[ Index ]

title

Body

[close]

## /se3-unattended/var/se3/unattended/install/linuxaux/opt/perl/lib/5.10.0/pod/ -> perlport.pod (source)

   1  =head1 NAME
2
3  perlport - Writing portable Perl
4
6
7  Perl runs on numerous operating systems.  While most of them share
8  much in common, they also have their own unique features.
9
10  This document is meant to help you to find out what constitutes portable
11  Perl code.  That way once you make a decision to write portably,
12  you know where the lines are drawn, and you can stay within them.
13
14  There is a tradeoff between taking full advantage of one particular
15  type of computer and taking advantage of a full range of them.
16  Naturally, as you broaden your range and become more diverse, the
17  common factors drop, and you are left with an increasingly smaller
18  area of common ground in which you can operate to accomplish a
19  particular task.  Thus, when you begin attacking a problem, it is
20  important to consider under which part of the tradeoff curve you
21  want to operate.  Specifically, you must decide whether it is
22  important that the task that you are coding have the full generality
23  of being portable, or whether to just get the job done right now.
24  This is the hardest choice to be made.  The rest is easy, because
25  Perl provides many choices, whichever way you want to approach your
26  problem.
27
28  Looking at it another way, writing portable code is usually about
29  willfully limiting your available choices.  Naturally, it takes
30  discipline and sacrifice to do that.  The product of portability
31  and convenience may be a constant.  You have been warned.
32
33  Be aware of two important points:
34
35  =over 4
36
37  =item Not all Perl programs have to be portable
38
39  There is no reason you should not use Perl as a language to glue Unix
40  tools together, or to prototype a Macintosh application, or to manage the
41  Windows registry.  If it makes no sense to aim for portability for one
42  reason or another in a given program, then don't bother.
43
44  =item Nearly all of Perl already I<is> portable
45
46  Don't be fooled into thinking that it is hard to create portable Perl
47  code.  It isn't.  Perl tries its level-best to bridge the gaps between
48  what's available on different platforms, and all the means available to
49  use those features.  Thus almost all Perl code runs on any machine
50  without modification.  But there are some significant issues in
51  writing portable code, and this document is entirely about those issues.
52
53  =back
54
55  Here's the general rule: When you approach a task commonly done
56  using a whole range of platforms, think about writing portable
57  code.  That way, you don't sacrifice much by way of the implementation
58  choices you can avail yourself of, and at the same time you can give
59  your users lots of platform choices.  On the other hand, when you have to
60  take advantage of some unique feature of a particular platform, as is
61  often the case with systems programming (whether for Unix, Windows,
62  S<Mac OS>, VMS, etc.), consider writing platform-specific code.
63
64  When the code will run on only two or three operating systems, you
65  may need to consider only the differences of those particular systems.
66  The important thing is to decide where the code will run and to be
68
69  The material below is separated into three main sections: main issues of
70  portability (L<"ISSUES">), platform-specific issues (L<"PLATFORMS">), and
71  built-in perl functions that behave differently on various ports
72  (L<"FUNCTION IMPLEMENTATIONS">).
73
74  This information should not be considered complete; it includes possibly
75  transient information about idiosyncrasies of some of the ports, almost
76  all of which are in a state of constant evolution.  Thus, this material
77  should be considered a perpetual work in progress
78  (C<< <IMG SRC="yellow_sign.gif" ALT="Under Construction"> >>).
79
81
83
84  In most operating systems, lines in files are terminated by newlines.
85  Just what is used as a newline may vary from OS to OS.  Unix
86  traditionally uses C<\012>, one type of DOSish I/O uses C<\015\012>,
87  and S<Mac OS> uses C<\015>.
88
89  Perl uses C<\n> to represent the "logical" newline, where what is
90  logical may depend on the platform in use.  In MacPerl, C<\n> always
91  means C<\015>.  In DOSish perls, C<\n> usually means C<\012>, but
92  when accessing a file in "text" mode, STDIO translates it to (or
93  from) C<\015\012>, depending on whether you're reading or writing.
94  Unix does the same thing on ttys in canonical mode.  C<\015\012>
95  is commonly referred to as CRLF.
96
97  To trim trailing newlines from text lines use chomp().  With default
98  settings that function looks for a trailing C<\n> character and thus
99  trims in a portable way.
100
101  When dealing with binary files (or text files in binary mode) be sure
102  to explicitly set $/ to the appropriate value for your file format 103 before using chomp(). 104 105 Because of the "text" mode translation, DOSish perls have limitations 106 in using C<seek> and C<tell> on a file accessed in "text" mode. 107 Stick to C<seek>-ing to locations you got from C<tell> (and no 108 others), and you are usually free to use C<seek> and C<tell> even 109 in "text" mode. Using C<seek> or C<tell> or other file operations 110 may be non-portable. If you use C<binmode> on a file, however, you 111 can usually C<seek> and C<tell> with arbitrary values in safety. 112 113 A common misconception in socket programming is that C<\n> eq C<\012> 114 everywhere. When using protocols such as common Internet protocols, 115 C<\012> and C<\015> are called for specifically, and the values of 116 the logical C<\n> and C<\r> (carriage return) are not reliable. 117 118 print SOCKET "Hi there, client!\r\n"; # WRONG 119 print SOCKET "Hi there, client!\015\012"; # RIGHT 120 121 However, using C<\015\012> (or C<\cM\cJ>, or C<\x0D\x0A>) can be tedious 122 and unsightly, as well as confusing to those maintaining the code. As 123 such, the Socket module supplies the Right Thing for those who want it. 124 125 use Socket qw(:DEFAULT :crlf); 126 print SOCKET "Hi there, client!$CRLF"      # RIGHT
127
128  When reading from a socket, remember that the default input record
129  separator C<$/> is C<\n>, but robust socket code will recognize as 130 either C<\012> or C<\015\012> as end of line: 131 132 while (<SOCKET>) { 133 # ... 134 } 135 136 Because both CRLF and LF end in LF, the input record separator can 137 be set to LF and any CR stripped later. Better to write: 138 139 use Socket qw(:DEFAULT :crlf); 140 local($/) = LF;      # not needed if $/ is already \012 141 142 while (<SOCKET>) { 143 s/$CR?$LF/\n/; # not sure if socket uses LF or CRLF, OK 144 # s/\015?\012/\n/; # same thing 145 } 146 147 This example is preferred over the previous one--even for Unix 148 platforms--because now any C<\015>'s (C<\cM>'s) are stripped out 149 (and there was much rejoicing). 150 151 Similarly, functions that return text data--such as a function that 152 fetches a web page--should sometimes translate newlines before 153 returning the data, if they've not yet been translated to the local 154 newline representation. A single line of code will often suffice: 155 156$data =~ s/\015?\012/\n/g;
157      return $data; 158 159 Some of this may be confusing. Here's a handy reference to the ASCII CR 160 and LF characters. You can print it out and stick it in your wallet. 161 162 LF eq \012 eq \x0A eq \cJ eq chr(10) eq ASCII 10 163 CR eq \015 eq \x0D eq \cM eq chr(13) eq ASCII 13 164 165 | Unix | DOS | Mac | 166 --------------------------- 167 \n | LF | LF | CR | 168 \r | CR | CR | LF | 169 \n * | LF | CRLF | CR | 170 \r * | CR | CR | LF | 171 --------------------------- 172 * text-mode STDIO 173 174 The Unix column assumes that you are not accessing a serial line 175 (like a tty) in canonical mode. If you are, then CR on input becomes 176 "\n", and "\n" on output becomes CRLF. 177 178 These are just the most common definitions of C<\n> and C<\r> in Perl. 179 There may well be others. For example, on an EBCDIC implementation 180 such as z/OS (OS/390) or OS/400 (using the ILE, the PASE is ASCII-based) 181 the above material is similar to "Unix" but the code numbers change: 182 183 LF eq \025 eq \x15 eq \cU eq chr(21) eq CP-1047 21 184 LF eq \045 eq \x25 eq chr(37) eq CP-0037 37 185 CR eq \015 eq \x0D eq \cM eq chr(13) eq CP-1047 13 186 CR eq \015 eq \x0D eq \cM eq chr(13) eq CP-0037 13 187 188 | z/OS | OS/400 | 189 ---------------------- 190 \n | LF | LF | 191 \r | CR | CR | 192 \n * | LF | LF | 193 \r * | CR | CR | 194 ---------------------- 195 * text-mode STDIO 196 197 =head2 Numbers endianness and Width 198 199 Different CPUs store integers and floating point numbers in different 200 orders (called I<endianness>) and widths (32-bit and 64-bit being the 201 most common today). This affects your programs when they attempt to transfer 202 numbers in binary format from one CPU architecture to another, 203 usually either "live" via network connection, or by storing the 204 numbers to secondary storage such as a disk file or tape. 205 206 Conflicting storage orders make utter mess out of the numbers. If a 207 little-endian host (Intel, VAX) stores 0x12345678 (305419896 in 208 decimal), a big-endian host (Motorola, Sparc, PA) reads it as 209 0x78563412 (2018915346 in decimal). Alpha and MIPS can be either: 210 Digital/Compaq used/uses them in little-endian mode; SGI/Cray uses 211 them in big-endian mode. To avoid this problem in network (socket) 212 connections use the C<pack> and C<unpack> formats C<n> and C<N>, the 213 "network" orders. These are guaranteed to be portable. 214 215 As of perl 5.9.2, you can also use the C<E<gt>> and C<E<lt>> modifiers 216 to force big- or little-endian byte-order. This is useful if you want 217 to store signed integers or 64-bit integers, for example. 218 219 You can explore the endianness of your platform by unpacking a 220 data structure packed in native format such as: 221 222 print unpack("h*", pack("s2", 1, 2)), "\n"; 223 # '10002000' on e.g. Intel x86 or Alpha 21064 in little-endian mode 224 # '00100020' on e.g. Motorola 68040 225 226 If you need to distinguish between endian architectures you could use 227 either of the variables set like so: 228 229$is_big_endian   = unpack("h*", pack("s", 1)) =~ /01/;
230      $is_little_endian = unpack("h*", pack("s", 1)) =~ /^1/; 231 232 Differing widths can cause truncation even between platforms of equal 233 endianness. The platform of shorter width loses the upper parts of the 234 number. There is no good solution for this problem except to avoid 235 transferring or storing raw binary numbers. 236 237 One can circumnavigate both these problems in two ways. Either 238 transfer and store numbers always in text format, instead of raw 239 binary, or else consider using modules like Data::Dumper (included in 240 the standard distribution as of Perl 5.005) and Storable (included as 241 of perl 5.8). Keeping all data as text significantly simplifies matters. 242 243 The v-strings are portable only up to v2147483647 (0x7FFFFFFF), that's 244 how far EBCDIC, or more precisely UTF-EBCDIC will go. 245 246 =head2 Files and Filesystems 247 248 Most platforms these days structure files in a hierarchical fashion. 249 So, it is reasonably safe to assume that all platforms support the 250 notion of a "path" to uniquely identify a file on the system. How 251 that path is really written, though, differs considerably. 252 253 Although similar, file path specifications differ between Unix, 254 Windows, S<Mac OS>, OS/2, VMS, VOS, S<RISC OS>, and probably others. 255 Unix, for example, is one of the few OSes that has the elegant idea 256 of a single root directory. 257 258 DOS, OS/2, VMS, VOS, and Windows can work similarly to Unix with C</> 259 as path separator, or in their own idiosyncratic ways (such as having 260 several root directories and various "unrooted" device files such NIL: 261 and LPT:). 262 263 S<Mac OS> uses C<:> as a path separator instead of C</>. 264 265 The filesystem may support neither hard links (C<link>) nor 266 symbolic links (C<symlink>, C<readlink>, C<lstat>). 267 268 The filesystem may support neither access timestamp nor change 269 timestamp (meaning that about the only portable timestamp is the 270 modification timestamp), or one second granularity of any timestamps 271 (e.g. the FAT filesystem limits the time granularity to two seconds). 272 273 The "inode change timestamp" (the C<-C> filetest) may really be the 274 "creation timestamp" (which it is not in UNIX). 275 276 VOS perl can emulate Unix filenames with C</> as path separator. The 277 native pathname characters greater-than, less-than, number-sign, and 278 percent-sign are always accepted. 279 280 S<RISC OS> perl can emulate Unix filenames with C</> as path 281 separator, or go native and use C<.> for path separator and C<:> to 282 signal filesystems and disk names. 283 284 Don't assume UNIX filesystem access semantics: that read, write, 285 and execute are all the permissions there are, and even if they exist, 286 that their semantics (for example what do r, w, and x mean on 287 a directory) are the UNIX ones. The various UNIX/POSIX compatibility 288 layers usually try to make interfaces like chmod() work, but sometimes 289 there simply is no good mapping. 290 291 If all this is intimidating, have no (well, maybe only a little) 292 fear. There are modules that can help. The File::Spec modules 293 provide methods to do the Right Thing on whatever platform happens 294 to be running the program. 295 296 use File::Spec::Functions; 297 chdir(updir()); # go up one directory 298$file = catfile(curdir(), 'temp', 'file.txt');
299      # on Unix and Win32, './temp/file.txt'
300      # on Mac OS, ':temp:file.txt'
301      # on VMS, '[.temp]file.txt'
302
303  File::Spec is available in the standard distribution as of version
304  5.004_05.  File::Spec::Functions is only in File::Spec 0.7 and later,
305  and some versions of perl come with version 0.6.  If File::Spec
306  is not updated to 0.7 or later, you must use the object-oriented
307  interface from File::Spec (or upgrade File::Spec).
