3 perlport - Writing portable Perl
7 Perl runs on numerous operating systems. While most of them share
8 much in common, they also have their own unique features.
10 This document is meant to help you to find out what constitutes portable
11 Perl code. That way once you make a decision to write portably,
12 you know where the lines are drawn, and you can stay within them.
14 There is a tradeoff between taking full advantage of one particular
15 type of computer and taking advantage of a full range of them.
16 Naturally, as you broaden your range and become more diverse, the
17 common factors drop, and you are left with an increasingly smaller
18 area of common ground in which you can operate to accomplish a
19 particular task. Thus, when you begin attacking a problem, it is
20 important to consider under which part of the tradeoff curve you
21 want to operate. Specifically, you must decide whether it is
22 important that the task that you are coding have the full generality
23 of being portable, or whether to just get the job done right now.
24 This is the hardest choice to be made. The rest is easy, because
25 Perl provides many choices, whichever way you want to approach your
28 Looking at it another way, writing portable code is usually about
29 willfully limiting your available choices. Naturally, it takes
30 discipline and sacrifice to do that. The product of portability
31 and convenience may be a constant. You have been warned.
33 Be aware of two important points:
37 =item Not all Perl programs have to be portable
39 There is no reason you should not use Perl as a language to glue Unix
40 tools together, or to prototype a Macintosh application, or to manage the
41 Windows registry. If it makes no sense to aim for portability for one
42 reason or another in a given program, then don't bother.
44 =item Nearly all of Perl already I<is> portable
46 Don't be fooled into thinking that it is hard to create portable Perl
47 code. It isn't. Perl tries its level-best to bridge the gaps between
48 what's available on different platforms, and all the means available to
49 use those features. Thus almost all Perl code runs on any machine
50 without modification. But there are some significant issues in
51 writing portable code, and this document is entirely about those issues.
55 Here's the general rule: When you approach a task commonly done
56 using a whole range of platforms, think about writing portable
57 code. That way, you don't sacrifice much by way of the implementation
58 choices you can avail yourself of, and at the same time you can give
59 your users lots of platform choices. On the other hand, when you have to
60 take advantage of some unique feature of a particular platform, as is
61 often the case with systems programming (whether for Unix, Windows,
62 S<Mac OS>, VMS, etc.), consider writing platform-specific code.
64 When the code will run on only two or three operating systems, you
65 may need to consider only the differences of those particular systems.
66 The important thing is to decide where the code will run and to be
67 deliberate in your decision.
69 The material below is separated into three main sections: main issues of
70 portability (L<"ISSUES">), platform-specific issues (L<"PLATFORMS">), and
71 built-in perl functions that behave differently on various ports
72 (L<"FUNCTION IMPLEMENTATIONS">).
74 This information should not be considered complete; it includes possibly
75 transient information about idiosyncrasies of some of the ports, almost
76 all of which are in a state of constant evolution. Thus, this material
77 should be considered a perpetual work in progress
78 (C<< <IMG SRC="yellow_sign.gif" ALT="Under Construction"> >>).
84 In most operating systems, lines in files are terminated by newlines.
85 Just what is used as a newline may vary from OS to OS. Unix
86 traditionally uses C<\012>, one type of DOSish I/O uses C<\015\012>,
87 and S<Mac OS> uses C<\015>.
89 Perl uses C<\n> to represent the "logical" newline, where what is
90 logical may depend on the platform in use. In MacPerl, C<\n> always
91 means C<\015>. In DOSish perls, C<\n> usually means C<\012>, but
92 when accessing a file in "text" mode, STDIO translates it to (or
93 from) C<\015\012>, depending on whether you're reading or writing.
94 Unix does the same thing on ttys in canonical mode. C<\015\012>
95 is commonly referred to as CRLF.
97 A common cause of unportable programs is the misuse of chop() to trim
107 You can get away with this on Unix and Mac OS (they have a single
108 character end-of-line), but the same program will break under DOSish
109 perls because you're only chop()ing half the end-of-line. Instead,
110 chomp() should be used to trim newlines. The L<Dunce::Files> module
111 can help audit your code for misuses of chop().
113 When dealing with binary files (or text files in binary mode) be sure
114 to explicitly set $/ to the appropriate value for your file format
115 before using chomp().
117 Because of the "text" mode translation, DOSish perls have limitations
118 in using C<seek> and C<tell> on a file accessed in "text" mode.
119 Stick to C<seek>-ing to locations you got from C<tell> (and no
120 others), and you are usually free to use C<seek> and C<tell> even
121 in "text" mode. Using C<seek> or C<tell> or other file operations
122 may be non-portable. If you use C<binmode> on a file, however, you
123 can usually C<seek> and C<tell> with arbitrary values in safety.
125 A common misconception in socket programming is that C<\n> eq C<\012>
126 everywhere. When using protocols such as common Internet protocols,
127 C<\012> and C<\015> are called for specifically, and the values of
128 the logical C<\n> and C<\r> (carriage return) are not reliable.
130 print SOCKET "Hi there, client!\r\n"; # WRONG
131 print SOCKET "Hi there, client!\015\012"; # RIGHT
133 However, using C<\015\012> (or C<\cM\cJ>, or C<\x0D\x0A>) can be tedious
134 and unsightly, as well as confusing to those maintaining the code. As
135 such, the Socket module supplies the Right Thing for those who want it.
137 use Socket qw(:DEFAULT :crlf);
138 print SOCKET "Hi there, client!$CRLF" # RIGHT
140 When reading from a socket, remember that the default input record
141 separator C<$/> is C<\n>, but robust socket code will recognize as
142 either C<\012> or C<\015\012> as end of line:
148 Because both CRLF and LF end in LF, the input record separator can
149 be set to LF and any CR stripped later. Better to write:
151 use Socket qw(:DEFAULT :crlf);
152 local($/) = LF; # not needed if $/ is already \012
155 s/$CR?$LF/\n/; # not sure if socket uses LF or CRLF, OK
156 # s/\015?\012/\n/; # same thing
159 This example is preferred over the previous one--even for Unix
160 platforms--because now any C<\015>'s (C<\cM>'s) are stripped out
161 (and there was much rejoicing).
163 Similarly, functions that return text data--such as a function that
164 fetches a web page--should sometimes translate newlines before
165 returning the data, if they've not yet been translated to the local
166 newline representation. A single line of code will often suffice:
168 $data =~ s/\015?\012/\n/g;
171 Some of this may be confusing. Here's a handy reference to the ASCII CR
172 and LF characters. You can print it out and stick it in your wallet.
174 LF eq \012 eq \x0A eq \cJ eq chr(10) eq ASCII 10
175 CR eq \015 eq \x0D eq \cM eq chr(13) eq ASCII 13
178 ---------------------------
181 \n * | LF | CRLF | CR |
182 \r * | CR | CR | LF |
183 ---------------------------
186 The Unix column assumes that you are not accessing a serial line
187 (like a tty) in canonical mode. If you are, then CR on input becomes
188 "\n", and "\n" on output becomes CRLF.
190 These are just the most common definitions of C<\n> and C<\r> in Perl.
191 There may well be others. For example, on an EBCDIC implementation
192 such as z/OS (OS/390) or OS/400 (using the ILE, the PASE is ASCII-based)
193 the above material is similar to "Unix" but the code numbers change:
195 LF eq \025 eq \x15 eq \cU eq chr(21) eq CP-1047 21
196 LF eq \045 eq \x25 eq chr(37) eq CP-0037 37
197 CR eq \015 eq \x0D eq \cM eq chr(13) eq CP-1047 13
198 CR eq \015 eq \x0D eq \cM eq chr(13) eq CP-0037 13
201 ----------------------
206 ----------------------
209 =head2 Numbers endianness and Width
211 Different CPUs store integers and floating point numbers in different
212 orders (called I<endianness>) and widths (32-bit and 64-bit being the
213 most common today). This affects your programs when they attempt to transfer
214 numbers in binary format from one CPU architecture to another,
215 usually either "live" via network connection, or by storing the
216 numbers to secondary storage such as a disk file or tape.
218 Conflicting storage orders make utter mess out of the numbers. If a
219 little-endian host (Intel, VAX) stores 0x12345678 (305419896 in
220 decimal), a big-endian host (Motorola, Sparc, PA) reads it as
221 0x78563412 (2018915346 in decimal). Alpha and MIPS can be either:
222 Digital/Compaq used/uses them in little-endian mode; SGI/Cray uses
223 them in big-endian mode. To avoid this problem in network (socket)
224 connections use the C<pack> and C<unpack> formats C<n> and C<N>, the
225 "network" orders. These are guaranteed to be portable.
227 As of perl 5.8.5, you can also use the C<E<gt>> and C<E<lt>> modifiers
228 to force big- or little-endian byte-order. This is useful if you want
229 to store signed integers or 64-bit integers, for example.
231 You can explore the endianness of your platform by unpacking a
232 data structure packed in native format such as:
234 print unpack("h*", pack("s2", 1, 2)), "\n";
235 # '10002000' on e.g. Intel x86 or Alpha 21064 in little-endian mode
236 # '00100020' on e.g. Motorola 68040
238 If you need to distinguish between endian architectures you could use
239 either of the variables set like so:
241 $is_big_endian = unpack("h*", pack("s", 1)) =~ /01/;
242 $is_little_endian = unpack("h*", pack("s", 1)) =~ /^1/;
244 Differing widths can cause truncation even between platforms of equal
245 endianness. The platform of shorter width loses the upper parts of the
246 number. There is no good solution for this problem except to avoid
247 transferring or storing raw binary numbers.
249 One can circumnavigate both these problems in two ways. Either
250 transfer and store numbers always in text format, instead of raw
251 binary, or else consider using modules like Data::Dumper (included in
252 the standard distribution as of Perl 5.005) and Storable (included as
253 of perl 5.8). Keeping all data as text significantly simplifies matters.
