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 B<a> particular type
17 of computer, and taking advantage of a full B<range> of them. Naturally,
18 as you make your range bigger (and thus more diverse), the common denominators
19 drop, and you are left with fewer areas of common ground in which
20 you can operate to accomplish a particular task. Thus, when you begin
21 attacking a problem, it is important to consider which part of the tradeoff
22 curve you want to operate under. Specifically, whether it is important to
23 you that the task that you are coding needs the full generality of being
24 portable, or if it is sufficient to just get the job done. This is the
25 hardest choice to be made. The rest is easy, because Perl provides lots
26 of choices, whichever way you want to approach your problem.
28 Looking at it another way, writing portable code is usually about willfully
29 limiting your available choices. Naturally, it takes discipline to do that.
31 Be aware of two important points:
35 =item Not all Perl programs have to be portable
37 There is no reason why you should not use Perl as a language to glue Unix
38 tools together, or to prototype a Macintosh application, or to manage the
39 Windows registry. If it makes no sense to aim for portability for one
40 reason or another in a given program, then don't bother.
42 =item The vast majority of Perl B<is> portable
44 Don't be fooled into thinking that it is hard to create portable Perl
45 code. It isn't. Perl tries its level-best to bridge the gaps between
46 what's available on different platforms, and all the means available to
47 use those features. Thus almost all Perl code runs on any machine
48 without modification. But there I<are> some significant issues in
49 writing portable code, and this document is entirely about those issues.
53 Here's the general rule: When you approach a task that is commonly done
54 using a whole range of platforms, think in terms of writing portable
55 code. That way, you don't sacrifice much by way of the implementation
56 choices you can avail yourself of, and at the same time you can give
57 your users lots of platform choices. On the other hand, when you have to
58 take advantage of some unique feature of a particular platform, as is
59 often the case with systems programming (whether for Unix, Windows,
60 S<Mac OS>, VMS, etc.), consider writing platform-specific code.
62 When the code will run on only two or three operating systems, then you may
63 only need to consider the differences of those particular systems. The
64 important thing is to decide where the code will run, and to be deliberate
67 This information should not be considered complete; it includes possibly
68 transient information about idiosyncrasies of some of the ports, almost
69 all of which are in a state of constant evolution. Thus this material
70 should be considered a perpetual work in progress
71 (E<lt>IMG SRC="yellow_sign.gif" ALT="Under Construction"E<gt>).
78 In most operating systems, lines in files are separated with newlines.
79 Just what is used as a newline may vary from OS to OS. Unix
80 traditionally uses C<\012>, one kind of Windows I/O uses C<\015\012>,
81 and S<Mac OS> uses C<\015>.
83 Perl uses C<\n> to represent the "logical" newline, where what
84 is logical may depend on the platform in use. In MacPerl, C<\n>
85 always means C<\015>. In DOSish perls, C<\n> usually means C<\012>, but
86 when accessing a file in "text" mode, STDIO translates it to (or from)
89 Due to the "text" mode translation, DOSish perls have limitations
90 of using C<seek> and C<tell> when a file is being accessed in "text"
91 mode. Specifically, if you stick to C<seek>-ing to locations you got
92 from C<tell> (and no others), you are usually free to use C<seek> and
93 C<tell> even in "text" mode. In general, using C<seek> or C<tell> or
94 other file operations that count bytes instead of characters, without
95 considering the length of C<\n>, may be non-portable. If you use
96 C<binmode> on a file, however, you can usually use C<seek> and C<tell>
97 with arbitrary values quite safely.
99 A common misconception in socket programming is that C<\n> eq C<\012>
100 everywhere. When using protocols, such as common Internet protocols,
101 C<\012> and C<\015> are called for specifically, and the values of
102 the logical C<\n> and C<\r> (carriage return) are not reliable.
104 print SOCKET "Hi there, client!\r\n"; # WRONG
105 print SOCKET "Hi there, client!\015\012"; # RIGHT
107 [NOTE: this does not necessarily apply to communications that are
108 filtered by another program or module before sending to the socket; the
109 the most popular EBCDIC webserver, for instance, accepts C<\r\n>,
110 which translates those characters, along with all other
111 characters in text streams, from EBCDIC to ASCII.]
113 However, C<\015\012> (or C<\cM\cJ>, or C<\x0D\x0A>) can be tedious and
114 unsightly, as well as confusing to those maintaining the code. As such,
115 the C<Socket> module supplies the Right Thing for those who want it.
117 use Socket qw(:DEFAULT :crlf);
118 print SOCKET "Hi there, client!$CRLF" # RIGHT
120 When reading I<from> a socket, remember that the default input record
121 separator (C<$/>) is C<\n>, but code like this should recognize C<$/> as
122 C<\012> or C<\015\012>:
130 use Socket qw(:DEFAULT :crlf);
131 local($/) = LF; # not needed if $/ is already \012
134 s/$CR?$LF/\n/; # not sure if socket uses LF or CRLF, OK
135 # s/\015?\012/\n/; # same thing
138 And this example is actually better than the previous one even for Unix
139 platforms, because now any C<\015>'s (C<\cM>'s) are stripped out
140 (and there was much rejoicing).
145 Most platforms these days structure files in a hierarchical fashion.
146 So, it is reasonably safe to assume that any platform supports the
147 notion of a "path" to uniquely identify a file on the system. Just
148 how that path is actually written, differs.
150 While they are similar, file path specifications differ between Unix,
151 Windows, S<Mac OS>, OS/2, VMS, S<RISC OS> and probably others. Unix,
152 for example, is one of the few OSes that has the idea of a root directory.
153 S<Mac OS> uses C<:> as a path separator instead of C</>. VMS, Windows, and
154 OS/2 can work similarly to Unix with C</> as path separator, or in their own
155 idiosyncratic ways. C<RISC OS> perl can emulate Unix filenames with C</>
156 as path separator, or go native and use C<.> for path separator and C<:>
157 to signal filing systems and disc names.