308
309  In general, production code should not have file paths hardcoded.
310  Making them user-supplied or read from a configuration file is
311  better, keeping in mind that file path syntax varies on different
312  machines.
313
314  This is especially noticeable in scripts like Makefiles and test suites,
315  which often assume C</> as a path separator for subdirectories.
316
317  Also of use is File::Basename from the standard distribution, which
318  splits a pathname into pieces (base filename, full path to directory,
319  and file suffix).
320
321  Even when on a single platform (if you can call Unix a single platform),
322  remember not to count on the existence or the contents of particular
323  system-specific files or directories, like F</etc/passwd>,
324  F</etc/sendmail.conf>, F</etc/resolv.conf>, or even F</tmp/>.  For
325  example, F</etc/passwd> may exist but not contain the encrypted
326  passwords, because the system is using some form of enhanced security.
327  Or it may not contain all the accounts, because the system is using NIS.
328  If code does need to rely on such a file, include a description of the
329  file and its format in the code's documentation, then make it easy for
330  the user to override the default location of the file.
331
332  Don't assume a text file will end with a newline.  They should,
333  but people forget.
334
335  Do not have two files or directories of the same name with different
336  case, like F<test.pl> and F<Test.pl>, as many platforms have
337  case-insensitive (or at least case-forgiving) filenames.  Also, try
338  not to have non-word characters (except for C<.>) in the names, and
339  keep them to the 8.3 convention, for maximum portability, onerous a
340  burden though this may appear.
341
342  Likewise, when using the AutoSplit module, try to keep your functions to
343  8.3 naming and case-insensitive conventions; or, at the least,
344  make it so the resulting files have a unique (case-insensitively)
345  first 8 characters.
346
347  Whitespace in filenames is tolerated on most systems, but not all,
348  and even on systems where it might be tolerated, some utilities
349  might become confused by such whitespace.
350
351  Many systems (DOS, VMS ODS-2) cannot have more than one C<.> in their
352  filenames.
353
354  Don't assume C<< > >> won't be the first character of a filename.
355  Always use C<< < >> explicitly to open a file for reading, or even
356  better, use the three-arg version of open, unless you want the user to
357  be able to specify a pipe open.
358
359      open(FILE, '<', $existing_file) or die$!;
360
361  If filenames might use strange characters, it is safest to open it
362  with C<sysopen> instead of C<open>.  C<open> is magic and can
363  translate characters like C<< > >>, C<< < >>, and C<|>, which may
364  be the wrong thing to do.  (Sometimes, though, it's the right thing.)
365  Three-arg open can also help protect against this translation in cases
366  where it is undesirable.
367
368  Don't use C<:> as a part of a filename since many systems use that for
369  their own semantics (Mac OS Classic for separating pathname components,
370  many networking schemes and utilities for separating the nodename and
371  the pathname, and so on).  For the same reasons, avoid C<@>, C<;> and
372  C<|>.
373
374  Don't assume that in pathnames you can collapse two leading slashes
375  C<//> into one: some networking and clustering filesystems have special
376  semantics for that.  Let the operating system to sort it out.
377
378  The I<portable filename characters> as defined by ANSI C are
379
380   a b c d e f g h i j k l m n o p q r t u v w x y z
381   A B C D E F G H I J K L M N O P Q R T U V W X Y Z
382   0 1 2 3 4 5 6 7 8 9
383   . _ -
384
385  and the "-" shouldn't be the first character.  If you want to be
386  hypercorrect, stay case-insensitive and within the 8.3 naming
387  convention (all the files and directories have to be unique within one
388  directory if their names are lowercased and truncated to eight
389  characters before the C<.>, if any, and to three characters after the
390  C<.>, if any).  (And do not use C<.>s in directory names.)
391
393
394  Not all platforms provide a command line.  These are usually platforms
395  that rely primarily on a Graphical User Interface (GUI) for user
396  interaction.  A program requiring a command line interface might
397  not work everywhere.  This is probably for the user of the program
398  to deal with, so don't stay up late worrying about it.
399
400  Some platforms can't delete or rename files held open by the system,
401  this limitation may also apply to changing filesystem metainformation
402  like file permissions or owners.  Remember to C<close> files when you
403  are done with them.  Don't C<unlink> or C<rename> an open file.  Don't
404  C<tie> or C<open> a file already tied or opened; C<untie> or C<close>
405  it first.
406
407  Don't open the same file more than once at a time for writing, as some
408  operating systems put mandatory locks on such files.
409
410  Don't assume that write/modify permission on a directory gives the
411  right to add or delete files/directories in that directory.  That is
412  filesystem specific: in some filesystems you need write/modify
413  permission also (or even just) in the file/directory itself.  In some
414  filesystems (AFS, DFS) the permission to add/delete directory entries
415  is a completely separate permission.
416
417  Don't assume that a single C<unlink> completely gets rid of the file:
418  some filesystems (most notably the ones in VMS) have versioned
419  filesystems, and unlink() removes only the most recent one (it doesn't
420  remove all the versions because by default the native tools on those
422  idiom to remove all the versions of a file is
423
425
426  This will terminate if the file is undeleteable for some reason
427  (protected, not there, and so on).
428
429  Don't count on a specific environment variable existing in C<%ENV>.
430  Don't count on C<%ENV> entries being case-sensitive, or even
431  case-preserving.  Don't try to clear %ENV by saying C<%ENV = ();>, or,
432  if you really have to, make it conditional on C<$^O ne 'VMS'> since in 433 VMS the C<%ENV> table is much more than a per-process key-value string 434 table. 435 436 On VMS, some entries in the %ENV hash are dynamically created when 437 their key is used on a read if they did not previously exist. The 438 values for C<$ENV{HOME}>, C<$ENV{TERM}>, C<$ENV{HOME}>, and C<$ENV{USER}>, 439 are known to be dynamically generated. The specific names that are 440 dynamically generated may vary with the version of the C library on VMS, 441 and more may exist than is documented. 442 443 On VMS by default, changes to the %ENV hash are persistent after the process 444 exits. This can cause unintended issues. 445 446 Don't count on signals or C<%SIG> for anything. 447 448 Don't count on filename globbing. Use C<opendir>, C<readdir>, and 449 C<closedir> instead. 450 451 Don't count on per-program environment variables, or per-program current 452 directories. 453 454 Don't count on specific values of C<$!>, neither numeric nor
455  especially the strings values-- users may switch their locales causing
456  error messages to be translated into their languages.  If you can
457  trust a POSIXish environment, you can portably use the symbols defined
458  by the Errno module, like ENOENT.  And don't trust on the values of C<$!> 459 at all except immediately after a failed system call. 460 461 =head2 Command names versus file pathnames 462 463 Don't assume that the name used to invoke a command or program with 464 C<system> or C<exec> can also be used to test for the existence of the 465 file that holds the executable code for that command or program. 466 First, many systems have "internal" commands that are built-in to the 467 shell or OS and while these commands can be invoked, there is no 468 corresponding file. Second, some operating systems (e.g., Cygwin, 469 DJGPP, OS/2, and VOS) have required suffixes for executable files; 470 these suffixes are generally permitted on the command name but are not 471 required. Thus, a command like "perl" might exist in a file named 472 "perl", "perl.exe", or "perl.pm", depending on the operating system. 473 The variable "_exe" in the Config module holds the executable suffix, 474 if any. Third, the VMS port carefully sets up$^X and
475  $Config{perlpath} so that no further processing is required. This is 476 just as well, because the matching regular expression used below would 477 then have to deal with a possible trailing version number in the VMS 478 file name. 479 480 To convert$^X to a file pathname, taking account of the requirements
481  of the various operating system possibilities, say:
482
483    use Config;
484    $thisperl =$^X;
485    if ($^O ne 'VMS') 486 {$thisperl .= $Config{_exe} unless$thisperl =~ m/$Config{_exe}$/i;}
487
488  To convert $Config{perlpath} to a file pathname, say: 489 490 use Config; 491$thisperl = $Config{perlpath}; 492 if ($^O ne 'VMS')
493       {$thisperl .=$Config{_exe} unless $thisperl =~ m/$Config{_exe}$/i;} 494 495 =head2 Networking 496 497 Don't assume that you can reach the public Internet. 498 499 Don't assume that there is only one way to get through firewalls 500 to the public Internet. 501 502 Don't assume that you can reach outside world through any other port 503 than 80, or some web proxy. ftp is blocked by many firewalls. 504 505 Don't assume that you can send email by connecting to the local SMTP port. 506 507 Don't assume that you can reach yourself or any node by the name 508 'localhost'. The same goes for '127.0.0.1'. You will have to try both. 509 510 Don't assume that the host has only one network card, or that it 511 can't bind to many virtual IP addresses. 512 513 Don't assume a particular network device name. 514 515 Don't assume a particular set of ioctl()s will work. 516 517 Don't assume that you can ping hosts and get replies. 518 519 Don't assume that any particular port (service) will respond. 520 521 Don't assume that Sys::Hostname (or any other API or command) 522 returns either a fully qualified hostname or a non-qualified hostname: 523 it all depends on how the system had been configured. Also remember 524 things like DHCP and NAT-- the hostname you get back might not be very 525 useful. 526 527 All the above "don't":s may look daunting, and they are -- but the key 528 is to degrade gracefully if one cannot reach the particular network 529 service one wants. Croaking or hanging do not look very professional. 530 531 =head2 Interprocess Communication (IPC) 532 533 In general, don't directly access the system in code meant to be 534 portable. That means, no C<system>, C<exec>, C<fork>, C<pipe>, 535 C<>, C<qx//>, C<open> with a C<|>, nor any of the other things 536 that makes being a perl hacker worth being. 537 538 Commands that launch external processes are generally supported on 539 most platforms (though many of them do not support any type of 540 forking). The problem with using them arises from what you invoke 541 them on. External tools are often named differently on different 542 platforms, may not be available in the same location, might accept 543 different arguments, can behave differently, and often present their 544 results in a platform-dependent way. Thus, you should seldom depend 545 on them to produce consistent results. (Then again, if you're calling 546 I<netstat -a>, you probably don't expect it to run on both Unix and CP/M.) 547 548 One especially common bit of Perl code is opening a pipe to B<sendmail>: 549 550 open(MAIL, '|/usr/lib/sendmail -t') 551 or die "cannot fork sendmail:$!";
552
553  This is fine for systems programming when sendmail is known to be
554  available.  But it is not fine for many non-Unix systems, and even
555  some Unix systems that may not have sendmail installed.  If a portable
556  solution is needed, see the various distributions on CPAN that deal
557  with it.  Mail::Mailer and Mail::Send in the MailTools distribution are
558  commonly used, and provide several mailing methods, including mail,
559  sendmail, and direct SMTP (via Net::SMTP) if a mail transfer agent is
560  not available.  Mail::Sendmail is a standalone module that provides
561  simple, platform-independent mailing.