255 The v-strings are portable only up to v2147483647 (0x7FFFFFFF), that's
256 how far EBCDIC, or more precisely UTF-EBCDIC will go.
258 =head2 Files and Filesystems
260 Most platforms these days structure files in a hierarchical fashion.
261 So, it is reasonably safe to assume that all platforms support the
262 notion of a "path" to uniquely identify a file on the system. How
263 that path is really written, though, differs considerably.
265 Although similar, file path specifications differ between Unix,
266 Windows, S<Mac OS>, OS/2, VMS, VOS, S<RISC OS>, and probably others.
267 Unix, for example, is one of the few OSes that has the elegant idea
268 of a single root directory.
270 DOS, OS/2, VMS, VOS, and Windows can work similarly to Unix with C</>
271 as path separator, or in their own idiosyncratic ways (such as having
272 several root directories and various "unrooted" device files such NIL:
275 S<Mac OS> uses C<:> as a path separator instead of C</>.
277 The filesystem may support neither hard links (C<link>) nor
278 symbolic links (C<symlink>, C<readlink>, C<lstat>).
280 The filesystem may support neither access timestamp nor change
281 timestamp (meaning that about the only portable timestamp is the
282 modification timestamp), or one second granularity of any timestamps
283 (e.g. the FAT filesystem limits the time granularity to two seconds).
285 The "inode change timestamp" (the C<-C> filetest) may really be the
286 "creation timestamp" (which it is not in UNIX).
288 VOS perl can emulate Unix filenames with C</> as path separator. The
289 native pathname characters greater-than, less-than, number-sign, and
290 percent-sign are always accepted.
292 S<RISC OS> perl can emulate Unix filenames with C</> as path
293 separator, or go native and use C<.> for path separator and C<:> to
294 signal filesystems and disk names.
296 Don't assume UNIX filesystem access semantics: that read, write,
297 and execute are all the permissions there are, and even if they exist,
298 that their semantics (for example what do r, w, and x mean on
299 a directory) are the UNIX ones. The various UNIX/POSIX compatibility
300 layers usually try to make interfaces like chmod() work, but sometimes
301 there simply is no good mapping.
303 If all this is intimidating, have no (well, maybe only a little)
304 fear. There are modules that can help. The File::Spec modules
305 provide methods to do the Right Thing on whatever platform happens
306 to be running the program.
308 use File::Spec::Functions;
309 chdir(updir()); # go up one directory
310 $file = catfile(curdir(), 'temp', 'file.txt');
311 # on Unix and Win32, './temp/file.txt'
312 # on Mac OS, ':temp:file.txt'
313 # on VMS, '[.temp]file.txt'
315 File::Spec is available in the standard distribution as of version
316 5.004_05. File::Spec::Functions is only in File::Spec 0.7 and later,
317 and some versions of perl come with version 0.6. If File::Spec
318 is not updated to 0.7 or later, you must use the object-oriented
319 interface from File::Spec (or upgrade File::Spec).
321 In general, production code should not have file paths hardcoded.
322 Making them user-supplied or read from a configuration file is
323 better, keeping in mind that file path syntax varies on different
326 This is especially noticeable in scripts like Makefiles and test suites,
327 which often assume C</> as a path separator for subdirectories.
329 Also of use is File::Basename from the standard distribution, which
330 splits a pathname into pieces (base filename, full path to directory,
333 Even when on a single platform (if you can call Unix a single platform),
334 remember not to count on the existence or the contents of particular
335 system-specific files or directories, like F</etc/passwd>,
336 F</etc/sendmail.conf>, F</etc/resolv.conf>, or even F</tmp/>. For
337 example, F</etc/passwd> may exist but not contain the encrypted
338 passwords, because the system is using some form of enhanced security.
339 Or it may not contain all the accounts, because the system is using NIS.
340 If code does need to rely on such a file, include a description of the
341 file and its format in the code's documentation, then make it easy for
342 the user to override the default location of the file.
344 Don't assume a text file will end with a newline. They should,
347 Do not have two files or directories of the same name with different
348 case, like F<test.pl> and F<Test.pl>, as many platforms have
349 case-insensitive (or at least case-forgiving) filenames. Also, try
350 not to have non-word characters (except for C<.>) in the names, and
351 keep them to the 8.3 convention, for maximum portability, onerous a
352 burden though this may appear.
354 Likewise, when using the AutoSplit module, try to keep your functions to
355 8.3 naming and case-insensitive conventions; or, at the least,
356 make it so the resulting files have a unique (case-insensitively)
359 Whitespace in filenames is tolerated on most systems, but not all,
360 and even on systems where it might be tolerated, some utilities
361 might become confused by such whitespace.
363 Many systems (DOS, VMS) cannot have more than one C<.> in their filenames.
365 Don't assume C<< > >> won't be the first character of a filename.
366 Always use C<< < >> explicitly to open a file for reading, or even
367 better, use the three-arg version of open, unless you want the user to
368 be able to specify a pipe open.
370 open(FILE, '<', $existing_file) or die $!;
372 If filenames might use strange characters, it is safest to open it
373 with C<sysopen> instead of C<open>. C<open> is magic and can
374 translate characters like C<< > >>, C<< < >>, and C<|>, which may
375 be the wrong thing to do. (Sometimes, though, it's the right thing.)
376 Three-arg open can also help protect against this translation in cases
377 where it is undesirable.
379 Don't use C<:> as a part of a filename since many systems use that for
380 their own semantics (Mac OS Classic for separating pathname components,
381 many networking schemes and utilities for separating the nodename and
382 the pathname, and so on). For the same reasons, avoid C<@>, C<;> and
385 Don't assume that in pathnames you can collapse two leading slashes
386 C<//> into one: some networking and clustering filesystems have special
387 semantics for that. Let the operating system to sort it out.
389 The I<portable filename characters> as defined by ANSI C are
391 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
392 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
396 and the "-" shouldn't be the first character. If you want to be
397 hypercorrect, stay case-insensitive and within the 8.3 naming
398 convention (all the files and directories have to be unique within one
399 directory if their names are lowercased and truncated to eight
400 characters before the C<.>, if any, and to three characters after the
401 C<.>, if any). (And do not use C<.>s in directory names.)
403 =head2 System Interaction
405 Not all platforms provide a command line. These are usually platforms
406 that rely primarily on a Graphical User Interface (GUI) for user
407 interaction. A program requiring a command line interface might
408 not work everywhere. This is probably for the user of the program
409 to deal with, so don't stay up late worrying about it.
411 Some platforms can't delete or rename files held open by the system,
412 this limitation may also apply to changing filesystem metainformation
413 like file permissions or owners. Remember to C<close> files when you
414 are done with them. Don't C<unlink> or C<rename> an open file. Don't
415 C<tie> or C<open> a file already tied or opened; C<untie> or C<close>
418 Don't open the same file more than once at a time for writing, as some
419 operating systems put mandatory locks on such files.
421 Don't assume that write/modify permission on a directory gives the
422 right to add or delete files/directories in that directory. That is
423 filesystem specific: in some filesystems you need write/modify
424 permission also (or even just) in the file/directory itself. In some
425 filesystems (AFS, DFS) the permission to add/delete directory entries
426 is a completely separate permission.
428 Don't assume that a single C<unlink> completely gets rid of the file:
429 some filesystems (most notably the ones in VMS) have versioned
430 filesystems, and unlink() removes only the most recent one (it doesn't
431 remove all the versions because by default the native tools on those
432 platforms remove just the most recent version, too). The portable
433 idiom to remove all the versions of a file is
435 1 while unlink "file";
437 This will terminate if the file is undeleteable for some reason
438 (protected, not there, and so on).
440 Don't count on a specific environment variable existing in C<%ENV>.
441 Don't count on C<%ENV> entries being case-sensitive, or even
442 case-preserving. Don't try to clear %ENV by saying C<%ENV = ();>, or,
443 if you really have to, make it conditional on C<$^O ne 'VMS'> since in
444 VMS the C<%ENV> table is much more than a per-process key-value string
447 Don't count on signals or C<%SIG> for anything.
449 Don't count on filename globbing. Use C<opendir>, C<readdir>, and
452 Don't count on per-program environment variables, or per-program current
455 Don't count on specific values of C<$!>, neither numeric nor
456 especially the strings values-- users may switch their locales causing
457 error messages to be translated into their languages. If you can
458 trust a POSIXish environment, you can portably use the symbols defined
459 by the Errno module, like ENOENT. And don't trust on the values of C<$!>
460 at all except immediately after a failed system call.
462 =head2 Command names versus file pathnames
464 Don't assume that the name used to invoke a command or program with
465 C<system> or C<exec> can also be used to test for the existence of the
466 file that holds the executable code for that command or program.
467 First, many systems have "internal" commands that are built-in to the
468 shell or OS and while these commands can be invoked, there is no
469 corresponding file. Second, some operating systems (e.g., Cygwin,
470 DJGPP, OS/2, and VOS) have required suffixes for executable files;
471 these suffixes are generally permitted on the command name but are not
472 required. Thus, a command like "perl" might exist in a file named
473 "perl", "perl.exe", or "perl.pm", depending on the operating system.
474 The variable "_exe" in the Config module holds the executable suffix,
475 if any. Third, the VMS port carefully sets up $^X and
476 $Config{perlpath} so that no further processing is required. This is
477 just as well, because the matching regular expression used below would
478 then have to deal with a possible trailing version number in the VMS
481 To convert $^X to a file pathname, taking account of the requirements
482 of the various operating system possibilities, say:
487 {$thisperl .= $Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;}
489 To convert $Config{perlpath} to a file pathname, say:
492 $thisperl = $Config{perlpath};
494 {$thisperl .= $Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;}
498 Don't assume that you can reach the public Internet.
500 Don't assume that there is only one way to get through firewalls
501 to the public Internet.
