1Inline(3) User Contributed Perl Documentation Inline(3)
2
6 Inline - Write Perl Subroutines in Other Programming Languages
7
9 This document describes Inline version 0.83.
10
12 use Inline C;
13 print "9 + 16 = ", add(9, 16), "\n";
15 print "9 - 16 = ", subtract(9, 16), "\n";
16 __END__
18 __C__
19 int add(int x, int y) {
20 return x + y;
21 }
22 int subtract(int x, int y) {
24 return x - y;
25 }
26
28 The Inline module allows you to put source code from other programming
29 languages directly "inline" in a Perl script or module. The code is
30 automatically compiled as needed, and then loaded for immediate access
31 from Perl.
32 Inline saves you from the hassle of having to write and compile your
34 own glue code using facilities like XS or SWIG. Simply type the code
35 where you want it and run your Perl as normal. All the hairy details
36 are handled for you. The compilation and installation of your code
37 chunks all happen transparently; all you will notice is the delay of
38 compilation on the first run.
39 The Inline code only gets compiled the first time you run it (or
41 whenever it is modified) so you only take the performance hit once.
42 Code that is Inlined into distributed modules (like on the CPAN) will
43 get compiled when the module is installed, so the end user will never
44 notice the compilation time.
45 Best of all, it works the same on both Unix and Microsoft Windows. See
47 Inline- Support for support information.
48 Why Inline?
50 Do you want to know "Why would I use other languages in Perl?" or "Why
51 should I use Inline to do it?"? I'll try to answer both.
52 Why would I use other languages in Perl?
54 The most obvious reason is performance. For an interpreted
55 language, Perl is very fast. Many people will say "Anything Perl
56 can do, C can do faster". (They never mention the development time
57 :-) Anyway, you may be able to remove a bottleneck in your Perl
58 code by using another language, without having to write the entire
59 program in that language. This keeps your overall development time
60 down, because you're using Perl for all of the non-critical code.
61 Another reason is to access functionality from existing API-s that
63 use the language. Some of this code may only be available in binary
64 form. But by creating small subroutines in the native language, you
65 can "glue" existing libraries to your Perl. As a user of the CPAN,
66 you know that code reuse is a good thing. So why throw away those
67 Fortran libraries just yet?
68 If you are using Inline with the C language, then you can access
70 the full internals of Perl itself. This opens up the floodgates to
71 both extreme power and peril.
72 Maybe the best reason is "Because you want to!". Diversity keeps
74 the world interesting. TMTOWTDI!
75 Why should I use Inline to do it?
77 There are already two major facilities for extending Perl with C.
78 They are XS and SWIG. Both are similar in their capabilities, at
79 least as far as Perl is concerned. And both of them are quite
80 difficult to learn compared to Inline.
81 There is a big fat learning curve involved with setting up and
83 using the XS environment. You need to get quite intimate with the
84 following docs:
85 · perlxs
87 · perlxstut
89 · perlapi
91 · perlguts
93 · perlmod
95 · h2xs
97 · xsubpp
99 · ExtUtils::MakeMaker
101 With Inline you can be up and running in minutes. There is a C
103 Cookbook with lots of short but complete programs that you can
104 extend to your real-life problems. No need to learn about the
105 complicated build process going on in the background. You don't
106 even need to compile the code yourself. Inline takes care of every
107 last detail except writing the C code.
108 Perl programmers cannot be bothered with silly things like
110 compiling. "Tweak, Run, Tweak, Run" is our way of life. Inline does
111 all the dirty work for you.
112 Another advantage of Inline is that you can use it directly in a
114 script. You can even use it in a Perl one-liner. With XS and SWIG,
115 you always set up an entirely separate module. Even if you only
116 have one or two functions. Inline makes easy things easy, and hard
117 things possible. Just like Perl.
118 Finally, Inline supports several programming languages (not just C
120 and C++). As of this writing, Inline has support for C, C++, Java,
121 Python, Ruby, Tcl, Assembler, Basic, Guile, Befunge, Octave, Awk,
122 BC, TT (Template Toolkit), WebChat and even PERL. New Inline
123 Language Support Modules (ILSMs) are regularly being added. See
124 Inline-API for details on how to create your own ILSM.
125
127 Inline is a little bit different than most of the Perl modules that you
128 are used to. It doesn't import any functions into your namespace and it
129 doesn't have any object oriented methods. Its entire interface (with
130 two minor exceptions) is specified through the 'use Inline ...'
131 command.
132 This section will explain all of the different ways to "use Inline". If
134 you want to begin using C with Inline immediately, see
135 Inline::C-Cookbook.
136 The Basics
138 The most basic form for using Inline is:
139 use Inline X => "X source code";
141 where 'X' is one of the supported Inline programming languages. The
143 second parameter identifies the source code that you want to bind to
144 Perl. The source code can be specified using any of the following
145 syntaxes:
146 The DATA Keyword.
148 use Inline Java => 'DATA';
149 # Perl code goes here ...
