[DOCPATCH] perlfunc/read

[perl5.git] / pod / perlport.pod

=head1 NAME

perlport - Writing portable Perl

=head1 DESCRIPTION

Perl runs on numerous operating systems.  While most of them share
much in common, they also have their own unique features.

This document is meant to help you to find out what constitutes portable
Perl code.  That way once you make a decision to write portably,
you know where the lines are drawn, and you can stay within them.

There is a tradeoff between taking full advantage of one particular
type of computer and taking advantage of a full range of them.
Naturally, as you broaden your range and become more diverse, the
common factors drop, and you are left with an increasingly smaller
area of common ground in which you can operate to accomplish a
particular task.  Thus, when you begin attacking a problem, it is
important to consider under which part of the tradeoff curve you
want to operate.  Specifically, you must decide whether it is
important that the task that you are coding have the full generality
of being portable, or whether to just get the job done right now.
This is the hardest choice to be made.  The rest is easy, because
Perl provides many choices, whichever way you want to approach your
problem.

Looking at it another way, writing portable code is usually about
willfully limiting your available choices.  Naturally, it takes
discipline and sacrifice to do that.  The product of portability
and convenience may be a constant.  You have been warned.

Be aware of two important points:

=over 4

=item Not all Perl programs have to be portable

There is no reason you should not use Perl as a language to glue Unix
tools together, or to prototype a Macintosh application, or to manage the
Windows registry.  If it makes no sense to aim for portability for one
reason or another in a given program, then don't bother.

=item Nearly all of Perl already I<is> portable

Don't be fooled into thinking that it is hard to create portable Perl
code.  It isn't.  Perl tries its level-best to bridge the gaps between
what's available on different platforms, and all the means available to
use those features.  Thus almost all Perl code runs on any machine
without modification.  But there are some significant issues in
writing portable code, and this document is entirely about those issues.

=back

Here's the general rule: When you approach a task commonly done
using a whole range of platforms, think about writing portable
code.  That way, you don't sacrifice much by way of the implementation
choices you can avail yourself of, and at the same time you can give
your users lots of platform choices.  On the other hand, when you have to
take advantage of some unique feature of a particular platform, as is
often the case with systems programming (whether for Unix, Windows,
S<Mac OS>, VMS, etc.), consider writing platform-specific code.

When the code will run on only two or three operating systems, you
may need to consider only the differences of those particular systems.
The important thing is to decide where the code will run and to be
deliberate in your decision.

The material below is separated into three main sections: main issues of
portability (L<"ISSUES">, platform-specific issues (L<"PLATFORMS">, and
built-in perl functions that behave differently on various ports
(L<"FUNCTION IMPLEMENTATIONS">.

This information should not be considered complete; it includes possibly
transient information about idiosyncrasies of some of the ports, almost
all of which are in a state of constant evolution.  Thus, this material
should be considered a perpetual work in progress
(C<< <IMG SRC="yellow_sign.gif" ALT="Under Construction"> >>).

=head1 ISSUES

=head2 Newlines

In most operating systems, lines in files are terminated by newlines.
Just what is used as a newline may vary from OS to OS.  Unix
traditionally uses C<\012>, one type of DOSish I/O uses C<\015\012>,
and S<Mac OS> uses C<\015>.

Perl uses C<\n> to represent the "logical" newline, where what is
logical may depend on the platform in use.  In MacPerl, C<\n> always
means C<\015>.  In DOSish perls, C<\n> usually means C<\012>, but
when accessing a file in "text" mode, STDIO translates it to (or
from) C<\015\012>, depending on whether you're reading or writing.
Unix does the same thing on ttys in canonical mode.  C<\015\012>
is commonly referred to as CRLF.

A common cause of unportable programs is the misuse of chop() to trim
newlines:

    # XXX UNPORTABLE!
    while(<FILE>) {
        chop;
        @array = split(/:/);
        #...
    }

You can get away with this on Unix and Mac OS (they have a single
character end-of-line), but the same program will break under DOSish
perls because you're only chop()ing half the end-of-line.  Instead,
chomp() should be used to trim newlines.  The Dunce::Files module can
help audit your code for misuses of chop().

When dealing with binary files (or text files in binary mode) be sure
to explicitly set $/ to the appropriate value for your file format
before using chomp().

Because of the "text" mode translation, DOSish perls have limitations
in using C<seek> and C<tell> on a file accessed in "text" mode.
Stick to C<seek>-ing to locations you got from C<tell> (and no
others), and you are usually free to use C<seek> and C<tell> even
in "text" mode.  Using C<seek> or C<tell> or other file operations
may be non-portable.  If you use C<binmode> on a file, however, you
can usually C<seek> and C<tell> with arbitrary values in safety.

A common misconception in socket programming is that C<\n> eq C<\012>
everywhere.  When using protocols such as common Internet protocols,
C<\012> and C<\015> are called for specifically, and the values of
the logical C<\n> and C<\r> (carriage return) are not reliable.

    print SOCKET "Hi there, client!\r\n";      # WRONG
    print SOCKET "Hi there, client!\015\012";  # RIGHT

However, using C<\015\012> (or C<\cM\cJ>, or C<\x0D\x0A>) can be tedious
and unsightly, as well as confusing to those maintaining the code.  As
such, the Socket module supplies the Right Thing for those who want it.

    use Socket qw(:DEFAULT :crlf);
    print SOCKET "Hi there, client!$CRLF"      # RIGHT

When reading from a socket, remember that the default input record
separator C<$/> is C<\n>, but robust socket code will recognize as
either C<\012> or C<\015\012> as end of line:

    while (<SOCKET>) {
        # ...
    }

Because both CRLF and LF end in LF, the input record separator can
be set to LF and any CR stripped later.  Better to write:

    use Socket qw(:DEFAULT :crlf);
    local($/) = LF;      # not needed if $/ is already \012

    while (<SOCKET>) {
        s/$CR?$LF/\n/;   # not sure if socket uses LF or CRLF, OK
    #   s/\015?\012/\n/; # same thing
    }

This example is preferred over the previous one--even for Unix
platforms--because now any C<\015>'s (C<\cM>'s) are stripped out
(and there was much rejoicing).

Similarly, functions that return text data--such as a function that
fetches a web page--should sometimes translate newlines before
returning the data, if they've not yet been translated to the local
newline representation.  A single line of code will often suffice:

    $data =~ s/\015?\012/\n/g;
    return $data;

Some of this may be confusing.  Here's a handy reference to the ASCII CR
and LF characters.  You can print it out and stick it in your wallet.

    LF  eq  \012  eq  \x0A  eq  \cJ  eq  chr(10)  eq  ASCII 10
    CR  eq  \015  eq  \x0D  eq  \cM  eq  chr(13)  eq  ASCII 13

             | Unix | DOS  | Mac  |
        ---------------------------
        \n   |  LF  |  LF  |  CR  |
        \r   |  CR  |  CR  |  LF  |
        \n * |  LF  | CRLF |  CR  |
        \r * |  CR  |  CR  |  LF  |
        ---------------------------
        * text-mode STDIO

The Unix column assumes that you are not accessing a serial line
(like a tty) in canonical mode.  If you are, then CR on input becomes
"\n", and "\n" on output becomes CRLF.

These are just the most common definitions of C<\n> and C<\r> in Perl.
There may well be others.  For example, on an EBCDIC implementation
such as z/OS (OS/390) or OS/400 (using the ILE, the PASE is ASCII-based)
the above material is similar to "Unix" but the code numbers change:

    LF  eq  \025  eq  \x15  eq  \cU  eq  chr(21)  eq  CP-1047 21
    LF  eq  \045  eq  \x25  eq           chr(37)  eq  CP-0037 37
    CR  eq  \015  eq  \x0D  eq  \cM  eq  chr(13)  eq  CP-1047 13
    CR  eq  \015  eq  \x0D  eq  \cM  eq  chr(13)  eq  CP-0037 13

             | z/OS | OS/400 |
        ----------------------
        \n   |  LF  |  LF    |
        \r   |  CR  |  CR    |
        \n * |  LF  |  LF    |
        \r * |  CR  |  CR    |
        ----------------------
        * text-mode STDIO

=head2 Numbers endianness and Width

Different CPUs store integers and floating point numbers in different
orders (called I<endianness>) and widths (32-bit and 64-bit being the
most common today).  This affects your programs when they attempt to transfer
numbers in binary format from one CPU architecture to another,
usually either "live" via network connection, or by storing the
numbers to secondary storage such as a disk file or tape.

Conflicting storage orders make utter mess out of the numbers.  If a
little-endian host (Intel, VAX) stores 0x12345678 (305419896 in
decimal), a big-endian host (Motorola, Sparc, PA) reads it as
0x78563412 (2018915346 in decimal).  Alpha and MIPS can be either:
Digital/Compaq used/uses them in little-endian mode; SGI/Cray uses
them in big-endian mode.  To avoid this problem in network (socket)
connections use the C<pack> and C<unpack> formats C<n> and C<N>, the
"network" orders.  These are guaranteed to be portable.

You can explore the endianness of your platform by unpacking a
data structure packed in native format such as:

    print unpack("h*", pack("s2", 1, 2)), "\n";
    # '10002000' on e.g. Intel x86 or Alpha 21064 in little-endian mode
    # '00100020' on e.g. Motorola 68040

If you need to distinguish between endian architectures you could use
either of the variables set like so:

    $is_big_endian   = unpack("h*", pack("s", 1)) =~ /01/;
    $is_little_endian = unpack("h*", pack("s", 1)) =~ /^1/;

Differing widths can cause truncation even between platforms of equal
endianness.  The platform of shorter width loses the upper parts of the
number.  There is no good solution for this problem except to avoid
transferring or storing raw binary numbers.

One can circumnavigate both these problems in two ways.  Either
transfer and store numbers always in text format, instead of raw
binary, or else consider using modules like Data::Dumper (included in
the standard distribution as of Perl 5.005) and Storable (included as
of perl 5.8).  Keeping all data as text significantly simplifies matters.

