| 1 | If you read this file _as_is_, just ignore the funny characters you |
| 2 | see. It is written in the POD format (see pod/perlpod.pod) which is |
| 3 | specially designed to be readable as is. |
| 4 | |
| 5 | =head1 NAME |
| 6 | |
| 7 | perlwin32 - Perl under Windows |
| 8 | |
| 9 | =head1 SYNOPSIS |
| 10 | |
| 11 | These are instructions for building Perl under Windows 9x/NT/2000/XP |
| 12 | on the Intel x86 and Itanium architectures. |
| 13 | |
| 14 | =head1 DESCRIPTION |
| 15 | |
| 16 | Before you start, you should glance through the README file |
| 17 | found in the top-level directory to which the Perl distribution |
| 18 | was extracted. Make sure you read and understand the terms under |
| 19 | which this software is being distributed. |
| 20 | |
| 21 | Also make sure you read L<BUGS AND CAVEATS> below for the |
| 22 | known limitations of this port. |
| 23 | |
| 24 | The INSTALL file in the perl top-level has much information that is |
| 25 | only relevant to people building Perl on Unix-like systems. In |
| 26 | particular, you can safely ignore any information that talks about |
| 27 | "Configure". |
| 28 | |
| 29 | You may also want to look at two other options for building |
| 30 | a perl that will work on Windows NT: the README.cygwin and |
| 31 | README.os2 files, each of which give a different set of rules to |
| 32 | build a Perl that will work on Win32 platforms. Those two methods |
| 33 | will probably enable you to build a more Unix-compatible perl, but |
| 34 | you will also need to download and use various other build-time and |
| 35 | run-time support software described in those files. |
| 36 | |
| 37 | This set of instructions is meant to describe a so-called "native" |
| 38 | port of Perl to Win32 platforms. This includes both 32-bit and |
| 39 | 64-bit Windows operating systems. The resulting Perl requires no |
| 40 | additional software to run (other than what came with your operating |
| 41 | system). Currently, this port is capable of using one of the |
| 42 | following compilers on the Intel x86 architecture: |
| 43 | |
| 44 | Borland C++ version 5.02 or later |
| 45 | Microsoft Visual C++ version 4.2 or later |
| 46 | Mingw32 with GCC version 2.95.2 or better |
| 47 | |
| 48 | The last of these is a high quality freeware compiler. Support |
| 49 | for it is still experimental. (Older versions of GCC are known |
| 50 | not to work.) |
| 51 | |
| 52 | This port can also be built on the Intel IA64 using: |
| 53 | |
| 54 | Microsoft Platform SDK Nov 2001 (64-bit compiler and tools) |
| 55 | |
| 56 | The MS Platform SDK can be downloaded from http://www.microsoft.com/. |
| 57 | |
| 58 | This port fully supports MakeMaker (the set of modules that |
| 59 | is used to build extensions to perl). Therefore, you should be |
| 60 | able to build and install most extensions found in the CPAN sites. |
| 61 | See L<Usage Hints for Perl on Win32> below for general hints about this. |
| 62 | |
| 63 | =head2 Setting Up Perl on Win32 |
| 64 | |
| 65 | =over 4 |
| 66 | |
| 67 | =item Make |
| 68 | |
| 69 | You need a "make" program to build the sources. If you are using |
| 70 | Visual C++ or the Platform SDK tools under Windows NT/2000/XP, nmake |
| 71 | will work. All other builds need dmake. |
| 72 | |
| 73 | dmake is a freely available make that has very nice macro features |
| 74 | and parallelability. |
| 75 | |
| 76 | A port of dmake for Windows is available from: |
| 77 | |
| 78 | http://www.cpan.org/authors/id/GSAR/dmake-4.1pl1-win32.zip |
| 79 | |
| 80 | (This is a fixed version of the original dmake sources obtained from |
| 81 | http://www.wticorp.com/ As of version 4.1PL1, the original |
| 82 | sources did not build as shipped and had various other problems. |
| 83 | A patch is included in the above fixed version.) |
| 84 | |
| 85 | Fetch and install dmake somewhere on your path (follow the instructions |
| 86 | in the README.NOW file). |
| 87 | |
| 88 | There exists a minor coexistence problem with dmake and Borland C++ |
| 89 | compilers. Namely, if a distribution has C files named with mixed |
| 90 | case letters, they will be compiled into appropriate .obj-files named |
| 91 | with all lowercase letters, and every time dmake is invoked |
| 92 | to bring files up to date, it will try to recompile such files again. |
| 93 | For example, Tk distribution has a lot of such files, resulting in |
| 94 | needless recompiles every time dmake is invoked. To avoid this, you |
| 95 | may use the script "sync_ext.pl" after a successful build. It is |
| 96 | available in the win32 subdirectory of the Perl source distribution. |
| 97 | |
| 98 | =item Command Shell |
| 99 | |
