This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Re: [PATCH] Added wisdom to perlhack.pod
[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
3e148164
JH
221Keeping in sync with the most recent branch can be done in several ways,
222but the most convenient and reliable way is using B<rsync>, available at
223ftp://rsync.samba.org/pub/rsync/ . (You can also get the most recent
224branch by FTP.)
a1f349fd
MB
225
226If you choose to keep in sync using rsync, there are two approaches
3e148164 227to doing so:
a1f349fd
MB
228
229=over 4
230
231=item rsync'ing the source tree
232
3e148164 233Presuming you are in the directory where your perl source resides
a1f349fd
MB
234and you have rsync installed and available, you can `upgrade' to
235the bleadperl using:
236
237 # rsync -avz rsync://ftp.linux.activestate.com/perl-current/ .
238
239This takes care of updating every single item in the source tree to
240the latest applied patch level, creating files that are new (to your
241distribution) and setting date/time stamps of existing files to
242reflect the bleadperl status.
243
c6d0653e
LC
244Note that this will not delete any files that were in '.' before
245the rsync. Once you are sure that the rsync is running correctly,
246run it with the --delete and the --dry-run options like this:
247
248 # rsync -avz --delete --dry-run rsync://ftp.linux.activestate.com/perl-current/ .
249
250This will I<simulate> an rsync run that also deletes files not
251present in the bleadperl master copy. Observe the results from
252this run closely. If you are sure that the actual run would delete
253no files precious to you, you could remove the '--dry-run' option.
254
a1f349fd
MB
255You can than check what patch was the latest that was applied by
256looking in the file B<.patch>, which will show the number of the
257latest patch.
258
259If you have more than one machine to keep in sync, and not all of
260them have access to the WAN (so you are not able to rsync all the
261source trees to the real source), there are some ways to get around
262this problem.
263
264=over 4
265
266=item Using rsync over the LAN
267
268Set up a local rsync server which makes the rsynced source tree
3e148164 269available to the LAN and sync the other machines against this
a1f349fd
MB
270directory.
271
272From http://rsync.samba.org/README.html:
273
274 "Rsync uses rsh or ssh for communication. It does not need to be
275 setuid and requires no special privileges for installation. It
3958b146 276 does not require an inetd entry or a daemon. You must, however,
a1f349fd
MB
277 have a working rsh or ssh system. Using ssh is recommended for
278 its security features."
279
280=item Using pushing over the NFS
281
282Having the other systems mounted over the NFS, you can take an
3e148164
JH
283active pushing approach by checking the just updated tree against
284the other not-yet synced trees. An example would be
285
286 #!/usr/bin/perl -w
287
288 use strict;
289 use File::Copy;
290
291 my %MF = map {
292 m/(\S+)/;
293 $1 => [ (stat $1)[2, 7, 9] ]; # mode, size, mtime
294 } `cat MANIFEST`;
295
296 my %remote = map { $_ => "/$_/pro/3gl/CPAN/perl-5.7.1" } qw(host1 host2);
297
298 foreach my $host (keys %remote) {
299 unless (-d $remote{$host}) {
300 print STDERR "Cannot Xsync for host $host\n";
301 next;
302 }
303 foreach my $file (keys %MF) {
304 my $rfile = "$remote{$host}/$file";
305 my ($mode, $size, $mtime) = (stat $rfile)[2, 7, 9];
306 defined $size or ($mode, $size, $mtime) = (0, 0, 0);
307 $size == $MF{$file}[1] && $mtime == $MF{$file}[2] and next;
308 printf "%4s %-34s %8d %9d %8d %9d\n",
309 $host, $file, $MF{$file}[1], $MF{$file}[2], $size, $mtime;
310 unlink $rfile;
311 copy ($file, $rfile);
312 utime time, $MF{$file}[2], $rfile;
313 chmod $MF{$file}[0], $rfile;
314 }
315 }
316
317though this is not perfect. It could be improved with checking
a1f349fd
MB
318file checksums before updating. Not all NFS systems support
319reliable utime support (when used over the NFS).
320
321=back
322
323=item rsync'ing the patches
324
325The source tree is maintained by the pumpking who applies patches to
326the files in the tree. These patches are either created by the
327pumpking himself using C<diff -c> after updating the file manually or
328by applying patches sent in by posters on the perl5-porters list.
