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 5.9.2, 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 ODS-2) cannot have more than one C<.> in their
366 Don't assume C<< > >> won't be the first character of a filename.
367 Always use C<< < >> explicitly to open a file for reading, or even
368 better, use the three-arg version of open, unless you want the user to
369 be able to specify a pipe open.
371 open(FILE, '<', $existing_file) or die $!;
373 If filenames might use strange characters, it is safest to open it
374 with C<sysopen> instead of C<open>. C<open> is magic and can
375 translate characters like C<< > >>, C<< < >>, and C<|>, which may
376 be the wrong thing to do. (Sometimes, though, it's the right thing.)
377 Three-arg open can also help protect against this translation in cases
378 where it is undesirable.
380 Don't use C<:> as a part of a filename since many systems use that for
381 their own semantics (Mac OS Classic for separating pathname components,
382 many networking schemes and utilities for separating the nodename and
383 the pathname, and so on). For the same reasons, avoid C<@>, C<;> and
386 Don't assume that in pathnames you can collapse two leading slashes
387 C<//> into one: some networking and clustering filesystems have special
388 semantics for that. Let the operating system to sort it out.
390 The I<portable filename characters> as defined by ANSI C are
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
393 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
397 and the "-" shouldn't be the first character. If you want to be
398 hypercorrect, stay case-insensitive and within the 8.3 naming
399 convention (all the files and directories have to be unique within one
400 directory if their names are lowercased and truncated to eight
401 characters before the C<.>, if any, and to three characters after the
402 C<.>, if any). (And do not use C<.>s in directory names.)
404 =head2 System Interaction
406 Not all platforms provide a command line. These are usually platforms
407 that rely primarily on a Graphical User Interface (GUI) for user
408 interaction. A program requiring a command line interface might
409 not work everywhere. This is probably for the user of the program
410 to deal with, so don't stay up late worrying about it.
412 Some platforms can't delete or rename files held open by the system,
413 this limitation may also apply to changing filesystem metainformation
414 like file permissions or owners. Remember to C<close> files when you
415 are done with them. Don't C<unlink> or C<rename> an open file. Don't
416 C<tie> or C<open> a file already tied or opened; C<untie> or C<close>
419 Don't open the same file more than once at a time for writing, as some
420 operating systems put mandatory locks on such files.
422 Don't assume that write/modify permission on a directory gives the
423 right to add or delete files/directories in that directory. That is
424 filesystem specific: in some filesystems you need write/modify
425 permission also (or even just) in the file/directory itself. In some
426 filesystems (AFS, DFS) the permission to add/delete directory entries
427 is a completely separate permission.
429 Don't assume that a single C<unlink> completely gets rid of the file:
430 some filesystems (most notably the ones in VMS) have versioned
431 filesystems, and unlink() removes only the most recent one (it doesn't
432 remove all the versions because by default the native tools on those
433 platforms remove just the most recent version, too). The portable
434 idiom to remove all the versions of a file is
436 1 while unlink "file";
438 This will terminate if the file is undeleteable for some reason
439 (protected, not there, and so on).
441 Don't count on a specific environment variable existing in C<%ENV>.
442 Don't count on C<%ENV> entries being case-sensitive, or even
443 case-preserving. Don't try to clear %ENV by saying C<%ENV = ();>, or,
444 if you really have to, make it conditional on C<$^O ne 'VMS'> since in
445 VMS the C<%ENV> table is much more than a per-process key-value string
448 Don't count on signals or C<%SIG> for anything.
450 Don't count on filename globbing. Use C<opendir>, C<readdir>, and
453 Don't count on per-program environment variables, or per-program current
456 Don't count on specific values of C<$!>, neither numeric nor
457 especially the strings values-- users may switch their locales causing
458 error messages to be translated into their languages. If you can
459 trust a POSIXish environment, you can portably use the symbols defined
460 by the Errno module, like ENOENT. And don't trust on the values of C<$!>
461 at all except immediately after a failed system call.
463 =head2 Command names versus file pathnames
465 Don't assume that the name used to invoke a command or program with
466 C<system> or C<exec> can also be used to test for the existence of the
467 file that holds the executable code for that command or program.
468 First, many systems have "internal" commands that are built-in to the
469 shell or OS and while these commands can be invoked, there is no
470 corresponding file. Second, some operating systems (e.g., Cygwin,
471 DJGPP, OS/2, and VOS) have required suffixes for executable files;
472 these suffixes are generally permitted on the command name but are not
473 required. Thus, a command like "perl" might exist in a file named
474 "perl", "perl.exe", or "perl.pm", depending on the operating system.
475 The variable "_exe" in the Config module holds the executable suffix,
476 if any. Third, the VMS port carefully sets up $^X and
477 $Config{perlpath} so that no further processing is required. This is
478 just as well, because the matching regular expression used below would
479 then have to deal with a possible trailing version number in the VMS
482 To convert $^X to a file pathname, taking account of the requirements
483 of the various operating system possibilities, say:
488 {$thisperl .= $Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;}
490 To convert $Config{perlpath} to a file pathname, say:
493 $thisperl = $Config{perlpath};
495 {$thisperl .= $Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;}
499 Don't assume that you can reach the public Internet.
501 Don't assume that there is only one way to get through firewalls
502 to the public Internet.
504 Don't assume that you can reach outside world through any other port
505 than 80, or some web proxy. ftp is blocked by many firewalls.
507 Don't assume that you can send email by connecting to the local SMTP port.
509 Don't assume that you can reach yourself or any node by the name
510 'localhost'. The same goes for '127.0.0.1'. You will have to try both.
512 Don't assume that the host has only one network card, or that it
513 can't bind to many virtual IP addresses.
515 Don't assume a particular network device name.
517 Don't assume a particular set of ioctl()s will work.
519 Don't assume that you can ping hosts and get replies.
521 Don't assume that any particular port (service) will respond.
523 Don't assume that Sys::Hostname (or any other API or command)
524 returns either a fully qualified hostname or a non-qualified hostname:
525 it all depends on how the system had been configured. Also remember
526 things like DHCP and NAT-- the hostname you get back might not be very
529 All the above "don't":s may look daunting, and they are -- but the key
530 is to degrade gracefully if one cannot reach the particular network
531 service one wants. Croaking or hanging do not look very professional.
533 =head2 Interprocess Communication (IPC)
535 In general, don't directly access the system in code meant to be
536 portable. That means, no C<system>, C<exec>, C<fork>, C<pipe>,
537 C<``>, C<qx//>, C<open> with a C<|>, nor any of the other things
538 that makes being a perl hacker worth being.
540 Commands that launch external processes are generally supported on
541 most platforms (though many of them do not support any type of
542 forking). The problem with using them arises from what you invoke
543 them on. External tools are often named differently on different
544 platforms, may not be available in the same location, might accept
545 different arguments, can behave differently, and often present their
546 results in a platform-dependent way. Thus, you should seldom depend
547 on them to produce consistent results. (Then again, if you're calling
548 I<netstat -a>, you probably don't expect it to run on both Unix and CP/M.)
550 One especially common bit of Perl code is opening a pipe to B<sendmail>:
552 open(MAIL, '|/usr/lib/sendmail -t')
553 or die "cannot fork sendmail: $!";
555 This is fine for systems programming when sendmail is known to be
556 available. But it is not fine for many non-Unix systems, and even
557 some Unix systems that may not have sendmail installed. If a portable
558 solution is needed, see the various distributions on CPAN that deal
559 with it. Mail::Mailer and Mail::Send in the MailTools distribution are
560 commonly used, and provide several mailing methods, including mail,
561 sendmail, and direct SMTP (via Net::SMTP) if a mail transfer agent is
562 not available. Mail::Sendmail is a standalone module that provides
563 simple, platform-independent mailing.
565 The Unix System V IPC (C<msg*(), sem*(), shm*()>) is not available
566 even on all Unix platforms.
568 Do not use either the bare result of C<pack("N", 10, 20, 30, 40)> or
569 bare v-strings (such as C<v10.20.30.40>) to represent IPv4 addresses:
570 both forms just pack the four bytes into network order. That this
571 would be equal to the C language C<in_addr> struct (which is what the
572 socket code internally uses) is not guaranteed. To be portable use
573 the routines of the Socket extension, such as C<inet_aton()>,
574 C<inet_ntoa()>, and C<sockaddr_in()>.
576 The rule of thumb for portable code is: Do it all in portable Perl, or
577 use a module (that may internally implement it with platform-specific
578 code, but expose a common interface).