| 100 | Use the default "cmd" shell that comes with NT. Some versions of the |
| 101 | popular 4DOS/NT shell have incompatibilities that may cause you trouble. |
| 102 | If the build fails under that shell, try building again with the cmd |
| 103 | shell. |
| 104 | |
| 105 | The nmake Makefile also has known incompatibilities with the |
| 106 | "command.com" shell that comes with Windows 9x. You will need to |
| 107 | use dmake and makefile.mk to build under Windows 9x. |
| 108 | |
| 109 | The surest way to build it is on Windows NT/2000/XP, using the cmd shell. |
| 110 | |
| 111 | Make sure the path to the build directory does not contain spaces. The |
| 112 | build usually works in this circumstance, but some tests will fail. |
| 113 | |
| 114 | =item Borland C++ |
| 115 | |
| 116 | If you are using the Borland compiler, you will need dmake. |
| 117 | (The make that Borland supplies is seriously crippled and will not |
| 118 | work for MakeMaker builds.) |
| 119 | |
| 120 | See L</"Make"> above. |
| 121 | |
| 122 | =item Microsoft Visual C++ |
| 123 | |
| 124 | The nmake that comes with Visual C++ will suffice for building. |
| 125 | You will need to run the VCVARS32.BAT file, usually found somewhere |
| 126 | like C:\MSDEV4.2\BIN. This will set your build environment. |
| 127 | |
| 128 | You can also use dmake to build using Visual C++; provided, however, |
| 129 | you set OSRELEASE to "microsft" (or whatever the directory name |
| 130 | under which the Visual C dmake configuration lives) in your environment |
| 131 | and edit win32/config.vc to change "make=nmake" into "make=dmake". The |
| 132 | latter step is only essential if you want to use dmake as your default |
| 133 | make for building extensions using MakeMaker. |
| 134 | |
| 135 | =item Microsoft Platform SDK 64-bit Compiler |
| 136 | |
| 137 | The nmake that comes with the Platform SDK will suffice for building |
| 138 | Perl. Make sure you are building within one of the "Build Environment" |
| 139 | shells available after you install the Platform SDK from the Start Menu. |
| 140 | |
| 141 | =item Mingw32 with GCC |
| 142 | |
| 143 | GCC-2.95.2 binaries can be downloaded from: |
| 144 | |
| 145 | ftp://ftp.xraylith.wisc.edu/pub/khan/gnu-win32/mingw32/ |
| 146 | |
| 147 | You also need dmake. See L</"Make"> above on how to get it. |
| 148 | |
| 149 | The GCC-2.95.2 bundle comes with Mingw32 libraries and headers. |
| 150 | |
| 151 | Make sure you install the binaries that work with MSVCRT.DLL as indicated |
| 152 | in the README for the GCC bundle. You may need to set up a few environment |
| 153 | variables (usually ran from a batch file). |
| 154 | |
| 155 | There are a couple of problems with the version of gcc-2.95.2-msvcrt.exe |
| 156 | released 7 November 1999: |
| 157 | |
| 158 | =over |
| 159 | |
| 160 | =item * |
| 161 | |
| 162 | It left out a fix for certain command line quotes. To fix this, be sure |
| 163 | to download and install the file fixes/quote-fix-msvcrt.exe from the above |
| 164 | ftp location. |
| 165 | |
| 166 | =item * |
| 167 | |
| 168 | The definition of the fpos_t type in stdio.h may be wrong. If your |
| 169 | stdio.h has this problem, you will see an exception when running the |
| 170 | test t/lib/io_xs.t. To fix this, change the typedef for fpos_t from |
| 171 | "long" to "long long" in the file i386-mingw32msvc/include/stdio.h, |
| 172 | and rebuild. |
| 173 | |
| 174 | =back |
| 175 | |
| 176 | A potentially simpler to install (but probably soon-to-be-outdated) bundle |
| 177 | of the above package with the mentioned fixes already applied is available |
| 178 | here: |
| 179 | |
| 180 | http://downloads.ActiveState.com/pub/staff/gsar/gcc-2.95.2-msvcrt.zip |
| 181 | ftp://ftp.ActiveState.com/pub/staff/gsar/gcc-2.95.2-msvcrt.zip |
| 182 | |
| 183 | =back |
| 184 | |
| 185 | =head2 Building |
| 186 | |
| 187 | =over 4 |
| 188 | |
| 189 | =item * |
| 190 | |
| 191 | Make sure you are in the "win32" subdirectory under the perl toplevel. |
| 192 | This directory contains a "Makefile" that will work with |
| 193 | versions of nmake that come with Visual C++ or the Platform SDK, and |
| 194 | a dmake "makefile.mk" that will work for all supported compilers. The |
| 195 | defaults in the dmake makefile are setup to build using the GCC compiler. |
| 196 | |
| 197 | =item * |
| 198 | |
| 199 | Edit the makefile.mk (or Makefile, if you're using nmake) and change |
| 200 | the values of INST_DRV and INST_TOP. You can also enable various |
| 201 | build flags. These are explained in the makefiles. |
| 202 | |
| 203 | You will have to make sure that CCTYPE is set correctly and that |
| 204 | CCHOME points to wherever you installed your compiler. |
