Re: slow non-blocking reads

2006-07-05 Thread Lawrence D'Oliveiro
In article <[EMAIL PROTECTED]>, "Mark Dufour" <[EMAIL PROTECTED]> wrote: >interestingly, leaving out the fcntl stuff makes it work much faster. No surprises there... -- http://mail.python.org/mailman/listinfo/python-list

Re: slow non-blocking reads

2006-07-03 Thread Mark Dufour
interestingly, leaving out the fcntl stuff makes it work much faster. it seems to block only sometimes now, for just a moment, but on the whole the performance is acceptable now. On 6/29/06, Mark Dufour <[EMAIL PROTECTED]> wrote: > hello all, > > I am trying to fire up a child process using os.pop

Re: slow non-blocking reads

2006-07-01 Thread Lawrence D'Oliveiro
In article <[EMAIL PROTECTED]>, "Mark Dufour" <[EMAIL PROTECTED]> wrote: >any thoughts about why this runs extremely slowly? Because both processes are chewing up most of your CPU each waiting for the other to do something. -- http://mail.python.org/mailman/listinfo/python-list

slow non-blocking reads

2006-06-29 Thread Mark Dufour
hello all, I am trying to fire up a child process using os.popen2, and have the parent and child communicate in a non-blocking fashion. it works, but somehow it's unbearably slow. the following simulates a blocking readline: import os, fcntl fi, fo = os.popen2('./child') fcntl.fcntl(fo.fileno(),