580 =head2 External Subroutines (XS)
582 XS code can usually be made to work with any platform, but dependent
583 libraries, header files, etc., might not be readily available or
584 portable, or the XS code itself might be platform-specific, just as Perl
585 code might be. If the libraries and headers are portable, then it is
586 normally reasonable to make sure the XS code is portable, too.
588 A different type of portability issue arises when writing XS code:
589 availability of a C compiler on the end-user's system. C brings
590 with it its own portability issues, and writing XS code will expose
591 you to some of those. Writing purely in Perl is an easier way to
594 =head2 Standard Modules
596 In general, the standard modules work across platforms. Notable
597 exceptions are the CPAN module (which currently makes connections to external
598 programs that may not be available), platform-specific modules (like
599 ExtUtils::MM_VMS), and DBM modules.
601 There is no one DBM module available on all platforms.
602 SDBM_File and the others are generally available on all Unix and DOSish
603 ports, but not in MacPerl, where only NBDM_File and DB_File are
606 The good news is that at least some DBM module should be available, and
607 AnyDBM_File will use whichever module it can find. Of course, then
608 the code needs to be fairly strict, dropping to the greatest common
609 factor (e.g., not exceeding 1K for each record), so that it will
610 work with any DBM module. See L<AnyDBM_File> for more details.
614 The system's notion of time of day and calendar date is controlled in
615 widely different ways. Don't assume the timezone is stored in C<$ENV{TZ}>,
616 and even if it is, don't assume that you can control the timezone through
617 that variable. Don't assume anything about the three-letter timezone
618 abbreviations (for example that MST would be the Mountain Standard Time,
619 it's been known to stand for Moscow Standard Time). If you need to
620 use timezones, express them in some unambiguous format like the
621 exact number of minutes offset from UTC, or the POSIX timezone
624 Don't assume that the epoch starts at 00:00:00, January 1, 1970,
625 because that is OS- and implementation-specific. It is better to
626 store a date in an unambiguous representation. The ISO 8601 standard
627 defines YYYY-MM-DD as the date format, or YYYY-MM-DDTHH-MM-SS
628 (that's a literal "T" separating the date from the time).
629 Please do use the ISO 8601 instead of making us to guess what
630 date 02/03/04 might be. ISO 8601 even sorts nicely as-is.
631 A text representation (like "1987-12-18") can be easily converted
632 into an OS-specific value using a module like Date::Parse.
633 An array of values, such as those returned by C<localtime>, can be
634 converted to an OS-specific representation using Time::Local.
636 When calculating specific times, such as for tests in time or date modules,
637 it may be appropriate to calculate an offset for the epoch.
640 $offset = Time::Local::timegm(0, 0, 0, 1, 0, 70);
642 The value for C<$offset> in Unix will be C<0>, but in Mac OS will be
643 some large number. C<$offset> can then be added to a Unix time value
644 to get what should be the proper value on any system.
646 On Windows (at least), you shouldn't pass a negative value to C<gmtime> or
649 =head2 Character sets and character encoding
651 Assume very little about character sets.
653 Assume nothing about numerical values (C<ord>, C<chr>) of characters.
654 Do not use explicit code point ranges (like \xHH-\xHH); use for
655 example symbolic character classes like C<[:print:]>.
657 Do not assume that the alphabetic characters are encoded contiguously
658 (in the numeric sense). There may be gaps.
660 Do not assume anything about the ordering of the characters.
661 The lowercase letters may come before or after the uppercase letters;
662 the lowercase and uppercase may be interlaced so that both "a" and "A"
663 come before "b"; the accented and other international characters may
664 be interlaced so that E<auml> comes before "b".
666 =head2 Internationalisation
668 If you may assume POSIX (a rather large assumption), you may read
669 more about the POSIX locale system from L<perllocale>. The locale
670 system at least attempts to make things a little bit more portable,
671 or at least more convenient and native-friendly for non-English
672 users. The system affects character sets and encoding, and date
673 and time formatting--amongst other things.
675 If you really want to be international, you should consider Unicode.
676 See L<perluniintro> and L<perlunicode> for more information.
678 If you want to use non-ASCII bytes (outside the bytes 0x00..0x7f) in
679 the "source code" of your code, to be portable you have to be explicit
680 about what bytes they are. Someone might for example be using your
681 code under a UTF-8 locale, in which case random native bytes might be
682 illegal ("Malformed UTF-8 ...") This means that for example embedding
683 ISO 8859-1 bytes beyond 0x7f into your strings might cause trouble
684 later. If the bytes are native 8-bit bytes, you can use the C<bytes>
685 pragma. If the bytes are in a string (regular expression being a
686 curious string), you can often also use the C<\xHH> notation instead
687 of embedding the bytes as-is. If they are in some particular legacy
688 encoding (ether single-byte or something more complicated), you can
689 use the C<encoding> pragma. (If you want to write your code in UTF-8,
690 you can use either the C<utf8> pragma, or the C<encoding> pragma.)
691 The C<bytes> and C<utf8> pragmata are available since Perl 5.6.0, and
692 the C<encoding> pragma since Perl 5.8.0.
694 =head2 System Resources
696 If your code is destined for systems with severely constrained (or
697 missing!) virtual memory systems then you want to be I<especially> mindful
698 of avoiding wasteful constructs such as:
700 # NOTE: this is no longer "bad" in perl5.005
701 for (0..10000000) {} # bad
702 for (my $x = 0; $x <= 10000000; ++$x) {} # good
704 @lines = <VERY_LARGE_FILE>; # bad
706 while (<FILE>) {$file .= $_} # sometimes bad
707 $file = join('', <FILE>); # better
709 The last two constructs may appear unintuitive to most people. The
710 first repeatedly grows a string, whereas the second allocates a
711 large chunk of memory in one go. On some systems, the second is
712 more efficient that the first.
716 Most multi-user platforms provide basic levels of security, usually
717 implemented at the filesystem level. Some, however, do
718 not-- unfortunately. Thus the notion of user id, or "home" directory,
719 or even the state of being logged-in, may be unrecognizable on many
720 platforms. If you write programs that are security-conscious, it
721 is usually best to know what type of system you will be running
722 under so that you can write code explicitly for that platform (or
725 Don't assume the UNIX filesystem access semantics: the operating
726 system or the filesystem may be using some ACL systems, which are
727 richer languages than the usual rwx. Even if the rwx exist,
728 their semantics might be different.
730 (From security viewpoint testing for permissions before attempting to
731 do something is silly anyway: if one tries this, there is potential
732 for race conditions-- someone or something might change the
733 permissions between the permissions check and the actual operation.
734 Just try the operation.)
736 Don't assume the UNIX user and group semantics: especially, don't
737 expect the C<< $< >> and C<< $> >> (or the C<$(> and C<$)>) to work
738 for switching identities (or memberships).
740 Don't assume set-uid and set-gid semantics. (And even if you do,
741 think twice: set-uid and set-gid are a known can of security worms.)
745 For those times when it is necessary to have platform-specific code,
746 consider keeping the platform-specific code in one place, making porting
747 to other platforms easier. Use the Config module and the special
748 variable C<$^O> to differentiate platforms, as described in
751 Be careful in the tests you supply with your module or programs.
752 Module code may be fully portable, but its tests might not be. This
753 often happens when tests spawn off other processes or call external
754 programs to aid in the testing, or when (as noted above) the tests
755 assume certain things about the filesystem and paths. Be careful not
756 to depend on a specific output style for errors, such as when checking
757 C<$!> after a failed system call. Using C<$!> for anything else than
758 displaying it as output is doubtful (though see the Errno module for
759 testing reasonably portably for error value). Some platforms expect
760 a certain output format, and Perl on those platforms may have been
761 adjusted accordingly. Most specifically, don't anchor a regex when
762 testing an error value.
766 Modules uploaded to CPAN are tested by a variety of volunteers on
767 different platforms. These CPAN testers are notified by mail of each
768 new upload, and reply to the list with PASS, FAIL, NA (not applicable to
769 this platform), or UNKNOWN (unknown), along with any relevant notations.