159 As with the newline problem above, there are modules that can help. The
160 C<File::Spec> modules provide methods to do the Right Thing on whatever
161 platform happens to be running the program.
164 chdir(File::Spec->updir()); # go up one directory
165 $file = File::Spec->catfile(
166 File::Spec->curdir(), 'temp', 'file.txt'
168 # on Unix and Win32, './temp/file.txt'
169 # on Mac OS, ':temp:file.txt'
171 File::Spec is available in the standard distribution, as of version
174 In general, production code should not have file paths hardcoded; making
175 them user supplied or from a configuration file is better, keeping in mind
176 that file path syntax varies on different machines.
178 This is especially noticeable in scripts like Makefiles and test suites,
179 which often assume C</> as a path separator for subdirectories.
181 Also of use is C<File::Basename>, from the standard distribution, which
182 splits a pathname into pieces (base filename, full path to directory,
185 Remember not to count on the existence of system-specific files, like
186 F</etc/resolv.conf>. If code does need to rely on such a file, include a
187 description of the file and its format in the code's documentation, and
188 make it easy for the user to override the default location of the file.
190 Don't assume that a you can open a full pathname for input with
191 C<open (FILE, $name)>, as some platforms can use characters such as C<E<lt>>
192 which will perl C<open> will interpret and eat.
194 Do not have two files of the same name with different case, like
195 F<test.pl> and <Test.pl>, as many platforms have case-insensitive
196 filenames. Also, try not to have non-word characters (except for C<.>)
197 in the names, and keep them to the 8.3 convention, for maximum portability.
199 Likewise, if using C<AutoSplit>, try to keep the split functions to
200 8.3 naming and case-insensitive conventions; or, at the very least,
201 make it so the resulting files have a unique (case-insensitively)
205 =head2 System Interaction
207 Not all platforms provide for the notion of a command line, necessarily.
208 These are usually platforms that rely on a Graphical User Interface (GUI)
209 for user interaction. So a program requiring command lines might not work
210 everywhere. But this is probably for the user of the program to deal
213 Some platforms can't delete or rename files that are being held open by
214 the system. Remember to C<close> files when you are done with them.
215 Don't C<unlink> or C<rename> an open file. Don't C<tie> to or C<open> a
216 file that is already tied to or opened; C<untie> or C<close> first.
218 Don't count on a specific environment variable existing in C<%ENV>.
219 Don't even count on C<%ENV> entries being case-sensitive, or even
222 Don't count on signals in portable programs.
224 Don't count on filename globbing. Use C<opendir>, C<readdir>, and
227 Don't count on per-program environment variables, or per-program current
231 =head2 Interprocess Communication (IPC)
233 In general, don't directly access the system in code that is meant to be
234 portable. That means, no: C<system>, C<exec>, C<fork>, C<pipe>, C<``>,
235 C<qx//>, C<open> with a C<|>, or any of the other things that makes being
236 a Unix perl hacker worth being.
238 Commands that launch external processes are generally supported on
239 most platforms (though many of them do not support any type of forking),
240 but the problem with using them arises from what you invoke with them.
241 External tools are often named differently on different platforms, often
242 not available in the same location, often accept different arguments,
243 often behave differently, and often represent their results in a
244 platform-dependent way. Thus you should seldom depend on them to produce
247 One especially common bit of Perl code is opening a pipe to sendmail:
249 open(MAIL, '|/usr/lib/sendmail -t') or die $!;
251 This is fine for systems programming when sendmail is known to be
252 available. But it is not fine for many non-Unix systems, and even
253 some Unix systems that may not have sendmail installed. If a portable
254 solution is needed, see the C<Mail::Send> and C<Mail::Mailer> modules
255 in the C<MailTools> distribution. C<Mail::Mailer> provides several
256 mailing methods, including mail, sendmail, and direct SMTP
257 (via C<Net::SMTP>) if a mail transfer agent is not available.
259 The rule of thumb for portable code is: Do it all in portable Perl, or
260 use a module that may internally implement it with platform-specific code,
261 but expose a common interface. By portable Perl, we mean code that
262 avoids the constructs described in this document as being non-portable.
265 =head2 External Subroutines (XS)
267 XS code, in general, can be made to work with any platform; but dependent
268 libraries, header files, etc., might not be readily available or
269 portable, or the XS code itself might be platform-specific, just as Perl
270 code might be. If the libraries and headers are portable, then it is
271 normally reasonable to make sure the XS code is portable, too.
273 There is a different kind of portability issue with writing XS
274 code: availability of a C compiler on the end-user's system. C brings with
275 it its own portability issues, and writing XS code will expose you to
276 some of those. Writing purely in perl is a comparatively easier way to
280 =head2 Standard Modules
282 In general, the standard modules work across platforms. Notable
283 exceptions are C<CPAN.pm> (which currently makes connections to external
284 programs that may not be available), platform-specific modules (like
285 C<ExtUtils::MM_VMS>), and DBM modules.
287 There is no one DBM module that is available on all platforms.
288 C<SDBM_File> and the others are generally available on all Unix and DOSish
289 ports, but not in MacPerl, where C<NBDM_File> and C<DB_File> are available.
291 The good news is that at least some DBM module should be available, and
292 C<AnyDBM_File> will use whichever module it can find. Of course, then
293 the code needs to be fairly strict, dropping to the lowest common
294 denominator (e.g., not exceeding 1K for each record).
299 The system's notion of time of day and calendar date is controlled in widely
300 different ways. Don't assume the timezone is stored in C<$ENV{TZ}>, and even
301 if it is, don't assume that you can control the timezone through that
304 Don't assume that the epoch starts at January 1, 1970, because that is
305 OS-specific. Better to store a date in an unambiguous representation.
306 A text representation (like C<1 Jan 1970>) can be easily converted into an
307 OS-specific value using a module like C<Date::Parse>. An array of values,
308 such as those returned by C<localtime>, can be converted to an OS-specific
309 representation using C<Time::Local>.