151 __DATA__
153 __Java__
154 /* Java code goes here ... */
155 The easiest and most visually clean way to specify your source code
157 in an Inline Perl program is to use the special "DATA" keyword.
158 This tells Inline to look for a special marker in your "DATA"
159 filehandle's input stream. In this example the special marker is
160 "__Java__", which is the programming language surrounded by double
161 underscores.
162 In case you've forgotten, the "DATA" pseudo file is comprised of
164 all the text after the "__END__" or "__DATA__" section of your
165 program. If you're working outside the "main" package, you'd best
166 use the "__DATA__" marker or else Inline will not find your code.
167 Using this scheme keeps your Perl code at the top, and all the ugly
169 Java stuff down below where it belongs. This is visually clean and
170 makes for more maintainable code. An excellent side benefit is that
171 you don't have to escape any characters like you might in a Perl
172 string. The source code is verbatim. For these reasons, I prefer
173 this method the most.
174 The only problem with this style is that since Perl can't read the
176 "DATA" filehandle until runtime, it obviously can't bind your
177 functions until runtime. The net effect of this is that you can't
178 use your Inline functions as barewords (without predeclaring them)
179 because Perl has no idea they exist during compile time.
180 The FILE and BELOW keywords.
182 use Inline::Files;
183 use Inline Java => 'file';
184 # Perl code goes here ...
186 __JAVA__
188 /* Java code goes here ... */
189 This is the newest method of specifying your source code. It makes
191 use of the Perl module "Inline::Files" written by Damian Conway.
192 The basic style and meaning are the same as for the "DATA" keyword,
193 but there are a few syntactic and semantic twists.
194 First, you must say 'use Inline::Files' before you 'use Inline'
196 code that needs those files. The special '"DATA"' keyword is
197 replaced by either '"file"' or '"below"'. This allows for the bad
198 pun idiom of:
199 use Inline C => 'below';
201 You can omit the "__DATA__" tag now. Inline::Files is a source
203 filter that will remove these sections from your program before
204 Perl compiles it. They are then available for Inline to make use
205 of. And since this can all be done at compile time, you don't have
206 to worry about the caveats of the 'DATA' keyword.
207 This module has a couple small gotchas. Since Inline::Files only
209 recognizes file markers with capital letters, you must specify the
210 capital form of your language name. Also, there is a startup time
211 penalty for using a source code filter.
212 At this point Inline::Files is alpha software and use of it is
214 experimental. Inline's integration of this module is also
215 fledgling at the time being. One of things I plan to do with
216 Inline::Files is to get line number info so when an extension
217 doesn't compile, the error messages will point to the correct
218 source file and line number.
219 My best advice is to use Inline::Files for testing (especially as
221 support for it improves), but use DATA for production and
222 distributed/CPAN code.
223 Strings
225 use Inline Java => <<'END';
226 /* Java code goes here ... */
228 END
229 # Perl code goes here ...
231 You also just specify the source code as a single string. A handy
233 way to write the string is to use Perl's "here document" style of
234 quoting. This is ok for small functions but can get unwieldy in the
235 large. On the other hand, the string variant probably has the least
236 startup penalty and all functions are bound at compile time.
237 If you wish to put the string into a scalar variable, please be
239 aware that the "use" statement is a compile time directive. As
240 such, all the variables it uses must also be set at compile time,
241 "before" the 'use Inline' statement. Here is one way to do it:
242 my $code;
244 BEGIN {
245 $code = <<END;
246 /* Java code goes here ... */
248 END
249 }
250 use Inline Java => $code;
251 # Perl code goes here ...
253 The bind() Function
255 An alternative to using the BEGIN block method is to specify the
256 source code at run time using the 'Inline->bind()' method. (This is
257 one of the interface exceptions mentioned above) The "bind()"
258 method takes the same arguments as 'use Inline ...'.
259 my $code = <<END;
261 /* Java code goes here ... */
263 END
264 Inline->bind(Java => $code);
266 You can think of "bind()" as a way to "eval()" code in other
268 programming languages.
269 Although bind() is a powerful feature, it is not recommended for
271 use in Inline based modules. In fact, it won't work at all for
272 installable modules. See instructions below for creating modules
273 with Inline.
274 Other Methods
276 The source code for Inline can also be specified as an external
277 filename, a reference to a subroutine that returns source code, or
278 a reference to an array that contains lines of source code. (Note
279 that if the external source file is in the current directory it
280 must be specified with a leading '.' - ie '.file.ext' instead of
281 simply 'file.ext'.) These methods are less frequently used but may
282 be useful in some situations.