The v-strings are portable only up to v2147483647 (0x7FFFFFFF), that's
how far EBCDIC, or more precisely UTF-EBCDIC will go.

=head2 Files and Filesystems

Most platforms these days structure files in a hierarchical fashion.
So, it is reasonably safe to assume that all platforms support the
notion of a "path" to uniquely identify a file on the system.  How
that path is really written, though, differs considerably.

Although similar, file path specifications differ between Unix,
Windows, S<Mac OS>, OS/2, VMS, VOS, S<RISC OS>, and probably others.
Unix, for example, is one of the few OSes that has the elegant idea
of a single root directory.

DOS, OS/2, VMS, VOS, and Windows can work similarly to Unix with C</>
as path separator, or in their own idiosyncratic ways (such as having
several root directories and various "unrooted" device files such NIL:
and LPT:).

S<Mac OS> uses C<:> as a path separator instead of C</>.

The filesystem may support neither hard links (C<link>) nor
symbolic links (C<symlink>, C<readlink>, C<lstat>).

The filesystem may support neither access timestamp nor change
timestamp (meaning that about the only portable timestamp is the
modification timestamp), or one second granularity of any timestamps
(e.g. the FAT filesystem limits the time granularity to two seconds).

The "inode change timestamp" (the C<-C> filetest) may really be the
"creation timestamp" (which it is not in UNIX).

VOS perl can emulate Unix filenames with C</> as path separator.  The
native pathname characters greater-than, less-than, number-sign, and
percent-sign are always accepted.

S<RISC OS> perl can emulate Unix filenames with C</> as path
separator, or go native and use C<.> for path separator and C<:> to
signal filesystems and disk names.

Don't assume UNIX filesystem access semantics: that read, write,
and execute are all the permissions there are, and even if they exist,
that their semantics (for example what do r, w, and x mean on
a directory) are the UNIX ones.  The various UNIX/POSIX compatibility
layers usually try to make interfaces like chmod() work, but sometimes
there simply is no good mapping.

If all this is intimidating, have no (well, maybe only a little)
fear.  There are modules that can help.  The File::Spec modules
provide methods to do the Right Thing on whatever platform happens
to be running the program.

    use File::Spec::Functions;
    chdir(updir());        # go up one directory
    $file = catfile(curdir(), 'temp', 'file.txt');
    # on Unix and Win32, './temp/file.txt'
    # on Mac OS, ':temp:file.txt'
    # on VMS, '[.temp]file.txt'

File::Spec is available in the standard distribution as of version
5.004_05.  File::Spec::Functions is only in File::Spec 0.7 and later,
and some versions of perl come with version 0.6.  If File::Spec
is not updated to 0.7 or later, you must use the object-oriented
interface from File::Spec (or upgrade File::Spec).

In general, production code should not have file paths hardcoded.
Making them user-supplied or read from a configuration file is
better, keeping in mind that file path syntax varies on different
machines.

This is especially noticeable in scripts like Makefiles and test suites,
which often assume C</> as a path separator for subdirectories.

Also of use is File::Basename from the standard distribution, which
splits a pathname into pieces (base filename, full path to directory,
and file suffix).

Even when on a single platform (if you can call Unix a single platform),
remember not to count on the existence or the contents of particular
system-specific files or directories, like F</etc/passwd>,
F</etc/sendmail.conf>, F</etc/resolv.conf>, or even F</tmp/>.  For
example, F</etc/passwd> may exist but not contain the encrypted
passwords, because the system is using some form of enhanced security.
Or it may not contain all the accounts, because the system is using NIS. 
If code does need to rely on such a file, include a description of the
file and its format in the code's documentation, then make it easy for
the user to override the default location of the file.

Don't assume a text file will end with a newline.  They should,
but people forget.

Do not have two files or directories of the same name with different
case, like F<test.pl> and F<Test.pl>, as many platforms have
case-insensitive (or at least case-forgiving) filenames.  Also, try
not to have non-word characters (except for C<.>) in the names, and
keep them to the 8.3 convention, for maximum portability, onerous a
burden though this may appear.

Likewise, when using the AutoSplit module, try to keep your functions to
8.3 naming and case-insensitive conventions; or, at the least,
make it so the resulting files have a unique (case-insensitively)
first 8 characters.

Whitespace in filenames is tolerated on most systems, but not all,
and even on systems where it might be tolerated, some utilities
might become confused by such whitespace.

Many systems (DOS, VMS) cannot have more than one C<.> in their filenames.

Don't assume C<< > >> won't be the first character of a filename.
Always use C<< < >> explicitly to open a file for reading, or even
better, use the three-arg version of open, unless you want the user to
be able to specify a pipe open.

    open(FILE, '<', $existing_file) or die $!;

If filenames might use strange characters, it is safest to open it
with C<sysopen> instead of C<open>.  C<open> is magic and can
translate characters like C<< > >>, C<< < >>, and C<|>, which may
be the wrong thing to do.  (Sometimes, though, it's the right thing.)
Three-arg open can also help protect against this translation in cases
where it is undesirable.

Don't use C<:> as a part of a filename since many systems use that for
their own semantics (Mac OS Classic for separating pathname components,
many networking schemes and utilities for separating the nodename and
the pathname, and so on).  For the same reasons, avoid C<@>, C<;> and
C<|>.

Don't assume that in pathnames you can collapse two leading slashes
C<//> into one: some networking and clustering filesystems have special
semantics for that.  Let the operating system to sort it out.

The I<portable filename characters> as defined by ANSI C are

 a b c d e f g h i j k l m n o p q r t u v w x y z
 A B C D E F G H I J K L M N O P Q R T U V W X Y Z
 0 1 2 3 4 5 6 7 8 9
 . _ -

and the "-" shouldn't be the first character.  If you want to be
hypercorrect, stay case-insensitive and within the 8.3 naming
convention (all the files and directories have to be unique within one
directory if their names are lowercased and truncated to eight
characters before the C<.>, if any, and to three characters after the
C<.>, if any).  (And do not use C<.>s in directory names.)

=head2 System Interaction

Not all platforms provide a command line.  These are usually platforms
that rely primarily on a Graphical User Interface (GUI) for user
interaction.  A program requiring a command line interface might
not work everywhere.  This is probably for the user of the program
to deal with, so don't stay up late worrying about it.

Some platforms can't delete or rename files held open by the system,
this limitation may also apply to changing filesystem metainformation
like file permissions or owners.  Remember to C<close> files when you
are done with them.  Don't C<unlink> or C<rename> an open file.  Don't
C<tie> or C<open> a file already tied or opened; C<untie> or C<close>
it first.

Don't open the same file more than once at a time for writing, as some
operating systems put mandatory locks on such files.

Don't assume that write/modify permission on a directory gives the
right to add or delete files/directories in that directory.  That is
filesystem specific: in some filesystems you need write/modify
permission also (or even just) in the file/directory itself.  In some
filesystems (AFS, DFS) the permission to add/delete directory entries
is a completely separate permission.

Don't assume that a single C<unlink> completely gets rid of the file:
some filesystems (most notably the ones in VMS) have versioned
filesystems, and unlink() removes only the most recent one (it doesn't
remove all the versions because by default the native tools on those
platforms remove just the most recent version, too).  The portable
idiom to remove all the versions of a file is

    1 while unlink "file";

This will terminate if the file is undeleteable for some reason
(protected, not there, and so on).

Don't count on a specific environment variable existing in C<%ENV>.
Don't count on C<%ENV> entries being case-sensitive, or even
case-preserving.  Don't try to clear %ENV by saying C<%ENV = ();>, or,
if you really have to, make it conditional on C<$^O ne 'VMS'> since in
VMS the C<%ENV> table is much more than a per-process key-value string
table.

Don't count on signals or C<%SIG> for anything.

Don't count on filename globbing.  Use C<opendir>, C<readdir>, and
C<closedir> instead.

Don't count on per-program environment variables, or per-program current
directories.

Don't count on specific values of C<$!>, neither numeric nor
especially the strings values-- users may switch their locales causing
error messages to be translated into their languages.  If you can
trust a POSIXish environment, you can portably use the symbols defined
by the Errno module, like ENOENT.  And don't trust on the values of C<$!>
at all except immediately after a failed system call.

=head2 Command names versus file pathnames

Don't assume that the name used to invoke a command or program with
C<system> or C<exec> can also be used to test for the existence of the
file that holds the executable code for that command or program.
First, many systems have "internal" commands that are built-in to the
shell or OS and while these commands can be invoked, there is no
corresponding file.  Second, some operating systems (e.g., Cygwin,
DJGPP, OS/2, and VOS) have required suffixes for executable files;
these suffixes are generally permitted on the command name but are not
required.  Thus, a command like "perl" might exist in a file named
"perl", "perl.exe", or "perl.pm", depending on the operating system.
The variable "_exe" in the Config module holds the executable suffix,
if any.  Third, the VMS port carefully sets up $^X and
$Config{perlpath} so that no further processing is required.  This is
just as well, because the matching regular expression used below would
then have to deal with a possible trailing version number in the VMS
file name.

To convert $^X to a file pathname, taking account of the requirements
of the various operating system possibilities, say:
  use Config;
  $thisperl = $^X;
  if ($^O ne 'VMS')
     {$thisperl .= $Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;}

To convert $Config{perlpath} to a file pathname, say:
  use Config;
  $thisperl = $Config{perlpath};
  if ($^O ne 'VMS')
     {$thisperl .= $Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;}

=head2 Networking

Don't assume that you can reach the public Internet.

Don't assume that there is only one way to get through firewalls
to the public Internet.

Don't assume that you can reach yourself or any node by the name
'localhost'.  The same goes for '127.0.0.1'.

Don't assume that any particular port (service) will respond.

Don't assume that you can ping hosts and get replies.

