3 perlhack - How to hack at the Perl internals
7 This document attempts to explain how Perl development takes place,
8 and ends with some suggestions for people wanting to become bona fide
11 The perl5-porters mailing list is where the Perl standard distribution
12 is maintained and developed. The list can get anywhere from 10 to 150
13 messages a day, depending on the heatedness of the debate. Most days
14 there are two or three patches, extensions, features, or bugs being
17 A searchable archive of the list is at either:
19 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
23 http://archive.develooper.com/perl5-porters@perl.org/
25 List subscribers (the porters themselves) come in several flavours.
26 Some are quiet curious lurkers, who rarely pitch in and instead watch
27 the ongoing development to ensure they're forewarned of new changes or
28 features in Perl. Some are representatives of vendors, who are there
29 to make sure that Perl continues to compile and work on their
30 platforms. Some patch any reported bug that they know how to fix,
31 some are actively patching their pet area (threads, Win32, the regexp
32 engine), while others seem to do nothing but complain. In other
33 words, it's your usual mix of technical people.
35 Over this group of porters presides Larry Wall. He has the final word
36 in what does and does not change in the Perl language. Various
37 releases of Perl are shepherded by a "pumpking", a porter
38 responsible for gathering patches, deciding on a patch-by-patch,
39 feature-by-feature basis what will and will not go into the release.
40 For instance, Gurusamy Sarathy was the pumpking for the 5.6 release of
41 Perl, and Jarkko Hietaniemi was the pumpking for the 5.8 release, and
42 Rafael Garcia-Suarez holds the pumpking crown for the 5.10 release.
44 In addition, various people are pumpkings for different things. For
45 instance, Andy Dougherty and Jarkko Hietaniemi did a grand job as the
46 I<Configure> pumpkin up till the 5.8 release. For the 5.10 release
47 H.Merijn Brand took over.
49 Larry sees Perl development along the lines of the US government:
50 there's the Legislature (the porters), the Executive branch (the
51 pumpkings), and the Supreme Court (Larry). The legislature can
52 discuss and submit patches to the executive branch all they like, but
53 the executive branch is free to veto them. Rarely, the Supreme Court
54 will side with the executive branch over the legislature, or the
55 legislature over the executive branch. Mostly, however, the
56 legislature and the executive branch are supposed to get along and
57 work out their differences without impeachment or court cases.
59 You might sometimes see reference to Rule 1 and Rule 2. Larry's power
60 as Supreme Court is expressed in The Rules:
66 Larry is always by definition right about how Perl should behave.
67 This means he has final veto power on the core functionality.
71 Larry is allowed to change his mind about any matter at a later date,
72 regardless of whether he previously invoked Rule 1.
76 Got that? Larry is always right, even when he was wrong. It's rare
77 to see either Rule exercised, but they are often alluded to.
79 New features and extensions to the language are contentious, because
80 the criteria used by the pumpkings, Larry, and other porters to decide
81 which features should be implemented and incorporated are not codified
82 in a few small design goals as with some other languages. Instead,
83 the heuristics are flexible and often difficult to fathom. Here is
84 one person's list, roughly in decreasing order of importance, of
85 heuristics that new features have to be weighed against:
89 =item Does concept match the general goals of Perl?
91 These haven't been written anywhere in stone, but one approximation
94 1. Keep it fast, simple, and useful.
95 2. Keep features/concepts as orthogonal as possible.
96 3. No arbitrary limits (platforms, data sizes, cultures).
97 4. Keep it open and exciting to use/patch/advocate Perl everywhere.
98 5. Either assimilate new technologies, or build bridges to them.
100 =item Where is the implementation?
102 All the talk in the world is useless without an implementation. In
103 almost every case, the person or people who argue for a new feature
104 will be expected to be the ones who implement it. Porters capable
105 of coding new features have their own agendas, and are not available
106 to implement your (possibly good) idea.
108 =item Backwards compatibility
110 It's a cardinal sin to break existing Perl programs. New warnings are
111 contentious--some say that a program that emits warnings is not
112 broken, while others say it is. Adding keywords has the potential to
113 break programs, changing the meaning of existing token sequences or
114 functions might break programs.
116 =item Could it be a module instead?
118 Perl 5 has extension mechanisms, modules and XS, specifically to avoid
119 the need to keep changing the Perl interpreter. You can write modules
120 that export functions, you can give those functions prototypes so they
121 can be called like built-in functions, you can even write XS code to
122 mess with the runtime data structures of the Perl interpreter if you
123 want to implement really complicated things. If it can be done in a
124 module instead of in the core, it's highly unlikely to be added.
126 =item Is the feature generic enough?
128 Is this something that only the submitter wants added to the language,
129 or would it be broadly useful? Sometimes, instead of adding a feature
130 with a tight focus, the porters might decide to wait until someone
131 implements the more generalized feature. For instance, instead of
132 implementing a "delayed evaluation" feature, the porters are waiting
133 for a macro system that would permit delayed evaluation and much more.
135 =item Does it potentially introduce new bugs?
137 Radical rewrites of large chunks of the Perl interpreter have the
138 potential to introduce new bugs. The smaller and more localized the
141 =item Does it preclude other desirable features?
143 A patch is likely to be rejected if it closes off future avenues of
144 development. For instance, a patch that placed a true and final
145 interpretation on prototypes is likely to be rejected because there
146 are still options for the future of prototypes that haven't been
149 =item Is the implementation robust?
151 Good patches (tight code, complete, correct) stand more chance of
152 going in. Sloppy or incorrect patches might be placed on the back
153 burner until the pumpking has time to fix, or might be discarded
154 altogether without further notice.
156 =item Is the implementation generic enough to be portable?
158 The worst patches make use of a system-specific features. It's highly
159 unlikely that non-portable additions to the Perl language will be
162 =item Is the implementation tested?
164 Patches which change behaviour (fixing bugs or introducing new features)
165 must include regression tests to verify that everything works as expected.
166 Without tests provided by the original author, how can anyone else changing
167 perl in the future be sure that they haven't unwittingly broken the behaviour
168 the patch implements? And without tests, how can the patch's author be
169 confident that his/her hard work put into the patch won't be accidentally
170 thrown away by someone in the future?
172 =item Is there enough documentation?
174 Patches without documentation are probably ill-thought out or
175 incomplete. Nothing can be added without documentation, so submitting
176 a patch for the appropriate manpages as well as the source code is
179 =item Is there another way to do it?
181 Larry said "Although the Perl Slogan is I<There's More Than One Way
182 to Do It>, I hesitate to make 10 ways to do something". This is a
183 tricky heuristic to navigate, though--one man's essential addition is
184 another man's pointless cruft.
186 =item Does it create too much work?
188 Work for the pumpking, work for Perl programmers, work for module
189 authors, ... Perl is supposed to be easy.
191 =item Patches speak louder than words
193 Working code is always preferred to pie-in-the-sky ideas. A patch to
194 add a feature stands a much higher chance of making it to the language
195 than does a random feature request, no matter how fervently argued the
196 request might be. This ties into "Will it be useful?", as the fact
197 that someone took the time to make the patch demonstrates a strong
198 desire for the feature.
202 If you're on the list, you might hear the word "core" bandied
203 around. It refers to the standard distribution. "Hacking on the
204 core" means you're changing the C source code to the Perl
205 interpreter. "A core module" is one that ships with Perl.
207 =head2 Keeping in sync
209 The source code to the Perl interpreter, in its different versions, is
210 kept in a repository managed by the git revision control system. The
211 pumpkings and a few others have write access to the repository to check in
214 How to clone and use the git perl repository is described in L<perlrepository>.
216 You can also choose to use rsync to get a copy of the current source tree
217 for the bleadperl branch and all maintenance branches :
219 $ rsync -avz rsync://perl5.git.perl.org/APC/perl-current .
220 $ rsync -avz rsync://perl5.git.perl.org/APC/perl-5.10.x .
221 $ rsync -avz rsync://perl5.git.perl.org/APC/perl-5.8.x .
222 $ rsync -avz rsync://perl5.git.perl.org/APC/perl-5.6.x .
223 $ rsync -avz rsync://perl5.git.perl.org/APC/perl-5.005xx .
225 (Add the C<--delete> option to remove leftover files)
227 You may also want to subscribe to the perl5-changes mailing list to
228 receive a copy of each patch that gets submitted to the maintenance
229 and development "branches" of the perl repository. See
230 http://lists.perl.org/ for subscription information.
232 If you are a member of the perl5-porters mailing list, it is a good
233 thing to keep in touch with the most recent changes. If not only to
234 verify if what you would have posted as a bug report isn't already
235 solved in the most recent available perl development branch, also
236 known as perl-current, bleading edge perl, bleedperl or bleadperl.
238 Needless to say, the source code in perl-current is usually in a perpetual
239 state of evolution. You should expect it to be very buggy. Do B<not> use
240 it for any purpose other than testing and development.
242 =head2 Perlbug administration
244 There is a single remote administrative interface for modifying bug status,
245 category, open issues etc. using the B<RT> bugtracker system, maintained
246 by Robert Spier. Become an administrator, and close any bugs you can get
247 your sticky mitts on:
249 http://bugs.perl.org/
251 To email the bug system administrators:
253 "perlbug-admin" <perlbug-admin@perl.org>
255 =head2 Submitting patches
257 Always submit patches to I<perl5-porters@perl.org>. If you're
258 patching a core module and there's an author listed, send the author a
259 copy (see L<Patching a core module>). This lets other porters review
260 your patch, which catches a surprising number of errors in patches.
261 Please patch against the latest B<development> version. (e.g., even if
262 you're fixing a bug in the 5.8 track, patch against the C<blead> branch in
265 If changes are accepted, they are applied to the development branch. Then
266 the maintenance pumpking decides which of those patches is to be
267 backported to the maint branch. Only patches that survive the heat of the
268 development branch get applied to maintenance versions.
270 Your patch should update the documentation and test suite. See
271 L<Writing a test>. If you have added or removed files in the distribution,
272 edit the MANIFEST file accordingly, sort the MANIFEST file using
273 C<make manisort>, and include those changes as part of your patch.
275 Patching documentation also follows the same order: if accepted, a patch
276 is first applied to B<development>, and if relevant then it's backported
277 to B<maintenance>. (With an exception for some patches that document
278 behaviour that only appears in the maintenance branch, but which has
279 changed in the development version.)
281 To report a bug in Perl, use the program I<perlbug> which comes with
282 Perl (if you can't get Perl to work, send mail to the address
283 I<perlbug@perl.org> or I<perlbug@perl.com>). Reporting bugs through
284 I<perlbug> feeds into the automated bug-tracking system, access to
285 which is provided through the web at http://rt.perl.org/rt3/ . It
286 often pays to check the archives of the perl5-porters mailing list to
287 see whether the bug you're reporting has been reported before, and if
288 so whether it was considered a bug. See above for the location of
289 the searchable archives.
291 The CPAN testers ( http://testers.cpan.org/ ) are a group of
292 volunteers who test CPAN modules on a variety of platforms. Perl
293 Smokers ( http://www.nntp.perl.org/group/perl.daily-build and
294 http://www.nntp.perl.org/group/perl.daily-build.reports/ )
295 automatically test Perl source releases on platforms with various
296 configurations. Both efforts welcome volunteers. In order to get
297 involved in smoke testing of the perl itself visit
298 L<http://search.cpan.org/dist/Test-Smoke>. In order to start smoke
299 testing CPAN modules visit L<http://search.cpan.org/dist/CPANPLUS-YACSmoke/>
300 or L<http://search.cpan.org/dist/minismokebox/> or
301 L<http://search.cpan.org/dist/CPAN-Reporter/>.
303 It's a good idea to read and lurk for a while before chipping in.
304 That way you'll get to see the dynamic of the conversations, learn the
305 personalities of the players, and hopefully be better prepared to make
306 a useful contribution when do you speak up.
308 If after all this you still think you want to join the perl5-porters
309 mailing list, send mail to I<perl5-porters-subscribe@perl.org>. To
310 unsubscribe, send mail to I<perl5-porters-unsubscribe@perl.org>.
312 To hack on the Perl guts, you'll need to read the following things:
318 This is of paramount importance, since it's the documentation of what
319 goes where in the Perl source. Read it over a couple of times and it
320 might start to make sense - don't worry if it doesn't yet, because the
321 best way to study it is to read it in conjunction with poking at Perl
322 source, and we'll do that later on.
324 Gisle Aas's illustrated perlguts (also known as I<illguts>) is wonderful,
325 although a little out of date with regard to some size details; the
326 various SV structures have since been reworked for smaller memory footprint.
327 The fundamentals are right however, and the pictures are very helpful.
329 L<http://www.perl.org/tpc/1998/Perl_Language_and_Modules/Perl%20Illustrated/>
331 =item L<perlxstut> and L<perlxs>
333 A working knowledge of XSUB programming is incredibly useful for core
334 hacking; XSUBs use techniques drawn from the PP code, the portion of the
335 guts that actually executes a Perl program. It's a lot gentler to learn
336 those techniques from simple examples and explanation than from the core
341 The documentation for the Perl API explains what some of the internal
342 functions do, as well as the many macros used in the source.
344 =item F<Porting/pumpkin.pod>
346 This is a collection of words of wisdom for a Perl porter; some of it is
347 only useful to the pumpkin holder, but most of it applies to anyone
348 wanting to go about Perl development.
350 =item The perl5-porters FAQ
352 This should be available from http://dev.perl.org/perl5/docs/p5p-faq.html .
353 It contains hints on reading perl5-porters, information on how
354 perl5-porters works and how Perl development in general works.
358 =head2 Finding Your Way Around
360 Perl maintenance can be split into a number of areas, and certain people
361 (pumpkins) will have responsibility for each area. These areas sometimes
362 correspond to files or directories in the source kit. Among the areas are:
368 Modules shipped as part of the Perl core live in various subdirectories, where
369 two are dedicated to core-only modules, and two are for the dual-life modules
370 which live on CPAN and may be maintained separately with respect to the Perl
373 lib/ is for pure-Perl modules, which exist in the core only.
375 ext/ is for XS extensions, and modules with special Makefile.PL requirements, which exist in the core only.