283 For instance, to load your C++ code from a file named the same as
285 your perl module with a swapped file extension, you can use:
286 use Inline CPP => (__FILE__ =~ s/\.pm$/.cpp/r);
288 Shorthand
290 If you are using the 'DATA' or 'file' methods described above and
291 there are no extra parameters, you can omit the keyword altogether.
292 For example:
293 use Inline 'Java';
295 # Perl code goes here ...
297 __DATA__
299 __Java__
300 /* Java code goes here ... */
301 or
303 use Inline::Files;
305 use Inline 'Java';
306 # Perl code goes here ...
308 __JAVA__
310 /* Java code goes here ... */
311 More about the DATA Section
313 If you are writing a module, you can also use the DATA section for POD
314 and AutoLoader subroutines. Just be sure to put them before the first
315 Inline marker. If you install the helper module "Inline::Filters", you
316 can even use POD inside your Inline code. You just have to specify a
317 filter to strip it out.
318 You can also specify multiple Inline sections, possibly in different
320 programming languages. Here is another example:
321 # The module Foo.pm
323 package Foo;
324 use AutoLoader;
325 use Inline C;
327 use Inline C => DATA => filters => 'Strip_POD';
328 use Inline Python;
329 1;
331 __DATA__
333 sub marine {
335 # This is an autoloaded subroutine
336 }
337 =head1 External subroutines
339 =cut
341 __C__
343 /* First C section */
344 __C__
346 /* Second C section */
347 =head1 My C Function
348 Some POD doc.
350 =cut
352 __Python__
354 """A Python Section"""
355 An important thing to remember is that you need to have one "use Inline
357 Foo => 'DATA'" for each "__Foo__" marker, and they must be in the same
358 order. This allows you to apply different configuration options to
359 each section.
360 Configuration Options
362 Inline tries to do the right thing as often as possible. But sometimes
363 you may need to override the default actions. This is easy to do.
364 Simply list the Inline configuration options after the regular Inline
365 parameters. All configuration options are specified as (key, value)
366 pairs.
367 use Inline (C => 'DATA',
369 directory => './inline_dir',
370 libs => '-lfoo',
371 inc => '-I/foo/include',
372 prefix => 'XXX_',
373 warnings => 0,
374 );
375 You can also specify the configuration options on a separate Inline
377 call like this:
378 use Inline (C => Config =>
380 directory => './inline_dir',
381 libs => '-lfoo',
382 inc => '-I/foo/include',
383 prefix => 'XXX_',
384 warnings => 0,
385 );
386 use Inline C => <<'END_OF_C_CODE';
387 The special keyword 'Config' tells Inline that this is a configuration-
389 only call. No source code will be compiled or bound to Perl.
390 If you want to specify global configuration options that don't apply to
392 a particular language, just leave the language out of the call. Like
393 this:
394 use Inline Config => warnings => 0;
396 The Config options are inherited and additive. You can use as many
398 Config calls as you want. And you can apply different options to
399 different code sections. When a source code section is passed in,
400 Inline will apply whichever options have been specified up to that
401 point. Here is a complex configuration example:
402 use Inline (Config =>
404 directory => './inline_dir',
405 );
406 use Inline (C => Config =>
407 libs => '-lglobal',
408 );
409 use Inline (C => 'DATA', # First C Section
410 libs => ['-llocal1', '-llocal2'],
411 );
412 use Inline (Config =>
413 warnings => 0,
414 );
415 use Inline (Python => 'DATA', # First Python Section
416 libs => '-lmypython1',
417 );
418 use Inline (C => 'DATA', # Second C Section
419 libs => [undef, '-llocal3'],
420 );
421 The first "Config" applies to all subsequent calls. The second "Config"
423 applies to all subsequent "C" sections (but not "Python" sections). In
424 the first "C" section, the external libraries "global", "local1" and
425 "local2" are used. (Most options allow either string or array ref
426 forms, and do the right thing.) The "Python" section does not use the
427 "global" library, but does use the same "DIRECTORY", and has warnings
428 turned off. The second "C" section only uses the "local3" library.
429 That's because a value of "undef" resets the additive behavior.
430 The "directory" and "warnings" options are generic Inline options. All
432 other options are language specific. To find out what the "C" options
433 do, see "Inline::C".
434 On and Off
436 If a particular config option has value options of 1 and 0, you can use
437 the 'enable' and 'disable' modifiers. In other words, this:
438 use Inline Config =>
440 force_build => 1,
441 clean_after_build => 0;
442 could be reworded as:
444 use Inline Config =>
446 enable => force_build =>
447 disable => clean_after_build;
448 Playing 'with' Others
450 Inline has a special configuration syntax that tells it to get more
451 configuration options from other Perl modules. Here is an example:
452 use Inline with => 'Event';
454 This tells Inline to load the module "Event.pm" and ask it for
456 configuration information. Since "Event" has a C API of its own, it can
457 pass Inline all of the information it needs to be able to use "Event" C
458 callbacks seamlessly.