=head2 Interprocess Communication (IPC)

In general, don't directly access the system in code meant to be
portable.  That means, no C<system>, C<exec>, C<fork>, C<pipe>,
C<``>, C<qx//>, C<open> with a C<|>, nor any of the other things
that makes being a perl hacker worth being.

Commands that launch external processes are generally supported on
most platforms (though many of them do not support any type of
forking).  The problem with using them arises from what you invoke
them on.  External tools are often named differently on different
platforms, may not be available in the same location, might accept
different arguments, can behave differently, and often present their
results in a platform-dependent way.  Thus, you should seldom depend
on them to produce consistent results. (Then again, if you're calling 
I<netstat -a>, you probably don't expect it to run on both Unix and CP/M.)

One especially common bit of Perl code is opening a pipe to B<sendmail>:

    open(MAIL, '|/usr/lib/sendmail -t') 
        or die "cannot fork sendmail: $!";

This is fine for systems programming when sendmail is known to be
available.  But it is not fine for many non-Unix systems, and even
some Unix systems that may not have sendmail installed.  If a portable
solution is needed, see the various distributions on CPAN that deal
with it.  Mail::Mailer and Mail::Send in the MailTools distribution are
commonly used, and provide several mailing methods, including mail,
sendmail, and direct SMTP (via Net::SMTP) if a mail transfer agent is
not available.  Mail::Sendmail is a standalone module that provides
simple, platform-independent mailing.

The Unix System V IPC (C<msg*(), sem*(), shm*()>) is not available
even on all Unix platforms.

Do not use either the bare result of C<pack("N", 10, 20, 30, 40)> or
bare v-strings (such as C<v10.20.30.40>) to represent IPv4 addresses:
both forms just pack the four bytes into network order.  That this
would be equal to the C language C<in_addr> struct (which is what the
socket code internally uses) is not guaranteed.  To be portable use
the routines of the Socket extension, such as C<inet_aton()>,
C<inet_ntoa()>, and C<sockaddr_in()>.

The rule of thumb for portable code is: Do it all in portable Perl, or
use a module (that may internally implement it with platform-specific
code, but expose a common interface).

=head2 External Subroutines (XS)

XS code can usually be made to work with any platform, but dependent
libraries, header files, etc., might not be readily available or
portable, or the XS code itself might be platform-specific, just as Perl
code might be.  If the libraries and headers are portable, then it is
normally reasonable to make sure the XS code is portable, too.

A different type of portability issue arises when writing XS code:
availability of a C compiler on the end-user's system.  C brings
with it its own portability issues, and writing XS code will expose
you to some of those.  Writing purely in Perl is an easier way to
achieve portability.

=head2 Standard Modules

In general, the standard modules work across platforms.  Notable
exceptions are the CPAN module (which currently makes connections to external
programs that may not be available), platform-specific modules (like
ExtUtils::MM_VMS), and DBM modules.

There is no one DBM module available on all platforms.
SDBM_File and the others are generally available on all Unix and DOSish
ports, but not in MacPerl, where only NBDM_File and DB_File are
available.

The good news is that at least some DBM module should be available, and
AnyDBM_File will use whichever module it can find.  Of course, then
the code needs to be fairly strict, dropping to the greatest common
factor (e.g., not exceeding 1K for each record), so that it will
work with any DBM module.  See L<AnyDBM_File> for more details.

=head2 Time and Date

The system's notion of time of day and calendar date is controlled in
widely different ways.  Don't assume the timezone is stored in C<$ENV{TZ}>,
and even if it is, don't assume that you can control the timezone through
that variable.  Don't assume anything about the three-letter timezone
abbreviations (for example that MST would be the Mountain Standard Time,
it's been known to stand for Moscow Standard Time).  If you need to
use timezones, express them in some unambiguous format like the
exact number of minutes offset from UTC, or the POSIX timezone
format.

Don't assume that the epoch starts at 00:00:00, January 1, 1970,
because that is OS- and implementation-specific.  It is better to
store a date in an unambiguous representation.  The ISO 8601 standard
defines YYYY-MM-DD as the date format, or YYYY-MM-DDTHH-MM-SS
(that's a literal "T" separating the date from the time).
Please do use the ISO 8601 instead of making us to guess what
date 02/03/04 might be.  ISO 8601 even sorts nicely as-is.
A text representation (like "1987-12-18") can be easily converted
into an OS-specific value using a module like Date::Parse.
An array of values, such as those returned by C<localtime>, can be
converted to an OS-specific representation using Time::Local.

When calculating specific times, such as for tests in time or date modules,
it may be appropriate to calculate an offset for the epoch.

    require Time::Local;
    $offset = Time::Local::timegm(0, 0, 0, 1, 0, 70);

The value for C<$offset> in Unix will be C<0>, but in Mac OS will be
some large number.  C<$offset> can then be added to a Unix time value
to get what should be the proper value on any system.

On Windows (at least), you shouldn't pass a negative value to C<gmtime> or
C<localtime>.

=head2 Character sets and character encoding

Assume very little about character sets.

Assume nothing about numerical values (C<ord>, C<chr>) of characters.
Do not use explicit code point ranges (like \xHH-\xHH); use for
example symbolic character classes like C<[:print:]>.

Do not assume that the alphabetic characters are encoded contiguously
(in the numeric sense).  There may be gaps.

Do not assume anything about the ordering of the characters.
The lowercase letters may come before or after the uppercase letters;
the lowercase and uppercase may be interlaced so that both `a' and `A'
come before `b'; the accented and other international characters may
be interlaced so that E<auml> comes before `b'.

=head2 Internationalisation

If you may assume POSIX (a rather large assumption), you may read
more about the POSIX locale system from L<perllocale>.  The locale
system at least attempts to make things a little bit more portable,
or at least more convenient and native-friendly for non-English
users.  The system affects character sets and encoding, and date
and time formatting--amongst other things.

If you really want to be international, you should consider Unicode.
See L<perluniintro> and L<perlunicode> for more information.

If you want to use non-ASCII bytes (outside the bytes 0x00..0x7f) in
the "source code" of your code, to be portable you have to be explicit
about what bytes they are.  Someone might for example be using your
code under a UTF-8 locale, in which case random native bytes might be
illegal ("Malformed UTF-8 ...")  This means that for example embedding
ISO 8859-1 bytes beyond 0x7f into your strings might cause trouble
later.  If the bytes are native 8-bit bytes, you can use the C<bytes>
pragma.  If the bytes are in a string (regular expression being a
curious string), you can often also use the C<\xHH> notation instead
of embedding the bytes as-is.  If they are in some particular legacy
encoding (ether single-byte or something more complicated), you can
use the C<encoding> pragma.  (If you want to write your code in UTF-8,
you can use either the C<utf8> pragma, or the C<encoding> pragma.)
The C<bytes> and C<utf8> pragmata are available since Perl 5.6.0, and
the C<encoding> pragma since Perl 5.8.0.

=head2 System Resources

If your code is destined for systems with severely constrained (or
missing!) virtual memory systems then you want to be I<especially> mindful
of avoiding wasteful constructs such as:

    # NOTE: this is no longer "bad" in perl5.005
    for (0..10000000) {}                       # bad
    for (my $x = 0; $x <= 10000000; ++$x) {}   # good

    @lines = <VERY_LARGE_FILE>;                # bad

    while (<FILE>) {$file .= $_}               # sometimes bad
    $file = join('', <FILE>);                  # better

The last two constructs may appear unintuitive to most people.  The
first repeatedly grows a string, whereas the second allocates a
large chunk of memory in one go.  On some systems, the second is
more efficient that the first.

=head2 Security

Most multi-user platforms provide basic levels of security, usually
implemented at the filesystem level.  Some, however, do
not-- unfortunately.  Thus the notion of user id, or "home" directory,
or even the state of being logged-in, may be unrecognizable on many
platforms.  If you write programs that are security-conscious, it
is usually best to know what type of system you will be running
under so that you can write code explicitly for that platform (or
class of platforms).

Don't assume the UNIX filesystem access semantics: the operating
system or the filesystem may be using some ACL systems, which are
richer languages than the usual rwx.  Even if the rwx exist,
their semantics might be different.

(From security viewpoint testing for permissions before attempting to
do something is silly anyway: if one tries this, there is potential
for race conditions-- someone or something might change the
permissions between the permissions check and the actual operation.
Just try the operation.)

Don't assume the UNIX user and group semantics: especially, don't
expect the C<< $< >> and C<< $> >> (or the C<$(> and C<$)>) to work
for switching identities (or memberships).

Don't assume set-uid and set-gid semantics. (And even if you do,
think twice: set-uid and set-gid are a known can of security worms.)

=head2 Style

For those times when it is necessary to have platform-specific code,
consider keeping the platform-specific code in one place, making porting
to other platforms easier.  Use the Config module and the special
variable C<$^O> to differentiate platforms, as described in
L<"PLATFORMS">.

Be careful in the tests you supply with your module or programs.
Module code may be fully portable, but its tests might not be.  This
often happens when tests spawn off other processes or call external
programs to aid in the testing, or when (as noted above) the tests
assume certain things about the filesystem and paths.  Be careful not
to depend on a specific output style for errors, such as when checking
C<$!> after a failed system call.  Using C<$!> for anything else than
displaying it as output is doubtful (though see the Errno module for
testing reasonably portably for error value). Some platforms expect
a certain output format, and Perl on those platforms may have been
adjusted accordingly.  Most specifically, don't anchor a regex when
testing an error value.

=head1 CPAN Testers

Modules uploaded to CPAN are tested by a variety of volunteers on
different platforms.  These CPAN testers are notified by mail of each
new upload, and reply to the list with PASS, FAIL, NA (not applicable to
this platform), or UNKNOWN (unknown), along with any relevant notations.

The purpose of the testing is twofold: one, to help developers fix any
problems in their code that crop up because of lack of testing on other
platforms; two, to provide users with information about whether
a given module works on a given platform.

