This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Update Changes.
[perl5.git] / pod / perlhack.pod
CommitLineData
e8cd7eae
GS
1=head1 NAME
2
3perlhack - How to hack at the Perl internals
4
5=head1 DESCRIPTION
6
7This document attempts to explain how Perl development takes place,
8and ends with some suggestions for people wanting to become bona fide
9porters.
10
11The perl5-porters mailing list is where the Perl standard distribution
12is maintained and developed. The list can get anywhere from 10 to 150
13messages a day, depending on the heatedness of the debate. Most days
14there are two or three patches, extensions, features, or bugs being
15discussed at a time.
16
17A searchable archive of the list is at:
18
19 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
20
21The list is also archived under the usenet group name
22C<perl.porters-gw> at:
23
24 http://www.deja.com/
25
26List subscribers (the porters themselves) come in several flavours.
27Some are quiet curious lurkers, who rarely pitch in and instead watch
28the ongoing development to ensure they're forewarned of new changes or
29features in Perl. Some are representatives of vendors, who are there
30to make sure that Perl continues to compile and work on their
31platforms. Some patch any reported bug that they know how to fix,
32some are actively patching their pet area (threads, Win32, the regexp
33engine), while others seem to do nothing but complain. In other
34words, it's your usual mix of technical people.
35
36Over this group of porters presides Larry Wall. He has the final word
f6c51b38
GS
37in what does and does not change in the Perl language. Various
38releases of Perl are shepherded by a ``pumpking'', a porter
39responsible for gathering patches, deciding on a patch-by-patch
40feature-by-feature basis what will and will not go into the release.
41For instance, Gurusamy Sarathy is the pumpking for the 5.6 release of
42Perl.
e8cd7eae
GS
43
44In addition, various people are pumpkings for different things. For
45instance, Andy Dougherty and Jarkko Hietaniemi share the I<Configure>
46pumpkin, and Tom Christiansen is the documentation pumpking.
47
48Larry sees Perl development along the lines of the US government:
49there's the Legislature (the porters), the Executive branch (the
50pumpkings), and the Supreme Court (Larry). The legislature can
51discuss and submit patches to the executive branch all they like, but
52the executive branch is free to veto them. Rarely, the Supreme Court
53will side with the executive branch over the legislature, or the
54legislature over the executive branch. Mostly, however, the
55legislature and the executive branch are supposed to get along and
56work out their differences without impeachment or court cases.
57
58You might sometimes see reference to Rule 1 and Rule 2. Larry's power
59as Supreme Court is expressed in The Rules:
60
61=over 4
62
63=item 1
64
65Larry is always by definition right about how Perl should behave.
66This means he has final veto power on the core functionality.
67
68=item 2
69
70Larry is allowed to change his mind about any matter at a later date,
71regardless of whether he previously invoked Rule 1.
72
73=back
74
75Got that? Larry is always right, even when he was wrong. It's rare
76to see either Rule exercised, but they are often alluded to.
77
78New features and extensions to the language are contentious, because
79the criteria used by the pumpkings, Larry, and other porters to decide
80which features should be implemented and incorporated are not codified
81in a few small design goals as with some other languages. Instead,
82the heuristics are flexible and often difficult to fathom. Here is
83one person's list, roughly in decreasing order of importance, of
84heuristics that new features have to be weighed against:
85
86=over 4
87
88=item Does concept match the general goals of Perl?
89
90These haven't been written anywhere in stone, but one approximation
91is:
92
93 1. Keep it fast, simple, and useful.
94 2. Keep features/concepts as orthogonal as possible.
95 3. No arbitrary limits (platforms, data sizes, cultures).
96 4. Keep it open and exciting to use/patch/advocate Perl everywhere.
97 5. Either assimilate new technologies, or build bridges to them.
98
99=item Where is the implementation?
100
101All the talk in the world is useless without an implementation. In
102almost every case, the person or people who argue for a new feature
103will be expected to be the ones who implement it. Porters capable
104of coding new features have their own agendas, and are not available
105to implement your (possibly good) idea.
106
107=item Backwards compatibility
108
109It's a cardinal sin to break existing Perl programs. New warnings are
110contentious--some say that a program that emits warnings is not
111broken, while others say it is. Adding keywords has the potential to
112break programs, changing the meaning of existing token sequences or
113functions might break programs.
114
115=item Could it be a module instead?
116
117Perl 5 has extension mechanisms, modules and XS, specifically to avoid
118the need to keep changing the Perl interpreter. You can write modules
119that export functions, you can give those functions prototypes so they
120can be called like built-in functions, you can even write XS code to
121mess with the runtime data structures of the Perl interpreter if you
122want to implement really complicated things. If it can be done in a
123module instead of in the core, it's highly unlikely to be added.
124
125=item Is the feature generic enough?
126
127Is this something that only the submitter wants added to the language,
128or would it be broadly useful? Sometimes, instead of adding a feature
129with a tight focus, the porters might decide to wait until someone
130implements the more generalized feature. For instance, instead of
131implementing a ``delayed evaluation'' feature, the porters are waiting
132for a macro system that would permit delayed evaluation and much more.
133
134=item Does it potentially introduce new bugs?
135
136Radical rewrites of large chunks of the Perl interpreter have the
137potential to introduce new bugs. The smaller and more localized the
138change, the better.
139
140=item Does it preclude other desirable features?
141
142A patch is likely to be rejected if it closes off future avenues of
143development. For instance, a patch that placed a true and final
144interpretation on prototypes is likely to be rejected because there
145are still options for the future of prototypes that haven't been
146addressed.
147
148=item Is the implementation robust?
149
150Good patches (tight code, complete, correct) stand more chance of
151going in. Sloppy or incorrect patches might be placed on the back
152burner until the pumpking has time to fix, or might be discarded
153altogether without further notice.
154
155=item Is the implementation generic enough to be portable?
156
157The worst patches make use of a system-specific features. It's highly
158unlikely that nonportable additions to the Perl language will be
159accepted.
160
161=item Is there enough documentation?
162
163Patches without documentation are probably ill-thought out or
164incomplete. Nothing can be added without documentation, so submitting
165a patch for the appropriate manpages as well as the source code is
166always a good idea. If appropriate, patches should add to the test
167suite as well.
168
169=item Is there another way to do it?
170
171Larry said ``Although the Perl Slogan is I<There's More Than One Way
172to Do It>, I hesitate to make 10 ways to do something''. This is a
173tricky heuristic to navigate, though--one man's essential addition is
174another man's pointless cruft.
175
176=item Does it create too much work?
177
178Work for the pumpking, work for Perl programmers, work for module
179authors, ... Perl is supposed to be easy.
180
f6c51b38
GS
181=item Patches speak louder than words
182
183Working code is always preferred to pie-in-the-sky ideas. A patch to
184add a feature stands a much higher chance of making it to the language
185than does a random feature request, no matter how fervently argued the
186request might be. This ties into ``Will it be useful?'', as the fact
187that someone took the time to make the patch demonstrates a strong
188desire for the feature.
189
e8cd7eae
GS
190=back
191
192If you're on the list, you might hear the word ``core'' bandied
193around. It refers to the standard distribution. ``Hacking on the
194core'' means you're changing the C source code to the Perl
195interpreter. ``A core module'' is one that ships with Perl.
196
a1f349fd
MB
197=head2 Keeping in sync
198
e8cd7eae
GS
199The source code to the Perl interpreter, in its different versions, is
200kept in a repository managed by a revision control system (which is
201currently the Perforce program, see http://perforce.com/). The
202pumpkings and a few others have access to the repository to check in
203changes. Periodically the pumpking for the development version of Perl
204will release a new version, so the rest of the porters can see what's
2be4c08b
GS
205changed. The current state of the main trunk of repository, and patches
206that describe the individual changes that have happened since the last
207public release are available at this location:
208
209 ftp://ftp.linux.activestate.com/pub/staff/gsar/APC/
210
a1f349fd
MB
211If you are a member of the perl5-porters mailing list, it is a good
212thing to keep in touch with the most recent changes. If not only to
213verify if what you would have posted as a bug report isn't already
214solved in the most recent available perl development branch, also
215known as perl-current, bleading edge perl, bleedperl or bleadperl.
