Perl6::Slurp.3pm

Langue: en

Autres versions - même langue

Version: 2008-03-12 (debian - 07/07/09)

Section: 3 (Bibliothèques de fonctions)

NAME

Perl6::Slurp - Implements the Perl 6 'slurp' built-in

SYNOPSIS

     use Perl6::Slurp;
 
 
         # Slurp a file by name...
 
 
         $file_contents = slurp 'filename';
         $file_contents = slurp '<filename';
         $file_contents = slurp '<', 'filename';
         $file_contents = slurp '+<', 'filename';
 
 
         # Slurp a file via an (already open!) handle...
 
 
         $file_contents = slurp \*STDIN;
         $file_contents = slurp $filehandle;
         $file_contents = slurp IO::File->new('filename');
 
 
         # Slurp a string...
 
 
         $str_contents = slurp \$string;
         $str_contents = slurp '<', \$string;
 
 
         # Slurp a pipe...
 
 
         $str_contents = slurp 'tail -20 $filename |';
         $str_contents = slurp '-|', 'tail', -20, $filename;
 
 
         # Slurp with no source slurps from whatever $_ indicates...
 
 
         for (@files) {
                 $contents .= slurp;
         }
 
 
         # ...or from the entire ARGV list, if $_ is undefined...
 
 
         $_ = undef;
         $ARGV_contents = slurp;
 
 
         # Specify I/O layers as part of mode...
 
 
         $file_contents = slurp '<:raw', $file;
         $file_contents = slurp '<:utf8', $file;
         $file_contents = slurp '<:raw :utf8', $file;
 
 
         # Specify I/O layers as separate options...
 
 
         $file_contents = slurp $file, {raw=>1};
         $file_contents = slurp $file, {utf8=>1};
         $file_contents = slurp $file, {raw=>1}, {utf8=>1};
         $file_contents = slurp $file, [raw=>1, utf8=>1];
 
 
         # Specify input record separator...
 
 
         $file_contents = slurp $file, {irs=>"\n\n"};
         $file_contents = slurp '<', $file, {irs=>"\n\n"};
         $file_contents = slurp {irs=>"\n\n"}, $file;
 
 
         # Input record separator can be regex...
 
 
         $file_contents = slurp $file, {irs=>qr/\n+/};
         $file_contents = slurp '<', $file, {irs=>qr/\n+|\t{2,}};
 
 
         # Specify autochomping...
 
 
         $file_contents = slurp $file, {chomp=>1};
         $file_contents = slurp {chomp=>1}, $file;
         $file_contents = slurp $file, {chomp=>1, irs=>"\n\n"};
         $file_contents = slurp $file, {chomp=>1, irs=>qr/\n+/};
 
 
         # Specify autochomping that replaces irs
         # with another string...
 
 
         $file_contents = slurp $file, {irs=>"\n\n", chomp=>"\n"};
         $file_contents = slurp $file, {chomp=>"\n\n"}, {irs=>qr/\n+/};
 
 
         # Specify autochomping that replaces
         # irs with a dynamically computed string...
 
 
         my $n = 1;
         $file_contents = slurp $file, {chomp=>sub{ "\n#line ".$n++."\n"};
 
 
         # Slurp in a list context...
 
 
         @lines = slurp 'filename';
         @lines = slurp $filehandle;
         @lines = slurp \$string;
         @lines = slurp '<:utf8', 'filename', {irs=>"\x{2020}", chomp=>"\n"};
 
 

DESCRIPTION

"slurp" takes:
a filename,
a filehandle,
a typeglob reference,
an IO::File object, or
a scalar reference,

converts it to an input stream if necessary, and reads in the entire stream. If "slurp" fails to set up or read the stream, it throws an exception.

If no data source is specified "slurp" uses the value of $_ as the source. If $_ is undefined, "slurp" uses the @ARGV list, and magically slurps the contents of all the sources listed in @ARGV. Note that the same magic is also applied if you explicitly slurp <*ARGV>, so the following three input operations:

         $contents = join "", <ARGV>;
 
 
         $contents = slurp \*ARGV;
 
 
         $/ = undef;
         $contents = slurp;
 
 

are identical in effect.

In a scalar context "slurp" returns the stream contents as a single string. If the stream is at EOF, it returns an empty string. In a list context, it splits the contents after the appropriate input record separator and returns the resulting list of strings.

You can set the input record separator ("{ irs => $your_irs_here}") for the input operation. The separator can be specified as a string or a regex. Note that an explicit input record separator has no effect in a scalar context, since "slurp" always reads in everything anyway.

In a list context, changing the separator can change how the input is broken up within the list that is returned.

If an input record separator is not explicitly specified, "slurp" defaults to "\n" (not to the current value of $/ E<ndash> since Perl 6 doesn't have a $/);

You can also tell "slurp" to automagically "chomp" the input as it is read in, by specifying: ("{ chomp => 1 }")

Better still, you can tell "slurp" to automagically "chomp" the input and replace what it chomps with another string, by specifying: ("{ chomp => "another string" }")

You can also tell "slurp" to compute the replacement string on-the-fly by specifying a subroutine as the "chomp" value: ("{ chomp => sub{...} }"). This subroutine is passed the string being chomped off, so for example you could squeeze single newlines to a single space and multiple conseqcutive newlines to a two newlines with:

         sub squeeze { 
                 my ($removed) = @_;
                 if ($removed =~ tr/\n/\n/ == 1) { return " " }
                 else                            { return "\n\n"; }
         }
 
 
         print slurp(\*DATA, {irs=>qr/[ \t]*\n+/, chomp=>\&squeeze}), "\n";
 
 

Which would transform:

         This is the 
         first paragraph
 
 
         This is the 
         second
         paragraph
 
 
         This, the
         third
 
 
         This one is
         the
         very 
         last
 
 

to:

         This is the first paragraph
 
 
         This is the second paragraph
 
 
         This, the third
 
 
         This one is the very last
 
 

Autochomping works in both scalar and list contexts. In scalar contexts every instance of the input record separator will be removed (or replaced) within the returned string. In list context, each list item returned with its terminating separator removed (or replaced).

You can specify I/O layers, either using the Perl 5 notation:

     slurp "<:layer1 :layer2 :etc", $filename;
 
 

or as an array of options:

     slurp $filename, [layer1=>1, layer2=>1, etc=>1];
     slurp [layer1=>1, layer2=>1, etc=>1], $filename;
 
 

or as individual options (each of which must be in a separate hash):

     slurp $filename, {layer1=>1}, {layer2=>1}, {etc=>1};
     slurp {layer1=>1}, {layer2=>1}, {etc=>1}, $filename;
 
 

(...which, of course, would look much cooler in Perl 6:

     # Perl 6 only :-(
 
 
     slurp $filename, :layer1 :layer2 :etc;
     slurp :layer1 :layer2 :etc, $filename;
 
 

)

A common mistake is to put all the options together in one hash:

     slurp $filename, {layer1=>1, layer2=>1, etc=>1};
 
 

This is almost always a disaster, since the order of I/O layers is usually critical, and placing them all in one hash effectively randomizes that order. Use an array instead:

     slurp $filename, [layer1=>1, layer2=>1, etc=>1];
 
 

WARNING

The syntax and semantics of Perl 6 is still being finalized and consequently is at any time subject to change. That means the same caveat applies to this module.

DEPENDENCIES

Requires: Perl 5.8.0, Perl6::Export

AUTHOR

Damian Conway (damian@conway.org)
  Copyright (c) 2003, Damian Conway. All Rights Reserved.
  This module is free software. It may be used, redistributed
     and/or modified under the same terms as Perl itself.