503 Don't assume that you can reach outside world through any other port
504 than 80, or some web proxy. ftp is blocked by many firewalls.
506 Don't assume that you can send email by connecting to the local SMTP port.
508 Don't assume that you can reach yourself or any node by the name
509 'localhost'. The same goes for '127.0.0.1'. You will have to try both.
511 Don't assume that the host has only one network card, or that it
512 can't bind to many virtual IP addresses.
514 Don't assume a particular network device name.
516 Don't assume a particular set of ioctl()s will work.
518 Don't assume that you can ping hosts and get replies.
520 Don't assume that any particular port (service) will respond.
522 Don't assume that Sys::Hostname (or any other API or command)
523 returns either a fully qualified hostname or a non-qualified hostname:
524 it all depends on how the system had been configured. Also remember
525 things like DHCP and NAT-- the hostname you get back might not be very
528 All the above "don't":s may look daunting, and they are -- but the key
529 is to degrade gracefully if one cannot reach the particular network
530 service one wants. Croaking or hanging do not look very professional.
532 =head2 Interprocess Communication (IPC)
534 In general, don't directly access the system in code meant to be
535 portable. That means, no C<system>, C<exec>, C<fork>, C<pipe>,
536 C<``>, C<qx//>, C<open> with a C<|>, nor any of the other things
537 that makes being a perl hacker worth being.
539 Commands that launch external processes are generally supported on
540 most platforms (though many of them do not support any type of
541 forking). The problem with using them arises from what you invoke
542 them on. External tools are often named differently on different
543 platforms, may not be available in the same location, might accept
544 different arguments, can behave differently, and often present their
545 results in a platform-dependent way. Thus, you should seldom depend
546 on them to produce consistent results. (Then again, if you're calling
547 I<netstat -a>, you probably don't expect it to run on both Unix and CP/M.)
549 One especially common bit of Perl code is opening a pipe to B<sendmail>:
551 open(MAIL, '|/usr/lib/sendmail -t')
552 or die "cannot fork sendmail: $!";
554 This is fine for systems programming when sendmail is known to be
555 available. But it is not fine for many non-Unix systems, and even
556 some Unix systems that may not have sendmail installed. If a portable
557 solution is needed, see the various distributions on CPAN that deal
558 with it. Mail::Mailer and Mail::Send in the MailTools distribution are
559 commonly used, and provide several mailing methods, including mail,
560 sendmail, and direct SMTP (via Net::SMTP) if a mail transfer agent is
561 not available. Mail::Sendmail is a standalone module that provides
562 simple, platform-independent mailing.
564 The Unix System V IPC (C<msg*(), sem*(), shm*()>) is not available
565 even on all Unix platforms.
567 Do not use either the bare result of C<pack("N", 10, 20, 30, 40)> or
568 bare v-strings (such as C<v10.20.30.40>) to represent IPv4 addresses:
569 both forms just pack the four bytes into network order. That this
570 would be equal to the C language C<in_addr> struct (which is what the
571 socket code internally uses) is not guaranteed. To be portable use
572 the routines of the Socket extension, such as C<inet_aton()>,
573 C<inet_ntoa()>, and C<sockaddr_in()>.
575 The rule of thumb for portable code is: Do it all in portable Perl, or
576 use a module (that may internally implement it with platform-specific
577 code, but expose a common interface).
579 =head2 External Subroutines (XS)
581 XS code can usually be made to work with any platform, but dependent
582 libraries, header files, etc., might not be readily available or
583 portable, or the XS code itself might be platform-specific, just as Perl
584 code might be. If the libraries and headers are portable, then it is
585 normally reasonable to make sure the XS code is portable, too.
587 A different type of portability issue arises when writing XS code:
588 availability of a C compiler on the end-user's system. C brings
589 with it its own portability issues, and writing XS code will expose
590 you to some of those. Writing purely in Perl is an easier way to
593 =head2 Standard Modules
595 In general, the standard modules work across platforms. Notable
596 exceptions are the CPAN module (which currently makes connections to external
597 programs that may not be available), platform-specific modules (like
598 ExtUtils::MM_VMS), and DBM modules.
600 There is no one DBM module available on all platforms.
601 SDBM_File and the others are generally available on all Unix and DOSish
602 ports, but not in MacPerl, where only NBDM_File and DB_File are
605 The good news is that at least some DBM module should be available, and
606 AnyDBM_File will use whichever module it can find. Of course, then
607 the code needs to be fairly strict, dropping to the greatest common
608 factor (e.g., not exceeding 1K for each record), so that it will
609 work with any DBM module. See L<AnyDBM_File> for more details.
613 The system's notion of time of day and calendar date is controlled in
614 widely different ways. Don't assume the timezone is stored in C<$ENV{TZ}>,
615 and even if it is, don't assume that you can control the timezone through
616 that variable. Don't assume anything about the three-letter timezone
617 abbreviations (for example that MST would be the Mountain Standard Time,
618 it's been known to stand for Moscow Standard Time). If you need to
619 use timezones, express them in some unambiguous format like the
620 exact number of minutes offset from UTC, or the POSIX timezone
623 Don't assume that the epoch starts at 00:00:00, January 1, 1970,
624 because that is OS- and implementation-specific. It is better to
625 store a date in an unambiguous representation. The ISO 8601 standard
626 defines YYYY-MM-DD as the date format, or YYYY-MM-DDTHH-MM-SS
627 (that's a literal "T" separating the date from the time).
628 Please do use the ISO 8601 instead of making us to guess what
629 date 02/03/04 might be. ISO 8601 even sorts nicely as-is.
630 A text representation (like "1987-12-18") can be easily converted
631 into an OS-specific value using a module like Date::Parse.
632 An array of values, such as those returned by C<localtime>, can be
633 converted to an OS-specific representation using Time::Local.
635 When calculating specific times, such as for tests in time or date modules,
636 it may be appropriate to calculate an offset for the epoch.
639 $offset = Time::Local::timegm(0, 0, 0, 1, 0, 70);
641 The value for C<$offset> in Unix will be C<0>, but in Mac OS will be
642 some large number. C<$offset> can then be added to a Unix time value
643 to get what should be the proper value on any system.
645 On Windows (at least), you shouldn't pass a negative value to C<gmtime> or
648 =head2 Character sets and character encoding
650 Assume very little about character sets.
652 Assume nothing about numerical values (C<ord>, C<chr>) of characters.
653 Do not use explicit code point ranges (like \xHH-\xHH); use for
654 example symbolic character classes like C<[:print:]>.
656 Do not assume that the alphabetic characters are encoded contiguously
657 (in the numeric sense). There may be gaps.
659 Do not assume anything about the ordering of the characters.
660 The lowercase letters may come before or after the uppercase letters;
661 the lowercase and uppercase may be interlaced so that both "a" and "A"
662 come before "b"; the accented and other international characters may
663 be interlaced so that E<auml> comes before "b".
665 =head2 Internationalisation
667 If you may assume POSIX (a rather large assumption), you may read
668 more about the POSIX locale system from L<perllocale>. The locale
669 system at least attempts to make things a little bit more portable,
670 or at least more convenient and native-friendly for non-English
671 users. The system affects character sets and encoding, and date
672 and time formatting--amongst other things.
674 If you really want to be international, you should consider Unicode.
675 See L<perluniintro> and L<perlunicode> for more information.
677 If you want to use non-ASCII bytes (outside the bytes 0x00..0x7f) in
678 the "source code" of your code, to be portable you have to be explicit
679 about what bytes they are. Someone might for example be using your
680 code under a UTF-8 locale, in which case random native bytes might be
681 illegal ("Malformed UTF-8 ...") This means that for example embedding
682 ISO 8859-1 bytes beyond 0x7f into your strings might cause trouble
683 later. If the bytes are native 8-bit bytes, you can use the C<bytes>
684 pragma. If the bytes are in a string (regular expression being a
685 curious string), you can often also use the C<\xHH> notation instead
686 of embedding the bytes as-is. If they are in some particular legacy
687 encoding (ether single-byte or something more complicated), you can
688 use the C<encoding> pragma. (If you want to write your code in UTF-8,
689 you can use either the C<utf8> pragma, or the C<encoding> pragma.)
690 The C<bytes> and C<utf8> pragmata are available since Perl 5.6.0, and
691 the C<encoding> pragma since Perl 5.8.0.
693 =head2 System Resources
695 If your code is destined for systems with severely constrained (or
696 missing!) virtual memory systems then you want to be I<especially> mindful
697 of avoiding wasteful constructs such as:
699 # NOTE: this is no longer "bad" in perl5.005
700 for (0..10000000) {} # bad
701 for (my $x = 0; $x <= 10000000; ++$x) {} # good
703 @lines = <VERY_LARGE_FILE>; # bad
705 while (<FILE>) {$file .= $_} # sometimes bad
706 $file = join('', <FILE>); # better
708 The last two constructs may appear unintuitive to most people. The
709 first repeatedly grows a string, whereas the second allocates a
710 large chunk of memory in one go. On some systems, the second is
711 more efficient that the first.
715 Most multi-user platforms provide basic levels of security, usually
716 implemented at the filesystem level. Some, however, do
717 not-- unfortunately. Thus the notion of user id, or "home" directory,
718 or even the state of being logged-in, may be unrecognizable on many
719 platforms. If you write programs that are security-conscious, it
720 is usually best to know what type of system you will be running
721 under so that you can write code explicitly for that platform (or
724 Don't assume the UNIX filesystem access semantics: the operating
725 system or the filesystem may be using some ACL systems, which are
726 richer languages than the usual rwx. Even if the rwx exist,
727 their semantics might be different.
729 (From security viewpoint testing for permissions before attempting to
730 do something is silly anyway: if one tries this, there is potential
731 for race conditions-- someone or something might change the
732 permissions between the permissions check and the actual operation.