2be4c08b
GS
216
217Needless to say, the source code in perl-current is usually in a perpetual
218state of evolution. You should expect it to be very buggy. Do B<not> use
219it for any purpose other than testing and development.
e8cd7eae 220
a1f349fd
MB
221To keep in sync with the most recent branch can be done in several
222ways, but the most convenient and reliable way is using B<rsync>,
223available at ftp://rsync.samba.org/pub/rsync/ (other ways include ftp).
224
225If you choose to keep in sync using rsync, there are two approaches
226to do so:
227
228=over 4
229
230=item rsync'ing the source tree
231
232Presuming you are in the directory where your perl source resides,
233and you have rsync installed and available, you can `upgrade' to
234the bleadperl using:
235
236 # rsync -avz rsync://ftp.linux.activestate.com/perl-current/ .
237
238This takes care of updating every single item in the source tree to
239the latest applied patch level, creating files that are new (to your
240distribution) and setting date/time stamps of existing files to
241reflect the bleadperl status.
242
243You can than check what patch was the latest that was applied by
244looking in the file B<.patch>, which will show the number of the
245latest patch.
246
247If you have more than one machine to keep in sync, and not all of
248them have access to the WAN (so you are not able to rsync all the
249source trees to the real source), there are some ways to get around
250this problem.
251
252=over 4
253
254=item Using rsync over the LAN
255
256Set up a local rsync server which makes the rsynced source tree
257available to the LAN, and sync the other machines towards this
258directory.
259
260From http://rsync.samba.org/README.html:
261
262 "Rsync uses rsh or ssh for communication. It does not need to be
263 setuid and requires no special privileges for installation. It
264 does not require a inetd entry or a deamon. You must, however,
265 have a working rsh or ssh system. Using ssh is recommended for
266 its security features."
267
268=item Using pushing over the NFS
269
270Having the other systems mounted over the NFS, you can take an
271active pushing approach, in checking the just updated tree against
272the other not-yet synced trees. An example would be:
273
274 ...... Xsync
275
276Though this is not perfect. It could be improved with checking
277file checksums before updating. Not all NFS systems support
278reliable utime support (when used over the NFS).
279
280=back
281
282=item rsync'ing the patches
283
284The source tree is maintained by the pumpking who applies patches to
285the files in the tree. These patches are either created by the
286pumpking himself using C<diff -c> after updating the file manually or
287by applying patches sent in by posters on the perl5-porters list.
288These patches are also saved and rsync'able, so you can apply them
289yourself to the source files.
290
291Presuming you are in a directory where your patches reside, you can
292get them in sync with:
293
294 # rsync -avz rsync://ftp.linux.activestate.com/perl-current-diffs/ .
295
296This makes sure the latest available patch is downloaded to your
297patch directory.
298
299It's then up to you to apply these patches, using something like:
300
301 # last=`ls -rt1 *.gz | tail -1`
302 # rsync -avz rsync://ftp.linux.activestate.com/perl-current-diffs/ .
303 # find . -name '*.gz' -newer $last -exec gzcat {} \; >blead.patch
304 # cd ../perl-current
305 # patch -p1 -N <../perl-current-diffs/blead.patch
306
307or, since this is only a hint towards how it works, use CPAN-patchaperl
308from Andreas K├Ânig to have better control over the patching process.
309
310=back
311
312=head3 Why rsync the source tree
313
314=over 4
315
316=item It's easier
317
318Since you don't have to apply the patches yourself, you are sure all
319files in the source tree are in the right state.
320
321=item It's more recent
322
323According to Gurusamy Sarathy:
324
325 "... The rsync mirror is automatic and syncs with the repository
326 every five minutes.
327
328 Updating the patch area still requires manual intervention
329 (with all the goofiness that implies, which you've noted) and
330 is typically on a daily cycle. Making this process automatic
331 is on my tuit list, but don't ask me when."
332
333=item It's more reliable
334
335Well, since the patches are updated by hand, I don't have to say no
336more ... (see Sarathy's remark).
337
338=back
339
340=head3 Why rsync the patches
341
342=over 4
343
344=item It's easier
345
346If you have more than one machine that you want to keep in track with
347bleadperl, it's easier to rsync the patches only once and than apply
348them to all the source trees on the different machines.
349
350In case you try to keep in pace on 5 different machines, for which
351only one of them has access to the WAN, rsync'ing all the source
352tree's should than be done 5 times over the NFS, whereas having
353rsync'ed the patches only once, I can apply them to all the source
354trees automatically. Need I say more ;-)
355
356=item It's a good reference
357
358If you do not only like to have the most recent development branch,
359but also like to B<fix> bugs, or extend features, you want to dive
360into the sources. If you are a seasoned perl core diver, you don't
361need no manuals, tips, roadmaps, perlguts.pod or other aids to find
362your way around. But if you are a starter, the patches may help you
363in finding where you should start and how to change the bits that
364bug you.
365
366The file B<Changes> is updated on occasions the pumpking sees as his
367own little sync points. On those occasions, he releases a tar-ball of
368the current source tree (i.e. perl@7582.tar.gz), which will be an
369excellent point to start with when choosing to use the 'rsync the
370patches' scheme. Starting with perl@7582, which means a set of source
371files on which the latest applied patch is number 7582, you apply all
372succeeding patches available from than on (7583, 7584, ...).
373
374You can use the patches later as a kind of search archive.
375
376=over 4
377
378=item Finding a start point
379
380If you want to fix/change the behaviour of function/feature Foo, just
381scan the patches for patches that mention Foo either in the subject,
382the comments, or the body of the fix. A good change the patch shows
383you the files that are affected by that patch which are very likely
384to be the starting point of your journey into the guts of perl.
385
386=item Finding how to fix a bug
387
388If you've found I<where> the function/feature Foo misbehaves, but you
389don't know how to fix it (but you do know the change you want to
390make), you can, again, peruse the patches for similar changes and
391look how others apply the fix.
392
393=item Finding the source of misbehaviour
394
395When you keep in sync with bleadperl, the pumpking would love to
396I<see> that the community efforts realy work. So after each of his
397sync points, you are to 'make test' to check if everything is still
398in working order. If it is, you do 'make ok', which will send an OK
399report to perlbug@perl.org. (If you do not have access to a mailer
400from the sytem you just finished successfully 'make test', you can
401do 'make okfile', which creates the file C<perl.ok>, which you can
402than take to your favourite mailer and mail yourself).
403
404But of course, as allways, things will not allways lead to a success
405path, and one or more test do not pass the 'make test'. Before
406sending in a bug report (using 'make nok' or 'make nokfile'), check
407the mailing list if someone else has reported the bug already and if
408so, confirm it by replying to that message. If not, you might want to
409trace the source of that misbehaviour B<before> sending in the bug,
410which will help all the other porters in finding the solution.
411
412Here the saved patches come in very handy. You can check in there
413which patch changed what file and what change caused the
414misbehaviour. If you note that in the bug report, it saves the one
415trying to solve it, looking for that point.
416
417=back
418
419If searching the patches is too bothersome, you might consider using
420perl's bugtron to find more information about discussions and
421ramblings on posted bugs.
422
423=back
424
425=head2 Submitting patches
426
e8cd7eae
GS
427Always submit patches to I<perl5-porters@perl.org>. This lets other
428porters review your patch, which catches a surprising number of errors
f6c51b38
GS
429in patches. Either use the diff program (available in source code
430form from I<ftp://ftp.gnu.org/pub/gnu/>), or use Johan Vromans'
431I<makepatch> (available from I<CPAN/authors/id/JV/>). Unified diffs
432are preferred, but context diffs are accepted. Do not send RCS-style
433diffs or diffs without context lines. More information is given in
434the I<Porting/patching.pod> file in the Perl source distribution.
435Please patch against the latest B<development> version (e.g., if
436you're fixing a bug in the 5.005 track, patch against the latest
4375.005_5x version). Only patches that survive the heat of the
438development branch get applied to maintenance versions.
e8cd7eae
GS
439
440Your patch should update the documentation and test suite.
