Difference between can_read() and select() in IO::Select

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

Dear Monks,

I'm using IO::Select in order to switch between different clients connected to the same server program. At first, I used a sample code to understand how IO::Select works. That code uses the select() method in order to find the clients that have produced some output, like this:

$read_set = new IO::Select(); # create handle set for reading
$read_set->add($socket);      # add the main socket to the set
# Get a set of readable handles (blocks until at least one handle is r
+eady)
my ($rh_set) = IO::Select->select($read_set, undef, undef, 0);
[download]

Now, after reading the module's documentation, I'm trying to understand what's the difference between using can_read() and select(). Can the last line in the above extract be written as my @rh_set = $read_set->can_read(); ?

I can see that one method is static whole the other is not, and that one returns references to arrays, while the second returns one array - but are there any differences in the results themselves? Is one of them somehow "better" to use than the other?

Thanks a lot!

Comment on Difference between can_read() and select() in IO::Select Select or Download Code

Replies are listed 'Best First'.
Re: Difference between can_read() and select() in IO::Select by balakrishnan (Monk) on Feb 25, 2009 at 15:31 UTC
To get clear idea about the difference between these two function, you can better go through the following node, IO::Select: select vs. can_read	[reply]
Re^2: Difference between can_read() and select() in IO::Select by mr_mischief (Monsignor) on Feb 25, 2009 at 16:29 UTC
When you say `<a href="http://www.perlmonks.org/?node_id=255590">IO::Select: select vs. can_read</a>` you really mean `[id://255590]`. The former sends people to one of four hostnames through which PerlMonks is available, each with its own cookie domain. The latter will leave a monk logged in after following the link no matter which hostname they use.^* See Markup in the Monastery (which is linked near every node input box) for more information on markup on PerlMonks. Read more... (356 Bytes)	[reply] [d/l] [select]
Re: Difference between can_read() and select() in IO::Select by zentara (Cardinal) on Feb 25, 2009 at 15:39 UTC
From my perspective, the select object just holds some bits indicating the state of the connection(s). So your $read_set can add many sockets....a whole giant internal array of them. It can loop over them constantly to see which sockets are open, in a recv/send state, etc. There is also the accept condition, so you can monitor everytime someone wants to connect to the socket. . So you will see pseudo logic like: `while( my $sel = $select->can_read ){ if($sel->accept){spawn a new connection maybe by forking/threading or addi +ng to the select array } if( $sel->is_connected){ #check for connection status if ($sel->can_write){ etc } } }` [download] the problem comes with making the select work in a bi-directional manner, usually a forking server and or client is used so the socket can both send and recv simultaneously. If both client and server are set to either send or recv, it locks up the socket. I'm not really a human, but I play one on earth My Petition to the Great Cosmic Conciousness	[reply] [d/l]
Re: Difference between can_read() and select() in IO::Select by targetsmart (Curate) on Feb 25, 2009 at 16:42 UTC
I have also had confusions like this in the earlier stages, but now if I have confusion I will just open the module code and understand it. For example, if I want to know what can_read is doing in IO::Select, I will just do this on my system `my system details OS:Debian GNU/Linux 4.0 Kernel:2.6.18-4-686 GNU/Linux Perl: 5.8.8` [download] steps I will follow $ locate Select.pm /usr/lib/perl/5.8.8/IO/Select.pm /usr/share/perl/5.8.8/Pod/Select.pm /usr/share/perl5/Debconf/AutoSelect.pm /usr/share/perl5/Debconf/Element/Dialog/Select.pm /usr/share/perl5/Debconf/Element/Editor/Select.pm /usr/share/perl5/Debconf/Element/Gnome/Select.pm /usr/share/perl5/Debconf/Element/Kde/Select.pm /usr/share/perl5/Debconf/Element/Noninteractive/Select.pm /usr/share/perl5/Debconf/Element/Select.pm /usr/share/perl5/Debconf/Element/Teletype/Select.pm /usr/share/perl5/Debconf/Element/Web/Select.pm /usr/share/perl5/HTML/Widget/Element/Select.pm I will open the below module in vim editor $vim /usr/lib/perl/5.8.8/IO/Select.pm just search the pattern /b can_read I have got this sub can_read { my $vec = shift; my $timeout = shift; my $r = $vec->[VEC_BITS]; defined($r) && (select($r,undef,undef,$timeout) > 0) ? handles($vec, $r) : (); } after seeing this I will know what can_read can do. I have done this for lot of perl modules and I have learnt some effici +ent techniques in this process, I am still learning, this is just my +tips to you. [download] Vivek -- In accordance with the prarabdha of each, the One whose function it is to ordain makes each to act. What will not happen will never happen, whatever effort one may put forth. And what will happen will not fail to happen, however much one may seek to prevent it. This is certain. The part of wisdom therefore is to stay quiet.	[reply] [d/l] [select]
Re^2: Difference between can_read() and select() in IO::Select by chromatic (Archbishop) on Feb 25, 2009 at 19:24 UTC
If you don't mind using a pager instead of an editor, `perldoc -m IO::Select` is easier.	[reply] [d/l]
Re^2: Difference between can_read() and select() in IO::Select by zentara (Cardinal) on Feb 25, 2009 at 16:57 UTC
Sometimes reading the actual code is more confusing.:-) So you are trying to understand Perl's select, and you get code using select on vec? Simplifies things for sure. !! :-) I'm not really a human, but I play one on earth My Petition to the Great Cosmic Conciousness	[reply]
Re: Difference between can_read() and select() in IO::Select by Anonymous Monk on Feb 25, 2009 at 15:29 UTC
Can the last line in the above extract be written as `my @rh_set = $read_set->can_read();`? Yes. but are there any differences in the results themselves? Is one of them somehow "better" to use than the other? No difference in results.	[reply] [d/l]
Re^2: Difference between can_read() and select() in IO::Select by ikegami (Patriarch) on Feb 25, 2009 at 15:41 UTC
No, they're not the same. `my ($rh_set) = IO::Select->select($read_set, undef, undef, 0);` is actually equivalent to `my @rh_set = $read_set->can_read(0);` These are non-blocking. `my @rh_set = $read_set->can_read();` and `my @rh_set = $read_set->can_read(undef);` block just like `my ($rh_set) = IO::Select->select($read_set, undef, undef);` and `my ($rh_set) = IO::Select->select($read_set, undef, undef, undef);` To the OP, `can_read` is just a thin wrapper around `IO::Select::select` that saves you from specifying undefs and dealing with the more complex return value.	[reply] [d/l] [select]