3 perlport - Writing portable Perl
8 Perl runs on a variety of operating systems. While most of them share
9 a lot in common, they also have their own very particular and unique
12 This document is meant to help you to find out what constitutes portable
13 Perl code, so that once you have made your decision to write portably,
14 you know where the lines are drawn, and you can stay within them.
16 There is a tradeoff between taking full advantage of one particular type
17 of computer, and taking advantage of a full range of them. Naturally,
18 as you make your range bigger (and thus more diverse), the common
19 denominators drop, and you are left with fewer areas of common ground in
20 which you can operate to accomplish a particular task. Thus, when you
21 begin attacking a problem, it is important to consider which part of the
22 tradeoff curve you want to operate under. Specifically, whether it is
23 important to you that the task that you are coding needs the full
24 generality of being portable, or if it is sufficient to just get the job
25 done. This is the hardest choice to be made. The rest is easy, because
26 Perl provides lots of choices, whichever way you want to approach your
29 Looking at it another way, writing portable code is usually about
30 willfully limiting your available choices. Naturally, it takes discipline
33 Be aware of two important points:
38 =item Not all Perl programs have to be portable
40 There is no reason why you should not use Perl as a language to glue Unix
41 tools together, or to prototype a Macintosh application, or to manage the
42 Windows registry. If it makes no sense to aim for portability for one
43 reason or another in a given program, then don't bother.
45 =item The vast majority of Perl I<is> portable
47 Don't be fooled into thinking that it is hard to create portable Perl
48 code. It isn't. Perl tries its level-best to bridge the gaps between
49 what's available on different platforms, and all the means available to
50 use those features. Thus almost all Perl code runs on any machine
51 without modification. But there are some significant issues in
52 writing portable code, and this document is entirely about those issues.
57 Here's the general rule: When you approach a task that is commonly done
58 using a whole range of platforms, think in terms of writing portable
59 code. That way, you don't sacrifice much by way of the implementation
60 choices you can avail yourself of, and at the same time you can give
61 your users lots of platform choices. On the other hand, when you have to
62 take advantage of some unique feature of a particular platform, as is
63 often the case with systems programming (whether for Unix, Windows,
64 S<Mac OS>, VMS, etc.), consider writing platform-specific code.
66 When the code will run on only two or three operating systems, then you
67 may only need to consider the differences of those particular systems.
68 The important thing is to decide where the code will run, and to be
69 deliberate in your decision.
71 The material below is separated into three main sections: main issues of
72 portability (L<"ISSUES">, platform-specific issues (L<"PLATFORMS">, and
73 builtin perl functions that behave differently on various ports
74 (L<"FUNCTION IMPLEMENTATIONS">.
76 This information should not be considered complete; it includes possibly
77 transient information about idiosyncrasies of some of the ports, almost
78 all of which are in a state of constant evolution. Thus this material
79 should be considered a perpetual work in progress
80 (E<lt>IMG SRC="yellow_sign.gif" ALT="Under Construction"E<gt>).
87 In most operating systems, lines in files are terminated by newlines.
88 Just what is used as a newline may vary from OS to OS. Unix
89 traditionally uses C<\012>, one kind of Windows I/O uses C<\015\012>,
90 and S<Mac OS> uses C<\015>.
92 Perl uses C<\n> to represent the "logical" newline, where what
93 is logical may depend on the platform in use. In MacPerl, C<\n>
94 always means C<\015>. In DOSish perls, C<\n> usually means C<\012>, but
95 when accessing a file in "text" mode, STDIO translates it to (or from)
96 C<\015\012>. C<\015\012> is commonly referred to as CRLF.
98 Due to the "text" mode translation, DOSish perls have limitations
99 of using C<seek> and C<tell> when a file is being accessed in "text"
100 mode. Specifically, if you stick to C<seek>-ing to locations you got
101 from C<tell> (and no others), you are usually free to use C<seek> and
102 C<tell> even in "text" mode. In general, using C<seek> or C<tell> or
103 other file operations that count bytes instead of characters, without
104 considering the length of C<\n>, may be non-portable. If you use
105 C<binmode> on a file, however, you can usually use C<seek> and C<tell>
106 with arbitrary values quite safely.
108 A common misconception in socket programming is that C<\n> eq C<\012>
109 everywhere. When using protocols such as common Internet protocols,
110 C<\012> and C<\015> are called for specifically, and the values of
111 the logical C<\n> and C<\r> (carriage return) are not reliable.
113 print SOCKET "Hi there, client!\r\n"; # WRONG
114 print SOCKET "Hi there, client!\015\012"; # RIGHT
116 However, using C<\015\012> (or C<\cM\cJ>, or C<\x0D\x0A>) can be tedious
117 and unsightly, as well as confusing to those maintaining the code. As
118 such, the Socket module supplies the Right Thing for those who want it.
120 use Socket qw(:DEFAULT :crlf);
121 print SOCKET "Hi there, client!$CRLF" # RIGHT
123 When reading from a socket, remember that the default input record
124 separator C<$/> is C<\n>, but code like this should recognize C<$/> as
125 C<\012> or C<\015\012>:
131 Since both CRLF and LF end in LF, the input record separator can
132 be set to LF, and the CR can be stripped later, if present. Better:
134 use Socket qw(:DEFAULT :crlf);
135 local($/) = LF; # not needed if $/ is already \012
138 s/$CR?$LF/\n/; # not sure if socket uses LF or CRLF, OK
139 # s/\015?\012/\n/; # same thing
142 And this example is actually better than the previous one even for Unix
143 platforms, because now any C<\015>'s (C<\cM>'s) are stripped out
144 (and there was much rejoicing).
146 Similarly, functions that return text data--such as a function that
147 fetches a web page--should, in some cases, translate newlines before
148 returning the data, if they've not yet been trsnalted to the local
149 newline. Often one line of code will suffice:
151 $data =~ s/\015?\012/\n/g;
154 Some of this may be confusing. Here's a handy reference to the ASCII CR
155 and LF characters. You can print it out and stick it in your wallet.
157 LF == \012 == \x0A == \cJ == ASCII 10
158 CR == \015 == \x0D == \cM == ASCII 13
161 ---------------------------
164 \n * | LF | CRLF | CR |
165 \r * | CR | CR | LF |
166 ---------------------------
169 These are just the most common definitions of C<\n> and C<\r> in Perl.
170 There may well be others.
173 =head2 Numbers endianness and Width
175 Different CPUs store integers and floating point numbers in different
176 orders (called I<endianness>) and widths (32-bit and 64-bit being the
177 most common). This affects your programs if they attempt to transfer
178 numbers in binary format from one CPU architecture to another over some
179 channel, usually either "live" via network connection, or by storing the
180 numbers to secondary storage such as a disk file.
182 Conflicting storage orders make utter mess out of the numbers: if a
183 little-endian host (Intel, Alpha) stores 0x12345678 (305419896 in
184 decimal), a big-endian host (Motorola, MIPS, Sparc, PA) reads it as
185 0x78563412 (2018915346 in decimal). To avoid this problem in network
186 (socket) connections use the C<pack> and C<unpack> formats C<n>
187 and C<N>, the "network" orders. They are guaranteed to be portable.
189 Different widths can cause truncation even between platforms of equal
190 endianness: the platform of shorter width loses the upper parts of the
191 number. There is no good solution for this problem except to avoid
192 transferring or storing raw binary numbers.
194 One can circumnavigate both these problems in two ways: either
195 transfer and store numbers always in text format, instead of raw
196 binary, or consider using modules like Data::Dumper (included in
197 the standard distribution as of Perl 5.005) and Storable.