312 =head2 System Resources
314 If your code is destined for systems with severely constrained (or missing!)
315 virtual memory systems then you want to be especially mindful of avoiding
316 wasteful constructs such as:
318 # NOTE: this is no longer "bad" in perl5.005
319 for (0..10000000) {} # bad
320 for (my $x = 0; $x <= 10000000; ++$x) {} # good
322 @lines = <VERY_LARGE_FILE>; # bad
324 while (<FILE>) {$file .= $_} # sometimes bad
325 $file = join '', <FILE>; # better
327 The last two may appear unintuitive to most people. The first of those
328 two constructs repeatedly grows a string, while the second allocates a
329 large chunk of memory in one go. On some systems, the latter is more
330 efficient that the former.
334 Most multi-user platforms provide basic levels of security that is usually felt
335 at the file-system level. Other platforms usually don't (unfortunately).
336 Thus the notion of User-ID, or "home" directory, or even the state of
337 being logged-in may be unrecognizable on many platforms. If you write
338 programs that are security conscious, it is usually best to know what
339 type of system you will be operating under, and write code explicitly
340 for that platform (or class of platforms).
344 For those times when it is necessary to have platform-specific code,
345 consider keeping the platform-specific code in one place, making porting
346 to other platforms easier. Use the C<Config> module and the special
347 variable C<$^O> to differentiate platforms, as described in L<"PLATFORMS">.
352 Module uploaded to CPAN are tested by a variety of volunteers on
353 different platforms. These CPAN testers are notified by e-mail of each
354 new upload, and reply to the list with PASS, FAIL, NA (not applicable to
355 this platform), or ???? (unknown), along with any relevant notations.
357 The purpose of the testing is twofold: one, to help developers fix any
358 problems in their code; two, to provide users with information about
359 whether or not a given module works on a given platform.
363 =item Mailing list: cpan-testers@perl.org
365 =item Testing results: C<http://www.connect.net/gbarr/cpan-test/>
372 As of version 5.002, Perl is built with a C<$^O> variable that
373 indicates the operating system it was built on. This was implemented
374 to help speed up code that would otherwise have to C<use Config;> and
375 use the value of C<$Config{'osname'}>. Of course, to get
376 detailed information about the system, looking into C<%Config> is
377 certainly recommended.
381 Perl works on a bewildering variety of Unix and Unix-like platforms (see
382 e.g. most of the files in the F<hints/> directory in the source code kit).
383 On most of these systems, the value of C<$^O> (hence C<$Config{'osname'}>,
384 too) is determined by lowercasing and stripping punctuation from the first
385 field of the string returned by typing
389 (or a similar command) at the shell prompt. Here, for example, are a few
390 of the more popular Unix flavors:
392 uname $^O $Config{'archname'}
393 -------------------------------------------
399 SunOS solaris sun4-solaris
400 SunOS solaris i86pc-solaris
404 =head2 DOS and Derivatives
406 Perl has long been ported to PC style microcomputers running under
407 systems like PC-DOS, MS-DOS, OS/2, and most Windows platforms you can
408 bring yourself to mention (except for Windows CE, if you count that).
409 Users familiar with I<COMMAND.COM> and/or I<CMD.EXE> style shells should
410 be aware that each of these file specifications may have subtle
413 $filespec0 = "c:/foo/bar/file.txt";
414 $filespec1 = "c:\\foo\\bar\\file.txt";
415 $filespec2 = 'c:\foo\bar\file.txt';
416 $filespec3 = 'c:\\foo\\bar\\file.txt';
418 System calls accept either C</> or C<\> as the path separator. However,
419 many command-line utilities of DOS vintage treat C</> as the option
420 prefix, so they may get confused by filenames containing C</>. Aside
421 from calling any external programs, C</> will work just fine, and
422 probably better, as it is more consistent with popular usage, and avoids
423 the problem of remembering what to backwhack and what not to.
425 The DOS FAT file system can only accommodate "8.3" style filenames. Under
426 the "case insensitive, but case preserving" HPFS (OS/2) and NTFS (NT)
427 file systems you may have to be careful about case returned with functions
428 like C<readdir> or used with functions like C<open> or C<opendir>.
430 DOS also treats several filenames as special, such as AUX, PRN, NUL, CON,
431 COM1, LPT1, LPT2 etc. Unfortunately these filenames won't even work
432 if you include an explicit directory prefix, in some cases. It is best
433 to avoid such filenames, if you want your code to be portable to DOS
436 Users of these operating systems may also wish to make use of
437 scripts such as I<pl2bat.bat> or I<pl2cmd> as appropriate to
438 put wrappers around your scripts.
440 Newline (C<\n>) is translated as C<\015\012> by STDIO when reading from
441 and writing to files. C<binmode(FILEHANDLE)> will keep C<\n> translated
442 as C<\012> for that filehandle. Since it is a noop on other systems,
443 C<binmode> should be used for cross-platform code that deals with binary
446 The C<$^O> variable and the C<$Config{'archname'}> values for various
447 DOSish perls are as follows:
449 OS $^O $Config{'archname'}
450 --------------------------------------------
454 Windows 95 MSWin32 MSWin32-x86
455 Windows NT MSWin32 MSWin32-x86
456 Windows NT MSWin32 MSWin32-alpha
457 Windows NT MSWin32 MSWin32-ppc
463 =item The djgpp environment for DOS, C<http://www.delorie.com/djgpp/>
465 =item The EMX environment for DOS, OS/2, etc. C<emx@iaehv.nl>,
466 C<http://www.juge.com/bbs/Hobb.19.html>
468 =item Build instructions for Win32, L<perlwin32>.