377 cpan/ is for dual-life modules, where the CPAN module is canonical (should be patched first).
379 dist/ is for dual-life modules, where the blead source is canonical.
383 There are tests for nearly all the modules, built-ins and major bits
384 of functionality. Test files all have a .t suffix. Module tests live
385 in the F<lib/> and F<ext/> directories next to the module being
386 tested. Others live in F<t/>. See L<Writing a test>
390 Documentation maintenance includes looking after everything in the
391 F<pod/> directory, (as well as contributing new documentation) and
392 the documentation to the modules in core.
396 The Configure process is the way we make Perl portable across the
397 myriad of operating systems it supports. Responsibility for the
398 Configure, build and installation process, as well as the overall
399 portability of the core code rests with the Configure pumpkin -
400 others help out with individual operating systems.
402 The three files that fall under his/her resposibility are Configure,
403 config_h.SH, and Porting/Glossary (and a whole bunch of small related
404 files that are less important here). The Configure pumpkin decides how
405 patches to these are dealt with. Currently, the Configure pumpkin will
406 accept patches in most common formats, even directly to these files.
407 Other committers are allowed to commit to these files under the strict
408 condition that they will inform the Configure pumpkin, either on IRC
409 (if he/she happens to be around) or through (personal) e-mail.
411 The files involved are the operating system directories, (F<win32/>,
412 F<os2/>, F<vms/> and so on) the shell scripts which generate F<config.h>
413 and F<Makefile>, as well as the metaconfig files which generate
414 F<Configure>. (metaconfig isn't included in the core distribution.)
416 See http://perl5.git.perl.org/metaconfig.git/blob/HEAD:/README for a
417 description of the full process involved.
421 And of course, there's the core of the Perl interpreter itself. Let's
422 have a look at that in a little more detail.
426 Before we leave looking at the layout, though, don't forget that
427 F<MANIFEST> contains not only the file names in the Perl distribution,
428 but short descriptions of what's in them, too. For an overview of the
429 important files, try this:
431 perl -lne 'print if /^[^\/]+\.[ch]\s+/' MANIFEST
433 =head2 Elements of the interpreter
435 The work of the interpreter has two main stages: compiling the code
436 into the internal representation, or bytecode, and then executing it.
437 L<perlguts/Compiled code> explains exactly how the compilation stage
440 Here is a short breakdown of perl's operation:
446 The action begins in F<perlmain.c>. (or F<miniperlmain.c> for miniperl)
447 This is very high-level code, enough to fit on a single screen, and it
448 resembles the code found in L<perlembed>; most of the real action takes
451 F<perlmain.c> is generated by L<writemain> from F<miniperlmain.c> at
452 make time, so you should make perl to follow this along.
454 First, F<perlmain.c> allocates some memory and constructs a Perl
455 interpreter, along these lines:
457 1 PERL_SYS_INIT3(&argc,&argv,&env);
459 3 if (!PL_do_undump) {
460 4 my_perl = perl_alloc();
463 7 perl_construct(my_perl);
464 8 PL_perl_destruct_level = 0;
467 Line 1 is a macro, and its definition is dependent on your operating
468 system. Line 3 references C<PL_do_undump>, a global variable - all
469 global variables in Perl start with C<PL_>. This tells you whether the
470 current running program was created with the C<-u> flag to perl and then
471 F<undump>, which means it's going to be false in any sane context.
473 Line 4 calls a function in F<perl.c> to allocate memory for a Perl
474 interpreter. It's quite a simple function, and the guts of it looks like
477 my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));
479 Here you see an example of Perl's system abstraction, which we'll see
480 later: C<PerlMem_malloc> is either your system's C<malloc>, or Perl's
481 own C<malloc> as defined in F<malloc.c> if you selected that option at
484 Next, in line 7, we construct the interpreter using perl_construct,
485 also in F<perl.c>; this sets up all the special variables that Perl
486 needs, the stacks, and so on.
488 Now we pass Perl the command line options, and tell it to go:
490 exitstatus = perl_parse(my_perl, xs_init, argc, argv, (char **)NULL);
494 exitstatus = perl_destruct(my_perl);
498 C<perl_parse> is actually a wrapper around C<S_parse_body>, as defined
499 in F<perl.c>, which processes the command line options, sets up any
500 statically linked XS modules, opens the program and calls C<yyparse> to
505 The aim of this stage is to take the Perl source, and turn it into an op
506 tree. We'll see what one of those looks like later. Strictly speaking,
507 there's three things going on here.
509 C<yyparse>, the parser, lives in F<perly.c>, although you're better off
510 reading the original YACC input in F<perly.y>. (Yes, Virginia, there
511 B<is> a YACC grammar for Perl!) The job of the parser is to take your
512 code and "understand" it, splitting it into sentences, deciding which
513 operands go with which operators and so on.
515 The parser is nobly assisted by the lexer, which chunks up your input
516 into tokens, and decides what type of thing each token is: a variable
517 name, an operator, a bareword, a subroutine, a core function, and so on.
518 The main point of entry to the lexer is C<yylex>, and that and its
519 associated routines can be found in F<toke.c>. Perl isn't much like
520 other computer languages; it's highly context sensitive at times, it can
521 be tricky to work out what sort of token something is, or where a token
522 ends. As such, there's a lot of interplay between the tokeniser and the
523 parser, which can get pretty frightening if you're not used to it.
525 As the parser understands a Perl program, it builds up a tree of
526 operations for the interpreter to perform during execution. The routines
527 which construct and link together the various operations are to be found
528 in F<op.c>, and will be examined later.
532 Now the parsing stage is complete, and the finished tree represents
533 the operations that the Perl interpreter needs to perform to execute our
534 program. Next, Perl does a dry run over the tree looking for
535 optimisations: constant expressions such as C<3 + 4> will be computed
536 now, and the optimizer will also see if any multiple operations can be
537 replaced with a single one. For instance, to fetch the variable C<$foo>,
538 instead of grabbing the glob C<*foo> and looking at the scalar
539 component, the optimizer fiddles the op tree to use a function which
540 directly looks up the scalar in question. The main optimizer is C<peep>
541 in F<op.c>, and many ops have their own optimizing functions.
545 Now we're finally ready to go: we have compiled Perl byte code, and all
546 that's left to do is run it. The actual execution is done by the
547 C<runops_standard> function in F<run.c>; more specifically, it's done by
548 these three innocent looking lines:
550 while ((PL_op = CALL_FPTR(PL_op->op_ppaddr)(aTHX))) {
554 You may be more comfortable with the Perl version of that:
556 PERL_ASYNC_CHECK() while $Perl::op = &{$Perl::op->{function}};
558 Well, maybe not. Anyway, each op contains a function pointer, which
559 stipulates the function which will actually carry out the operation.
560 This function will return the next op in the sequence - this allows for
561 things like C<if> which choose the next op dynamically at run time.
562 The C<PERL_ASYNC_CHECK> makes sure that things like signals interrupt
563 execution if required.
565 The actual functions called are known as PP code, and they're spread
566 between four files: F<pp_hot.c> contains the "hot" code, which is most
567 often used and highly optimized, F<pp_sys.c> contains all the
568 system-specific functions, F<pp_ctl.c> contains the functions which
569 implement control structures (C<if>, C<while> and the like) and F<pp.c>
570 contains everything else. These are, if you like, the C code for Perl's
571 built-in functions and operators.
573 Note that each C<pp_> function is expected to return a pointer to the next
574 op. Calls to perl subs (and eval blocks) are handled within the same
575 runops loop, and do not consume extra space on the C stack. For example,
576 C<pp_entersub> and C<pp_entertry> just push a C<CxSUB> or C<CxEVAL> block
577 struct onto the context stack which contain the address of the op
578 following the sub call or eval. They then return the first op of that sub
579 or eval block, and so execution continues of that sub or block. Later, a
580 C<pp_leavesub> or C<pp_leavetry> op pops the C<CxSUB> or C<CxEVAL>,
581 retrieves the return op from it, and returns it.
583 =item Exception handing
585 Perl's exception handing (i.e. C<die> etc.) is built on top of the low-level
586 C<setjmp()>/C<longjmp()> C-library functions. These basically provide a
587 way to capture the current PC and SP registers and later restore them; i.e.
588 a C<longjmp()> continues at the point in code where a previous C<setjmp()>
589 was done, with anything further up on the C stack being lost. This is why
590 code should always save values using C<SAVE_FOO> rather than in auto
593 The perl core wraps C<setjmp()> etc in the macros C<JMPENV_PUSH> and
594 C<JMPENV_JUMP>. The basic rule of perl exceptions is that C<exit>, and
595 C<die> (in the absence of C<eval>) perform a C<JMPENV_JUMP(2)>, while
596 C<die> within C<eval> does a C<JMPENV_JUMP(3)>.
598 At entry points to perl, such as C<perl_parse()>, C<perl_run()> and
599 C<call_sv(cv, G_EVAL)> each does a C<JMPENV_PUSH>, then enter a runops
600 loop or whatever, and handle possible exception returns. For a 2 return,
601 final cleanup is performed, such as popping stacks and calling C<CHECK> or
602 C<END> blocks. Amongst other things, this is how scope cleanup still
603 occurs during an C<exit>.
605 If a C<die> can find a C<CxEVAL> block on the context stack, then the
606 stack is popped to that level and the return op in that block is assigned
607 to C<PL_restartop>; then a C<JMPENV_JUMP(3)> is performed. This normally
608 passes control back to the guard. In the case of C<perl_run> and
609 C<call_sv>, a non-null C<PL_restartop> triggers re-entry to the runops
610 loop. The is the normal way that C<die> or C<croak> is handled within an
613 Sometimes ops are executed within an inner runops loop, such as tie, sort
614 or overload code. In this case, something like
616 sub FETCH { eval { die } }
618 would cause a longjmp right back to the guard in C<perl_run>, popping both
619 runops loops, which is clearly incorrect. One way to avoid this is for the
620 tie code to do a C<JMPENV_PUSH> before executing C<FETCH> in the inner
621 runops loop, but for efficiency reasons, perl in fact just sets a flag,
622 using C<CATCH_SET(TRUE)>. The C<pp_require>, C<pp_entereval> and
623 C<pp_entertry> ops check this flag, and if true, they call C<docatch>,
624 which does a C<JMPENV_PUSH> and starts a new runops level to execute the
625 code, rather than doing it on the current loop.
627 As a further optimisation, on exit from the eval block in the C<FETCH>,
628 execution of the code following the block is still carried on in the inner
629 loop. When an exception is raised, C<docatch> compares the C<JMPENV>
630 level of the C<CxEVAL> with C<PL_top_env> and if they differ, just
631 re-throws the exception. In this way any inner loops get popped.
635 1: eval { tie @a, 'A' };
641 To run this code, C<perl_run> is called, which does a C<JMPENV_PUSH> then
642 enters a runops loop. This loop executes the eval and tie ops on line 1,
643 with the eval pushing a C<CxEVAL> onto the context stack.
645 The C<pp_tie> does a C<CATCH_SET(TRUE)>, then starts a second runops loop
646 to execute the body of C<TIEARRAY>. When it executes the entertry op on
647 line 3, C<CATCH_GET> is true, so C<pp_entertry> calls C<docatch> which
648 does a C<JMPENV_PUSH> and starts a third runops loop, which then executes
649 the die op. At this point the C call stack looks like this:
652 Perl_runops # third loop
656 Perl_runops # second loop
660 Perl_runops # first loop
665 and the context and data stacks, as shown by C<-Dstv>, look like:
669 CX 1: EVAL => AV() PV("A"\0)
677 The die pops the first C<CxEVAL> off the context stack, sets
678 C<PL_restartop> from it, does a C<JMPENV_JUMP(3)>, and control returns to
679 the top C<docatch>. This then starts another third-level runops level,
680 which executes the nextstate, pushmark and die ops on line 4. At the point
681 that the second C<pp_die> is called, the C call stack looks exactly like
682 that above, even though we are no longer within an inner eval; this is
683 because of the optimization mentioned earlier. However, the context stack
684 now looks like this, ie with the top CxEVAL popped:
688 CX 1: EVAL => AV() PV("A"\0)
694 The die on line 4 pops the context stack back down to the CxEVAL, leaving
700 As usual, C<PL_restartop> is extracted from the C<CxEVAL>, and a
701 C<JMPENV_JUMP(3)> done, which pops the C stack back to the docatch:
705 Perl_runops # second loop
709 Perl_runops # first loop
714 In this case, because the C<JMPENV> level recorded in the C<CxEVAL>
715 differs from the current one, C<docatch> just does a C<JMPENV_JUMP(3)>
716 and the C stack unwinds to:
721 Because C<PL_restartop> is non-null, C<run_body> starts a new runops loop
722 and execution continues.
726 =head2 Internal Variable Types
728 You should by now have had a look at L<perlguts>, which tells you about
729 Perl's internal variable types: SVs, HVs, AVs and the rest. If not, do
732 These variables are used not only to represent Perl-space variables, but
733 also any constants in the code, as well as some structures completely
734 internal to Perl. The symbol table, for instance, is an ordinary Perl
735 hash. Your code is represented by an SV as it's read into the parser;
736 any program files you call are opened via ordinary Perl filehandles, and
739 The core L<Devel::Peek|Devel::Peek> module lets us examine SVs from a
740 Perl program. Let's see, for instance, how Perl treats the constant
743 % perl -MDevel::Peek -e 'Dump("hello")'
744 1 SV = PV(0xa041450) at 0xa04ecbc
746 3 FLAGS = (POK,READONLY,pPOK)
747 4 PV = 0xa0484e0 "hello"\0
751 Reading C<Devel::Peek> output takes a bit of practise, so let's go
752 through it line by line.
754 Line 1 tells us we're looking at an SV which lives at C<0xa04ecbc> in
755 memory. SVs themselves are very simple structures, but they contain a
756 pointer to a more complex structure. In this case, it's a PV, a
757 structure which holds a string value, at location C<0xa041450>. Line 2
758 is the reference count; there are no other references to this data, so
761 Line 3 are the flags for this SV - it's OK to use it as a PV, it's a
762 read-only SV (because it's a constant) and the data is a PV internally.
