[Spread-users] prob. related to spread

[John Robinson]

Isn't Spread's Event subsystem designed to help users deal with this?

There are other packages that provide similar support.

/jr
---
John Schultz wrote:
> Pankaj,
> 
> Currently, the spread library ultimately is potentially blocking in both 
> SP_receive and SP_*cast.
> 
> You can mitigate this issue to some extent by using select() on the 
> mailbox, which is a file descriptor, to test for incoming data and when 
> there is room to write on the socket.  However, select() will signal 
> that the file descriptor is ready when there is *any* data to read, 
> which can cause SP_receive to block if more still has to come across the 
> client-daemon connection.  Similarly, select() will signal write 
> readiness the first time a single byte can be written to the file 
> descriptor without blocking.
> 
> You can further mitigate this problem by using either SP_poll or 
> ioctl(FIONREAD) to test the amount of data ready to be read in a 
> non-blocking manner and ioctl(FIONWRITE) (I think) to see how much can 
> be written in a non-blocking manner.  On the read side, you have no way 
> of knowing how big the next message will be so you could still block but 
> you can at least wait until there is a significant amount of data.  On 
> the write side, you should be able to wait until the file descriptor is 
> open enough to send your entire message and not block at all.  The 
> harder question in those scenerios is how do you prevent your select() 
> loop from spinning because from select()'s POV the file descriptor is 
> ready for action.  I don't have a good idea on that right this second.  
> Maybe someone else on the list can chime in with a good idea?
> 
> Good luck!
> John
> 
> ---
> John Schultz
> Spread Concepts
> Phn: 443 838 2200
> 
> On Fri, 8 Sep 2006, Zahir Koradia wrote:
> 
>> Hi Pankaj,
>> I am not very clear as to what you are looking for, but one thing that
>> did understand from your mail was that you dont want to block on
>> receive. There is function "SP_poll" which will allow you to poll your
>> mailbox if there are any mails in it. This is a non blocking call. So
>> once your "poll" says that you have something in you mailbox you can
>> go ahead and receive it. That way your receive will not block since
>> you know there is something waiting for you in the mailbox. I am not a
>> frequent user of spread and so am not aware if the send also blocks.
>> Logically speaking a send should not block.
>>
>> Hope this helps
>>
>> Zahir Koradia
>>
>> On 9/8/06, Pankaj Bhardwaj <pankajb83 at yahoo.co.in> wrote:
>>
>>> Hi Sir/Madam,
>>> >
>>> > I am having some prob. working with spread.
>>> >
>>> > I am using the spread server for communication.I have to develop
>>> > client code which can send and receive data anytime.
>>> >
>>> > My probs.
>>> > 1) I will be able to receive data and send that data
>>> > 2) I will be able to create data and send that data.
>>> >
>>> > but these two things will be able to work simultaneously.
>>> >
>>> > like receive is a blocking call , if at that time i want to send data
>>> > by creating myself i am able to do that.
>>> >
>>> > The prob i faced is that my program blocks in receive call. or in
>>> > send.call
>>> > until i do not send any data i am not able to receive any data.
>>> >
>>> > plz help us
>>> >
>>> >
>>>
>>>
>>>  ________________________________
>>>  Here's a new way to find what you're looking for - Yahoo! Answers
>>>  Send FREE SMS to your friend's mobile from Yahoo! Messenger Version 
>>> 8. Get
>>> it NOW
>>>
>>>
>>> _______________________________________________
>>> Spread-users mailing list
>>> Spread-users at lists.spread.org
>>> http://lists.spread.org/mailman/listinfo/spread-users
>>>
>>>
>>>
>>
>> _______________________________________________
>> Spread-users mailing list
>> Spread-users at lists.spread.org
>> http://lists.spread.org/mailman/listinfo/spread-users
>>
> 
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users