=over 4

=item Mailing list: cpan-testers@perl.org

=item Testing results: http://testers.cpan.org/

=back

=head1 PLATFORMS

As of version 5.002, Perl is built with a C<$^O> variable that
indicates the operating system it was built on.  This was implemented
to help speed up code that would otherwise have to C<use Config>
and use the value of C<$Config{osname}>.  Of course, to get more
detailed information about the system, looking into C<%Config> is
certainly recommended.

C<%Config> cannot always be trusted, however, because it was built
at compile time.  If perl was built in one place, then transferred
elsewhere, some values may be wrong.  The values may even have been
edited after the fact.

=head2 Unix

Perl works on a bewildering variety of Unix and Unix-like platforms (see
e.g. most of the files in the F<hints/> directory in the source code kit).
On most of these systems, the value of C<$^O> (hence C<$Config{'osname'}>,
too) is determined either by lowercasing and stripping punctuation from the
first field of the string returned by typing C<uname -a> (or a similar command)
at the shell prompt or by testing the file system for the presence of
uniquely named files such as a kernel or header file.  Here, for example,
are a few of the more popular Unix flavors:

    uname         $^O        $Config{'archname'}
    --------------------------------------------
    AIX           aix        aix
    BSD/OS        bsdos      i386-bsdos
    Darwin        darwin     darwin
    dgux          dgux       AViiON-dgux
    DYNIX/ptx     dynixptx   i386-dynixptx
    FreeBSD       freebsd    freebsd-i386    
    Linux         linux      arm-linux
    Linux         linux      i386-linux
    Linux         linux      i586-linux
    Linux         linux      ppc-linux
    HP-UX         hpux       PA-RISC1.1
    IRIX          irix       irix
    Mac OS X      darwin     darwin
    MachTen PPC   machten    powerpc-machten
    NeXT 3        next       next-fat
    NeXT 4        next       OPENSTEP-Mach
    openbsd       openbsd    i386-openbsd
    OSF1          dec_osf    alpha-dec_osf
    reliantunix-n svr4       RM400-svr4
    SCO_SV        sco_sv     i386-sco_sv
    SINIX-N       svr4       RM400-svr4
    sn4609        unicos     CRAY_C90-unicos
    sn6521        unicosmk   t3e-unicosmk
    sn9617        unicos     CRAY_J90-unicos
    SunOS         solaris    sun4-solaris
    SunOS         solaris    i86pc-solaris
    SunOS4        sunos      sun4-sunos

Because the value of C<$Config{archname}> may depend on the
hardware architecture, it can vary more than the value of C<$^O>.

=head2 DOS and Derivatives

Perl has long been ported to Intel-style microcomputers running under
systems like PC-DOS, MS-DOS, OS/2, and most Windows platforms you can
bring yourself to mention (except for Windows CE, if you count that).
Users familiar with I<COMMAND.COM> or I<CMD.EXE> style shells should
be aware that each of these file specifications may have subtle
differences:

    $filespec0 = "c:/foo/bar/file.txt";
    $filespec1 = "c:\\foo\\bar\\file.txt";
    $filespec2 = 'c:\foo\bar\file.txt';
    $filespec3 = 'c:\\foo\\bar\\file.txt';

System calls accept either C</> or C<\> as the path separator.
However, many command-line utilities of DOS vintage treat C</> as
the option prefix, so may get confused by filenames containing C</>.
Aside from calling any external programs, C</> will work just fine,
and probably better, as it is more consistent with popular usage,
and avoids the problem of remembering what to backwhack and what
not to.

The DOS FAT filesystem can accommodate only "8.3" style filenames.  Under
the "case-insensitive, but case-preserving" HPFS (OS/2) and NTFS (NT)
filesystems you may have to be careful about case returned with functions
like C<readdir> or used with functions like C<open> or C<opendir>.

DOS also treats several filenames as special, such as AUX, PRN,
NUL, CON, COM1, LPT1, LPT2, etc.  Unfortunately, sometimes these
filenames won't even work if you include an explicit directory
prefix.  It is best to avoid such filenames, if you want your code
to be portable to DOS and its derivatives.  It's hard to know what
these all are, unfortunately.

Users of these operating systems may also wish to make use of
scripts such as I<pl2bat.bat> or I<pl2cmd> to
put wrappers around your scripts.

Newline (C<\n>) is translated as C<\015\012> by STDIO when reading from
and writing to files (see L<"Newlines">).  C<binmode(FILEHANDLE)>
will keep C<\n> translated as C<\012> for that filehandle.  Since it is a
no-op on other systems, C<binmode> should be used for cross-platform code
that deals with binary data.  That's assuming you realize in advance
that your data is in binary.  General-purpose programs should
often assume nothing about their data.

The C<$^O> variable and the C<$Config{archname}> values for various
DOSish perls are as follows:

     OS            $^O      $Config{archname}   ID    Version
     --------------------------------------------------------
     MS-DOS        dos        ?                 
     PC-DOS        dos        ?                 
     OS/2          os2        ?
     Windows 3.1   ?          ?                 0      3 01
     Windows 95    MSWin32    MSWin32-x86       1      4 00
     Windows 98    MSWin32    MSWin32-x86       1      4 10
     Windows ME    MSWin32    MSWin32-x86       1      ?
     Windows NT    MSWin32    MSWin32-x86       2      4 xx
     Windows NT    MSWin32    MSWin32-ALPHA     2      4 xx
     Windows NT    MSWin32    MSWin32-ppc       2      4 xx
     Windows 2000  MSWin32    MSWin32-x86       2      5 xx
     Windows XP    MSWin32    MSWin32-x86       2      ?
     Windows CE    MSWin32    ?                 3           
     Cygwin        cygwin     ?                 

The various MSWin32 Perl's can distinguish the OS they are running on
via the value of the fifth element of the list returned from 
Win32::GetOSVersion().  For example:

    if ($^O eq 'MSWin32') {
        my @os_version_info = Win32::GetOSVersion();
        print +('3.1','95','NT')[$os_version_info[4]],"\n";
    }

There are also Win32::IsWinNT() and Win32::IsWin95(), try C<perldoc Win32>,
and as of libwin32 0.19 (not part of the core Perl distribution)
Win32::GetOSName().  The very portable POSIX::uname() will work too:

    c:\> perl -MPOSIX -we "print join '|', uname"
    Windows NT|moonru|5.0|Build 2195 (Service Pack 2)|x86

Also see:

=over 4

=item *

The djgpp environment for DOS, http://www.delorie.com/djgpp/
and L<perldos>.

=item *

The EMX environment for DOS, OS/2, etc. emx@iaehv.nl,
http://www.leo.org/pub/comp/os/os2/leo/gnu/emx+gcc/index.html or
ftp://hobbes.nmsu.edu/pub/os2/dev/emx/  Also L<perlos2>.

=item *

Build instructions for Win32 in L<perlwin32>, or under the Cygnus environment
in L<perlcygwin>.  

=item *

The C<Win32::*> modules in L<Win32>.

=item *

The ActiveState Pages, http://www.activestate.com/

=item *

The Cygwin environment for Win32; F<README.cygwin> (installed 
as L<perlcygwin>), http://www.cygwin.com/

=item *

The U/WIN environment for Win32,
http://www.research.att.com/sw/tools/uwin/

=item *

Build instructions for OS/2, L<perlos2>

=back

=head2 S<Mac OS>

Any module requiring XS compilation is right out for most people, because
MacPerl is built using non-free (and non-cheap!) compilers.  Some XS
modules that can work with MacPerl are built and distributed in binary
form on CPAN.

Directories are specified as:

    volume:folder:file              for absolute pathnames
    volume:folder:                  for absolute pathnames
    :folder:file                    for relative pathnames
    :folder:                        for relative pathnames
    :file                           for relative pathnames
    file                            for relative pathnames

Files are stored in the directory in alphabetical order.  Filenames are
limited to 31 characters, and may include any character except for
null and C<:>, which is reserved as the path separator.

Instead of C<flock>, see C<FSpSetFLock> and C<FSpRstFLock> in the
Mac::Files module, or C<chmod(0444, ...)> and C<chmod(0666, ...)>.

In the MacPerl application, you can't run a program from the command line;
programs that expect C<@ARGV> to be populated can be edited with something
like the following, which brings up a dialog box asking for the command
line arguments.

    if (!@ARGV) {
        @ARGV = split /\s+/, MacPerl::Ask('Arguments?');
    }

A MacPerl script saved as a "droplet" will populate C<@ARGV> with the full
pathnames of the files dropped onto the script.

Mac users can run programs under a type of command line interface
under MPW (Macintosh Programmer's Workshop, a free development
environment from Apple).  MacPerl was first introduced as an MPW
tool, and MPW can be used like a shell:

    perl myscript.plx some arguments

ToolServer is another app from Apple that provides access to MPW tools
from MPW and the MacPerl app, which allows MacPerl programs to use
C<system>, backticks, and piped C<open>.

"S<Mac OS>" is the proper name for the operating system, but the value
in C<$^O> is "MacOS".  To determine architecture, version, or whether
the application or MPW tool version is running, check:

    $is_app    = $MacPerl::Version =~ /App/;
    $is_tool   = $MacPerl::Version =~ /MPW/;
    ($version) = $MacPerl::Version =~ /^(\S+)/;
    $is_ppc    = $MacPerl::Architecture eq 'MacPPC';
    $is_68k    = $MacPerl::Architecture eq 'Mac68K';

S<Mac OS X>, based on NeXT's OpenStep OS, runs MacPerl natively, under the
"Classic" environment.  There is no "Carbon" version of MacPerl to run
under the primary Mac OS X environment.  S<Mac OS X> and its Open Source
version, Darwin, both run Unix perl natively.

Also see:

=over 4

=item *

MacPerl Development, http://dev.macperl.org/ .

=item *

The MacPerl Pages, http://www.macperl.com/ .

=item *

The MacPerl mailing lists, http://lists.perl.org/ .