763 Next we've got the contents of the string, starting at location
766 Line 5 gives us the current length of the string - note that this does
767 B<not> include the null terminator. Line 6 is not the length of the
768 string, but the length of the currently allocated buffer; as the string
769 grows, Perl automatically extends the available storage via a routine
772 You can get at any of these quantities from C very easily; just add
773 C<Sv> to the name of the field shown in the snippet, and you've got a
774 macro which will return the value: C<SvCUR(sv)> returns the current
775 length of the string, C<SvREFCOUNT(sv)> returns the reference count,
776 C<SvPV(sv, len)> returns the string itself with its length, and so on.
777 More macros to manipulate these properties can be found in L<perlguts>.
779 Let's take an example of manipulating a PV, from C<sv_catpvn>, in F<sv.c>
782 2 Perl_sv_catpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
787 6 junk = SvPV_force(sv, tlen);
788 7 SvGROW(sv, tlen + len + 1);
791 10 Move(ptr,SvPVX(sv)+tlen,len,char);
793 12 *SvEND(sv) = '\0';
794 13 (void)SvPOK_only_UTF8(sv); /* validate pointer */
798 This is a function which adds a string, C<ptr>, of length C<len> onto
799 the end of the PV stored in C<sv>. The first thing we do in line 6 is
800 make sure that the SV B<has> a valid PV, by calling the C<SvPV_force>
801 macro to force a PV. As a side effect, C<tlen> gets set to the current
802 value of the PV, and the PV itself is returned to C<junk>.
804 In line 7, we make sure that the SV will have enough room to accommodate
805 the old string, the new string and the null terminator. If C<LEN> isn't
806 big enough, C<SvGROW> will reallocate space for us.
808 Now, if C<junk> is the same as the string we're trying to add, we can
809 grab the string directly from the SV; C<SvPVX> is the address of the PV
812 Line 10 does the actual catenation: the C<Move> macro moves a chunk of
813 memory around: we move the string C<ptr> to the end of the PV - that's
814 the start of the PV plus its current length. We're moving C<len> bytes
815 of type C<char>. After doing so, we need to tell Perl we've extended the
816 string, by altering C<CUR> to reflect the new length. C<SvEND> is a
817 macro which gives us the end of the string, so that needs to be a
820 Line 13 manipulates the flags; since we've changed the PV, any IV or NV
821 values will no longer be valid: if we have C<$a=10; $a.="6";> we don't
822 want to use the old IV of 10. C<SvPOK_only_utf8> is a special UTF-8-aware
823 version of C<SvPOK_only>, a macro which turns off the IOK and NOK flags
824 and turns on POK. The final C<SvTAINT> is a macro which launders tainted
825 data if taint mode is turned on.
827 AVs and HVs are more complicated, but SVs are by far the most common
828 variable type being thrown around. Having seen something of how we
829 manipulate these, let's go on and look at how the op tree is
834 First, what is the op tree, anyway? The op tree is the parsed
835 representation of your program, as we saw in our section on parsing, and
836 it's the sequence of operations that Perl goes through to execute your
837 program, as we saw in L</Running>.
839 An op is a fundamental operation that Perl can perform: all the built-in
840 functions and operators are ops, and there are a series of ops which
841 deal with concepts the interpreter needs internally - entering and
842 leaving a block, ending a statement, fetching a variable, and so on.
844 The op tree is connected in two ways: you can imagine that there are two
845 "routes" through it, two orders in which you can traverse the tree.
846 First, parse order reflects how the parser understood the code, and
847 secondly, execution order tells perl what order to perform the
850 The easiest way to examine the op tree is to stop Perl after it has
851 finished parsing, and get it to dump out the tree. This is exactly what
852 the compiler backends L<B::Terse|B::Terse>, L<B::Concise|B::Concise>
853 and L<B::Debug|B::Debug> do.
855 Let's have a look at how Perl sees C<$a = $b + $c>:
857 % perl -MO=Terse -e '$a=$b+$c'
858 1 LISTOP (0x8179888) leave
859 2 OP (0x81798b0) enter
860 3 COP (0x8179850) nextstate
861 4 BINOP (0x8179828) sassign
862 5 BINOP (0x8179800) add [1]
863 6 UNOP (0x81796e0) null [15]
864 7 SVOP (0x80fafe0) gvsv GV (0x80fa4cc) *b
865 8 UNOP (0x81797e0) null [15]
866 9 SVOP (0x8179700) gvsv GV (0x80efeb0) *c
867 10 UNOP (0x816b4f0) null [15]
868 11 SVOP (0x816dcf0) gvsv GV (0x80fa460) *a
870 Let's start in the middle, at line 4. This is a BINOP, a binary
871 operator, which is at location C<0x8179828>. The specific operator in
872 question is C<sassign> - scalar assignment - and you can find the code
873 which implements it in the function C<pp_sassign> in F<pp_hot.c>. As a
874 binary operator, it has two children: the add operator, providing the
875 result of C<$b+$c>, is uppermost on line 5, and the left hand side is on
878 Line 10 is the null op: this does exactly nothing. What is that doing
879 there? If you see the null op, it's a sign that something has been
880 optimized away after parsing. As we mentioned in L</Optimization>,
881 the optimization stage sometimes converts two operations into one, for
882 example when fetching a scalar variable. When this happens, instead of
883 rewriting the op tree and cleaning up the dangling pointers, it's easier
884 just to replace the redundant operation with the null op. Originally,
885 the tree would have looked like this:
887 10 SVOP (0x816b4f0) rv2sv [15]
888 11 SVOP (0x816dcf0) gv GV (0x80fa460) *a
890 That is, fetch the C<a> entry from the main symbol table, and then look
891 at the scalar component of it: C<gvsv> (C<pp_gvsv> into F<pp_hot.c>)
892 happens to do both these things.
894 The right hand side, starting at line 5 is similar to what we've just
895 seen: we have the C<add> op (C<pp_add> also in F<pp_hot.c>) add together
898 Now, what's this about?
900 1 LISTOP (0x8179888) leave
901 2 OP (0x81798b0) enter
902 3 COP (0x8179850) nextstate
904 C<enter> and C<leave> are scoping ops, and their job is to perform any
905 housekeeping every time you enter and leave a block: lexical variables
906 are tidied up, unreferenced variables are destroyed, and so on. Every
907 program will have those first three lines: C<leave> is a list, and its
908 children are all the statements in the block. Statements are delimited
909 by C<nextstate>, so a block is a collection of C<nextstate> ops, with
910 the ops to be performed for each statement being the children of
911 C<nextstate>. C<enter> is a single op which functions as a marker.
913 That's how Perl parsed the program, from top to bottom:
926 However, it's impossible to B<perform> the operations in this order:
927 you have to find the values of C<$b> and C<$c> before you add them
928 together, for instance. So, the other thread that runs through the op
929 tree is the execution order: each op has a field C<op_next> which points
930 to the next op to be run, so following these pointers tells us how perl
931 executes the code. We can traverse the tree in this order using
932 the C<exec> option to C<B::Terse>:
934 % perl -MO=Terse,exec -e '$a=$b+$c'
935 1 OP (0x8179928) enter
936 2 COP (0x81798c8) nextstate
937 3 SVOP (0x81796c8) gvsv GV (0x80fa4d4) *b
938 4 SVOP (0x8179798) gvsv GV (0x80efeb0) *c
939 5 BINOP (0x8179878) add [1]
940 6 SVOP (0x816dd38) gvsv GV (0x80fa468) *a
941 7 BINOP (0x81798a0) sassign
942 8 LISTOP (0x8179900) leave
944 This probably makes more sense for a human: enter a block, start a
945 statement. Get the values of C<$b> and C<$c>, and add them together.
946 Find C<$a>, and assign one to the other. Then leave.
948 The way Perl builds up these op trees in the parsing process can be
949 unravelled by examining F<perly.y>, the YACC grammar. Let's take the
950 piece we need to construct the tree for C<$a = $b + $c>
952 1 term : term ASSIGNOP term
953 2 { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
955 4 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
957 If you're not used to reading BNF grammars, this is how it works: You're
958 fed certain things by the tokeniser, which generally end up in upper
959 case. Here, C<ADDOP>, is provided when the tokeniser sees C<+> in your
960 code. C<ASSIGNOP> is provided when C<=> is used for assigning. These are
961 "terminal symbols", because you can't get any simpler than them.
963 The grammar, lines one and three of the snippet above, tells you how to
964 build up more complex forms. These complex forms, "non-terminal symbols"
965 are generally placed in lower case. C<term> here is a non-terminal
966 symbol, representing a single expression.
968 The grammar gives you the following rule: you can make the thing on the
969 left of the colon if you see all the things on the right in sequence.
970 This is called a "reduction", and the aim of parsing is to completely
971 reduce the input. There are several different ways you can perform a
972 reduction, separated by vertical bars: so, C<term> followed by C<=>
973 followed by C<term> makes a C<term>, and C<term> followed by C<+>
974 followed by C<term> can also make a C<term>.
976 So, if you see two terms with an C<=> or C<+>, between them, you can
977 turn them into a single expression. When you do this, you execute the
978 code in the block on the next line: if you see C<=>, you'll do the code
979 in line 2. If you see C<+>, you'll do the code in line 4. It's this code
980 which contributes to the op tree.
983 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
985 What this does is creates a new binary op, and feeds it a number of
986 variables. The variables refer to the tokens: C<$1> is the first token in
987 the input, C<$2> the second, and so on - think regular expression
988 backreferences. C<$$> is the op returned from this reduction. So, we
989 call C<newBINOP> to create a new binary operator. The first parameter to
990 C<newBINOP>, a function in F<op.c>, is the op type. It's an addition
991 operator, so we want the type to be C<ADDOP>. We could specify this
992 directly, but it's right there as the second token in the input, so we
993 use C<$2>. The second parameter is the op's flags: 0 means "nothing
994 special". Then the things to add: the left and right hand side of our
995 expression, in scalar context.
999 When perl executes something like C<addop>, how does it pass on its
1000 results to the next op? The answer is, through the use of stacks. Perl
1001 has a number of stacks to store things it's currently working on, and
1002 we'll look at the three most important ones here.
1006 =item Argument stack
1008 Arguments are passed to PP code and returned from PP code using the
1009 argument stack, C<ST>. The typical way to handle arguments is to pop
1010 them off the stack, deal with them how you wish, and then push the result
1011 back onto the stack. This is how, for instance, the cosine operator
1016 value = Perl_cos(value);
1019 We'll see a more tricky example of this when we consider Perl's macros
1020 below. C<POPn> gives you the NV (floating point value) of the top SV on
1021 the stack: the C<$x> in C<cos($x)>. Then we compute the cosine, and push
1022 the result back as an NV. The C<X> in C<XPUSHn> means that the stack
1023 should be extended if necessary - it can't be necessary here, because we
1024 know there's room for one more item on the stack, since we've just
1025 removed one! The C<XPUSH*> macros at least guarantee safety.
1027 Alternatively, you can fiddle with the stack directly: C<SP> gives you
1028 the first element in your portion of the stack, and C<TOP*> gives you
1029 the top SV/IV/NV/etc. on the stack. So, for instance, to do unary
1030 negation of an integer:
1034 Just set the integer value of the top stack entry to its negation.
1036 Argument stack manipulation in the core is exactly the same as it is in
1037 XSUBs - see L<perlxstut>, L<perlxs> and L<perlguts> for a longer
1038 description of the macros used in stack manipulation.
1042 I say "your portion of the stack" above because PP code doesn't
1043 necessarily get the whole stack to itself: if your function calls
1044 another function, you'll only want to expose the arguments aimed for the
1045 called function, and not (necessarily) let it get at your own data. The
1046 way we do this is to have a "virtual" bottom-of-stack, exposed to each
1047 function. The mark stack keeps bookmarks to locations in the argument
1048 stack usable by each function. For instance, when dealing with a tied
1049 variable, (internally, something with "P" magic) Perl has to call
1050 methods for accesses to the tied variables. However, we need to separate
1051 the arguments exposed to the method to the argument exposed to the
1052 original function - the store or fetch or whatever it may be. Here's
1053 roughly how the tied C<push> is implemented; see C<av_push> in F<av.c>:
1057 3 PUSHs(SvTIED_obj((SV*)av, mg));
1061 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1064 Let's examine the whole implementation, for practice:
1068 Push the current state of the stack pointer onto the mark stack. This is
1069 so that when we've finished adding items to the argument stack, Perl
1070 knows how many things we've added recently.
1073 3 PUSHs(SvTIED_obj((SV*)av, mg));
1076 We're going to add two more items onto the argument stack: when you have
1077 a tied array, the C<PUSH> subroutine receives the object and the value
1078 to be pushed, and that's exactly what we have here - the tied object,
1079 retrieved with C<SvTIED_obj>, and the value, the SV C<val>.
1083 Next we tell Perl to update the global stack pointer from our internal
1084 variable: C<dSP> only gave us a local copy, not a reference to the global.
1087 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1090 C<ENTER> and C<LEAVE> localise a block of code - they make sure that all
1091 variables are tidied up, everything that has been localised gets
1092 its previous value returned, and so on. Think of them as the C<{> and
1093 C<}> of a Perl block.
1095 To actually do the magic method call, we have to call a subroutine in
1096 Perl space: C<call_method> takes care of that, and it's described in
1097 L<perlcall>. We call the C<PUSH> method in scalar context, and we're
1098 going to discard its return value. The call_method() function
1099 removes the top element of the mark stack, so there is nothing for
1100 the caller to clean up.