199 Also, don't depend on specific values returned from C<printf> and
200 C<sprintf>. These in some cases may differ from platform to platform.
202 =head2 Files and Filesystems
204 Most platforms these days structure files in a hierarchical fashion.
205 So, it is reasonably safe to assume that any platform supports the
206 notion of a "path" to uniquely identify a file on the system. How
207 that path is actually written differs.
209 While they are similar, file path specifications differ between Unix,
210 Windows, S<Mac OS>, OS/2, VMS, VOS, S<RISC OS> and probably others.
211 Unix, for example, is one of the few OSes that has the idea of a single
214 DOS, OS/2, VMS, VOS, and Windows can work similarly to Unix with C</>
215 as path separator, or in their own idiosyncratic ways (such as having
216 several root directories and various "unrooted" device files such NIL:
219 S<Mac OS> uses C<:> as a path separator instead of C</>.
221 The filesystem may support neither hard links (C<link>) nor
222 symbolic links (C<symlink>, C<readlink>, C<lstat>).
224 The filesystem may support neither access timestamp nor change
225 timestamp (meaning that about the only portable timestamp is the
226 modification timestamp), or one second granularity of any timestamps
227 (e.g. the FAT filesystem limits the time granularity to two seconds).
229 VOS perl can emulate Unix filenames with C</> as path separator. The
230 native pathname characters greater-than, less-than, number-sign, and
231 percent-sign are always accepted.
233 S<RISC OS> perl can emulate Unix filenames with C</> as path
234 separator, or go native and use C<.> for path separator and C<:> to
235 signal filesystems and disk names.
237 If all this is intimidating, have no (well, maybe only a little) fear.
238 There are modules that can help. The File::Spec modules provide
239 methods to do the Right Thing on whatever
240 platform happens to be running the program.
242 use File::Spec::Functions;
243 chdir(updir()); # go up one directory
244 $file = catfile(curdir(), 'temp', 'file.txt');
245 # on Unix and Win32, './temp/file.txt'
246 # on Mac OS, ':temp:file.txt'
248 File::Spec is available in the standard distribution, as of version
251 In general, production code should not have file paths hardcoded; making
252 them user supplied or from a configuration file is better, keeping in mind
253 that file path syntax varies on different machines.
255 This is especially noticeable in scripts like Makefiles and test suites,
256 which often assume C</> as a path separator for subdirectories.
258 Also of use is File::Basename, from the standard distribution, which
259 splits a pathname into pieces (base filename, full path to directory,
262 Even when on a single platform (if you can call Unix a single platform),
263 remember not to count on the existence or the contents of
264 system-specific files or directories, like F</etc/passwd>,
265 F</etc/sendmail.conf>, F</etc/resolv.conf>, or even F</tmp/>. For
266 example, F</etc/passwd> may exist but it may not contain the encrypted
267 passwords because the system is using some form of enhanced security,
268 or it may not contain all the accounts because the system is using NIS.
269 If code does need to rely on such a file, include a description of the
270 file and its format in the code's documentation, and make it easy for
271 the user to override the default location of the file.
273 Don't assume a text file will end with a newline.
275 Do not have two files of the same name with different case, like
276 F<test.pl> and F<Test.pl>, as many platforms have case-insensitive
277 filenames. Also, try not to have non-word characters (except for C<.>)
278 in the names, and keep them to the 8.3 convention, for maximum
281 Likewise, if using the AutoSplit module, try to keep the split functions to
282 8.3 naming and case-insensitive conventions; or, at the very least,
283 make it so the resulting files have a unique (case-insensitively)
286 There certainly can be whitespace in filenames on most systems, but
287 some may not allow it. Many systems (DOS, VMS) cannot have more than
288 one C<.> in their filenames.
290 Don't assume C<E<gt>> won't be the first character of a filename.
291 Always use C<E<lt>> explicitly to open a file for reading.
293 open(FILE, "< $existing_file") or die $!;
295 If filenames might use strange characters, it is safest to open it
296 with C<sysopen> instead of C<open>. C<open> is magic and can
297 translate characters like C<E<gt>>, C<E<lt>>, and C<|>, which may
298 be the wrong thing to do.
301 =head2 System Interaction
303 Not all platforms provide for the notion of a command line, necessarily.
304 These are usually platforms that rely on a Graphical User Interface (GUI)
305 for user interaction. So a program requiring command lines might not work
306 everywhere. But this is probably for the user of the program to deal
307 with, so don't stay up late worrying about it.
309 Some platforms can't delete or rename files that are being held open by
310 the system. Remember to C<close> files when you are done with them.
311 Don't C<unlink> or C<rename> an open file. Don't C<tie> or C<open> a
312 file that is already tied or opened; C<untie> or C<close> first.
314 Don't open the same file more than once at a time for writing, as some
315 operating systems put mandatory locks on such files.
317 Don't count on a specific environment variable existing in C<%ENV>.
318 Don't count on C<%ENV> entries being case-sensitive, or even
321 Don't count on signals for anything.
323 Don't count on filename globbing. Use C<opendir>, C<readdir>, and
326 Don't count on per-program environment variables, or per-program current
329 Don't count on specific values of C<$!>.
332 =head2 Interprocess Communication (IPC)
334 In general, don't directly access the system in code that is meant to be
335 portable. That means, no C<system>, C<exec>, C<fork>, C<pipe>, C<``>,
336 C<qx//>, C<open> with a C<|>, nor any of the other things that makes being
337 a Unix perl hacker worth being.
339 Commands that launch external processes are generally supported on
340 most platforms (though many of them do not support any type of forking),
341 but the problem with using them arises from what you invoke with them.
342 External tools are often named differently on different platforms, often
343 not available in the same location, often accept different arguments,
344 often behave differently, and often represent their results in a
345 platform-dependent way. Thus you should seldom depend on them to produce
348 One especially common bit of Perl code is opening a pipe to sendmail:
350 open(MAIL, '|/usr/lib/sendmail -t') or die $!;
352 This is fine for systems programming when sendmail is known to be
353 available. But it is not fine for many non-Unix systems, and even
354 some Unix systems that may not have sendmail installed. If a portable
355 solution is needed, see the C<Mail::Send> and C<Mail::Mailer> modules
356 in the C<MailTools> distribution. C<Mail::Mailer> provides several
357 mailing methods, including mail, sendmail, and direct SMTP
358 (via C<Net::SMTP>) if a mail transfer agent is not available.
360 The rule of thumb for portable code is: Do it all in portable Perl, or
361 use a module (that may internally implement it with platform-specific
362 code, but expose a common interface).
364 The Unix System V IPC (C<msg*(), sem*(), shm*()>) is not available
365 even in all Unix platforms.
368 =head2 External Subroutines (XS)
370 XS code, in general, can be made to work with any platform; but dependent
371 libraries, header files, etc., might not be readily available or
372 portable, or the XS code itself might be platform-specific, just as Perl
373 code might be. If the libraries and headers are portable, then it is
374 normally reasonable to make sure the XS code is portable, too.
376 There is a different kind of portability issue with writing XS
377 code: availability of a C compiler on the end-user's system. C brings
378 with it its own portability issues, and writing XS code will expose you to
379 some of those. Writing purely in perl is a comparatively easier way to
383 =head2 Standard Modules
385 In general, the standard modules work across platforms. Notable
386 exceptions are the CPAN module (which currently makes connections to external
387 programs that may not be available), platform-specific modules (like
388 ExtUtils::MM_VMS), and DBM modules.
390 There is no one DBM module that is available on all platforms.
391 SDBM_File and the others are generally available on all Unix and DOSish
392 ports, but not in MacPerl, where only NBDM_File and DB_File are
395 The good news is that at least some DBM module should be available, and
396 AnyDBM_File will use whichever module it can find. Of course, then
397 the code needs to be fairly strict, dropping to the lowest common
398 denominator (e.g., not exceeding 1K for each record), so that it will
399 work with any DBM module. See L<AnyDBM_File> for more details.