562
563  The Unix System V IPC (C<msg*(), sem*(), shm*()>) is not available
564  even on all Unix platforms.
565
566  Do not use either the bare result of C<pack("N", 10, 20, 30, 40)> or
567  bare v-strings (such as C<v10.20.30.40>) to represent IPv4 addresses:
568  both forms just pack the four bytes into network order.  That this
569  would be equal to the C language C<in_addr> struct (which is what the
570  socket code internally uses) is not guaranteed.  To be portable use
571  the routines of the Socket extension, such as C<inet_aton()>,
573
574  The rule of thumb for portable code is: Do it all in portable Perl, or
575  use a module (that may internally implement it with platform-specific
576  code, but expose a common interface).
577
579
580  XS code can usually be made to work with any platform, but dependent
582  portable, or the XS code itself might be platform-specific, just as Perl
583  code might be.  If the libraries and headers are portable, then it is
584  normally reasonable to make sure the XS code is portable, too.
585
586  A different type of portability issue arises when writing XS code:
587  availability of a C compiler on the end-user's system.  C brings
588  with it its own portability issues, and writing XS code will expose
589  you to some of those.  Writing purely in Perl is an easier way to
590  achieve portability.
591
593
594  In general, the standard modules work across platforms.  Notable
595  exceptions are the CPAN module (which currently makes connections to external
596  programs that may not be available), platform-specific modules (like
597  ExtUtils::MM_VMS), and DBM modules.
598
599  There is no one DBM module available on all platforms.
600  SDBM_File and the others are generally available on all Unix and DOSish
601  ports, but not in MacPerl, where only NBDM_File and DB_File are
602  available.
603
604  The good news is that at least some DBM module should be available, and
605  AnyDBM_File will use whichever module it can find.  Of course, then
606  the code needs to be fairly strict, dropping to the greatest common
607  factor (e.g., not exceeding 1K for each record), so that it will
608  work with any DBM module.  See L<AnyDBM_File> for more details.
609
611
612  The system's notion of time of day and calendar date is controlled in
613  widely different ways.  Don't assume the timezone is stored in C<$ENV{TZ}>, 614 and even if it is, don't assume that you can control the timezone through 615 that variable. Don't assume anything about the three-letter timezone 616 abbreviations (for example that MST would be the Mountain Standard Time, 617 it's been known to stand for Moscow Standard Time). If you need to 618 use timezones, express them in some unambiguous format like the 619 exact number of minutes offset from UTC, or the POSIX timezone 620 format. 621 622 Don't assume that the epoch starts at 00:00:00, January 1, 1970, 623 because that is OS- and implementation-specific. It is better to 624 store a date in an unambiguous representation. The ISO 8601 standard 625 defines YYYY-MM-DD as the date format, or YYYY-MM-DDTHH-MM-SS 626 (that's a literal "T" separating the date from the time). 627 Please do use the ISO 8601 instead of making us to guess what 628 date 02/03/04 might be. ISO 8601 even sorts nicely as-is. 629 A text representation (like "1987-12-18") can be easily converted 630 into an OS-specific value using a module like Date::Parse. 631 An array of values, such as those returned by C<localtime>, can be 632 converted to an OS-specific representation using Time::Local. 633 634 When calculating specific times, such as for tests in time or date modules, 635 it may be appropriate to calculate an offset for the epoch. 636 637 require Time::Local; 638$offset = Time::Local::timegm(0, 0, 0, 1, 0, 70);
639
640  The value for C<$offset> in Unix will be C<0>, but in Mac OS will be 641 some large number. C<$offset> can then be added to a Unix time value
642  to get what should be the proper value on any system.
643
644  On Windows (at least), you shouldn't pass a negative value to C<gmtime> or
645  C<localtime>.
646
647  =head2 Character sets and character encoding
648
649  Assume very little about character sets.
650
651  Assume nothing about numerical values (C<ord>, C<chr>) of characters.
652  Do not use explicit code point ranges (like \xHH-\xHH); use for
653  example symbolic character classes like C<[:print:]>.
654
655  Do not assume that the alphabetic characters are encoded contiguously
656  (in the numeric sense).  There may be gaps.
657
658  Do not assume anything about the ordering of the characters.
659  The lowercase letters may come before or after the uppercase letters;
660  the lowercase and uppercase may be interlaced so that both "a" and "A"
661  come before "b"; the accented and other international characters may
662  be interlaced so that E<auml> comes before "b".
663
665
666  If you may assume POSIX (a rather large assumption), you may read
667  more about the POSIX locale system from L<perllocale>.  The locale
668  system at least attempts to make things a little bit more portable,
669  or at least more convenient and native-friendly for non-English
670  users.  The system affects character sets and encoding, and date
671  and time formatting--amongst other things.
672
673  If you really want to be international, you should consider Unicode.
675
676  If you want to use non-ASCII bytes (outside the bytes 0x00..0x7f) in
677  the "source code" of your code, to be portable you have to be explicit
678  about what bytes they are.  Someone might for example be using your
679  code under a UTF-8 locale, in which case random native bytes might be
680  illegal ("Malformed UTF-8 ...")  This means that for example embedding
681  ISO 8859-1 bytes beyond 0x7f into your strings might cause trouble
682  later.  If the bytes are native 8-bit bytes, you can use the C<bytes>
683  pragma.  If the bytes are in a string (regular expression being a
684  curious string), you can often also use the C<\xHH> notation instead
685  of embedding the bytes as-is.  (If you want to write your code in UTF-8,
686  you can use the C<utf8>.) The C<bytes> and C<utf8> pragmata are
687  available since Perl 5.6.0.
688
690
691  If your code is destined for systems with severely constrained (or
692  missing!) virtual memory systems then you want to be I<especially> mindful
693  of avoiding wasteful constructs such as:
694
695      # NOTE: this is no longer "bad" in perl5.005
696      for (0..10000000) {}                       # bad
697      for (my $x = 0;$x <= 10000000; ++$x) {} # good 698 699 @lines = <VERY_LARGE_FILE>; # bad 700 701 while (<FILE>) {$file .= $_} # sometimes bad 702$file = join('', <FILE>);                  # better
703
704  The last two constructs may appear unintuitive to most people.  The
705  first repeatedly grows a string, whereas the second allocates a
706  large chunk of memory in one go.  On some systems, the second is
707  more efficient that the first.
708
710
711  Most multi-user platforms provide basic levels of security, usually
712  implemented at the filesystem level.  Some, however, do
713  not-- unfortunately.  Thus the notion of user id, or "home" directory,
714  or even the state of being logged-in, may be unrecognizable on many
715  platforms.  If you write programs that are security-conscious, it
716  is usually best to know what type of system you will be running
717  under so that you can write code explicitly for that platform (or
718  class of platforms).
719
720  Don't assume the UNIX filesystem access semantics: the operating
721  system or the filesystem may be using some ACL systems, which are
722  richer languages than the usual rwx.  Even if the rwx exist,
723  their semantics might be different.
724
725  (From security viewpoint testing for permissions before attempting to
726  do something is silly anyway: if one tries this, there is potential
727  for race conditions-- someone or something might change the
728  permissions between the permissions check and the actual operation.
729  Just try the operation.)
730
731  Don't assume the UNIX user and group semantics: especially, don't
732  expect the C<< $< >> and C<<$> >> (or the C<$(> and C<$)>) to work
733  for switching identities (or memberships).
734
735  Don't assume set-uid and set-gid semantics. (And even if you do,
736  think twice: set-uid and set-gid are a known can of security worms.)