441
442To report a bug in Perl, use the program I<perlbug> which comes with
443Perl (if you can't get Perl to work, send mail to the address
444I<perlbug@perl.com> or I<perlbug@perl.org>). Reporting bugs through
445I<perlbug> feeds into the automated bug-tracking system, access to
446which is provided through the web at I<http://bugs.perl.org/>. It
447often pays to check the archives of the perl5-porters mailing list to
448see whether the bug you're reporting has been reported before, and if
449so whether it was considered a bug. See above for the location of
450the searchable archives.
451
452The CPAN testers (I<http://testers.cpan.org/>) are a group of
453volunteers who test CPAN modules on a variety of platforms. Perl Labs
f6c51b38
GS
454(I<http://labs.perl.org/>) automatically tests Perl source releases on
455platforms and gives feedback to the CPAN testers mailing list. Both
456efforts welcome volunteers.
e8cd7eae 457
e8cd7eae
GS
458It's a good idea to read and lurk for a while before chipping in.
459That way you'll get to see the dynamic of the conversations, learn the
460personalities of the players, and hopefully be better prepared to make
461a useful contribution when do you speak up.
462
463If after all this you still think you want to join the perl5-porters
f6c51b38
GS
464mailing list, send mail to I<perl5-porters-subscribe@perl.org>. To
465unsubscribe, send mail to I<perl5-porters-unsubscribe@perl.org>.
e8cd7eae 466
a422fd2d
SC
467To hack on the Perl guts, you'll need to read the following things:
468
469=over 3
470
471=item L<perlguts>
472
473This is of paramount importance, since it's the documentation of what
474goes where in the Perl source. Read it over a couple of times and it
475might start to make sense - don't worry if it doesn't yet, because the
476best way to study it is to read it in conjunction with poking at Perl
477source, and we'll do that later on.
478
479You might also want to look at Gisle Aas's illustrated perlguts -
480there's no guarantee that this will be absolutely up-to-date with the
481latest documentation in the Perl core, but the fundamentals will be
482right. (http://gisle.aas.no/perl/illguts/)
483
484=item L<perlxstut> and L<perlxs>
485
486A working knowledge of XSUB programming is incredibly useful for core
487hacking; XSUBs use techniques drawn from the PP code, the portion of the
488guts that actually executes a Perl program. It's a lot gentler to learn
489those techniques from simple examples and explanation than from the core
490itself.
491
492=item L<perlapi>
493
494The documentation for the Perl API explains what some of the internal
495functions do, as well as the many macros used in the source.
496
497=item F<Porting/pumpkin.pod>
498
499This is a collection of words of wisdom for a Perl porter; some of it is
500only useful to the pumpkin holder, but most of it applies to anyone
501wanting to go about Perl development.
502
503=item The perl5-porters FAQ
504
505This is posted to perl5-porters at the beginning on every month, and
506should be available from http://perlhacker.org/p5p-faq; alternatively,
507you can get the FAQ emailed to you by sending mail to
508C<perl5-porters-faq@perl.org>. It contains hints on reading
509perl5-porters, information on how perl5-porters works and how Perl
510development in general works.
511
512=back
513
514=head2 Finding Your Way Around
515
516Perl maintenance can be split into a number of areas, and certain people
517(pumpkins) will have responsibility for each area. These areas sometimes
518correspond to files or directories in the source kit. Among the areas are:
519
520=over 3
521
522=item Core modules
523
524Modules shipped as part of the Perl core live in the F<lib/> and F<ext/>
525subdirectories: F<lib/> is for the pure-Perl modules, and F<ext/>
526contains the core XS modules.
527
528=item Documentation
529
530Documentation maintenance includes looking after everything in the
531F<pod/> directory, (as well as contributing new documentation) and
532the documentation to the modules in core.
533
534=item Configure
535
536The configure process is the way we make Perl portable across the
537myriad of operating systems it supports. Responsibility for the
538configure, build and installation process, as well as the overall
539portability of the core code rests with the configure pumpkin - others
540help out with individual operating systems.
541
542The files involved are the operating system directories, (F<win32/>,
543F<os2/>, F<vms/> and so on) the shell scripts which generate F<config.h>
544and F<Makefile>, as well as the metaconfig files which generate
545F<Configure>. (metaconfig isn't included in the core distribution.)
546
547=item Interpreter
548
549And of course, there's the core of the Perl interpreter itself. Let's
550have a look at that in a little more detail.
551
552=back
553
554Before we leave looking at the layout, though, don't forget that
555F<MANIFEST> contains not only the file names in the Perl distribution,
556but short descriptions of what's in them, too. For an overview of the
557important files, try this:
558
559 perl -lne 'print if /^[^\/]+\.[ch]\s+/' MANIFEST
560
561=head2 Elements of the interpreter
562
563The work of the interpreter has two main stages: compiling the code
564into the internal representation, or bytecode, and then executing it.
565L<perlguts/Compiled code> explains exactly how the compilation stage
566happens.
567
568Here is a short breakdown of perl's operation:
569
570=over 3
571
572=item Startup
573
574The action begins in F<perlmain.c>. (or F<miniperlmain.c> for miniperl)
575This is very high-level code, enough to fit on a single screen, and it
576resembles the code found in L<perlembed>; most of the real action takes
577place in F<perl.c>
578
579First, F<perlmain.c> allocates some memory and constructs a Perl
580interpreter:
581
582 1 PERL_SYS_INIT3(&argc,&argv,&env);
583 2
584 3 if (!PL_do_undump) {
585 4 my_perl = perl_alloc();
586 5 if (!my_perl)
587 6 exit(1);
588 7 perl_construct(my_perl);
589 8 PL_perl_destruct_level = 0;
590 9 }
591
592Line 1 is a macro, and its definition is dependent on your operating
593system. Line 3 references C<PL_do_undump>, a global variable - all
594global variables in Perl start with C<PL_>. This tells you whether the
595current running program was created with the C<-u> flag to perl and then
596F<undump>, which means it's going to be false in any sane context.
597
598Line 4 calls a function in F<perl.c> to allocate memory for a Perl
599interpreter. It's quite a simple function, and the guts of it looks like
600this:
601
602 my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));
603
604Here you see an example of Perl's system abstraction, which we'll see
605later: C<PerlMem_malloc> is either your system's C<malloc>, or Perl's
606own C<malloc> as defined in F<malloc.c> if you selected that option at
607configure time.
608
609Next, in line 7, we construct the interpreter; this sets up all the
610special variables that Perl needs, the stacks, and so on.
611
612Now we pass Perl the command line options, and tell it to go:
613
614 exitstatus = perl_parse(my_perl, xs_init, argc, argv, (char **)NULL);
615 if (!exitstatus) {
616 exitstatus = perl_run(my_perl);
617 }
618
619
620C<perl_parse> is actually a wrapper around C<S_parse_body>, as defined
621in F<perl.c>, which processes the command line options, sets up any
622statically linked XS modules, opens the program and calls C<yyparse> to
623parse it.
624
625=item Parsing
626
627The aim of this stage is to take the Perl source, and turn it into an op
628tree. We'll see what one of those looks like later. Strictly speaking,
629there's three things going on here.
630
631C<yyparse>, the parser, lives in F<perly.c>, although you're better off
632reading the original YACC input in F<perly.y>. (Yes, Virginia, there
633B<is> a YACC grammar for Perl!) The job of the parser is to take your
634code and `understand' it, splitting it into sentences, deciding which
635operands go with which operators and so on.
636
637The parser is nobly assisted by the lexer, which chunks up your input
638into tokens, and decides what type of thing each token is: a variable
639name, an operator, a bareword, a subroutine, a core function, and so on.
640The main point of entry to the lexer is C<yylex>, and that and its
641associated routines can be found in F<toke.c>. Perl isn't much like
642other computer languages; it's highly context sensitive at times, it can
643be tricky to work out what sort of token something is, or where a token
644ends. As such, there's a lot of interplay between the tokeniser and the
645parser, which can get pretty frightening if you're not used to it.
646
647As the parser understands a Perl program, it builds up a tree of
648operations for the interpreter to perform during execution. The routines
649which construct and link together the various operations are to be found
650in F<op.c>, and will be examined later.