| 205 | |
| 206 | The default value for CCHOME in the makefiles for Visual C++ |
| 207 | may not be correct for some versions. Make sure the default exists |
| 208 | and is valid. |
| 209 | |
| 210 | If you have either the source or a library that contains des_fcrypt(), |
| 211 | enable the appropriate option in the makefile. des_fcrypt() is not |
| 212 | bundled with the distribution due to US Government restrictions |
| 213 | on the export of cryptographic software. Nevertheless, this routine |
| 214 | is part of the "libdes" library (written by Eric Young) which is widely |
| 215 | available worldwide, usually along with SSLeay ( for example, |
| 216 | ftp://ftp.funet.fi/pub/crypt/mirrors/dsi/libdes/ ). Set CRYPT_SRC to the |
| 217 | name of the file that implements des_fcrypt(). Alternatively, if |
| 218 | you have built a library that contains des_fcrypt(), you can set |
| 219 | CRYPT_LIB to point to the library name. The location above contains |
| 220 | many versions of the "libdes" library, all with slightly different |
| 221 | implementations of des_fcrypt(). Older versions have a single, |
| 222 | self-contained file (fcrypt.c) that implements crypt(), so they may be |
| 223 | easier to use. A patch against the fcrypt.c found in libdes-3.06 is |
| 224 | in des_fcrypt.patch. |
| 225 | |
| 226 | An easier alternative may be to get the pre-patched and ready-to-use |
| 227 | fcrypt.c that can be found here: |
| 228 | |
| 229 | http://downloads.ActiveState.com/pub/staff/gsar/fcrypt.c |
| 230 | ftp://ftp.ActiveState.com/pub/staff/gsar/fcrypt.c |
| 231 | |
| 232 | Perl will also build without des_fcrypt(), but the crypt() builtin will |
| 233 | fail at run time. |
| 234 | |
| 235 | Be sure to read the instructions near the top of the makefiles carefully. |
| 236 | |
| 237 | =item * |
| 238 | |
| 239 | Type "dmake" (or "nmake" if you are using that make). |
| 240 | |
| 241 | This should build everything. Specifically, it will create perl.exe, |
| 242 | perl58.dll at the perl toplevel, and various other extension dll's |
| 243 | under the lib\auto directory. If the build fails for any reason, make |
| 244 | sure you have done the previous steps correctly. |
| 245 | |
| 246 | =back |
| 247 | |
| 248 | =head2 Testing Perl on Win32 |
| 249 | |
| 250 | Type "dmake test" (or "nmake test"). This will run most of the tests from |
| 251 | the testsuite (many tests will be skipped). |
| 252 | |
| 253 | There should be no test failures when running under Windows NT/2000/XP. |
| 254 | Many tests I<will> fail under Windows 9x due to the inferior command shell. |
| 255 | |
| 256 | Some test failures may occur if you use a command shell other than the |
| 257 | native "cmd.exe", or if you are building from a path that contains |
| 258 | spaces. So don't do that. |
| 259 | |
| 260 | If you are running the tests from a emacs shell window, you may see |
| 261 | failures in op/stat.t. Run "dmake test-notty" in that case. |
| 262 | |
| 263 | If you're using the Borland compiler, you may see a failure in op/taint.t |
| 264 | arising from the inability to find the Borland Runtime DLLs on the system |
| 265 | default path. You will need to copy the DLLs reported by the messages |
| 266 | from where Borland chose to install it, into the Windows system directory |
| 267 | (usually somewhere like C:\WINNT\SYSTEM32) and rerun the test. |
| 268 | |
| 269 | If you're using Borland compiler versions 5.2 and below, you may run into |
| 270 | problems finding the correct header files when building extensions. For |
| 271 | example, building the "Tk" extension may fail because both perl and Tk |
| 272 | contain a header file called "patchlevel.h". The latest Borland compiler |
| 273 | (v5.5) is free of this misbehaviour, and it even supports an |
| 274 | option -VI- for backward (bugward) compatibility for using the old Borland |
| 275 | search algorithm to locate header files. |
| 276 | |
| 277 | Please report any other failures as described under L<BUGS AND CAVEATS>. |
| 278 | |
| 279 | =head2 Installation of Perl on Win32 |
| 280 | |
| 281 | Type "dmake install" (or "nmake install"). This will put the newly |
| 282 | built perl and the libraries under whatever C<INST_TOP> points to in the |
| 283 | Makefile. It will also install the pod documentation under |
| 284 | C<$INST_TOP\$VERSION\lib\pod> and HTML versions of the same under |
| 285 | C<$INST_TOP\$VERSION\lib\pod\html>. To use the Perl you just installed, |
| 286 | you will need to add two components to your PATH environment variable, |
| 287 | C<$INST_TOP\$VERSION\bin> and C<$INST_TOP\$VERSION\bin\$ARCHNAME>. |
| 288 | For example: |
| 289 | |