733 Just try the operation.)
735 Don't assume the UNIX user and group semantics: especially, don't
736 expect the C<< $< >> and C<< $> >> (or the C<$(> and C<$)>) to work
737 for switching identities (or memberships).
739 Don't assume set-uid and set-gid semantics. (And even if you do,
740 think twice: set-uid and set-gid are a known can of security worms.)
744 For those times when it is necessary to have platform-specific code,
745 consider keeping the platform-specific code in one place, making porting
746 to other platforms easier. Use the Config module and the special
747 variable C<$^O> to differentiate platforms, as described in
750 Be careful in the tests you supply with your module or programs.
751 Module code may be fully portable, but its tests might not be. This
752 often happens when tests spawn off other processes or call external
753 programs to aid in the testing, or when (as noted above) the tests
754 assume certain things about the filesystem and paths. Be careful not
755 to depend on a specific output style for errors, such as when checking
756 C<$!> after a failed system call. Using C<$!> for anything else than
757 displaying it as output is doubtful (though see the Errno module for
758 testing reasonably portably for error value). Some platforms expect
759 a certain output format, and Perl on those platforms may have been
760 adjusted accordingly. Most specifically, don't anchor a regex when
761 testing an error value.
765 Modules uploaded to CPAN are tested by a variety of volunteers on
766 different platforms. These CPAN testers are notified by mail of each
767 new upload, and reply to the list with PASS, FAIL, NA (not applicable to
768 this platform), or UNKNOWN (unknown), along with any relevant notations.
770 The purpose of the testing is twofold: one, to help developers fix any
771 problems in their code that crop up because of lack of testing on other
772 platforms; two, to provide users with information about whether
773 a given module works on a given platform.
781 Mailing list: cpan-testers@perl.org
785 Testing results: http://testers.cpan.org/
791 As of version 5.002, Perl is built with a C<$^O> variable that
792 indicates the operating system it was built on. This was implemented
793 to help speed up code that would otherwise have to C<use Config>
794 and use the value of C<$Config{osname}>. Of course, to get more
795 detailed information about the system, looking into C<%Config> is
796 certainly recommended.
798 C<%Config> cannot always be trusted, however, because it was built
799 at compile time. If perl was built in one place, then transferred
800 elsewhere, some values may be wrong. The values may even have been
801 edited after the fact.
805 Perl works on a bewildering variety of Unix and Unix-like platforms (see
806 e.g. most of the files in the F<hints/> directory in the source code kit).
807 On most of these systems, the value of C<$^O> (hence C<$Config{'osname'}>,
808 too) is determined either by lowercasing and stripping punctuation from the
809 first field of the string returned by typing C<uname -a> (or a similar command)
810 at the shell prompt or by testing the file system for the presence of
811 uniquely named files such as a kernel or header file. Here, for example,
812 are a few of the more popular Unix flavors:
814 uname $^O $Config{'archname'}
815 --------------------------------------------
817 BSD/OS bsdos i386-bsdos
819 dgux dgux AViiON-dgux
820 DYNIX/ptx dynixptx i386-dynixptx
821 FreeBSD freebsd freebsd-i386
822 Linux linux arm-linux
823 Linux linux i386-linux
824 Linux linux i586-linux
825 Linux linux ppc-linux
826 HP-UX hpux PA-RISC1.1
828 Mac OS X darwin darwin
829 MachTen PPC machten powerpc-machten
831 NeXT 4 next OPENSTEP-Mach
832 openbsd openbsd i386-openbsd
833 OSF1 dec_osf alpha-dec_osf
834 reliantunix-n svr4 RM400-svr4
835 SCO_SV sco_sv i386-sco_sv
836 SINIX-N svr4 RM400-svr4
837 sn4609 unicos CRAY_C90-unicos
838 sn6521 unicosmk t3e-unicosmk
839 sn9617 unicos CRAY_J90-unicos
840 SunOS solaris sun4-solaris
841 SunOS solaris i86pc-solaris
842 SunOS4 sunos sun4-sunos
844 Because the value of C<$Config{archname}> may depend on the
845 hardware architecture, it can vary more than the value of C<$^O>.
847 =head2 DOS and Derivatives
849 Perl has long been ported to Intel-style microcomputers running under
850 systems like PC-DOS, MS-DOS, OS/2, and most Windows platforms you can
851 bring yourself to mention (except for Windows CE, if you count that).
852 Users familiar with I<COMMAND.COM> or I<CMD.EXE> style shells should
853 be aware that each of these file specifications may have subtle
856 $filespec0 = "c:/foo/bar/file.txt";
857 $filespec1 = "c:\\foo\\bar\\file.txt";
858 $filespec2 = 'c:\foo\bar\file.txt';
859 $filespec3 = 'c:\\foo\\bar\\file.txt';
861 System calls accept either C</> or C<\> as the path separator.
862 However, many command-line utilities of DOS vintage treat C</> as
863 the option prefix, so may get confused by filenames containing C</>.
864 Aside from calling any external programs, C</> will work just fine,
865 and probably better, as it is more consistent with popular usage,
866 and avoids the problem of remembering what to backwhack and what
869 The DOS FAT filesystem can accommodate only "8.3" style filenames. Under
870 the "case-insensitive, but case-preserving" HPFS (OS/2) and NTFS (NT)
871 filesystems you may have to be careful about case returned with functions
872 like C<readdir> or used with functions like C<open> or C<opendir>.
874 DOS also treats several filenames as special, such as AUX, PRN,
875 NUL, CON, COM1, LPT1, LPT2, etc. Unfortunately, sometimes these
876 filenames won't even work if you include an explicit directory
877 prefix. It is best to avoid such filenames, if you want your code
878 to be portable to DOS and its derivatives. It's hard to know what
879 these all are, unfortunately.
881 Users of these operating systems may also wish to make use of
882 scripts such as I<pl2bat.bat> or I<pl2cmd> to
883 put wrappers around your scripts.
885 Newline (C<\n>) is translated as C<\015\012> by STDIO when reading from
886 and writing to files (see L<"Newlines">). C<binmode(FILEHANDLE)>
887 will keep C<\n> translated as C<\012> for that filehandle. Since it is a
888 no-op on other systems, C<binmode> should be used for cross-platform code
889 that deals with binary data. That's assuming you realize in advance
890 that your data is in binary. General-purpose programs should
891 often assume nothing about their data.
893 The C<$^O> variable and the C<$Config{archname}> values for various
894 DOSish perls are as follows:
896 OS $^O $Config{archname} ID Version
897 --------------------------------------------------------
901 Windows 3.1 ? ? 0 3 01
902 Windows 95 MSWin32 MSWin32-x86 1 4 00
903 Windows 98 MSWin32 MSWin32-x86 1 4 10
904 Windows ME MSWin32 MSWin32-x86 1 ?
905 Windows NT MSWin32 MSWin32-x86 2 4 xx
906 Windows NT MSWin32 MSWin32-ALPHA 2 4 xx
907 Windows NT MSWin32 MSWin32-ppc 2 4 xx
908 Windows 2000 MSWin32 MSWin32-x86 2 5 00
909 Windows XP MSWin32 MSWin32-x86 2 5 01
910 Windows 2003 MSWin32 MSWin32-x86 2 5 02
911 Windows CE MSWin32 ? 3
914 The various MSWin32 Perl's can distinguish the OS they are running on
915 via the value of the fifth element of the list returned from
916 Win32::GetOSVersion(). For example:
918 if ($^O eq 'MSWin32') {
919 my @os_version_info = Win32::GetOSVersion();
920 print +('3.1','95','NT')[$os_version_info[4]],"\n";
923 There are also Win32::IsWinNT() and Win32::IsWin95(), try C<perldoc Win32>,
924 and as of libwin32 0.19 (not part of the core Perl distribution)
925 Win32::GetOSName(). The very portable POSIX::uname() will work too:
927 c:\> perl -MPOSIX -we "print join '|', uname"
928 Windows NT|moonru|5.0|Build 2195 (Service Pack 2)|x86
936 The djgpp environment for DOS, http://www.delorie.com/djgpp/
941 The EMX environment for DOS, OS/2, etc. emx@iaehv.nl,
942 http://www.leo.org/pub/comp/os/os2/leo/gnu/emx+gcc/index.html or
943 ftp://hobbes.nmsu.edu/pub/os2/dev/emx/ Also L<perlos2>.
947 Build instructions for Win32 in L<perlwin32>, or under the Cygnus environment
952 The C<Win32::*> modules in L<Win32>.
956 The ActiveState Pages, http://www.activestate.com/
960 The Cygwin environment for Win32; F<README.cygwin> (installed
961 as L<perlcygwin>), http://www.cygwin.com/
965 The U/WIN environment for Win32,
966 http://www.research.att.com/sw/tools/uwin/
970 Build instructions for OS/2, L<perlos2>
976 Any module requiring XS compilation is right out for most people, because
977 MacPerl is built using non-free (and non-cheap!) compilers. Some XS
978 modules that can work with MacPerl are built and distributed in binary
981 Directories are specified as:
983 volume:folder:file for absolute pathnames
984 volume:folder: for absolute pathnames
985 :folder:file for relative pathnames
986 :folder: for relative pathnames
987 :file for relative pathnames
988 file for relative pathnames
990 Files are stored in the directory in alphabetical order. Filenames are
991 limited to 31 characters, and may include any character except for
992 null and C<:>, which is reserved as the path separator.
994 Instead of C<flock>, see C<FSpSetFLock> and C<FSpRstFLock> in the
995 Mac::Files module, or C<chmod(0444, ...)> and C<chmod(0666, ...)>.
997 In the MacPerl application, you can't run a program from the command line;
998 programs that expect C<@ARGV> to be populated can be edited with something
999 like the following, which brings up a dialog box asking for the command
1003 @ARGV = split /\s+/, MacPerl::Ask('Arguments?');
1006 A MacPerl script saved as a "droplet" will populate C<@ARGV> with the full
1007 pathnames of the files dropped onto the script.