651
652=item Optimization
653
654Now the parsing stage is complete, and the finished tree represents
655the operations that the Perl interpreter needs to perform to execute our
656program. Next, Perl does a dry run over the tree looking for
657optimisations: constant expressions such as C<3 + 4> will be computed
658now, and the optimizer will also see if any multiple operations can be
659replaced with a single one. For instance, to fetch the variable C<$foo>,
660instead of grabbing the glob C<*foo> and looking at the scalar
661component, the optimizer fiddles the op tree to use a function which
662directly looks up the scalar in question. The main optimizer is C<peep>
663in F<op.c>, and many ops have their own optimizing functions.
664
665=item Running
666
667Now we're finally ready to go: we have compiled Perl byte code, and all
668that's left to do is run it. The actual execution is done by the
669C<runops_standard> function in F<run.c>; more specifically, it's done by
670these three innocent looking lines:
671
672 while ((PL_op = CALL_FPTR(PL_op->op_ppaddr)(aTHX))) {
673 PERL_ASYNC_CHECK();
674 }
675
676You may be more comfortable with the Perl version of that:
677
678 PERL_ASYNC_CHECK() while $Perl::op = &{$Perl::op->{function}};
679
680Well, maybe not. Anyway, each op contains a function pointer, which
681stipulates the function which will actually carry out the operation.
682This function will return the next op in the sequence - this allows for
683things like C<if> which choose the next op dynamically at run time.
684The C<PERL_ASYNC_CHECK> makes sure that things like signals interrupt
685execution if required.
686
687The actual functions called are known as PP code, and they're spread
688between four files: F<pp_hot.c> contains the `hot' code, which is most
689often used and highly optimized, F<pp_sys.c> contains all the
690system-specific functions, F<pp_ctl.c> contains the functions which
691implement control structures (C<if>, C<while> and the like) and F<pp.c>
692contains everything else. These are, if you like, the C code for Perl's
693built-in functions and operators.
694
695=back
696
697=head2 Internal Variable Types
698
699You should by now have had a look at L<perlguts>, which tells you about
700Perl's internal variable types: SVs, HVs, AVs and the rest. If not, do
701that now.
702
703These variables are used not only to represent Perl-space variables, but
704also any constants in the code, as well as some structures completely
705internal to Perl. The symbol table, for instance, is an ordinary Perl
706hash. Your code is represented by an SV as it's read into the parser;
707any program files you call are opened via ordinary Perl filehandles, and
708so on.
709
710The core L<Devel::Peek|Devel::Peek> module lets us examine SVs from a
711Perl program. Let's see, for instance, how Perl treats the constant
712C<"hello">.
713
714 % perl -MDevel::Peek -e 'Dump("hello")'
715 1 SV = PV(0xa041450) at 0xa04ecbc
716 2 REFCNT = 1
717 3 FLAGS = (POK,READONLY,pPOK)
718 4 PV = 0xa0484e0 "hello"\0
719 5 CUR = 5
720 6 LEN = 6
721
722Reading C<Devel::Peek> output takes a bit of practise, so let's go
723through it line by line.
724
725Line 1 tells us we're looking at an SV which lives at C<0xa04ecbc> in
726memory. SVs themselves are very simple structures, but they contain a
727pointer to a more complex structure. In this case, it's a PV, a
728structure which holds a string value, at location C<0xa041450>. Line 2
729is the reference count; there are no other references to this data, so
730it's 1.
731
732Line 3 are the flags for this SV - it's OK to use it as a PV, it's a
733read-only SV (because it's a constant) and the data is a PV internally.
734Next we've got the contents of the string, starting at location
735C<0xa0484e0>.
736
737Line 5 gives us the current length of the string - note that this does
738B<not> include the null terminator. Line 6 is not the length of the
739string, but the length of the currently allocated buffer; as the string
740grows, Perl automatically extends the available storage via a routine
741called C<SvGROW>.
742
743You can get at any of these quantities from C very easily; just add
744C<Sv> to the name of the field shown in the snippet, and you've got a
745macro which will return the value: C<SvCUR(sv)> returns the current
746length of the string, C<SvREFCOUNT(sv)> returns the reference count,
747C<SvPV(sv, len)> returns the string itself with its length, and so on.
748More macros to manipulate these properties can be found in L<perlguts>.
749
750Let's take an example of manipulating a PV, from C<sv_catpvn>, in F<sv.c>
751
752 1 void
753 2 Perl_sv_catpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
754 3 {
755 4 STRLEN tlen;
756 5 char *junk;
757
758 6 junk = SvPV_force(sv, tlen);
759 7 SvGROW(sv, tlen + len + 1);
760 8 if (ptr == junk)
761 9 ptr = SvPVX(sv);
762 10 Move(ptr,SvPVX(sv)+tlen,len,char);
763 11 SvCUR(sv) += len;
764 12 *SvEND(sv) = '\0';
765 13 (void)SvPOK_only_UTF8(sv); /* validate pointer */
766 14 SvTAINT(sv);
767 15 }
768
769This is a function which adds a string, C<ptr>, of length C<len> onto
770the end of the PV stored in C<sv>. The first thing we do in line 6 is
771make sure that the SV B<has> a valid PV, by calling the C<SvPV_force>
772macro to force a PV. As a side effect, C<tlen> gets set to the current
773value of the PV, and the PV itself is returned to C<junk>.
774
b1866b2d 775In line 7, we make sure that the SV will have enough room to accommodate
a422fd2d
SC
776the old string, the new string and the null terminator. If C<LEN> isn't
777big enough, C<SvGROW> will reallocate space for us.
778
779Now, if C<junk> is the same as the string we're trying to add, we can
780grab the string directly from the SV; C<SvPVX> is the address of the PV
781in the SV.
782
783Line 10 does the actual catenation: the C<Move> macro moves a chunk of
784memory around: we move the string C<ptr> to the end of the PV - that's
785the start of the PV plus its current length. We're moving C<len> bytes
786of type C<char>. After doing so, we need to tell Perl we've extended the
787string, by altering C<CUR> to reflect the new length. C<SvEND> is a
788macro which gives us the end of the string, so that needs to be a
789C<"\0">.
790
791Line 13 manipulates the flags; since we've changed the PV, any IV or NV
792values will no longer be valid: if we have C<$a=10; $a.="6";> we don't
793want to use the old IV of 10. C<SvPOK_only_utf8> is a special UTF8-aware
794version of C<SvPOK_only>, a macro which turns off the IOK and NOK flags
795and turns on POK. The final C<SvTAINT> is a macro which launders tainted
796data if taint mode is turned on.
797
798AVs and HVs are more complicated, but SVs are by far the most common
799variable type being thrown around. Having seen something of how we
800manipulate these, let's go on and look at how the op tree is
801constructed.
802
803=head2 Op Trees
804
805First, what is the op tree, anyway? The op tree is the parsed
806representation of your program, as we saw in our section on parsing, and
807it's the sequence of operations that Perl goes through to execute your
808program, as we saw in L</Running>.
809
810An op is a fundamental operation that Perl can perform: all the built-in
811functions and operators are ops, and there are a series of ops which
812deal with concepts the interpreter needs internally - entering and
813leaving a block, ending a statement, fetching a variable, and so on.
814
815The op tree is connected in two ways: you can imagine that there are two
816"routes" through it, two orders in which you can traverse the tree.
817First, parse order reflects how the parser understood the code, and
818secondly, execution order tells perl what order to perform the
819operations in.
820
821The easiest way to examine the op tree is to stop Perl after it has
822finished parsing, and get it to dump out the tree. This is exactly what
823the compiler backends L<B::Terse|B::Terse> and L<B::Debug|B::Debug> do.