| 290 | set PATH c:\perl\5.6.0\bin;c:\perl\5.6.0\bin\MSWin32-x86;%PATH% |
| 291 | |
| 292 | If you opt to comment out INST_VER and INST_ARCH in the makefiles, the |
| 293 | installation structure is much simpler. In that case, it will be |
| 294 | sufficient to add a single entry to the path, for instance: |
| 295 | |
| 296 | set PATH c:\perl\bin;%PATH% |
| 297 | |
| 298 | =head2 Usage Hints for Perl on Win32 |
| 299 | |
| 300 | =over 4 |
| 301 | |
| 302 | =item Environment Variables |
| 303 | |
| 304 | The installation paths that you set during the build get compiled |
| 305 | into perl, so you don't have to do anything additional to start |
| 306 | using that perl (except add its location to your PATH variable). |
| 307 | |
| 308 | If you put extensions in unusual places, you can set PERL5LIB |
| 309 | to a list of paths separated by semicolons where you want perl |
| 310 | to look for libraries. Look for descriptions of other environment |
| 311 | variables you can set in L<perlrun>. |
| 312 | |
| 313 | You can also control the shell that perl uses to run system() and |
| 314 | backtick commands via PERL5SHELL. See L<perlrun>. |
| 315 | |
| 316 | Perl does not depend on the registry, but it can look up certain default |
| 317 | values if you choose to put them there. Perl attempts to read entries from |
| 318 | C<HKEY_CURRENT_USER\Software\Perl> and C<HKEY_LOCAL_MACHINE\Software\Perl>. |
| 319 | Entries in the former override entries in the latter. One or more of the |
| 320 | following entries (of type REG_SZ or REG_EXPAND_SZ) may be set: |
| 321 | |
| 322 | lib-$] version-specific standard library path to add to @INC |
| 323 | lib standard library path to add to @INC |
| 324 | sitelib-$] version-specific site library path to add to @INC |
| 325 | sitelib site library path to add to @INC |
| 326 | vendorlib-$] version-specific vendor library path to add to @INC |
| 327 | vendorlib vendor library path to add to @INC |
| 328 | PERL* fallback for all %ENV lookups that begin with "PERL" |
| 329 | |
| 330 | Note the C<$]> in the above is not literal. Substitute whatever version |
| 331 | of perl you want to honor that entry, e.g. C<5.6.0>. Paths must be |
| 332 | separated with semicolons, as usual on win32. |
| 333 | |
| 334 | =item File Globbing |
| 335 | |
| 336 | By default, perl handles file globbing using the File::Glob extension, |
| 337 | which provides portable globbing. |
| 338 | |
| 339 | If you want perl to use globbing that emulates the quirks of DOS |
| 340 | filename conventions, you might want to consider using File::DosGlob |
| 341 | to override the internal glob() implementation. See L<File::DosGlob> for |
| 342 | details. |
| 343 | |
| 344 | =item Using perl from the command line |
| 345 | |
| 346 | If you are accustomed to using perl from various command-line |
| 347 | shells found in UNIX environments, you will be less than pleased |
| 348 | with what Windows offers by way of a command shell. |
| 349 | |
| 350 | The crucial thing to understand about the Windows environment is that |
| 351 | the command line you type in is processed twice before Perl sees it. |
| 352 | First, your command shell (usually CMD.EXE on Windows NT, and |
| 353 | COMMAND.COM on Windows 9x) preprocesses the command line, to handle |
| 354 | redirection, environment variable expansion, and location of the |
| 355 | executable to run. Then, the perl executable splits the remaining |
| 356 | command line into individual arguments, using the C runtime library |
| 357 | upon which Perl was built. |
| 358 | |
| 359 | It is particularly important to note that neither the shell nor the C |
| 360 | runtime do any wildcard expansions of command-line arguments (so |
| 361 | wildcards need not be quoted). Also, the quoting behaviours of the |
| 362 | shell and the C runtime are rudimentary at best (and may, if you are |
| 363 | using a non-standard shell, be inconsistent). The only (useful) quote |
| 364 | character is the double quote ("). It can be used to protect spaces |
| 365 | and other special characters in arguments. |
| 366 | |
| 367 | The Windows NT documentation has almost no description of how the |
| 368 | quoting rules are implemented, but here are some general observations |
| 369 | based on experiments: The C runtime breaks arguments at spaces and |
| 370 | passes them to programs in argc/argv. Double quotes can be used to |
| 371 | prevent arguments with spaces in them from being split up. You can |
| 372 | put a double quote in an argument by escaping it with a backslash and |
| 373 | enclosing the whole argument within double quotes. The backslash and |
| 374 | the pair of double quotes surrounding the argument will be stripped by |
