plpythonu and "hello concurrent world"

Hi all. Im having some "problems" with a small concurrent plpython function.

Based on a small example [1] about concurrent programming, there is some
code which works fine under python:
#! /usr/bin/python

import threading, random
import time

def myPrint(str):
print 'searching...', str
time.sleep(random.randint(10, 10000) / 1000.0)
print str, 'OK!'

myThreads = (threading.Timer(random.random(), myPrint, ["hello"]), \
threading.Timer(random.random(), myPrint, ["concurrent"]), \
threading.Timer(random.random(), myPrint, ["world"]))

for thr in myThreads:
thr.start()

gherzig(at)linux: python pp.py
searching... concurrent
searching... world
searching... hello
hello OK!
concurrent OK!
world OK!

So far, so good. Almost the same example in plpythonu:
CREATE OR REPLACE FUNCTION search_t()
returns bigint
security definer
as
$$
import threading, random
import time

def myPrint(str):
plpy.notice ('searching...', str)
time.sleep(random.randint(10, 10000) / 1000.0)
plpy.notice(str, 'OK!')

myThreads = (threading.Timer(random.random(), myPrint, ["hello"]), \
threading.Timer(random.random(), myPrint, ["concurrent"]), \
threading.Timer(random.random(), myPrint, ["world"]))

for thr in myThreads:
thr.start()

return 10000
$$ language plpythonu;

gse_new_version=# select * From search_t();
search_t
----------
10000
(1 row)

Looks like myPrint() is not executing at all!!
Have no idea why, so i decided writing both on python and postgres
forums. Any ideas??

Postgres 8.1.3
python 2.5.1

Thanks!!
Gerardo

[1]
http://forums.hostrocket.com/showthread.php?t=13325

Gerardo Herzig <gherzig(at)fmed(dot)uba(dot)ar> writes:
> Hi all. Im having some "problems" with a small concurrent plpython function.

Don't even *think* about starting multiple threads inside the Postgres
backend. It's an excellent way to break things.

regards, tom lane

Tom Lane wrote:

>Gerardo Herzig <gherzig(at)fmed(dot)uba(dot)ar> writes:
>
>
>>Hi all. Im having some "problems" with a small concurrent plpython function.
>>
>>
>
>Don't even *think* about starting multiple threads inside the Postgres
>backend. It's an excellent way to break things.
>
> regards, tom lane
>
>---------------------------(end of broadcast)---------------------------
>TIP 6: explain analyze is your friend
>
>
>
>
Damn!! Well, comming from you Tom, i guess i will just look for another
approach without complaining ^^.
Thanks!

Gerardo