1104 C doesn't have a concept of local scope, so perl provides one. We've
1105 seen that C<ENTER> and C<LEAVE> are used as scoping braces; the save
1106 stack implements the C equivalent of, for example:
1113 See L<perlguts/Localising Changes> for how to use the save stack.
1117 =head2 Millions of Macros
1119 One thing you'll notice about the Perl source is that it's full of
1120 macros. Some have called the pervasive use of macros the hardest thing
1121 to understand, others find it adds to clarity. Let's take an example,
1122 the code which implements the addition operator:
1126 3 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
1129 6 SETn( left + right );
1134 Every line here (apart from the braces, of course) contains a macro. The
1135 first line sets up the function declaration as Perl expects for PP code;
1136 line 3 sets up variable declarations for the argument stack and the
1137 target, the return value of the operation. Finally, it tries to see if
1138 the addition operation is overloaded; if so, the appropriate subroutine
1141 Line 5 is another variable declaration - all variable declarations start
1142 with C<d> - which pops from the top of the argument stack two NVs (hence
1143 C<nn>) and puts them into the variables C<right> and C<left>, hence the
1144 C<rl>. These are the two operands to the addition operator. Next, we
1145 call C<SETn> to set the NV of the return value to the result of adding
1146 the two values. This done, we return - the C<RETURN> macro makes sure
1147 that our return value is properly handled, and we pass the next operator
1148 to run back to the main run loop.
1150 Most of these macros are explained in L<perlapi>, and some of the more
1151 important ones are explained in L<perlxs> as well. Pay special attention
1152 to L<perlguts/Background and PERL_IMPLICIT_CONTEXT> for information on
1153 the C<[pad]THX_?> macros.
1155 =head2 The .i Targets
1157 You can expand the macros in a F<foo.c> file by saying
1161 which will expand the macros using cpp. Don't be scared by the results.
1163 =head1 SOURCE CODE STATIC ANALYSIS
1165 Various tools exist for analysing C source code B<statically>, as
1166 opposed to B<dynamically>, that is, without executing the code.
1167 It is possible to detect resource leaks, undefined behaviour, type
1168 mismatches, portability problems, code paths that would cause illegal
1169 memory accesses, and other similar problems by just parsing the C code
1170 and looking at the resulting graph, what does it tell about the
1171 execution and data flows. As a matter of fact, this is exactly
1172 how C compilers know to give warnings about dubious code.
1176 The good old C code quality inspector, C<lint>, is available in
1177 several platforms, but please be aware that there are several
1178 different implementations of it by different vendors, which means that
1179 the flags are not identical across different platforms.
1181 There is a lint variant called C<splint> (Secure Programming Lint)
1182 available from http://www.splint.org/ that should compile on any
1185 There are C<lint> and <splint> targets in Makefile, but you may have
1186 to diddle with the flags (see above).
1190 Coverity (http://www.coverity.com/) is a product similar to lint and
1191 as a testbed for their product they periodically check several open
1192 source projects, and they give out accounts to open source developers
1193 to the defect databases.
1195 =head2 cpd (cut-and-paste detector)
1197 The cpd tool detects cut-and-paste coding. If one instance of the
1198 cut-and-pasted code changes, all the other spots should probably be
1199 changed, too. Therefore such code should probably be turned into a
1200 subroutine or a macro.
1202 cpd (http://pmd.sourceforge.net/cpd.html) is part of the pmd project
1203 (http://pmd.sourceforge.net/). pmd was originally written for static
1204 analysis of Java code, but later the cpd part of it was extended to
1205 parse also C and C++.
1207 Download the pmd-bin-X.Y.zip () from the SourceForge site, extract the
1208 pmd-X.Y.jar from it, and then run that on source code thusly:
1210 java -cp pmd-X.Y.jar net.sourceforge.pmd.cpd.CPD --minimum-tokens 100 --files /some/where/src --language c > cpd.txt
1212 You may run into memory limits, in which case you should use the -Xmx option:
1218 Though much can be written about the inconsistency and coverage
1219 problems of gcc warnings (like C<-Wall> not meaning "all the
1220 warnings", or some common portability problems not being covered by
1221 C<-Wall>, or C<-ansi> and C<-pedantic> both being a poorly defined
1222 collection of warnings, and so forth), gcc is still a useful tool in
1223 keeping our coding nose clean.
1225 The C<-Wall> is by default on.
1227 The C<-ansi> (and its sidekick, C<-pedantic>) would be nice to be on
1228 always, but unfortunately they are not safe on all platforms, they can
1229 for example cause fatal conflicts with the system headers (Solaris
1230 being a prime example). If Configure C<-Dgccansipedantic> is used,
1231 the C<cflags> frontend selects C<-ansi -pedantic> for the platforms
1232 where they are known to be safe.
1234 Starting from Perl 5.9.4 the following extra flags are added:
1248 C<-Wdeclaration-after-statement>
1252 The following flags would be nice to have but they would first need
1253 their own Augean stablemaster:
1267 C<-Wstrict-prototypes>
1271 The C<-Wtraditional> is another example of the annoying tendency of
1272 gcc to bundle a lot of warnings under one switch -- it would be
1273 impossible to deploy in practice because it would complain a lot -- but
1274 it does contain some warnings that would be beneficial to have available
1275 on their own, such as the warning about string constants inside macros
1276 containing the macro arguments: this behaved differently pre-ANSI
1277 than it does in ANSI, and some C compilers are still in transition,
1278 AIX being an example.
1280 =head2 Warnings of other C compilers
1282 Other C compilers (yes, there B<are> other C compilers than gcc) often
1283 have their "strict ANSI" or "strict ANSI with some portability extensions"
1284 modes on, like for example the Sun Workshop has its C<-Xa> mode on
1285 (though implicitly), or the DEC (these days, HP...) has its C<-std1>
1290 You can compile a special debugging version of Perl, which allows you
1291 to use the C<-D> option of Perl to tell more about what Perl is doing.
1292 But sometimes there is no alternative than to dive in with a debugger,
1293 either to see the stack trace of a core dump (very useful in a bug
1294 report), or trying to figure out what went wrong before the core dump
1295 happened, or how did we end up having wrong or unexpected results.
1297 =head2 Poking at Perl
1299 To really poke around with Perl, you'll probably want to build Perl for
1300 debugging, like this:
1302 ./Configure -d -D optimize=-g
1305 C<-g> is a flag to the C compiler to have it produce debugging
1306 information which will allow us to step through a running program,
1307 and to see in which C function we are at (without the debugging
1308 information we might see only the numerical addresses of the functions,
1309 which is not very helpful).
1311 F<Configure> will also turn on the C<DEBUGGING> compilation symbol which
1312 enables all the internal debugging code in Perl. There are a whole bunch
1313 of things you can debug with this: L<perlrun> lists them all, and the
1314 best way to find out about them is to play about with them. The most
1315 useful options are probably
1317 l Context (loop) stack processing
1319 o Method and overloading resolution
1320 c String/numeric conversions
1322 Some of the functionality of the debugging code can be achieved using XS
1325 -Dr => use re 'debug'
1326 -Dx => use O 'Debug'
1328 =head2 Using a source-level debugger
1330 If the debugging output of C<-D> doesn't help you, it's time to step
1331 through perl's execution with a source-level debugger.
1337 We'll use C<gdb> for our examples here; the principles will apply to
1338 any debugger (many vendors call their debugger C<dbx>), but check the
1339 manual of the one you're using.
1343 To fire up the debugger, type
1347 Or if you have a core dump:
1351 You'll want to do that in your Perl source tree so the debugger can read
1352 the source code. You should see the copyright message, followed by the
1357 C<help> will get you into the documentation, but here are the most
1364 Run the program with the given arguments.
1366 =item break function_name
1368 =item break source.c:xxx
1370 Tells the debugger that we'll want to pause execution when we reach
1371 either the named function (but see L<perlguts/Internal Functions>!) or the given
1372 line in the named source file.
1376 Steps through the program a line at a time.
1380 Steps through the program a line at a time, without descending into
1385 Run until the next breakpoint.
1389 Run until the end of the current function, then stop again.
1393 Just pressing Enter will do the most recent operation again - it's a
1394 blessing when stepping through miles of source code.
1398 Execute the given C code and print its results. B<WARNING>: Perl makes
1399 heavy use of macros, and F<gdb> does not necessarily support macros
1400 (see later L</"gdb macro support">). You'll have to substitute them
1401 yourself, or to invoke cpp on the source code files
1402 (see L</"The .i Targets">)
1403 So, for instance, you can't say
1405 print SvPV_nolen(sv)
1409 print Perl_sv_2pv_nolen(sv)
1413 You may find it helpful to have a "macro dictionary", which you can
1414 produce by saying C<cpp -dM perl.c | sort>. Even then, F<cpp> won't
1415 recursively apply those macros for you.
1417 =head2 gdb macro support
1419 Recent versions of F<gdb> have fairly good macro support, but
1420 in order to use it you'll need to compile perl with macro definitions
1421 included in the debugging information. Using F<gcc> version 3.1, this
1422 means configuring with C<-Doptimize=-g3>. Other compilers might use a
1423 different switch (if they support debugging macros at all).
1425 =head2 Dumping Perl Data Structures
1427 One way to get around this macro hell is to use the dumping functions in
1428 F<dump.c>; these work a little like an internal
1429 L<Devel::Peek|Devel::Peek>, but they also cover OPs and other structures
1430 that you can't get at from Perl. Let's take an example. We'll use the
1431 C<$a = $b + $c> we used before, but give it a bit of context:
1432 C<$b = "6XXXX"; $c = 2.3;>. Where's a good place to stop and poke around?
1434 What about C<pp_add>, the function we examined earlier to implement the
1437 (gdb) break Perl_pp_add
1438 Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.
1440 Notice we use C<Perl_pp_add> and not C<pp_add> - see L<perlguts/Internal Functions>.
1441 With the breakpoint in place, we can run our program:
1443 (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'
1445 Lots of junk will go past as gdb reads in the relevant source files and
1446 libraries, and then:
1448 Breakpoint 1, Perl_pp_add () at pp_hot.c:309
1449 309 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
1454 We looked at this bit of code before, and we said that C<dPOPTOPnnrl_ul>
1455 arranges for two C<NV>s to be placed into C<left> and C<right> - let's
1458 #define dPOPTOPnnrl_ul NV right = POPn; \
1459 SV *leftsv = TOPs; \
1460 NV left = USE_LEFT(leftsv) ? SvNV(leftsv) : 0.0
1462 C<POPn> takes the SV from the top of the stack and obtains its NV either
1463 directly (if C<SvNOK> is set) or by calling the C<sv_2nv> function.
1464 C<TOPs> takes the next SV from the top of the stack - yes, C<POPn> uses
1465 C<TOPs> - but doesn't remove it. We then use C<SvNV> to get the NV from
1466 C<leftsv> in the same way as before - yes, C<POPn> uses C<SvNV>.
1468 Since we don't have an NV for C<$b>, we'll have to use C<sv_2nv> to
1469 convert it. If we step again, we'll find ourselves there:
1471 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
1475 We can now use C<Perl_sv_dump> to investigate the SV:
1477 SV = PV(0xa057cc0) at 0xa0675d0
1480 PV = 0xa06a510 "6XXXX"\0
1485 We know we're going to get C<6> from this, so let's finish the
1489 Run till exit from #0 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
1490 0x462669 in Perl_pp_add () at pp_hot.c:311
1493 We can also dump out this op: the current op is always stored in
1494 C<PL_op>, and we can dump it with C<Perl_op_dump>. This'll give us
1495 similar output to L<B::Debug|B::Debug>.
1498 13 TYPE = add ===> 14
1500 FLAGS = (SCALAR,KIDS)
1502 TYPE = null ===> (12)
1504 FLAGS = (SCALAR,KIDS)
1506 11 TYPE = gvsv ===> 12
1512 # finish this later #
1516 All right, we've now had a look at how to navigate the Perl sources and
1517 some things you'll need to know when fiddling with them. Let's now get
1518 on and create a simple patch. Here's something Larry suggested: if a
1519 C<U> is the first active format during a C<pack>, (for example,
1520 C<pack "U3C8", @stuff>) then the resulting string should be treated as
1523 If you are working with a git clone of the Perl repository, you will want to
1524 create a branch for your changes. This will make creating a proper patch much
1525 simpler. See the L<perlrepository> for details on how to do this.
1527 How do we prepare to fix this up? First we locate the code in question -
1528 the C<pack> happens at runtime, so it's going to be in one of the F<pp>
1529 files. Sure enough, C<pp_pack> is in F<pp.c>. Since we're going to be
1530 altering this file, let's copy it to F<pp.c~>.
1532 [Well, it was in F<pp.c> when this tutorial was written. It has now been
1533 split off with C<pp_unpack> to its own file, F<pp_pack.c>]
1535 Now let's look over C<pp_pack>: we take a pattern into C<pat>, and then
1536 loop over the pattern, taking each format character in turn into
1537 C<datum_type>. Then for each possible format character, we swallow up
1538 the other arguments in the pattern (a field width, an asterisk, and so
1539 on) and convert the next chunk input into the specified format, adding
1540 it onto the output SV C<cat>.
1542 How do we know if the C<U> is the first format in the C<pat>? Well, if
1543 we have a pointer to the start of C<pat> then, if we see a C<U> we can
1544 test whether we're still at the start of the string. So, here's where
1548 register char *pat = SvPVx(*++MARK, fromlen);
1549 register char *patend = pat + fromlen;
1554 We'll have another string pointer in there:
1557 register char *pat = SvPVx(*++MARK, fromlen);
1558 register char *patend = pat + fromlen;
1564 And just before we start the loop, we'll set C<patcopy> to be the start
1569 sv_setpvn(cat, "", 0);
1571 while (pat < patend) {
1573 Now if we see a C<U> which was at the start of the string, we turn on
1574 the C<UTF8> flag for the output SV, C<cat>:
1576 + if (datumtype == 'U' && pat==patcopy+1)
1578 if (datumtype == '#') {
1579 while (pat < patend && *pat != '\n')
1582 Remember that it has to be C<patcopy+1> because the first character of
1583 the string is the C<U> which has been swallowed into C<datumtype!>
1585 Oops, we forgot one thing: what if there are spaces at the start of the
1586 pattern? C<pack(" U*", @stuff)> will have C<U> as the first active
1587 character, even though it's not the first thing in the pattern. In this
1588 case, we have to advance C<patcopy> along with C<pat> when we see spaces:
1590 if (isSPACE(datumtype))
1595 if (isSPACE(datumtype)) {
1600 OK. That's the C part done. Now we must do two additional things before
1601 this patch is ready to go: we've changed the behaviour of Perl, and so
1602 we must document that change. We must also provide some more regression
1603 tests to make sure our patch works and doesn't create a bug somewhere
1604 else along the line.