| 375 | the C runtime. |
| 376 | |
| 377 | The file redirection characters "<", ">", and "|" can be quoted by |
| 378 | double quotes (although there are suggestions that this may not always |
| 379 | be true). Single quotes are not treated as quotes by the shell or |
| 380 | the C runtime, they don't get stripped by the shell (just to make |
| 381 | this type of quoting completely useless). The caret "^" has also |
| 382 | been observed to behave as a quoting character, but this appears |
| 383 | to be a shell feature, and the caret is not stripped from the command |
| 384 | line, so Perl still sees it (and the C runtime phase does not treat |
| 385 | the caret as a quote character). |
| 386 | |
| 387 | Here are some examples of usage of the "cmd" shell: |
| 388 | |
| 389 | This prints two doublequotes: |
| 390 | |
| 391 | perl -e "print '\"\"' " |
| 392 | |
| 393 | This does the same: |
| 394 | |
| 395 | perl -e "print \"\\\"\\\"\" " |
| 396 | |
| 397 | This prints "bar" and writes "foo" to the file "blurch": |
| 398 | |
| 399 | perl -e "print 'foo'; print STDERR 'bar'" > blurch |
| 400 | |
| 401 | This prints "foo" ("bar" disappears into nowhereland): |
| 402 | |
| 403 | perl -e "print 'foo'; print STDERR 'bar'" 2> nul |
| 404 | |
| 405 | This prints "bar" and writes "foo" into the file "blurch": |
| 406 | |
| 407 | perl -e "print 'foo'; print STDERR 'bar'" 1> blurch |
| 408 | |
| 409 | This pipes "foo" to the "less" pager and prints "bar" on the console: |
| 410 | |
| 411 | perl -e "print 'foo'; print STDERR 'bar'" | less |
| 412 | |
| 413 | This pipes "foo\nbar\n" to the less pager: |
| 414 | |
| 415 | perl -le "print 'foo'; print STDERR 'bar'" 2>&1 | less |
| 416 | |
| 417 | This pipes "foo" to the pager and writes "bar" in the file "blurch": |
| 418 | |
| 419 | perl -e "print 'foo'; print STDERR 'bar'" 2> blurch | less |
| 420 | |
| 421 | |
| 422 | Discovering the usefulness of the "command.com" shell on Windows 9x |
| 423 | is left as an exercise to the reader :) |
| 424 | |
| 425 | One particularly pernicious problem with the 4NT command shell for |
| 426 | Windows NT is that it (nearly) always treats a % character as indicating |
| 427 | that environment variable expansion is needed. Under this shell, it is |
| 428 | therefore important to always double any % characters which you want |
| 429 | Perl to see (for example, for hash variables), even when they are |
| 430 | quoted. |
| 431 | |
| 432 | =item Building Extensions |
| 433 | |
| 434 | The Comprehensive Perl Archive Network (CPAN) offers a wealth |
| 435 | of extensions, some of which require a C compiler to build. |
| 436 | Look in http://www.cpan.org/ for more information on CPAN. |
| 437 | |
| 438 | Note that not all of the extensions available from CPAN may work |
| 439 | in the Win32 environment; you should check the information at |
| 440 | http://testers.cpan.org/ before investing too much effort into |
| 441 | porting modules that don't readily build. |
| 442 | |
| 443 | Most extensions (whether they require a C compiler or not) can |
| 444 | be built, tested and installed with the standard mantra: |
| 445 | |
| 446 | perl Makefile.PL |
| 447 | $MAKE |
| 448 | $MAKE test |
| 449 | $MAKE install |
| 450 | |
| 451 | where $MAKE is whatever 'make' program you have configured perl to |
| 452 | use. Use "perl -V:make" to find out what this is. Some extensions |
| 453 | may not provide a testsuite (so "$MAKE test" may not do anything or |
| 454 | fail), but most serious ones do. |
| 455 | |
| 456 | It is important that you use a supported 'make' program, and |
| 457 | ensure Config.pm knows about it. If you don't have nmake, you can |
| 458 | either get dmake from the location mentioned earlier or get an |
| 459 | old version of nmake reportedly available from: |
| 460 | |
| 461 | ftp://ftp.microsoft.com/Softlib/MSLFILES/nmake15.exe |
| 462 | |
| 463 | Another option is to use the make written in Perl, available from |
| 464 | CPAN. |
| 465 | |
| 466 | http://www.cpan.org/modules/by-module/Make/ |
| 467 | |
| 468 | You may also use dmake. See L</"Make"> above on how to get it. |
| 469 | |
| 470 | Note that MakeMaker actually emits makefiles with different syntax |
| 471 | depending on what 'make' it thinks you are using. Therefore, it is |
| 472 | important that one of the following values appears in Config.pm: |
| 473 | |
| 474 | make='nmake' # MakeMaker emits nmake syntax |
| 475 | make='dmake' # MakeMaker emits dmake syntax |
| 476 | any other value # MakeMaker emits generic make syntax |
| 477 | (e.g GNU make, or Perl make) |
| 478 | |
| 479 | If the value doesn't match the 'make' program you want to use, |