737
739
740  For those times when it is necessary to have platform-specific code,
741  consider keeping the platform-specific code in one place, making porting
742  to other platforms easier.  Use the Config module and the special
743  variable C<$^O> to differentiate platforms, as described in 744 L<"PLATFORMS">. 745 746 Be careful in the tests you supply with your module or programs. 747 Module code may be fully portable, but its tests might not be. This 748 often happens when tests spawn off other processes or call external 749 programs to aid in the testing, or when (as noted above) the tests 750 assume certain things about the filesystem and paths. Be careful not 751 to depend on a specific output style for errors, such as when checking 752 C<$!> after a failed system call.  Using C<$!> for anything else than 753 displaying it as output is doubtful (though see the Errno module for 754 testing reasonably portably for error value). Some platforms expect 755 a certain output format, and Perl on those platforms may have been 756 adjusted accordingly. Most specifically, don't anchor a regex when 757 testing an error value. 758 759 =head1 CPAN Testers 760 761 Modules uploaded to CPAN are tested by a variety of volunteers on 762 different platforms. These CPAN testers are notified by mail of each 763 new upload, and reply to the list with PASS, FAIL, NA (not applicable to 764 this platform), or UNKNOWN (unknown), along with any relevant notations. 765 766 The purpose of the testing is twofold: one, to help developers fix any 767 problems in their code that crop up because of lack of testing on other 768 platforms; two, to provide users with information about whether 769 a given module works on a given platform. 770 771 Also see: 772 773 =over 4 774 775 =item * 776 777 Mailing list: cpan-testers@perl.org 778 779 =item * 780 781 Testing results: http://testers.cpan.org/ 782 783 =back 784 785 =head1 PLATFORMS 786 787 As of version 5.002, Perl is built with a C<$^O> variable that
788  indicates the operating system it was built on.  This was implemented
789  to help speed up code that would otherwise have to C<use Config>
790  and use the value of C<$Config{osname}>. Of course, to get more 791 detailed information about the system, looking into C<%Config> is 792 certainly recommended. 793 794 C<%Config> cannot always be trusted, however, because it was built 795 at compile time. If perl was built in one place, then transferred 796 elsewhere, some values may be wrong. The values may even have been 797 edited after the fact. 798 799 =head2 Unix 800 801 Perl works on a bewildering variety of Unix and Unix-like platforms (see 802 e.g. most of the files in the F<hints/> directory in the source code kit). 803 On most of these systems, the value of C<$^O> (hence C<$Config{'osname'}>, 804 too) is determined either by lowercasing and stripping punctuation from the 805 first field of the string returned by typing C<uname -a> (or a similar command) 806 at the shell prompt or by testing the file system for the presence of 807 uniquely named files such as a kernel or header file. Here, for example, 808 are a few of the more popular Unix flavors: 809 810 uname$^O        $Config{'archname'} 811 -------------------------------------------- 812 AIX aix aix 813 BSD/OS bsdos i386-bsdos 814 Darwin darwin darwin 815 dgux dgux AViiON-dgux 816 DYNIX/ptx dynixptx i386-dynixptx 817 FreeBSD freebsd freebsd-i386 818 Linux linux arm-linux 819 Linux linux i386-linux 820 Linux linux i586-linux 821 Linux linux ppc-linux 822 HP-UX hpux PA-RISC1.1 823 IRIX irix irix 824 Mac OS X darwin darwin 825 MachTen PPC machten powerpc-machten 826 NeXT 3 next next-fat 827 NeXT 4 next OPENSTEP-Mach 828 openbsd openbsd i386-openbsd 829 OSF1 dec_osf alpha-dec_osf 830 reliantunix-n svr4 RM400-svr4 831 SCO_SV sco_sv i386-sco_sv 832 SINIX-N svr4 RM400-svr4 833 sn4609 unicos CRAY_C90-unicos 834 sn6521 unicosmk t3e-unicosmk 835 sn9617 unicos CRAY_J90-unicos 836 SunOS solaris sun4-solaris 837 SunOS solaris i86pc-solaris 838 SunOS4 sunos sun4-sunos 839 840 Because the value of C<$Config{archname}> may depend on the
841  hardware architecture, it can vary more than the value of C<$^O>. 842 843 =head2 DOS and Derivatives 844 845 Perl has long been ported to Intel-style microcomputers running under 846 systems like PC-DOS, MS-DOS, OS/2, and most Windows platforms you can 847 bring yourself to mention (except for Windows CE, if you count that). 848 Users familiar with I<COMMAND.COM> or I<CMD.EXE> style shells should 849 be aware that each of these file specifications may have subtle 850 differences: 851 852$filespec0 = "c:/foo/bar/file.txt";
853      $filespec1 = "c:\\foo\\bar\\file.txt"; 854$filespec2 = 'c:\foo\bar\file.txt';
855      $filespec3 = 'c:\\foo\\bar\\file.txt'; 856 857 System calls accept either C</> or C<\> as the path separator. 858 However, many command-line utilities of DOS vintage treat C</> as 859 the option prefix, so may get confused by filenames containing C</>. 860 Aside from calling any external programs, C</> will work just fine, 861 and probably better, as it is more consistent with popular usage, 862 and avoids the problem of remembering what to backwhack and what 863 not to. 864 865 The DOS FAT filesystem can accommodate only "8.3" style filenames. Under 866 the "case-insensitive, but case-preserving" HPFS (OS/2) and NTFS (NT) 867 filesystems you may have to be careful about case returned with functions 868 like C<readdir> or used with functions like C<open> or C<opendir>. 869 870 DOS also treats several filenames as special, such as AUX, PRN, 871 NUL, CON, COM1, LPT1, LPT2, etc. Unfortunately, sometimes these 872 filenames won't even work if you include an explicit directory 873 prefix. It is best to avoid such filenames, if you want your code 874 to be portable to DOS and its derivatives. It's hard to know what 875 these all are, unfortunately. 876 877 Users of these operating systems may also wish to make use of 878 scripts such as I<pl2bat.bat> or I<pl2cmd> to 879 put wrappers around your scripts. 880 881 Newline (C<\n>) is translated as C<\015\012> by STDIO when reading from 882 and writing to files (see L<"Newlines">). C<binmode(FILEHANDLE)> 883 will keep C<\n> translated as C<\012> for that filehandle. Since it is a 884 no-op on other systems, C<binmode> should be used for cross-platform code 885 that deals with binary data. That's assuming you realize in advance 886 that your data is in binary. General-purpose programs should 887 often assume nothing about their data. 888 889 The C<$^O> variable and the C<$Config{archname}> values for various 890 DOSish perls are as follows: 891 892 OS$^O      $Config{archname} ID Version 893 -------------------------------------------------------- 894 MS-DOS dos ? 895 PC-DOS dos ? 896 OS/2 os2 ? 897 Windows 3.1 ? ? 0 3 01 898 Windows 95 MSWin32 MSWin32-x86 1 4 00 899 Windows 98 MSWin32 MSWin32-x86 1 4 10 900 Windows ME MSWin32 MSWin32-x86 1 ? 901 Windows NT MSWin32 MSWin32-x86 2 4 xx 902 Windows NT MSWin32 MSWin32-ALPHA 2 4 xx 903 Windows NT MSWin32 MSWin32-ppc 2 4 xx 904 Windows 2000 MSWin32 MSWin32-x86 2 5 00 905 Windows XP MSWin32 MSWin32-x86 2 5 01 906 Windows 2003 MSWin32 MSWin32-x86 2 5 02 907 Windows CE MSWin32 ? 3 908 Cygwin cygwin cygwin 909 910 The various MSWin32 Perl's can distinguish the OS they are running on 911 via the value of the fifth element of the list returned from 912 Win32::GetOSVersion(). For example: 913 914 if ($^O eq 'MSWin32') {
915          my @os_version_info = Win32::GetOSVersion();
916          print +('3.1','95','NT')[$os_version_info[4]],"\n"; 917 } 918 919 There are also Win32::IsWinNT() and Win32::IsWin95(), try C<perldoc Win32>, 920 and as of libwin32 0.19 (not part of the core Perl distribution) 921 Win32::GetOSName(). The very portable POSIX::uname() will work too: 922 923 c:\> perl -MPOSIX -we "print join '|', uname" 924 Windows NT|moonru|5.0|Build 2195 (Service Pack 2)|x86 925 926 Also see: 927 928 =over 4 929 930 =item * 931 932 The djgpp environment for DOS, http://www.delorie.com/djgpp/ 933 and L<perldos>. 934 935 =item * 936 937 The EMX environment for DOS, OS/2, etc. emx@iaehv.nl, 938 http://www.leo.org/pub/comp/os/os2/leo/gnu/emx+gcc/index.html or 939 ftp://hobbes.nmsu.edu/pub/os2/dev/emx/ Also L<perlos2>. 940 941 =item * 942 943 Build instructions for Win32 in L<perlwin32>, or under the Cygnus environment 944 in L<perlcygwin>. 945 946 =item * 947 948 The C<Win32::*> modules in L<Win32>. 949 950 =item * 951 952 The ActiveState Pages, http://www.activestate.com/ 953 954 =item * 955 956 The Cygwin environment for Win32; F<README.cygwin> (installed 957 as L<perlcygwin>), http://www.cygwin.com/ 958 959 =item * 960 961 The U/WIN environment for Win32, 962 http://www.research.att.com/sw/tools/uwin/ 963 964 =item * 965 966 Build instructions for OS/2, L<perlos2> 967 968 =back 969 970 =head2 S<Mac OS> 971 972 Any module requiring XS compilation is right out for most people, because 973 MacPerl is built using non-free (and non-cheap!) compilers. Some XS 974 modules that can work with MacPerl are built and distributed in binary 975 form on CPAN. 976 977 Directories are specified as: 978 979 volume:folder:file for absolute pathnames 980 volume:folder: for absolute pathnames 981 :folder:file for relative pathnames 982 :folder: for relative pathnames 983 :file for relative pathnames 984 file for relative pathnames 985 986 Files are stored in the directory in alphabetical order. Filenames are 987 limited to 31 characters, and may include any character except for 988 null and C<:>, which is reserved as the path separator. 989 990 Instead of C<flock>, see C<FSpSetFLock> and C<FSpRstFLock> in the 991 Mac::Files module, or C<chmod(0444, ...)> and C<chmod(0666, ...)>. 992 993 In the MacPerl application, you can't run a program from the command line; 994 programs that expect C<@ARGV> to be populated can be edited with something 995 like the following, which brings up a dialog box asking for the command 996 line arguments. 997 998 if (!@ARGV) { 999 @ARGV = split /\s+/, MacPerl::Ask('Arguments?'); 1000 } 1001 1002 A MacPerl script saved as a "droplet" will populate C<@ARGV> with the full 1003 pathnames of the files dropped onto the script. 1004 1005 Mac users can run programs under a type of command line interface 1006 under MPW (Macintosh Programmer's Workshop, a free development 1007 environment from Apple). MacPerl was first introduced as an MPW 1008 tool, and MPW can be used like a shell: 1009 1010 perl myscript.plx some arguments 1011 1012 ToolServer is another app from Apple that provides access to MPW tools 1013 from MPW and the MacPerl app, which allows MacPerl programs to use 1014 C<system>, backticks, and piped C<open>. 1015 1016 "S<Mac OS>" is the proper name for the operating system, but the value 1017 in C<$^O> is "MacOS".  To determine architecture, version, or whether
1018  the application or MPW tool version is running, check:
1019
1020      $is_app =$MacPerl::Version =~ /App/;
1021      $is_tool =$MacPerl::Version =~ /MPW/;
1022      ($version) =$MacPerl::Version =~ /^(\S+)/;
1023      $is_ppc =$MacPerl::Architecture eq 'MacPPC';
1024      $is_68k =$MacPerl::Architecture eq 'Mac68K';
1025
1026  S<Mac OS X>, based on NeXT's OpenStep OS, runs MacPerl natively, under the
1027  "Classic" environment.  There is no "Carbon" version of MacPerl to run
1028  under the primary Mac OS X environment.  S<Mac OS X> and its Open Source
1029  version, Darwin, both run Unix perl natively.
1030
1031  Also see:
1032
1033  =over 4
1034
1035  =item *
1036
1037  MacPerl Development, http://dev.macperl.org/ .
1038
1039  =item *
1040
1041  The MacPerl Pages, http://www.macperl.com/ .
1042
1043  =item *
1044
1045  The MacPerl mailing lists, http://lists.perl.org/ .
1046
1047  =item *
1048
1049  MPW, ftp://ftp.apple.com/developer/Tool_Chest/Core_Mac_OS_Tools/
1050
1051  =back
1052
1054
1055  Perl on VMS is discussed in L<perlvms> in the perl distribution.
1056
1057  The official name of VMS as of this writing is OpenVMS.
1058
1059  Perl on VMS can accept either VMS- or Unix-style file
1060  specifications as in either of the following:
1061
1062      $perl -ne "print if /perl_setup/i" SYS$LOGIN:LOGIN.COM
1063      $perl -ne "print if /perl_setup/i" /sys$login/login.com
1064
1065  but not a mixture of both as in:
1066
1067      $perl -ne "print if /perl_setup/i" sys$login:/login.com
1068      Can't open sys$login:/login.com: file specification syntax error 1069 1070 Interacting with Perl from the Digital Command Language (DCL) shell 1071 often requires a different set of quotation marks than Unix shells do. 1072 For example: 1073 1074$ perl -e "print ""Hello, world.\n"""
1075      Hello, world.
1076
1077  There are several ways to wrap your perl scripts in DCL F<.COM> files, if
1078  you are so inclined.  For example:
1079
1080      $write sys$output "Hello from DCL!"
1081      $if p1 .eqs. "" 1082$ then perl -x 'f$environment("PROCEDURE") 1083$ else perl -x - 'p1 'p2 'p3 'p4 'p5 'p6 'p7 'p8
1084      $deck/dollars="__END__" 1085 #!/usr/bin/perl 1086 1087 print "Hello from Perl!\n"; 1088 1089 __END__ 1090$ endif
1091
1092  Do take care with C<$ASSIGN/nolog/user SYS$COMMAND: SYS$INPUT> if your 1093 perl-in-DCL script expects to do things like C<<$read = <STDIN>; >>.