470 =item The ActiveState Pages, C<http://www.activestate.com/>
477 Any module requiring XS compilation is right out for most people, because
478 MacPerl is built using non-free (and non-cheap!) compilers. Some XS
479 modules that can work with MacPerl are built and distributed in binary
480 form on CPAN. See I<MacPerl: Power and Ease> for more details.
482 Directories are specified as:
484 volume:folder:file for absolute pathnames
485 volume:folder: for absolute pathnames
486 :folder:file for relative pathnames
487 :folder: for relative pathnames
488 :file for relative pathnames
489 file for relative pathnames
491 Files in a directory are stored in alphabetical order. Filenames are
492 limited to 31 characters, and may include any character except C<:>,
493 which is reserved as a path separator.
495 Instead of C<flock>, see C<FSpSetFLock> and C<FSpRstFLock> in
498 In the MacPerl application, you can't run a program from the command line;
499 programs that expect C<@ARGV> to be populated can be edited with something
500 like the following, which brings up a dialog box asking for the command
504 @ARGV = split /\s+/, MacPerl::Ask('Arguments?');
507 A MacPerl script saved as a droplet will populate C<@ARGV> with the full
508 pathnames of the files dropped onto the script.
510 Mac users can use programs on a kind of command line under MPW (Macintosh
511 Programmer's Workshop, a free development environment from Apple).
512 MacPerl was first introduced as an MPW tool, and MPW can be used like a
515 perl myscript.plx some arguments
517 ToolServer is another app from Apple that provides access to MPW tools
518 from MPW and the MacPerl app, which allows MacPerl program to use
519 C<system>, backticks, and piped C<open>.
521 "S<Mac OS>" is the proper name for the operating system, but the value
522 in C<$^O> is "MacOS". To determine architecture, version, or whether
523 the application or MPW tool version is running, check:
525 $is_app = $MacPerl::Version =~ /App/;
526 $is_tool = $MacPerl::Version =~ /MPW/;
527 ($version) = $MacPerl::Version =~ /^(\S+)/;
528 $is_ppc = $MacPerl::Architecture eq 'MacPPC';
529 $is_68k = $MacPerl::Architecture eq 'Mac68K';
531 S<Mac OS X>, to be based on NeXT's OpenStep OS, will be able to run MacPerl
532 natively (in the Blue Box, and even in the Yellow Box, once some changes
533 to the toolbox calls are made), but Unix perl will also run natively.
539 =item The MacPerl Pages, C<http://www.ptf.com/macperl/>.
541 =item The MacPerl mailing list, C<mac-perl-request@iis.ee.ethz.ch>.
548 Perl on VMS is discussed in F<vms/perlvms.pod> in the perl distribution.
549 Note that perl on VMS can accept either VMS or Unix style file
550 specifications as in either of the following:
552 $ perl -ne "print if /perl_setup/i" SYS$LOGIN:LOGIN.COM
553 $ perl -ne "print if /perl_setup/i" /sys$login/login.com
555 but not a mixture of both as in:
557 $ perl -ne "print if /perl_setup/i" sys$login:/login.com
558 Can't open sys$login:/login.com: file specification syntax error
560 Interacting with Perl from the Digital Command Language (DCL) shell
561 often requires a different set of quotation marks than Unix shells do.
564 $ perl -e "print ""Hello, world.\n"""
567 There are a number of ways to wrap your perl scripts in DCL .COM files if
568 you are so inclined. For example:
570 $ write sys$output "Hello from DCL!"
572 $ then perl -x 'f$environment("PROCEDURE")
573 $ else perl -x - 'p1 'p2 'p3 'p4 'p5 'p6 'p7 'p8
574 $ deck/dollars="__END__"
577 print "Hello from Perl!\n";
582 Do take care with C<$ ASSIGN/nolog/user SYS$COMMAND: SYS$INPUT> if your
583 perl-in-DCL script expects to do things like C<$read = E<lt>STDINE<gt>;>.
585 Filenames are in the format "name.extension;version". The maximum
586 length for filenames is 39 characters, and the maximum length for
587 extensions is also 39 characters. Version is a number from 1 to
588 32767. Valid characters are C</[A-Z0-9$_-]/>.
590 VMS' RMS filesystem is case insensitive and does not preserve case.
591 C<readdir> returns lowercased filenames, but specifying a file for
592 opening remains case insensitive. Files without extensions have a
593 trailing period on them, so doing a C<readdir> with a file named F<A.;5>
594 will return F<a.> (though that file could be opened with C<open(FH, 'A')>.
596 RMS had an eight level limit on directory depths from any rooted logical
597 (allowing 16 levels overall) prior to VMS 7.2. Hence
598 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8]> is a valid directory specification but
599 C<PERL_ROOT:[LIB.2.3.4.5.6.7.8.9]> is not. F<Makefile.PL> authors might
600 have to take this into account, but at least they can refer to the former
601 as C</PERL_ROOT/lib/2/3/4/5/6/7/8/>.
603 The C<VMS::Filespec> module, which gets installed as part
604 of the build process on VMS, is a pure Perl module that can easily be
605 installed on non-VMS platforms and can be helpful for conversions to
606 and from RMS native formats.
608 What C<\n> represents depends on the type of file that is open. It could
609 be C<\015>, C<\012>, C<\015\012>, or nothing. Reading from a file
610 translates newlines to C<\012>, unless C<binmode> was executed on that
611 handle, just like DOSish perls.
613 TCP/IP stacks are optional on VMS, so socket routines might not be
614 implemented. UDP sockets may not be supported.