771 The purpose of the testing is twofold: one, to help developers fix any
772 problems in their code that crop up because of lack of testing on other
773 platforms; two, to provide users with information about whether
774 a given module works on a given platform.
782 Mailing list: cpan-testers@perl.org
786 Testing results: http://testers.cpan.org/
792 As of version 5.002, Perl is built with a C<$^O> variable that
793 indicates the operating system it was built on. This was implemented
794 to help speed up code that would otherwise have to C<use Config>
795 and use the value of C<$Config{osname}>. Of course, to get more
796 detailed information about the system, looking into C<%Config> is
797 certainly recommended.
799 C<%Config> cannot always be trusted, however, because it was built
800 at compile time. If perl was built in one place, then transferred
801 elsewhere, some values may be wrong. The values may even have been
802 edited after the fact.
806 Perl works on a bewildering variety of Unix and Unix-like platforms (see
807 e.g. most of the files in the F<hints/> directory in the source code kit).
808 On most of these systems, the value of C<$^O> (hence C<$Config{'osname'}>,
809 too) is determined either by lowercasing and stripping punctuation from the
810 first field of the string returned by typing C<uname -a> (or a similar command)
811 at the shell prompt or by testing the file system for the presence of
812 uniquely named files such as a kernel or header file. Here, for example,
813 are a few of the more popular Unix flavors:
815 uname $^O $Config{'archname'}
816 --------------------------------------------
818 BSD/OS bsdos i386-bsdos
820 dgux dgux AViiON-dgux
821 DYNIX/ptx dynixptx i386-dynixptx
822 FreeBSD freebsd freebsd-i386
823 Linux linux arm-linux
824 Linux linux i386-linux
825 Linux linux i586-linux
826 Linux linux ppc-linux
827 HP-UX hpux PA-RISC1.1
829 Mac OS X darwin darwin
830 MachTen PPC machten powerpc-machten
832 NeXT 4 next OPENSTEP-Mach
833 openbsd openbsd i386-openbsd
834 OSF1 dec_osf alpha-dec_osf
835 reliantunix-n svr4 RM400-svr4
836 SCO_SV sco_sv i386-sco_sv
837 SINIX-N svr4 RM400-svr4
838 sn4609 unicos CRAY_C90-unicos
839 sn6521 unicosmk t3e-unicosmk
840 sn9617 unicos CRAY_J90-unicos
841 SunOS solaris sun4-solaris
842 SunOS solaris i86pc-solaris
843 SunOS4 sunos sun4-sunos
845 Because the value of C<$Config{archname}> may depend on the
846 hardware architecture, it can vary more than the value of C<$^O>.
848 =head2 DOS and Derivatives
850 Perl has long been ported to Intel-style microcomputers running under
851 systems like PC-DOS, MS-DOS, OS/2, and most Windows platforms you can
852 bring yourself to mention (except for Windows CE, if you count that).
853 Users familiar with I<COMMAND.COM> or I<CMD.EXE> style shells should
854 be aware that each of these file specifications may have subtle
857 $filespec0 = "c:/foo/bar/file.txt";
858 $filespec1 = "c:\\foo\\bar\\file.txt";
859 $filespec2 = 'c:\foo\bar\file.txt';
860 $filespec3 = 'c:\\foo\\bar\\file.txt';
862 System calls accept either C</> or C<\> as the path separator.
863 However, many command-line utilities of DOS vintage treat C</> as
864 the option prefix, so may get confused by filenames containing C</>.
865 Aside from calling any external programs, C</> will work just fine,
866 and probably better, as it is more consistent with popular usage,
867 and avoids the problem of remembering what to backwhack and what
870 The DOS FAT filesystem can accommodate only "8.3" style filenames. Under
871 the "case-insensitive, but case-preserving" HPFS (OS/2) and NTFS (NT)
872 filesystems you may have to be careful about case returned with functions
873 like C<readdir> or used with functions like C<open> or C<opendir>.
875 DOS also treats several filenames as special, such as AUX, PRN,
876 NUL, CON, COM1, LPT1, LPT2, etc. Unfortunately, sometimes these
877 filenames won't even work if you include an explicit directory
878 prefix. It is best to avoid such filenames, if you want your code
879 to be portable to DOS and its derivatives. It's hard to know what
880 these all are, unfortunately.
882 Users of these operating systems may also wish to make use of
883 scripts such as I<pl2bat.bat> or I<pl2cmd> to
884 put wrappers around your scripts.
886 Newline (C<\n>) is translated as C<\015\012> by STDIO when reading from
887 and writing to files (see L<"Newlines">). C<binmode(FILEHANDLE)>
888 will keep C<\n> translated as C<\012> for that filehandle. Since it is a
889 no-op on other systems, C<binmode> should be used for cross-platform code
890 that deals with binary data. That's assuming you realize in advance
891 that your data is in binary. General-purpose programs should
892 often assume nothing about their data.
894 The C<$^O> variable and the C<$Config{archname}> values for various
895 DOSish perls are as follows:
897 OS $^O $Config{archname} ID Version
898 --------------------------------------------------------
902 Windows 3.1 ? ? 0 3 01
903 Windows 95 MSWin32 MSWin32-x86 1 4 00
904 Windows 98 MSWin32 MSWin32-x86 1 4 10
905 Windows ME MSWin32 MSWin32-x86 1 ?
906 Windows NT MSWin32 MSWin32-x86 2 4 xx
907 Windows NT MSWin32 MSWin32-ALPHA 2 4 xx
908 Windows NT MSWin32 MSWin32-ppc 2 4 xx
909 Windows 2000 MSWin32 MSWin32-x86 2 5 00
910 Windows XP MSWin32 MSWin32-x86 2 5 01
911 Windows 2003 MSWin32 MSWin32-x86 2 5 02
912 Windows CE MSWin32 ? 3
915 The various MSWin32 Perl's can distinguish the OS they are running on
916 via the value of the fifth element of the list returned from
917 Win32::GetOSVersion(). For example:
919 if ($^O eq 'MSWin32') {
920 my @os_version_info = Win32::GetOSVersion();
921 print +('3.1','95','NT')[$os_version_info[4]],"\n";
924 There are also Win32::IsWinNT() and Win32::IsWin95(), try C<perldoc Win32>,
925 and as of libwin32 0.19 (not part of the core Perl distribution)
926 Win32::GetOSName(). The very portable POSIX::uname() will work too:
928 c:\> perl -MPOSIX -we "print join '|', uname"
929 Windows NT|moonru|5.0|Build 2195 (Service Pack 2)|x86
937 The djgpp environment for DOS, http://www.delorie.com/djgpp/
942 The EMX environment for DOS, OS/2, etc. emx@iaehv.nl,
943 http://www.leo.org/pub/comp/os/os2/leo/gnu/emx+gcc/index.html or
944 ftp://hobbes.nmsu.edu/pub/os2/dev/emx/ Also L<perlos2>.
948 Build instructions for Win32 in L<perlwin32>, or under the Cygnus environment
953 The C<Win32::*> modules in L<Win32>.
957 The ActiveState Pages, http://www.activestate.com/
961 The Cygwin environment for Win32; F<README.cygwin> (installed
962 as L<perlcygwin>), http://www.cygwin.com/
966 The U/WIN environment for Win32,
967 http://www.research.att.com/sw/tools/uwin/
971 Build instructions for OS/2, L<perlos2>
977 Any module requiring XS compilation is right out for most people, because
978 MacPerl is built using non-free (and non-cheap!) compilers. Some XS
979 modules that can work with MacPerl are built and distributed in binary
982 Directories are specified as:
984 volume:folder:file for absolute pathnames
985 volume:folder: for absolute pathnames
986 :folder:file for relative pathnames
987 :folder: for relative pathnames
988 :file for relative pathnames
989 file for relative pathnames
991 Files are stored in the directory in alphabetical order. Filenames are
992 limited to 31 characters, and may include any character except for
993 null and C<:>, which is reserved as the path separator.
995 Instead of C<flock>, see C<FSpSetFLock> and C<FSpRstFLock> in the
996 Mac::Files module, or C<chmod(0444, ...)> and C<chmod(0666, ...)>.
998 In the MacPerl application, you can't run a program from the command line;
999 programs that expect C<@ARGV> to be populated can be edited with something
1000 like the following, which brings up a dialog box asking for the command
1004 @ARGV = split /\s+/, MacPerl::Ask('Arguments?');
1007 A MacPerl script saved as a "droplet" will populate C<@ARGV> with the full
1008 pathnames of the files dropped onto the script.