404 The system's notion of time of day and calendar date is controlled in
405 widely different ways. Don't assume the timezone is stored in C<$ENV{TZ}>,
406 and even if it is, don't assume that you can control the timezone through
409 Don't assume that the epoch starts at 00:00:00, January 1, 1970,
410 because that is OS- and implementation-specific. It is better to store a date
411 in an unambiguous representation. The ISO-8601 standard defines
412 "YYYY-MM-DD" as the date format. A text representation (like "1987-12-18")
413 can be easily converted into an OS-specific value using a module like
414 Date::Parse. An array of values, such as those returned by
415 C<localtime>, can be converted to an OS-specific representation using
418 When calculating specific times, such as for tests in time or date modules,
419 it may be appropriate to calculate an offset for the epoch.
421 $offset = Time::Local::timegm(0, 0, 0, 1, 0, 70);
422 The value for C<$offset> in Unix will be C<0>, but in Mac OS will be
423 some large number. C<$offset> can then be added to a Unix time value
424 to get what should be the proper value on any system.
426 =head2 Character sets and character encoding
428 Assume very little about character sets. Do not assume anything about
429 the numerical values (C<ord>, C<chr>) of characters. Do not
430 assume that the alphabetic characters are encoded contiguously (in
431 numerical sense). Do not assume anything about the ordering of the
432 characters. The lowercase letters may come before or after the
433 uppercase letters, the lowercase and uppercase may be interlaced so
434 that both 'a' and 'A' come before the 'b', the accented and other
435 international characters may be interlaced so that E<auml> comes
439 =head2 Internationalisation
441 If you may assume POSIX (a rather large assumption, that in practice
442 means Unix), you may read more about the POSIX locale system from
443 L<perllocale>. The locale system at least attempts to make things a
444 little bit more portable, or at least more convenient and
445 native-friendly for non-English users. The system affects character
446 sets and encoding, and date and time formatting, among other things.
449 =head2 System Resources
451 If your code is destined for systems with severely constrained (or
452 missing!) virtual memory systems then you want to be I<especially> mindful
453 of avoiding wasteful constructs such as:
455 # NOTE: this is no longer "bad" in perl5.005
456 for (0..10000000) {} # bad
457 for (my $x = 0; $x <= 10000000; ++$x) {} # good
459 @lines = <VERY_LARGE_FILE>; # bad
461 while (<FILE>) {$file .= $_} # sometimes bad
462 $file = join('', <FILE>); # better
464 The last two may appear unintuitive to most people. The first of those
465 two constructs repeatedly grows a string, while the second allocates a
466 large chunk of memory in one go. On some systems, the latter is more
467 efficient that the former.
472 Most multi-user platforms provide basic levels of security that is usually
473 felt at the file-system level. Other platforms usually don't
474 (unfortunately). Thus the notion of user id, or "home" directory, or even
475 the state of being logged-in, may be unrecognizable on many platforms. If
476 you write programs that are security-conscious, it is usually best to know
477 what type of system you will be operating under, and write code explicitly
478 for that platform (or class of platforms).
483 For those times when it is necessary to have platform-specific code,
484 consider keeping the platform-specific code in one place, making porting
485 to other platforms easier. Use the Config module and the special
486 variable C<$^O> to differentiate platforms, as described in
489 Be careful in the tests you supply with your module or programs. Often
490 module code is fully portable, but the tests are not. This occurs
491 often when tests spawn off other processes or call external programs
492 to aid in the testing, or when (as noted above) the tests assume certain
493 things about the filesystem and paths.
494 Be careful not to depend on a specific output style for errors,
495 such as when checking C<$@> after an C<eval>. Some platforms
496 expect a certain output format, and perl on those platforms may
497 have been adjusted accordingly. Most specifically, don't anchor
498 a regex when testing an error value.
500 $@ =~ /^I got an error!/ # may fail
501 $@ =~ /I got an error!/ # probably better
506 Modules uploaded to CPAN are tested by a variety of volunteers on
507 different platforms. These CPAN testers are notified by mail of each
508 new upload, and reply to the list with PASS, FAIL, NA (not applicable to
509 this platform), or UNKNOWN (unknown), along with any relevant notations.
511 The purpose of the testing is twofold: one, to help developers fix any
512 problems in their code that crop up because of lack of testing on other
513 platforms; two, to provide users with information about whether or not
514 a given module works on a given platform.
518 =item Mailing list: cpan-testers@perl.org
520 =item Testing results: C<http://www.perl.org/cpan-testers/>
527 As of version 5.002, Perl is built with a C<$^O> variable that
528 indicates the operating system it was built on. This was implemented
529 to help speed up code that would otherwise have to C<use Config;> and
530 use the value of C<$Config{'osname'}>. Of course, to get
531 detailed information about the system, looking into C<%Config> is
532 certainly recommended.
534 C<%Config> cannot always be trusted, however,
535 because it is built at compile time, and if perl was built in once
536 place and transferred elsewhere, some values may be off, or the
537 values may have been edited after the fact.
542 Perl works on a bewildering variety of Unix and Unix-like platforms (see
543 e.g. most of the files in the F<hints/> directory in the source code kit).
544 On most of these systems, the value of C<$^O> (hence C<$Config{'osname'}>,
545 too) is determined by lowercasing and stripping punctuation from the first
546 field of the string returned by typing C<uname -a> (or a similar command)
547 at the shell prompt. Here, for example, are a few of the more popular
550 uname $^O $Config{'archname'}
551 --------------------------------------------
553 BSD/OS bsdos i386-bsdos
554 dgux dgux AViiON-dgux
555 DYNIX/ptx dynixptx i386-dynixptx
556 FreeBSD freebsd freebsd-i386
557 Linux linux i386-linux
558 Linux linux i586-linux
559 Linux linux ppc-linux
560 HP-UX hpux PA-RISC1.1
562 openbsd openbsd i386-openbsd
563 OSF1 dec_osf alpha-dec_osf
564 reliantunix-n svr4 RM400-svr4
565 SCO_SV sco_sv i386-sco_sv
566 SINIX-N svr4 RM400-svr4
567 sn4609 unicos CRAY_C90-unicos
568 sn6521 unicosmk t3e-unicosmk
569 sn9617 unicos CRAY_J90-unicos
570 SunOS solaris sun4-solaris
571 SunOS solaris i86pc-solaris
572 SunOS4 sunos sun4-sunos
574 Note that because the C<$Config{'archname'}> may depend on the hardware
575 architecture it may vary quite a lot, much more than the C<$^O>.
578 =head2 DOS and Derivatives
580 Perl has long been ported to PC style microcomputers running under
581 systems like PC-DOS, MS-DOS, OS/2, and most Windows platforms you can
582 bring yourself to mention (except for Windows CE, if you count that).