1009 Mac users can run programs under a type of command line interface
1010 under MPW (Macintosh Programmer's Workshop, a free development
1011 environment from Apple). MacPerl was first introduced as an MPW
1012 tool, and MPW can be used like a shell:
1014 perl myscript.plx some arguments
1016 ToolServer is another app from Apple that provides access to MPW tools
1017 from MPW and the MacPerl app, which allows MacPerl programs to use
1018 C<system>, backticks, and piped C<open>.
1020 "S<Mac OS>" is the proper name for the operating system, but the value
1021 in C<$^O> is "MacOS". To determine architecture, version, or whether
1022 the application or MPW tool version is running, check:
1024 $is_app = $MacPerl::Version =~ /App/;
1025 $is_tool = $MacPerl::Version =~ /MPW/;
1026 ($version) = $MacPerl::Version =~ /^(\S+)/;
1027 $is_ppc = $MacPerl::Architecture eq 'MacPPC';
1028 $is_68k = $MacPerl::Architecture eq 'Mac68K';
1030 S<Mac OS X>, based on NeXT's OpenStep OS, runs MacPerl natively, under the
1031 "Classic" environment. There is no "Carbon" version of MacPerl to run
1032 under the primary Mac OS X environment. S<Mac OS X> and its Open Source
1033 version, Darwin, both run Unix perl natively.
1041 MacPerl Development, http://dev.macperl.org/ .
1045 The MacPerl Pages, http://www.macperl.com/ .
1049 The MacPerl mailing lists, http://lists.perl.org/ .
1053 MPW, ftp://ftp.apple.com/developer/Tool_Chest/Core_Mac_OS_Tools/
1059 Perl on VMS is discussed in L<perlvms> in the perl distribution.
1060 Perl on VMS can accept either VMS- or Unix-style file
1061 specifications as in either of the following:
1063 $ perl -ne "print if /perl_setup/i" SYS$LOGIN:LOGIN.COM
1064 $ perl -ne "print if /perl_setup/i" /sys$login/login.com
1066 but not a mixture of both as in:
1068 $ perl -ne "print if /perl_setup/i" sys$login:/login.com
1069 Can't open sys$login:/login.com: file specification syntax error
1071 Interacting with Perl from the Digital Command Language (DCL) shell
1072 often requires a different set of quotation marks than Unix shells do.
1075 $ perl -e "print ""Hello, world.\n"""
1078 There are several ways to wrap your perl scripts in DCL F<.COM> files, if
1079 you are so inclined. For example:
1081 $ write sys$output "Hello from DCL!"
1083 $ then perl -x 'f$environment("PROCEDURE")
1084 $ else perl -x - 'p1 'p2 'p3 'p4 'p5 'p6 'p7 'p8
1085 $ deck/dollars="__END__"
1088 print "Hello from Perl!\n";
1093 Do take care with C<$ ASSIGN/nolog/user SYS$COMMAND: SYS$INPUT> if your
1094 perl-in-DCL script expects to do things like C<< $read = <STDIN>; >>.
1096 Filenames are in the format "name.extension;version". The maximum
1097 length for filenames is 39 characters, and the maximum length for
1098 extensions is also 39 characters. Version is a number from 1 to
1099 32767. Valid characters are C</[A-Z0-9$_-]/>.
1101 VMS's RMS filesystem is case-insensitive and does not preserve case.
1102 C<readdir> returns lowercased filenames, but specifying a file for
1103 opening remains case-insensitive. Files without extensions have a
1104 trailing period on them, so doing a C<readdir> with a file named F<A.;5>
1105 will return F<a.> (though that file could be opened with
1108 RMS had an eight level limit on directory depths from any rooted logical
1109 (allowing 16 levels overall) prior to VMS 7.2. Hence
1110 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8]> is a valid directory specification but
1111 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8.9]> is not. F<Makefile.PL> authors might
1112 have to take this into account, but at least they can refer to the former
1113 as C</PERL_ROOT/lib/2/3/4/5/6/7/8/>.
1115 The VMS::Filespec module, which gets installed as part of the build
1116 process on VMS, is a pure Perl module that can easily be installed on
1117 non-VMS platforms and can be helpful for conversions to and from RMS
1120 What C<\n> represents depends on the type of file opened. It usually
1121 represents C<\012> but it could also be C<\015>, C<\012>, C<\015\012>,
1122 C<\000>, C<\040>, or nothing depending on the file organization and
1123 record format. The VMS::Stdio module provides access to the
1124 special fopen() requirements of files with unusual attributes on VMS.
1126 TCP/IP stacks are optional on VMS, so socket routines might not be
1127 implemented. UDP sockets may not be supported.
1129 The value of C<$^O> on OpenVMS is "VMS". To determine the architecture
1130 that you are running on without resorting to loading all of C<%Config>
1131 you can examine the content of the C<@INC> array like so:
1133 if (grep(/VMS_AXP/, @INC)) {
1134 print "I'm on Alpha!\n";
1136 } elsif (grep(/VMS_VAX/, @INC)) {
1137 print "I'm on VAX!\n";
1140 print "I'm not so sure about where $^O is...\n";
1143 On VMS, perl determines the UTC offset from the C<SYS$TIMEZONE_DIFFERENTIAL>
1144 logical name. Although the VMS epoch began at 17-NOV-1858 00:00:00.00,
1145 calls to C<localtime> are adjusted to count offsets from
1146 01-JAN-1970 00:00:00.00, just like Unix.
1154 F<README.vms> (installed as L<README_vms>), L<perlvms>
1158 vmsperl list, majordomo@perl.org
1160 (Put the words C<subscribe vmsperl> in message body.)
1164 vmsperl on the web, http://www.sidhe.org/vmsperl/index.html
1170 Perl on VOS is discussed in F<README.vos> in the perl distribution
1171 (installed as L<perlvos>). Perl on VOS can accept either VOS- or
1172 Unix-style file specifications as in either of the following:
1174 C<< $ perl -ne "print if /perl_setup/i" >system>notices >>
1175 C<< $ perl -ne "print if /perl_setup/i" /system/notices >>
1177 or even a mixture of both as in:
1179 C<< $ perl -ne "print if /perl_setup/i" >system/notices >>
1181 Even though VOS allows the slash character to appear in object
1182 names, because the VOS port of Perl interprets it as a pathname
1183 delimiting character, VOS files, directories, or links whose names
1184 contain a slash character cannot be processed. Such files must be
1185 renamed before they can be processed by Perl. Note that VOS limits
1186 file names to 32 or fewer characters.
1188 Perl on VOS can be built using two different compilers and two different
1189 versions of the POSIX runtime. The recommended method for building full
1190 Perl is with the GNU C compiler and the generally-available version of
1191 VOS POSIX support. See F<README.vos> (installed as L<perlvos>) for
1192 restrictions that apply when Perl is built using the VOS Standard C
1193 compiler or the alpha version of VOS POSIX support.
1195 The value of C<$^O> on VOS is "VOS". To determine the architecture that
1196 you are running on without resorting to loading all of C<%Config> you
1197 can examine the content of the @INC array like so:
1200 print "I'm on a Stratus box!\n";
1202 print "I'm not on a Stratus box!\n";
1206 if (grep(/860/, @INC)) {
1207 print "This box is a Stratus XA/R!\n";
1209 } elsif (grep(/7100/, @INC)) {
1210 print "This box is a Stratus HP 7100 or 8xxx!\n";
1212 } elsif (grep(/8000/, @INC)) {
1213 print "This box is a Stratus HP 8xxx!\n";
1216 print "This box is a Stratus 68K!\n";
1225 F<README.vos> (installed as L<perlvos>)
1229 The VOS mailing list.
1231 There is no specific mailing list for Perl on VOS. You can post
1232 comments to the comp.sys.stratus newsgroup, or subscribe to the general
1233 Stratus mailing list. Send a letter with "subscribe Info-Stratus" in
1234 the message body to majordomo@list.stratagy.com.
1238 VOS Perl on the web at http://ftp.stratus.com/pub/vos/posix/posix.html
1242 =head2 EBCDIC Platforms
1244 Recent versions of Perl have been ported to platforms such as OS/400 on
1245 AS/400 minicomputers as well as OS/390, VM/ESA, and BS2000 for S/390
1246 Mainframes. Such computers use EBCDIC character sets internally (usually
1247 Character Code Set ID 0037 for OS/400 and either 1047 or POSIX-BC for S/390
1248 systems). On the mainframe perl currently works under the "Unix system
1249 services for OS/390" (formerly known as OpenEdition), VM/ESA OpenEdition, or
1250 the BS200 POSIX-BC system (BS2000 is supported in perl 5.6 and greater).
1251 See L<perlos390> for details. Note that for OS/400 there is also a port of
1252 Perl 5.8.1/5.9.0 or later to the PASE which is ASCII-based (as opposed to
1253 ILE which is EBCDIC-based), see L<perlos400>.
1255 As of R2.5 of USS for OS/390 and Version 2.3 of VM/ESA these Unix
1256 sub-systems do not support the C<#!> shebang trick for script invocation.