616 The value of C<$^O> on OpenVMS is "VMS". To determine the architecture
617 that you are running on without resorting to loading all of C<%Config>
618 you can examine the content of the C<@INC> array like so:
620 if (grep(/VMS_AXP/, @INC)) {
621 print "I'm on Alpha!\n";
622 } elsif (grep(/VMS_VAX/, @INC)) {
623 print "I'm on VAX!\n";
625 print "I'm not so sure about where $^O is...\n";
634 =item vmsperl list, C<vmsperl-request@newman.upenn.edu>
636 Put words C<SUBSCRIBE VMSPERL> in message body.
638 =item vmsperl on the web, C<http://www.sidhe.org/vmsperl/index.html>
643 =head2 EBCDIC Platforms
645 Recent versions of Perl have been ported to platforms such as OS/400 on
646 AS/400 minicomputers as well as OS/390 for IBM Mainframes. Such computers
647 use EBCDIC character sets internally (usually Character Code Set ID 00819
648 for OS/400 and IBM-1047 for OS/390). Note that on the mainframe perl
649 currently works under the "Unix system services for OS/390" (formerly
650 known as OpenEdition).
652 As of R2.5 of USS for OS/390 that Unix sub-system did not support the
653 C<#!> shebang trick for script invocation. Hence, on OS/390 perl scripts
654 can executed with a header similar to the following simple script:
657 eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
659 #!/usr/local/bin/perl # just a comment really
661 print "Hello from perl!\n";
663 On these platforms, bear in mind that the EBCDIC character set may have
664 an effect on what happens with perl functions such as C<chr>, C<pack>,
665 C<print>, C<printf>, C<ord>, C<sort>, C<sprintf>, C<unpack>; as well as
666 bit-fiddling with ASCII constants using operators like C<^>, C<&> and
667 C<|>; not to mention dealing with socket interfaces to ASCII computers
670 Fortunately, most web servers for the mainframe will correctly translate
671 the C<\n> in the following statement to its ASCII equivalent (note that
672 C<\r> is the same under both ASCII and EBCDIC):
674 print "Content-type: text/html\r\n\r\n";
676 The value of C<$^O> on OS/390 is "os390".
678 Some simple tricks for determining if you are running on an EBCDIC
679 platform could include any of the following (perhaps all):
681 if ("\t" eq "\05") { print "EBCDIC may be spoken here!\n"; }
683 if (ord('A') == 193) { print "EBCDIC may be spoken here!\n"; }
685 if (chr(169) eq 'z') { print "EBCDIC may be spoken here!\n"; }
687 Note that one thing you may not want to rely on is the EBCDIC encoding
688 of punctuation characters since these may differ from code page to code page
689 (and once your module or script is rumoured to work with EBCDIC, folks will
690 want it to work with all EBCDIC character sets).
698 The perl-mvs@perl.org list is for discussion of porting issues as well as
699 general usage issues for all EBCDIC Perls. Send a message body of
700 "subscribe perl-mvs" to majordomo@perl.org.
702 =item AS/400 Perl information at C<http://as400.rochester.ibm.com>
709 As Acorns use ASCII with newlines (C<\n>) in text files as C<\012> like Unix
710 and Unix filename emulation is turned on by default, it is quite likely that
711 most simple scripts will work "out of the box". The native filing system is
712 modular, and individual filing systems are free to be case sensitive or
713 insensitive, usually case preserving. Some native filing systems have name
714 length limits which file and directory names are silently truncated to fit -
715 scripts should be aware that the standard disc filing system currently has
716 a name length limit of B<10> characters, with up to 77 items in a directory,
717 but other filing systems may not impose such limitations.
719 Native filenames are of the form
721 Filesystem#Special_Field::DiscName.$.Directory.Directory.File
725 Special_Field is not usually present, but may contain . and $ .
726 Filesystem =~ m|[A-Za-z0-9_]|
727 DsicName =~ m|[A-Za-z0-9_/]|
728 $ represents the root directory
729 . is the path separator
730 @ is the current directory (per filesystem but machine global)
731 ^ is the parent directory
732 Directory and File =~ m|[^\0- "\.\$\%\&:\@\\^\|\177]+|
734 The default filename translation is roughly C<tr|/.|./|;>
736 Note that C<"ADFS::HardDisc.$.File" ne 'ADFS::HardDisc.$.File'> and that
737 the second stage of $ interpolation in regular expressions will fall foul
738 of the C<$.> if scripts are not careful.
740 Logical paths specified by system variables containing comma separated
741 search lists are also allowed, hence C<System:Modules> is a valid filename,
742 and the filesystem will prefix C<Modules> with each section of C<System$Path>
743 until a name is made that points to an object on disc. Writing to a new
744 file C<System:Modules> would only be allowed if C<System$Path> contains a
745 single item list. The filesystem will also expand system variables in
746 filenames if enclosed in angle brackets, so C<E<lt>System$DirE<gt>.Modules>
747 would look for the file S<C<$ENV{'System$Dir'} . 'Modules'>>. The obvious
748 implication of this is that B<fully qualified filenames can start with C<E<lt>E<gt>>>
749 and should be protected when C<open> is used for input.
751 Because C<.> was in use as a directory separator and filenames could not
752 be assumed to be unique after 10 characters, Acorn implemented the C
753 compiler to strip the trailing C<.c> C<.h> C<.s> and C<.o> suffix from
754 filenames specified in source code and store the respective files in
755 subdirectories named after the suffix. Hence files are translated:
758 C:foo.h C:h.foo (logical path variable)
759 sys/os.h sys.h.os (C compiler groks Unix-speak)
760 10charname.c c.10charname
761 10charname.o o.10charname
762 11charname_.c c.11charname (assuming filesystem truncates at 10)
764 The Unix emulation library's translation of filenames to native assumes
765 that this sort of translation is required, and allows a user defined list of
766 known suffixes which it will transpose in this fashion. This may appear
767 transparent, but consider that with these rules C<foo/bar/baz.h> and
768 C<foo/bar/h/baz> both map to C<foo.bar.h.baz>, and that C<readdir> and
769 C<glob> cannot and do not attempt to emulate the reverse mapping. Other '.'s
770 in filenames are translated to '/'.
