IO::Select can_read and windowing

Marcello has asked for the wisdom of the Perl Monks concerning the following question:

Hi,

I have a problem with my Perl server. Clients connect and send command terminated by a newline. The server uses IO::Select's can_read to determine if there's any data to read. If so, it reads all data up to the first newline.

The problem is when a client sends two or more commands at once. The server reads the first command, but after that can_read returns false although there is still data available. Why is can_read returning false when I only read a part of the total data (one command at a time)?

More info: it's a Linux machine, the perl server uses IO::Socket. Reading is done byte for byte, because commands can have different end characters (e.g. newline, ETX etc.)

Thanks for any help,

Marcel

Comment on IO::Select can_read and windowing

Replies are listed 'Best First'.
Re: IO::Select can_read and windowing by aquarium (Curate) on Jun 19, 2003 at 13:35 UTC
from memory...when you access can_read and get a true it then automatically resets to false until your next client sends data...so, just keep on reading until eof or whatever from the socket once you have can_read = true.	[reply]
Re: Re: IO::Select can_read and windowing by Marcello (Hermit) on Jun 19, 2003 at 14:25 UTC
Thanks aquarium, But the problem is that I need to use blocking reads, so if I have read one command and issue another read, the server blocks until the next command is completely read. Is there a way to tell how many bytes can be read? I then can use this to determine if I should do another read? Or maybe another IO::Select function to determine if I've read all the data? Kind regards	[reply]
Re^2: IO::Select can_read and windowing by Aristotle (Chancellor) on Jun 20, 2003 at 20:18 UTC
I quote `man 2 select`: `Three independent sets of descriptors are watched. Those listed in readfds will be watched to see if characters become available for reading (more precisely, to see if a read will not block - in particular, a file descriptor is also ready on end-of-file), those in writefds will be watched to see if a write will not block, and those in exceptfds will be watched for exceptions. On exit, the sets are modifiedin place to indicate which descriptors actually changed status.` That seems to indicate behaviour as Marcello expected. Makeshifts last the longest.	[reply]