1606 The regression tests for each operator live in F<t/op/>, and so we
1607 make a copy of F<t/op/pack.t> to F<t/op/pack.t~>. Now we can add our
1608 tests to the end. First, we'll test that the C<U> does indeed create
1611 t/op/pack.t has a sensible ok() function, but if it didn't we could
1612 use the one from t/test.pl.
1614 require './test.pl';
1615 plan( tests => 159 );
1619 print 'not ' unless "1.20.300.4000" eq sprintf "%vd", pack("U*",1,20,300,4000);
1620 print "ok $test\n"; $test++;
1622 we can write the more sensible (see L<Test::More> for a full
1623 explanation of is() and other testing functions).
1625 is( "1.20.300.4000", sprintf "%vd", pack("U*",1,20,300,4000),
1626 "U* produces Unicode" );
1628 Now we'll test that we got that space-at-the-beginning business right:
1630 is( "1.20.300.4000", sprintf "%vd", pack(" U*",1,20,300,4000),
1631 " with spaces at the beginning" );
1633 And finally we'll test that we don't make Unicode strings if C<U> is B<not>
1634 the first active format:
1636 isnt( v1.20.300.4000, sprintf "%vd", pack("C0U*",1,20,300,4000),
1637 "U* not first isn't Unicode" );
1639 Mustn't forget to change the number of tests which appears at the top,
1640 or else the automated tester will get confused. This will either look
1647 plan( tests => 156 );
1649 We now compile up Perl, and run it through the test suite. Our new
1652 Finally, the documentation. The job is never done until the paperwork is
1653 over, so let's describe the change we've just made. The relevant place
1654 is F<pod/perlfunc.pod>; again, we make a copy, and then we'll insert
1655 this text in the description of C<pack>:
1659 If the pattern begins with a C<U>, the resulting string will be treated
1660 as UTF-8-encoded Unicode. You can force UTF-8 encoding on in a string
1661 with an initial C<U0>, and the bytes that follow will be interpreted as
1662 Unicode characters. If you don't want this to happen, you can begin your
1663 pattern with C<C0> (or anything else) to force Perl not to UTF-8 encode your
1664 string, and then follow this with a C<U*> somewhere in your pattern.
1666 =head2 Patching a core module
1668 This works just like patching anything else, with an extra
1669 consideration. Many core modules also live on CPAN. If this is so,
1670 patch the CPAN version instead of the core and send the patch off to
1671 the module maintainer (with a copy to p5p). This will help the module
1672 maintainer keep the CPAN version in sync with the core version without
1673 constantly scanning p5p.
1675 The list of maintainers of core modules is usefully documented in
1676 F<Porting/Maintainers.pl>.
1678 =head2 Adding a new function to the core
1680 If, as part of a patch to fix a bug, or just because you have an
1681 especially good idea, you decide to add a new function to the core,
1682 discuss your ideas on p5p well before you start work. It may be that
1683 someone else has already attempted to do what you are considering and
1684 can give lots of good advice or even provide you with bits of code
1685 that they already started (but never finished).
1687 You have to follow all of the advice given above for patching. It is
1688 extremely important to test any addition thoroughly and add new tests
1689 to explore all boundary conditions that your new function is expected
1690 to handle. If your new function is used only by one module (e.g. toke),
1691 then it should probably be named S_your_function (for static); on the
1692 other hand, if you expect it to accessible from other functions in
1693 Perl, you should name it Perl_your_function. See L<perlguts/Internal Functions>
1696 The location of any new code is also an important consideration. Don't
1697 just create a new top level .c file and put your code there; you would
1698 have to make changes to Configure (so the Makefile is created properly),
1699 as well as possibly lots of include files. This is strictly pumpking
1702 It is better to add your function to one of the existing top level
1703 source code files, but your choice is complicated by the nature of
1704 the Perl distribution. Only the files that are marked as compiled
1705 static are located in the perl executable. Everything else is located
1706 in the shared library (or DLL if you are running under WIN32). So,
1707 for example, if a function was only used by functions located in
1708 toke.c, then your code can go in toke.c. If, however, you want to call
1709 the function from universal.c, then you should put your code in another
1710 location, for example util.c.
1712 In addition to writing your c-code, you will need to create an
1713 appropriate entry in embed.pl describing your function, then run
1714 'make regen_headers' to create the entries in the numerous header
1715 files that perl needs to compile correctly. See L<perlguts/Internal Functions>
1716 for information on the various options that you can set in embed.pl.
1717 You will forget to do this a few (or many) times and you will get
1718 warnings during the compilation phase. Make sure that you mention
1719 this when you post your patch to P5P; the pumpking needs to know this.
1721 When you write your new code, please be conscious of existing code
1722 conventions used in the perl source files. See L<perlstyle> for
1723 details. Although most of the guidelines discussed seem to focus on
1724 Perl code, rather than c, they all apply (except when they don't ;).
1725 Also see I<perlrepository> for lots of details about both formatting and
1726 submitting patches of your changes.
1728 Lastly, TEST TEST TEST TEST TEST any code before posting to p5p.
1729 Test on as many platforms as you can find. Test as many perl
1730 Configure options as you can (e.g. MULTIPLICITY). If you have
1731 profiling or memory tools, see L<EXTERNAL TOOLS FOR DEBUGGING PERL>
1732 below for how to use them to further test your code. Remember that
1733 most of the people on P5P are doing this on their own time and
1734 don't have the time to debug your code.
1736 =head2 Writing a test
1738 Every module and built-in function has an associated test file (or
1739 should...). If you add or change functionality, you have to write a
1740 test. If you fix a bug, you have to write a test so that bug never
1741 comes back. If you alter the docs, it would be nice to test what the
1742 new documentation says.
1744 In short, if you submit a patch you probably also have to patch the
1747 For modules, the test file is right next to the module itself.
1748 F<lib/strict.t> tests F<lib/strict.pm>. This is a recent innovation,
1749 so there are some snags (and it would be wonderful for you to brush
1750 them out), but it basically works that way. Everything else lives in
1753 If you add a new test directory under F<t/>, it is imperative that you
1754 add that directory to F<t/HARNESS> and F<t/TEST>.
1760 Testing of the absolute basic functionality of Perl. Things like
1761 C<if>, basic file reads and writes, simple regexes, etc. These are
1762 run first in the test suite and if any of them fail, something is
1767 These test the basic control structures, C<if/else>, C<while>,
1772 Tests basic issues of how Perl parses and compiles itself.
1776 Tests for built-in IO functions, including command line arguments.
1780 The old home for the module tests, you shouldn't put anything new in
1781 here. There are still some bits and pieces hanging around in here
1782 that need to be moved. Perhaps you could move them? Thanks!
1786 Tests for perl's method resolution order implementations
1791 Tests for perl's built in functions that don't fit into any of the
1796 Tests for regex related functions or behaviour. (These used to live
1801 Testing features of how perl actually runs, including exit codes and
1802 handling of PERL* environment variables.
1806 Tests for the core support of Unicode.
1810 Windows-specific tests.
1814 A test suite for the s2p converter.
1818 The core uses the same testing style as the rest of Perl, a simple
1819 "ok/not ok" run through Test::Harness, but there are a few special
1822 There are three ways to write a test in the core. Test::More,
1823 t/test.pl and ad hoc C<print $test ? "ok 42\n" : "not ok 42\n">. The
1824 decision of which to use depends on what part of the test suite you're
1825 working on. This is a measure to prevent a high-level failure (such
1826 as Config.pm breaking) from causing basic functionality tests to fail.
1832 Since we don't know if require works, or even subroutines, use ad hoc
1833 tests for these two. Step carefully to avoid using the feature being
1836 =item t/cmd t/run t/io t/op
1838 Now that basic require() and subroutines are tested, you can use the
1839 t/test.pl library which emulates the important features of Test::More
1840 while using a minimum of core features.
1842 You can also conditionally use certain libraries like Config, but be
1843 sure to skip the test gracefully if it's not there.
1847 Now that the core of Perl is tested, Test::More can be used. You can
1848 also use the full suite of core modules in the tests.
1852 When you say "make test" Perl uses the F<t/TEST> program to run the
1853 test suite (except under Win32 where it uses F<t/harness> instead.)
1854 All tests are run from the F<t/> directory, B<not> the directory
1855 which contains the test. This causes some problems with the tests
1856 in F<lib/>, so here's some opportunity for some patching.
1858 You must be triply conscious of cross-platform concerns. This usually
1859 boils down to using File::Spec and avoiding things like C<fork()> and
1860 C<system()> unless absolutely necessary.
1862 =head2 Special Make Test Targets
1864 There are various special make targets that can be used to test Perl
1865 slightly differently than the standard "test" target. Not all them
1866 are expected to give a 100% success rate. Many of them have several
1867 aliases, and many of them are not available on certain operating
1874 Run F<perl> on all core tests (F<t/*> and F<lib/[a-z]*> pragma tests).
1876 (Not available on Win32)
1880 Run all the tests through B::Deparse. Not all tests will succeed.
1882 (Not available on Win32)
1884 =item test.taintwarn
1886 Run all tests with the B<-t> command-line switch. Not all tests
1887 are expected to succeed (until they're specifically fixed, of course).
1889 (Not available on Win32)
1893 Run F<miniperl> on F<t/base>, F<t/comp>, F<t/cmd>, F<t/run>, F<t/io>,
1894 F<t/op>, F<t/uni> and F<t/mro> tests.
1896 =item test.valgrind check.valgrind utest.valgrind ucheck.valgrind
1898 (Only in Linux) Run all the tests using the memory leak + naughty
1899 memory access tool "valgrind". The log files will be named
1900 F<testname.valgrind>.
1902 =item test.third check.third utest.third ucheck.third
1904 (Only in Tru64) Run all the tests using the memory leak + naughty
1905 memory access tool "Third Degree". The log files will be named
1906 F<perl.3log.testname>.
1908 =item test.torture torturetest
1910 Run all the usual tests and some extra tests. As of Perl 5.8.0 the
1911 only extra tests are Abigail's JAPHs, F<t/japh/abigail.t>.
1913 You can also run the torture test with F<t/harness> by giving
1914 C<-torture> argument to F<t/harness>.
1916 =item utest ucheck test.utf8 check.utf8
1918 Run all the tests with -Mutf8. Not all tests will succeed.
1920 (Not available on Win32)
1922 =item minitest.utf16 test.utf16
1924 Runs the tests with UTF-16 encoded scripts, encoded with different
1925 versions of this encoding.
1927 C<make utest.utf16> runs the test suite with a combination of C<-utf8> and
1928 C<-utf16> arguments to F<t/TEST>.
1930 (Not available on Win32)
1934 Run the test suite with the F<t/harness> controlling program, instead of
1935 F<t/TEST>. F<t/harness> is more sophisticated, and uses the
1936 L<Test::Harness> module, thus using this test target supposes that perl
1937 mostly works. The main advantage for our purposes is that it prints a
1938 detailed summary of failed tests at the end. Also, unlike F<t/TEST>, it
1939 doesn't redirect stderr to stdout.
1941 Note that under Win32 F<t/harness> is always used instead of F<t/TEST>, so
1942 there is no special "test_harness" target.
1944 Under Win32's "test" target you may use the TEST_SWITCHES and TEST_FILES
1945 environment variables to control the behaviour of F<t/harness>. This means
1948 nmake test TEST_FILES="op/*.t"
1949 nmake test TEST_SWITCHES="-torture" TEST_FILES="op/*.t"
1951 =item Parallel tests
1953 The core distribution can now run its regression tests in parallel on
1954 Unix-like platforms. Instead of running C<make test>, set C<TEST_JOBS> in
1955 your environment to the number of tests to run in parallel, and run
1956 C<make test_harness>. On a Bourne-like shell, this can be done as
1958 TEST_JOBS=3 make test_harness # Run 3 tests in parallel
1960 An environment variable is used, rather than parallel make itself, because
1961 L<TAP::Harness> needs to be able to schedule individual non-conflicting test
1962 scripts itself, and there is no standard interface to C<make> utilities to
1963 interact with their job schedulers.
1965 Note that currently some test scripts may fail when run in parallel (most
1966 notably C<ext/IO/t/io_dir.t>). If necessary run just the failing scripts
1967 again sequentially and see if the failures go away.
1968 =item test-notty test_notty
1970 Sets PERL_SKIP_TTY_TEST to true before running normal test.
1974 =head2 Running tests by hand
1976 You can run part of the test suite by hand by using one the following
1977 commands from the F<t/> directory :
1979 ./perl -I../lib TEST list-of-.t-files
1983 ./perl -I../lib harness list-of-.t-files
1985 (if you don't specify test scripts, the whole test suite will be run.)
1987 =head3 Using t/harness for testing
1989 If you use C<harness> for testing you have several command line options
1990 available to you. The arguments are as follows, and are in the order
1991 that they must appear if used together.
1993 harness -v -torture -re=pattern LIST OF FILES TO TEST
1994 harness -v -torture -re LIST OF PATTERNS TO MATCH
1996 If C<LIST OF FILES TO TEST> is omitted the file list is obtained from
1997 the manifest. The file list may include shell wildcards which will be
2004 Run the tests under verbose mode so you can see what tests were run,
2009 Run the torture tests as well as the normal set.
2013 Filter the file list so that all the test files run match PATTERN.
2014 Note that this form is distinct from the B<-re LIST OF PATTERNS> form below
2015 in that it allows the file list to be provided as well.
2017 =item -re LIST OF PATTERNS
2019 Filter the file list so that all the test files run match
2020 /(LIST|OF|PATTERNS)/. Note that with this form the patterns
2021 are joined by '|' and you cannot supply a list of files, instead
2022 the test files are obtained from the MANIFEST.
2026 You can run an individual test by a command similar to
2028 ./perl -I../lib patho/to/foo.t
2030 except that the harnesses set up some environment variables that may
2031 affect the execution of the test :
2037 indicates that we're running this test part of the perl core test suite.