772 S<RISC OS> has "image files", files that behave as directories. For
773 example with suitable software this allows the contents of a zip file to
774 be treated as a directory at command line (and therefore script) level,
775 with full read-write random access. At present the perl port treats images
776 as directories: C<-d> returns true, C<-f> false, and C<unlink> checks to
777 ensure that recognised images are empty before deleting them. In theory
778 images should never trouble a script, but in practice they may do so if
779 the software to deal with an image file is loaded and registered while the
780 script is running, as suddenly "files" that it had cached information on
781 metamorphose into directories.
783 As implied above the environment accessed through C<%ENV> is global, and the
784 convention is that program specific environment variables are of the form
785 C<Program$Name>. Each filing system maintains a current directory, and
786 the current filing system's current directory is the B<global> current
787 directory. Consequently sociable scripts don't change the current directory
788 but rely on full pathnames, and scripts (and Makefiles) cannot assume that
789 they can spawn a child process which can change the current directory
790 without affecting its parent (and everyone else for that matter).
792 As native operating system filehandles are global and currently are allocated
793 down from 255, with 0 being a reserved value the Unix emulation library
794 emulates Unix filehandles. Consequently you can't rely on passing C<STDIN>
795 C<STDOUT> or C<STDERR> to your children. Run time libraries perform
796 command line processing to emulate Unix shell style C<>> redirection, but
797 the core operating system is written in assembler and has its own private,
798 obscure and somewhat broken convention. All this is further complicated by
799 the desire of users to express filenames of the form C<E<lt>Foo$DirE<gt>.Bar> on
800 the command line unquoted. (Oh yes, it's run time libraries interpreting the
801 quoting convention.) Hence C<``> command output capture has to perform
802 a guessing game as to how the command is going to interpret the command line
803 so that it can bodge it correctly to capture output. It assumes that a
804 string C<E<lt>[^E<lt>E<gt>]+\$[^E<lt>E<gt>]E<gt>> is a reference to an environment
805 variable, whereas anything else involving C<E<lt>> or C<E<gt>> is redirection,
806 and generally manages to be 99% right. Despite all this the problem remains
807 that scripts cannot rely on any Unix tools being available, or that any tools
808 found have Unix-like command line arguments.
810 Extensions and XS are in theory buildable by anyone using free tools. In
811 practice many don't as the Acorn platform is used to binary distribution.
812 MakeMaker does itself run, but no make currently copes with MakeMaker's
813 makefiles! Even if (when) this is fixed os that the lack of a Unix-like
814 shell can cause problems with makefile rules, especially lines of the form
815 C<cd sdbm && make all> and anything using quoting.
817 "S<RISC OS>" is the proper name for the operating system, but the value
818 in C<$^O> is "riscos" (because we don't like shouting).
831 Perl has been ported to a variety of platforms that do not fit into any of
832 the above categories. Some, such as AmigaOS, BeOS, QNX, and Plan 9, have
833 been well integrated into the standard Perl source code kit. You may need
834 to see the F<ports/> directory on CPAN for information, and possibly
835 binaries, for the likes of: aos, atari, lynxos, HP-MPE/iX, riscos,
836 Tandem Guardian, vos, I<etc.> (yes we know that some of these OSes may fall
837 under the Unix category but we are not a standards body.)
843 =item Atari, Guido Flohr's page C<http://stud.uni-sb.de/~gufl0000/>
845 =item HP 300 MPE/iX C<http://www.cccd.edu/~markb/perlix.html>
849 A free Perl 5 based PERL.NLM for Novell Netware is available from
850 C<http://www.novell.com/>
855 =head1 FUNCTION IMPLEMENTATIONS
857 Listed below are functions unimplemented or implemented differently on
858 various platforms. Following each description will be, in parentheses, a
859 list of platforms that the description applies to.
861 The list may very well be incomplete, or wrong in some places. When in
862 doubt, consult the platform-specific README files in the Perl source
863 distribution, and other documentation resources for a given port.
865 Be aware, moreover, that even among Unix-ish systems there are variations,
866 and not all functions listed here are necessarily available, though
869 For many functions, you can also query C<%Config>, exported by default
870 from C<Config.pm>. For example, to check if the platform has the C<lstat>
871 call, check C<$Config{'d_lstat'}>. See L<Config> for a full description
872 of available variables.
875 =head2 Alphabetical Listing of Perl Functions
885 C<-r>, C<-w>, and C<-x> have only a very limited meaning; directories
886 and applications are executable, and there are no uid/gid
887 considerations. C<-o> is not supported. (S<Mac OS>)
889 C<-r>, C<-w>, C<-x>, and C<-o> tell whether or not file is accessible,
890 which may not reflect UIC-based file protections. (VMS)
892 C<-s> returns the size of the data fork, not the total size of data fork
893 plus resource fork. (S<Mac OS>).
895 C<-s> by name on an open file will return the space reserved on disk,
896 rather than the current extent. C<-s> on an open filehandle returns the
897 current size. (S<RISC OS>)
899 C<-R>, C<-W>, C<-X>, C<-O> are indistinguishable from C<-r>, C<-w>,
900 C<-x>, C<-o>. (S<Mac OS>, Win32, VMS, S<RISC OS>)
902 C<-b>, C<-c>, C<-k>, C<-g>, C<-p>, C<-u>, C<-A> are not implemented.
905 C<-g>, C<-k>, C<-l>, C<-p>, C<-u>, C<-A> are not particularly meaningful.
906 (Win32, VMS, S<RISC OS>)
908 C<-d> is true if passed a device spec without an explicit directory.
911 C<-T> and C<-B> are implemented, but might misclassify Mac text files
912 with foreign characters; this is the case will all platforms, but
913 affects S<Mac OS> a lot. (S<Mac OS>)
915 C<-x> (or C<-X>) determine if a file ends in one of the executable
916 suffixes. C<-S> is meaningless. (Win32)
918 C<-x> (or C<-X>) determine if a file has an executable file type.
921 =item binmode FILEHANDLE
923 Meaningless. (S<Mac OS>, S<RISC OS>)
925 Reopens file and restores pointer; if function fails, underlying
926 filehandle may be closed, or pointer may be in a different position.
