Re: PerlIO::via silent failure

[p5sagit/p5-mst-13.2.git] / ext / PerlIO / via / via.pm

package PerlIO::via;
our $VERSION = '0.01';
use XSLoader ();
XSLoader::load 'PerlIO::via';
1;
__END__

=head1 NAME

PerlIO::via - Helper class for PerlIO layers implemented in perl

=head1 SYNOPSIS

   use PerlIO::via::Layer;
   open($fh,"<:via(Layer)",...);

   use Some::Other::Package;
   open($fh,">:via(Some::Other::Package)",...);

=head1 DESCRIPTION

The PerlIO::via module allows you to develop PerlIO layers in Perl, without
having to go into the nitty gritty of programming C with XS as the interface
to Perl.

One example module, L<PerlIO::via::QuotedPrint>, is included with Perl
5.8.0, and more example modules are available from CPAN, such as
L<PerlIO::via::StripHTML> and L<PerlIO::via::Base64>.  The
PerlIO::via::StripHTML module for instance, allows you to say:

	use PerlIO::via::StripHTML;
	open( my $fh, "<:via(StripHTML)", "index.html" );
        my @line = <$fh>;

to obtain the text of an HTML-file in an array with all the HTML-tags
automagically removed.

Please note that if the layer is created in the PerlIO::via:: namespace, it
does B<not> have to be fully qualified.  The PerlIO::via module will prefix
the PerlIO::via:: namespace if the specified modulename does not exist as a
fully qualified module name.

=head1 EXPECTED METHODS

To create a Perl module that implements a PerlIO layer in Perl (as opposed to
in C using XS as the interface to Perl), you need to supply some of the
following subroutines.  It is recommended to create these Perl modules in the
PerlIO::via:: namespace, so that they can easily be located on CPAN and use
the default namespace feature of the PerlIO::via module itself.

Please note that this is an area of recent development in Perl and that the
interface described here is therefore still subject to change (and hopefully
will have better documentation and more examples).

In the method descriptions below I<$fh> will be
a reference to a glob which can be treated as a perl file handle.
It refers to the layer below. I<$fh> is not passed if the layer
is at the bottom of the stack, for this reason and to maintain
some level of "compatibility" with TIEHANDLE classes it is passed last.
  
=over 4

=item $class->PUSHED([$mode[,$fh]])

Should return an object or the class, or -1 on failure.  (Compare
TIEHANDLE.)  The arguments are an optional mode string ("r", "w",
"w+", ...) and a filehandle for the PerlIO layer below.  Mandatory.

When layer is pushed as part of an C<open> call, C<PUSHED> will be called 
I<before> the actual open occurs whether than be via C<OPEN>, C<SYSOPEN>,
C<FDOPEN> or by letting lower layer do the open. 

=item $obj->POPPED([$fh])

Optional - layer is about to be removed.

=item $obj->OPEN($path,$mode[,$fh])

Optional - if not present lower layer does open.
If present called for normal opens after layer is pushed.
This function is subject to change as there is no easy way 
to get lower layer to do open and then regain control.

=item $obj->BINMODE([,$fh])

Optional - if not available layer is popped on binmode($fh) or when C<:raw>
is pushed. If present it should return 0 on success -1 on error and undef
to pop the layer.

=item $obj->FDOPEN($fd[,$fh])

Optional - if not present lower layer does open.
If present called for opens which pass a numeric file 
descriptor after layer is pushed. 
This function is subject to change as there is no easy way 
to get lower layer to do open and then regain control.

=item $obj->SYSOPEN($path,$imode,$perm,[,$fh])

Optional - if not present lower layer does open.
If present called for sysopen style opens which pass a numeric mode 
and permissions after layer is pushed.
This function is subject to change as there is no easy way 
to get lower layer to do open and then regain control.

=item $obj->FILENO($fh)

Returns a numeric value for Unix-like file descriptor. Return -1 if
there isn't one.  Optional.  Default is fileno($fh).

=item $obj->READ($buffer,$len,$fh)

Returns the number of octets placed in $buffer (must be less than or
equal to $len).  Optional.  Default is to use FILL instead.

=item $obj->WRITE($buffer,$fh)

Returns the number of octets from buffer that have been sucessfully written.

=item $obj->FILL($fh)

Should return a string to be placed in the buffer.  Optional. If not
provided must provide READ or reject handles open for reading in
PUSHED.

=item $obj->CLOSE($fh)

Should return 0 on success, -1 on error.
Optional.

=item $obj->SEEK($posn,$whence,$fh)

Should return 0 on success, -1 on error.
Optional.  Default is to fail, but that is likely to be changed
in future.

=item $obj->TELL($fh)

Returns file postion.
Optional.  Default to be determined.

=item $obj->UNREAD($buffer,$fh)

Returns the number of octets from buffer that have been sucessfully
saved to be returned on future FILL/READ calls.  Optional.  Default is
to push data into a temporary layer above this one.

=item $obj->FLUSH($fh)

Flush any buffered write data.  May possibly be called on readable
handles too.  Should return 0 on success, -1 on error.

=item $obj->SETLINEBUF($fh)

Optional. No return.

=item $obj->CLEARERR($fh)

Optional. No return.

=item $obj->ERROR($fh)

Optional. Returns error state. Default is no error until a mechanism
to signal error (die?) is worked out.

=item $obj->EOF($fh)

Optional. Returns end-of-file state. Default is function of return
value of FILL or READ.

=back

=head1 EXAMPLES

Check the PerlIO::via:: namespace on CPAN for examples of PerlIO layers
implemented in Perl.  To give you an idea how simple the implementation of
a PerlIO layer can look, as simple example is included here.

=head2 Example - a Hexadecimal Handle

Given the following module, PerlIO::via::Hex :

    package PerlIO::via::Hex;

    sub PUSHED
    {
     my ($class,$mode,$fh) = @_;
     # When writing we buffer the data
     my $buf = '';
     return bless \$buf,$class;
    }

    sub FILL
    {
     my ($obj,$fh) = @_;
     my $line = <$fh>;
     return (defined $line) ? pack("H*", $line) : undef;
    }

    sub WRITE
    {
     my ($obj,$buf,$fh) = @_;
     $$obj .= unpack("H*", $buf);
     return length($buf);
    }

    sub FLUSH
    {
     my ($obj,$fh) = @_;
     print $fh $$obj or return -1;
     $$obj = '';
     return 0;
    }

    1;

the following code opens up an output handle that will convert any
output to hexadecimal dump of the output bytes: for example "A" will
be converted to "41" (on ASCII-based machines, on EBCDIC platforms
the "A" will become "c1")

    use PerlIO::via::Hex;
    open(my $fh, ">:via(Hex)", "foo.hex");

and the following code will read the hexdump in and convert it
on the fly back into bytes:

    open(my $fh, "<:via(Hex)", "foo.hex");

=cut