=back

=head2 VMS

Perl on VMS is discussed in L<perlvms> in the perl distribution.
Perl on VMS can accept either VMS- or Unix-style file
specifications as in either of the following:

    $ perl -ne "print if /perl_setup/i" SYS$LOGIN:LOGIN.COM
    $ perl -ne "print if /perl_setup/i" /sys$login/login.com

but not a mixture of both as in:

    $ perl -ne "print if /perl_setup/i" sys$login:/login.com
    Can't open sys$login:/login.com: file specification syntax error

Interacting with Perl from the Digital Command Language (DCL) shell
often requires a different set of quotation marks than Unix shells do.
For example:

    $ perl -e "print ""Hello, world.\n"""
    Hello, world.

There are several ways to wrap your perl scripts in DCL F<.COM> files, if
you are so inclined.  For example:

    $ write sys$output "Hello from DCL!"
    $ if p1 .eqs. ""
    $ then perl -x 'f$environment("PROCEDURE")
    $ else perl -x - 'p1 'p2 'p3 'p4 'p5 'p6 'p7 'p8
    $ deck/dollars="__END__"
    #!/usr/bin/perl

    print "Hello from Perl!\n";

    __END__
    $ endif

Do take care with C<$ ASSIGN/nolog/user SYS$COMMAND: SYS$INPUT> if your
perl-in-DCL script expects to do things like C<< $read = <STDIN>; >>.

Filenames are in the format "name.extension;version".  The maximum
length for filenames is 39 characters, and the maximum length for
extensions is also 39 characters.  Version is a number from 1 to
32767.  Valid characters are C</[A-Z0-9$_-]/>.

VMS's RMS filesystem is case-insensitive and does not preserve case.
C<readdir> returns lowercased filenames, but specifying a file for
opening remains case-insensitive.  Files without extensions have a
trailing period on them, so doing a C<readdir> with a file named F<A.;5>
will return F<a.> (though that file could be opened with
C<open(FH, 'A')>).

RMS had an eight level limit on directory depths from any rooted logical
(allowing 16 levels overall) prior to VMS 7.2.  Hence
C<PERL_ROOT:[LIB.2.3.4.5.6.7.8]> is a valid directory specification but
C<PERL_ROOT:[LIB.2.3.4.5.6.7.8.9]> is not.  F<Makefile.PL> authors might
have to take this into account, but at least they can refer to the former
as C</PERL_ROOT/lib/2/3/4/5/6/7/8/>.

The VMS::Filespec module, which gets installed as part of the build
process on VMS, is a pure Perl module that can easily be installed on
non-VMS platforms and can be helpful for conversions to and from RMS
native formats.

What C<\n> represents depends on the type of file opened.  It usually
represents C<\012> but it could also be C<\015>, C<\012>, C<\015\012>, 
C<\000>, C<\040>, or nothing depending on the file organiztion and 
record format.  The VMS::Stdio module provides access to the 
special fopen() requirements of files with unusual attributes on VMS.

TCP/IP stacks are optional on VMS, so socket routines might not be
implemented.  UDP sockets may not be supported.

The value of C<$^O> on OpenVMS is "VMS".  To determine the architecture
that you are running on without resorting to loading all of C<%Config>
you can examine the content of the C<@INC> array like so:

    if (grep(/VMS_AXP/, @INC)) {
        print "I'm on Alpha!\n";

    } elsif (grep(/VMS_VAX/, @INC)) {
        print "I'm on VAX!\n";

    } else {
        print "I'm not so sure about where $^O is...\n";
    }

On VMS, perl determines the UTC offset from the C<SYS$TIMEZONE_DIFFERENTIAL>
logical name.  Although the VMS epoch began at 17-NOV-1858 00:00:00.00,
calls to C<localtime> are adjusted to count offsets from
01-JAN-1970 00:00:00.00, just like Unix.

Also see:

=over 4

=item *

F<README.vms> (installed as L<README_vms>), L<perlvms>

=item *

vmsperl list, majordomo@perl.org

(Put the words C<subscribe vmsperl> in message body.)

=item *

vmsperl on the web, http://www.sidhe.org/vmsperl/index.html

=back

=head2 VOS

Perl on VOS is discussed in F<README.vos> in the perl distribution
(installed as L<perlvos>).  Perl on VOS can accept either VOS- or
Unix-style file specifications as in either of the following:

    C<< $ perl -ne "print if /perl_setup/i" >system>notices >>
    C<< $ perl -ne "print if /perl_setup/i" /system/notices >>

or even a mixture of both as in:

    C<< $ perl -ne "print if /perl_setup/i" >system/notices >>

Even though VOS allows the slash character to appear in object
names, because the VOS port of Perl interprets it as a pathname
delimiting character, VOS files, directories, or links whose names
contain a slash character cannot be processed.  Such files must be
renamed before they can be processed by Perl.  Note that VOS limits
file names to 32 or fewer characters.

The value of C<$^O> on VOS is "VOS".  To determine the architecture that
you are running on without resorting to loading all of C<%Config> you
can examine the content of the @INC array like so:

    if ($^O =~ /VOS/) {
        print "I'm on a Stratus box!\n";
    } else {
        print "I'm not on a Stratus box!\n";
        die;
    }

Also see:

=over 4

=item *

F<README.vos> (installed as L<perlvos>)

=item *

The VOS mailing list.

There is no specific mailing list for Perl on VOS.  You can post
comments to the comp.sys.stratus newsgroup, or subscribe to the general
Stratus mailing list.  Send a letter with "subscribe Info-Stratus" in
the message body to majordomo@list.stratagy.com.

=item *

VOS Perl on the web at http://ftp.stratus.com/pub/vos/posix/posix.html

=back

=head2 EBCDIC Platforms

Recent versions of Perl have been ported to platforms such as OS/400 on
AS/400 minicomputers as well as OS/390, VM/ESA, and BS2000 for S/390
Mainframes.  Such computers use EBCDIC character sets internally (usually
Character Code Set ID 0037 for OS/400 and either 1047 or POSIX-BC for S/390
systems).  On the mainframe perl currently works under the "Unix system
services for OS/390" (formerly known as OpenEdition), VM/ESA OpenEdition, or
the BS200 POSIX-BC system (BS2000 is supported in perl 5.6 and greater).
See L<perlos390> for details.  Note that for OS/400 there is also a port of
Perl 5.8.1/5.9.0 or later to the PASE which is ASCII-based (as opposed to
ILE which is EBCDIC-based), see L<perlos400>. 

As of R2.5 of USS for OS/390 and Version 2.3 of VM/ESA these Unix
sub-systems do not support the C<#!> shebang trick for script invocation.
Hence, on OS/390 and VM/ESA perl scripts can be executed with a header
similar to the following simple script:

    : # use perl
        eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
            if 0;
    #!/usr/local/bin/perl     # just a comment really

    print "Hello from perl!\n";

OS/390 will support the C<#!> shebang trick in release 2.8 and beyond.
Calls to C<system> and backticks can use POSIX shell syntax on all
S/390 systems.

On the AS/400, if PERL5 is in your library list, you may need
to wrap your perl scripts in a CL procedure to invoke them like so:

    BEGIN
      CALL PGM(PERL5/PERL) PARM('/QOpenSys/hello.pl')
    ENDPGM

This will invoke the perl script F<hello.pl> in the root of the
QOpenSys file system.  On the AS/400 calls to C<system> or backticks
must use CL syntax.

On these platforms, bear in mind that the EBCDIC character set may have
an effect on what happens with some perl functions (such as C<chr>,
C<pack>, C<print>, C<printf>, C<ord>, C<sort>, C<sprintf>, C<unpack>), as
well as bit-fiddling with ASCII constants using operators like C<^>, C<&>
and C<|>, not to mention dealing with socket interfaces to ASCII computers
(see L<"Newlines">).

Fortunately, most web servers for the mainframe will correctly
translate the C<\n> in the following statement to its ASCII equivalent
(C<\r> is the same under both Unix and OS/390 & VM/ESA):

    print "Content-type: text/html\r\n\r\n";

The values of C<$^O> on some of these platforms includes:

    uname         $^O        $Config{'archname'}
    --------------------------------------------
    OS/390        os390      os390
    OS400         os400      os400
    POSIX-BC      posix-bc   BS2000-posix-bc
    VM/ESA        vmesa      vmesa

Some simple tricks for determining if you are running on an EBCDIC
platform could include any of the following (perhaps all):

    if ("\t" eq "\05")   { print "EBCDIC may be spoken here!\n"; }

    if (ord('A') == 193) { print "EBCDIC may be spoken here!\n"; }

    if (chr(169) eq 'z') { print "EBCDIC may be spoken here!\n"; }

One thing you may not want to rely on is the EBCDIC encoding
of punctuation characters since these may differ from code page to code
page (and once your module or script is rumoured to work with EBCDIC,
folks will want it to work with all EBCDIC character sets).

Also see:

=over 4

=item *

*

L<perlos390>, F<README.os390>, F<perlbs2000>, F<README.vmesa>,
L<perlebcdic>.

=item *

The perl-mvs@perl.org list is for discussion of porting issues as well as
general usage issues for all EBCDIC Perls.  Send a message body of
"subscribe perl-mvs" to majordomo@perl.org.

=item  *

AS/400 Perl information at
http://as400.rochester.ibm.com/
as well as on CPAN in the F<ports/> directory.

=back

=head2 Acorn RISC OS

Because Acorns use ASCII with newlines (C<\n>) in text files as C<\012> like
Unix, and because Unix filename emulation is turned on by default, 
most simple scripts will probably work "out of the box".  The native
filesystem is modular, and individual filesystems are free to be
case-sensitive or insensitive, and are usually case-preserving.  Some
native filesystems have name length limits, which file and directory
names are silently truncated to fit.  Scripts should be aware that the
standard filesystem currently has a name length limit of B<10>
characters, with up to 77 items in a directory, but other filesystems
may not impose such limitations.