1257 Hence, on OS/390 and VM/ESA perl scripts can be executed with a header
1258 similar to the following simple script:
1261 eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
1263 #!/usr/local/bin/perl # just a comment really
1265 print "Hello from perl!\n";
1267 OS/390 will support the C<#!> shebang trick in release 2.8 and beyond.
1268 Calls to C<system> and backticks can use POSIX shell syntax on all
1271 On the AS/400, if PERL5 is in your library list, you may need
1272 to wrap your perl scripts in a CL procedure to invoke them like so:
1275 CALL PGM(PERL5/PERL) PARM('/QOpenSys/hello.pl')
1278 This will invoke the perl script F<hello.pl> in the root of the
1279 QOpenSys file system. On the AS/400 calls to C<system> or backticks
1282 On these platforms, bear in mind that the EBCDIC character set may have
1283 an effect on what happens with some perl functions (such as C<chr>,
1284 C<pack>, C<print>, C<printf>, C<ord>, C<sort>, C<sprintf>, C<unpack>), as
1285 well as bit-fiddling with ASCII constants using operators like C<^>, C<&>
1286 and C<|>, not to mention dealing with socket interfaces to ASCII computers
1287 (see L<"Newlines">).
1289 Fortunately, most web servers for the mainframe will correctly
1290 translate the C<\n> in the following statement to its ASCII equivalent
1291 (C<\r> is the same under both Unix and OS/390 & VM/ESA):
1293 print "Content-type: text/html\r\n\r\n";
1295 The values of C<$^O> on some of these platforms includes:
1297 uname $^O $Config{'archname'}
1298 --------------------------------------------
1301 POSIX-BC posix-bc BS2000-posix-bc
1304 Some simple tricks for determining if you are running on an EBCDIC
1305 platform could include any of the following (perhaps all):
1307 if ("\t" eq "\05") { print "EBCDIC may be spoken here!\n"; }
1309 if (ord('A') == 193) { print "EBCDIC may be spoken here!\n"; }
1311 if (chr(169) eq 'z') { print "EBCDIC may be spoken here!\n"; }
1313 One thing you may not want to rely on is the EBCDIC encoding
1314 of punctuation characters since these may differ from code page to code
1315 page (and once your module or script is rumoured to work with EBCDIC,
1316 folks will want it to work with all EBCDIC character sets).
1324 L<perlos390>, F<README.os390>, F<perlbs2000>, F<README.vmesa>,
1329 The perl-mvs@perl.org list is for discussion of porting issues as well as
1330 general usage issues for all EBCDIC Perls. Send a message body of
1331 "subscribe perl-mvs" to majordomo@perl.org.
1335 AS/400 Perl information at
1336 http://as400.rochester.ibm.com/
1337 as well as on CPAN in the F<ports/> directory.
1341 =head2 Acorn RISC OS
1343 Because Acorns use ASCII with newlines (C<\n>) in text files as C<\012> like
1344 Unix, and because Unix filename emulation is turned on by default,
1345 most simple scripts will probably work "out of the box". The native
1346 filesystem is modular, and individual filesystems are free to be
1347 case-sensitive or insensitive, and are usually case-preserving. Some
1348 native filesystems have name length limits, which file and directory
1349 names are silently truncated to fit. Scripts should be aware that the
1350 standard filesystem currently has a name length limit of B<10>
1351 characters, with up to 77 items in a directory, but other filesystems
1352 may not impose such limitations.
1354 Native filenames are of the form
1356 Filesystem#Special_Field::DiskName.$.Directory.Directory.File
1360 Special_Field is not usually present, but may contain . and $ .
1361 Filesystem =~ m|[A-Za-z0-9_]|
1362 DsicName =~ m|[A-Za-z0-9_/]|
1363 $ represents the root directory
1364 . is the path separator
1365 @ is the current directory (per filesystem but machine global)
1366 ^ is the parent directory
1367 Directory and File =~ m|[^\0- "\.\$\%\&:\@\\^\|\177]+|
1369 The default filename translation is roughly C<tr|/.|./|;>
1371 Note that C<"ADFS::HardDisk.$.File" ne 'ADFS::HardDisk.$.File'> and that
1372 the second stage of C<$> interpolation in regular expressions will fall
1373 foul of the C<$.> if scripts are not careful.
1375 Logical paths specified by system variables containing comma-separated
1376 search lists are also allowed; hence C<System:Modules> is a valid
1377 filename, and the filesystem will prefix C<Modules> with each section of
1378 C<System$Path> until a name is made that points to an object on disk.
1379 Writing to a new file C<System:Modules> would be allowed only if
1380 C<System$Path> contains a single item list. The filesystem will also
1381 expand system variables in filenames if enclosed in angle brackets, so
1382 C<< <System$Dir>.Modules >> would look for the file
1383 S<C<$ENV{'System$Dir'} . 'Modules'>>. The obvious implication of this is
1384 that B<fully qualified filenames can start with C<< <> >>> and should
1385 be protected when C<open> is used for input.
1387 Because C<.> was in use as a directory separator and filenames could not
1388 be assumed to be unique after 10 characters, Acorn implemented the C
1389 compiler to strip the trailing C<.c> C<.h> C<.s> and C<.o> suffix from
1390 filenames specified in source code and store the respective files in
1391 subdirectories named after the suffix. Hence files are translated:
1394 C:foo.h C:h.foo (logical path variable)
1395 sys/os.h sys.h.os (C compiler groks Unix-speak)
1396 10charname.c c.10charname
1397 10charname.o o.10charname
1398 11charname_.c c.11charname (assuming filesystem truncates at 10)
1400 The Unix emulation library's translation of filenames to native assumes
1401 that this sort of translation is required, and it allows a user-defined list
1402 of known suffixes that it will transpose in this fashion. This may
1403 seem transparent, but consider that with these rules C<foo/bar/baz.h>
1404 and C<foo/bar/h/baz> both map to C<foo.bar.h.baz>, and that C<readdir> and
1405 C<glob> cannot and do not attempt to emulate the reverse mapping. Other
1406 C<.>'s in filenames are translated to C</>.
1408 As implied above, the environment accessed through C<%ENV> is global, and
1409 the convention is that program specific environment variables are of the
1410 form C<Program$Name>. Each filesystem maintains a current directory,
1411 and the current filesystem's current directory is the B<global> current
1412 directory. Consequently, sociable programs don't change the current
1413 directory but rely on full pathnames, and programs (and Makefiles) cannot
1414 assume that they can spawn a child process which can change the current
1415 directory without affecting its parent (and everyone else for that
1418 Because native operating system filehandles are global and are currently
1419 allocated down from 255, with 0 being a reserved value, the Unix emulation
1420 library emulates Unix filehandles. Consequently, you can't rely on
1421 passing C<STDIN>, C<STDOUT>, or C<STDERR> to your children.
1423 The desire of users to express filenames of the form
1424 C<< <Foo$Dir>.Bar >> on the command line unquoted causes problems,
1425 too: C<``> command output capture has to perform a guessing game. It
1426 assumes that a string C<< <[^<>]+\$[^<>]> >> is a
1427 reference to an environment variable, whereas anything else involving
1428 C<< < >> or C<< > >> is redirection, and generally manages to be 99%
1429 right. Of course, the problem remains that scripts cannot rely on any
1430 Unix tools being available, or that any tools found have Unix-like command
1433 Extensions and XS are, in theory, buildable by anyone using free
1434 tools. In practice, many don't, as users of the Acorn platform are
1435 used to binary distributions. MakeMaker does run, but no available
1436 make currently copes with MakeMaker's makefiles; even if and when
1437 this should be fixed, the lack of a Unix-like shell will cause
1438 problems with makefile rules, especially lines of the form C<cd
1439 sdbm && make all>, and anything using quoting.
1441 "S<RISC OS>" is the proper name for the operating system, but the value
1442 in C<$^O> is "riscos" (because we don't like shouting).
1446 Perl has been ported to many platforms that do not fit into any of
1447 the categories listed above. Some, such as AmigaOS, Atari MiNT,
1448 BeOS, HP MPE/iX, QNX, Plan 9, and VOS, have been well-integrated
1449 into the standard Perl source code kit. You may need to see the
1450 F<ports/> directory on CPAN for information, and possibly binaries,
1451 for the likes of: aos, Atari ST, lynxos, riscos, Novell Netware,
1452 Tandem Guardian, I<etc.> (Yes, we know that some of these OSes may
1453 fall under the Unix category, but we are not a standards body.)
1455 Some approximate operating system names and their C<$^O> values
1456 in the "OTHER" category include:
1458 OS $^O $Config{'archname'}
1459 ------------------------------------------
1460 Amiga DOS amigaos m68k-amigos
1462 MPE/iX mpeix PA-RISC1.1
1470 Amiga, F<README.amiga> (installed as L<perlamiga>).
1474 Atari, F<README.mint> and Guido Flohr's web page
1475 http://stud.uni-sb.de/~gufl0000/
1479 Be OS, F<README.beos>
1483 HP 300 MPE/iX, F<README.mpeix> and Mark Bixby's web page
1484 http://www.bixby.org/mark/perlix.html
1488 A free perl5-based PERL.NLM for Novell Netware is available in
1489 precompiled binary and source code form from http://www.novell.com/
1490 as well as from CPAN.
1494 S<Plan 9>, F<README.plan9>
1498 =head1 FUNCTION IMPLEMENTATIONS
1500 Listed below are functions that are either completely unimplemented
1501 or else have been implemented differently on various platforms.
1502 Following each description will be, in parentheses, a list of
1503 platforms that the description applies to.
1505 The list may well be incomplete, or even wrong in some places. When
1506 in doubt, consult the platform-specific README files in the Perl
1507 source distribution, and any other documentation resources accompanying
1510 Be aware, moreover, that even among Unix-ish systems there are variations.
1512 For many functions, you can also query C<%Config>, exported by
1513 default from the Config module. For example, to check whether the
1514 platform has the C<lstat> call, check C<$Config{d_lstat}>. See
1515 L<Config> for a full description of available variables.
1517 =head2 Alphabetical Listing of Perl Functions
1523 C<-r>, C<-w>, and C<-x> have a limited meaning only; directories
1524 and applications are executable, and there are no uid/gid
1525 considerations. C<-o> is not supported. (S<Mac OS>)
1527 C<-r>, C<-w>, C<-x>, and C<-o> tell whether the file is accessible,
1528 which may not reflect UIC-based file protections. (VMS)
1530 C<-s> returns the size of the data fork, not the total size of data fork
1531 plus resource fork. (S<Mac OS>).