| 480 | edit Config.pm to fix it. |
| 481 | |
| 482 | If a module implements XSUBs, you will need one of the supported |
| 483 | C compilers. You must make sure you have set up the environment for |
| 484 | the compiler for command-line compilation. |
| 485 | |
| 486 | If a module does not build for some reason, look carefully for |
| 487 | why it failed, and report problems to the module author. If |
| 488 | it looks like the extension building support is at fault, report |
| 489 | that with full details of how the build failed using the perlbug |
| 490 | utility. |
| 491 | |
| 492 | =item Command-line Wildcard Expansion |
| 493 | |
| 494 | The default command shells on DOS descendant operating systems (such |
| 495 | as they are) usually do not expand wildcard arguments supplied to |
| 496 | programs. They consider it the application's job to handle that. |
| 497 | This is commonly achieved by linking the application (in our case, |
| 498 | perl) with startup code that the C runtime libraries usually provide. |
| 499 | However, doing that results in incompatible perl versions (since the |
| 500 | behavior of the argv expansion code differs depending on the |
| 501 | compiler, and it is even buggy on some compilers). Besides, it may |
| 502 | be a source of frustration if you use such a perl binary with an |
| 503 | alternate shell that *does* expand wildcards. |
| 504 | |
| 505 | Instead, the following solution works rather well. The nice things |
| 506 | about it are 1) you can start using it right away; 2) it is more |
| 507 | powerful, because it will do the right thing with a pattern like |
| 508 | */*/*.c; 3) you can decide whether you do/don't want to use it; and |
| 509 | 4) you can extend the method to add any customizations (or even |
| 510 | entirely different kinds of wildcard expansion). |
| 511 | |
| 512 | C:\> copy con c:\perl\lib\Wild.pm |
| 513 | # Wild.pm - emulate shell @ARGV expansion on shells that don't |
| 514 | use File::DosGlob; |
| 515 | @ARGV = map { |
| 516 | my @g = File::DosGlob::glob($_) if /[*?]/; |
| 517 | @g ? @g : $_; |
| 518 | } @ARGV; |
| 519 | 1; |
| 520 | ^Z |
| 521 | C:\> set PERL5OPT=-MWild |
| 522 | C:\> perl -le "for (@ARGV) { print }" */*/perl*.c |
| 523 | p4view/perl/perl.c |
| 524 | p4view/perl/perlio.c |
| 525 | p4view/perl/perly.c |
| 526 | perl5.005/win32/perlglob.c |
| 527 | perl5.005/win32/perllib.c |
| 528 | perl5.005/win32/perlglob.c |
| 529 | perl5.005/win32/perllib.c |
| 530 | perl5.005/win32/perlglob.c |
| 531 | perl5.005/win32/perllib.c |
| 532 | |
| 533 | Note there are two distinct steps there: 1) You'll have to create |
| 534 | Wild.pm and put it in your perl lib directory. 2) You'll need to |
| 535 | set the PERL5OPT environment variable. If you want argv expansion |
| 536 | to be the default, just set PERL5OPT in your default startup |
| 537 | environment. |
| 538 | |
| 539 | If you are using the Visual C compiler, you can get the C runtime's |
| 540 | command line wildcard expansion built into perl binary. The resulting |
| 541 | binary will always expand unquoted command lines, which may not be |
| 542 | what you want if you use a shell that does that for you. The expansion |
| 543 | done is also somewhat less powerful than the approach suggested above. |
| 544 | |
| 545 | =item Win32 Specific Extensions |
| 546 | |
| 547 | A number of extensions specific to the Win32 platform are available |
| 548 | from CPAN. You may find that many of these extensions are meant to |
| 549 | be used under the Activeware port of Perl, which used to be the only |
| 550 | native port for the Win32 platform. Since the Activeware port does not |
| 551 | have adequate support for Perl's extension building tools, these |
| 552 | extensions typically do not support those tools either and, therefore, |
| 553 | cannot be built using the generic steps shown in the previous section. |
| 554 | |
| 555 | To ensure smooth transitioning of existing code that uses the |
| 556 | ActiveState port, there is a bundle of Win32 extensions that contains |
| 557 | all of the ActiveState extensions and most other Win32 extensions from |
| 558 | CPAN in source form, along with many added bugfixes, and with MakeMaker |
| 559 | support. This bundle is available at: |
| 560 | |
| 561 | http://www.cpan.org/authors/id/GSAR/libwin32-0.18.zip |
| 562 | |
| 563 | See the README in that distribution for building and installation |
| 564 | instructions. Look for later versions that may be available at the |
| 565 | same location. |
| 566 | |
| 567 | =item Notes on 64-bit Windows |
| 568 | |
| 569 | Windows .NET Server supports the LLP64 data model on the Intel Itanium |