1010 Mac users can run programs under a type of command line interface
1011 under MPW (Macintosh Programmer's Workshop, a free development
1012 environment from Apple). MacPerl was first introduced as an MPW
1013 tool, and MPW can be used like a shell:
1015 perl myscript.plx some arguments
1017 ToolServer is another app from Apple that provides access to MPW tools
1018 from MPW and the MacPerl app, which allows MacPerl programs to use
1019 C<system>, backticks, and piped C<open>.
1021 "S<Mac OS>" is the proper name for the operating system, but the value
1022 in C<$^O> is "MacOS". To determine architecture, version, or whether
1023 the application or MPW tool version is running, check:
1025 $is_app = $MacPerl::Version =~ /App/;
1026 $is_tool = $MacPerl::Version =~ /MPW/;
1027 ($version) = $MacPerl::Version =~ /^(\S+)/;
1028 $is_ppc = $MacPerl::Architecture eq 'MacPPC';
1029 $is_68k = $MacPerl::Architecture eq 'Mac68K';
1031 S<Mac OS X>, based on NeXT's OpenStep OS, runs MacPerl natively, under the
1032 "Classic" environment. There is no "Carbon" version of MacPerl to run
1033 under the primary Mac OS X environment. S<Mac OS X> and its Open Source
1034 version, Darwin, both run Unix perl natively.
1042 MacPerl Development, http://dev.macperl.org/ .
1046 The MacPerl Pages, http://www.macperl.com/ .
1050 The MacPerl mailing lists, http://lists.perl.org/ .
1054 MPW, ftp://ftp.apple.com/developer/Tool_Chest/Core_Mac_OS_Tools/
1060 Perl on VMS is discussed in L<perlvms> in the perl distribution.
1062 The official name of VMS as of this writing is OpenVMS.
1064 Perl on VMS can accept either VMS- or Unix-style file
1065 specifications as in either of the following:
1067 $ perl -ne "print if /perl_setup/i" SYS$LOGIN:LOGIN.COM
1068 $ perl -ne "print if /perl_setup/i" /sys$login/login.com
1070 but not a mixture of both as in:
1072 $ perl -ne "print if /perl_setup/i" sys$login:/login.com
1073 Can't open sys$login:/login.com: file specification syntax error
1075 Interacting with Perl from the Digital Command Language (DCL) shell
1076 often requires a different set of quotation marks than Unix shells do.
1079 $ perl -e "print ""Hello, world.\n"""
1082 There are several ways to wrap your perl scripts in DCL F<.COM> files, if
1083 you are so inclined. For example:
1085 $ write sys$output "Hello from DCL!"
1087 $ then perl -x 'f$environment("PROCEDURE")
1088 $ else perl -x - 'p1 'p2 'p3 'p4 'p5 'p6 'p7 'p8
1089 $ deck/dollars="__END__"
1092 print "Hello from Perl!\n";
1097 Do take care with C<$ ASSIGN/nolog/user SYS$COMMAND: SYS$INPUT> if your
1098 perl-in-DCL script expects to do things like C<< $read = <STDIN>; >>.
1100 The VMS operating system has two filesystems, known as ODS-2 and ODS-5.
1102 For ODS-2, filenames are in the format "name.extension;version". The
1103 maximum length for filenames is 39 characters, and the maximum length for
1104 extensions is also 39 characters. Version is a number from 1 to
1105 32767. Valid characters are C</[A-Z0-9$_-]/>.
1107 The ODS-2 filesystem is case-insensitive and does not preserve case.
1108 Perl simulates this by converting all filenames to lowercase internally.
1110 For ODS-5, filenames may have almost any character in them and can include
1111 Unicode characters. Characters that could be misinterpreted by the DCL
1112 shell or file parsing utilities need to be prefixed with the C<^>
1113 character, or replaced with hexadecimal characters prefixed with the
1114 C<^> character. Such prefixing is only needed with the pathnames are
1115 in VMS format in applications. Programs that can accept the UNIX format
1116 of pathnames do not need the escape characters. The maximum length for
1117 filenames is 255 characters. The ODS-5 file system can handle both
1118 a case preserved and a case sensitive mode.
1120 ODS-5 is only available on the OpenVMS for 64 bit platforms.
1122 Support for the extended file specifications is being done as optional
1123 settings to preserve backward compatibility with Perl scripts that
1124 assume the previous VMS limitations.
1126 In general routines on VMS that get a UNIX format file specification
1127 should return it in a UNIX format, and when they get a VMS format
1128 specification they should return a VMS format unless they are documented
1131 For routines that generate return a file specification, VMS allows setting
1132 if the C library which Perl is built on if it will be returned in VMS
1133 format or in UNIX format.
1135 With the ODS-2 file system, there is not much difference in syntax of
1136 filenames without paths for VMS or UNIX. With the extended character
1137 set available with ODS-5 there can be a significant difference.
1139 Because of this, existing Perl scripts written for VMS were sometimes
1140 treating VMS and UNIX filenames interchangeably. Without the extended
1141 character set enabled, this behavior will mostly be maintained for
1142 backwards compatibility.
1144 When extended characters are enabled with ODS-5, the handling of
1145 UNIX formatted file specifications is to that of a UNIX system.
1147 VMS file specifications without extensions have a trailing dot. An
1148 equivalent UNIX file specification should not show the trailing dot.
1150 The result of all of this, is that for VMS, for portable scripts, you
1151 can not depend on Perl to present the filenames in lowercase, to be
1152 case sensitive, and that the filenames could be returned in either
1155 And if a routine returns a file specification, unless it is intended to
1156 convert it, it should return it in the same format as it found it.
1158 C<readdir> by default has traditionally returned lowercased filenames.
1159 When the ODS-5 support is enabled, it will return the exact case of the
1160 filename on the disk.
1162 Files without extensions have a trailing period on them, so doing a
1163 C<readdir> in the default mode with a file named F<A.;5> will
1164 return F<a.> when VMS is (though that file could be opened with
1167 With support for extended file specifications and if C<opendir> was
1168 given a UNIX format directory, a file named F<A.;5> will return F<a>
1169 and optionally in the exact case on the disk. When C<opendir> is given
1170 a VMS format directory, then C<readdir> should return F<a.>, and
1171 again with the optionally the exact case.
1173 RMS had an eight level limit on directory depths from any rooted logical
1174 (allowing 16 levels overall) prior to VMS 7.2. Hence
1175 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8]> is a valid directory specification but
1176 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8.9]> is not. F<Makefile.PL> authors might
1177 have to take this into account, but at least they can refer to the former
1178 as C</PERL_ROOT/lib/2/3/4/5/6/7/8/>.
1180 The VMS::Filespec module, which gets installed as part of the build
1181 process on VMS, is a pure Perl module that can easily be installed on
1182 non-VMS platforms and can be helpful for conversions to and from RMS
1183 native formats. It is also now the only way that you should check to
1184 see if VMS is in a case sensitive mode.
1186 What C<\n> represents depends on the type of file opened. It usually
1187 represents C<\012> but it could also be C<\015>, C<\012>, C<\015\012>,
1188 C<\000>, C<\040>, or nothing depending on the file organization and
1189 record format. The VMS::Stdio module provides access to the
1190 special fopen() requirements of files with unusual attributes on VMS.
1192 TCP/IP stacks are optional on VMS, so socket routines might not be
1193 implemented. UDP sockets may not be supported.
1195 The TCP/IP library support for all current versions of VMS is dynamically
1196 loaded if present, so even if the routines are configured, they may
1197 return a status indicating that they are not implemented.
1199 The value of C<$^O> on OpenVMS is "VMS". To determine the architecture
1200 that you are running on without resorting to loading all of C<%Config>
1201 you can examine the content of the C<@INC> array like so:
1203 if (grep(/VMS_AXP/, @INC)) {
1204 print "I'm on Alpha!\n";
1206 } elsif (grep(/VMS_VAX/, @INC)) {
1207 print "I'm on VAX!\n";
1209 } elsif (grep(/VMS_IA64/, @INC)) {
1210 print "I'm on IA64!\n";
1213 print "I'm not so sure about where $^O is...\n";
1216 In general, the significant differences should only be if Perl is running
1217 on VMS_VAX or one of the 64 bit OpenVMS platforms.