583 Users familiar with I<COMMAND.COM> and/or I<CMD.EXE> style shells should
584 be aware that each of these file specifications may have subtle
587 $filespec0 = "c:/foo/bar/file.txt";
588 $filespec1 = "c:\\foo\\bar\\file.txt";
589 $filespec2 = 'c:\foo\bar\file.txt';
590 $filespec3 = 'c:\\foo\\bar\\file.txt';
592 System calls accept either C</> or C<\> as the path separator. However,
593 many command-line utilities of DOS vintage treat C</> as the option
594 prefix, so they may get confused by filenames containing C</>. Aside
595 from calling any external programs, C</> will work just fine, and
596 probably better, as it is more consistent with popular usage, and avoids
597 the problem of remembering what to backwhack and what not to.
599 The DOS FAT filesystem can only accommodate "8.3" style filenames. Under
600 the "case insensitive, but case preserving" HPFS (OS/2) and NTFS (NT)
601 filesystems you may have to be careful about case returned with functions
602 like C<readdir> or used with functions like C<open> or C<opendir>.
604 DOS also treats several filenames as special, such as AUX, PRN, NUL, CON,
605 COM1, LPT1, LPT2 etc. Unfortunately these filenames won't even work
606 if you include an explicit directory prefix, in some cases. It is best
607 to avoid such filenames, if you want your code to be portable to DOS
610 Users of these operating systems may also wish to make use of
611 scripts such as F<pl2bat.bat> or F<pl2cmd> as appropriate to
612 put wrappers around your scripts.
614 Newline (C<\n>) is translated as C<\015\012> by STDIO when reading from
615 and writing to files (see L<"Newlines">). C<binmode(FILEHANDLE)>
616 will keep C<\n> translated as C<\012> for that filehandle. Since it is a
617 no-op on other systems, C<binmode> should be used for cross-platform code
618 that deals with binary data.
620 The C<$^O> variable and the C<$Config{'archname'}> values for various
621 DOSish perls are as follows:
623 OS $^O $Config{'archname'}
624 --------------------------------------------
628 Windows 95 MSWin32 MSWin32-x86
629 Windows 98 MSWin32 MSWin32-x86
630 Windows NT MSWin32 MSWin32-x86
631 Windows NT MSWin32 MSWin32-ALPHA
632 Windows NT MSWin32 MSWin32-ppc
638 =item The djgpp environment for DOS, C<http://www.delorie.com/djgpp/>
640 =item The EMX environment for DOS, OS/2, etc. C<emx@iaehv.nl>,
641 C<http://www.leo.org/pub/comp/os/os2/leo/gnu/emx+gcc/index.html> or
642 C<ftp://hobbes.nmsu.edu/pub/os2/dev/emx>
644 =item Build instructions for Win32, L<perlwin32>.
646 =item The ActiveState Pages, C<http://www.activestate.com/>
653 Any module requiring XS compilation is right out for most people, because
654 MacPerl is built using non-free (and non-cheap!) compilers. Some XS
655 modules that can work with MacPerl are built and distributed in binary
658 Directories are specified as:
660 volume:folder:file for absolute pathnames
661 volume:folder: for absolute pathnames
662 :folder:file for relative pathnames
663 :folder: for relative pathnames
664 :file for relative pathnames
665 file for relative pathnames
667 Files in a directory are stored in alphabetical order. Filenames are
668 limited to 31 characters, and may include any character except for
669 null and C<:>, which is reserved as path separator.
671 Instead of C<flock>, see C<FSpSetFLock> and C<FSpRstFLock> in the
672 Mac::Files module, or C<chmod(0444, ...)> and C<chmod(0666, ...)>.
674 In the MacPerl application, you can't run a program from the command line;
675 programs that expect C<@ARGV> to be populated can be edited with something
676 like the following, which brings up a dialog box asking for the command
680 @ARGV = split /\s+/, MacPerl::Ask('Arguments?');
683 A MacPerl script saved as a droplet will populate C<@ARGV> with the full
684 pathnames of the files dropped onto the script.
686 Mac users can use programs on a kind of command line under MPW (Macintosh
687 Programmer's Workshop, a free development environment from Apple).
688 MacPerl was first introduced as an MPW tool, and MPW can be used like a
691 perl myscript.plx some arguments
693 ToolServer is another app from Apple that provides access to MPW tools
694 from MPW and the MacPerl app, which allows MacPerl programs to use
695 C<system>, backticks, and piped C<open>.
697 "S<Mac OS>" is the proper name for the operating system, but the value
698 in C<$^O> is "MacOS". To determine architecture, version, or whether
699 the application or MPW tool version is running, check:
701 $is_app = $MacPerl::Version =~ /App/;
702 $is_tool = $MacPerl::Version =~ /MPW/;
703 ($version) = $MacPerl::Version =~ /^(\S+)/;
704 $is_ppc = $MacPerl::Architecture eq 'MacPPC';
705 $is_68k = $MacPerl::Architecture eq 'Mac68K';
707 S<Mac OS X> and S<Mac OS X Server>, based on NeXT's OpenStep OS, will
708 (in theory) be able to run MacPerl natively, under the "Classic"
709 environment. The new "Cocoa" environment (formerly called the "Yellow Box")
710 may run a slightly modified version of MacPerl, using the Carbon interfaces.
712 S<Mac OS X Server> and its Open Source version, Darwin, both run Unix
713 perl natively (with a small number of patches). Full support for these
714 is slated for perl5.006.
721 =item The MacPerl Pages, C<http://www.macperl.com/>.
723 =item The MacPerl mailing lists, C<http://www.macperl.org/>.
725 =item MacPerl Module Porters, C<http://pudge.net/mmp/>.
732 Perl on VMS is discussed in F<vms/perlvms.pod> in the perl distribution.
733 Note that perl on VMS can accept either VMS- or Unix-style file
734 specifications as in either of the following:
736 $ perl -ne "print if /perl_setup/i" SYS$LOGIN:LOGIN.COM
737 $ perl -ne "print if /perl_setup/i" /sys$login/login.com
739 but not a mixture of both as in:
741 $ perl -ne "print if /perl_setup/i" sys$login:/login.com
742 Can't open sys$login:/login.com: file specification syntax error
744 Interacting with Perl from the Digital Command Language (DCL) shell
745 often requires a different set of quotation marks than Unix shells do.
748 $ perl -e "print ""Hello, world.\n"""
751 There are a number of ways to wrap your perl scripts in DCL .COM files if
752 you are so inclined. For example:
754 $ write sys$output "Hello from DCL!"
756 $ then perl -x 'f$environment("PROCEDURE")
757 $ else perl -x - 'p1 'p2 'p3 'p4 'p5 'p6 'p7 'p8
758 $ deck/dollars="__END__"
761 print "Hello from Perl!\n";
766 Do take care with C<$ ASSIGN/nolog/user SYS$COMMAND: SYS$INPUT> if your
767 perl-in-DCL script expects to do things like C<$read = E<lt>STDINE<gt>;>.
769 Filenames are in the format "name.extension;version". The maximum
770 length for filenames is 39 characters, and the maximum length for
771 extensions is also 39 characters. Version is a number from 1 to
772 32767. Valid characters are C</[A-Z0-9$_-]/>.
774 VMS' RMS filesystem is case insensitive and does not preserve case.
775 C<readdir> returns lowercased filenames, but specifying a file for
776 opening remains case insensitive. Files without extensions have a
777 trailing period on them, so doing a C<readdir> with a file named F<A.;5>
778 will return F<a.> (though that file could be opened with
781 RMS had an eight level limit on directory depths from any rooted logical
782 (allowing 16 levels overall) prior to VMS 7.2. Hence
783 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8]> is a valid directory specification but
784 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8.9]> is not. F<Makefile.PL> authors might
785 have to take this into account, but at least they can refer to the former
786 as C</PERL_ROOT/lib/2/3/4/5/6/7/8/>.
