1.5.2 and functools or similar

Troels Thomsen nejtak...
Sun Mar 9 17:26:59 EDT 2008


Hello,

I am writing a simple delayed-call mechanism , that is causing a bit of 
headache. Works like this:

myPrint(s)
  print "..." + s

myTimer.add(myPrint , "hello" , 15)

This means that the myprint function is called in 15 seconds with the 
parameter "hello".
The housekeeping of these timers is called by the main loop of the "os"

This works well but i would like to be able to use it with any number of 
parameters

Functools is not a part of the 1.5.2+ python that I am running on (embedded 
device),
so i tried to pass the parameters as a tuple like this

myTimer.add(myAdder , (3,6) , 15)

and the housekeeping function would then call the function like this

def updateTimers()
  for timerItm in timerTable:
  ...
    ....
      ....
        timerItm.func(*timerItm.parameters)

Works well on python 2.5 but not on 1.5.2 (?)


Current solution is to have default parameters None for the add function

def add( func , timeout , param1 = None , param2 = None)

And the update function then checks if parameters is specified

def updateTimers()
  for timerItm in timerTable:
  ...
    ....
      ....
      # ugly part :
      if timerItm.param1 is not None and timerItm.param2 is not None:
        timerItm.func(timerItm.param1, timerItm.param2) # two parameters
      elif ......
        timerItm.func(timerItm.param1) # one parameter
      else
        timerItm.func() # no parameters

This has the implication that I can not call a function with the parameter 
None if I wanted to.
(not a huge problem)

Right now it works quite well with up to two parameters, it covers 99% of 
usage. If I need to call a function with more parameters, i can always write 
a wrapper function for it. Wondering if anyone had some sugestions ?


By the way, is it bad style to check for object identity instead of value 
"None".
What aboutt integers ? if value is 0: ..
I guess it is implementation specific / could change in future versions ?


Thx,
Troels












More information about the Python-list mailing list