1094
1095  The VMS operating system has two filesystems, known as ODS-2 and ODS-5.
1096
1097  For ODS-2, filenames are in the format "name.extension;version".  The
1098  maximum length for filenames is 39 characters, and the maximum length for
1099  extensions is also 39 characters.  Version is a number from 1 to
1100  32767.  Valid characters are C</[A-Z0-9$_-]/>. 1101 1102 The ODS-2 filesystem is case-insensitive and does not preserve case. 1103 Perl simulates this by converting all filenames to lowercase internally. 1104 1105 For ODS-5, filenames may have almost any character in them and can include 1106 Unicode characters. Characters that could be misinterpreted by the DCL 1107 shell or file parsing utilities need to be prefixed with the C<^> 1108 character, or replaced with hexadecimal characters prefixed with the 1109 C<^> character. Such prefixing is only needed with the pathnames are 1110 in VMS format in applications. Programs that can accept the UNIX format 1111 of pathnames do not need the escape characters. The maximum length for 1112 filenames is 255 characters. The ODS-5 file system can handle both 1113 a case preserved and a case sensitive mode. 1114 1115 ODS-5 is only available on the OpenVMS for 64 bit platforms. 1116 1117 Support for the extended file specifications is being done as optional 1118 settings to preserve backward compatibility with Perl scripts that 1119 assume the previous VMS limitations. 1120 1121 In general routines on VMS that get a UNIX format file specification 1122 should return it in a UNIX format, and when they get a VMS format 1123 specification they should return a VMS format unless they are documented 1124 to do a conversion. 1125 1126 For routines that generate return a file specification, VMS allows setting 1127 if the C library which Perl is built on if it will be returned in VMS 1128 format or in UNIX format. 1129 1130 With the ODS-2 file system, there is not much difference in syntax of 1131 filenames without paths for VMS or UNIX. With the extended character 1132 set available with ODS-5 there can be a significant difference. 1133 1134 Because of this, existing Perl scripts written for VMS were sometimes 1135 treating VMS and UNIX filenames interchangeably. Without the extended 1136 character set enabled, this behavior will mostly be maintained for 1137 backwards compatibility. 1138 1139 When extended characters are enabled with ODS-5, the handling of 1140 UNIX formatted file specifications is to that of a UNIX system. 1141 1142 VMS file specifications without extensions have a trailing dot. An 1143 equivalent UNIX file specification should not show the trailing dot. 1144 1145 The result of all of this, is that for VMS, for portable scripts, you 1146 can not depend on Perl to present the filenames in lowercase, to be 1147 case sensitive, and that the filenames could be returned in either 1148 UNIX or VMS format. 1149 1150 And if a routine returns a file specification, unless it is intended to 1151 convert it, it should return it in the same format as it found it. 1152 1153 C<readdir> by default has traditionally returned lowercased filenames. 1154 When the ODS-5 support is enabled, it will return the exact case of the 1155 filename on the disk. 1156 1157 Files without extensions have a trailing period on them, so doing a 1158 C<readdir> in the default mode with a file named F<A.;5> will 1159 return F<a.> when VMS is (though that file could be opened with 1160 C<open(FH, 'A')>). 1161 1162 With support for extended file specifications and if C<opendir> was 1163 given a UNIX format directory, a file named F<A.;5> will return F<a> 1164 and optionally in the exact case on the disk. When C<opendir> is given 1165 a VMS format directory, then C<readdir> should return F<a.>, and 1166 again with the optionally the exact case. 1167 1168 RMS had an eight level limit on directory depths from any rooted logical 1169 (allowing 16 levels overall) prior to VMS 7.2, and even with versions of 1170 VMS on VAX up through 7.3. Hence C<PERL_ROOT:[LIB.2.3.4.5.6.7.8]> is a 1171 valid directory specification but C<PERL_ROOT:[LIB.2.3.4.5.6.7.8.9]> is 1172 not. F<Makefile.PL> authors might have to take this into account, but at 1173 least they can refer to the former as C</PERL_ROOT/lib/2/3/4/5/6/7/8/>. 1174 1175 Pumpkings and module integrators can easily see whether files with too many 1176 directory levels have snuck into the core by running the following in the 1177 top-level source directory: 1178 1179$ perl -ne "$_=~s/\s+.*//; print if scalar(split /\//) > 8;" < MANIFEST 1180 1181 1182 The VMS::Filespec module, which gets installed as part of the build 1183 process on VMS, is a pure Perl module that can easily be installed on 1184 non-VMS platforms and can be helpful for conversions to and from RMS 1185 native formats. It is also now the only way that you should check to 1186 see if VMS is in a case sensitive mode. 1187 1188 What C<\n> represents depends on the type of file opened. It usually 1189 represents C<\012> but it could also be C<\015>, C<\012>, C<\015\012>, 1190 C<\000>, C<\040>, or nothing depending on the file organization and 1191 record format. The VMS::Stdio module provides access to the 1192 special fopen() requirements of files with unusual attributes on VMS. 1193 1194 TCP/IP stacks are optional on VMS, so socket routines might not be 1195 implemented. UDP sockets may not be supported. 1196 1197 The TCP/IP library support for all current versions of VMS is dynamically 1198 loaded if present, so even if the routines are configured, they may 1199 return a status indicating that they are not implemented. 1200 1201 The value of C<$^O> on OpenVMS is "VMS".  To determine the architecture
1202  that you are running on without resorting to loading all of C<%Config>
1203  you can examine the content of the C<@INC> array like so:
1204
1205      if (grep(/VMS_AXP/, @INC)) {
1206          print "I'm on Alpha!\n";
1207
1208      } elsif (grep(/VMS_VAX/, @INC)) {
1209          print "I'm on VAX!\n";
1210
1211      } elsif (grep(/VMS_IA64/, @INC)) {
1212          print "I'm on IA64!\n";
1213
1214      } else {
1215          print "I'm not so sure about where $^O is...\n"; 1216 } 1217 1218 In general, the significant differences should only be if Perl is running 1219 on VMS_VAX or one of the 64 bit OpenVMS platforms. 1220 1221 On VMS, perl determines the UTC offset from the C<SYS$TIMEZONE_DIFFERENTIAL>
1222  logical name.  Although the VMS epoch began at 17-NOV-1858 00:00:00.00,
1223  calls to C<localtime> are adjusted to count offsets from
1224  01-JAN-1970 00:00:00.00, just like Unix.
1225
1226  Also see:
1227
1228  =over 4
1229
1230  =item *
1231
1233
1234  =item *
1235
1236  vmsperl list, vmsperl-subscribe@perl.org
1237
1238  =item *
1239
1240  vmsperl on the web, http://www.sidhe.org/vmsperl/index.html
1241
1242  =back
1243
1245
1246  Perl on VOS is discussed in F<README.vos> in the perl distribution
1247  (installed as L<perlvos>).  Perl on VOS can accept either VOS- or
1248  Unix-style file specifications as in either of the following:
1249
1250      C<< $perl -ne "print if /perl_setup/i" >system>notices >> 1251 C<<$ perl -ne "print if /perl_setup/i" /system/notices >>
1252
1253  or even a mixture of both as in:
1254
1255      C<< $perl -ne "print if /perl_setup/i" >system/notices >> 1256 1257 Even though VOS allows the slash character to appear in object 1258 names, because the VOS port of Perl interprets it as a pathname 1259 delimiting character, VOS files, directories, or links whose names 1260 contain a slash character cannot be processed. Such files must be 1261 renamed before they can be processed by Perl. Note that VOS limits 1262 file names to 32 or fewer characters. 1263 1264 The value of C<$^O> on VOS is "VOS".  To determine the architecture that
1265  you are running on without resorting to loading all of C<%Config> you
1266  can examine the content of the @INC array like so:
1267
1268      if ($^O =~ /VOS/) { 1269 print "I'm on a Stratus box!\n"; 1270 } else { 1271 print "I'm not on a Stratus box!\n"; 1272 die; 1273 } 1274 1275 Also see: 1276 1277 =over 4 1278 1279 =item * 1280 1281 F<README.vos> (installed as L<perlvos>) 1282 1283 =item * 1284 1285 The VOS mailing list. 1286 1287 There is no specific mailing list for Perl on VOS. You can post 1288 comments to the comp.sys.stratus newsgroup, or subscribe to the general 1289 Stratus mailing list. Send a letter with "subscribe Info-Stratus" in 1290 the message body to majordomo@list.stratagy.com. 1291 1292 =item * 1293 1294 VOS Perl on the web at http://ftp.stratus.com/pub/vos/posix/posix.html 1295 1296 =back 1297 1298 =head2 EBCDIC Platforms 1299 1300 Recent versions of Perl have been ported to platforms such as OS/400 on 1301 AS/400 minicomputers as well as OS/390, VM/ESA, and BS2000 for S/390 1302 Mainframes. Such computers use EBCDIC character sets internally (usually 1303 Character Code Set ID 0037 for OS/400 and either 1047 or POSIX-BC for S/390 1304 systems). On the mainframe perl currently works under the "Unix system 1305 services for OS/390" (formerly known as OpenEdition), VM/ESA OpenEdition, or 1306 the BS200 POSIX-BC system (BS2000 is supported in perl 5.6 and greater). 1307 See L<perlos390> for details. Note that for OS/400 there is also a port of 1308 Perl 5.8.1/5.9.0 or later to the PASE which is ASCII-based (as opposed to 1309 ILE which is EBCDIC-based), see L<perlos400>. 1310 1311 As of R2.5 of USS for OS/390 and Version 2.3 of VM/ESA these Unix 1312 sub-systems do not support the C<#!> shebang trick for script invocation. 1313 Hence, on OS/390 and VM/ESA perl scripts can be executed with a header 1314 similar to the following simple script: 1315 1316 : # use perl 1317 eval 'exec /usr/local/bin/perl -S$0 $1+"$@"}'
1318              if 0;
1319      #!/usr/local/bin/perl     # just a comment really
1320
1321      print "Hello from perl!\n";
1322
1323  OS/390 will support the C<#!> shebang trick in release 2.8 and beyond.
1324  Calls to C<system> and backticks can use POSIX shell syntax on all
1325  S/390 systems.
1326
1327  On the AS/400, if PERL5 is in your library list, you may need
1328  to wrap your perl scripts in a CL procedure to invoke them like so:
1329
1330      BEGIN
1331        CALL PGM(PERL5/PERL) PARM('/QOpenSys/hello.pl')
1332      ENDPGM
1333
1334  This will invoke the perl script F<hello.pl> in the root of the
1335  QOpenSys file system.  On the AS/400 calls to C<system> or backticks
1336  must use CL syntax.
1337
1338  On these platforms, bear in mind that the EBCDIC character set may have
1339  an effect on what happens with some perl functions (such as C<chr>,
1340  C<pack>, C<print>, C<printf>, C<ord>, C<sort>, C<sprintf>, C<unpack>), as
1341  well as bit-fiddling with ASCII constants using operators like C<^>, C<&>
1342  and C<|>, not to mention dealing with socket interfaces to ASCII computers
1343  (see L<"Newlines">).