1533 C<-s> by name on an open file will return the space reserved on disk,
1534 rather than the current extent. C<-s> on an open filehandle returns the
1535 current size. (S<RISC OS>)
1537 C<-R>, C<-W>, C<-X>, C<-O> are indistinguishable from C<-r>, C<-w>,
1538 C<-x>, C<-o>. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1540 C<-b>, C<-c>, C<-k>, C<-g>, C<-p>, C<-u>, C<-A> are not implemented.
1543 C<-g>, C<-k>, C<-l>, C<-p>, C<-u>, C<-A> are not particularly meaningful.
1544 (Win32, VMS, S<RISC OS>)
1546 C<-d> is true if passed a device spec without an explicit directory.
1549 C<-T> and C<-B> are implemented, but might misclassify Mac text files
1550 with foreign characters; this is the case will all platforms, but may
1551 affect S<Mac OS> often. (S<Mac OS>)
1553 C<-x> (or C<-X>) determine if a file ends in one of the executable
1554 suffixes. C<-S> is meaningless. (Win32)
1556 C<-x> (or C<-X>) determine if a file has an executable file type.
1561 Due to issues with various CPUs, math libraries, compilers, and standards,
1562 results for C<atan2()> may vary depending on any combination of the above.
1563 Perl attempts to conform to the Open Group/IEEE standards for the results
1564 returned from C<atan2()>, but cannot force the issue if the system Perl is
1565 run on does not allow it. (Tru64, HP-UX 10.20)
1567 The current version of the standards for C<atan2()> is available at
1568 L<http://www.opengroup.org/onlinepubs/009695399/functions/atan2.html>.
1572 Due to issues with various CPUs, math libraries, compilers, and standards,
1573 results for C<atan2()> may vary depending on any combination of the above.
1574 Perl attempts to conform to the Open Group/IEEE standards for the results
1575 returned from C<atan2()>, but cannot force the issue if the system Perl is
1576 run on does not allow it. (Tru64, HP-UX 10.20)
1578 The current version of the standards for C<atan2()> is available at
1579 L<http://www.opengroup.org/onlinepubs/009695399/functions/atan2.html>.
1583 Meaningless. (S<Mac OS>, S<RISC OS>)
1585 Reopens file and restores pointer; if function fails, underlying
1586 filehandle may be closed, or pointer may be in a different position.
1589 The value returned by C<tell> may be affected after the call, and
1590 the filehandle may be flushed. (Win32)
1594 Only limited meaning. Disabling/enabling write permission is mapped to
1595 locking/unlocking the file. (S<Mac OS>)
1597 Only good for changing "owner" read-write access, "group", and "other"
1598 bits are meaningless. (Win32)
1600 Only good for changing "owner" and "other" read-write access. (S<RISC OS>)
1602 Access permissions are mapped onto VOS access-control list changes. (VOS)
1604 The actual permissions set depend on the value of the C<CYGWIN>
1605 in the SYSTEM environment settings. (Cygwin)
1609 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>, VOS)
1611 Does nothing, but won't fail. (Win32)
1615 Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS, VM/ESA)
1619 May not be available if library or source was not provided when building
1622 Not implemented. (VOS)
1626 Not implemented. (VMS, S<Plan 9>, VOS)
1630 Not implemented. (VMS, S<Plan 9>, VOS)
1634 Not useful. (S<Mac OS>, S<RISC OS>)
1636 Not implemented. (Win32)
1638 Invokes VMS debugger. (VMS)
1642 Not implemented. (S<Mac OS>)
1644 Implemented via Spawn. (VM/ESA)
1646 Does not automatically flush output handles on some platforms.
1647 (SunOS, Solaris, HP-UX)
1651 Emulates UNIX exit() (which considers C<exit 1> to indicate an error) by
1652 mapping the C<1> to SS$_ABORT (C<44>). This behavior may be overridden
1653 with the pragma C<use vmsish 'exit'>. As with the CRTL's exit()
1654 function, C<exit 0> is also mapped to an exit status of SS$_NORMAL
1655 (C<1>); this mapping cannot be overridden. Any other argument to exit()
1656 is used directly as Perl's exit status. (VMS)
1660 Not implemented. (Win32, VMS)
1664 Not implemented (S<Mac OS>, VMS, S<RISC OS>, VOS).
1666 Available only on Windows NT (not on Windows 95). (Win32)
1670 Not implemented. (S<Mac OS>, AmigaOS, S<RISC OS>, VOS, VM/ESA, VMS)
1672 Emulated using multiple interpreters. See L<perlfork>. (Win32)
1674 Does not automatically flush output handles on some platforms.
1675 (SunOS, Solaris, HP-UX)
1679 Not implemented. (S<Mac OS>, S<RISC OS>)
1683 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1687 Not implemented. (S<Mac OS>, Win32, S<RISC OS>)
1691 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1695 Not implemented. (S<Mac OS>, Win32)
1697 Not useful. (S<RISC OS>)
1701 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1705 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1709 Not implemented. (S<Mac OS>, Win32)
1711 Not useful. (S<RISC OS>)
1715 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1719 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1721 =item getprotobynumber
1723 Not implemented. (S<Mac OS>)
1727 Not implemented. (S<Mac OS>)
1731 Not implemented. (S<Mac OS>, Win32, VM/ESA)
1735 Not implemented. (S<Mac OS>, Win32, VMS, VM/ESA)
1739 C<gethostbyname('localhost')> does not work everywhere: you may have
1740 to use C<gethostbyname('127.0.0.1')>. (S<Mac OS>, S<Irix 5>)
1744 Not implemented. (S<Mac OS>, Win32)
1748 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1752 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1756 Not implemented. (Win32, S<Plan 9>)
1760 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1764 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1768 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1772 Not implemented. (S<Plan 9>, Win32, S<RISC OS>)
1776 Not implemented. (S<Mac OS>, MPE/iX, VM/ESA, Win32)
1780 Not implemented. (S<Mac OS>, MPE/iX, S<RISC OS>, VM/ESA, VMS, Win32)
1784 Not implemented. (S<Mac OS>, Win32)
1788 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1792 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1796 Not implemented. (S<Plan 9>, Win32)
1798 =item getsockopt SOCKET,LEVEL,OPTNAME
1800 Not implemented. (S<Plan 9>)
1804 This operator is implemented via the File::Glob extension on most
1805 platforms. See L<File::Glob> for portability information.
1809 Same portability caveats as L<localtime>.
1811 =item ioctl FILEHANDLE,FUNCTION,SCALAR
1813 Not implemented. (VMS)
1815 Available only for socket handles, and it does what the ioctlsocket() call
1816 in the Winsock API does. (Win32)
1818 Available only for socket handles. (S<RISC OS>)
1822 C<kill(0, LIST)> is implemented for the sake of taint checking;
1823 use with other signals is unimplemented. (S<Mac OS>)
1825 Not implemented, hence not useful for taint checking. (S<RISC OS>)
1827 C<kill()> doesn't have the semantics of C<raise()>, i.e. it doesn't send
1828 a signal to the identified process like it does on Unix platforms.
1829 Instead C<kill($sig, $pid)> terminates the process identified by $pid,
1830 and makes it exit immediately with exit status $sig. As in Unix, if
1831 $sig is 0 and the specified process exists, it returns true without
1832 actually terminating it. (Win32)
1836 Not implemented. (S<Mac OS>, MPE/iX, VMS, S<RISC OS>)
1838 Link count not updated because hard links are not quite that hard
1839 (They are sort of half-way between hard and soft links). (AmigaOS)
1841 Hard links are implemented on Win32 (Windows NT and Windows 2000)
1846 Because Perl currently relies on the native standard C localtime()
1847 function, it is only safe to use times between 0 and (2**31)-1. Times
1848 outside this range may result in unexpected behavior depending on your
1849 operating system's implementation of localtime().
1853 Not implemented. (VMS, S<RISC OS>)
1855 Return values (especially for device and inode) may be bogus. (Win32)
1865 Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS)
1869 The C<|> variants are supported only if ToolServer is installed.
1872 open to C<|-> and C<-|> are unsupported. (S<Mac OS>, Win32, S<RISC OS>)
1874 Opening a process does not automatically flush output handles on some
1875 platforms. (SunOS, Solaris, HP-UX)
1879 Very limited functionality. (MiNT)
1883 Not implemented. (Win32, VMS, S<RISC OS>)
1887 Can't move directories between directories on different logical volumes. (Win32)
1891 Only implemented on sockets. (Win32, VMS)
1893 Only reliable on sockets. (S<RISC OS>)
1895 Note that the C<select FILEHANDLE> form is generally portable.
1903 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1907 Not implemented. (S<Mac OS>, MPE/iX, VMS, Win32, S<RISC OS>)
1911 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1915 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1919 Not implemented. (S<Mac OS>, MPE/iX, Win32, S<RISC OS>)
1923 Not implemented. (S<Plan 9>)
1933 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1937 A relatively recent addition to socket functions, may not
1938 be implemented even in UNIX platforms.
1942 Not implemented. (Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1946 Platforms that do not have rdev, blksize, or blocks will return these
1947 as '', so numeric comparison or manipulation of these fields may cause
1948 'not numeric' warnings.
1950 mtime and atime are the same thing, and ctime is creation time instead of
1951 inode change time. (S<Mac OS>).
1953 ctime not supported on UFS (S<Mac OS X>).
1955 ctime is creation time instead of inode change time (Win32).