929 The value returned by C<tell> may be affected after the call, and
930 the filehandle may be flushed. (Win32)
934 Only limited meaning. Disabling/enabling write permission is mapped to
935 locking/unlocking the file. (S<Mac OS>)
937 Only good for changing "owner" read-write access, "group", and "other"
938 bits are meaningless. (Win32)
940 Only good for changing "owner" and "other" read-write access. (S<RISC OS>)
944 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
946 Does nothing, but won't fail. (Win32)
948 =item chroot FILENAME
952 Not implemented. (S<Mac OS>, Win32, VMS, Plan9, S<RISC OS>)
954 =item crypt PLAINTEXT,SALT
956 May not be available if library or source was not provided when building
961 Not implemented. (VMS, Plan9)
963 =item dbmopen HASH,DBNAME,MODE
965 Not implemented. (VMS, Plan9)
969 Not useful. (S<Mac OS>, S<RISC OS>)
971 Not implemented. (Win32)
973 Invokes VMS debugger. (VMS)
977 Not implemented. (S<Mac OS>)
979 =item fcntl FILEHANDLE,FUNCTION,SCALAR
981 Not implemented. (Win32, VMS)
983 =item flock FILEHANDLE,OPERATION
985 Not implemented (S<Mac OS>, VMS, S<RISC OS>).
987 Available only on Windows NT (not on Windows 95). (Win32)
991 Not implemented. (S<Mac OS>, Win32, AmigaOS, S<RISC OS>)
995 Not implemented. (S<Mac OS>, S<RISC OS>)
999 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1003 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1005 =item getpriority WHICH,WHO
1007 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1011 Not implemented. (S<Mac OS>, Win32)
1013 Not useful. (S<RISC OS>)
1017 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1019 =item getnetbyname NAME
1021 Not implemented. (S<Mac OS>, Win32, Plan9)
1025 Not implemented. (S<Mac OS>, Win32)
1027 Not useful. (S<RISC OS>)
1031 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1033 =item getnetbyaddr ADDR,ADDRTYPE
1035 Not implemented. (S<Mac OS>, Win32, Plan9)
1037 =item getprotobynumber NUMBER
1039 Not implemented. (S<Mac OS>)
1041 =item getservbyport PORT,PROTO
1043 Not implemented. (S<Mac OS>)
1047 Not implemented. (S<Mac OS>, Win32)
1051 Not implemented. (S<Mac OS>, Win32, VMS)
1055 Not implemented. (S<Mac OS>, Win32)
1059 Not implemented. (S<Mac OS>, Win32, Plan9)
1063 Not implemented. (S<Mac OS>, Win32, Plan9)
1067 Not implemented. (Win32, Plan9)
1071 Not implemented. (S<Mac OS>, Win32, S<RISC OS>)
1075 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1077 =item sethostent STAYOPEN
1079 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
1081 =item setnetent STAYOPEN
1083 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
1085 =item setprotoent STAYOPEN
1087 Not implemented. (S<Mac OS>, Win32, Plan9, S<RISC OS>)
1089 =item setservent STAYOPEN
1091 Not implemented. (Plan9, Win32, S<RISC OS>)
1095 Not implemented. (S<Mac OS>, Win32)
1099 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1103 Not implemented. (S<Mac OS>, Win32)
1107 Not implemented. (S<Mac OS>, Win32, Plan9)
1111 Not implemented. (S<Mac OS>, Win32, Plan9)
1115 Not implemented. (Plan9, Win32)
1117 =item getsockopt SOCKET,LEVEL,OPTNAME
1119 Not implemented. (S<Mac OS>, Plan9)
1125 Globbing built-in, but only C<*> and C<?> metacharacters are supported.
1128 Features depend on external perlglob.exe or perlglob.bat. May be overridden
1129 with something like File::DosGlob, which is recommended. (Win32)
1131 Globbing built-in, but only C<*> and C<?> metacharacters are supported.
1132 Globbing relies on operating system calls, which may return filenames in
1133 any order. As most filesystems are case insensitive even "sorted"
1134 filenames will not be in case sensitive order. (S<RISC OS>)
1136 =item ioctl FILEHANDLE,FUNCTION,SCALAR
1138 Not implemented. (VMS)
1140 Available only for socket handles, and it does what the ioctlsocket() call
1141 in the Winsock API does. (Win32)
1143 Available only for socket handles. (S<RISC OS>)
1147 Not implemented, hence not useful for taint checking. (S<Mac OS>, S<RISC OS>)
1149 Available only for process handles returned by the C<system(1, ...)> method of
1150 spawning a process. (Win32)
1152 =item link OLDFILE,NEWFILE
1154 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1156 =item lstat FILEHANDLE