788 The VMS::Filespec module, which gets installed as part of the build
789 process on VMS, is a pure Perl module that can easily be installed on
790 non-VMS platforms and can be helpful for conversions to and from RMS
793 What C<\n> represents depends on the type of file that is open. It could
794 be C<\015>, C<\012>, C<\015\012>, or nothing. Reading from a file
795 translates newlines to C<\012>, unless C<binmode> was executed on that
796 handle, just like DOSish perls.
798 TCP/IP stacks are optional on VMS, so socket routines might not be
799 implemented. UDP sockets may not be supported.
801 The value of C<$^O> on OpenVMS is "VMS". To determine the architecture
802 that you are running on without resorting to loading all of C<%Config>
803 you can examine the content of the C<@INC> array like so:
805 if (grep(/VMS_AXP/, @INC)) {
806 print "I'm on Alpha!\n";
808 } elsif (grep(/VMS_VAX/, @INC)) {
809 print "I'm on VAX!\n";
812 print "I'm not so sure about where $^O is...\n";
815 On VMS perl determines the UTC offset from the C<SYS$TIMEZONE_DIFFERENTIAL>
816 logical name. Though the VMS epoch began at 17-NOV-1858 00:00:00.00,
817 calls to C<localtime> are adjusted to count offsets from
818 01-JAN-1970 00:00:00.00 just like Unix.
826 =item vmsperl list, C<majordomo@perl.org>
828 Put the words C<subscribe vmsperl> in message body.
830 =item vmsperl on the web, C<http://www.sidhe.org/vmsperl/index.html>
837 Perl on VOS is discussed in F<README.vos> in the perl distribution.
838 Note that perl on VOS can accept either VOS- or Unix-style file
839 specifications as in either of the following:
841 $ perl -ne "print if /perl_setup/i" >system>notices
842 $ perl -ne "print if /perl_setup/i" /system/notices
844 or even a mixture of both as in:
846 $ perl -ne "print if /perl_setup/i" >system/notices
848 Note that even though VOS allows the slash character to appear in object
849 names, because the VOS port of Perl interprets it as a pathname
850 delimiting character, VOS files, directories, or links whose names
851 contain a slash character cannot be processed. Such files must be
852 renamed before they can be processed by Perl.
854 The following C functions are unimplemented on VOS, and any attempt by
855 Perl to use them will result in a fatal error message and an immediate
856 exit from Perl: dup, do_aspawn, do_spawn, fork, waitpid. Once these
857 functions become available in the VOS POSIX.1 implementation, you can
858 either recompile and rebind Perl, or you can download a newer port from
861 The value of C<$^O> on VOS is "VOS". To determine the architecture that
862 you are running on without resorting to loading all of C<%Config> you
863 can examine the content of the C<@INC> array like so:
865 if (grep(/VOS/, @INC)) {
866 print "I'm on a Stratus box!\n";
868 print "I'm not on a Stratus box!\n";
872 if (grep(/860/, @INC)) {
873 print "This box is a Stratus XA/R!\n";
875 } elsif (grep(/7100/, @INC)) {
876 print "This box is a Stratus HP 7100 or 8000!\n";
878 } elsif (grep(/8000/, @INC)) {
879 print "This box is a Stratus HP 8000!\n";
882 print "This box is a Stratus 68K...\n";
891 =item VOS mailing list
893 There is no specific mailing list for Perl on VOS. You can post
894 comments to the comp.sys.stratus newsgroup, or subscribe to the general
895 Stratus mailing list. Send a letter with "Subscribe Info-Stratus" in
896 the message body to majordomo@list.stratagy.com.
898 =item VOS Perl on the web at C<http://ftp.stratus.com/pub/vos/vos.html>
903 =head2 EBCDIC Platforms
905 Recent versions of Perl have been ported to platforms such as OS/400 on
906 AS/400 minicomputers as well as OS/390 & VM/ESA for IBM Mainframes. Such
907 computers use EBCDIC character sets internally (usually Character Code
908 Set ID 00819 for OS/400 and IBM-1047 for OS/390 & VM/ESA). Note that on
909 the mainframe perl currently works under the "Unix system services
910 for OS/390" (formerly known as OpenEdition) and VM/ESA OpenEdition.
912 As of R2.5 of USS for OS/390 and Version 2.3 of VM/ESA these Unix
913 sub-systems do not support the C<#!> shebang trick for script invocation.
914 Hence, on OS/390 and VM/ESA perl scripts can be executed with a header
915 similar to the following simple script:
918 eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
920 #!/usr/local/bin/perl # just a comment really
922 print "Hello from perl!\n";
924 On the AS/400, assuming that PERL5 is in your library list, you may need
925 to wrap your perl scripts in a CL procedure to invoke them like so:
928 CALL PGM(PERL5/PERL) PARM('/QOpenSys/hello.pl')
931 This will invoke the perl script F<hello.pl> in the root of the
932 QOpenSys file system. On the AS/400 calls to C<system> or backticks
935 On these platforms, bear in mind that the EBCDIC character set may have
936 an effect on what happens with some perl functions (such as C<chr>,
937 C<pack>, C<print>, C<printf>, C<ord>, C<sort>, C<sprintf>, C<unpack>), as
938 well as bit-fiddling with ASCII constants using operators like C<^>, C<&>
939 and C<|>, not to mention dealing with socket interfaces to ASCII computers
942 Fortunately, most web servers for the mainframe will correctly translate
943 the C<\n> in the following statement to its ASCII equivalent (note that
944 C<\r> is the same under both Unix and OS/390 & VM/ESA):
946 print "Content-type: text/html\r\n\r\n";
948 The value of C<$^O> on OS/390 is "os390".
950 The value of C<$^O> on VM/ESA is "vmesa".
952 Some simple tricks for determining if you are running on an EBCDIC
953 platform could include any of the following (perhaps all):
955 if ("\t" eq "\05") { print "EBCDIC may be spoken here!\n"; }
957 if (ord('A') == 193) { print "EBCDIC may be spoken here!\n"; }
959 if (chr(169) eq 'z') { print "EBCDIC may be spoken here!\n"; }
961 Note that one thing you may not want to rely on is the EBCDIC encoding
962 of punctuation characters since these may differ from code page to code
963 page (and once your module or script is rumoured to work with EBCDIC,
964 folks will want it to work with all EBCDIC character sets).
972 The perl-mvs@perl.org list is for discussion of porting issues as well as
973 general usage issues for all EBCDIC Perls. Send a message body of
974 "subscribe perl-mvs" to majordomo@perl.org.
976 =item AS/400 Perl information at C<http://as400.rochester.ibm.com/>
983 As Acorns use ASCII with newlines (C<\n>) in text files as C<\012> like
984 Unix and Unix filename emulation is turned on by default, it is quite
985 likely that most simple scripts will work "out of the box". The native
986 filesystem is modular, and individual filesystems are free to be
987 case-sensitive or insensitive, and are usually case-preserving. Some
988 native filesystems have name length limits which file and directory
989 names are silently truncated to fit. Scripts should be aware that the
990 standard filesystem currently has a name length limit of B<10>
991 characters, with up to 77 items in a directory, but other filesystems
992 may not impose such limitations.