459 That means that you don't need to specify the typemaps, shared
461 libraries, include files and other information required to get this to
462 work.
463 You can specify a single module or a list of them. Like:
465 use Inline with => qw(Event Foo Bar);
467 Currently, "Event" is the only module that works with Inline.
469 In order to make your module work with Inline in this way, your module
471 needs to provide a class method called "Inline" that takes an Inline
472 language as a parameter (e.g. "C"), and returns a reference to a hash
473 with configuration information that is acceptable to the relevant ILSM.
474 For C, see C Configuration Options. E.g.:
475 my $confighashref = Event->Inline('C'); # only supports C in 1.21
477 # hashref contains keys INC, TYPEMAPS, MYEXTLIB, AUTO_INCLUDE, BOOT
478 If your module uses ExtUtils::Depends version 0.400 or higher, your
480 module only needs this:
481 package Module;
483 use autouse Module::Install::Files => qw(Inline);
484 Inline Shortcuts
486 Inline lets you set many configuration options from the command line.
487 These options are called 'shortcuts'. They can be very handy,
488 especially when you only want to set the options temporarily, for say,
489 debugging.
490 For instance, to get some general information about your Inline code in
492 the script "Foo.pl", use the command:
493 perl -MInline=info Foo.pl
495 If you want to force your code to compile, even if its already done,
497 use:
498 perl -MInline=force Foo.pl
500 If you want to do both, use:
502 perl -MInline=info -MInline=force Foo.pl
504 or better yet:
506 perl -MInline=info,force Foo.pl
508 The Inline 'directory'
510 Inline needs a place to build your code and to install the results of
511 the build. It uses a single directory named '.Inline/' under normal
512 circumstances. If you create this directory in your home directory, the
513 current directory or in the directory where your program resides,
514 Inline will find and use it. You can also specify it in the environment
515 variable "PERL_INLINE_DIRECTORY" or directly in your program, by using
516 the "directory" keyword option. If Inline cannot find the directory in
517 any of these places it will create a '_Inline/' directory in either
518 your current directory or the directory where your script resides.
519 One of the key factors to using Inline successfully, is understanding
521 this directory. When developing code it is usually best to create this
522 directory (or let Inline do it) in your current directory. Remember
523 that there is nothing sacred about this directory except that it holds
524 your compiled code. Feel free to delete it at any time. Inline will
525 simply start from scratch and recompile your code on the next run. If
526 you have several programs that you want to force to recompile, just
527 delete your '.Inline/' directory.
528 It is probably best to have a separate '.Inline/' directory for each
530 project that you are working on. You may want to keep stable code in
531 the <.Inline/> in your home directory. On multi-user systems, each user
532 should have their own '.Inline/' directories. It could be a security
533 risk to put the directory in a shared place like "/tmp/".
534 Debugging Inline Errors
536 All programmers make mistakes. When you make a mistake with Inline,
537 like writing bad C code, you'll get a big error report on your screen.
538 This report tells you where to look to do the debugging. Some languages
539 may also dump out the error messages generated from the build.
540 When Inline needs to build something it creates a subdirectory under
542 your "DIRECTORY/build/" directory. This is where it writes all the
543 components it needs to build your extension. Things like XS files,
544 Makefiles and output log files.
545 If everything goes OK, Inline will delete this subdirectory. If there
547 is an error, Inline will leave the directory intact and print its
548 location. The idea is that you are supposed to go into that directory
549 and figure out what happened.
550 Read the doc for your particular Inline Language Support Module for
552 more information.
553 The 'config' Registry File
555 Inline keeps a cached file of all of the Inline Language Support
556 Module's meta data in a file called "config". This file can be found in
557 your "directory" directory. If the file does not exist, Inline creates
558 a new one. It will search your system for any module beginning with
559 "Inline::". It will then call that module's "register()" method to get
560 useful information for future invocations.
561 Whenever you add a new ILSM, you should delete this file so that Inline
563 will auto-discover your newly installed language module. (This should
564 no longer be necessary as of Inline-0.49.)
565
567 This section lists all of the generic Inline configuration options. For
568 language specific configuration, see the doc for that language.
569 "directory"
571 The "directory" config option is the directory that Inline uses to
572 both build and install an extension.
573 Normally Inline will search in a bunch of known places for a
575 directory called '.Inline/'. Failing that, it will create a
576 directory called '_Inline/'
577 If you want to specify your own directory, use this configuration
579 option.