1219 On VMS, perl determines the UTC offset from the C<SYS$TIMEZONE_DIFFERENTIAL>
1220 logical name. Although the VMS epoch began at 17-NOV-1858 00:00:00.00,
1221 calls to C<localtime> are adjusted to count offsets from
1222 01-JAN-1970 00:00:00.00, just like Unix.
1230 F<README.vms> (installed as L<README_vms>), L<perlvms>
1234 vmsperl list, majordomo@perl.org
1236 (Put the words C<subscribe vmsperl> in message body.)
1240 vmsperl on the web, http://www.sidhe.org/vmsperl/index.html
1246 Perl on VOS is discussed in F<README.vos> in the perl distribution
1247 (installed as L<perlvos>). Perl on VOS can accept either VOS- or
1248 Unix-style file specifications as in either of the following:
1250 C<< $ perl -ne "print if /perl_setup/i" >system>notices >>
1251 C<< $ perl -ne "print if /perl_setup/i" /system/notices >>
1253 or even a mixture of both as in:
1255 C<< $ perl -ne "print if /perl_setup/i" >system/notices >>
1257 Even though VOS allows the slash character to appear in object
1258 names, because the VOS port of Perl interprets it as a pathname
1259 delimiting character, VOS files, directories, or links whose names
1260 contain a slash character cannot be processed. Such files must be
1261 renamed before they can be processed by Perl. Note that VOS limits
1262 file names to 32 or fewer characters.
1264 The value of C<$^O> on VOS is "VOS". To determine the architecture that
1265 you are running on without resorting to loading all of C<%Config> you
1266 can examine the content of the @INC array like so:
1269 print "I'm on a Stratus box!\n";
1271 print "I'm not on a Stratus box!\n";
1281 F<README.vos> (installed as L<perlvos>)
1285 The VOS mailing list.
1287 There is no specific mailing list for Perl on VOS. You can post
1288 comments to the comp.sys.stratus newsgroup, or subscribe to the general
1289 Stratus mailing list. Send a letter with "subscribe Info-Stratus" in
1290 the message body to majordomo@list.stratagy.com.
1294 VOS Perl on the web at http://ftp.stratus.com/pub/vos/posix/posix.html
1298 =head2 EBCDIC Platforms
1300 Recent versions of Perl have been ported to platforms such as OS/400 on
1301 AS/400 minicomputers as well as OS/390, VM/ESA, and BS2000 for S/390
1302 Mainframes. Such computers use EBCDIC character sets internally (usually
1303 Character Code Set ID 0037 for OS/400 and either 1047 or POSIX-BC for S/390
1304 systems). On the mainframe perl currently works under the "Unix system
1305 services for OS/390" (formerly known as OpenEdition), VM/ESA OpenEdition, or
1306 the BS200 POSIX-BC system (BS2000 is supported in perl 5.6 and greater).
1307 See L<perlos390> for details. Note that for OS/400 there is also a port of
1308 Perl 5.8.1/5.9.0 or later to the PASE which is ASCII-based (as opposed to
1309 ILE which is EBCDIC-based), see L<perlos400>.
1311 As of R2.5 of USS for OS/390 and Version 2.3 of VM/ESA these Unix
1312 sub-systems do not support the C<#!> shebang trick for script invocation.
1313 Hence, on OS/390 and VM/ESA perl scripts can be executed with a header
1314 similar to the following simple script:
1317 eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
1319 #!/usr/local/bin/perl # just a comment really
1321 print "Hello from perl!\n";
1323 OS/390 will support the C<#!> shebang trick in release 2.8 and beyond.
1324 Calls to C<system> and backticks can use POSIX shell syntax on all
1327 On the AS/400, if PERL5 is in your library list, you may need
1328 to wrap your perl scripts in a CL procedure to invoke them like so:
1331 CALL PGM(PERL5/PERL) PARM('/QOpenSys/hello.pl')
1334 This will invoke the perl script F<hello.pl> in the root of the
1335 QOpenSys file system. On the AS/400 calls to C<system> or backticks
1338 On these platforms, bear in mind that the EBCDIC character set may have
1339 an effect on what happens with some perl functions (such as C<chr>,
1340 C<pack>, C<print>, C<printf>, C<ord>, C<sort>, C<sprintf>, C<unpack>), as
1341 well as bit-fiddling with ASCII constants using operators like C<^>, C<&>
1342 and C<|>, not to mention dealing with socket interfaces to ASCII computers
1343 (see L<"Newlines">).
1345 Fortunately, most web servers for the mainframe will correctly
1346 translate the C<\n> in the following statement to its ASCII equivalent
1347 (C<\r> is the same under both Unix and OS/390 & VM/ESA):
1349 print "Content-type: text/html\r\n\r\n";
1351 The values of C<$^O> on some of these platforms includes:
1353 uname $^O $Config{'archname'}
1354 --------------------------------------------
1357 POSIX-BC posix-bc BS2000-posix-bc
1360 Some simple tricks for determining if you are running on an EBCDIC
1361 platform could include any of the following (perhaps all):
1363 if ("\t" eq "\05") { print "EBCDIC may be spoken here!\n"; }
1365 if (ord('A') == 193) { print "EBCDIC may be spoken here!\n"; }
1367 if (chr(169) eq 'z') { print "EBCDIC may be spoken here!\n"; }
1369 One thing you may not want to rely on is the EBCDIC encoding
1370 of punctuation characters since these may differ from code page to code
1371 page (and once your module or script is rumoured to work with EBCDIC,
1372 folks will want it to work with all EBCDIC character sets).
1380 L<perlos390>, F<README.os390>, F<perlbs2000>, F<README.vmesa>,
1385 The perl-mvs@perl.org list is for discussion of porting issues as well as
1386 general usage issues for all EBCDIC Perls. Send a message body of
1387 "subscribe perl-mvs" to majordomo@perl.org.
1391 AS/400 Perl information at
1392 http://as400.rochester.ibm.com/
1393 as well as on CPAN in the F<ports/> directory.
1397 =head2 Acorn RISC OS
1399 Because Acorns use ASCII with newlines (C<\n>) in text files as C<\012> like
1400 Unix, and because Unix filename emulation is turned on by default,
1401 most simple scripts will probably work "out of the box". The native
1402 filesystem is modular, and individual filesystems are free to be
1403 case-sensitive or insensitive, and are usually case-preserving. Some
1404 native filesystems have name length limits, which file and directory
1405 names are silently truncated to fit. Scripts should be aware that the
1406 standard filesystem currently has a name length limit of B<10>
1407 characters, with up to 77 items in a directory, but other filesystems
1408 may not impose such limitations.
1410 Native filenames are of the form
1412 Filesystem#Special_Field::DiskName.$.Directory.Directory.File
1416 Special_Field is not usually present, but may contain . and $ .
1417 Filesystem =~ m|[A-Za-z0-9_]|
1418 DsicName =~ m|[A-Za-z0-9_/]|
1419 $ represents the root directory
1420 . is the path separator
1421 @ is the current directory (per filesystem but machine global)
1422 ^ is the parent directory
1423 Directory and File =~ m|[^\0- "\.\$\%\&:\@\\^\|\177]+|
1425 The default filename translation is roughly C<tr|/.|./|;>
1427 Note that C<"ADFS::HardDisk.$.File" ne 'ADFS::HardDisk.$.File'> and that
1428 the second stage of C<$> interpolation in regular expressions will fall
1429 foul of the C<$.> if scripts are not careful.
1431 Logical paths specified by system variables containing comma-separated
1432 search lists are also allowed; hence C<System:Modules> is a valid
1433 filename, and the filesystem will prefix C<Modules> with each section of
1434 C<System$Path> until a name is made that points to an object on disk.
1435 Writing to a new file C<System:Modules> would be allowed only if
1436 C<System$Path> contains a single item list. The filesystem will also
1437 expand system variables in filenames if enclosed in angle brackets, so
1438 C<< <System$Dir>.Modules >> would look for the file
1439 S<C<$ENV{'System$Dir'} . 'Modules'>>. The obvious implication of this is
1440 that B<fully qualified filenames can start with C<< <> >>> and should
1441 be protected when C<open> is used for input.