994 Native filenames are of the form
996 Filesystem#Special_Field::DiskName.$.Directory.Directory.File
1000 Special_Field is not usually present, but may contain . and $ .
1001 Filesystem =~ m|[A-Za-z0-9_]|
1002 DsicName =~ m|[A-Za-z0-9_/]|
1003 $ represents the root directory
1004 . is the path separator
1005 @ is the current directory (per filesystem but machine global)
1006 ^ is the parent directory
1007 Directory and File =~ m|[^\0- "\.\$\%\&:\@\\^\|\177]+|
1009 The default filename translation is roughly C<tr|/.|./|;>
1011 Note that C<"ADFS::HardDisk.$.File" ne 'ADFS::HardDisk.$.File'> and that
1012 the second stage of C<$> interpolation in regular expressions will fall
1013 foul of the C<$.> if scripts are not careful.
1015 Logical paths specified by system variables containing comma-separated
1016 search lists are also allowed, hence C<System:Modules> is a valid
1017 filename, and the filesystem will prefix C<Modules> with each section of
1018 C<System$Path> until a name is made that points to an object on disk.
1019 Writing to a new file C<System:Modules> would only be allowed if
1020 C<System$Path> contains a single item list. The filesystem will also
1021 expand system variables in filenames if enclosed in angle brackets, so
1022 C<E<lt>System$DirE<gt>.Modules> would look for the file
1023 S<C<$ENV{'System$Dir'} . 'Modules'>>. The obvious implication of this is
1024 that B<fully qualified filenames can start with C<E<lt>E<gt>>> and should
1025 be protected when C<open> is used for input.
1027 Because C<.> was in use as a directory separator and filenames could not
1028 be assumed to be unique after 10 characters, Acorn implemented the C
1029 compiler to strip the trailing C<.c> C<.h> C<.s> and C<.o> suffix from
1030 filenames specified in source code and store the respective files in
1031 subdirectories named after the suffix. Hence files are translated:
1034 C:foo.h C:h.foo (logical path variable)
1035 sys/os.h sys.h.os (C compiler groks Unix-speak)
1036 10charname.c c.10charname
1037 10charname.o o.10charname
1038 11charname_.c c.11charname (assuming filesystem truncates at 10)
1040 The Unix emulation library's translation of filenames to native assumes
1041 that this sort of translation is required, and allows a user defined list
1042 of known suffixes which it will transpose in this fashion. This may
1043 appear transparent, but consider that with these rules C<foo/bar/baz.h>
1044 and C<foo/bar/h/baz> both map to C<foo.bar.h.baz>, and that C<readdir> and
1045 C<glob> cannot and do not attempt to emulate the reverse mapping. Other
1046 C<.>'s in filenames are translated to C</>.
1048 As implied above the environment accessed through C<%ENV> is global, and
1049 the convention is that program specific environment variables are of the
1050 form C<Program$Name>. Each filesystem maintains a current directory,
1051 and the current filesystem's current directory is the B<global> current
1052 directory. Consequently, sociable scripts don't change the current
1053 directory but rely on full pathnames, and scripts (and Makefiles) cannot
1054 assume that they can spawn a child process which can change the current
1055 directory without affecting its parent (and everyone else for that
1058 As native operating system filehandles are global and currently are
1059 allocated down from 255, with 0 being a reserved value the Unix emulation
1060 library emulates Unix filehandles. Consequently, you can't rely on
1061 passing C<STDIN>, C<STDOUT>, or C<STDERR> to your children.
1063 The desire of users to express filenames of the form
1064 C<E<lt>Foo$DirE<gt>.Bar> on the command line unquoted causes problems,
1065 too: C<``> command output capture has to perform a guessing game. It
1066 assumes that a string C<E<lt>[^E<lt>E<gt>]+\$[^E<lt>E<gt>]E<gt>> is a
1067 reference to an environment variable, whereas anything else involving
1068 C<E<lt>> or C<E<gt>> is redirection, and generally manages to be 99%
1069 right. Of course, the problem remains that scripts cannot rely on any
1070 Unix tools being available, or that any tools found have Unix-like command
1073 Extensions and XS are, in theory, buildable by anyone using free tools.
1074 In practice, many don't, as users of the Acorn platform are used to binary
1075 distribution. MakeMaker does run, but no available make currently copes
1076 with MakeMaker's makefiles; even if/when this is fixed, the lack of a
1077 Unix-like shell can cause problems with makefile rules, especially lines
1078 of the form C<cd sdbm && make all>, and anything using quoting.
1080 "S<RISC OS>" is the proper name for the operating system, but the value
1081 in C<$^O> is "riscos" (because we don't like shouting).
1086 Perl has been ported to a variety of platforms that do not fit into any of
1087 the above categories. Some, such as AmigaOS, Atari MiNT, BeOS, HP MPE/iX,
1088 QNX, Plan 9, and VOS, have been well-integrated into the standard Perl source
1089 code kit. You may need to see the F<ports/> directory on CPAN for
1090 information, and possibly binaries, for the likes of: aos, Atari ST, lynxos,
1091 riscos, Novell Netware, Tandem Guardian, I<etc.> (yes we know that some of
1092 these OSes may fall under the Unix category, but we are not a standards body.)
1098 =item Atari, Guido Flohr's page C<http://stud.uni-sb.de/~gufl0000/>
1100 =item HP 300 MPE/iX C<http://www.cccd.edu/~markb/perlix.html>
1102 =item Novell Netware
1104 A free perl5-based PERL.NLM for Novell Netware is available in
1105 precompiled binary and source code form from C<http://www.novell.com/>
1106 as well as from CPAN.
1111 =head1 FUNCTION IMPLEMENTATIONS
1113 Listed below are functions unimplemented or implemented differently on
1114 various platforms. Following each description will be, in parentheses, a
1115 list of platforms that the description applies to.
1117 The list may very well be incomplete, or wrong in some places. When in
1118 doubt, consult the platform-specific README files in the Perl source
1119 distribution, and other documentation resources for a given port.
1121 Be aware, moreover, that even among Unix-ish systems there are variations.
1123 For many functions, you can also query C<%Config>, exported by default
1124 from the Config module. For example, to check if the platform has the C<lstat>
1125 call, check C<$Config{'d_lstat'}>. See L<Config> for a full
1126 description of available variables.
1129 =head2 Alphabetical Listing of Perl Functions
1139 C<-r>, C<-w>, and C<-x> have only a very limited meaning; directories
1140 and applications are executable, and there are no uid/gid
1141 considerations. C<-o> is not supported. (S<Mac OS>)
1143 C<-r>, C<-w>, C<-x>, and C<-o> tell whether or not file is accessible,
1144 which may not reflect UIC-based file protections. (VMS)
1146 C<-s> returns the size of the data fork, not the total size of data fork
1147 plus resource fork. (S<Mac OS>).
1149 C<-s> by name on an open file will return the space reserved on disk,
1150 rather than the current extent. C<-s> on an open filehandle returns the
1151 current size. (S<RISC OS>)
1153 C<-R>, C<-W>, C<-X>, C<-O> are indistinguishable from C<-r>, C<-w>,
1154 C<-x>, C<-o>. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1156 C<-b>, C<-c>, C<-k>, C<-g>, C<-p>, C<-u>, C<-A> are not implemented.
1159 C<-g>, C<-k>, C<-l>, C<-p>, C<-u>, C<-A> are not particularly meaningful.
1160 (Win32, VMS, S<RISC OS>)
1162 C<-d> is true if passed a device spec without an explicit directory.
1165 C<-T> and C<-B> are implemented, but might misclassify Mac text files
1166 with foreign characters; this is the case will all platforms, but may
1167 affect S<Mac OS> often. (S<Mac OS>)
1169 C<-x> (or C<-X>) determine if a file ends in one of the executable
1170 suffixes. C<-S> is meaningless. (Win32)
1172 C<-x> (or C<-X>) determine if a file has an executable file type.
1175 =item binmode FILEHANDLE
1177 Meaningless. (S<Mac OS>, S<RISC OS>)
1179 Reopens file and restores pointer; if function fails, underlying
1180 filehandle may be closed, or pointer may be in a different position.