824
825Let's have a look at how Perl sees C<$a = $b + $c>:
826
827 % perl -MO=Terse -e '$a=$b+$c'
828 1 LISTOP (0x8179888) leave
829 2 OP (0x81798b0) enter
830 3 COP (0x8179850) nextstate
831 4 BINOP (0x8179828) sassign
832 5 BINOP (0x8179800) add [1]
833 6 UNOP (0x81796e0) null [15]
834 7 SVOP (0x80fafe0) gvsv GV (0x80fa4cc) *b
835 8 UNOP (0x81797e0) null [15]
836 9 SVOP (0x8179700) gvsv GV (0x80efeb0) *c
837 10 UNOP (0x816b4f0) null [15]
838 11 SVOP (0x816dcf0) gvsv GV (0x80fa460) *a
839
840Let's start in the middle, at line 4. This is a BINOP, a binary
841operator, which is at location C<0x8179828>. The specific operator in
842question is C<sassign> - scalar assignment - and you can find the code
843which implements it in the function C<pp_sassign> in F<pp_hot.c>. As a
844binary operator, it has two children: the add operator, providing the
845result of C<$b+$c>, is uppermost on line 5, and the left hand side is on
846line 10.
847
848Line 10 is the null op: this does exactly nothing. What is that doing
849there? If you see the null op, it's a sign that something has been
850optimized away after parsing. As we mentioned in L</Optimization>,
851the optimization stage sometimes converts two operations into one, for
852example when fetching a scalar variable. When this happens, instead of
853rewriting the op tree and cleaning up the dangling pointers, it's easier
854just to replace the redundant operation with the null op. Originally,
855the tree would have looked like this:
856
857 10 SVOP (0x816b4f0) rv2sv [15]
858 11 SVOP (0x816dcf0) gv GV (0x80fa460) *a
859
860That is, fetch the C<a> entry from the main symbol table, and then look
861at the scalar component of it: C<gvsv> (C<pp_gvsv> into F<pp_hot.c>)
862happens to do both these things.
863
864The right hand side, starting at line 5 is similar to what we've just
865seen: we have the C<add> op (C<pp_add> also in F<pp_hot.c>) add together
866two C<gvsv>s.
867
868Now, what's this about?
869
870 1 LISTOP (0x8179888) leave
871 2 OP (0x81798b0) enter
872 3 COP (0x8179850) nextstate
873
874C<enter> and C<leave> are scoping ops, and their job is to perform any
875housekeeping every time you enter and leave a block: lexical variables
876are tidied up, unreferenced variables are destroyed, and so on. Every
877program will have those first three lines: C<leave> is a list, and its
878children are all the statements in the block. Statements are delimited
879by C<nextstate>, so a block is a collection of C<nextstate> ops, with
880the ops to be performed for each statement being the children of
881C<nextstate>. C<enter> is a single op which functions as a marker.
882
883That's how Perl parsed the program, from top to bottom:
884
885 Program
886 |
887 Statement
888 |
889 =
890 / \
891 / \
892 $a +
893 / \
894 $b $c
895
896However, it's impossible to B<perform> the operations in this order:
897you have to find the values of C<$b> and C<$c> before you add them
898together, for instance. So, the other thread that runs through the op
899tree is the execution order: each op has a field C<op_next> which points
900to the next op to be run, so following these pointers tells us how perl
901executes the code. We can traverse the tree in this order using
902the C<exec> option to C<B::Terse>:
903
904 % perl -MO=Terse,exec -e '$a=$b+$c'
905 1 OP (0x8179928) enter
906 2 COP (0x81798c8) nextstate
907 3 SVOP (0x81796c8) gvsv GV (0x80fa4d4) *b
908 4 SVOP (0x8179798) gvsv GV (0x80efeb0) *c
909 5 BINOP (0x8179878) add [1]
910 6 SVOP (0x816dd38) gvsv GV (0x80fa468) *a
911 7 BINOP (0x81798a0) sassign
912 8 LISTOP (0x8179900) leave
913
914This probably makes more sense for a human: enter a block, start a
915statement. Get the values of C<$b> and C<$c>, and add them together.
916Find C<$a>, and assign one to the other. Then leave.
917
918The way Perl builds up these op trees in the parsing process can be
919unravelled by examining F<perly.y>, the YACC grammar. Let's take the
920piece we need to construct the tree for C<$a = $b + $c>
921
922 1 term : term ASSIGNOP term
923 2 { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
924 3 | term ADDOP term
925 4 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
926
927If you're not used to reading BNF grammars, this is how it works: You're
928fed certain things by the tokeniser, which generally end up in upper
929case. Here, C<ADDOP>, is provided when the tokeniser sees C<+> in your
930code. C<ASSIGNOP> is provided when C<=> is used for assigning. These are
931`terminal symbols', because you can't get any simpler than them.
932
933The grammar, lines one and three of the snippet above, tells you how to
934build up more complex forms. These complex forms, `non-terminal symbols'
935are generally placed in lower case. C<term> here is a non-terminal
936symbol, representing a single expression.
937
938The grammar gives you the following rule: you can make the thing on the
939left of the colon if you see all the things on the right in sequence.
940This is called a "reduction", and the aim of parsing is to completely
941reduce the input. There are several different ways you can perform a
942reduction, separated by vertical bars: so, C<term> followed by C<=>
943followed by C<term> makes a C<term>, and C<term> followed by C<+>
944followed by C<term> can also make a C<term>.
945
946So, if you see two terms with an C<=> or C<+>, between them, you can
947turn them into a single expression. When you do this, you execute the
948code in the block on the next line: if you see C<=>, you'll do the code
949in line 2. If you see C<+>, you'll do the code in line 4. It's this code
950which contributes to the op tree.
951
952 | term ADDOP term
953 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
954
955What this does is creates a new binary op, and feeds it a number of
956variables. The variables refer to the tokens: C<$1> is the first token in
957the input, C<$2> the second, and so on - think regular expression
958backreferences. C<$$> is the op returned from this reduction. So, we
959call C<newBINOP> to create a new binary operator. The first parameter to
960C<newBINOP>, a function in F<op.c>, is the op type. It's an addition
961operator, so we want the type to be C<ADDOP>. We could specify this
962directly, but it's right there as the second token in the input, so we
963use C<$2>. The second parameter is the op's flags: 0 means `nothing
964special'. Then the things to add: the left and right hand side of our
965expression, in scalar context.
966
967=head2 Stacks
968
969When perl executes something like C<addop>, how does it pass on its
970results to the next op? The answer is, through the use of stacks. Perl
971has a number of stacks to store things it's currently working on, and
972we'll look at the three most important ones here.
973
974=over 3
975
976=item Argument stack
977
978Arguments are passed to PP code and returned from PP code using the
979argument stack, C<ST>. The typical way to handle arguments is to pop
980them off the stack, deal with them how you wish, and then push the result
981back onto the stack. This is how, for instance, the cosine operator
982works:
983
984 NV value;
985 value = POPn;
986 value = Perl_cos(value);
987 XPUSHn(value);
988
989We'll see a more tricky example of this when we consider Perl's macros
990below. C<POPn> gives you the NV (floating point value) of the top SV on
991the stack: the C<$x> in C<cos($x)>. Then we compute the cosine, and push
992the result back as an NV. The C<X> in C<XPUSHn> means that the stack
993should be extended if necessary - it can't be necessary here, because we
994know there's room for one more item on the stack, since we've just
995removed one! The C<XPUSH*> macros at least guarantee safety.
996
997Alternatively, you can fiddle with the stack directly: C<SP> gives you
998the first element in your portion of the stack, and C<TOP*> gives you
999the top SV/IV/NV/etc. on the stack. So, for instance, to do unary
1000negation of an integer:
1001
1002 SETi(-TOPi);
1003
1004Just set the integer value of the top stack entry to its negation.
1005
1006Argument stack manipulation in the core is exactly the same as it is in
1007XSUBs - see L<perlxstut>, L<perlxs> and L<perlguts> for a longer
1008description of the macros used in stack manipulation.