580 Note that you must create the "directory" directory yourself.
582 Inline will not do it for you.
583 "name"
585 You can use this option to set the name of your Inline extension
586 object module. For example:
587 use Inline C => 'DATA',
589 name => 'Foo::Bar';
590 would cause your C code to be compiled in to the object:
592 lib/auto/Foo/Bar/Bar.so
594 lib/auto/Foo/Bar/Bar.inl
595 (The .inl component contains dependency information to make sure
597 the source code is in sync with the executable)
598 If you don't use "name", Inline will pick a name for you based on
600 your program name or package name. In this case, Inline will also
601 enable the "autoname" option which mangles in a small piece of the
602 MD5 fingerprint into your object name, to make it unique.
603 "autoname"
605 This option is enabled whenever the "name" parameter is not
606 specified. To disable it say:
607 use Inline C => 'DATA',
609 disable => 'autoname';
610 "autoname" mangles in enough of the MD5 fingerprint to make your
612 module name unique. Objects created with "autoname" will never get
613 replaced. That also means they will never get cleaned up
614 automatically.
615 "autoname" is very useful for small throw away scripts. For more
617 serious things, always use the "name" option.
618 "version"
620 Specifies the version number of the Inline extension object. It is
621 used only for modules, and it must match the global variable
622 $VERSION. Additionally, this option should used if (and only if) a
623 module is being set up to be installed permanently into the Perl
624 sitelib tree using Inline::MakeMaker (NOT used by Inline::Module).
625 Inline will croak if you use it otherwise.
626 The presence of the "version" parameter is the official way to let
628 Inline know that your code is an installable/installed module.
629 Inline will never generate an object in the temporary cache
630 ("_Inline/" directory) if "version" is set. It will also never try
631 to recompile a module that was installed into someone's Perl site
632 tree.
633 So the basic rule is develop without "version", and deliver with
635 "version".
636 "with"
638 "with" can also be used as a configuration option instead of using
639 the special 'with' syntax. Do this if you want to use different
640 sections of Inline code with different modules. (Probably a very
641 rare usage)
642 use Event;
644 use Inline C => DATA => with => 'Event';
645 Modules specified using the config form of "with" will not be
647 automatically required. You must "use" them yourself.
648 "using"
650 You can override modules that get used by ILSMs with the "using"
651 option. This is typically used to override the default parser for
652 Inline::C, but might be used by any ILSM for any purpose.
653 use Inline config => using => '::Parser::RecDescent';
655 use Inline C => '...';
656 This would tell Inline::C to use Inline::C::Parser::RecDescent.
658 "global_load"
660 This option is for compiled languages only. It tells Inline to tell
661 DynaLoader to load an object file in such a way that its symbols
662 can be dynamically resolved by other object files. May not work on
663 all platforms. See the "global" shortcut below.
664 "untaint"
666 You can use this option whenever you use Perl's "-T" switch, for
667 taint checking. This option tells Inline to blindly untaint all
668 tainted variables. (This is generally considered to be an
669 appallingly insecure thing to do, and not to be recommended - but
670 the option is there for you to use if you want. Please consider
671 using something other than Inline for scripts that need taint
672 checking.) It also turns on "safemode" by default. See the
673 "untaint" shortcut below. You will see warnings about blindly
674 untainting fields in both %ENV and Inline objects. If you want to
675 silence these warnings, set the Config option "no_untaint_warn" =>
676 1. There can be some problems untainting Inline scripts where older
677 versions of Cwd, such as those that shipped with early versions of
678 perl-5.8 (and earlier), are installed. Updating Cwd will probably
679 solve these problems.
680 safemode
682 Perform extra safety checking, in an attempt to thwart malicious
683 code. This option cannot guarantee security, but it does turn on
684 all the currently implemented checks. (Currently, the only
685 "currently implemented check" is to ensure that the "directory"
686 option has also been used.)
687 There is a slight startup penalty by using "safemode". Also, using
689 "untaint" automatically turns this option on. If you need your code
690 to start faster under "-T" (taint) checking, you'll need to turn
691 this option off manually. Only do this if you are not worried
692 about security risks. See the "unsafe" shortcut below.
693 "force_build"
695 Makes Inline build (compile) the source code every time the program
696 is run. The default is 0. See the "force" shortcut below.
697 "build_noisy"
699 Tells ILSMs that they should dump build messages to the terminal
700 rather than be silent about all the build details.
701 "build_timers"
703 Tells ILSMs to print timing information about how long each build
704 phase took. Usually requires "Time::HiRes".
705 "clean_after_build"
707 Tells Inline to clean up the current build area if the build was
708 successful. Sometimes you want to "disable" this for debugging.
709 Default is 1. See the "noclean" shortcut below.
710 "clean_build_area"
712 Tells Inline to clean up the old build areas within the entire
713 Inline "directory". Default is 0. See the "clean" shortcut below.
714 "print_info"
716 Tells Inline to print various information about the source code.
717 Default is 0. See the "info" shortcut below.
718 "print_version"
720 Tells Inline to print version info about itself. Default is 0. See
721 the "version" shortcut below.