Native filenames are of the form

    Filesystem#Special_Field::DiskName.$.Directory.Directory.File

where

    Special_Field is not usually present, but may contain . and $ .
    Filesystem =~ m|[A-Za-z0-9_]|
    DsicName   =~ m|[A-Za-z0-9_/]|
    $ represents the root directory
    . is the path separator
    @ is the current directory (per filesystem but machine global)
    ^ is the parent directory
    Directory and File =~ m|[^\0- "\.\$\%\&:\@\\^\|\177]+|

The default filename translation is roughly C<tr|/.|./|;>

Note that C<"ADFS::HardDisk.$.File" ne 'ADFS::HardDisk.$.File'> and that
the second stage of C<$> interpolation in regular expressions will fall
foul of the C<$.> if scripts are not careful.

Logical paths specified by system variables containing comma-separated
search lists are also allowed; hence C<System:Modules> is a valid
filename, and the filesystem will prefix C<Modules> with each section of
C<System$Path> until a name is made that points to an object on disk.
Writing to a new file C<System:Modules> would be allowed only if
C<System$Path> contains a single item list.  The filesystem will also
expand system variables in filenames if enclosed in angle brackets, so
C<< <System$Dir>.Modules >> would look for the file
S<C<$ENV{'System$Dir'} . 'Modules'>>.  The obvious implication of this is
that B<fully qualified filenames can start with C<< <> >>> and should
be protected when C<open> is used for input.

Because C<.> was in use as a directory separator and filenames could not
be assumed to be unique after 10 characters, Acorn implemented the C
compiler to strip the trailing C<.c> C<.h> C<.s> and C<.o> suffix from
filenames specified in source code and store the respective files in
subdirectories named after the suffix.  Hence files are translated:

    foo.h           h.foo
    C:foo.h         C:h.foo        (logical path variable)
    sys/os.h        sys.h.os       (C compiler groks Unix-speak)
    10charname.c    c.10charname
    10charname.o    o.10charname
    11charname_.c   c.11charname   (assuming filesystem truncates at 10)

The Unix emulation library's translation of filenames to native assumes
that this sort of translation is required, and it allows a user-defined list
of known suffixes that it will transpose in this fashion.  This may
seem transparent, but consider that with these rules C<foo/bar/baz.h>
and C<foo/bar/h/baz> both map to C<foo.bar.h.baz>, and that C<readdir> and
C<glob> cannot and do not attempt to emulate the reverse mapping.  Other
C<.>'s in filenames are translated to C</>.

As implied above, the environment accessed through C<%ENV> is global, and
the convention is that program specific environment variables are of the
form C<Program$Name>.  Each filesystem maintains a current directory,
and the current filesystem's current directory is the B<global> current
directory.  Consequently, sociable programs don't change the current
directory but rely on full pathnames, and programs (and Makefiles) cannot
assume that they can spawn a child process which can change the current
directory without affecting its parent (and everyone else for that
matter).

Because native operating system filehandles are global and are currently 
allocated down from 255, with 0 being a reserved value, the Unix emulation
library emulates Unix filehandles.  Consequently, you can't rely on
passing C<STDIN>, C<STDOUT>, or C<STDERR> to your children.

The desire of users to express filenames of the form
C<< <Foo$Dir>.Bar >> on the command line unquoted causes problems,
too: C<``> command output capture has to perform a guessing game.  It
assumes that a string C<< <[^<>]+\$[^<>]> >> is a
reference to an environment variable, whereas anything else involving
C<< < >> or C<< > >> is redirection, and generally manages to be 99%
right.  Of course, the problem remains that scripts cannot rely on any
Unix tools being available, or that any tools found have Unix-like command
line arguments.

Extensions and XS are, in theory, buildable by anyone using free
tools.  In practice, many don't, as users of the Acorn platform are
used to binary distributions.  MakeMaker does run, but no available
make currently copes with MakeMaker's makefiles; even if and when
this should be fixed, the lack of a Unix-like shell will cause
problems with makefile rules, especially lines of the form C<cd
sdbm && make all>, and anything using quoting.

"S<RISC OS>" is the proper name for the operating system, but the value
in C<$^O> is "riscos" (because we don't like shouting).

=head2 Other perls

Perl has been ported to many platforms that do not fit into any of
the categories listed above.  Some, such as AmigaOS, Atari MiNT,
BeOS, HP MPE/iX, QNX, Plan 9, and VOS, have been well-integrated
into the standard Perl source code kit.  You may need to see the
F<ports/> directory on CPAN for information, and possibly binaries,
for the likes of: aos, Atari ST, lynxos, riscos, Novell Netware,
Tandem Guardian, I<etc.>  (Yes, we know that some of these OSes may
fall under the Unix category, but we are not a standards body.)

Some approximate operating system names and their C<$^O> values
in the "OTHER" category include:

    OS            $^O        $Config{'archname'}
    ------------------------------------------
    Amiga DOS     amigaos    m68k-amigos
    BeOS          beos
    MPE/iX        mpeix      PA-RISC1.1

See also:

=over 4

=item *

Amiga, F<README.amiga> (installed as L<perlamiga>).

=item *

Atari, F<README.mint> and Guido Flohr's web page
http://stud.uni-sb.de/~gufl0000/

=item *

Be OS, F<README.beos>

=item *

HP 300 MPE/iX, F<README.mpeix> and Mark Bixby's web page
http://www.bixby.org/mark/perlix.html

=item *

A free perl5-based PERL.NLM for Novell Netware is available in
precompiled binary and source code form from http://www.novell.com/
as well as from CPAN.

=item  *

S<Plan 9>, F<README.plan9>

=back

=head1 FUNCTION IMPLEMENTATIONS

Listed below are functions that are either completely unimplemented
or else have been implemented differently on various platforms.
Following each description will be, in parentheses, a list of
platforms that the description applies to.

The list may well be incomplete, or even wrong in some places.  When
in doubt, consult the platform-specific README files in the Perl
source distribution, and any other documentation resources accompanying
a given port.

Be aware, moreover, that even among Unix-ish systems there are variations.

For many functions, you can also query C<%Config>, exported by
default from the Config module.  For example, to check whether the
platform has the C<lstat> call, check C<$Config{d_lstat}>.  See
L<Config> for a full description of available variables.

=head2 Alphabetical Listing of Perl Functions

=over 8

=item -X FILEHANDLE

=item -X EXPR

=item -X

C<-r>, C<-w>, and C<-x> have a limited meaning only; directories
and applications are executable, and there are no uid/gid
considerations.  C<-o> is not supported.  (S<Mac OS>)

C<-r>, C<-w>, C<-x>, and C<-o> tell whether the file is accessible,
which may not reflect UIC-based file protections.  (VMS)

C<-s> returns the size of the data fork, not the total size of data fork
plus resource fork.  (S<Mac OS>).

C<-s> by name on an open file will return the space reserved on disk,
rather than the current extent.  C<-s> on an open filehandle returns the
current size.  (S<RISC OS>)

C<-R>, C<-W>, C<-X>, C<-O> are indistinguishable from C<-r>, C<-w>,
C<-x>, C<-o>. (S<Mac OS>, Win32, VMS, S<RISC OS>)

C<-b>, C<-c>, C<-k>, C<-g>, C<-p>, C<-u>, C<-A> are not implemented.
(S<Mac OS>)

C<-g>, C<-k>, C<-l>, C<-p>, C<-u>, C<-A> are not particularly meaningful.
(Win32, VMS, S<RISC OS>)

C<-d> is true if passed a device spec without an explicit directory.
(VMS)

C<-T> and C<-B> are implemented, but might misclassify Mac text files
with foreign characters; this is the case will all platforms, but may
affect S<Mac OS> often.  (S<Mac OS>)

C<-x> (or C<-X>) determine if a file ends in one of the executable
suffixes.  C<-S> is meaningless.  (Win32)

C<-x> (or C<-X>) determine if a file has an executable file type.
(S<RISC OS>)

=item alarm SECONDS

=item alarm

Not implemented. (Win32)

=item binmode FILEHANDLE

Meaningless.  (S<Mac OS>, S<RISC OS>)

Reopens file and restores pointer; if function fails, underlying
filehandle may be closed, or pointer may be in a different position.
(VMS)

The value returned by C<tell> may be affected after the call, and
the filehandle may be flushed. (Win32)

=item chmod LIST

Only limited meaning.  Disabling/enabling write permission is mapped to
locking/unlocking the file. (S<Mac OS>)

Only good for changing "owner" read-write access, "group", and "other"
bits are meaningless. (Win32)

Only good for changing "owner" and "other" read-write access. (S<RISC OS>)

Access permissions are mapped onto VOS access-control list changes. (VOS)

The actual permissions set depend on the value of the C<CYGWIN>
in the SYSTEM environment settings.  (Cygwin)

=item chown LIST

Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)

Does nothing, but won't fail. (Win32)

A little funky, because VOS's notion of ownership is a little funky (VOS).

=item chroot FILENAME

=item chroot

Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS, VM/ESA)

=item crypt PLAINTEXT,SALT

May not be available if library or source was not provided when building
perl. (Win32)

=item dbmclose HASH

Not implemented. (VMS, S<Plan 9>, VOS)

=item dbmopen HASH,DBNAME,MODE

Not implemented. (VMS, S<Plan 9>, VOS)

=item dump LABEL

Not useful. (S<Mac OS>, S<RISC OS>)

Not implemented. (Win32)

Invokes VMS debugger. (VMS)

=item exec LIST

Not implemented. (S<Mac OS>)

Implemented via Spawn. (VM/ESA)

Does not automatically flush output handles on some platforms.
(SunOS, Solaris, HP-UX)

=item exit EXPR

=item exit

Emulates UNIX exit() (which considers C<exit 1> to indicate an error) by
mapping the C<1> to SS$_ABORT (C<44>).  This behavior may be overridden
with the pragma C<use vmsish 'exit'>.  As with the CRTL's exit()
function, C<exit 0> is also mapped to an exit status of SS$_NORMAL
(C<1>); this mapping cannot be overridden.  Any other argument to exit()
is used directly as Perl's exit status. (VMS)

=item fcntl FILEHANDLE,FUNCTION,SCALAR

Not implemented. (Win32, VMS)

=item flock FILEHANDLE,OPERATION

Not implemented (S<Mac OS>, VMS, S<RISC OS>, VOS).