1009
1010=item Mark stack
1011
1012I say `your portion of the stack' above because PP code doesn't
1013necessarily get the whole stack to itself: if your function calls
1014another function, you'll only want to expose the arguments aimed for the
1015called function, and not (necessarily) let it get at your own data. The
1016way we do this is to have a `virtual' bottom-of-stack, exposed to each
1017function. The mark stack keeps bookmarks to locations in the argument
1018stack usable by each function. For instance, when dealing with a tied
1019variable, (internally, something with `P' magic) Perl has to call
1020methods for accesses to the tied variables. However, we need to separate
1021the arguments exposed to the method to the argument exposed to the
1022original function - the store or fetch or whatever it may be. Here's how
1023the tied C<push> is implemented; see C<av_push> in F<av.c>:
1024
1025 1 PUSHMARK(SP);
1026 2 EXTEND(SP,2);
1027 3 PUSHs(SvTIED_obj((SV*)av, mg));
1028 4 PUSHs(val);
1029 5 PUTBACK;
1030 6 ENTER;
1031 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1032 8 LEAVE;
1033 9 POPSTACK;
13a2d996 1034
a422fd2d
SC
1035The lines which concern the mark stack are the first, fifth and last
1036lines: they save away, restore and remove the current position of the
1037argument stack.
1038
1039Let's examine the whole implementation, for practice:
1040
1041 1 PUSHMARK(SP);
1042
1043Push the current state of the stack pointer onto the mark stack. This is
1044so that when we've finished adding items to the argument stack, Perl
1045knows how many things we've added recently.
1046
1047 2 EXTEND(SP,2);
1048 3 PUSHs(SvTIED_obj((SV*)av, mg));
1049 4 PUSHs(val);
1050
1051We're going to add two more items onto the argument stack: when you have
1052a tied array, the C<PUSH> subroutine receives the object and the value
1053to be pushed, and that's exactly what we have here - the tied object,
1054retrieved with C<SvTIED_obj>, and the value, the SV C<val>.
1055
1056 5 PUTBACK;
1057
1058Next we tell Perl to make the change to the global stack pointer: C<dSP>
1059only gave us a local copy, not a reference to the global.
1060
1061 6 ENTER;
1062 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1063 8 LEAVE;
1064
1065C<ENTER> and C<LEAVE> localise a block of code - they make sure that all
1066variables are tidied up, everything that has been localised gets
1067its previous value returned, and so on. Think of them as the C<{> and
1068C<}> of a Perl block.
1069
1070To actually do the magic method call, we have to call a subroutine in
1071Perl space: C<call_method> takes care of that, and it's described in
1072L<perlcall>. We call the C<PUSH> method in scalar context, and we're
1073going to discard its return value.
1074
1075 9 POPSTACK;
1076
1077Finally, we remove the value we placed on the mark stack, since we
1078don't need it any more.
1079
1080=item Save stack
1081
1082C doesn't have a concept of local scope, so perl provides one. We've
1083seen that C<ENTER> and C<LEAVE> are used as scoping braces; the save
1084stack implements the C equivalent of, for example:
1085
1086 {
1087 local $foo = 42;
1088 ...
1089 }
1090
1091See L<perlguts/Localising Changes> for how to use the save stack.
1092
1093=back
1094
1095=head2 Millions of Macros
1096
1097One thing you'll notice about the Perl source is that it's full of
1098macros. Some have called the pervasive use of macros the hardest thing
1099to understand, others find it adds to clarity. Let's take an example,
1100the code which implements the addition operator:
1101
1102 1 PP(pp_add)
1103 2 {
1104 3 djSP; dATARGET; tryAMAGICbin(add,opASSIGN);
1105 4 {
1106 5 dPOPTOPnnrl_ul;
1107 6 SETn( left + right );
1108 7 RETURN;
1109 8 }
1110 9 }
1111
1112Every line here (apart from the braces, of course) contains a macro. The
1113first line sets up the function declaration as Perl expects for PP code;
1114line 3 sets up variable declarations for the argument stack and the
1115target, the return value of the operation. Finally, it tries to see if
1116the addition operation is overloaded; if so, the appropriate subroutine
1117is called.
1118
1119Line 5 is another variable declaration - all variable declarations start
1120with C<d> - which pops from the top of the argument stack two NVs (hence
1121C<nn>) and puts them into the variables C<right> and C<left>, hence the
1122C<rl>. These are the two operands to the addition operator. Next, we
1123call C<SETn> to set the NV of the return value to the result of adding
1124the two values. This done, we return - the C<RETURN> macro makes sure
1125that our return value is properly handled, and we pass the next operator
1126to run back to the main run loop.
1127
1128Most of these macros are explained in L<perlapi>, and some of the more
1129important ones are explained in L<perlxs> as well. Pay special attention
1130to L<perlguts/Background and PERL_IMPLICIT_CONTEXT> for information on
1131the C<[pad]THX_?> macros.
1132
1133
1134=head2 Poking at Perl
1135
1136To really poke around with Perl, you'll probably want to build Perl for
1137debugging, like this:
1138
1139 ./Configure -d -D optimize=-g
1140 make
1141
1142C<-g> is a flag to the C compiler to have it produce debugging
1143information which will allow us to step through a running program.
1144F<Configure> will also turn on the C<DEBUGGING> compilation symbol which
1145enables all the internal debugging code in Perl. There are a whole bunch
1146of things you can debug with this: L<perlrun> lists them all, and the
1147best way to find out about them is to play about with them. The most
1148useful options are probably
1149
1150 l Context (loop) stack processing
1151 t Trace execution
1152 o Method and overloading resolution
1153 c String/numeric conversions
1154
1155Some of the functionality of the debugging code can be achieved using XS
1156modules.
13a2d996 1157
a422fd2d
SC
1158 -Dr => use re 'debug'
1159 -Dx => use O 'Debug'
1160
1161=head2 Using a source-level debugger
1162
1163If the debugging output of C<-D> doesn't help you, it's time to step
1164through perl's execution with a source-level debugger.
1165
1166=over 3
1167
1168=item *
1169
1170We'll use C<gdb> for our examples here; the principles will apply to any
1171debugger, but check the manual of the one you're using.
1172
1173=back
1174
1175To fire up the debugger, type
1176
1177 gdb ./perl
1178
1179You'll want to do that in your Perl source tree so the debugger can read
1180the source code. You should see the copyright message, followed by the
1181prompt.
1182
1183 (gdb)
1184
1185C<help> will get you into the documentation, but here are the most
1186useful commands:
1187
1188=over 3
1189
1190=item run [args]
1191
1192Run the program with the given arguments.
1193
1194=item break function_name
1195
1196=item break source.c:xxx
1197
1198Tells the debugger that we'll want to pause execution when we reach
1199either the named function (but see L</Function names>!) or the given
1200line in the named source file.
1201
1202=item step
1203
1204Steps through the program a line at a time.
1205
1206=item next
1207
1208Steps through the program a line at a time, without descending into
1209functions.
1210
1211=item continue
1212
1213Run until the next breakpoint.
1214
1215=item finish
1216
1217Run until the end of the current function, then stop again.
1218
13a2d996 1219=item 'enter'
a422fd2d
SC
1220
1221Just pressing Enter will do the most recent operation again - it's a
1222blessing when stepping through miles of source code.
1223
1224=item print
1225
1226Execute the given C code and print its results. B<WARNING>: Perl makes
1227heavy use of macros, and F<gdb> is not aware of macros. You'll have to
1228substitute them yourself. So, for instance, you can't say
1229
1230 print SvPV_nolen(sv)
1231
1232but you have to say
1233
1234 print Perl_sv_2pv_nolen(sv)
1235
1236You may find it helpful to have a "macro dictionary", which you can
1237produce by saying C<cpp -dM perl.c | sort>. Even then, F<cpp> won't
1238recursively apply the macros for you.
1239
1240=back
1241
1242=head2 Dumping Perl Data Structures
1243
1244One way to get around this macro hell is to use the dumping functions in
1245F<dump.c>; these work a little like an internal
1246L<Devel::Peek|Devel::Peek>, but they also cover OPs and other structures
1247that you can't get at from Perl. Let's take an example. We'll use the
1248C<$a = $b + $c> we used before, but give it a bit of context:
1249C<$b = "6XXXX"; $c = 2.3;>. Where's a good place to stop and poke around?
1250
1251What about C<pp_add>, the function we examined earlier to implement the
1252C<+> operator:
1253
1254 (gdb) break Perl_pp_add
1255 Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.
1256
1257Notice we use C<Perl_pp_add> and not C<pp_add> - see L<perlguts/Function Names>.