722 "reportbug"
724 Puts Inline into 'reportbug' mode, which is what you want if you
725 desire to report a bug.
726 "rewrite_config_file"
728 Default is 0, but setting "rewrite_config_file => 1" will mean that
729 the existing configuration file in the Inline "directory" will be
730 overwritten. (This is useful if the existing config file is not up
731 to date as regards supported languages.)
732 "warnings"
734 This option tells Inline whether to print certain warnings. Default
735 is 1.
736
738 This is a list of all the shortcut configuration options currently
739 available for Inline. Specify them from the command line when running
740 Inline scripts.
741 perl -MInline=noclean inline_script.pl
743 or
745 perl -MInline=info,force,noclean inline_script.pl
747 You can specify multiple shortcuts separated by commas. They are not
749 case sensitive. You can also specify shortcuts inside the Inline
750 program like this:
751 use Inline 'info', 'force', 'noclean';
753 NOTE: If a 'use Inline' statement is used to set shortcuts, it can not
755 be
756 used for additional purposes.
757 "clean"
759 Tells Inline to remove any build directories that may be lying
760 around in your build area. Normally these directories get removed
761 immediately after a successful build. Exceptions are when the build
762 fails, or when you use the "noclean" or "reportbug" options.
763 "force"
765 Forces the code to be recompiled, even if everything is up to date.
766 "global"
768 Turns on the "global_load" option.
769 "info"
771 This is a very useful option when you want to know what's going on
772 under the hood. It tells Inline to print helpful information to
773 "STDERR". Among the things that get printed is a list of which
774 Inline functions were successfully bound to Perl.
775 "noclean"
777 Tells Inline to leave the build files after compiling.
778 "noisy"
780 Use the "build_noisy" option to print messages during a build.
781 "reportbug"
783 Puts Inline into "reportbug" mode, which does special processing
784 when you want to report a bug. "reportbug" also automatically
785 forces a build, and doesn't clean up afterwards. This is so that
786 you can tar and mail the build directory to me. "reportbug" will
787 print exact instructions on what to do. Please read and follow
788 them carefully.
789 NOTE: "reportbug" informs you to use the tar command. If your
791 system does not
792 have tar, please use the equivalent "zip" command.
793 "safe"
795 Turns "safemode" on. "untaint" will turn this on automatically.
796 While this mode performs extra security checking, it does not
797 guarantee safety.
798 "site_install"
800 This parameter used to be used for creating installable Inline
801 modules. It has been removed from Inline altogether and replaced
802 with a much simpler and more powerful mechanism,
803 "Inline::MakeMaker". See the section below on how to create modules
804 with Inline.
805 "_testing"
807 Used internally by Ct09parser.t and Ct10callback.t(in the Inline::C
808 test suite). Setting this option with Inline::C will mean that
809 files named "parser_id" and "void_test" are created in the
810 "./Inline_test" directory, creating that directory if it doesn't
811 already exist. The files (but not the "./Inline_test directory")
812 are cleaned up by calling "Inline::C::_testing_cleanup()". Also
813 used by "t/06rewrite_config.t" to trigger a warning.
814 "timers"
816 Turn on "build_timers" to get extra diagnostic info about builds.
817 "unsafe"
819 Turns "safemode" off. Use this in combination with "untaint" for
820 slightly faster startup time under "-T". Only use this if you are
821 sure the environment is safe.
822 "untaint"
824 Turn the "untaint" option on. Used with "-T" switch. In terms of
825 secure practices, this is definitely not a recommended way of
826 dealing with taint checking, but it's the only option currently
827 available with Inline. Use it at your own risk.
828 "version"
830 Tells Inline to report its release version.
831
833 The current preferred way to author CPAN modules with Inline is to use
834 Inline::Module (distributed separately). Inline ships with
835 Inline::MakeMaker, which helps you set up a Makefile.PL that invokes
836 Inline at install time to compile all the code before it gets
837 installed, but the resulting module still depends on Inline and the
838 language support module like Inline::C. In order to avoid this
839 dependency, what you really want to do is convert your distribution to
840 plain XS before uploading it to CPAN. Inline::Module fills that role,
841 and also integrates well with more modern authoring tools.
842 See Inline::Module for details on that approach, or continue reading
844 below for the older Inline::MakeMaker technique.
845 Let's say that you wanted to write a module called "Math::Simple".
847 Start by using the following command:
848 h2xs -PAXn Math::Simple
850 This will generate a bunch of files that form a skeleton of what you
852 need for a distributable module. (Read the h2xs manpage to find out
853 what the options do) Next, modify the "Simple.pm" file to look like
854 this:
855 package Math::Simple;
857 $VERSION = '1.23';
858 use base 'Exporter';
860 @EXPORT_OK = qw(add subtract);
861 use strict;
862 use Inline C => 'DATA',
864 version => '1.23',
865 name => 'Math::Simple';
866 # The following Inline->init() call is optional - see below for more info.