Available only on Windows NT (not on Windows 95). (Win32)

=item fork

Not implemented. (S<Mac OS>, AmigaOS, S<RISC OS>, VM/ESA, VMS)

Emulated using multiple interpreters.  See L<perlfork>.  (Win32)

Does not automatically flush output handles on some platforms.
(SunOS, Solaris, HP-UX)

=item getlogin

Not implemented. (S<Mac OS>, S<RISC OS>)

=item getpgrp PID

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)

=item getppid

Not implemented. (S<Mac OS>, Win32, S<RISC OS>)

=item getpriority WHICH,WHO

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)

=item getpwnam NAME

Not implemented. (S<Mac OS>, Win32)

Not useful. (S<RISC OS>)

=item getgrnam NAME

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)

=item getnetbyname NAME

Not implemented. (S<Mac OS>, Win32, S<Plan 9>)

=item getpwuid UID

Not implemented. (S<Mac OS>, Win32)

Not useful. (S<RISC OS>)

=item getgrgid GID

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)

=item getnetbyaddr ADDR,ADDRTYPE

Not implemented. (S<Mac OS>, Win32, S<Plan 9>)

=item getprotobynumber NUMBER

Not implemented. (S<Mac OS>)

=item getservbyport PORT,PROTO

Not implemented. (S<Mac OS>)

=item getpwent

Not implemented. (S<Mac OS>, Win32, VM/ESA)

=item getgrent

Not implemented. (S<Mac OS>, Win32, VMS, VM/ESA)

=item gethostbyname

C<gethostbyname('localhost')> does not work everywhere: you may have
to use C<gethostbyname('127.0.0.1')>. (S<Mac OS>, S<Irix 5>)

=item gethostent

Not implemented. (S<Mac OS>, Win32)

=item getnetent

Not implemented. (S<Mac OS>, Win32, S<Plan 9>)

=item getprotoent

Not implemented. (S<Mac OS>, Win32, S<Plan 9>)

=item getservent

Not implemented. (Win32, S<Plan 9>)

=item sethostent STAYOPEN

Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)

=item setnetent STAYOPEN

Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)

=item setprotoent STAYOPEN

Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)

=item setservent STAYOPEN

Not implemented. (S<Plan 9>, Win32, S<RISC OS>)

=item endpwent

Not implemented. (S<Mac OS>, MPE/iX, VM/ESA, Win32)

=item endgrent

Not implemented. (S<Mac OS>, MPE/iX, S<RISC OS>, VM/ESA, VMS, Win32)

=item endhostent

Not implemented. (S<Mac OS>, Win32)

=item endnetent

Not implemented. (S<Mac OS>, Win32, S<Plan 9>)

=item endprotoent

Not implemented. (S<Mac OS>, Win32, S<Plan 9>)

=item endservent

Not implemented. (S<Plan 9>, Win32)

=item getsockopt SOCKET,LEVEL,OPTNAME

Not implemented. (S<Plan 9>)

=item glob EXPR

=item glob

This operator is implemented via the File::Glob extension on most
platforms.  See L<File::Glob> for portability information.

=item ioctl FILEHANDLE,FUNCTION,SCALAR

Not implemented. (VMS)

Available only for socket handles, and it does what the ioctlsocket() call
in the Winsock API does. (Win32)

Available only for socket handles. (S<RISC OS>)

=item kill SIGNAL, LIST

C<kill(0, LIST)> is implemented for the sake of taint checking;
use with other signals is unimplemented. (S<Mac OS>)

Not implemented, hence not useful for taint checking. (S<RISC OS>)

C<kill()> doesn't have the semantics of C<raise()>, i.e. it doesn't send
a signal to the identified process like it does on Unix platforms.
Instead C<kill($sig, $pid)> terminates the process identified by $pid,
and makes it exit immediately with exit status $sig.  As in Unix, if
$sig is 0 and the specified process exists, it returns true without
actually terminating it. (Win32)

=item link OLDFILE,NEWFILE

Not implemented. (S<Mac OS>, MPE/iX, VMS, S<RISC OS>)

Link count not updated because hard links are not quite that hard
(They are sort of half-way between hard and soft links). (AmigaOS)

Hard links are implemented on Win32 (Windows NT and Windows 2000)
under NTFS only.

=item lstat FILEHANDLE

=item lstat EXPR

=item lstat

Not implemented. (VMS, S<RISC OS>)

Return values (especially for device and inode) may be bogus. (Win32)

=item msgctl ID,CMD,ARG

=item msgget KEY,FLAGS

=item msgsnd ID,MSG,FLAGS

=item msgrcv ID,VAR,SIZE,TYPE,FLAGS

Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS)

=item open FILEHANDLE,EXPR

=item open FILEHANDLE

The C<|> variants are supported only if ToolServer is installed.
(S<Mac OS>)

open to C<|-> and C<-|> are unsupported. (S<Mac OS>, Win32, S<RISC OS>)

Opening a process does not automatically flush output handles on some
platforms.  (SunOS, Solaris, HP-UX)

=item pipe READHANDLE,WRITEHANDLE

Very limited functionality. (MiNT)

=item readlink EXPR

=item readlink

Not implemented. (Win32, VMS, S<RISC OS>)

=item select RBITS,WBITS,EBITS,TIMEOUT

Only implemented on sockets. (Win32, VMS)

Only reliable on sockets. (S<RISC OS>)

Note that the C<select FILEHANDLE> form is generally portable.

=item semctl ID,SEMNUM,CMD,ARG

=item semget KEY,NSEMS,FLAGS

=item semop KEY,OPSTRING

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)

=item setgrent

Not implemented. (S<Mac OS>, MPE/iX, VMS, Win32, S<RISC OS>, VOS)

=item setpgrp PID,PGRP

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)

=item setpriority WHICH,WHO,PRIORITY

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)

=item setpwent

Not implemented. (S<Mac OS>, MPE/iX, Win32, S<RISC OS>, VOS)

=item setsockopt SOCKET,LEVEL,OPTNAME,OPTVAL

Not implemented. (S<Plan 9>)

=item shmctl ID,CMD,ARG

=item shmget KEY,SIZE,FLAGS

=item shmread ID,VAR,POS,SIZE

=item shmwrite ID,STRING,POS,SIZE

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)

=item sockatmark SOCKET

A relatively recent addition to socket functions, may not
be implemented even in UNIX platforms.

=item socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL

Not implemented. (Win32, VMS, S<RISC OS>, VOS, VM/ESA)

=item stat FILEHANDLE

=item stat EXPR

=item stat

Platforms that do not have rdev, blksize, or blocks will return these
as '', so numeric comparison or manipulation of these fields may cause
'not numeric' warnings.

mtime and atime are the same thing, and ctime is creation time instead of
inode change time. (S<Mac OS>).

ctime not supported on UFS (S<Mac OS X>).

ctime is creation time instead of inode change time  (Win32).

device and inode are not meaningful.  (Win32)

device and inode are not necessarily reliable.  (VMS)

mtime, atime and ctime all return the last modification time.  Device and
inode are not necessarily reliable.  (S<RISC OS>)

dev, rdev, blksize, and blocks are not available.  inode is not
meaningful and will differ between stat calls on the same file.  (os2)

some versions of cygwin when doing a stat("foo") and if not finding it
may then attempt to stat("foo.exe") (Cygwin)

=item symlink OLDFILE,NEWFILE

Not implemented. (Win32, VMS, S<RISC OS>)

=item syscall LIST

Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)

=item sysopen FILEHANDLE,FILENAME,MODE,PERMS

The traditional "0", "1", and "2" MODEs are implemented with different
numeric values on some systems.  The flags exported by C<Fcntl>
(O_RDONLY, O_WRONLY, O_RDWR) should work everywhere though.  (S<Mac
OS>, OS/390, VM/ESA)

=item system LIST

In general, do not assume the UNIX/POSIX semantics that you can shift
C<$?> right by eight to get the exit value, or that C<$? & 127>
would give you the number of the signal that terminated the program,
or that C<$? & 128> would test true if the program was terminated by a
coredump.  Instead, use the POSIX W*() interfaces: for example, use
WIFEXITED($?) and WEXITVALUE($?) to test for a normal exit and the exit
value, WIFSIGNALED($?) and WTERMSIG($?) for a signal exit and the
signal.  Core dumping is not a portable concept, so there's no portable
way to test for that.

Only implemented if ToolServer is installed. (S<Mac OS>)

As an optimization, may not call the command shell specified in
C<$ENV{PERL5SHELL}>.  C<system(1, @args)> spawns an external
process and immediately returns its process designator, without
waiting for it to terminate.  Return value may be used subsequently
in C<wait> or C<waitpid>.  Failure to spawn() a subprocess is indicated
by setting $? to "255 << 8".  C<$?> is set in a way compatible with
Unix (i.e. the exitstatus of the subprocess is obtained by "$? >> 8",
as described in the documentation).  (Win32)

There is no shell to process metacharacters, and the native standard is
to pass a command line terminated by "\n" "\r" or "\0" to the spawned
program.  Redirection such as C<< > foo >> is performed (if at all) by
the run time library of the spawned program.  C<system> I<list> will call
the Unix emulation library's C<exec> emulation, which attempts to provide
emulation of the stdin, stdout, stderr in force in the parent, providing
the child program uses a compatible version of the emulation library.
I<scalar> will call the native command line direct and no such emulation
of a child Unix program will exists.  Mileage B<will> vary.  (S<RISC OS>)

Far from being POSIX compliant.  Because there may be no underlying
/bin/sh tries to work around the problem by forking and execing the
first token in its argument string.  Handles basic redirection
("<" or ">") on its own behalf. (MiNT)

Does not automatically flush output handles on some platforms.
(SunOS, Solaris, HP-UX)

The return value is POSIX-like (shifted up by 8 bits), which only allows
room for a made-up value derived from the severity bits of the native
32-bit condition code (unless overridden by C<use vmsish 'status'>). 
For more details see L<perlvms/$?>. (VMS)

=item times

Only the first entry returned is nonzero. (S<Mac OS>)

"cumulative" times will be bogus.  On anything other than Windows NT
or Windows 2000, "system" time will be bogus, and "user" time is
actually the time returned by the clock() function in the C runtime
library. (Win32)

Not useful. (S<RISC OS>)

=item truncate FILEHANDLE,LENGTH

=item truncate EXPR,LENGTH

Not implemented. (Older versions of VMS)

Truncation to same-or-shorter lengths only. (VOS)

If a FILEHANDLE is supplied, it must be writable and opened in append
mode (i.e., use C<<< open(FH, '>>filename') >>>
or C<sysopen(FH,...,O_APPEND|O_RDWR)>.  If a filename is supplied, it
should not be held open elsewhere. (Win32)

=item umask EXPR

=item umask

Returns undef where unavailable, as of version 5.005.