2038 This is useful for modules that have a dual life on CPAN.
2040 =item PERL_DESTRUCT_LEVEL=2
2042 is set to 2 if it isn't set already (see L</PERL_DESTRUCT_LEVEL>)
2046 (used only by F<t/TEST>) if set, overrides the path to the perl executable
2047 that should be used to run the tests (the default being F<./perl>).
2049 =item PERL_SKIP_TTY_TEST
2051 if set, tells to skip the tests that need a terminal. It's actually set
2052 automatically by the Makefile, but can also be forced artificially by
2053 running 'make test_notty'.
2057 =head3 Other environment variables that may influence tests
2061 =item PERL_TEST_Net_Ping
2063 Setting this variable runs all the Net::Ping modules tests,
2064 otherwise some tests that interact with the outside world are skipped.
2067 =item PERL_TEST_NOVREXX
2069 Setting this variable skips the vrexx.t tests for OS2::REXX.
2071 =item PERL_TEST_NUMCONVERTS
2073 This sets a variable in op/numconvert.t.
2077 See also the documentation for the Test and Test::Harness modules,
2078 for more environment variables that affect testing.
2080 =head2 Common problems when patching Perl source code
2082 Perl source plays by ANSI C89 rules: no C99 (or C++) extensions. In
2083 some cases we have to take pre-ANSI requirements into consideration.
2084 You don't care about some particular platform having broken Perl?
2085 I hear there is still a strong demand for J2EE programmers.
2087 =head2 Perl environment problems
2093 Not compiling with threading
2095 Compiling with threading (-Duseithreads) completely rewrites
2096 the function prototypes of Perl. You better try your changes
2097 with that. Related to this is the difference between "Perl_-less"
2098 and "Perl_-ly" APIs, for example:
2100 Perl_sv_setiv(aTHX_ ...);
2103 The first one explicitly passes in the context, which is needed for e.g.
2104 threaded builds. The second one does that implicitly; do not get them
2105 mixed. If you are not passing in a aTHX_, you will need to do a dTHX
2106 (or a dVAR) as the first thing in the function.
2108 See L<perlguts/"How multiple interpreters and concurrency are supported">
2109 for further discussion about context.
2113 Not compiling with -DDEBUGGING
2115 The DEBUGGING define exposes more code to the compiler,
2116 therefore more ways for things to go wrong. You should try it.
2120 Introducing (non-read-only) globals
2122 Do not introduce any modifiable globals, truly global or file static.
2123 They are bad form and complicate multithreading and other forms of
2124 concurrency. The right way is to introduce them as new interpreter
2125 variables, see F<intrpvar.h> (at the very end for binary compatibility).
2127 Introducing read-only (const) globals is okay, as long as you verify
2128 with e.g. C<nm libperl.a|egrep -v ' [TURtr] '> (if your C<nm> has
2129 BSD-style output) that the data you added really is read-only.
2130 (If it is, it shouldn't show up in the output of that command.)
2132 If you want to have static strings, make them constant:
2134 static const char etc[] = "...";
2136 If you want to have arrays of constant strings, note carefully
2137 the right combination of C<const>s:
2139 static const char * const yippee[] =
2140 {"hi", "ho", "silver"};
2142 There is a way to completely hide any modifiable globals (they are all
2143 moved to heap), the compilation setting C<-DPERL_GLOBAL_STRUCT_PRIVATE>.
2144 It is not normally used, but can be used for testing, read more
2145 about it in L<perlguts/"Background and PERL_IMPLICIT_CONTEXT">.
2149 Not exporting your new function
2151 Some platforms (Win32, AIX, VMS, OS/2, to name a few) require any
2152 function that is part of the public API (the shared Perl library)
2153 to be explicitly marked as exported. See the discussion about
2154 F<embed.pl> in L<perlguts>.
2158 Exporting your new function
2160 The new shiny result of either genuine new functionality or your
2161 arduous refactoring is now ready and correctly exported. So what
2162 could possibly go wrong?
2164 Maybe simply that your function did not need to be exported in the
2165 first place. Perl has a long and not so glorious history of exporting
2166 functions that it should not have.
2168 If the function is used only inside one source code file, make it
2169 static. See the discussion about F<embed.pl> in L<perlguts>.
2171 If the function is used across several files, but intended only for
2172 Perl's internal use (and this should be the common case), do not
2173 export it to the public API. See the discussion about F<embed.pl>
2178 =head2 Portability problems
2180 The following are common causes of compilation and/or execution
2181 failures, not common to Perl as such. The C FAQ is good bedtime
2182 reading. Please test your changes with as many C compilers and
2183 platforms as possible -- we will, anyway, and it's nice to save
2184 oneself from public embarrassment.
2186 If using gcc, you can add the C<-std=c89> option which will hopefully
2187 catch most of these unportabilities. (However it might also catch
2188 incompatibilities in your system's header files.)
2190 Use the Configure C<-Dgccansipedantic> flag to enable the gcc
2191 C<-ansi -pedantic> flags which enforce stricter ANSI rules.
2193 If using the C<gcc -Wall> note that not all the possible warnings
2194 (like C<-Wunitialized>) are given unless you also compile with C<-O>.
2196 Note that if using gcc, starting from Perl 5.9.5 the Perl core source
2197 code files (the ones at the top level of the source code distribution,
2198 but not e.g. the extensions under ext/) are automatically compiled
2199 with as many as possible of the C<-std=c89>, C<-ansi>, C<-pedantic>,
2200 and a selection of C<-W> flags (see cflags.SH).
2202 Also study L<perlport> carefully to avoid any bad assumptions
2203 about the operating system, filesystems, and so forth.
2205 You may once in a while try a "make microperl" to see whether we
2206 can still compile Perl with just the bare minimum of interfaces.
2209 Do not assume an operating system indicates a certain compiler.
2215 Casting pointers to integers or casting integers to pointers
2217 void castaway(U8* p)
2223 void castaway(U8* p)
2227 Both are bad, and broken, and unportable. Use the PTR2IV()
2228 macro that does it right. (Likewise, there are PTR2UV(), PTR2NV(),
2229 INT2PTR(), and NUM2PTR().)
2233 Casting between data function pointers and data pointers
2235 Technically speaking casting between function pointers and data
2236 pointers is unportable and undefined, but practically speaking
2237 it seems to work, but you should use the FPTR2DPTR() and DPTR2FPTR()
2238 macros. Sometimes you can also play games with unions.
2242 Assuming sizeof(int) == sizeof(long)
2244 There are platforms where longs are 64 bits, and platforms where ints
2245 are 64 bits, and while we are out to shock you, even platforms where
2246 shorts are 64 bits. This is all legal according to the C standard.
2247 (In other words, "long long" is not a portable way to specify 64 bits,
2248 and "long long" is not even guaranteed to be any wider than "long".)
2250 Instead, use the definitions IV, UV, IVSIZE, I32SIZE, and so forth.
2251 Avoid things like I32 because they are B<not> guaranteed to be
2252 I<exactly> 32 bits, they are I<at least> 32 bits, nor are they
2253 guaranteed to be B<int> or B<long>. If you really explicitly need
2254 64-bit variables, use I64 and U64, but only if guarded by HAS_QUAD.
2258 Assuming one can dereference any type of pointer for any type of data
2261 long pony = *p; /* BAD */
2263 Many platforms, quite rightly so, will give you a core dump instead
2264 of a pony if the p happens not be correctly aligned.
2270 (int)*p = ...; /* BAD */
2272 Simply not portable. Get your lvalue to be of the right type,
2273 or maybe use temporary variables, or dirty tricks with unions.
2277 Assume B<anything> about structs (especially the ones you
2278 don't control, like the ones coming from the system headers)
2284 That a certain field exists in a struct
2288 That no other fields exist besides the ones you know of
2292 That a field is of certain signedness, sizeof, or type
2296 That the fields are in a certain order
2302 While C guarantees the ordering specified in the struct definition,
2303 between different platforms the definitions might differ
2309 That the sizeof(struct) or the alignments are the same everywhere
2315 There might be padding bytes between the fields to align the fields -
2316 the bytes can be anything
2320 Structs are required to be aligned to the maximum alignment required
2321 by the fields - which for native types is for usually equivalent to
2322 sizeof() of the field
2330 Assuming the character set is ASCIIish
2332 Perl can compile and run under EBCDIC platforms. See L<perlebcdic>.
2333 This is transparent for the most part, but because the character sets
2334 differ, you shouldn't use numeric (decimal, octal, nor hex) constants
2335 to refer to characters. You can safely say 'A', but not 0x41.
2336 You can safely say '\n', but not \012.
2337 If a character doesn't have a trivial input form, you can
2338 create a #define for it in both C<utfebcdic.h> and C<utf8.h>, so that
2339 it resolves to different values depending on the character set being used.
2340 (There are three different EBCDIC character sets defined in C<utfebcdic.h>,
2341 so it might be best to insert the #define three times in that file.)
2343 Also, the range 'A' - 'Z' in ASCII is an unbroken sequence of 26 upper case
2344 alphabetic characters. That is not true in EBCDIC. Nor for 'a' to 'z'.
2345 But '0' - '9' is an unbroken range in both systems. Don't assume anything
2348 Many of the comments in the existing code ignore the possibility of EBCDIC,
2349 and may be wrong therefore, even if the code works.
2350 This is actually a tribute to the successful transparent insertion of being
2351 able to handle EBCDIC without having to change pre-existing code.
2353 UTF-8 and UTF-EBCDIC are two different encodings used to represent Unicode
2354 code points as sequences of bytes. Macros
2355 with the same names (but different definitions)
2356 in C<utf8.h> and C<utfebcdic.h>
2357 are used to allow the calling code to think that there is only one such
2359 This is almost always referred to as C<utf8>, but it means the EBCDIC version
2360 as well. Again, comments in the code may well be wrong even if the code itself
2362 For example, the concept of C<invariant characters> differs between ASCII and
2364 On ASCII platforms, only characters that do not have the high-order
2365 bit set (i.e. whose ordinals are strict ASCII, 0 - 127)
2366 are invariant, and the documentation and comments in the code
2368 often referring to something like, say, C<hibit>.
2369 The situation differs and is not so simple on EBCDIC machines, but as long as
2370 the code itself uses the C<NATIVE_IS_INVARIANT()> macro appropriately, it
2371 works, even if the comments are wrong.
2375 Assuming the character set is just ASCII
2377 ASCII is a 7 bit encoding, but bytes have 8 bits in them. The 128 extra
2378 characters have different meanings depending on the locale. Absent a locale,
2379 currently these extra characters are generally considered to be unassigned,
2380 and this has presented some problems.
2381 This is scheduled to be changed in 5.12 so that these characters will
2382 be considered to be Latin-1 (ISO-8859-1).
2386 Mixing #define and #ifdef
2388 #define BURGLE(x) ... \
2389 #ifdef BURGLE_OLD_STYLE /* BAD */
2390 ... do it the old way ... \
2392 ... do it the new way ... \
2395 You cannot portably "stack" cpp directives. For example in the above
2396 you need two separate BURGLE() #defines, one for each #ifdef branch.
2400 Adding non-comment stuff after #endif or #else
2404 #else !SNOSH /* BAD */
2406 #endif SNOSH /* BAD */
2408 The #endif and #else cannot portably have anything non-comment after
2409 them. If you want to document what is going (which is a good idea
2410 especially if the branches are long), use (C) comments:
2418 The gcc option C<-Wendif-labels> warns about the bad variant
2419 (by default on starting from Perl 5.9.4).
2423 Having a comma after the last element of an enum list
2431 is not portable. Leave out the last comma.
2433 Also note that whether enums are implicitly morphable to ints
2434 varies between compilers, you might need to (int).
2440 // This function bamfoodles the zorklator. /* BAD */
2442 That is C99 or C++. Perl is C89. Using the //-comments is silently
2443 allowed by many C compilers but cranking up the ANSI C89 strictness
2444 (which we like to do) causes the compilation to fail.
2448 Mixing declarations and code
2453 set_zorkmids(n); /* BAD */
2456 That is C99 or C++. Some C compilers allow that, but you shouldn't.
2458 The gcc option C<-Wdeclaration-after-statements> scans for such problems
2459 (by default on starting from Perl 5.9.4).
2463 Introducing variables inside for()
2465 for(int i = ...; ...; ...) { /* BAD */
2467 That is C99 or C++. While it would indeed be awfully nice to have that
2468 also in C89, to limit the scope of the loop variable, alas, we cannot.
2472 Mixing signed char pointers with unsigned char pointers
2474 int foo(char *s) { ... }
2476 unsigned char *t = ...; /* Or U8* t = ... */
2479 While this is legal practice, it is certainly dubious, and downright
2480 fatal in at least one platform: for example VMS cc considers this a
2481 fatal error. One cause for people often making this mistake is that a
2482 "naked char" and therefore dereferencing a "naked char pointer" have
2483 an undefined signedness: it depends on the compiler and the flags of
2484 the compiler and the underlying platform whether the result is signed
2485 or unsigned. For this very same reason using a 'char' as an array
2490 Macros that have string constants and their arguments as substrings of
2491 the string constants
2493 #define FOO(n) printf("number = %d\n", n) /* BAD */
2496 Pre-ANSI semantics for that was equivalent to
2498 printf("10umber = %d\10");
2500 which is probably not what you were expecting. Unfortunately at least
2501 one reasonably common and modern C compiler does "real backward
2502 compatibility" here, in AIX that is what still happens even though the
2503 rest of the AIX compiler is very happily C89.
2507 Using printf formats for non-basic C types
2510 printf("i = %d\n", i); /* BAD */
2512 While this might by accident work in some platform (where IV happens
2513 to be an C<int>), in general it cannot. IV might be something larger.
2514 Even worse the situation is with more specific types (defined by Perl's
2515 configuration step in F<config.h>):
2518 printf("who = %d\n", who); /* BAD */
2520 The problem here is that Uid_t might be not only not C<int>-wide
2521 but it might also be unsigned, in which case large uids would be
2522 printed as negative values.