1344
1345  Fortunately, most web servers for the mainframe will correctly
1346  translate the C<\n> in the following statement to its ASCII equivalent
1347  (C<\r> is the same under both Unix and OS/390 & VM/ESA):
1348
1349      print "Content-type: text/html\r\n\r\n";
1350
1351  The values of C<$^O> on some of these platforms includes: 1352 1353 uname$^O        $Config{'archname'} 1354 -------------------------------------------- 1355 OS/390 os390 os390 1356 OS400 os400 os400 1357 POSIX-BC posix-bc BS2000-posix-bc 1358 VM/ESA vmesa vmesa 1359 1360 Some simple tricks for determining if you are running on an EBCDIC 1361 platform could include any of the following (perhaps all): 1362 1363 if ("\t" eq "\05") { print "EBCDIC may be spoken here!\n"; } 1364 1365 if (ord('A') == 193) { print "EBCDIC may be spoken here!\n"; } 1366 1367 if (chr(169) eq 'z') { print "EBCDIC may be spoken here!\n"; } 1368 1369 One thing you may not want to rely on is the EBCDIC encoding 1370 of punctuation characters since these may differ from code page to code 1371 page (and once your module or script is rumoured to work with EBCDIC, 1372 folks will want it to work with all EBCDIC character sets). 1373 1374 Also see: 1375 1376 =over 4 1377 1378 =item * 1379 1380 L<perlos390>, F<README.os390>, F<perlbs2000>, F<README.vmesa>, 1381 L<perlebcdic>. 1382 1383 =item * 1384 1385 The perl-mvs@perl.org list is for discussion of porting issues as well as 1386 general usage issues for all EBCDIC Perls. Send a message body of 1387 "subscribe perl-mvs" to majordomo@perl.org. 1388 1389 =item * 1390 1391 AS/400 Perl information at 1392 http://as400.rochester.ibm.com/ 1393 as well as on CPAN in the F<ports/> directory. 1394 1395 =back 1396 1397 =head2 Acorn RISC OS 1398 1399 Because Acorns use ASCII with newlines (C<\n>) in text files as C<\012> like 1400 Unix, and because Unix filename emulation is turned on by default, 1401 most simple scripts will probably work "out of the box". The native 1402 filesystem is modular, and individual filesystems are free to be 1403 case-sensitive or insensitive, and are usually case-preserving. Some 1404 native filesystems have name length limits, which file and directory 1405 names are silently truncated to fit. Scripts should be aware that the 1406 standard filesystem currently has a name length limit of B<10> 1407 characters, with up to 77 items in a directory, but other filesystems 1408 may not impose such limitations. 1409 1410 Native filenames are of the form 1411 1412 Filesystem#Special_Field::DiskName.$.Directory.Directory.File
1413
1414  where
1415
1416      Special_Field is not usually present, but may contain . and $. 1417 Filesystem =~ m|[A-Za-z0-9_]| 1418 DsicName =~ m|[A-Za-z0-9_/]| 1419$ represents the root directory
1420      . is the path separator
1421      @ is the current directory (per filesystem but machine global)
1422      ^ is the parent directory
1423      Directory and File =~ m|[^\0- "\.\$\%\&:\@\\^\|\177]+| 1424 1425 The default filename translation is roughly C<tr|/.|./|;> 1426 1427 Note that C<"ADFS::HardDisk.$.File" ne 'ADFS::HardDisk.$.File'> and that 1428 the second stage of C<$> interpolation in regular expressions will fall
1429  foul of the C<$.> if scripts are not careful. 1430 1431 Logical paths specified by system variables containing comma-separated 1432 search lists are also allowed; hence C<System:Modules> is a valid 1433 filename, and the filesystem will prefix C<Modules> with each section of 1434 C<System$Path> until a name is made that points to an object on disk.
1435  Writing to a new file C<System:Modules> would be allowed only if
1436  C<System$Path> contains a single item list. The filesystem will also 1437 expand system variables in filenames if enclosed in angle brackets, so 1438 C<< <System$Dir>.Modules >> would look for the file
1439  S<C<$ENV{'System$Dir'} . 'Modules'>>.  The obvious implication of this is
1440  that B<fully qualified filenames can start with C<< <> >>> and should
1441  be protected when C<open> is used for input.
1442
1443  Because C<.> was in use as a directory separator and filenames could not
1444  be assumed to be unique after 10 characters, Acorn implemented the C
1445  compiler to strip the trailing C<.c> C<.h> C<.s> and C<.o> suffix from
1446  filenames specified in source code and store the respective files in
1447  subdirectories named after the suffix.  Hence files are translated:
1448
1449      foo.h           h.foo
1450      C:foo.h         C:h.foo        (logical path variable)
1451      sys/os.h        sys.h.os       (C compiler groks Unix-speak)
1452      10charname.c    c.10charname
1453      10charname.o    o.10charname
1454      11charname_.c   c.11charname   (assuming filesystem truncates at 10)
1455
1456  The Unix emulation library's translation of filenames to native assumes
1457  that this sort of translation is required, and it allows a user-defined list
1458  of known suffixes that it will transpose in this fashion.  This may
1459  seem transparent, but consider that with these rules C<foo/bar/baz.h>
1460  and C<foo/bar/h/baz> both map to C<foo.bar.h.baz>, and that C<readdir> and
1461  C<glob> cannot and do not attempt to emulate the reverse mapping.  Other
1462  C<.>'s in filenames are translated to C</>.
1463
1464  As implied above, the environment accessed through C<%ENV> is global, and
1465  the convention is that program specific environment variables are of the
1466  form C<Program$Name>. Each filesystem maintains a current directory, 1467 and the current filesystem's current directory is the B<global> current 1468 directory. Consequently, sociable programs don't change the current 1469 directory but rely on full pathnames, and programs (and Makefiles) cannot 1470 assume that they can spawn a child process which can change the current 1471 directory without affecting its parent (and everyone else for that 1472 matter). 1473 1474 Because native operating system filehandles are global and are currently 1475 allocated down from 255, with 0 being a reserved value, the Unix emulation 1476 library emulates Unix filehandles. Consequently, you can't rely on 1477 passing C<STDIN>, C<STDOUT>, or C<STDERR> to your children. 1478 1479 The desire of users to express filenames of the form 1480 C<< <Foo$Dir>.Bar >> on the command line unquoted causes problems,
1481  too: C<> command output capture has to perform a guessing game.  It
1482  assumes that a string C<< <[^<>]+\$[^<>]> >> is a 1483 reference to an environment variable, whereas anything else involving 1484 C<< < >> or C<< > >> is redirection, and generally manages to be 99% 1485 right. Of course, the problem remains that scripts cannot rely on any 1486 Unix tools being available, or that any tools found have Unix-like command 1487 line arguments. 1488 1489 Extensions and XS are, in theory, buildable by anyone using free 1490 tools. In practice, many don't, as users of the Acorn platform are 1491 used to binary distributions. MakeMaker does run, but no available 1492 make currently copes with MakeMaker's makefiles; even if and when 1493 this should be fixed, the lack of a Unix-like shell will cause 1494 problems with makefile rules, especially lines of the form C<cd 1495 sdbm && make all>, and anything using quoting. 1496 1497 "S<RISC OS>" is the proper name for the operating system, but the value 1498 in C<$^O> is "riscos" (because we don't like shouting).
1499
1501
1502  Perl has been ported to many platforms that do not fit into any of
1503  the categories listed above.  Some, such as AmigaOS, Atari MiNT,
1504  BeOS, HP MPE/iX, QNX, Plan 9, and VOS, have been well-integrated
1505  into the standard Perl source code kit.  You may need to see the
1506  F<ports/> directory on CPAN for information, and possibly binaries,
1507  for the likes of: aos, Atari ST, lynxos, riscos, Novell Netware,
1508  Tandem Guardian, I<etc.>  (Yes, we know that some of these OSes may
1509  fall under the Unix category, but we are not a standards body.)
1510
1511  Some approximate operating system names and their C<$^O> values 1512 in the "OTHER" category include: 1513 1514 OS$^O        $Config{'archname'} 1515 ------------------------------------------ 1516 Amiga DOS amigaos m68k-amigos 1517 BeOS beos 1518 MPE/iX mpeix PA-RISC1.1 1519 1520 See also: 1521 1522 =over 4 1523 1524 =item * 1525 1526 Amiga, F<README.amiga> (installed as L<perlamiga>). 1527 1528 =item * 1529 1530 Atari, F<README.mint> and Guido Flohr's web page 1531 http://stud.uni-sb.de/~gufl0000/ 1532 1533 =item * 1534 1535 Be OS, F<README.beos> 1536 1537 =item * 1538 1539 HP 300 MPE/iX, F<README.mpeix> and Mark Bixby's web page 1540 http://www.bixby.org/mark/perlix.html 1541 1542 =item * 1543 1544 A free perl5-based PERL.NLM for Novell Netware is available in 1545 precompiled binary and source code form from http://www.novell.com/ 1546 as well as from CPAN. 1547 1548 =item * 1549 1550 S<Plan 9>, F<README.plan9> 1551 1552 =back 1553 1554 =head1 FUNCTION IMPLEMENTATIONS 1555 1556 Listed below are functions that are either completely unimplemented 1557 or else have been implemented differently on various platforms. 1558 Following each description will be, in parentheses, a list of 1559 platforms that the description applies to. 1560 1561 The list may well be incomplete, or even wrong in some places. When 1562 in doubt, consult the platform-specific README files in the Perl 1563 source distribution, and any other documentation resources accompanying 1564 a given port. 1565 1566 Be aware, moreover, that even among Unix-ish systems there are variations. 1567 1568 For many functions, you can also query C<%Config>, exported by 1569 default from the Config module. For example, to check whether the 1570 platform has the C<lstat> call, check C<$Config{d_lstat}>.  See
1571  L<Config> for a full description of available variables.
1572
1573  =head2 Alphabetical Listing of Perl Functions
1574
1575  =over 8
1576
1577  =item -X
1578
1579  C<-r>, C<-w>, and C<-x> have a limited meaning only; directories
1580  and applications are executable, and there are no uid/gid
1581  considerations.  C<-o> is not supported.  (S<Mac OS>)
1582
1583  C<-r>, C<-w>, C<-x>, and C<-o> tell whether the file is accessible,
1584  which may not reflect UIC-based file protections.  (VMS)
1585
1586  C<-s> returns the size of the data fork, not the total size of data fork
1587  plus resource fork.  (S<Mac OS>).
1588
1589  C<-s> by name on an open file will return the space reserved on disk,
1590  rather than the current extent.  C<-s> on an open filehandle returns the
1591  current size.  (S<RISC OS>)
1592
1593  C<-R>, C<-W>, C<-X>, C<-O> are indistinguishable from C<-r>, C<-w>,
1594  C<-x>, C<-o>. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1595
1596  C<-b>, C<-c>, C<-k>, C<-g>, C<-p>, C<-u>, C<-A> are not implemented.
1597  (S<Mac OS>)
1598
1599  C<-g>, C<-k>, C<-l>, C<-p>, C<-u>, C<-A> are not particularly meaningful.
1600  (Win32, VMS, S<RISC OS>)
1601
1602  C<-d> is true if passed a device spec without an explicit directory.
1603  (VMS)
1604
1605  C<-T> and C<-B> are implemented, but might misclassify Mac text files
1606  with foreign characters; this is the case will all platforms, but may
1607  affect S<Mac OS> often.  (S<Mac OS>)
1608
1609  C<-x> (or C<-X>) determine if a file ends in one of the executable
1610  suffixes.  C<-S> is meaningless.  (Win32)
1611
1612  C<-x> (or C<-X>) determine if a file has an executable file type.
1613  (S<RISC OS>)
1614
1615  =item atan2
1616
1617  Due to issues with various CPUs, math libraries, compilers, and standards,
1618  results for C<atan2()> may vary depending on any combination of the above.
1619  Perl attempts to conform to the Open Group/IEEE standards for the results
1620  returned from C<atan2()>, but cannot force the issue if the system Perl is
1621  run on does not allow it.  (Tru64, HP-UX 10.20)
1622
1623  The current version of the standards for C<atan2()> is available at
1624  L<http://www.opengroup.org/onlinepubs/009695399/functions/atan2.html>.
1625
1626  =item binmode
1627
1628  Meaningless.  (S<Mac OS>, S<RISC OS>)
1629
1630  Reopens file and restores pointer; if function fails, underlying
1631  filehandle may be closed, or pointer may be in a different position.