1183 The value returned by C<tell> may be affected after the call, and
1184 the filehandle may be flushed. (Win32)
1188 Only limited meaning. Disabling/enabling write permission is mapped to
1189 locking/unlocking the file. (S<Mac OS>)
1191 Only good for changing "owner" read-write access, "group", and "other"
1192 bits are meaningless. (Win32)
1194 Only good for changing "owner" and "other" read-write access. (S<RISC OS>)
1196 Access permissions are mapped onto VOS access-control list changes. (VOS)
1200 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>, VOS)
1202 Does nothing, but won't fail. (Win32)
1204 =item chroot FILENAME
1208 Not implemented. (S<Mac OS>, Win32, VMS, Plan9, S<RISC OS>, VOS, VM/ESA)
1210 =item crypt PLAINTEXT,SALT
1212 May not be available if library or source was not provided when building
1215 Not implemented. (VOS)
1219 Not implemented. (VMS, Plan9, VOS)
1221 =item dbmopen HASH,DBNAME,MODE
1223 Not implemented. (VMS, Plan9, VOS)
1227 Not useful. (S<Mac OS>, S<RISC OS>)
1229 Not implemented. (Win32)
1231 Invokes VMS debugger. (VMS)
1235 Not implemented. (S<Mac OS>)
1237 Implemented via Spawn. (VM/ESA)
1239 =item fcntl FILEHANDLE,FUNCTION,SCALAR
1241 Not implemented. (Win32, VMS)
1243 =item flock FILEHANDLE,OPERATION
1245 Not implemented (S<Mac OS>, VMS, S<RISC OS>, VOS).
1247 Available only on Windows NT (not on Windows 95). (Win32)
1251 Not implemented. (S<Mac OS>, Win32, AmigaOS, S<RISC OS>, VOS, VM/ESA)
1255 Not implemented. (S<Mac OS>, S<RISC OS>)
1259 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1263 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1265 =item getpriority WHICH,WHO
1267 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1271 Not implemented. (S<Mac OS>, Win32)
1273 Not useful. (S<RISC OS>)
1277 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1279 =item getnetbyname NAME
1281 Not implemented. (S<Mac OS>, Win32, Plan9)
1285 Not implemented. (S<Mac OS>, Win32)
1287 Not useful. (S<RISC OS>)
1291 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1293 =item getnetbyaddr ADDR,ADDRTYPE
1295 Not implemented. (S<Mac OS>, Win32, Plan9)
1297 =item getprotobynumber NUMBER
1299 Not implemented. (S<Mac OS>)
1301 =item getservbyport PORT,PROTO
1303 Not implemented. (S<Mac OS>)
1307 Not implemented. (S<Mac OS>, Win32, VM/ESA)
1311 Not implemented. (S<Mac OS>, Win32, VMS, VM/ESA)
1315 Not implemented. (S<Mac OS>, Win32)
1319 Not implemented. (S<Mac OS>, Win32, Plan9)
1323 Not implemented. (S<Mac OS>, Win32, Plan9)
1327 Not implemented. (Win32, Plan9)
1331 Not implemented. (S<Mac OS>, Win32, S<RISC OS>)
1335 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1337 =item sethostent STAYOPEN
1339 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
1341 =item setnetent STAYOPEN
1343 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
1345 =item setprotoent STAYOPEN
1347 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
1349 =item setservent STAYOPEN
1351 Not implemented. (Plan9, Win32, S<RISC OS>)
1355 Not implemented. (S<Mac OS>, Win32, VM/ESA)
1359 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VM/ESA)
1363 Not implemented. (S<Mac OS>, Win32)
1367 Not implemented. (S<Mac OS>, Win32, Plan9)
1371 Not implemented. (S<Mac OS>, Win32, Plan9)
1375 Not implemented. (Plan9, Win32)
1377 =item getsockopt SOCKET,LEVEL,OPTNAME
1379 Not implemented. (S<Mac OS>, Plan9)
1385 Globbing built-in, but only C<*> and C<?> metacharacters are supported.
1388 Features depend on external perlglob.exe or perlglob.bat. May be
1389 overridden with something like File::DosGlob, which is recommended.
1392 Globbing built-in, but only C<*> and C<?> metacharacters are supported.
1393 Globbing relies on operating system calls, which may return filenames
1394 in any order. As most filesystems are case-insensitive, even "sorted"
1395 filenames will not be in case-sensitive order. (S<RISC OS>)
1397 =item ioctl FILEHANDLE,FUNCTION,SCALAR
1399 Not implemented. (VMS)
1401 Available only for socket handles, and it does what the ioctlsocket() call
1402 in the Winsock API does. (Win32)
1404 Available only for socket handles. (S<RISC OS>)
1408 Not implemented, hence not useful for taint checking. (S<Mac OS>,
1411 Available only for process handles returned by the C<system(1, ...)>
1412 method of spawning a process. (Win32)
1414 =item link OLDFILE,NEWFILE
1416 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1418 Link count not updated because hard links are not quite that hard
1419 (They are sort of half-way between hard and soft links). (AmigaOS)
1421 =item lstat FILEHANDLE
1427 Not implemented. (VMS, S<RISC OS>)
1429 Return values may be bogus. (Win32)
1431 =item msgctl ID,CMD,ARG
1433 =item msgget KEY,FLAGS
1435 =item msgsnd ID,MSG,FLAGS
1437 =item msgrcv ID,VAR,SIZE,TYPE,FLAGS
1439 Not implemented. (S<Mac OS>, Win32, VMS, Plan9, S<RISC OS>, VOS)
1441 =item open FILEHANDLE,EXPR
1443 =item open FILEHANDLE
1445 The C<|> variants are only supported if ToolServer is installed.
1448 open to C<|-> and C<-|> are unsupported. (S<Mac OS>, Win32, S<RISC OS>)
1450 =item pipe READHANDLE,WRITEHANDLE
1452 Not implemented. (S<Mac OS>)
1454 Very limited functionality. (MiNT)
1460 Not implemented. (Win32, VMS, S<RISC OS>)
1462 =item select RBITS,WBITS,EBITS,TIMEOUT
1464 Only implemented on sockets. (Win32)
1466 Only reliable on sockets. (S<RISC OS>)
1468 =item semctl ID,SEMNUM,CMD,ARG
1470 =item semget KEY,NSEMS,FLAGS
1472 =item semop KEY,OPSTRING
1474 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1476 =item setpgrp PID,PGRP
1478 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1480 =item setpriority WHICH,WHO,PRIORITY
1482 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1484 =item setsockopt SOCKET,LEVEL,OPTNAME,OPTVAL
1486 Not implemented. (S<Mac OS>, Plan9)
1488 =item shmctl ID,CMD,ARG
1490 =item shmget KEY,SIZE,FLAGS
1492 =item shmread ID,VAR,POS,SIZE
1494 =item shmwrite ID,STRING,POS,SIZE
1496 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1498 =item socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL
1500 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1502 =item stat FILEHANDLE
1508 mtime and atime are the same thing, and ctime is creation time instead of
1509 inode change time. (S<Mac OS>)
1511 device and inode are not meaningful. (Win32)
1513 device and inode are not necessarily reliable. (VMS)
1515 mtime, atime and ctime all return the last modification time. Device and
1516 inode are not necessarily reliable. (S<RISC OS>)
1518 =item symlink OLDFILE,NEWFILE
1520 Not implemented. (Win32, VMS, S<RISC OS>)
1524 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1526 =item sysopen FILEHANDLE,FILENAME,MODE,PERMS
1528 The traditional "0", "1", and "2" MODEs are implemented with different
1529 numeric values on some systems. The flags exported by C<Fcntl>
1530 (O_RDONLY, O_WRONLY, O_RDWR) should work everywhere though. (S<Mac
1531 OS>, OS/390, VM/ESA)
1535 Only implemented if ToolServer is installed. (S<Mac OS>)
1537 As an optimization, may not call the command shell specified in
1538 C<$ENV{PERL5SHELL}>. C<system(1, @args)> spawns an external
1539 process and immediately returns its process designator, without
1540 waiting for it to terminate. Return value may be used subsequently
1541 in C<wait> or C<waitpid>. (Win32)
1543 There is no shell to process metacharacters, and the native standard is
1544 to pass a command line terminated by "\n" "\r" or "\0" to the spawned
1545 program. Redirection such as C<E<gt> foo> is performed (if at all) by
1546 the run time library of the spawned program. C<system> I<list> will call
1547 the Unix emulation library's C<exec> emulation, which attempts to provide
1548 emulation of the stdin, stdout, stderr in force in the parent, providing
1549 the child program uses a compatible version of the emulation library.