2524 There is no simple solution to this because of printf()'s limited
2525 intelligence, but for many types the right format is available as
2526 with either 'f' or '_f' suffix, for example:
2528 IVdf /* IV in decimal */
2529 UVxf /* UV is hexadecimal */
2531 printf("i = %"IVdf"\n", i); /* The IVdf is a string constant. */
2533 Uid_t_f /* Uid_t in decimal */
2535 printf("who = %"Uid_t_f"\n", who);
2537 Or you can try casting to a "wide enough" type:
2539 printf("i = %"IVdf"\n", (IV)something_very_small_and_signed);
2541 Also remember that the C<%p> format really does require a void pointer:
2544 printf("p = %p\n", (void*)p);
2546 The gcc option C<-Wformat> scans for such problems.
2550 Blindly using variadic macros
2552 gcc has had them for a while with its own syntax, and C99 brought
2553 them with a standardized syntax. Don't use the former, and use
2554 the latter only if the HAS_C99_VARIADIC_MACROS is defined.
2558 Blindly passing va_list
2560 Not all platforms support passing va_list to further varargs (stdarg)
2561 functions. The right thing to do is to copy the va_list using the
2562 Perl_va_copy() if the NEED_VA_COPY is defined.
2566 Using gcc statement expressions
2568 val = ({...;...;...}); /* BAD */
2570 While a nice extension, it's not portable. The Perl code does
2571 admittedly use them if available to gain some extra speed
2572 (essentially as a funky form of inlining), but you shouldn't.
2576 Binding together several statements in a macro
2578 Use the macros STMT_START and STMT_END.
2586 Testing for operating systems or versions when should be testing for features
2588 #ifdef __FOONIX__ /* BAD */
2592 Unless you know with 100% certainty that quux() is only ever available
2593 for the "Foonix" operating system B<and> that is available B<and>
2594 correctly working for B<all> past, present, B<and> future versions of
2595 "Foonix", the above is very wrong. This is more correct (though still
2596 not perfect, because the below is a compile-time check):
2602 How does the HAS_QUUX become defined where it needs to be? Well, if
2603 Foonix happens to be UNIXy enough to be able to run the Configure
2604 script, and Configure has been taught about detecting and testing
2605 quux(), the HAS_QUUX will be correctly defined. In other platforms,
2606 the corresponding configuration step will hopefully do the same.
2608 In a pinch, if you cannot wait for Configure to be educated,
2609 or if you have a good hunch of where quux() might be available,
2610 you can temporarily try the following:
2612 #if (defined(__FOONIX__) || defined(__BARNIX__))
2622 But in any case, try to keep the features and operating systems separate.
2626 =head2 Problematic System Interfaces
2632 malloc(0), realloc(0), calloc(0, 0) are non-portable. To be portable
2633 allocate at least one byte. (In general you should rarely need to
2634 work at this low level, but instead use the various malloc wrappers.)
2638 snprintf() - the return type is unportable. Use my_snprintf() instead.
2642 =head2 Security problems
2644 Last but not least, here are various tips for safer coding.
2652 Or we will publicly ridicule you. Seriously.
2656 Do not use strcpy() or strcat() or strncpy() or strncat()
2658 Use my_strlcpy() and my_strlcat() instead: they either use the native
2659 implementation, or Perl's own implementation (borrowed from the public
2660 domain implementation of INN).
2664 Do not use sprintf() or vsprintf()
2666 If you really want just plain byte strings, use my_snprintf()
2667 and my_vsnprintf() instead, which will try to use snprintf() and
2668 vsnprintf() if those safer APIs are available. If you want something
2669 fancier than a plain byte string, use SVs and Perl_sv_catpvf().
2673 =head1 EXTERNAL TOOLS FOR DEBUGGING PERL
2675 Sometimes it helps to use external tools while debugging and
2676 testing Perl. This section tries to guide you through using
2677 some common testing and debugging tools with Perl. This is
2678 meant as a guide to interfacing these tools with Perl, not
2679 as any kind of guide to the use of the tools themselves.
2681 B<NOTE 1>: Running under memory debuggers such as Purify, valgrind, or
2682 Third Degree greatly slows down the execution: seconds become minutes,
2683 minutes become hours. For example as of Perl 5.8.1, the
2684 ext/Encode/t/Unicode.t takes extraordinarily long to complete under
2685 e.g. Purify, Third Degree, and valgrind. Under valgrind it takes more
2686 than six hours, even on a snappy computer-- the said test must be
2687 doing something that is quite unfriendly for memory debuggers. If you
2688 don't feel like waiting, that you can simply kill away the perl
2691 B<NOTE 2>: To minimize the number of memory leak false alarms (see
2692 L</PERL_DESTRUCT_LEVEL> for more information), you have to have
2693 environment variable PERL_DESTRUCT_LEVEL set to 2. The F<TEST>
2694 and harness scripts do that automatically. But if you are running
2695 some of the tests manually-- for csh-like shells:
2697 setenv PERL_DESTRUCT_LEVEL 2
2699 and for Bourne-type shells:
2701 PERL_DESTRUCT_LEVEL=2
2702 export PERL_DESTRUCT_LEVEL
2704 or in UNIXy environments you can also use the C<env> command:
2706 env PERL_DESTRUCT_LEVEL=2 valgrind ./perl -Ilib ...
2708 B<NOTE 3>: There are known memory leaks when there are compile-time
2709 errors within eval or require, seeing C<S_doeval> in the call stack
2710 is a good sign of these. Fixing these leaks is non-trivial,
2711 unfortunately, but they must be fixed eventually.
2713 B<NOTE 4>: L<DynaLoader> will not clean up after itself completely
2714 unless Perl is built with the Configure option
2715 C<-Accflags=-DDL_UNLOAD_ALL_AT_EXIT>.
2717 =head2 Rational Software's Purify
2719 Purify is a commercial tool that is helpful in identifying
2720 memory overruns, wild pointers, memory leaks and other such
2721 badness. Perl must be compiled in a specific way for
2722 optimal testing with Purify. Purify is available under
2723 Windows NT, Solaris, HP-UX, SGI, and Siemens Unix.
2725 =head2 Purify on Unix
2727 On Unix, Purify creates a new Perl binary. To get the most
2728 benefit out of Purify, you should create the perl to Purify
2731 sh Configure -Accflags=-DPURIFY -Doptimize='-g' \
2732 -Uusemymalloc -Dusemultiplicity
2734 where these arguments mean:
2738 =item -Accflags=-DPURIFY
2740 Disables Perl's arena memory allocation functions, as well as
2741 forcing use of memory allocation functions derived from the
2744 =item -Doptimize='-g'
2746 Adds debugging information so that you see the exact source
2747 statements where the problem occurs. Without this flag, all
2748 you will see is the source filename of where the error occurred.
2752 Disable Perl's malloc so that Purify can more closely monitor
2753 allocations and leaks. Using Perl's malloc will make Purify
2754 report most leaks in the "potential" leaks category.
2756 =item -Dusemultiplicity
2758 Enabling the multiplicity option allows perl to clean up
2759 thoroughly when the interpreter shuts down, which reduces the
2760 number of bogus leak reports from Purify.
2764 Once you've compiled a perl suitable for Purify'ing, then you
2769 which creates a binary named 'pureperl' that has been Purify'ed.
2770 This binary is used in place of the standard 'perl' binary
2771 when you want to debug Perl memory problems.
2773 As an example, to show any memory leaks produced during the
2774 standard Perl testset you would create and run the Purify'ed
2779 ../pureperl -I../lib harness
2781 which would run Perl on test.pl and report any memory problems.
2783 Purify outputs messages in "Viewer" windows by default. If
2784 you don't have a windowing environment or if you simply
2785 want the Purify output to unobtrusively go to a log file
2786 instead of to the interactive window, use these following
2787 options to output to the log file "perl.log":
2789 setenv PURIFYOPTIONS "-chain-length=25 -windows=no \
2790 -log-file=perl.log -append-logfile=yes"
2792 If you plan to use the "Viewer" windows, then you only need this option:
2794 setenv PURIFYOPTIONS "-chain-length=25"
2796 In Bourne-type shells:
2799 export PURIFYOPTIONS
2801 or if you have the "env" utility:
2803 env PURIFYOPTIONS="..." ../pureperl ...
2807 Purify on Windows NT instruments the Perl binary 'perl.exe'
2808 on the fly. There are several options in the makefile you
2809 should change to get the most use out of Purify:
2815 You should add -DPURIFY to the DEFINES line so the DEFINES
2816 line looks something like:
2818 DEFINES = -DWIN32 -D_CONSOLE -DNO_STRICT $(CRYPT_FLAG) -DPURIFY=1
2820 to disable Perl's arena memory allocation functions, as
2821 well as to force use of memory allocation functions derived
2822 from the system malloc.
2824 =item USE_MULTI = define
2826 Enabling the multiplicity option allows perl to clean up
2827 thoroughly when the interpreter shuts down, which reduces the
2828 number of bogus leak reports from Purify.
2830 =item #PERL_MALLOC = define
2832 Disable Perl's malloc so that Purify can more closely monitor
2833 allocations and leaks. Using Perl's malloc will make Purify
2834 report most leaks in the "potential" leaks category.
2838 Adds debugging information so that you see the exact source
2839 statements where the problem occurs. Without this flag, all
2840 you will see is the source filename of where the error occurred.
2844 As an example, to show any memory leaks produced during the
2845 standard Perl testset you would create and run Purify as:
2850 purify ../perl -I../lib harness
2852 which would instrument Perl in memory, run Perl on test.pl,
2853 then finally report any memory problems.
2857 The excellent valgrind tool can be used to find out both memory leaks
2858 and illegal memory accesses. As of version 3.3.0, Valgrind only
2859 supports Linux on x86, x86-64 and PowerPC. The special "test.valgrind"
2860 target can be used to run the tests under valgrind. Found errors
2861 and memory leaks are logged in files named F<testfile.valgrind>.
2863 Valgrind also provides a cachegrind tool, invoked on perl as:
2865 VG_OPTS=--tool=cachegrind make test.valgrind
2867 As system libraries (most notably glibc) are also triggering errors,
2868 valgrind allows to suppress such errors using suppression files. The
2869 default suppression file that comes with valgrind already catches a lot
2870 of them. Some additional suppressions are defined in F<t/perl.supp>.
2872 To get valgrind and for more information see
2874 http://developer.kde.org/~sewardj/
2876 =head2 Compaq's/Digital's/HP's Third Degree
2878 Third Degree is a tool for memory leak detection and memory access checks.
2879 It is one of the many tools in the ATOM toolkit. The toolkit is only
2880 available on Tru64 (formerly known as Digital UNIX formerly known as
2883 When building Perl, you must first run Configure with -Doptimize=-g
2884 and -Uusemymalloc flags, after that you can use the make targets
2885 "perl.third" and "test.third". (What is required is that Perl must be
2886 compiled using the C<-g> flag, you may need to re-Configure.)
2888 The short story is that with "atom" you can instrument the Perl
2889 executable to create a new executable called F<perl.third>. When the
2890 instrumented executable is run, it creates a log of dubious memory
2891 traffic in file called F<perl.3log>. See the manual pages of atom and
2892 third for more information. The most extensive Third Degree
2893 documentation is available in the Compaq "Tru64 UNIX Programmer's
2894 Guide", chapter "Debugging Programs with Third Degree".
2896 The "test.third" leaves a lot of files named F<foo_bar.3log> in the t/
2897 subdirectory. There is a problem with these files: Third Degree is so
2898 effective that it finds problems also in the system libraries.
2899 Therefore you should used the Porting/thirdclean script to cleanup
2900 the F<*.3log> files.
2902 There are also leaks that for given certain definition of a leak,
2903 aren't. See L</PERL_DESTRUCT_LEVEL> for more information.
2905 =head2 PERL_DESTRUCT_LEVEL
2907 If you want to run any of the tests yourself manually using e.g.
2908 valgrind, or the pureperl or perl.third executables, please note that
2909 by default perl B<does not> explicitly cleanup all the memory it has
2910 allocated (such as global memory arenas) but instead lets the exit()
2911 of the whole program "take care" of such allocations, also known as
2912 "global destruction of objects".
2914 There is a way to tell perl to do complete cleanup: set the
2915 environment variable PERL_DESTRUCT_LEVEL to a non-zero value.
2916 The t/TEST wrapper does set this to 2, and this is what you
2917 need to do too, if you don't want to see the "global leaks":
2918 For example, for "third-degreed" Perl:
2920 env PERL_DESTRUCT_LEVEL=2 ./perl.third -Ilib t/foo/bar.t
2922 (Note: the mod_perl apache module uses also this environment variable
2923 for its own purposes and extended its semantics. Refer to the mod_perl
2924 documentation for more information. Also, spawned threads do the
2925 equivalent of setting this variable to the value 1.)
2927 If, at the end of a run you get the message I<N scalars leaked>, you can
2928 recompile with C<-DDEBUG_LEAKING_SCALARS>, which will cause the addresses
2929 of all those leaked SVs to be dumped along with details as to where each
2930 SV was originally allocated. This information is also displayed by
2931 Devel::Peek. Note that the extra details recorded with each SV increases
2932 memory usage, so it shouldn't be used in production environments. It also
2933 converts C<new_SV()> from a macro into a real function, so you can use
2934 your favourite debugger to discover where those pesky SVs were allocated.
2936 If you see that you're leaking memory at runtime, but neither valgrind
2937 nor C<-DDEBUG_LEAKING_SCALARS> will find anything, you're probably
2938 leaking SVs that are still reachable and will be properly cleaned up
2939 during destruction of the interpreter. In such cases, using the C<-Dm>
2940 switch can point you to the source of the leak. If the executable was
2941 built with C<-DDEBUG_LEAKING_SCALARS>, C<-Dm> will output SV allocations
2942 in addition to memory allocations. Each SV allocation has a distinct
2943 serial number that will be written on creation and destruction of the SV.
2944 So if you're executing the leaking code in a loop, you need to look for
2945 SVs that are created, but never destroyed between each cycle. If such an
2946 SV is found, set a conditional breakpoint within C<new_SV()> and make it
2947 break only when C<PL_sv_serial> is equal to the serial number of the
2948 leaking SV. Then you will catch the interpreter in exactly the state
2949 where the leaking SV is allocated, which is sufficient in many cases to
2950 find the source of the leak.
