> I was missing the windows build option totally.... If speeds between unix
> and Windows are at least on the same scale it can solve the
> mother-of-all-issues of windows deployments of python apps without using
> apache or iis.
>


I was not able to make some serious benchmark as all of the windows
machines i have are virtualized systems.

>From what i can measure the cygwin translation has a 10-15% impact in some
area (mainly in startup phases so it is nothing relevant), but we plan to
rewrite some critical part using the native windows api (we have already
done that for the locking part that instead of pthreads uses windows
mutexes).

Another limit i have not investigated is windows select() implementation
(on the others posix system we use more modern event facilities). But i
suppose until you do not use it with hundreds of concurrent file
descriptors you should not experiment problems.

-- 
Roberto De Ioris
http://unbit.it

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"web2py-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to web2py+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to