1550 I<scalar> will call the native command line direct and no such emulation
1551 of a child Unix program will exists. Mileage B<will> vary. (S<RISC OS>)
1553 Far from being POSIX compliant. Because there may be no underlying
1554 /bin/sh tries to work around the problem by forking and execing the
1555 first token in its argument string. Handles basic redirection
1556 ("E<lt>" or "E<gt>") on its own behalf. (MiNT)
1560 Only the first entry returned is nonzero. (S<Mac OS>)
1562 "cumulative" times will be bogus. On anything other than Windows NT,
1563 "system" time will be bogus, and "user" time is actually the time
1564 returned by the clock() function in the C runtime library. (Win32)
1566 Not useful. (S<RISC OS>)
1568 =item truncate FILEHANDLE,LENGTH
1570 =item truncate EXPR,LENGTH
1572 Not implemented. (VMS)
1574 Truncation to zero-length only. (VOS)
1576 If a FILEHANDLE is supplied, it must be writable and opened in append
1577 mode (i.e., use C<open(FH, '>>filename')>
1578 or C<sysopen(FH,...,O_APPEND|O_RDWR)>. If a filename is supplied, it
1579 should not be held open elsewhere. (Win32)
1585 Returns undef where unavailable, as of version 5.005.
1587 C<umask()> works but the correct permissions are only set when the file
1588 is finally close()d. (AmigaOS)
1592 Only the modification time is updated. (S<Mac OS>, VMS, S<RISC OS>)
1594 May not behave as expected. Behavior depends on the C runtime
1595 library's implementation of utime(), and the filesystem being
1596 used. The FAT filesystem typically does not support an "access
1597 time" field, and it may limit timestamps to a granularity of
1598 two seconds. (Win32)
1602 =item waitpid PID,FLAGS
1604 Not implemented. (S<Mac OS>, VOS)
1606 Can only be applied to process handles returned for processes spawned
1607 using C<system(1, ...)>. (Win32)
1609 Not useful. (S<RISC OS>)
1617 =item v1.42, 22 May 1999
1618 Added notes about tests, sprintf/printf, and epoch offsets.
1619 =item v1.41, 19 May 1999
1621 Lots more little changes to formatting and content.
1623 Added a bunch of <$^O> and related values
1624 for various platforms; fixed mail and web addresses, and added
1625 and changed miscellaneous notes. (Peter Prymmer)
1627 =item v1.40, 11 April 1999
1629 Miscellaneous changes.
1631 =item v1.39, 11 February 1999
1633 Changes from Jarkko and EMX URL fixes Michael Schwern. Additional
1634 note about newlines added.
1636 =item v1.38, 31 December 1998
1638 More changes from Jarkko.
1640 =item v1.37, 19 December 1998
1642 More minor changes. Merge two separate version 1.35 documents.
1644 =item v1.36, 9 September 1998
1646 Updated for Stratus VOS. Also known as version 1.35.
1648 =item v1.35, 13 August 1998
1650 Integrate more minor changes, plus addition of new sections under
1651 L<"ISSUES">: L<"Numbers endianness and Width">,
1652 L<"Character sets and character encoding">,
1653 L<"Internationalisation">.
1655 =item v1.33, 06 August 1998
1657 Integrate more minor changes.
1659 =item v1.32, 05 August 1998
1661 Integrate more minor changes.
1663 =item v1.30, 03 August 1998
1665 Major update for RISC OS, other minor changes.
1667 =item v1.23, 10 July 1998
1669 First public release with perl5.005.
1673 =head1 AUTHORS / CONTRIBUTORS
1675 Abigail E<lt>abigail@fnx.comE<gt>,
1676 Charles Bailey E<lt>bailey@newman.upenn.eduE<gt>,
1677 Graham Barr E<lt>gbarr@pobox.comE<gt>,
1678 Tom Christiansen E<lt>tchrist@perl.comE<gt>,
1679 Nicholas Clark E<lt>Nicholas.Clark@liverpool.ac.ukE<gt>,
1680 Andy Dougherty E<lt>doughera@lafcol.lafayette.eduE<gt>,
1681 Dominic Dunlop E<lt>domo@vo.luE<gt>,
1682 Neale Ferguson E<lt>neale@mailbox.tabnsw.com.auE<gt>
1683 Paul Green E<lt>Paul_Green@stratus.comE<gt>,
1684 M.J.T. Guy E<lt>mjtg@cus.cam.ac.ukE<gt>,
1685 Jarkko Hietaniemi E<lt>jhi@iki.fi<gt>,
1686 Luther Huffman E<lt>lutherh@stratcom.comE<gt>,
1687 Nick Ing-Simmons E<lt>nick@ni-s.u-net.comE<gt>,
1688 Andreas J. KE<ouml>nig E<lt>koenig@kulturbox.deE<gt>,
1689 Markus Laker E<lt>mlaker@contax.co.ukE<gt>,
1690 Andrew M. Langmead E<lt>aml@world.std.comE<gt>,
1691 Larry Moore E<lt>ljmoore@freespace.netE<gt>,
1692 Paul Moore E<lt>Paul.Moore@uk.origin-it.comE<gt>,
1693 Chris Nandor E<lt>pudge@pobox.comE<gt>,
1694 Matthias Neeracher E<lt>neeri@iis.ee.ethz.chE<gt>,
1695 Gary Ng E<lt>71564.1743@CompuServe.COME<gt>,
1696 Tom Phoenix E<lt>rootbeer@teleport.comE<gt>,
1697 Peter Prymmer E<lt>pvhp@forte.comE<gt>,
1698 Hugo van der Sanden E<lt>hv@crypt0.demon.co.ukE<gt>,
1699 Gurusamy Sarathy E<lt>gsar@umich.eduE<gt>,
1700 Paul J. Schinder E<lt>schinder@pobox.comE<gt>,
1701 Michael G Schwern E<lt>schwern@pobox.comE<gt>,
1702 Dan Sugalski E<lt>sugalskd@ous.eduE<gt>,
1703 Nathan Torkington E<lt>gnat@frii.comE<gt>.
1705 This document is maintained by Chris Nandor
1706 E<lt>pudge@pobox.comE<gt>.
1710 Version 1.42, last modified 22 May 1999