2952 As C<-Dm> is using the PerlIO layer for output, it will by itself
2953 allocate quite a bunch of SVs, which are hidden to avoid recursion.
2954 You can bypass the PerlIO layer if you use the SV logging provided
2955 by C<-DPERL_MEM_LOG> instead.
2959 If compiled with C<-DPERL_MEM_LOG>, both memory and SV allocations go
2960 through logging functions, which is handy for breakpoint setting.
2962 Unless C<-DPERL_MEM_LOG_NOIMPL> is also compiled, the logging
2963 functions read $ENV{PERL_MEM_LOG} to determine whether to log the
2964 event, and if so how:
2966 $ENV{PERL_MEM_LOG} =~ /m/ Log all memory ops
2967 $ENV{PERL_MEM_LOG} =~ /s/ Log all SV ops
2968 $ENV{PERL_MEM_LOG} =~ /t/ include timestamp in Log
2969 $ENV{PERL_MEM_LOG} =~ /^(\d+)/ write to FD given (default is 2)
2971 Memory logging is somewhat similar to C<-Dm> but is independent of
2972 C<-DDEBUGGING>, and at a higher level; all uses of Newx(), Renew(),
2973 and Safefree() are logged with the caller's source code file and line
2974 number (and C function name, if supported by the C compiler). In
2975 contrast, C<-Dm> is directly at the point of C<malloc()>. SV logging
2978 Since the logging doesn't use PerlIO, all SV allocations are logged
2979 and no extra SV allocations are introduced by enabling the logging.
2980 If compiled with C<-DDEBUG_LEAKING_SCALARS>, the serial number for
2981 each SV allocation is also logged.
2985 Depending on your platform there are various of profiling Perl.
2987 There are two commonly used techniques of profiling executables:
2988 I<statistical time-sampling> and I<basic-block counting>.
2990 The first method takes periodically samples of the CPU program
2991 counter, and since the program counter can be correlated with the code
2992 generated for functions, we get a statistical view of in which
2993 functions the program is spending its time. The caveats are that very
2994 small/fast functions have lower probability of showing up in the
2995 profile, and that periodically interrupting the program (this is
2996 usually done rather frequently, in the scale of milliseconds) imposes
2997 an additional overhead that may skew the results. The first problem
2998 can be alleviated by running the code for longer (in general this is a
2999 good idea for profiling), the second problem is usually kept in guard
3000 by the profiling tools themselves.
3002 The second method divides up the generated code into I<basic blocks>.
3003 Basic blocks are sections of code that are entered only in the
3004 beginning and exited only at the end. For example, a conditional jump
3005 starts a basic block. Basic block profiling usually works by
3006 I<instrumenting> the code by adding I<enter basic block #nnnn>
3007 book-keeping code to the generated code. During the execution of the
3008 code the basic block counters are then updated appropriately. The
3009 caveat is that the added extra code can skew the results: again, the
3010 profiling tools usually try to factor their own effects out of the
3013 =head2 Gprof Profiling
3015 gprof is a profiling tool available in many UNIX platforms,
3016 it uses F<statistical time-sampling>.
3018 You can build a profiled version of perl called "perl.gprof" by
3019 invoking the make target "perl.gprof" (What is required is that Perl
3020 must be compiled using the C<-pg> flag, you may need to re-Configure).
3021 Running the profiled version of Perl will create an output file called
3022 F<gmon.out> is created which contains the profiling data collected
3023 during the execution.
3025 The gprof tool can then display the collected data in various ways.
3026 Usually gprof understands the following options:
3032 Suppress statically defined functions from the profile.
3036 Suppress the verbose descriptions in the profile.
3040 Exclude the given routine and its descendants from the profile.
3044 Display only the given routine and its descendants in the profile.
3048 Generate a summary file called F<gmon.sum> which then may be given
3049 to subsequent gprof runs to accumulate data over several runs.
3053 Display routines that have zero usage.
3057 For more detailed explanation of the available commands and output
3058 formats, see your own local documentation of gprof.
3062 $ sh Configure -des -Dusedevel -Doptimize='-pg' && make perl.gprof
3063 $ ./perl.gprof someprog # creates gmon.out in current directory
3064 $ gprof ./perl.gprof > out
3067 =head2 GCC gcov Profiling
3069 Starting from GCC 3.0 I<basic block profiling> is officially available
3072 You can build a profiled version of perl called F<perl.gcov> by
3073 invoking the make target "perl.gcov" (what is required that Perl must
3074 be compiled using gcc with the flags C<-fprofile-arcs
3075 -ftest-coverage>, you may need to re-Configure).
3077 Running the profiled version of Perl will cause profile output to be
3078 generated. For each source file an accompanying ".da" file will be
3081 To display the results you use the "gcov" utility (which should
3082 be installed if you have gcc 3.0 or newer installed). F<gcov> is
3083 run on source code files, like this
3087 which will cause F<sv.c.gcov> to be created. The F<.gcov> files
3088 contain the source code annotated with relative frequencies of
3089 execution indicated by "#" markers.
3091 Useful options of F<gcov> include C<-b> which will summarise the
3092 basic block, branch, and function call coverage, and C<-c> which
3093 instead of relative frequencies will use the actual counts. For
3094 more information on the use of F<gcov> and basic block profiling
3095 with gcc, see the latest GNU CC manual, as of GCC 3.0 see
3097 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc.html
3099 and its section titled "8. gcov: a Test Coverage Program"
3101 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc_8.html#SEC132
3105 $ sh Configure -des -Doptimize='-g' -Accflags='-fprofile-arcs -ftest-coverage' \
3106 -Aldflags='-fprofile-arcs -ftest-coverage' && make perl.gcov
3107 $ rm -f regexec.c.gcov regexec.gcda
3110 $ view regexec.c.gcov
3112 =head2 Pixie Profiling
3114 Pixie is a profiling tool available on IRIX and Tru64 (aka Digital
3115 UNIX aka DEC OSF/1) platforms. Pixie does its profiling using
3116 I<basic-block counting>.
3118 You can build a profiled version of perl called F<perl.pixie> by
3119 invoking the make target "perl.pixie" (what is required is that Perl
3120 must be compiled using the C<-g> flag, you may need to re-Configure).
3122 In Tru64 a file called F<perl.Addrs> will also be silently created,
3123 this file contains the addresses of the basic blocks. Running the
3124 profiled version of Perl will create a new file called "perl.Counts"
3125 which contains the counts for the basic block for that particular
3128 To display the results you use the F<prof> utility. The exact
3129 incantation depends on your operating system, "prof perl.Counts" in
3130 IRIX, and "prof -pixie -all -L. perl" in Tru64.
3132 In IRIX the following prof options are available:
3138 Reports the most heavily used lines in descending order of use.
3139 Useful for finding the hotspot lines.
3143 Groups lines by procedure, with procedures sorted in descending order of use.
3144 Within a procedure, lines are listed in source order.
3145 Useful for finding the hotspots of procedures.
3149 In Tru64 the following options are available:
3155 Procedures sorted in descending order by the number of cycles executed
3156 in each procedure. Useful for finding the hotspot procedures.
3157 (This is the default option.)
3161 Lines sorted in descending order by the number of cycles executed in
3162 each line. Useful for finding the hotspot lines.
3164 =item -i[nvocations]
3166 The called procedures are sorted in descending order by number of calls
3167 made to the procedures. Useful for finding the most used procedures.
3171 Grouped by procedure, sorted by cycles executed per procedure.
3172 Useful for finding the hotspots of procedures.
3176 The compiler emitted code for these lines, but the code was unexecuted.
3180 Unexecuted procedures.
3184 For further information, see your system's manual pages for pixie and prof.
3186 =head2 Miscellaneous tricks
3192 Those debugging perl with the DDD frontend over gdb may find the
3195 You can extend the data conversion shortcuts menu, so for example you
3196 can display an SV's IV value with one click, without doing any typing.
3197 To do that simply edit ~/.ddd/init file and add after:
3199 ! Display shortcuts.
3200 Ddd*gdbDisplayShortcuts: \
3201 /t () // Convert to Bin\n\
3202 /d () // Convert to Dec\n\
3203 /x () // Convert to Hex\n\
3204 /o () // Convert to Oct(\n\
3206 the following two lines:
3208 ((XPV*) (())->sv_any )->xpv_pv // 2pvx\n\
3209 ((XPVIV*) (())->sv_any )->xiv_iv // 2ivx
3211 so now you can do ivx and pvx lookups or you can plug there the
3212 sv_peek "conversion":
3214 Perl_sv_peek(my_perl, (SV*)()) // sv_peek
3216 (The my_perl is for threaded builds.)
3217 Just remember that every line, but the last one, should end with \n\
3219 Alternatively edit the init file interactively via:
3220 3rd mouse button -> New Display -> Edit Menu
3222 Note: you can define up to 20 conversion shortcuts in the gdb
3227 If you see in a debugger a memory area mysteriously full of 0xABABABAB
3228 or 0xEFEFEFEF, you may be seeing the effect of the Poison() macros,
3233 Under ithreads the optree is read only. If you want to enforce this, to check
3234 for write accesses from buggy code, compile with C<-DPL_OP_SLAB_ALLOC> to
3235 enable the OP slab allocator and C<-DPERL_DEBUG_READONLY_OPS> to enable code
3236 that allocates op memory via C<mmap>, and sets it read-only at run time.
3237 Any write access to an op results in a C<SIGBUS> and abort.
3239 This code is intended for development only, and may not be portable even to
3240 all Unix variants. Also, it is an 80% solution, in that it isn't able to make
3241 all ops read only. Specifically it
3247 Only sets read-only on all slabs of ops at C<CHECK> time, hence ops allocated
3248 later via C<require> or C<eval> will be re-write
3252 Turns an entire slab of ops read-write if the refcount of any op in the slab
3253 needs to be decreased.
3257 Turns an entire slab of ops read-write if any op from the slab is freed.
3261 It's not possible to turn the slabs to read-only after an action requiring
3262 read-write access, as either can happen during op tree building time, so
3263 there may still be legitimate write access.
3265 However, as an 80% solution it is still effective, as currently it catches
3266 a write access during the generation of F<Config.pm>, which means that we
3267 can't yet build F<perl> with this enabled.
3274 We've had a brief look around the Perl source, how to maintain quality
3275 of the source code, an overview of the stages F<perl> goes through
3276 when it's running your code, how to use debuggers to poke at the Perl
3277 guts, and finally how to analyse the execution of Perl. We took a very
3278 simple problem and demonstrated how to solve it fully - with
3279 documentation, regression tests, and finally a patch for submission to
3280 p5p. Finally, we talked about how to use external tools to debug and
3283 I'd now suggest you read over those references again, and then, as soon
3284 as possible, get your hands dirty. The best way to learn is by doing,
3291 Subscribe to perl5-porters, follow the patches and try and understand
3292 them; don't be afraid to ask if there's a portion you're not clear on -
3293 who knows, you may unearth a bug in the patch...
3297 Keep up to date with the bleeding edge Perl distributions and get
3298 familiar with the changes. Try and get an idea of what areas people are
3299 working on and the changes they're making.
3303 Do read the README associated with your operating system, e.g. README.aix
3304 on the IBM AIX OS. Don't hesitate to supply patches to that README if
3305 you find anything missing or changed over a new OS release.
3309 Find an area of Perl that seems interesting to you, and see if you can
3310 work out how it works. Scan through the source, and step over it in the
3311 debugger. Play, poke, investigate, fiddle! You'll probably get to
3312 understand not just your chosen area but a much wider range of F<perl>'s
3313 activity as well, and probably sooner than you'd think.
3319 =item I<The Road goes ever on and on, down from the door where it began.>
3323 If you can do these things, you've started on the long road to Perl porting.
3324 Thanks for wanting to help make Perl better - and happy hacking!
3326 =head2 Metaphoric Quotations
3328 If you recognized the quote about the Road above, you're in luck.
3330 Most software projects begin each file with a literal description of each
3331 file's purpose. Perl instead begins each with a literary allusion to that
3334 Like chapters in many books, all top-level Perl source files (along with a
3335 few others here and there) begin with an epigramic inscription that alludes,
3336 indirectly and metaphorically, to the material you're about to read.
3338 Quotations are taken from writings of J.R.R Tolkien pertaining to his
3339 Legendarium, almost always from I<The Lord of the Rings>. Chapters and
3340 page numbers are given using the following editions:
3346 I<The Hobbit>, by J.R.R. Tolkien. The hardcover, 70th-anniversary
3347 edition of 2007 was used, published in the UK by Harper Collins Publishers
3348 and in the US by the Houghton Mifflin Company.
3352 I<The Lord of the Rings>, by J.R.R. Tolkien. The hardcover,
3353 50th-anniversary edition of 2004 was used, published in the UK by Harper
3354 Collins Publishers and in the US by the Houghton Mifflin Company.
3358 I<The Lays of Beleriand>, by J.R.R. Tolkien and published posthumously by his
3359 son and literary executor, C.J.R. Tolkien, being the 3rd of the 12 volumes
3360 in Christopher's mammoth I<History of Middle Earth>. Page numbers derive
3361 from the hardcover edition, first published in 1983 by George Allen &
3362 Unwin; no page numbers changed for the special 3-volume omnibus edition of
3363 2002 or the various trade-paper editions, all again now by Harper Collins
3364 or Houghton Mifflin.
3368 Other JRRT books fair game for quotes would thus include I<The Adventures of
3369 Tom Bombadil>, I<The Silmarillion>, I<Unfinished Tales>, and I<The Tale of
3370 the Children of Hurin>, all but the first posthumously assembled by CJRT.
3371 But I<The Lord of the Rings> itself is perfectly fine and probably best to
3372 quote from, provided you can find a suitable quote there.
3374 So if you were to supply a new, complete, top-level source file to add to
3375 Perl, you should conform to this peculiar practice by yourself selecting an
3376 appropriate quotation from Tolkien, retaining the original spelling and
3377 punctuation and using the same format the rest of the quotes are in.
3378 Indirect and oblique is just fine; remember, it's a metaphor, so being meta
3379 is, after all, what it's for.
3383 This document was written by Nathan Torkington, and is maintained by
3384 the perl5-porters mailing list.