Replacement for locals() ???

Christian Tismer tismer at stackless.com
Sat Jul 31 13:35:28 CEST 2004


Harald Massa wrote:
> Terry,
> 
> 
> 
>>>So two of the greatest PythonBrains have problems in optimizing
>>>Python because of locals(), the most common usage of Python in my
>>>practice is peripheral; so just by way of precaution I am looking for
>>>a 
>>>r eplacement.
>>
>>I suspect you are very prematurely optimizing.  Since you are filling
>>out SQL statements, I would expect that your bottlenecks are and will
>>be the execution of the statements, and not the construction of the
>>statements. 
> 
> 
> you are partially right. I do not have ANY performance problems in that
> Python code. Most of the time the program is either waiting for a
> COM-Server, a database-Server or user-typing. 
> 
> As Skip stated: using locals() prevents the EASY use of psyco.  
> (import psyco; psyco full

Ok, I see. Note that the EASY use is almost always much slower
than a partial, controlled usage of Psyco, for non-trivial
applications.

The bad thing about locals() is that it simply goes wrong
with Psyco. I think it would be nicer if psyco would
simply refuse optimization in this case and leave things intact.

There are some constructs which just disable Psyco: If the
code contains certain opcodes, Psyco will not touch the whole
function. Examples are:

- generators: an existing yield disables Psyco for the func
- usage of nested scopes.

On the scopes case: If you have a construct like this:

def outer():
     something = 42
     def inner():
         print something
     inner()

Then both outer and inner will contain opcodes which disable
Psyco from touching things at all.
So if you can manage to inject such a construct, your locals
will be fine with Psyco.full().

Anyway, I think we should hint Armin (the other one) to
change the locals behavior to disable Psyco.

ciao - chris

-- 
Christian Tismer             :^)   <mailto:tismer at stackless.com>
Mission Impossible 5oftware  :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9a     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 89 09 53 34  home +49 30 802 86 56  mobile +49 173 24 18 776
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
      whom do you want to sponsor today?   http://www.stackless.com/




More information about the Python-list mailing list