1258With the breakpoint in place, we can run our program:
1259
1260 (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'
1261
1262Lots of junk will go past as gdb reads in the relevant source files and
1263libraries, and then:
1264
1265 Breakpoint 1, Perl_pp_add () at pp_hot.c:309
1266 309 djSP; dATARGET; tryAMAGICbin(add,opASSIGN);
1267 (gdb) step
1268 311 dPOPTOPnnrl_ul;
1269 (gdb)
1270
1271We looked at this bit of code before, and we said that C<dPOPTOPnnrl_ul>
1272arranges for two C<NV>s to be placed into C<left> and C<right> - let's
1273slightly expand it:
1274
1275 #define dPOPTOPnnrl_ul NV right = POPn; \
1276 SV *leftsv = TOPs; \
1277 NV left = USE_LEFT(leftsv) ? SvNV(leftsv) : 0.0
1278
1279C<POPn> takes the SV from the top of the stack and obtains its NV either
1280directly (if C<SvNOK> is set) or by calling the C<sv_2nv> function.
1281C<TOPs> takes the next SV from the top of the stack - yes, C<POPn> uses
1282C<TOPs> - but doesn't remove it. We then use C<SvNV> to get the NV from
1283C<leftsv> in the same way as before - yes, C<POPn> uses C<SvNV>.
1284
1285Since we don't have an NV for C<$b>, we'll have to use C<sv_2nv> to
1286convert it. If we step again, we'll find ourselves there:
1287
1288 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
1289 1669 if (!sv)
1290 (gdb)
1291
1292We can now use C<Perl_sv_dump> to investigate the SV:
1293
1294 SV = PV(0xa057cc0) at 0xa0675d0
1295 REFCNT = 1
1296 FLAGS = (POK,pPOK)
1297 PV = 0xa06a510 "6XXXX"\0
1298 CUR = 5
1299 LEN = 6
1300 $1 = void
1301
1302We know we're going to get C<6> from this, so let's finish the
1303subroutine:
1304
1305 (gdb) finish
1306 Run till exit from #0 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
1307 0x462669 in Perl_pp_add () at pp_hot.c:311
1308 311 dPOPTOPnnrl_ul;
1309
1310We can also dump out this op: the current op is always stored in
1311C<PL_op>, and we can dump it with C<Perl_op_dump>. This'll give us
1312similar output to L<B::Debug|B::Debug>.
1313
1314 {
1315 13 TYPE = add ===> 14
1316 TARG = 1
1317 FLAGS = (SCALAR,KIDS)
1318 {
1319 TYPE = null ===> (12)
1320 (was rv2sv)
1321 FLAGS = (SCALAR,KIDS)
1322 {
1323 11 TYPE = gvsv ===> 12
1324 FLAGS = (SCALAR)
1325 GV = main::b
1326 }
1327 }
1328
1329< finish this later >
1330
1331=head2 Patching
1332
1333All right, we've now had a look at how to navigate the Perl sources and
1334some things you'll need to know when fiddling with them. Let's now get
1335on and create a simple patch. Here's something Larry suggested: if a
1336C<U> is the first active format during a C<pack>, (for example,
1337C<pack "U3C8", @stuff>) then the resulting string should be treated as
1338UTF8 encoded.
1339
1340How do we prepare to fix this up? First we locate the code in question -
1341the C<pack> happens at runtime, so it's going to be in one of the F<pp>
1342files. Sure enough, C<pp_pack> is in F<pp.c>. Since we're going to be
1343altering this file, let's copy it to F<pp.c~>.
1344
1345Now let's look over C<pp_pack>: we take a pattern into C<pat>, and then
1346loop over the pattern, taking each format character in turn into
1347C<datum_type>. Then for each possible format character, we swallow up
1348the other arguments in the pattern (a field width, an asterisk, and so
1349on) and convert the next chunk input into the specified format, adding
1350it onto the output SV C<cat>.
1351
1352How do we know if the C<U> is the first format in the C<pat>? Well, if
1353we have a pointer to the start of C<pat> then, if we see a C<U> we can
1354test whether we're still at the start of the string. So, here's where
1355C<pat> is set up:
1356
1357 STRLEN fromlen;
1358 register char *pat = SvPVx(*++MARK, fromlen);
1359 register char *patend = pat + fromlen;
1360 register I32 len;
1361 I32 datumtype;
1362 SV *fromstr;
1363
1364We'll have another string pointer in there:
1365
1366 STRLEN fromlen;
1367 register char *pat = SvPVx(*++MARK, fromlen);
1368 register char *patend = pat + fromlen;
1369 + char *patcopy;
1370 register I32 len;
1371 I32 datumtype;
1372 SV *fromstr;
1373
1374And just before we start the loop, we'll set C<patcopy> to be the start
1375of C<pat>:
1376
1377 items = SP - MARK;
1378 MARK++;
1379 sv_setpvn(cat, "", 0);
1380 + patcopy = pat;
1381 while (pat < patend) {
1382
1383Now if we see a C<U> which was at the start of the string, we turn on
1384the UTF8 flag for the output SV, C<cat>:
1385
1386 + if (datumtype == 'U' && pat==patcopy+1)
1387 + SvUTF8_on(cat);
1388 if (datumtype == '#') {
1389 while (pat < patend && *pat != '\n')
1390 pat++;
1391
1392Remember that it has to be C<patcopy+1> because the first character of
1393the string is the C<U> which has been swallowed into C<datumtype!>
1394
1395Oops, we forgot one thing: what if there are spaces at the start of the
1396pattern? C<pack(" U*", @stuff)> will have C<U> as the first active
1397character, even though it's not the first thing in the pattern. In this
1398case, we have to advance C<patcopy> along with C<pat> when we see spaces:
1399
1400 if (isSPACE(datumtype))
1401 continue;
1402
1403needs to become
1404
1405 if (isSPACE(datumtype)) {
1406 patcopy++;
1407 continue;
1408 }
1409
1410OK. That's the C part done. Now we must do two additional things before
1411this patch is ready to go: we've changed the behaviour of Perl, and so
1412we must document that change. We must also provide some more regression
1413tests to make sure our patch works and doesn't create a bug somewhere
1414else along the line.
1415
1416The regression tests for each operator live in F<t/op/>, and so we make
1417a copy of F<t/op/pack.t> to F<t/op/pack.t~>. Now we can add our tests
1418to the end. First, we'll test that the C<U> does indeed create Unicode
1419strings:
1420
1421 print 'not ' unless "1.20.300.4000" eq sprintf "%vd", pack("U*",1,20,300,4000);
1422 print "ok $test\n"; $test++;
1423
1424Now we'll test that we got that space-at-the-beginning business right:
1425
1426 print 'not ' unless "1.20.300.4000" eq
1427 sprintf "%vd", pack(" U*",1,20,300,4000);
1428 print "ok $test\n"; $test++;
1429
1430And finally we'll test that we don't make Unicode strings if C<U> is B<not>
1431the first active format:
1432
1433 print 'not ' unless v1.20.300.4000 ne
1434 sprintf "%vd", pack("C0U*",1,20,300,4000);
1435 print "ok $test\n"; $test++;
1436
b1866b2d 1437Mustn't forget to change the number of tests which appears at the top, or
a422fd2d
SC
1438else the automated tester will get confused:
1439
1440 -print "1..156\n";
1441 +print "1..159\n";
1442
1443We now compile up Perl, and run it through the test suite. Our new
1444tests pass, hooray!
1445
1446Finally, the documentation. The job is never done until the paperwork is
1447over, so let's describe the change we've just made. The relevant place
1448is F<pod/perlfunc.pod>; again, we make a copy, and then we'll insert
1449this text in the description of C<pack>:
1450
1451 =item *
1452
1453 If the pattern begins with a C<U>, the resulting string will be treated
1454 as Unicode-encoded. You can force UTF8 encoding on in a string with an
1455 initial C<U0>, and the bytes that follow will be interpreted as Unicode
1456 characters. If you don't want this to happen, you can begin your pattern
1457 with C<C0> (or anything else) to force Perl not to UTF8 encode your
1458 string, and then follow this with a C<U*> somewhere in your pattern.