1443 Because C<.> was in use as a directory separator and filenames could not
1444 be assumed to be unique after 10 characters, Acorn implemented the C
1445 compiler to strip the trailing C<.c> C<.h> C<.s> and C<.o> suffix from
1446 filenames specified in source code and store the respective files in
1447 subdirectories named after the suffix. Hence files are translated:
1450 C:foo.h C:h.foo (logical path variable)
1451 sys/os.h sys.h.os (C compiler groks Unix-speak)
1452 10charname.c c.10charname
1453 10charname.o o.10charname
1454 11charname_.c c.11charname (assuming filesystem truncates at 10)
1456 The Unix emulation library's translation of filenames to native assumes
1457 that this sort of translation is required, and it allows a user-defined list
1458 of known suffixes that it will transpose in this fashion. This may
1459 seem transparent, but consider that with these rules C<foo/bar/baz.h>
1460 and C<foo/bar/h/baz> both map to C<foo.bar.h.baz>, and that C<readdir> and
1461 C<glob> cannot and do not attempt to emulate the reverse mapping. Other
1462 C<.>'s in filenames are translated to C</>.
1464 As implied above, the environment accessed through C<%ENV> is global, and
1465 the convention is that program specific environment variables are of the
1466 form C<Program$Name>. Each filesystem maintains a current directory,
1467 and the current filesystem's current directory is the B<global> current
1468 directory. Consequently, sociable programs don't change the current
1469 directory but rely on full pathnames, and programs (and Makefiles) cannot
1470 assume that they can spawn a child process which can change the current
1471 directory without affecting its parent (and everyone else for that
1474 Because native operating system filehandles are global and are currently
1475 allocated down from 255, with 0 being a reserved value, the Unix emulation
1476 library emulates Unix filehandles. Consequently, you can't rely on
1477 passing C<STDIN>, C<STDOUT>, or C<STDERR> to your children.
1479 The desire of users to express filenames of the form
1480 C<< <Foo$Dir>.Bar >> on the command line unquoted causes problems,
1481 too: C<``> command output capture has to perform a guessing game. It
1482 assumes that a string C<< <[^<>]+\$[^<>]> >> is a
1483 reference to an environment variable, whereas anything else involving
1484 C<< < >> or C<< > >> is redirection, and generally manages to be 99%
1485 right. Of course, the problem remains that scripts cannot rely on any
1486 Unix tools being available, or that any tools found have Unix-like command
1489 Extensions and XS are, in theory, buildable by anyone using free
1490 tools. In practice, many don't, as users of the Acorn platform are
1491 used to binary distributions. MakeMaker does run, but no available
1492 make currently copes with MakeMaker's makefiles; even if and when
1493 this should be fixed, the lack of a Unix-like shell will cause
1494 problems with makefile rules, especially lines of the form C<cd
1495 sdbm && make all>, and anything using quoting.
1497 "S<RISC OS>" is the proper name for the operating system, but the value
1498 in C<$^O> is "riscos" (because we don't like shouting).
1502 Perl has been ported to many platforms that do not fit into any of
1503 the categories listed above. Some, such as AmigaOS, Atari MiNT,
1504 BeOS, HP MPE/iX, QNX, Plan 9, and VOS, have been well-integrated
1505 into the standard Perl source code kit. You may need to see the
1506 F<ports/> directory on CPAN for information, and possibly binaries,
1507 for the likes of: aos, Atari ST, lynxos, riscos, Novell Netware,
1508 Tandem Guardian, I<etc.> (Yes, we know that some of these OSes may
1509 fall under the Unix category, but we are not a standards body.)
1511 Some approximate operating system names and their C<$^O> values
1512 in the "OTHER" category include:
1514 OS $^O $Config{'archname'}
1515 ------------------------------------------
1516 Amiga DOS amigaos m68k-amigos
1518 MPE/iX mpeix PA-RISC1.1
1526 Amiga, F<README.amiga> (installed as L<perlamiga>).
1530 Atari, F<README.mint> and Guido Flohr's web page
1531 http://stud.uni-sb.de/~gufl0000/
1535 Be OS, F<README.beos>
1539 HP 300 MPE/iX, F<README.mpeix> and Mark Bixby's web page
1540 http://www.bixby.org/mark/perlix.html
1544 A free perl5-based PERL.NLM for Novell Netware is available in
1545 precompiled binary and source code form from http://www.novell.com/
1546 as well as from CPAN.
1550 S<Plan 9>, F<README.plan9>
1554 =head1 FUNCTION IMPLEMENTATIONS
1556 Listed below are functions that are either completely unimplemented
1557 or else have been implemented differently on various platforms.
1558 Following each description will be, in parentheses, a list of
1559 platforms that the description applies to.
1561 The list may well be incomplete, or even wrong in some places. When
1562 in doubt, consult the platform-specific README files in the Perl
1563 source distribution, and any other documentation resources accompanying
1566 Be aware, moreover, that even among Unix-ish systems there are variations.
1568 For many functions, you can also query C<%Config>, exported by
1569 default from the Config module. For example, to check whether the
1570 platform has the C<lstat> call, check C<$Config{d_lstat}>. See
1571 L<Config> for a full description of available variables.
1573 =head2 Alphabetical Listing of Perl Functions
1579 C<-r>, C<-w>, and C<-x> have a limited meaning only; directories
1580 and applications are executable, and there are no uid/gid
1581 considerations. C<-o> is not supported. (S<Mac OS>)
1583 C<-r>, C<-w>, C<-x>, and C<-o> tell whether the file is accessible,
1584 which may not reflect UIC-based file protections. (VMS)
1586 C<-s> returns the size of the data fork, not the total size of data fork
1587 plus resource fork. (S<Mac OS>).
1589 C<-s> by name on an open file will return the space reserved on disk,
1590 rather than the current extent. C<-s> on an open filehandle returns the
1591 current size. (S<RISC OS>)
1593 C<-R>, C<-W>, C<-X>, C<-O> are indistinguishable from C<-r>, C<-w>,
1594 C<-x>, C<-o>. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1596 C<-b>, C<-c>, C<-k>, C<-g>, C<-p>, C<-u>, C<-A> are not implemented.
1599 C<-g>, C<-k>, C<-l>, C<-p>, C<-u>, C<-A> are not particularly meaningful.
1600 (Win32, VMS, S<RISC OS>)
1602 C<-d> is true if passed a device spec without an explicit directory.
1605 C<-T> and C<-B> are implemented, but might misclassify Mac text files
1606 with foreign characters; this is the case will all platforms, but may
1607 affect S<Mac OS> often. (S<Mac OS>)
1609 C<-x> (or C<-X>) determine if a file ends in one of the executable
1610 suffixes. C<-S> is meaningless. (Win32)
1612 C<-x> (or C<-X>) determine if a file has an executable file type.
1617 Due to issues with various CPUs, math libraries, compilers, and standards,
1618 results for C<atan2()> may vary depending on any combination of the above.
1619 Perl attempts to conform to the Open Group/IEEE standards for the results
1620 returned from C<atan2()>, but cannot force the issue if the system Perl is
1621 run on does not allow it. (Tru64, HP-UX 10.20)
1623 The current version of the standards for C<atan2()> is available at
1624 L<http://www.opengroup.org/onlinepubs/009695399/functions/atan2.html>.
1628 Meaningless. (S<Mac OS>, S<RISC OS>)
1630 Reopens file and restores pointer; if function fails, underlying
1631 filehandle may be closed, or pointer may be in a different position.