| 570 | architecture. |
| 571 | |
| 572 | The LLP64 data model is different from the LP64 data model that is the |
| 573 | norm on 64-bit Unix platforms. In the former, C<int> and C<long> are |
| 574 | both 32-bit data types, while pointers are 64 bits wide. In addition, |
| 575 | there is a separate 64-bit wide integral type, C<__int64>. In contrast, |
| 576 | the LP64 data model that is pervasive on Unix platforms provides C<int> |
| 577 | as the 32-bit type, while both the C<long> type and pointers are of |
| 578 | 64-bit precision. Note that both models provide for 64-bits of |
| 579 | addressability. |
| 580 | |
| 581 | 64-bit Windows running on Itanium is capable of running 32-bit x86 |
| 582 | binaries transparently. This means that you could use a 32-bit build |
| 583 | of Perl on a 64-bit system. Given this, why would one want to build |
| 584 | a 64-bit build of Perl? Here are some reasons why you would bother: |
| 585 | |
| 586 | =item * |
| 587 | |
| 588 | A 64-bit native application will run much more efficiently on |
| 589 | Itanium hardware. |
| 590 | |
| 591 | =item * |
| 592 | |
| 593 | There is no 2GB limit on process size. |
| 594 | |
| 595 | =item * |
| 596 | |
| 597 | Perl automatically provides large file support when built under |
| 598 | 64-bit Windows. |
| 599 | |
| 600 | =item * |
| 601 | |
| 602 | Embedding Perl inside a 64-bit application. |
| 603 | |
| 604 | =back |
| 605 | |
| 606 | =item Running Perl Scripts |
| 607 | |
| 608 | Perl scripts on UNIX use the "#!" (a.k.a "shebang") line to |
| 609 | indicate to the OS that it should execute the file using perl. |
| 610 | Win32 has no comparable means to indicate arbitrary files are |
| 611 | executables. |
| 612 | |
| 613 | Instead, all available methods to execute plain text files on |
| 614 | Win32 rely on the file "extension". There are three methods |
| 615 | to use this to execute perl scripts: |
| 616 | |
| 617 | =over 8 |
| 618 | |
| 619 | =item 1 |
| 620 | |
| 621 | There is a facility called "file extension associations" that will |
| 622 | work in Windows NT 4.0. This can be manipulated via the two |
| 623 | commands "assoc" and "ftype" that come standard with Windows NT |
| 624 | 4.0. Type "ftype /?" for a complete example of how to set this |
| 625 | up for perl scripts (Say what? You thought Windows NT wasn't |
| 626 | perl-ready? :). |
| 627 | |
| 628 | =item 2 |
| 629 | |
| 630 | Since file associations don't work everywhere, and there are |
| 631 | reportedly bugs with file associations where it does work, the |
| 632 | old method of wrapping the perl script to make it look like a |
| 633 | regular batch file to the OS, may be used. The install process |
| 634 | makes available the "pl2bat.bat" script which can be used to wrap |
| 635 | perl scripts into batch files. For example: |
| 636 | |
| 637 | pl2bat foo.pl |
| 638 | |
| 639 | will create the file "FOO.BAT". Note "pl2bat" strips any |
| 640 | .pl suffix and adds a .bat suffix to the generated file. |
| 641 | |
| 642 | If you use the 4DOS/NT or similar command shell, note that |
| 643 | "pl2bat" uses the "%*" variable in the generated batch file to |
| 644 | refer to all the command line arguments, so you may need to make |
| 645 | sure that construct works in batch files. As of this writing, |
| 646 | 4DOS/NT users will need a "ParameterChar = *" statement in their |
| 647 | 4NT.INI file or will need to execute "setdos /p*" in the 4DOS/NT |
| 648 | startup file to enable this to work. |
| 649 | |
| 650 | =item 3 |
| 651 | |
| 652 | Using "pl2bat" has a few problems: the file name gets changed, |
| 653 | so scripts that rely on C<$0> to find what they must do may not |
| 654 | run properly; running "pl2bat" replicates the contents of the |
| 655 | original script, and so this process can be maintenance intensive |
| 656 | if the originals get updated often. A different approach that |
| 657 | avoids both problems is possible. |
| 658 | |
| 659 | A script called "runperl.bat" is available that can be copied |
| 660 | to any filename (along with the .bat suffix). For example, |
| 661 | if you call it "foo.bat", it will run the file "foo" when it is |
| 662 | executed. Since you can run batch files on Win32 platforms simply |
| 663 | by typing the name (without the extension), this effectively |
| 664 | runs the file "foo", when you type either "foo" or "foo.bat". |
| 665 | With this method, "foo.bat" can even be in a different location |
| 666 | than the file "foo", as long as "foo" is available somewhere on |
| 667 | the PATH. If your scripts are on a filesystem that allows symbolic |
| 668 | links, you can even avoid copying "runperl.bat". |
| 669 | |