1162 Not implemented. (VMS, S<RISC OS>)
1164 Return values may be bogus. (Win32)
1166 =item msgctl ID,CMD,ARG
1168 =item msgget KEY,FLAGS
1170 =item msgsnd ID,MSG,FLAGS
1172 =item msgrcv ID,VAR,SIZE,TYPE,FLAGS
1174 Not implemented. (S<Mac OS>, Win32, VMS, Plan9, S<RISC OS>)
1176 =item open FILEHANDLE,EXPR
1178 =item open FILEHANDLE
1180 The C<|> variants are only supported if ToolServer is installed.
1183 open to C<|-> and C<-|> are unsupported. (S<Mac OS>, Win32, S<RISC OS>)
1185 =item pipe READHANDLE,WRITEHANDLE
1187 Not implemented. (S<Mac OS>)
1193 Not implemented. (Win32, VMS, S<RISC OS>)
1195 =item select RBITS,WBITS,EBITS,TIMEOUT
1197 Only implemented on sockets. (Win32)
1199 Only reliable on sockets. (S<RISC OS>)
1201 =item semctl ID,SEMNUM,CMD,ARG
1203 =item semget KEY,NSEMS,FLAGS
1205 =item semop KEY,OPSTRING
1207 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1209 =item setpgrp PID,PGRP
1211 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1213 =item setpriority WHICH,WHO,PRIORITY
1215 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1217 =item setsockopt SOCKET,LEVEL,OPTNAME,OPTVAL
1219 Not implemented. (S<Mac OS>, Plan9)
1221 =item shmctl ID,CMD,ARG
1223 =item shmget KEY,SIZE,FLAGS
1225 =item shmread ID,VAR,POS,SIZE
1227 =item shmwrite ID,STRING,POS,SIZE
1229 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1231 =item socketpair SOCKET1,SOCKET2,DOMAIN,TYPE,PROTOCOL
1233 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1235 =item stat FILEHANDLE
1241 mtime and atime are the same thing, and ctime is creation time instead of
1242 inode change time. (S<Mac OS>)
1244 device and inode are not meaningful. (Win32)
1246 device and inode are not necessarily reliable. (VMS)
1248 mtime, atime and ctime all return the last modification time. Device and
1249 inode are not necessarily reliable. (S<RISC OS>)
1251 =item symlink OLDFILE,NEWFILE
1253 Not implemented. (Win32, VMS, S<RISC OS>)
1257 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1259 =item sysopen FILEHANDLE,FILENAME,MODE,PERMS
1261 The traditional "0", "1", and "2" MODEs are implemented with different
1262 numeric values on some systems. The flags exported by C<Fcntl> should work
1263 everywhere though. (S<Mac OS>, OS/390)
1267 Only implemented if ToolServer is installed. (S<Mac OS>)
1269 As an optimization, may not call the command shell specified in
1270 C<$ENV{PERL5SHELL}>. C<system(1, @args)> spawns an external
1271 process and immediately returns its process designator, without
1272 waiting for it to terminate. Return value may be used subsequently
1273 in C<wait> or C<waitpid>. (Win32)
1275 There is no shell to process metacharacters, and the native standard is
1276 to pass a command line terminated by "\n" "\r" or "\0" to the spawned
1277 program. Redirection such as C<E<gt> foo> is performed (if at all) by
1278 the run time library of the spawned program. C<system> I<list> will call
1279 the Unix emulation library's C<exec> emulation, which attempts to provide
1280 emulation of the stdin, stdout, stderr in force in the parent, providing
1281 the child program uses a compatible version of the emulation library.
1282 I<scalar> will call the native command line direct and no such emulation
1283 of a child Unix program will exists. Mileage B<will> vary. (S<RISC OS>)
1287 Only the first entry returned is nonzero. (S<Mac OS>)
1289 "cumulative" times will be bogus. On anything other than Windows NT,
1290 "system" time will be bogus, and "user" time is actually the time
1291 returned by the clock() function in the C runtime library. (Win32)
1293 Not useful. (S<RISC OS>)
1295 =item truncate FILEHANDLE,LENGTH
1297 =item truncate EXPR,LENGTH
1299 Not implemented. (VMS)
1305 Returns undef where unavailable, as of version 5.005.
1309 Only the modification time is updated. (S<Mac OS>, VMS, S<RISC OS>)
1311 May not behave as expected. (Win32)
1315 =item waitpid PID,FLAGS
1317 Not implemented. (S<Mac OS>)
1319 Can only be applied to process handles returned for processes spawned
1320 using C<system(1, ...)>. (Win32)
1322 Not useful. (S<RISC OS>)
1330 =item 1.32, 05 August 1998
1332 Integrate more minor changes.
1334 =item 1.30, 03 August 1998
1336 Major update for RISC OS, other minor changes.
1338 =item 1.23, 10 July 1998
1340 First public release with perl5.005.
1344 =head1 AUTHORS / CONTRIBUTORS
1346 Abigail E<lt>abigail@fnx.comE<gt>,
1347 Charles Bailey E<lt>bailey@genetics.upenn.eduE<gt>,
1348 Graham Barr E<lt>gbarr@pobox.comE<gt>,
1349 Tom Christiansen E<lt>tchrist@perl.comE<gt>,
1350 Nicholas Clark E<lt>Nicholas.Clark@liverpool.ac.ukE<gt>,
1351 Andy Dougherty E<lt>doughera@lafcol.lafayette.eduE<gt>,
1352 Dominic Dunlop E<lt>domo@vo.luE<gt>,
1353 M.J.T. Guy E<lt>mjtg@cus.cam.ac.ukE<gt>,
1354 Luther Huffman E<lt>lutherh@stratcom.comE<gt>,
1355 Nick Ing-Simmons E<lt>nick@ni-s.u-net.comE<gt>,
1356 Andreas J. Koenig E<lt>koenig@kulturbox.deE<gt>,
1357 Andrew M. Langmead E<lt>aml@world.std.comE<gt>,
1358 Paul Moore E<lt>Paul.Moore@uk.origin-it.comE<gt>,
1359 Chris Nandor E<lt>pudge@pobox.comE<gt>,
1360 Matthias Neercher E<lt>neeri@iis.ee.ethz.chE<gt>,
1361 Gary Ng E<lt>71564.1743@CompuServe.COME<gt>,
1362 Tom Phoenix E<lt>rootbeer@teleport.comE<gt>,
1363 Peter Prymmer E<lt>pvhp@forte.comE<gt>,
1364 Hugo van der Sanden E<lt>h.sanden@elsevier.nlE<gt>,
1365 Gurusamy Sarathy E<lt>gsar@umich.eduE<gt>,
1366 Paul J. Schinder E<lt>schinder@pobox.comE<gt>,
1367 Dan Sugalski E<lt>sugalskd@ous.eduE<gt>,
1368 Nathan Torkington E<lt>gnat@frii.comE<gt>.
1370 This document is maintained by Chris Nandor.
1374 Version 1.32, last modified 05 August 1998.