329These patches are also saved and rsync'able, so you can apply them
330yourself to the source files.
331
332Presuming you are in a directory where your patches reside, you can
3e148164 333get them in sync with
a1f349fd
MB
334
335 # rsync -avz rsync://ftp.linux.activestate.com/perl-current-diffs/ .
336
337This makes sure the latest available patch is downloaded to your
338patch directory.
339
3e148164 340It's then up to you to apply these patches, using something like
a1f349fd
MB
341
342 # last=`ls -rt1 *.gz | tail -1`
343 # rsync -avz rsync://ftp.linux.activestate.com/perl-current-diffs/ .
344 # find . -name '*.gz' -newer $last -exec gzcat {} \; >blead.patch
345 # cd ../perl-current
346 # patch -p1 -N <../perl-current-diffs/blead.patch
347
348or, since this is only a hint towards how it works, use CPAN-patchaperl
349from Andreas K├Ânig to have better control over the patching process.
350
351=back
352
f7e1e956 353=head2 Why rsync the source tree
a1f349fd
MB
354
355=over 4
356
10f58044 357=item It's easier to rsync the source tree
a1f349fd
MB
358
359Since you don't have to apply the patches yourself, you are sure all
360files in the source tree are in the right state.
361
362=item It's more recent
363
364According to Gurusamy Sarathy:
365
366 "... The rsync mirror is automatic and syncs with the repository
367 every five minutes.
368
3e148164 369 "Updating the patch area still requires manual intervention
a1f349fd
MB
370 (with all the goofiness that implies, which you've noted) and
371 is typically on a daily cycle. Making this process automatic
372 is on my tuit list, but don't ask me when."
373
374=item It's more reliable
375
3e148164 376Well, since the patches are updated by hand, I don't have to say any
a1f349fd
MB
377more ... (see Sarathy's remark).
378
379=back
380
f7e1e956 381=head2 Why rsync the patches
a1f349fd
MB
382
383=over 4
384
10f58044 385=item It's easier to rsync the patches
a1f349fd
MB
386
387If you have more than one machine that you want to keep in track with
3e148164 388bleadperl, it's easier to rsync the patches only once and then apply
a1f349fd
MB
389them to all the source trees on the different machines.
390
391In case you try to keep in pace on 5 different machines, for which
392only one of them has access to the WAN, rsync'ing all the source
3e148164 393trees should than be done 5 times over the NFS. Having
a1f349fd 394rsync'ed the patches only once, I can apply them to all the source
3e148164 395trees automatically. Need you say more ;-)
a1f349fd
MB
396
397=item It's a good reference
398
399If you do not only like to have the most recent development branch,
400but also like to B<fix> bugs, or extend features, you want to dive
401into the sources. If you are a seasoned perl core diver, you don't
402need no manuals, tips, roadmaps, perlguts.pod or other aids to find
403your way around. But if you are a starter, the patches may help you
404in finding where you should start and how to change the bits that
405bug you.
406
407The file B<Changes> is updated on occasions the pumpking sees as his
408own little sync points. On those occasions, he releases a tar-ball of
409the current source tree (i.e. perl@7582.tar.gz), which will be an
410excellent point to start with when choosing to use the 'rsync the
411patches' scheme. Starting with perl@7582, which means a set of source
412files on which the latest applied patch is number 7582, you apply all
f18956b7 413succeeding patches available from then on (7583, 7584, ...).
a1f349fd
MB
414
415You can use the patches later as a kind of search archive.
416
417=over 4
418
419=item Finding a start point
420
421If you want to fix/change the behaviour of function/feature Foo, just
422scan the patches for patches that mention Foo either in the subject,
3e148164 423the comments, or the body of the fix. A good chance the patch shows
a1f349fd
MB
424you the files that are affected by that patch which are very likely
425to be the starting point of your journey into the guts of perl.
426
427=item Finding how to fix a bug
428
429If you've found I<where> the function/feature Foo misbehaves, but you
430don't know how to fix it (but you do know the change you want to
431make), you can, again, peruse the patches for similar changes and
432look how others apply the fix.