1634 The value returned by C<tell> may be affected after the call, and
1635 the filehandle may be flushed. (Win32)
1639 Only limited meaning. Disabling/enabling write permission is mapped to
1640 locking/unlocking the file. (S<Mac OS>)
1642 Only good for changing "owner" read-write access, "group", and "other"
1643 bits are meaningless. (Win32)
1645 Only good for changing "owner" and "other" read-write access. (S<RISC OS>)
1647 Access permissions are mapped onto VOS access-control list changes. (VOS)
1649 The actual permissions set depend on the value of the C<CYGWIN>
1650 in the SYSTEM environment settings. (Cygwin)
1654 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1656 Does nothing, but won't fail. (Win32)
1658 A little funky, because VOS's notion of ownership is a little funky (VOS).
1662 Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS, VM/ESA)
1666 May not be available if library or source was not provided when building
1671 Not implemented. (VMS, S<Plan 9>, VOS)
1675 Not implemented. (VMS, S<Plan 9>, VOS)
1679 Not useful. (S<Mac OS>, S<RISC OS>)
1681 Not implemented. (Win32)
1683 Invokes VMS debugger. (VMS)
1687 Not implemented. (S<Mac OS>)
1689 Implemented via Spawn. (VM/ESA)
1691 Does not automatically flush output handles on some platforms.
1692 (SunOS, Solaris, HP-UX)
1696 Emulates UNIX exit() (which considers C<exit 1> to indicate an error) by
1697 mapping the C<1> to SS$_ABORT (C<44>). This behavior may be overridden
1698 with the pragma C<use vmsish 'exit'>. As with the CRTL's exit()
1699 function, C<exit 0> is also mapped to an exit status of SS$_NORMAL
1700 (C<1>); this mapping cannot be overridden. Any other argument to exit()
1701 is used directly as Perl's exit status. On VMS, unless the future
1702 POSIX_EXIT mode is enabled, the exit code should always be a valid
1703 VMS exit code and not a generic number. When the POSIX_EXIT mode is
1704 enabled, a generic number will be encoded in a method compatible with
1705 the C library _POSIX_EXIT macro so that it can be decoded by other
1706 programs, particularly ones written in C, like the GNV package. (VMS)
1710 Not implemented. (Win32)
1711 Some functions available based on the version of VMS. (VMS)
1715 Not implemented (S<Mac OS>, VMS, S<RISC OS>, VOS).
1717 Available only on Windows NT (not on Windows 95). (Win32)
1721 Not implemented. (S<Mac OS>, AmigaOS, S<RISC OS>, VM/ESA, VMS)
1723 Emulated using multiple interpreters. See L<perlfork>. (Win32)
1725 Does not automatically flush output handles on some platforms.
1726 (SunOS, Solaris, HP-UX)
1730 Not implemented. (S<Mac OS>, S<RISC OS>)
1734 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1738 Not implemented. (S<Mac OS>, Win32, S<RISC OS>)
1742 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1746 Not implemented. (S<Mac OS>, Win32)
1748 Not useful. (S<RISC OS>)
1752 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1756 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1760 Not implemented. (S<Mac OS>, Win32)
1762 Not useful. (S<RISC OS>)
1766 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1770 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1772 =item getprotobynumber
1774 Not implemented. (S<Mac OS>)
1778 Not implemented. (S<Mac OS>)
1782 Not implemented. (S<Mac OS>, Win32, VM/ESA)
1786 Not implemented. (S<Mac OS>, Win32, VMS, VM/ESA)
1790 C<gethostbyname('localhost')> does not work everywhere: you may have
1791 to use C<gethostbyname('127.0.0.1')>. (S<Mac OS>, S<Irix 5>)
1795 Not implemented. (S<Mac OS>, Win32)
1799 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1803 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1807 Not implemented. (Win32, S<Plan 9>)
1811 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1815 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1819 Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1823 Not implemented. (S<Plan 9>, Win32, S<RISC OS>)
1827 Not implemented. (S<Mac OS>, MPE/iX, VM/ESA, Win32)
1831 Not implemented. (S<Mac OS>, MPE/iX, S<RISC OS>, VM/ESA, VMS, Win32)
1835 Not implemented. (S<Mac OS>, Win32)
1839 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1843 Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1847 Not implemented. (S<Plan 9>, Win32)
1849 =item getsockopt SOCKET,LEVEL,OPTNAME
1851 Not implemented. (S<Plan 9>)
1855 This operator is implemented via the File::Glob extension on most
1856 platforms. See L<File::Glob> for portability information.
1860 Same portability caveats as L<localtime>.
1862 =item ioctl FILEHANDLE,FUNCTION,SCALAR
1864 Not implemented. (VMS)
1866 Available only for socket handles, and it does what the ioctlsocket() call
1867 in the Winsock API does. (Win32)
1869 Available only for socket handles. (S<RISC OS>)
1873 C<kill(0, LIST)> is implemented for the sake of taint checking;
1874 use with other signals is unimplemented. (S<Mac OS>)
1876 Not implemented, hence not useful for taint checking. (S<RISC OS>)
1878 C<kill()> doesn't have the semantics of C<raise()>, i.e. it doesn't send
1879 a signal to the identified process like it does on Unix platforms.
1880 Instead C<kill($sig, $pid)> terminates the process identified by $pid,
1881 and makes it exit immediately with exit status $sig. As in Unix, if
1882 $sig is 0 and the specified process exists, it returns true without
1883 actually terminating it. (Win32)
1885 Is not supported for process identification number of 0 or negative
1890 Not implemented. (S<Mac OS>, MPE/iX, S<RISC OS>)
1892 Link count not updated because hard links are not quite that hard
1893 (They are sort of half-way between hard and soft links). (AmigaOS)
1895 Hard links are implemented on Win32 (Windows NT and Windows 2000)
1898 Available on 64 bit OpenVMS 8.2 and later. (VMS)
1902 Because Perl currently relies on the native standard C localtime()
1903 function, it is only safe to use times between 0 and (2**31)-1. Times
1904 outside this range may result in unexpected behavior depending on your
1905 operating system's implementation of localtime().
1909 Not implemented. (S<RISC OS>)
1911 Return values (especially for device and inode) may be bogus. (Win32)
1921 Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS)
1925 The C<|> variants are supported only if ToolServer is installed.
1928 open to C<|-> and C<-|> are unsupported. (S<Mac OS>, Win32, S<RISC OS>)
1930 Opening a process does not automatically flush output handles on some
1931 platforms. (SunOS, Solaris, HP-UX)
1935 Very limited functionality. (MiNT)
1939 Not implemented. (Win32, VMS, S<RISC OS>)
1943 Can't move directories between directories on different logical volumes. (Win32)
1947 Only implemented on sockets. (Win32, VMS)
1949 Only reliable on sockets. (S<RISC OS>)
1951 Note that the C<select FILEHANDLE> form is generally portable.
1959 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1963 Not implemented. (S<Mac OS>, MPE/iX, VMS, Win32, S<RISC OS>, VOS)
1967 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1971 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1975 Not implemented. (S<Mac OS>, MPE/iX, Win32, S<RISC OS>, VOS)
1979 Not implemented. (S<Plan 9>)
1989 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1993 A relatively recent addition to socket functions, may not
1994 be implemented even in UNIX platforms.
1998 Not implemented. (Win32, S<RISC OS>, VOS, VM/ESA)
2000 Available on 64 bit OpenVMS 8.2 and later. (VMS)
2004 Platforms that do not have rdev, blksize, or blocks will return these
2005 as '', so numeric comparison or manipulation of these fields may cause
2006 'not numeric' warnings.
2008 mtime and atime are the same thing, and ctime is creation time instead of
2009 inode change time. (S<Mac OS>).
2011 ctime not supported on UFS (S<Mac OS X>).
2013 ctime is creation time instead of inode change time (Win32).
2015 device and inode are not meaningful. (Win32)
2017 device and inode are not necessarily reliable. (VMS)
2019 mtime, atime and ctime all return the last modification time. Device and
2020 inode are not necessarily reliable. (S<RISC OS>)
2022 dev, rdev, blksize, and blocks are not available. inode is not
2023 meaningful and will differ between stat calls on the same file. (os2)
2025 some versions of cygwin when doing a stat("foo") and if not finding it
2026 may then attempt to stat("foo.exe") (Cygwin)
2030 Not implemented. (Win32, VMS, S<RISC OS>)
2034 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
2038 The traditional "0", "1", and "2" MODEs are implemented with different
2039 numeric values on some systems. The flags exported by C<Fcntl>
2040 (O_RDONLY, O_WRONLY, O_RDWR) should work everywhere though. (S<Mac
2041 OS>, OS/390, VM/ESA)
2045 Only implemented if ToolServer is installed. (S<Mac OS>)
2047 As an optimization, may not call the command shell specified in
2048 C<$ENV{PERL5SHELL}>. C<system(1, @args)> spawns an external
2049 process and immediately returns its process designator, without
2050 waiting for it to terminate. Return value may be used subsequently
2051 in C<wait> or C<waitpid>. Failure to spawn() a subprocess is indicated
2052 by setting $? to "255 << 8". C<$?> is set in a way compatible with
2053 Unix (i.e. the exitstatus of the subprocess is obtained by "$? >> 8",
2054 as described in the documentation). (Win32)
2056 There is no shell to process metacharacters, and the native standard is
2057 to pass a command line terminated by "\n" "\r" or "\0" to the spawned
2058 program. Redirection such as C<< > foo >> is performed (if at all) by
2059 the run time library of the spawned program. C<system> I<list> will call
2060 the Unix emulation library's C<exec> emulation, which attempts to provide
2061 emulation of the stdin, stdout, stderr in force in the parent, providing
2062 the child program uses a compatible version of the emulation library.
