software/)
>> together with a python scgi server installed as a service.
>> However, SCGI-CGI needs heavy patching to compile under mingw32 and
>> IIS has some annoying CGI bugs.
>>
>> For example, if the status code of the http response is anything but
>> 200, then IIS 5 will replace the response with its own. IIS 6, on the
>> other hand, also replaces the response but doesn't touch the
>> content-length header, leaving your browser waiting for data that won't
>> arrive if your response happens to be longer than the IIS default one.
>>
>> The workaround is to strip away the content-length header before
>> passing the response to IIS. This way your redirects will work
>> again. You'll have to use status code 200 for all the other error
>> pages. Also note http://support.microsoft.com/kb/q176113/
>
> I wasn't for the scgi-cgi solution both for the lack of tests under win32
> and because, even if the script is short, Python is not so quick to start.
> ISAPI is more natural in IIS environment and i'm for it (if Apache is not a
> solution, of course)

Actually, it's a C program. My timings with Linux and Apache show it
to be on the order of only 2x slower that mod_scgi. It's not tested on
Windows, because I don't have that :-) (Works on Macs, though.)

cgi2scgi included with scgi does pretty much the same thing.

--
|>|\/|<
/--------------------------------------------------------------------------\
|David M. Cooke                      http://arbutus.physics.mcmaster.ca/dmc/
|cookedm@physics.mcmaster.ca
reply