Emile van Sebille wrote:
Lie Ryan wrote:
In a much simpler situation, even a communicating from a plain file
could be enough. In the daemon's program folder, there'll be two files:
input and output. You write to input to instruct the server and read the
response from output. This model is in respect to Unix's philosophy:
"make program to handle text streams, because it's the universal
interface".
I've done this and it works well... one thing to watch out for though is
snagging a file before it's completely written. Setting up a semaphore
or pausing to allow the file write to complete once seeing the file
fixes it adequately.
If instead of an input/output file you use a directory this is easily
solved. You write your request to a uniquely named file (with .in
extension) and you read the response out of an equally named file with a
.out extension. The server polls the directory for new .in files and
processes them in order of creation date (you can even include a
priorities scheme coded in the extension (e.g. .in1 to .in9 for the
different priorities).
_______________________________________________
Tutor maillist - Tutor@python.org
http://mail.python.org/mailman/listinfo/tutor