| 670 | Here's a diversion: copy "runperl.bat" to "runperl", and type |
| 671 | "runperl". Explain the observed behavior, or lack thereof. :) |
| 672 | Hint: .gnidnats llits er'uoy fi ,"lrepnur" eteled :tniH |
| 673 | |
| 674 | =item Miscellaneous Things |
| 675 | |
| 676 | A full set of HTML documentation is installed, so you should be |
| 677 | able to use it if you have a web browser installed on your |
| 678 | system. |
| 679 | |
| 680 | C<perldoc> is also a useful tool for browsing information contained |
| 681 | in the documentation, especially in conjunction with a pager |
| 682 | like C<less> (recent versions of which have Win32 support). You may |
| 683 | have to set the PAGER environment variable to use a specific pager. |
| 684 | "perldoc -f foo" will print information about the perl operator |
| 685 | "foo". |
| 686 | |
| 687 | If you find bugs in perl, you can run C<perlbug> to create a |
| 688 | bug report (you may have to send it manually if C<perlbug> cannot |
| 689 | find a mailer on your system). |
| 690 | |
| 691 | =back |
| 692 | |
| 693 | =head1 BUGS AND CAVEATS |
| 694 | |
| 695 | Norton AntiVirus interferes with the build process, particularly if |
| 696 | set to "AutoProtect, All Files, when Opened". Unlike large applications |
| 697 | the perl build process opens and modifies a lot of files. Having the |
| 698 | the AntiVirus scan each and every one slows build the process significantly. |
| 699 | Worse, with PERLIO=stdio the build process fails with peculiar messages |
| 700 | as the virus checker interacts badly with miniperl.exe writing configure |
| 701 | files (it seems to either catch file part written and treat it as suspicious, |
| 702 | or virus checker may have it "locked" in a way which inhibits miniperl |
| 703 | updating it). The build does complete with |
| 704 | |
| 705 | set PERLIO=perlio |
| 706 | |
| 707 | but that may be just luck. Other AntiVirus software may have similar issues. |
| 708 | |
| 709 | Some of the built-in functions do not act exactly as documented in |
| 710 | L<perlfunc>, and a few are not implemented at all. To avoid |
| 711 | surprises, particularly if you have had prior exposure to Perl |
| 712 | in other operating environments or if you intend to write code |
| 713 | that will be portable to other environments. See L<perlport> |
| 714 | for a reasonably definitive list of these differences. |
| 715 | |
| 716 | Not all extensions available from CPAN may build or work properly |
| 717 | in the Win32 environment. See L</"Building Extensions">. |
| 718 | |
| 719 | Most C<socket()> related calls are supported, but they may not |
| 720 | behave as on Unix platforms. See L<perlport> for the full list. |
| 721 | |
| 722 | Signal handling may not behave as on Unix platforms (where it |
| 723 | doesn't exactly "behave", either :). For instance, calling C<die()> |
| 724 | or C<exit()> from signal handlers will cause an exception, since most |
| 725 | implementations of C<signal()> on Win32 are severely crippled. |
| 726 | Thus, signals may work only for simple things like setting a flag |
| 727 | variable in the handler. Using signals under this port should |
| 728 | currently be considered unsupported. |
| 729 | |
| 730 | Please send detailed descriptions of any problems and solutions that |
| 731 | you may find to <F<perlbug@perl.com>>, along with the output produced |
| 732 | by C<perl -V>. |
| 733 | |
| 734 | =head1 AUTHORS |
| 735 | |
| 736 | =over 4 |
| 737 | |
| 738 | =item Gary Ng E<lt>71564.1743@CompuServe.COME<gt> |
| 739 | |
| 740 | =item Gurusamy Sarathy E<lt>gsar@activestate.comE<gt> |
| 741 | |
| 742 | =item Nick Ing-Simmons E<lt>nick@ing-simmons.netE<gt> |
| 743 | |
| 744 | =back |
| 745 | |
| 746 | This document is maintained by Gurusamy Sarathy. |
| 747 | |
| 748 | =head1 SEE ALSO |
| 749 | |
| 750 | L<perl> |
| 751 | |
| 752 | =head1 HISTORY |
| 753 | |
| 754 | This port was originally contributed by Gary Ng around 5.003_24, |
| 755 | and borrowed from the Hip Communications port that was available |
| 756 | at the time. Various people have made numerous and sundry hacks |
| 757 | since then. |
| 758 | |
| 759 | Borland support was added in 5.004_01 (Gurusamy Sarathy). |
| 760 | |
| 761 | GCC/mingw32 support was added in 5.005 (Nick Ing-Simmons). |
| 762 | |
| 763 | Support for PERL_OBJECT was added in 5.005 (ActiveState Tool Corp). |
| 764 | |
| 765 | Support for fork() emulation was added in 5.6 (ActiveState Tool Corp). |
| 766 | |
| 767 | Win9x support was added in 5.6 (Benjamin Stuhl). |
| 768 | |
| 769 | Support for 64-bit Windows added in 5.8 (ActiveState Corp). |
| 770 | |
| 771 | Last updated: 20 April 2002 |
| 772 | |
| 773 | =cut |