1632  (VMS)
1633
1634  The value returned by C<tell> may be affected after the call, and
1635  the filehandle may be flushed. (Win32)
1636
1637  =item chmod
1638
1639  Only limited meaning.  Disabling/enabling write permission is mapped to
1640  locking/unlocking the file. (S<Mac OS>)
1641
1642  Only good for changing "owner" read-write access, "group", and "other"
1643  bits are meaningless. (Win32)
1644
1645  Only good for changing "owner" and "other" read-write access. (S<RISC OS>)
1646
1647  Access permissions are mapped onto VOS access-control list changes. (VOS)
1648
1649  The actual permissions set depend on the value of the C<CYGWIN>
1650  in the SYSTEM environment settings.  (Cygwin)
1651
1652  =item chown
1653
1654  Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1655
1656  Does nothing, but won't fail. (Win32)
1657
1658  A little funky, because VOS's notion of ownership is a little funky (VOS).
1659
1660  =item chroot
1661
1662  Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS, VM/ESA)
1663
1664  =item crypt
1665
1666  May not be available if library or source was not provided when building
1667  perl. (Win32)
1668
1669  =item dbmclose
1670
1671  Not implemented. (VMS, S<Plan 9>, VOS)
1672
1673  =item dbmopen
1674
1675  Not implemented. (VMS, S<Plan 9>, VOS)
1676
1677  =item dump
1678
1679  Not useful. (S<Mac OS>, S<RISC OS>)
1680
1681  Not supported. (Cygwin, Win32)
1682
1683  Invokes VMS debugger. (VMS)
1684
1685  =item exec
1686
1687  Not implemented. (S<Mac OS>)
1688
1689  Implemented via Spawn. (VM/ESA)
1690
1691  Does not automatically flush output handles on some platforms.
1692  (SunOS, Solaris, HP-UX)
1693
1694  =item exit
1695
1696  Emulates UNIX exit() (which considers C<exit 1> to indicate an error) by
1697  mapping the C<1> to SS$_ABORT (C<44>). This behavior may be overridden 1698 with the pragma C<use vmsish 'exit'>. As with the CRTL's exit() 1699 function, C<exit 0> is also mapped to an exit status of SS$_NORMAL
1700  (C<1>); this mapping cannot be overridden.  Any other argument to exit()
1701  is used directly as Perl's exit status.  On VMS, unless the future
1702  POSIX_EXIT mode is enabled, the exit code should always be a valid
1703  VMS exit code and not a generic number.  When the POSIX_EXIT mode is
1704  enabled, a generic number will be encoded in a method compatible with
1705  the C library _POSIX_EXIT macro so that it can be decoded by other
1706  programs, particularly ones written in C, like the GNV package.  (VMS)
1707
1708  =item fcntl
1709
1710  Not implemented. (Win32)
1711  Some functions available based on the version of VMS. (VMS)
1712
1713  =item flock
1714
1715  Not implemented (S<Mac OS>, VMS, S<RISC OS>, VOS).
1716
1717  Available only on Windows NT (not on Windows 95). (Win32)
1718
1719  =item fork
1720
1721  Not implemented. (S<Mac OS>, AmigaOS, S<RISC OS>, VM/ESA, VMS)
1722
1723  Emulated using multiple interpreters.  See L<perlfork>.  (Win32)
1724
1725  Does not automatically flush output handles on some platforms.
1726  (SunOS, Solaris, HP-UX)
1727
1729
1730  Not implemented. (S<Mac OS>, S<RISC OS>)
1731
1732  =item getpgrp
1733
1734  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1735
1736  =item getppid
1737
1738  Not implemented. (S<Mac OS>, Win32, S<RISC OS>)
1739
1740  =item getpriority
1741
1742  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA)
1743
1744  =item getpwnam
1745
1746  Not implemented. (S<Mac OS>, Win32)
1747
1748  Not useful. (S<RISC OS>)
1749
1750  =item getgrnam
1751
1752  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1753
1754  =item getnetbyname
1755
1756  Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1757
1758  =item getpwuid
1759
1760  Not implemented. (S<Mac OS>, Win32)
1761
1762  Not useful. (S<RISC OS>)
1763
1764  =item getgrgid
1765
1766  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>)
1767
1769
1770  Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1771
1772  =item getprotobynumber
1773
1774  Not implemented. (S<Mac OS>)
1775
1776  =item getservbyport
1777
1778  Not implemented. (S<Mac OS>)
1779
1780  =item getpwent
1781
1782  Not implemented. (S<Mac OS>, Win32, VM/ESA)
1783
1784  =item getgrent
1785
1786  Not implemented. (S<Mac OS>, Win32, VMS, VM/ESA)
1787
1788  =item gethostbyname
1789
1790  C<gethostbyname('localhost')> does not work everywhere: you may have
1791  to use C<gethostbyname('127.0.0.1')>. (S<Mac OS>, S<Irix 5>)
1792
1793  =item gethostent
1794
1795  Not implemented. (S<Mac OS>, Win32)
1796
1797  =item getnetent
1798
1799  Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1800
1801  =item getprotoent
1802
1803  Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1804
1805  =item getservent
1806
1807  Not implemented. (Win32, S<Plan 9>)
1808
1809  =item sethostent
1810
1811  Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1812
1813  =item setnetent
1814
1815  Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1816
1817  =item setprotoent
1818
1819  Not implemented. (S<Mac OS>, Win32, S<Plan 9>, S<RISC OS>)
1820
1821  =item setservent
1822
1823  Not implemented. (S<Plan 9>, Win32, S<RISC OS>)
1824
1825  =item endpwent
1826
1827  Not implemented. (S<Mac OS>, MPE/iX, VM/ESA, Win32)
1828
1829  =item endgrent
1830
1831  Not implemented. (S<Mac OS>, MPE/iX, S<RISC OS>, VM/ESA, VMS, Win32)
1832
1833  =item endhostent
1834
1835  Not implemented. (S<Mac OS>, Win32)
1836
1837  =item endnetent
1838
1839  Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1840
1841  =item endprotoent
1842
1843  Not implemented. (S<Mac OS>, Win32, S<Plan 9>)
1844
1845  =item endservent
1846
1847  Not implemented. (S<Plan 9>, Win32)
1848
1849  =item getsockopt SOCKET,LEVEL,OPTNAME
1850
1851  Not implemented. (S<Plan 9>)
1852
1853  =item glob
1854
1855  This operator is implemented via the File::Glob extension on most
1856  platforms.  See L<File::Glob> for portability information.
1857
1858  =item gmtime
1859
1860  Same portability caveats as L<localtime>.
1861
1862  =item ioctl FILEHANDLE,FUNCTION,SCALAR
1863
1864  Not implemented. (VMS)
1865
1866  Available only for socket handles, and it does what the ioctlsocket() call
1867  in the Winsock API does. (Win32)
1868
1869  Available only for socket handles. (S<RISC OS>)
1870
1871  =item kill
1872
1873  C<kill(0, LIST)> is implemented for the sake of taint checking;
1874  use with other signals is unimplemented. (S<Mac OS>)
1875
1876  Not implemented, hence not useful for taint checking. (S<RISC OS>)
1877
1878  C<kill()> doesn't have the semantics of C<raise()>, i.e. it doesn't send
1879  a signal to the identified process like it does on Unix platforms.
1880  Instead C<kill($sig,$pid)> terminates the process identified by $pid, 1881 and makes it exit immediately with exit status$sig.  As in Unix, if
1882  $sig is 0 and the specified process exists, it returns true without 1883 actually terminating it. (Win32) 1884 1885 C<kill(-9,$pid)> will terminate the process specified by $pid and 1886 recursively all child processes owned by it. This is different from 1887 the Unix semantics, where the signal will be delivered to all 1888 processes in the same process group as the process specified by 1889$pid. (Win32)
1890
1891  Is not supported for process identification number of 0 or negative
1892  numbers. (VMS)
1893
1895
1896  Not implemented. (S<Mac OS>, MPE/iX, S<RISC OS>)
1897
1898  Link count not updated because hard links are not quite that hard
1899  (They are sort of half-way between hard and soft links). (AmigaOS)
1900
1901  Hard links are implemented on Win32 under NTFS only. They are
1902  natively supported on Windows 2000 and later.  On Windows NT they
1903  are implemented using the Windows POSIX subsystem support and the
1904  Perl process will need Administrator or Backup Operator privileges
1906
1907  Available on 64 bit OpenVMS 8.2 and later.  (VMS)
1908
1909  =item localtime
1910
1911  Because Perl currently relies on the native standard C localtime()
1912  function, it is only safe to use times between 0 and (2**31)-1.  Times
1913  outside this range may result in unexpected behavior depending on your
1914  operating system's implementation of localtime().
1915
1916  =item lstat
1917
1918  Not implemented. (S<RISC OS>)
1919
1920  Return values (especially for device and inode) may be bogus. (Win32)
1921
1922  =item msgctl
1923
1924  =item msgget
1925
1926  =item msgsnd
1927
1928  =item msgrcv
1929
1930  Not implemented. (S<Mac OS>, Win32, VMS, S<Plan 9>, S<RISC OS>, VOS)
1931
1932  =item open
1933
1934  The C<|> variants are supported only if ToolServer is installed.
1935  (S<Mac OS>)
1936
1937  open to C<|-> and C<-|> are unsupported. (S<Mac OS>, Win32, S<RISC OS>)
1938
1939  Opening a process does not automatically flush output handles on some
1940  platforms.  (SunOS, Solaris, HP-UX)
1941
1942  =item pipe
1943
1944  Very limited functionality. (MiNT)
1945
1947
1948  Not implemented. (Win32, VMS, S<RISC OS>)
1949
1950  =item rename
1951
1952  Can't move directories between directories on different logical volumes. (Win32)
1953
1954  =item select
1955
1956  Only implemented on sockets. (Win32, VMS)
1957
1958  Only reliable on sockets. (S<RISC OS>)
1959
1960  Note that the C<select FILEHANDLE> form is generally portable.
1961
1962  =item semctl
1963
1964  =item semget
1965
1966  =item semop
1967
1968  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1969
1970  =item setgrent
1971
1972  Not implemented. (S<Mac OS>, MPE/iX, VMS, Win32, S<RISC OS>, VOS)
1973
1974  =item setpgrp
1975
1976  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1977
1978  =item setpriority
1979
1980  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1981
1982  =item setpwent
1983
1984  Not implemented. (S<Mac OS>, MPE/iX, Win32, S<RISC OS>, VOS)
1985
1986  =item setsockopt
1987
1988  Not implemented. (S<Plan 9>)
1989
1990  =item shmctl
1991
1992  =item shmget
1993
1995
1996  =item shmwrite
1997
1998  Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS)
1999
2000  =item sockatmark
2001
2002  A relatively recent addition to socket functions, may not
2003  be implemented even in UNIX platforms.
2004
2005  =item socketpair
2006
2007  Not implemented. (Win32, S<RISC OS>, VOS, VM/ESA)
2008
2009  Available on 64 bit OpenVMS 8.2 and later.  (VMS)
2010
2011  =item stat
2012
2013  Platforms that do not have rdev, blksize, or blocks will return these
2014  as '', so numeric comparison or manipulation of these fields may cause
2015  'not numeric' warnings.
2016
2017  mtime and atime are the same thing, and ctime is creation time instead of
2018  inode change time. (S<Mac OS>).
2019
2020  ctime not supported on UFS (S<Mac OS X>).
2021
2022  ctime is creation time instead of inode change time  (Win32).
2023
2024  device and inode are not meaningful.  (Win32)
2025
2026  device and inode are not necessarily reliable.  (VMS)
2027
2028  mtime, atime and ctime all return the last modification time.  Device and
2029  inode are not necessarily reliable.  (S<RISC OS>)
2030
2031  dev, rdev, blksize, and blocks are not available.  inode is not
2032  meaningful and will differ between stat calls on the same file.  (os2)
2033
2034  some versions of cygwin when doing a stat("foo") and if not finding it
2035  may then attempt to stat("foo.exe") (Cygwin)
2036
2037  On Win32 stat() needs to open the file to determine the link count
2038  and update attributes that may have been changed through hard links.