868 #Inline->init();
869 1;
871 __DATA__
873 =pod
875 =cut
877 __C__
879 int add(int x, int y) {
880 return x + y;
881 }
882 int subtract(int x, int y) {
884 return x - y;
885 }
886 The important things to note here are that you must specify a "name"
888 and "version" parameter. The "name" must match your module's package
889 name. The "version" parameter must match your module's $VERSION
890 variable and they must be considered valid by "version::parse".
891 NOTE: These are Inline's sanity checks to make sure you know what
893 you're doing
894 before uploading your code to CPAN. They insure that once the
895 module has
896 been installed on someone's system, the module would not get
897 automatically recompiled for any reason. This makes Inline based
898 modules
899 work in exactly the same manner as XS based ones.
900 Finally, you need to modify the Makefile.PL. Simply change:
902 use ExtUtils::MakeMaker;
904 to
906 use Inline::MakeMaker;
908 And, in order that the module build work correctly in the cpan shell,
910 add the following directive to the Makefile.PL's WriteMakefile():
911 CONFIGURE_REQUIRES => {
913 'Inline::MakeMaker' => 0.45,
914 'ExtUtils::MakeMaker' => 6.52,
915 },
916 This "CONFIGURE_REQUIRES" directive ensures that the cpan shell will
918 install Inline on the user's machine (if it's not already present)
919 before building your Inline-based module. Specifying of
920 "ExtUtils::MakeMaker => 6.52," is optional, and can be omitted if you
921 like. It ensures only that some harmless warnings relating to the
922 "CONFIGURE_REQUIRES" directive won't be emitted during the building of
923 the module. It also means, of course, that ExtUtils::Makemaker will
924 first be updated on the user's machine unless the user already has
925 version 6.52 or later.
926 If the "Inline->init();" is not done then, having installed
928 Math::Simple, a warning that "One or more DATA sections were not
929 processed by Inline" will appear when (and only when) Math::Simple is
930 loaded by a "require call. It's a harmless warning - and if you're
931 prepared to live with it, then there's no need to make the
932 "Inline->init();" call.
933 When the person installing "Math::Simple" does a ""make"", the
935 generated Makefile will invoke Inline in such a way that the C code
936 will be compiled and the executable code will be placed into the
937 "./blib" directory. Then when a ""make install"" is done, the module
938 will be copied into the appropriate Perl sitelib directory (which is
939 where an installed module should go).
940 Now all you need to do is:
942 perl Makefile.PL
944 make dist
945 That will generate the file "Math-Simple-0.20.tar.gz" which is a
947 distributable package. (It will also generate some harmless warnings in
948 relation to "CONFIGURE_REQUIRES" unless the version of your
949 ExtUtils::MakeMaker is 6.52 or later.) That's all there is to it.
950 IMPORTANT NOTE: Although the above steps will produce a workable
952 module, you still have a few more responsibilities as a budding new
953 CPAN author. You need to write lots of documentation and write lots of
954 tests. Take a look at some of the better CPAN modules for ideas on
955 creating a killer test harness. Actually, don't listen to me, go read
956 these:
957 · perldoc perlnewmod
959 · <http://www.cpan.org/modules/04pause.html>
961 · <http://www.cpan.org/modules/00modlist.long.html>
963
965 In reality, Inline just automates everything you would need to do if
966 you were going to do it by hand (using XS, etc).
967 Inline performs the following steps:
969 · Receive the Source Code
971 Inline gets the source code from your script or module with a
973 statements like the following:
974 use Inline C => "Source-Code";
976 or
978 use Inline;
980 bind Inline C => "Source-Code";
981 where "C" is the programming language of the source code, and
983 "Source- Code" is a string, a file name, an array reference, or the
984 special 'DATA' keyword.
985 Since Inline is coded in a ""use"" statement, everything is done
987 during Perl's compile time. If anything needs to be done that will
988 affect the "Source- Code", it needs to be done in a "BEGIN" block
989 that is before the ""use Inline ..."" statement. If you really need
990 to specify code to Inline at runtime, you can use the "bind()"
991 method.
992 Source code that is stowed in the 'DATA' section of your code, is
994 read in by an "INIT" subroutine in Inline. That's because the
995 "DATA" filehandle is not available at compile time.
