[albgarse]

Hi!

I'm developing an application that makes an http connection every minute. It seems to work fine, but when it's working for a while, it stop to make connections. In the app log, I can see that the function urllib.urlopen is called, but it never returns. There is not timeouts or exceptions, what seems very rare.

Any idea of what could be happening?

Best regards.

[marcelobarrosalmeida]

Hi albgarse

Are you using PyS60 1.4.5 or PyS60 1.9.5 ? Even we do not have timeout option in urllib in 1.9.5 as well (I tested today), at least it is newer and probably more stable.

[albgarse]

I'm working with 1.4.5. I'm going to try with 1.9.5.

EDIT:
Just tried, but my app won't start with the new runtime. I use some extensions that propably are no compatible.

Best regards.

Originally Posted by marcelobarrosalmeida

Hi albgarse

Are you using PyS60 1.4.5 or PyS60 1.9.5 ? Even we do not have timeout option in urllib in 1.9.5 as well (I tested today), at least it is newer and probably more stable.

[gaba88]

Originally Posted by albgarse

I'm working with 1.4.5. I'm going to try with 1.9.5.

EDIT:
Just tried, but my app won't start with the new runtime. I use some extensions that propably are no compatible.

Best regards.

hello albgarse

ya all extensions are not working with 1.9.x releases so always make a note of that.

Regards

[albgarse]

I found a possible solution here:
http://mail.python.org/pipermail/pyt...il/001307.html

It uses a code like this to force a timeout:

import signal

def alarmHandler(*args):
    """
    signal handler for SIGALRM, just raise an exception
    """
    raise "TimeOut"

....
    signal.signal(signal.SIGALRM, alarmHandler)
    try:
        # set timeout
        signal.alarm(120)

        #... urllib.urlretrieve pages

    except "TimeOut":
        # some error handling
    signal.alarm(0)

But it seems like the signal module is not present in py60.

I'm trying with e32.Ao_Timer()

def timeoutHandler():
    raise IOError, "Timeout error"
    
try:
    timer = e32.Ao_timer()
    timer.after(120,timeoutHandler)
    # comms. function here...
    timer.cancel()
except:
    # capture exception here...

The timer calls timeoutHandler, and the exception is raised (I can see it in the python console), but not captured by the try. Seems like it's launched in a different thread.
is there a solution to raise the exception to the thread from where the timer was triggered?

regards.

[albgarse]

I've been making some more tryings. It's a bit rare.
The phone opens a communication every minute with urllib.urlopen() and it works fine. Some hours after, one communication fails. it don't return from urlopen() and every call to urlopen() after this fails in the same way.
There is not any memory problem. The free ram don't go down.

Is urlopen() function reentrant? so if one fails, could I make another urlopen() from other thread (have two urlopen() working at the same time)?
It seems a little rare that once one communication fails, every communication after fails too.

Regards.