C<umask> works but the correct permissions are set only when the file
is finally closed. (AmigaOS)

=item utime LIST

Only the modification time is updated. (S<BeOS>, S<Mac OS>, VMS, S<RISC OS>)

May not behave as expected.  Behavior depends on the C runtime
library's implementation of utime(), and the filesystem being
used.  The FAT filesystem typically does not support an "access
time" field, and it may limit timestamps to a granularity of
two seconds. (Win32)

=item wait

=item waitpid PID,FLAGS

Not implemented. (S<Mac OS>)

Can only be applied to process handles returned for processes spawned
using C<system(1, ...)> or pseudo processes created with C<fork()>. (Win32)

Not useful. (S<RISC OS>)

=back

=head1 CHANGES

=over 4

=item v1.49, 12 August 2002

Updates for VOS from Paul Green.

=item v1.48, 02 February 2001

Various updates from perl5-porters over the past year, supported
platforms update from Jarkko Hietaniemi.

=item v1.47, 22 March 2000

Various cleanups from Tom Christiansen, including migration of 
long platform listings from L<perl>.

=item v1.46, 12 February 2000

Updates for VOS and MPE/iX. (Peter Prymmer)  Other small changes.

=item v1.45, 20 December 1999

Small changes from 5.005_63 distribution, more changes to EBCDIC info.

=item v1.44, 19 July 1999

A bunch of updates from Peter Prymmer for C<$^O> values,
endianness, File::Spec, VMS, BS2000, OS/400.

=item v1.43, 24 May 1999

Added a lot of cleaning up from Tom Christiansen.

=item v1.42, 22 May 1999

Added notes about tests, sprintf/printf, and epoch offsets.

=item v1.41, 19 May 1999

Lots more little changes to formatting and content.

Added a bunch of C<$^O> and related values
for various platforms; fixed mail and web addresses, and added
and changed miscellaneous notes.  (Peter Prymmer)

=item v1.40, 11 April 1999

Miscellaneous changes.

=item v1.39, 11 February 1999

Changes from Jarkko and EMX URL fixes Michael Schwern.  Additional
note about newlines added.

=item v1.38, 31 December 1998

More changes from Jarkko.

=item v1.37, 19 December 1998

More minor changes.  Merge two separate version 1.35 documents.

=item v1.36, 9 September 1998

Updated for Stratus VOS.  Also known as version 1.35.

=item v1.35, 13 August 1998

Integrate more minor changes, plus addition of new sections under
L<"ISSUES">: L<"Numbers endianness and Width">,
L<"Character sets and character encoding">,
L<"Internationalisation">.

=item v1.33, 06 August 1998

Integrate more minor changes.

=item v1.32, 05 August 1998

Integrate more minor changes.

=item v1.30, 03 August 1998

Major update for RISC OS, other minor changes.

=item v1.23, 10 July 1998

First public release with perl5.005.

=back

=head1 Supported Platforms

As of July 2002 (the Perl release 5.8.0), the following platforms are
able to build Perl from the standard source code distribution
available at http://www.cpan.org/src/index.html

        AIX
        BeOS
        BSD/OS          (BSDi)
        Cygwin
        DG/UX
        DOS DJGPP       1)
        DYNIX/ptx
        EPOC R5
        FreeBSD
        HI-UXMPP        (Hitachi) (5.8.0 worked but we didn't know it)
        HP-UX
        IRIX
        Linux
        Mac OS Classic
        Mac OS X        (Darwin)
        MPE/iX
        NetBSD
        NetWare
        NonStop-UX
        ReliantUNIX     (formerly SINIX)
        OpenBSD
        OpenVMS         (formerly VMS)
        Open UNIX       (Unixware) (since Perl 5.8.1/5.9.0)
        OS/2
        OS/400          (using the PASE) (since Perl 5.8.1/5.9.0)
        PowerUX
        POSIX-BC        (formerly BS2000)
        QNX
        Solaris
        SunOS 4
        SUPER-UX        (NEC)
        Tru64 UNIX      (formerly DEC OSF/1, Digital UNIX)
        UNICOS
        UNICOS/mk
        UTS
        VOS
        Win95/98/ME/2K/XP 2)
        WinCE
        z/OS            (formerly OS/390)
        VM/ESA

        1) in DOS mode either the DOS or OS/2 ports can be used
        2) compilers: Borland, MinGW (GCC), VC6

The following platforms worked with the previous releases (5.6 and
5.7), but we did not manage either to fix or to test these in time
for the 5.8.0 release.  There is a very good chance that many of these
will work fine with the 5.8.0.

        BSD/OS
        DomainOS
        Hurd
        LynxOS
        MachTen
        PowerMAX
        SCO SV
        SVR4
        Unixware
        Windows 3.1

Known to be broken for 5.8.0 (but 5.6.1 and 5.7.2 can be used):

        AmigaOS

The following platforms have been known to build Perl from source in
the past (5.005_03 and earlier), but we haven't been able to verify
their status for the current release, either because the
hardware/software platforms are rare or because we don't have an
active champion on these platforms--or both.  They used to work,
though, so go ahead and try compiling them, and let perlbug@perl.org
of any trouble.

        3b1
        A/UX
        ConvexOS
        CX/UX
        DC/OSx
        DDE SMES
        DOS EMX
        Dynix
        EP/IX
        ESIX
        FPS
        GENIX
        Greenhills
        ISC
        MachTen 68k
        MiNT
        MPC
        NEWS-OS
        NextSTEP
        OpenSTEP
        Opus
        Plan 9
        RISC/os
        SCO ODT/OSR
        Stellar
        SVR2
        TI1500
        TitanOS
        Ultrix
        Unisys Dynix

The following platforms have their own source code distributions and
binaries available via http://www.cpan.org/ports/

                                Perl release

        OS/400 (ILE)            5.005_02
        Tandem Guardian         5.004

The following platforms have only binaries available via
http://www.cpan.org/ports/index.html :

                                Perl release

        Acorn RISCOS            5.005_02
        AOS                     5.002
        LynxOS                  5.004_02

Although we do suggest that you always build your own Perl from
the source code, both for maximal configurability and for security,
in case you are in a hurry you can check
http://www.cpan.org/ports/index.html for binary distributions.

=head1 SEE ALSO

L<perlaix>, L<perlamiga>, L<perlapollo>, L<perlbeos>, L<perlbs2000>,
L<perlce>, L<perlcygwin>, L<perldgux>, L<perldos>, L<perlepoc>,
L<perlebcdic>, L<perlfreebsd>, L<perlhurd>, L<perlhpux>, L<perlirix>,
L<perlmachten>, L<perlmacos>, L<perlmint>, L<perlmpeix>,
L<perlnetware>, L<perlos2>, L<perlos390>, L<perlos400>,
L<perlplan9>, L<perlqnx>, L<perlsolaris>, L<perltru64>,
L<perlunicode>, L<perlvmesa>, L<perlvms>, L<perlvos>,
L<perlwin32>, and L<Win32>.

=head1 AUTHORS / CONTRIBUTORS

Abigail <abigail@foad.org>,
Charles Bailey <bailey@newman.upenn.edu>,
Graham Barr <gbarr@pobox.com>,
Tom Christiansen <tchrist@perl.com>,
Nicholas Clark <nick@ccl4.org>,
Thomas Dorner <Thomas.Dorner@start.de>,
Andy Dougherty <doughera@lafayette.edu>,
Dominic Dunlop <domo@computer.org>,
Neale Ferguson <neale@vma.tabnsw.com.au>,
David J. Fiander <davidf@mks.com>,
Paul Green <Paul.Green@stratus.com>,
M.J.T. Guy <mjtg@cam.ac.uk>,
Jarkko Hietaniemi <jhi@iki.fi>,
Luther Huffman <lutherh@stratcom.com>,
Nick Ing-Simmons <nick@ing-simmons.net>,
Andreas J. KE<ouml>nig <a.koenig@mind.de>,
Markus Laker <mlaker@contax.co.uk>,
Andrew M. Langmead <aml@world.std.com>,
Larry Moore <ljmoore@freespace.net>,
Paul Moore <Paul.Moore@uk.origin-it.com>,
Chris Nandor <pudge@pobox.com>,
Matthias Neeracher <neeracher@mac.com>,
Philip Newton <pne@cpan.org>,
Gary Ng <71564.1743@CompuServe.COM>,
Tom Phoenix <rootbeer@teleport.com>,
AndrE<eacute> Pirard <A.Pirard@ulg.ac.be>,
Peter Prymmer <pvhp@forte.com>,
Hugo van der Sanden <hv@crypt0.demon.co.uk>,
Gurusamy Sarathy <gsar@activestate.com>,
Paul J. Schinder <schinder@pobox.com>,
Michael G Schwern <schwern@pobox.com>,
Dan Sugalski <dan@sidhe.org>,
Nathan Torkington <gnat@frii.com>.