2063 I<scalar> will call the native command line direct and no such emulation
2064 of a child Unix program will exists. Mileage B<will> vary. (S<RISC OS>)
2066 Far from being POSIX compliant. Because there may be no underlying
2067 /bin/sh tries to work around the problem by forking and execing the
2068 first token in its argument string. Handles basic redirection
2069 ("<" or ">") on its own behalf. (MiNT)
2071 Does not automatically flush output handles on some platforms.
2072 (SunOS, Solaris, HP-UX)
2074 The return value is POSIX-like (shifted up by 8 bits), which only allows
2075 room for a made-up value derived from the severity bits of the native
2076 32-bit condition code (unless overridden by C<use vmsish 'status'>).
2077 If the native condition code is one that has a POSIX value encoded, the
2078 POSIX value will be decoded to extract the expected exit value.
2079 For more details see L<perlvms/$?>. (VMS)
2083 Only the first entry returned is nonzero. (S<Mac OS>)
2085 "cumulative" times will be bogus. On anything other than Windows NT
2086 or Windows 2000, "system" time will be bogus, and "user" time is
2087 actually the time returned by the clock() function in the C runtime
2090 Not useful. (S<RISC OS>)
2094 Not implemented. (Older versions of VMS)
2096 Truncation to same-or-shorter lengths only. (VOS)
2098 If a FILEHANDLE is supplied, it must be writable and opened in append
2099 mode (i.e., use C<<< open(FH, '>>filename') >>>
2100 or C<sysopen(FH,...,O_APPEND|O_RDWR)>. If a filename is supplied, it
2101 should not be held open elsewhere. (Win32)
2105 Returns undef where unavailable, as of version 5.005.
2107 C<umask> works but the correct permissions are set only when the file
2108 is finally closed. (AmigaOS)
2112 Only the modification time is updated. (S<BeOS>, S<Mac OS>, VMS, S<RISC OS>)
2114 May not behave as expected. Behavior depends on the C runtime
2115 library's implementation of utime(), and the filesystem being
2116 used. The FAT filesystem typically does not support an "access
2117 time" field, and it may limit timestamps to a granularity of
2118 two seconds. (Win32)
2124 Not implemented. (S<Mac OS>)
2126 Can only be applied to process handles returned for processes spawned
2127 using C<system(1, ...)> or pseudo processes created with C<fork()>. (Win32)
2129 Not useful. (S<RISC OS>)
2134 =head1 Supported Platforms
2136 As of July 2002 (the Perl release 5.8.0), the following platforms are
2137 able to build Perl from the standard source code distribution
2138 available at http://www.cpan.org/src/index.html
2149 HI-UXMPP (Hitachi) (5.8.0 worked but we didn't know it)
2159 ReliantUNIX (formerly SINIX)
2161 OpenVMS (formerly VMS)
2162 Open UNIX (Unixware) (since Perl 5.8.1/5.9.0)
2164 OS/400 (using the PASE) (since Perl 5.8.1/5.9.0)
2166 POSIX-BC (formerly BS2000)
2171 Tru64 UNIX (formerly DEC OSF/1, Digital UNIX)
2176 Win95/98/ME/2K/XP 2)
2178 z/OS (formerly OS/390)
2181 1) in DOS mode either the DOS or OS/2 ports can be used
2182 2) compilers: Borland, MinGW (GCC), VC6
2184 The following platforms worked with the previous releases (5.6 and
2185 5.7), but we did not manage either to fix or to test these in time
2186 for the 5.8.0 release. There is a very good chance that many of these
2187 will work fine with the 5.8.0.
2200 Known to be broken for 5.8.0 (but 5.6.1 and 5.7.2 can be used):
2204 The following platforms have been known to build Perl from source in
2205 the past (5.005_03 and earlier), but we haven't been able to verify
2206 their status for the current release, either because the
2207 hardware/software platforms are rare or because we don't have an
2208 active champion on these platforms--or both. They used to work,
2209 though, so go ahead and try compiling them, and let perlbug@perl.org
2243 The following platforms have their own source code distributions and
2244 binaries available via http://www.cpan.org/ports/
2248 OS/400 (ILE) 5.005_02
2249 Tandem Guardian 5.004
2251 The following platforms have only binaries available via
2252 http://www.cpan.org/ports/index.html :
2256 Acorn RISCOS 5.005_02
2260 Although we do suggest that you always build your own Perl from
2261 the source code, both for maximal configurability and for security,
2262 in case you are in a hurry you can check
2263 http://www.cpan.org/ports/index.html for binary distributions.
2267 L<perlaix>, L<perlamiga>, L<perlapollo>, L<perlbeos>, L<perlbs2000>,
2268 L<perlce>, L<perlcygwin>, L<perldgux>, L<perldos>, L<perlepoc>,
2269 L<perlebcdic>, L<perlfreebsd>, L<perlhurd>, L<perlhpux>, L<perlirix>,
2270 L<perlmachten>, L<perlmacos>, L<perlmacosx>, L<perlmint>, L<perlmpeix>,
2271 L<perlnetware>, L<perlos2>, L<perlos390>, L<perlos400>,
2272 L<perlplan9>, L<perlqnx>, L<perlsolaris>, L<perltru64>,
2273 L<perlunicode>, L<perlvmesa>, L<perlvms>, L<perlvos>,
2274 L<perlwin32>, and L<Win32>.
2276 =head1 AUTHORS / CONTRIBUTORS
2278 Abigail <abigail@foad.org>,
2279 Charles Bailey <bailey@newman.upenn.edu>,
2280 Graham Barr <gbarr@pobox.com>,
2281 Tom Christiansen <tchrist@perl.com>,
2282 Nicholas Clark <nick@ccl4.org>,
2283 Thomas Dorner <Thomas.Dorner@start.de>,
2284 Andy Dougherty <doughera@lafayette.edu>,
2285 Dominic Dunlop <domo@computer.org>,
2286 Neale Ferguson <neale@vma.tabnsw.com.au>,
2287 David J. Fiander <davidf@mks.com>,
2288 Paul Green <Paul.Green@stratus.com>,
2289 M.J.T. Guy <mjtg@cam.ac.uk>,
2290 Jarkko Hietaniemi <jhi@iki.fi>,
2291 Luther Huffman <lutherh@stratcom.com>,
2292 Nick Ing-Simmons <nick@ing-simmons.net>,
2293 Andreas J. KE<ouml>nig <a.koenig@mind.de>,
2294 Markus Laker <mlaker@contax.co.uk>,
2295 Andrew M. Langmead <aml@world.std.com>,
2296 Larry Moore <ljmoore@freespace.net>,
2297 Paul Moore <Paul.Moore@uk.origin-it.com>,
2298 Chris Nandor <pudge@pobox.com>,
2299 Matthias Neeracher <neeracher@mac.com>,
2300 Philip Newton <pne@cpan.org>,
2301 Gary Ng <71564.1743@CompuServe.COM>,
2302 Tom Phoenix <rootbeer@teleport.com>,
2303 AndrE<eacute> Pirard <A.Pirard@ulg.ac.be>,
2304 Peter Prymmer <pvhp@forte.com>,
2305 Hugo van der Sanden <hv@crypt0.demon.co.uk>,
2306 Gurusamy Sarathy <gsar@activestate.com>,
2307 Paul J. Schinder <schinder@pobox.com>,
2308 Michael G Schwern <schwern@pobox.com>,
2309 Dan Sugalski <dan@sidhe.org>,
2310 Nathan Torkington <gnat@frii.com>.
2311 John Malmberg <wb8tyw@qsl.net>