open - perl pragma to set default PerlIO layers for input and output

    use open IN  => ":crlf", OUT => ":bytes";
    use open OUT => ':utf8';
    use open IO  => ":encoding(iso-8859-7)";

    use open IO  => ':locale';

    use open ':utf8';
    use open ':locale';
    use open ':encoding(iso-8859-7)';

    use open ':std';

Full-fledged support for I/O layers is now implemented provided Perl is configured to use PerlIO as its IO system (which is now the default).

The CWopen pragma serves as one of the interfaces to declare default layers (also known as disciplines) for all I/O. Any two-argument open(), readpipe() (aka qx//) and similar operators found within the lexical scope of this pragma will use the declared defaults. Three-argument opens are not affected by this pragma since there you (can) explicitly specify the layers and are supposed to know what you are doing.

With the CWIN subpragma you can declare the default layers of input streams, and with the CWOUT subpragma you can declare the default layers of output streams. With the CWIO subpragma you can control both input and output streams simultaneously.

If you have a legacy encoding, you can use the CW:encoding(...) tag.

if you want to set your encoding layers based on your locale environment variables, you can use the CW:locale tag. For example:

    $ENV{LANG} = 'ru_RU.KOI8-R';
    # the :locale will probe the locale environment variables like LANG
    use open OUT => ':locale';
    open(O, ">koi8");
    print O chr(0); # Unicode CYRILLIC SMALL LETTER A = KOI8-R 0xc1
    close O;
    open(I, "<koi8");
    printf "%#x\n", ord(<I>), "\n"; # this should print 0xc1
    close I;

These are equivalent

    use open ':utf8';
    use open IO => ':utf8';

as are these

    use open ':locale';
    use open IO => ':locale';

and these

    use open ':encoding(iso-8859-7)';
    use open IO => ':encoding(iso-8859-7)';

The matching of encoding names is loose: case does not matter, and many encodings have several aliases. See Encode::Supported for details and the list of supported locales.

Note that CW:utf8 PerlIO layer must always be specified exactly like that, it is not subject to the loose matching of encoding names.

When open() is given an explicit list of layers they are appended to the list declared using this pragma.

The CW:std subpragma on its own has no effect, but if combined with the CW:utf8 or CW:encoding subpragmas, it converts the standard filehandles (STDIN, STDOUT, STDERR) to comply with encoding selected for input/output handles. For example, if both input and out are chosen to be CW:utf8, a CW:std will mean that STDIN, STDOUT, and STDERR are also in CW:utf8. On the other hand, if only output is chosen to be in CW:encoding(koi8r), a CW:std will cause only the STDOUT and STDERR to be in CWkoi8r. The CW:locale subpragma implicitly turns on CW:std.

The logic of CW:locale is described in full in encoding, but in short it is first trying nl_langinfo(CODESET) and then guessing from the LC_ALL and LANG locale environment variables.

Directory handles may also support PerlIO layers in the future.

If Perl is not built to use PerlIO as its IO system then only the two pseudo-layers CW:bytes and CW:crlf are available.

The CW:bytes layer corresponds to binary mode and the CW:crlf layer corresponds to text mode on platforms that distinguish between the two modes when opening files (which is many DOS-like platforms, including Windows). These two layers are no-ops on platforms where binmode() is a no-op, but perform their functions everywhere if PerlIO is enabled.

There is a class method in CWPerlIO::Layer CWfind which is implemented as XS code. It is called by CWimport to validate the layers:

   PerlIO::Layer::->find("perlio")

The return value (if defined) is a Perl object, of class CWPerlIO::Layer which is created by the C code in perlio.c. As yet there is nothing useful you can do with the object at the perl level.

binmode in perlfunc, open in perlfunc, perlunicode, PerlIO, encoding