996 · Check if the Source Code has been Built
998 Inline only needs to build the source code if it has not yet been
1000 built. It accomplishes this seemingly magical task in an extremely
1001 simple and straightforward manner. It runs the source text through
1002 the "Digest::MD5" module to produce a 128-bit "fingerprint" which
1003 is virtually unique. The fingerprint along with a bunch of other
1004 contingency information is stored in a ".inl" file that sits next
1005 to your executable object. For instance, the "C" code from a script
1006 called "example.pl" might create these files:
1007 example_pl_3a9a.so
1009 example_pl_3a9a.inl
1010 If all the contingency information matches the values stored in the
1012 ".inl" file, then proceed to step 8. (No compilation is necessary)
1013 · Find a Place to Build and Install
1015 At this point Inline knows it needs to build the source code. The
1017 first thing to figure out is where to create the great big mess
1018 associated with compilation, and where to put the object when it's
1019 done.
1020 By default Inline will try to build and install under the first
1022 place that meets one of the following conditions:
1023 1. The DIRECTORY= config option; if specified
1025 2. The "PERL_INLINE_DIRECTORY" environment variable; if set
1027 3. ".Inline/" (in current directory); if exists and "$PWD !=
1029 $HOME"
1030 4. bin.Inline (in directory of your script); if exists
1032 5. "~/.Inline/" - if exists
1034 6. "./_Inline/" - if exists
1036 7. "bin/_Inline" - if exists
1038 8. Create "./_Inline/" - if possible
1040 9. Create "bin/_Inline/" - if possible
1042 Failing that, Inline will croak. This is rare and easily remedied
1044 by just making a directory that Inline will use.
1045 If the "PERL_INSTALL_ROOT" Environment Variable has been set, you
1047 will need to make special provision for that if the 'make install'
1048 phase of your Inline scripts are to succeed.
1049 If the module option is being compiled for permanent installation,
1051 then Inline will only use "./_Inline/" to build in, and the
1052 $Config{installsitearch} directory to install the executable in.
1053 This action is caused by Inline::MakeMaker, and is intended to be
1054 used in modules that are to be distributed on the CPAN, so that
1055 they get installed in the proper place.
1056 · Parse the Source for Semantic Cues
1058 Inline::C uses the module "Parse::RecDescent" to parse through your
1060 chunks of C source code and look for things that it can create run-
1061 time bindings to. In "C" it looks for all of the function
1062 definitions and breaks them down into names and data types. These
1063 elements are used to correctly bind the "C" function to a "Perl"
1064 subroutine. Other Inline languages like Python and Java actually
1065 use the "python" and "javac" modules to parse the Inline code.
1066 · Create the Build Environment
1068 Now Inline can take all of the gathered information and create an
1070 environment to build your source code into an executable. Without
1071 going into all the details, it just creates the appropriate
1072 directories, creates the appropriate source files including an XS
1073 file (for C) and a "Makefile.PL".
1074 · Build the Code and Install the Executable
1076 The planets are in alignment. Now for the easy part. Inline just
1078 does what you would do to install a module. "`perl Makefile.PL &&
1079 make && make test && make install>". If something goes awry, Inline
1080 will croak with a message indicating where to look for more info.
1081 · Tidy Up
1083 By default, Inline will remove all of the mess created by the build
1085 process, assuming that everything worked. If the build fails,
1086 Inline will leave everything intact, so that you can debug your
1087 errors. Setting the "noclean" shortcut option will also stop Inline
1088 from cleaning up.
1089 · DynaLoad the Executable
1091 For C (and C++), Inline uses the "DynaLoader::bootstrap" method to
1093 pull your external module into "Perl" space. Now you can call all
1094 of your external functions like Perl subroutines.
1095 Other languages like Python and Java, provide their own loaders.
1097
1099 For information about using Inline with C see Inline::C.
1100 For sample programs using Inline with C see Inline::C-Cookbook.
1102 For "Formerly Answered Questions" about Inline, see Inline-FAQ.
1104 For information on supported languages and platforms see Inline-
1106 Support.
1107 For information on writing your own Inline Language Support Module, see
1109 Inline-API.
1110 Inline's mailing list is inline@perl.org
1112 To subscribe, send email to inline-subscribe@perl.org
1114
1116 When reporting a bug, please do the following:
1117 · Put "use Inline 'reportbug';" at the top of your code, or use the
1119 command line option "perl -MInline=reportbug ...".
1120 · Run your code.
1122 · Follow the printed directions.
1124
1126 Ingy döt Net <ingy@cpan.org>
1127 Sisyphus <sisyphus@cpan.org> fixed some bugs and is current co-
1129 maintainer.
1130
1132 · Copyright 2000-2019. Ingy döt Net.
1133 · Copyright 2008, 2010-2014. Sisyphus.
1135 This program is free software; you can redistribute it and/or modify it
1137 under the same terms as Perl itself.
1138 See <http://www.perl.com/perl/misc/Artistic.html>
1140perl v5.30.0 2019-07-26 Inline(3)