| [ Index ] |
PHP Cross Reference of Unnamed Project |
[Summary view] [Print] [Text view]
1 =head1 NAME 2 3 perlport - Writing portable Perl 4 5 =head1 DESCRIPTION 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 67 deliberate in your decision. 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 80 =head1 ISSUES 81 82 =head2 Newlines 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.) 391 392 =head2 System Interaction 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 421 platforms remove just the most recent version, too). The portable 422 idiom to remove all the versions of a file is 423 424 1 while unlink "file"; 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()>, 572 C<inet_ntoa()>, and C<sockaddr_in()>. 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). 577 578 =head2 External Subroutines (XS) 579 580 XS code can usually be made to work with any platform, but dependent 581 libraries, header files, etc., might not be readily available or 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. 591 592 =head2 Standard Modules 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. 609 610 =head2 Time and Date 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". 663 664 =head2 Internationalisation 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. 674 See L<perluniintro> and L<perlunicode> for more information. 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. 688 689 =head2 System Resources 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. 708 709 =head2 Security 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<$)>) to work 733 for switching identities (or memberships). 734 735 Don't assume set-uid and set-gid semantics. (And even if you do, 736 think twice: set-uid and set-gid are a known can of security worms.) 737 738 =head2 Style 739 740 For those times when it is necessary to have platform-specific code, 741 consider keeping the platform-specific code in one place, making porting 742 to other platforms easier. Use the Config module and the special 743 variable C<$^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 1053 =head2 VMS 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 1232 F<README.vms> (installed as L<README_vms>), L<perlvms> 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 1244 =head2 VOS 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 1500 =head2 Other perls 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 1728 =item getlogin 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 1768 =item getnetbyaddr 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 1894 =item link 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 1905 to create hard links. 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 1946 =item readlink 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 1994 =item shmread 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 2120 =item umask 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 2151 =head1 Supported Platforms 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 2282 =head1 SEE ALSO 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 2293 =head1 AUTHORS / CONTRIBUTORS 2294 2295 Abigail <abigail@foad.org>, 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>, 2312 Andrew M. Langmead <aml@world.std.com>, 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>
title
Description
Body
title
Description
Body
title
Description
Body
title
Body
| Generated: Tue Mar 17 22:47:18 2015 | Cross-referenced by PHPXref 0.7.1 |