2039  Setting ${^WIN32_SLOPPY_STAT} to a true value speeds up stat() by 2040 not performing this operation. (Win32) 2041 2042 =item symlink 2043 2044 Not implemented. (Win32, S<RISC OS>) 2045 2046 Implemented on 64 bit VMS 8.3. VMS requires the symbolic link to be in Unix 2047 syntax if it is intended to resolve to a valid path. 2048 2049 =item syscall 2050 2051 Not implemented. (S<Mac OS>, Win32, VMS, S<RISC OS>, VOS, VM/ESA) 2052 2053 =item sysopen 2054 2055 The traditional "0", "1", and "2" MODEs are implemented with different 2056 numeric values on some systems. The flags exported by C<Fcntl> 2057 (O_RDONLY, O_WRONLY, O_RDWR) should work everywhere though. (S<Mac 2058 OS>, OS/390, VM/ESA) 2059 2060 =item system 2061 2062 Only implemented if ToolServer is installed. (S<Mac OS>) 2063 2064 As an optimization, may not call the command shell specified in 2065 C<$ENV{PERL5SHELL}>.  C<system(1, @args)> spawns an external
2066  process and immediately returns its process designator, without
2067  waiting for it to terminate.  Return value may be used subsequently
2068  in C<wait> or C<waitpid>.  Failure to spawn() a subprocess is indicated
2069  by setting $? to "255 << 8". C<$?> is set in a way compatible with
2070  Unix (i.e. the exitstatus of the subprocess is obtained by "$? >> 8", 2071 as described in the documentation). (Win32) 2072 2073 There is no shell to process metacharacters, and the native standard is 2074 to pass a command line terminated by "\n" "\r" or "\0" to the spawned 2075 program. Redirection such as C<< > foo >> is performed (if at all) by 2076 the run time library of the spawned program. C<system> I<list> will call 2077 the Unix emulation library's C<exec> emulation, which attempts to provide 2078 emulation of the stdin, stdout, stderr in force in the parent, providing 2079 the child program uses a compatible version of the emulation library. 2080 I<scalar> will call the native command line direct and no such emulation 2081 of a child Unix program will exists. Mileage B<will> vary. (S<RISC OS>) 2082 2083 Far from being POSIX compliant. Because there may be no underlying 2084 /bin/sh tries to work around the problem by forking and execing the 2085 first token in its argument string. Handles basic redirection 2086 ("<" or ">") on its own behalf. (MiNT) 2087 2088 Does not automatically flush output handles on some platforms. 2089 (SunOS, Solaris, HP-UX) 2090 2091 The return value is POSIX-like (shifted up by 8 bits), which only allows 2092 room for a made-up value derived from the severity bits of the native 2093 32-bit condition code (unless overridden by C<use vmsish 'status'>). 2094 If the native condition code is one that has a POSIX value encoded, the 2095 POSIX value will be decoded to extract the expected exit value. 2096 For more details see L<perlvms/$?>. (VMS)
2097
2098  =item times
2099
2100  Only the first entry returned is nonzero. (S<Mac OS>)
2101
2102  "cumulative" times will be bogus.  On anything other than Windows NT
2103  or Windows 2000, "system" time will be bogus, and "user" time is
2104  actually the time returned by the clock() function in the C runtime
2105  library. (Win32)
2106
2107  Not useful. (S<RISC OS>)
2108
2109  =item truncate
2110
2111  Not implemented. (Older versions of VMS)
2112
2113  Truncation to same-or-shorter lengths only. (VOS)
2114
2115  If a FILEHANDLE is supplied, it must be writable and opened in append
2116  mode (i.e., use C<<< open(FH, '>>filename') >>>
2117  or C<sysopen(FH,...,O_APPEND|O_RDWR)>.  If a filename is supplied, it
2118  should not be held open elsewhere. (Win32)
2119
2121
2122  Returns undef where unavailable, as of version 5.005.
2123
2124  C<umask> works but the correct permissions are set only when the file
2125  is finally closed. (AmigaOS)
2126
2127  =item utime
2128
2129  Only the modification time is updated. (S<BeOS>, S<Mac OS>, VMS, S<RISC OS>)
2130
2131  May not behave as expected.  Behavior depends on the C runtime
2132  library's implementation of utime(), and the filesystem being
2133  used.  The FAT filesystem typically does not support an "access
2134  time" field, and it may limit timestamps to a granularity of
2135  two seconds. (Win32)
2136
2137  =item wait
2138
2139  =item waitpid
2140
2141  Not implemented. (S<Mac OS>)
2142
2143  Can only be applied to process handles returned for processes spawned
2144  using C<system(1, ...)> or pseudo processes created with C<fork()>. (Win32)
2145
2146  Not useful. (S<RISC OS>)
2147
2148  =back
2149
2150
2152
2153  As of July 2002 (the Perl release 5.8.0), the following platforms are
2154  able to build Perl from the standard source code distribution
2155  available at http://www.cpan.org/src/index.html
2156
2157          AIX
2158          BeOS
2159          BSD/OS          (BSDi)
2160          Cygwin
2161          DG/UX
2162          DOS DJGPP       1)
2163          DYNIX/ptx
2164          EPOC R5
2165          FreeBSD
2166          HI-UXMPP        (Hitachi) (5.8.0 worked but we didn't know it)
2167          HP-UX
2168          IRIX
2169          Linux
2170          Mac OS Classic
2171          Mac OS X        (Darwin)
2172          MPE/iX
2173          NetBSD
2174          NetWare
2175          NonStop-UX
2176          ReliantUNIX     (formerly SINIX)
2177          OpenBSD
2178          OpenVMS         (formerly VMS)
2179          Open UNIX       (Unixware) (since Perl 5.8.1/5.9.0)
2180          OS/2
2181          OS/400          (using the PASE) (since Perl 5.8.1/5.9.0)
2182          PowerUX
2183          POSIX-BC        (formerly BS2000)
2184          QNX
2185          Solaris
2186          SunOS 4
2187          SUPER-UX        (NEC)
2188          Tru64 UNIX      (formerly DEC OSF/1, Digital UNIX)
2189          UNICOS
2190          UNICOS/mk
2191          UTS
2192          VOS
2193          Win95/98/ME/2K/XP 2)
2194          WinCE
2195          z/OS            (formerly OS/390)
2196          VM/ESA
2197
2198          1) in DOS mode either the DOS or OS/2 ports can be used
2199          2) compilers: Borland, MinGW (GCC), VC6
2200
2201  The following platforms worked with the previous releases (5.6 and
2202  5.7), but we did not manage either to fix or to test these in time
2203  for the 5.8.0 release.  There is a very good chance that many of these
2204  will work fine with the 5.8.0.
2205
2206          BSD/OS
2207          DomainOS
2208          Hurd
2209          LynxOS
2210          MachTen
2211          PowerMAX
2212          SCO SV
2213          SVR4
2214          Unixware
2215          Windows 3.1
2216
2217  Known to be broken for 5.8.0 (but 5.6.1 and 5.7.2 can be used):
2218
2219      AmigaOS
2220
2221  The following platforms have been known to build Perl from source in
2222  the past (5.005_03 and earlier), but we haven't been able to verify
2223  their status for the current release, either because the
2224  hardware/software platforms are rare or because we don't have an
2225  active champion on these platforms--or both.  They used to work,
2226  though, so go ahead and try compiling them, and let perlbug@perl.org
2227  of any trouble.
2228
2229          3b1
2230          A/UX
2231          ConvexOS
2232          CX/UX
2233          DC/OSx
2234          DDE SMES
2235          DOS EMX
2236          Dynix
2237          EP/IX
2238          ESIX
2239          FPS
2240          GENIX
2241          Greenhills
2242          ISC
2243          MachTen 68k
2244          MiNT
2245          MPC
2246          NEWS-OS
2247          NextSTEP
2248          OpenSTEP
2249          Opus
2250          Plan 9
2251          RISC/os
2252          SCO ODT/OSR
2253          Stellar
2254          SVR2
2255          TI1500
2256          TitanOS
2257          Ultrix
2258          Unisys Dynix
2259
2260  The following platforms have their own source code distributions and
2261  binaries available via http://www.cpan.org/ports/
2262
2263                                  Perl release
2264
2265          OS/400 (ILE)            5.005_02
2266          Tandem Guardian         5.004
2267
2268  The following platforms have only binaries available via
2269  http://www.cpan.org/ports/index.html :
2270
2271                                  Perl release
2272
2273          Acorn RISCOS            5.005_02
2274          AOS                     5.002
2275          LynxOS                  5.004_02
2276
2277  Although we do suggest that you always build your own Perl from
2278  the source code, both for maximal configurability and for security,
2279  in case you are in a hurry you can check
2280  http://www.cpan.org/ports/index.html for binary distributions.
2281
2283
2284  L<perlaix>, L<perlamiga>, L<perlapollo>, L<perlbeos>, L<perlbs2000>,
2285  L<perlce>, L<perlcygwin>, L<perldgux>, L<perldos>, L<perlepoc>,
2286  L<perlebcdic>, L<perlfreebsd>, L<perlhurd>, L<perlhpux>, L<perlirix>,
2287  L<perlmachten>, L<perlmacos>, L<perlmacosx>, L<perlmint>, L<perlmpeix>,
2288  L<perlnetware>, L<perlos2>, L<perlos390>, L<perlos400>,
2289  L<perlplan9>, L<perlqnx>, L<perlsolaris>, L<perltru64>,
2290  L<perlunicode>, L<perlvmesa>, L<perlvms>, L<perlvos>,
2291  L<perlwin32>, and L<Win32>.
2292
2294
2296  Charles Bailey <bailey@newman.upenn.edu>,
2297  Graham Barr <gbarr@pobox.com>,
2298  Tom Christiansen <tchrist@perl.com>,
2299  Nicholas Clark <nick@ccl4.org>,
2300  Thomas Dorner <Thomas.Dorner@start.de>,
2301  Andy Dougherty <doughera@lafayette.edu>,
2302  Dominic Dunlop <domo@computer.org>,
2303  Neale Ferguson <neale@vma.tabnsw.com.au>,
2304  David J. Fiander <davidf@mks.com>,
2305  Paul Green <Paul.Green@stratus.com>,
2306  M.J.T. Guy <mjtg@cam.ac.uk>,
2307  Jarkko Hietaniemi <jhi@iki.fi>,
2308  Luther Huffman <lutherh@stratcom.com>,
2309  Nick Ing-Simmons <nick@ing-simmons.net>,
2310  Andreas J. KE<ouml>nig <a.koenig@mind.de>,
2311  Markus Laker <mlaker@contax.co.uk>,
2313  Larry Moore <ljmoore@freespace.net>,
2314  Paul Moore <Paul.Moore@uk.origin-it.com>,
2315  Chris Nandor <pudge@pobox.com>,
2316  Matthias Neeracher <neeracher@mac.com>,
2317  Philip Newton <pne@cpan.org>,
2318  Gary Ng <71564.1743@CompuServe.COM>,
2319  Tom Phoenix <rootbeer@teleport.com>,
2320  AndrE<eacute> Pirard <A.Pirard@ulg.ac.be>,
2321  Peter Prymmer <pvhp@forte.com>,
2322  Hugo van der Sanden <hv@crypt0.demon.co.uk>,
2323  Gurusamy Sarathy <gsar@activestate.com>,
2324  Paul J. Schinder <schinder@pobox.com>,
2325  Michael G Schwern <schwern@pobox.com>,
2326  Dan Sugalski <dan@sidhe.org>,
2327  Nathan Torkington <gnat@frii.com>.
2328  John Malmberg <wb8tyw@qsl.net>


 Generated: Tue Mar 17 22:47:18 2015 Cross-referenced by PHPXref 0.7.1