1459
1460All done. Now let's create the patch. F<Porting/patching.pod> tells us
1461that if we're making major changes, we should copy the entire directory
1462to somewhere safe before we begin fiddling, and then do
13a2d996 1463
a422fd2d
SC
1464 diff -ruN old new > patch
1465
1466However, we know which files we've changed, and we can simply do this:
1467
1468 diff -u pp.c~ pp.c > patch
1469 diff -u t/op/pack.t~ t/op/pack.t >> patch
1470 diff -u pod/perlfunc.pod~ pod/perlfunc.pod >> patch
1471
1472We end up with a patch looking a little like this:
1473
1474 --- pp.c~ Fri Jun 02 04:34:10 2000
1475 +++ pp.c Fri Jun 16 11:37:25 2000
1476 @@ -4375,6 +4375,7 @@
1477 register I32 items;
1478 STRLEN fromlen;
1479 register char *pat = SvPVx(*++MARK, fromlen);
1480 + char *patcopy;
1481 register char *patend = pat + fromlen;
1482 register I32 len;
1483 I32 datumtype;
1484 @@ -4405,6 +4406,7 @@
1485 ...
1486
1487And finally, we submit it, with our rationale, to perl5-porters. Job
1488done!
1489
902b9dbf
MF
1490=head1 EXTERNAL TOOLS FOR DEBUGGING PERL
1491
1492Sometimes it helps to use external tools while debugging and
1493testing Perl. This section tries to guide you through using
1494some common testing and debugging tools with Perl. This is
1495meant as a guide to interfacing these tools with Perl, not
1496as any kind of guide to the use of the tools themselves.
1497
1498=head2 Rational Software's Purify
1499
1500Purify is a commercial tool that is helpful in identifying
1501memory overruns, wild pointers, memory leaks and other such
1502badness. Perl must be compiled in a specific way for
1503optimal testing with Purify. Purify is available under
1504Windows NT, Solaris, HP-UX, SGI, and Siemens Unix.
1505
1506The only currently known leaks happen when there are
1507compile-time errors within eval or require. (Fixing these
1508is non-trivial, unfortunately, but they must be fixed
1509eventually.)
1510
1511=head2 Purify on Unix
1512
1513On Unix, Purify creates a new Perl binary. To get the most
1514benefit out of Purify, you should create the perl to Purify
1515using:
1516
1517 sh Configure -Accflags=-DPURIFY -Doptimize='-g' \
1518 -Uusemymalloc -Dusemultiplicity
1519
1520where these arguments mean:
1521
1522=over 4
1523
1524=item -Accflags=-DPURIFY
1525
1526Disables Perl's arena memory allocation functions, as well as
1527forcing use of memory allocation functions derived from the
1528system malloc.
1529
1530=item -Doptimize='-g'
1531
1532Adds debugging information so that you see the exact source
1533statements where the problem occurs. Without this flag, all
1534you will see is the source filename of where the error occurred.
1535
1536=item -Uusemymalloc
1537
1538Disable Perl's malloc so that Purify can more closely monitor
1539allocations and leaks. Using Perl's malloc will make Purify
1540report most leaks in the "potential" leaks category.
1541
1542=item -Dusemultiplicity
1543
1544Enabling the multiplicity option allows perl to clean up
1545thoroughly when the interpreter shuts down, which reduces the
1546number of bogus leak reports from Purify.
1547
1548=back
1549
1550Once you've compiled a perl suitable for Purify'ing, then you
1551can just:
1552
1553 make pureperl
1554
1555which creates a binary named 'pureperl' that has been Purify'ed.
1556This binary is used in place of the standard 'perl' binary
1557when you want to debug Perl memory problems.
1558
1559As an example, to show any memory leaks produced during the
1560standard Perl testset you would create and run the Purify'ed
1561perl as:
1562
1563 make pureperl
1564 cd t
1565 ../pureperl -I../lib harness
1566
1567which would run Perl on test.pl and report any memory problems.
1568
1569Purify outputs messages in "Viewer" windows by default. If
1570you don't have a windowing environment or if you simply
1571want the Purify output to unobtrusively go to a log file
1572instead of to the interactive window, use these following
1573options to output to the log file "perl.log":
1574
1575 setenv PURIFYOPTIONS "-chain-length=25 -windows=no \
1576 -log-file=perl.log -append-logfile=yes"
1577
1578If you plan to use the "Viewer" windows, then you only need this option:
1579
1580 setenv PURIFYOPTIONS "-chain-length=25"
1581
1582=head2 Purify on NT
1583
1584Purify on Windows NT instruments the Perl binary 'perl.exe'
1585on the fly. There are several options in the makefile you
1586should change to get the most use out of Purify:
1587
1588=over 4
1589
1590=item DEFINES
1591
1592You should add -DPURIFY to the DEFINES line so the DEFINES
1593line looks something like:
1594
1595 DEFINES = -DWIN32 -D_CONSOLE -DNO_STRICT $(CRYPT_FLAG) -DPURIFY=1
1596
1597to disable Perl's arena memory allocation functions, as
1598well as to force use of memory allocation functions derived
1599from the system malloc.
1600
1601=item USE_MULTI = define
1602
1603Enabling the multiplicity option allows perl to clean up
1604thoroughly when the interpreter shuts down, which reduces the
1605number of bogus leak reports from Purify.
1606
1607=item #PERL_MALLOC = define
1608
1609Disable Perl's malloc so that Purify can more closely monitor
1610allocations and leaks. Using Perl's malloc will make Purify
1611report most leaks in the "potential" leaks category.
1612
1613=item CFG = Debug
1614
1615Adds debugging information so that you see the exact source
1616statements where the problem occurs. Without this flag, all
1617you will see is the source filename of where the error occurred.
1618
1619=back
1620
1621As an example, to show any memory leaks produced during the
1622standard Perl testset you would create and run Purify as:
1623
1624 cd win32
1625 make
1626 cd ../t
1627 purify ../perl -I../lib harness
1628
1629which would instrument Perl in memory, run Perl on test.pl,
1630then finally report any memory problems.
1631
a422fd2d
SC
1632=head2 CONCLUSION
1633
1634We've had a brief look around the Perl source, an overview of the stages
1635F<perl> goes through when it's running your code, and how to use a
902b9dbf
MF
1636debugger to poke at the Perl guts. We took a very simple problem and
1637demonstrated how to solve it fully - with documentation, regression
1638tests, and finally a patch for submission to p5p. Finally, we talked
1639about how to use external tools to debug and test Perl.
a422fd2d
SC
1640
1641I'd now suggest you read over those references again, and then, as soon
1642as possible, get your hands dirty. The best way to learn is by doing,
1643so:
1644
1645=over 3
1646
1647=item *
1648
1649Subscribe to perl5-porters, follow the patches and try and understand
1650them; don't be afraid to ask if there's a portion you're not clear on -
1651who knows, you may unearth a bug in the patch...
1652
1653=item *
1654
1655Keep up to date with the bleeding edge Perl distributions and get
1656familiar with the changes. Try and get an idea of what areas people are
1657working on and the changes they're making.
1658
1659=item *
1660
a1f349fd
MB
1661Do read the README associated with your operating system, i.e. README.aix
1662on the IBM AIX OS. Don't hesitate to supply patches to that README if
1663you find anything missing or changed over a new OS release.
1664
1665=item *
1666
a422fd2d
SC
1667Find an area of Perl that seems interesting to you, and see if you can
1668work out how it works. Scan through the source, and step over it in the
1669debugger. Play, poke, investigate, fiddle! You'll probably get to
1670understand not just your chosen area but a much wider range of F<perl>'s
1671activity as well, and probably sooner than you'd think.
1672
1673=back
1674
1675=over 3
1676
1677=item I<The Road goes ever on and on, down from the door where it began.>
1678
1679=back
1680
1681If you can do these things, you've started on the long road to Perl porting.
1682Thanks for wanting to help make Perl better - and happy hacking!
1683
e8cd7eae
GS
1684=head1 AUTHOR
1685
1686This document was written by Nathan Torkington, and is maintained by
1687the perl5-porters mailing list.
1688