1957 device and inode are not meaningful. (Win32)
1959 device and inode are not necessarily reliable. (VMS)
1961 mtime, atime and ctime all return the last modification time. Device and
1962 inode are not necessarily reliable. (S<RISC OS>)
1964 dev, rdev, blksize, and blocks are not available. inode is not
1965 meaningful and will differ between stat calls on the same file. (os2)
1967 some versions of cygwin when doing a stat("foo") and if not finding it
1968 may then attempt to stat("foo.exe") (Cygwin)
1972 Not implemented. (Win32, VMS, S<RISC OS>)
1976 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1980 The traditional "0", "1", and "2" MODEs are implemented with different
1981 numeric values on some systems. The flags exported by C<Fcntl>
1982 (O_RDONLY, O_WRONLY, O_RDWR) should work everywhere though. (S<Mac
1983 OS>, OS/390, VM/ESA)
1987 In general, do not assume the UNIX/POSIX semantics that you can shift
1988 C<$?> right by eight to get the exit value, or that C<$? & 127>
1989 would give you the number of the signal that terminated the program,
1990 or that C<$? & 128> would test true if the program was terminated by a
1991 coredump. Instead, use the POSIX W*() interfaces: for example, use
1992 WIFEXITED($?) and WEXITVALUE($?) to test for a normal exit and the exit
1993 value, WIFSIGNALED($?) and WTERMSIG($?) for a signal exit and the
1994 signal. Core dumping is not a portable concept, so there's no portable
1995 way to test for that.
1997 Only implemented if ToolServer is installed. (S<Mac OS>)
1999 As an optimization, may not call the command shell specified in
2000 C<$ENV{PERL5SHELL}>. C<system(1, @args)> spawns an external
2001 process and immediately returns its process designator, without
2002 waiting for it to terminate. Return value may be used subsequently
2003 in C<wait> or C<waitpid>. Failure to spawn() a subprocess is indicated
2004 by setting $? to "255 << 8". C<$?> is set in a way compatible with
2005 Unix (i.e. the exitstatus of the subprocess is obtained by "$? >> 8",
2006 as described in the documentation). (Win32)
2008 There is no shell to process metacharacters, and the native standard is
2009 to pass a command line terminated by "\n" "\r" or "\0" to the spawned
2010 program. Redirection such as C<< > foo >> is performed (if at all) by
2011 the run time library of the spawned program. C<system> I<list> will call
2012 the Unix emulation library's C<exec> emulation, which attempts to provide
2013 emulation of the stdin, stdout, stderr in force in the parent, providing
2014 the child program uses a compatible version of the emulation library.
2015 I<scalar> will call the native command line direct and no such emulation
2016 of a child Unix program will exists. Mileage B<will> vary. (S<RISC OS>)
2018 Far from being POSIX compliant. Because there may be no underlying
2019 /bin/sh tries to work around the problem by forking and execing the
2020 first token in its argument string. Handles basic redirection
2021 ("<" or ">") on its own behalf. (MiNT)
2023 Does not automatically flush output handles on some platforms.
2024 (SunOS, Solaris, HP-UX)
2026 The return value is POSIX-like (shifted up by 8 bits), which only allows
2027 room for a made-up value derived from the severity bits of the native
2028 32-bit condition code (unless overridden by C<use vmsish 'status'>).
2029 For more details see L<perlvms/$?>. (VMS)
2033 Only the first entry returned is nonzero. (S<Mac OS>)
2035 "cumulative" times will be bogus. On anything other than Windows NT
2036 or Windows 2000, "system" time will be bogus, and "user" time is
2037 actually the time returned by the clock() function in the C runtime
2040 Not useful. (S<RISC OS>)
2044 Not implemented. (Older versions of VMS)
2046 Truncation to zero-length only. (VOS)
2048 If a FILEHANDLE is supplied, it must be writable and opened in append
2049 mode (i.e., use C<<< open(FH, '>>filename') >>>
2050 or C<sysopen(FH,...,O_APPEND|O_RDWR)>. If a filename is supplied, it
2051 should not be held open elsewhere. (Win32)
2055 Returns undef where unavailable, as of version 5.005.
2057 C<umask> works but the correct permissions are set only when the file
2058 is finally closed. (AmigaOS)
2062 Only the modification time is updated. (S<BeOS>, S<Mac OS>, VMS, S<RISC OS>)
2064 May not behave as expected. Behavior depends on the C runtime
2065 library's implementation of utime(), and the filesystem being
2066 used. The FAT filesystem typically does not support an "access
2067 time" field, and it may limit timestamps to a granularity of
2068 two seconds. (Win32)
2074 Not implemented. (S<Mac OS>, VOS)
2076 Can only be applied to process handles returned for processes spawned
2077 using C<system(1, ...)> or pseudo processes created with C<fork()>. (Win32)
2079 Not useful. (S<RISC OS>)
2084 =head1 Supported Platforms
2086 As of September 2003 (the Perl release 5.8.1), the following platforms
2087 are able to build Perl from the standard source code distribution
2088 available at http://www.cpan.org/src/index.html
2099 HI-UXMPP (Hitachi) (5.8.0 worked but we didn't know it)
2110 ReliantUNIX (formerly SINIX)
2112 OpenVMS (formerly VMS)
2113 Open UNIX (Unixware) (since Perl 5.8.1/5.9.0)
2115 OS/400 (using the PASE) (since Perl 5.8.1/5.9.0)
2117 POSIX-BC (formerly BS2000)
2123 Tru64 UNIX (formerly DEC OSF/1, Digital UNIX)
2128 Win95/98/ME/2K/XP 2)
2130 z/OS (formerly OS/390)
2133 1) in DOS mode either the DOS or OS/2 ports can be used
2134 2) compilers: Borland, MinGW (GCC), VC6
2136 The following platforms worked with the previous releases (5.6 and
2137 5.7), but we did not manage either to fix or to test these in time
2138 for the 5.8.1 release. There is a very good chance that many of these
2139 will work fine with the 5.8.1.
2149 Known to be broken for 5.8.0 and 5.8.1 (but 5.6.1 and 5.7.2 can be used):
2153 The following platforms have been known to build Perl from source in
2154 the past (5.005_03 and earlier), but we haven't been able to verify
2155 their status for the current release, either because the
2156 hardware/software platforms are rare or because we don't have an
2157 active champion on these platforms--or both. They used to work,
2158 though, so go ahead and try compiling them, and let perlbug@perl.org
2192 The following platforms have their own source code distributions and
2193 binaries available via http://www.cpan.org/ports/
2197 OS/400 (ILE) 5.005_02
2198 Tandem Guardian 5.004
2200 The following platforms have only binaries available via
2201 http://www.cpan.org/ports/index.html :
2205 Acorn RISCOS 5.005_02
2209 Although we do suggest that you always build your own Perl from
2210 the source code, both for maximal configurability and for security,
2211 in case you are in a hurry you can check
2212 http://www.cpan.org/ports/index.html for binary distributions.
2216 L<perlaix>, L<perlamiga>, L<perlapollo>, L<perlbeos>, L<perlbs2000>,
2217 L<perlce>, L<perlcygwin>, L<perldgux>, L<perldos>, L<perlepoc>,
2218 L<perlebcdic>, L<perlfreebsd>, L<perlhurd>, L<perlhpux>, L<perlirix>,
2219 L<perlmachten>, L<perlmacos>, L<perlmacosx>, L<perlmint>, L<perlmpeix>,
2220 L<perlnetware>, L<perlos2>, L<perlos390>, L<perlos400>,
2221 L<perlplan9>, L<perlqnx>, L<perlsolaris>, L<perltru64>,
2222 L<perlunicode>, L<perlvmesa>, L<perlvms>, L<perlvos>,
2223 L<perlwin32>, and L<Win32>.
2225 =head1 AUTHORS / CONTRIBUTORS
2227 Abigail <abigail@foad.org>,
2228 Charles Bailey <bailey@newman.upenn.edu>,
2229 Graham Barr <gbarr@pobox.com>,
2230 Tom Christiansen <tchrist@perl.com>,
2231 Nicholas Clark <nick@ccl4.org>,
2232 Thomas Dorner <Thomas.Dorner@start.de>,
2233 Andy Dougherty <doughera@lafayette.edu>,
2234 Dominic Dunlop <domo@computer.org>,
2235 Neale Ferguson <neale@vma.tabnsw.com.au>,
2236 David J. Fiander <davidf@mks.com>,
2237 Paul Green <Paul_Green@stratus.com>,
2238 M.J.T. Guy <mjtg@cam.ac.uk>,
2239 Jarkko Hietaniemi <jhi@iki.fi>,
2240 Luther Huffman <lutherh@stratcom.com>,
2241 Nick Ing-Simmons <nick@ing-simmons.net>,
2242 Andreas J. KE<ouml>nig <a.koenig@mind.de>,
2243 Markus Laker <mlaker@contax.co.uk>,
2244 Andrew M. Langmead <aml@world.std.com>,
2245 Larry Moore <ljmoore@freespace.net>,
2246 Paul Moore <Paul.Moore@uk.origin-it.com>,
2247 Chris Nandor <pudge@pobox.com>,
2248 Matthias Neeracher <neeracher@mac.com>,
2249 Philip Newton <pne@cpan.org>,
2250 Gary Ng <71564.1743@CompuServe.COM>,
2251 Tom Phoenix <rootbeer@teleport.com>,
2252 AndrE<eacute> Pirard <A.Pirard@ulg.ac.be>,
2253 Peter Prymmer <pvhp@forte.com>,
2254 Hugo van der Sanden <hv@crypt0.demon.co.uk>,
2255 Gurusamy Sarathy <gsar@activestate.com>,
2256 Paul J. Schinder <schinder@pobox.com>,
2257 Michael G Schwern <schwern@pobox.com>,
2258 Dan Sugalski <dan@sidhe.org>,
2259 Nathan Torkington <gnat@frii.com>.