433
434=item Finding the source of misbehaviour
435
436When you keep in sync with bleadperl, the pumpking would love to
3958b146 437I<see> that the community efforts really work. So after each of his
a1f349fd
MB
438sync points, you are to 'make test' to check if everything is still
439in working order. If it is, you do 'make ok', which will send an OK
440report to perlbug@perl.org. (If you do not have access to a mailer
3e148164 441from the system you just finished successfully 'make test', you can
a1f349fd
MB
442do 'make okfile', which creates the file C<perl.ok>, which you can
443than take to your favourite mailer and mail yourself).
444
3958b146 445But of course, as always, things will not always lead to a success
a1f349fd
MB
446path, and one or more test do not pass the 'make test'. Before
447sending in a bug report (using 'make nok' or 'make nokfile'), check
448the mailing list if someone else has reported the bug already and if
449so, confirm it by replying to that message. If not, you might want to
450trace the source of that misbehaviour B<before> sending in the bug,
451which will help all the other porters in finding the solution.
452
3e148164
JH
453Here the saved patches come in very handy. You can check the list of
454patches to see which patch changed what file and what change caused
455the misbehaviour. If you note that in the bug report, it saves the
456one trying to solve it, looking for that point.
a1f349fd
MB
457
458=back
459
460If searching the patches is too bothersome, you might consider using
461perl's bugtron to find more information about discussions and
462ramblings on posted bugs.
463
464=back
465
3e148164
JH
466If you want to get the best of both worlds, rsync both the source
467tree for convenience, reliability and ease and rsync the patches
468for reference.
469
a1f349fd
MB
470=head2 Submitting patches
471
f7e1e956
MS
472Always submit patches to I<perl5-porters@perl.org>. If you're
473patching a core module and there's an author listed, send the author a
474copy (see L<Patching a core module>). This lets other porters review
475your patch, which catches a surprising number of errors in patches.
476Either use the diff program (available in source code form from
477I<ftp://ftp.gnu.org/pub/gnu/>), or use Johan Vromans' I<makepatch>
478(available from I<CPAN/authors/id/JV/>). Unified diffs are preferred,
479but context diffs are accepted. Do not send RCS-style diffs or diffs
480without context lines. More information is given in the
481I<Porting/patching.pod> file in the Perl source distribution. Please
482patch against the latest B<development> version (e.g., if you're
483fixing a bug in the 5.005 track, patch against the latest 5.005_5x
484version). Only patches that survive the heat of the development
485branch get applied to maintenance versions.
486
487Your patch should update the documentation and test suite. See
488L<Writing a test>.
e8cd7eae
GS
489
490To report a bug in Perl, use the program I<perlbug> which comes with
491Perl (if you can't get Perl to work, send mail to the address
f18956b7 492I<perlbug@perl.org> or I<perlbug@perl.com>). Reporting bugs through
e8cd7eae
GS
493I<perlbug> feeds into the automated bug-tracking system, access to
494which is provided through the web at I<http://bugs.perl.org/>. It
495often pays to check the archives of the perl5-porters mailing list to
496see whether the bug you're reporting has been reported before, and if
497so whether it was considered a bug. See above for the location of
498the searchable archives.
499
500The CPAN testers (I<http://testers.cpan.org/>) are a group of
501volunteers who test CPAN modules on a variety of platforms. Perl Labs
f6c51b38
GS
502(I<http://labs.perl.org/>) automatically tests Perl source releases on
503platforms and gives feedback to the CPAN testers mailing list. Both
504efforts welcome volunteers.
e8cd7eae 505
e8cd7eae
GS
506It's a good idea to read and lurk for a while before chipping in.
507That way you'll get to see the dynamic of the conversations, learn the
508personalities of the players, and hopefully be better prepared to make
509a useful contribution when do you speak up.
510
511If after all this you still think you want to join the perl5-porters
f6c51b38
GS
512mailing list, send mail to I<perl5-porters-subscribe@perl.org>. To
513unsubscribe, send mail to I<perl5-porters-unsubscribe@perl.org>.
e8cd7eae 514
a422fd2d
SC
515To hack on the Perl guts, you'll need to read the following things:
516
517=over 3
518
519=item L<perlguts>
520
521This is of paramount importance, since it's the documentation of what
522goes where in the Perl source. Read it over a couple of times and it
523might start to make sense - don't worry if it doesn't yet, because the
524best way to study it is to read it in conjunction with poking at Perl
525source, and we'll do that later on.
