improving performance of python webserver running python scripts incgi-bin

[Dale]

I am using a simple python webserver (see code below) to serve up
python scripts located in my cgi-bin directory.

import BaseHTTPServer
import CGIHTTPServer
class Handler(CGIHTTP Server.CGIHTTPR equestHandler):
cgi_directories = ['/cgi-bin']
httpd = BaseHTTPServer. HTTPServer(('', 8000), Handler)
httpd.serve_for ever()
This works fine, but now I would like to combine the python scripts
into the server program to eliminate starting the python interpreter
on each script call. I am new to python, and was wondering if there
is a better techique that will be faster.

Also, can someone reccommend an alternative approach to
httpd.serve_for ever(). I would like to perform other python functions
(read a serial port, write to an Ethernet port, write to a file, etc.)
inside the web server program above. Is there an example of how to
modify the code for an event loop style of operation where the program
mostly performs my python I/O functions until an HTTP request comes
in, and then it breaks out of the I/O operations to handle the HTTP
request.

thanks
Dale

[Bruno Desthuilliers]

Dale a écrit :

I am using a simple python webserver (see code below) to serve up
python scripts located in my cgi-bin directory.

import BaseHTTPServer
import CGIHTTPServer
class Handler(CGIHTTP Server.CGIHTTPR equestHandler):
cgi_directories = ['/cgi-bin']
httpd = BaseHTTPServer. HTTPServer(('', 8000), Handler)
httpd.serve_for ever()
This works fine, but now I would like to combine the python scripts
into the server program to eliminate starting the python interpreter
on each script call. I am new to python, and was wondering if there
is a better techique that will be faster.

Also, can someone reccommend an alternative approach to
httpd.serve_for ever(). I would like to perform other python functions
(read a serial port, write to an Ethernet port, write to a file, etc.)
inside the web server program above. Is there an example of how to
modify the code for an event loop style of operation where the program
mostly performs my python I/O functions until an HTTP request comes
in, and then it breaks out of the I/O operations to handle the HTTP
request.

May I suggest that you take a look at more sophisticated solutions, like
either wsgi, CherryPy or Twisted ?

[Mike Meyer]

On Thu, 10 Jan 2008 23:17:28 -0800 (PST) Dale <da********@gma il.comwrote:

I am using a simple python webserver (see code below) to serve up
python scripts located in my cgi-bin directory.

import BaseHTTPServer
import CGIHTTPServer
class Handler(CGIHTTP Server.CGIHTTPR equestHandler):
cgi_directories = ['/cgi-bin']
httpd = BaseHTTPServer. HTTPServer(('', 8000), Handler)
httpd.serve_for ever()
This works fine, but now I would like to combine the python scripts
into the server program to eliminate starting the python interpreter
on each script call. I am new to python, and was wondering if there
is a better techique that will be faster.

You can use BaseHTTPRequest Handler and override do_GET to handle the
actual request.

Also, can someone reccommend an alternative approach to
httpd.serve_for ever(). I would like to perform other python functions
(read a serial port, write to an Ethernet port, write to a file, etc.)
inside the web server program above. Is there an example of how to
modify the code for an event loop style of operation where the program
mostly performs my python I/O functions until an HTTP request comes
in, and then it breaks out of the I/O operations to handle the HTTP
request.

Use poll/select/whatever to check if httpd.socket has data available,
and then invoke httpd.handle_re quest() when it does. Nothing else will
happen while the request is being handled.

The suggestion to check the various frameworks for doing async work is
a good one as well.

<mike
--
Mike Meyer <mw*@mired.or g> http://www.mired.org/consulting.html
Independent Network/Unix/Perforce consultant, email for more information.