Hi Python Gurus!
Here is a method I used before to receive data over a socket (with
Python 2.2):
SELECT_GRANULAR ITY = 0.1 # 0.1 seconds
def readdata(self,l ength,timeout):
res = ''
remain = length
lapsed = 0.0
while (remain > 0) and (lapsed < timeout):
i,o,e = select.select([self.socket],[],[],SELECT_GRANULA RITY)
lapsed = lapsed + SELECT_GRANULAR ITY
if i:
res = res + self.socket.rec v(remain)
remain = length - len(res)
if remain > 0:
raise TimeOutError
else:
return res
I used a very similar code to send data over a socket. Now in Python
2.3, we have
the wonderful settimeout() method. Here is my problem. Suppose that the
other
side (sender) is rather slow. I want to receive 1MB data in 20 seconds
(at most). Probably
the socket socket.recv() call will not return all data after the first
call. So my old method
is continuously trying to receive data until the timeout arrives (or it
will return sooner if all data
was received). This worked flawlessly with Python 2.2. With the new api,
it is possible to
use settimeout() instead of the select.select() call. However, recv()
will not return
all the 1MB data after the first call either. No matter what was the
timeout, it will give up
receiving when the input buffer becomes empty. So I still need to
receive data in a while loop.
What are the benefits and drawback of this solution with the new
settimeout() method?
I read the library reference documentation about compatibility. There is
not a word about
what platforms support the settimeout() method so I suppose all
platforms support this. As
opposed to select.select which is not available on every OS. There is a
note for the select module::
"This module provides access to the select() and poll() functions
available in most operating systems."
A proposal: It would be nice to have a recvall() method with a length
parameter. Suggestions,
comments are welcome.
Thank you in advance,
Laci 1.0
Here is a method I used before to receive data over a socket (with
Python 2.2):
SELECT_GRANULAR ITY = 0.1 # 0.1 seconds
def readdata(self,l ength,timeout):
res = ''
remain = length
lapsed = 0.0
while (remain > 0) and (lapsed < timeout):
i,o,e = select.select([self.socket],[],[],SELECT_GRANULA RITY)
lapsed = lapsed + SELECT_GRANULAR ITY
if i:
res = res + self.socket.rec v(remain)
remain = length - len(res)
if remain > 0:
raise TimeOutError
else:
return res
I used a very similar code to send data over a socket. Now in Python
2.3, we have
the wonderful settimeout() method. Here is my problem. Suppose that the
other
side (sender) is rather slow. I want to receive 1MB data in 20 seconds
(at most). Probably
the socket socket.recv() call will not return all data after the first
call. So my old method
is continuously trying to receive data until the timeout arrives (or it
will return sooner if all data
was received). This worked flawlessly with Python 2.2. With the new api,
it is possible to
use settimeout() instead of the select.select() call. However, recv()
will not return
all the 1MB data after the first call either. No matter what was the
timeout, it will give up
receiving when the input buffer becomes empty. So I still need to
receive data in a while loop.
What are the benefits and drawback of this solution with the new
settimeout() method?
I read the library reference documentation about compatibility. There is
not a word about
what platforms support the settimeout() method so I suppose all
platforms support this. As
opposed to select.select which is not available on every OS. There is a
note for the select module::
"This module provides access to the select() and poll() functions
available in most operating systems."
A proposal: It would be nice to have a recvall() method with a length
parameter. Suggestions,
comments are welcome.
Thank you in advance,
Laci 1.0