[mod_python] CGI handler and multithreading: Can i have multiple sys.stdout objects?

Gregory (Grisha) Trubetskoy wrote:
> Your best bet is the prefork MPM. Remember that the 10M per child is
> actually a lot less because pages for executables are shared between
> processes on (most) UNIXes, certainly on Solaris they are. I find it hard
> to believe that it is slower than regular CGI - it should be at least
> several times faster, though not nearly as efficient as using the
> publisher or a native handler.
> 
> The ultimate solution is not use CGI of course :-)
> 
> Grisha

I haven't investigated things in 3.x.x yet, so I don't know if this 
applies.  If this were for mod_python 2.x.x, I'd suggest writing a small 
custom handler.  Just do a bit of error checking, setup a couple local 
variables, and an execfile().  I have an example below.

Your CGIs would need to be modified to use req.write() which is somewhat 
annoying.  For maximum compatibility, you could write a custom write() 
function to detect between outputting via req.write() or via stdout.  On 
a side note, do any mod_python request objects support variable argument 
outputs?

# httpd.conf example
<Location /cgi/>
     SetHandler python-program
     PythonHandler your_handler
</Location>

# your_handler.py example
# called by any request in this directory
def handler( req ) :
     from mod_python import apache

     cur_dir = "/directory/of/this/handler/"
     filename = req.filename[ len( cur_dir ): ]

     # setup any local variables

     try :
         # all my scripts are extension-less, so
         # I can easily test for ".."-style hacking
         if filename.find( "." ) != -1 :
             raise IOError
         execfile( "/directory/of/your/scripts/" + filename )
     except IOError :
         # handle error condition

     return apache.OK