526
527You might also want to look at Gisle Aas's illustrated perlguts -
528there's no guarantee that this will be absolutely up-to-date with the
529latest documentation in the Perl core, but the fundamentals will be
530right. (http://gisle.aas.no/perl/illguts/)
531
532=item L<perlxstut> and L<perlxs>
533
534A working knowledge of XSUB programming is incredibly useful for core
535hacking; XSUBs use techniques drawn from the PP code, the portion of the
536guts that actually executes a Perl program. It's a lot gentler to learn
537those techniques from simple examples and explanation than from the core
538itself.
539
540=item L<perlapi>
541
542The documentation for the Perl API explains what some of the internal
543functions do, as well as the many macros used in the source.
544
545=item F<Porting/pumpkin.pod>
546
547This is a collection of words of wisdom for a Perl porter; some of it is
548only useful to the pumpkin holder, but most of it applies to anyone
549wanting to go about Perl development.
550
551=item The perl5-porters FAQ
552
553This is posted to perl5-porters at the beginning on every month, and
554should be available from http://perlhacker.org/p5p-faq; alternatively,
555you can get the FAQ emailed to you by sending mail to
556C<perl5-porters-faq@perl.org>. It contains hints on reading
557perl5-porters, information on how perl5-porters works and how Perl
558development in general works.
559
560=back
561
562=head2 Finding Your Way Around
563
564Perl maintenance can be split into a number of areas, and certain people
565(pumpkins) will have responsibility for each area. These areas sometimes
566correspond to files or directories in the source kit. Among the areas are:
567
568=over 3
569
570=item Core modules
571
572Modules shipped as part of the Perl core live in the F<lib/> and F<ext/>
573subdirectories: F<lib/> is for the pure-Perl modules, and F<ext/>
574contains the core XS modules.
575
f7e1e956
MS
576=item Tests
577
578There are tests for nearly all the modules, built-ins and major bits
579of functionality. Test files all have a .t suffix. Module tests live
580in the F<lib/> and F<ext/> directories next to the module being
581tested. Others live in F<t/>. See L<Writing a test>
582
a422fd2d
SC
583=item Documentation
584
585Documentation maintenance includes looking after everything in the
586F<pod/> directory, (as well as contributing new documentation) and
587the documentation to the modules in core.
588
589=item Configure
590
591The configure process is the way we make Perl portable across the
592myriad of operating systems it supports. Responsibility for the
593configure, build and installation process, as well as the overall
594portability of the core code rests with the configure pumpkin - others
595help out with individual operating systems.
596
597The files involved are the operating system directories, (F<win32/>,
598F<os2/>, F<vms/> and so on) the shell scripts which generate F<config.h>
599and F<Makefile>, as well as the metaconfig files which generate
600F<Configure>. (metaconfig isn't included in the core distribution.)
601
602=item Interpreter
603
604And of course, there's the core of the Perl interpreter itself. Let's
605have a look at that in a little more detail.
606
607=back
608
609Before we leave looking at the layout, though, don't forget that
610F<MANIFEST> contains not only the file names in the Perl distribution,
611but short descriptions of what's in them, too. For an overview of the
612important files, try this:
613
614 perl -lne 'print if /^[^\/]+\.[ch]\s+/' MANIFEST
615
616=head2 Elements of the interpreter
617
618The work of the interpreter has two main stages: compiling the code
619into the internal representation, or bytecode, and then executing it.
620L<perlguts/Compiled code> explains exactly how the compilation stage
621happens.
622
623Here is a short breakdown of perl's operation:
624
625=over 3
626
627=item Startup
628
629The action begins in F<perlmain.c>. (or F<miniperlmain.c> for miniperl)
630This is very high-level code, enough to fit on a single screen, and it
631resembles the code found in L<perlembed>; most of the real action takes
632place in F<perl.c>
633
634First, F<perlmain.c> allocates some memory and constructs a Perl
635interpreter:
636
637 1 PERL_SYS_INIT3(&argc,&argv,&env);
638 2
639 3 if (!PL_do_undump) {
640 4 my_perl = perl_alloc();
641 5 if (!my_perl)
642 6 exit(1);
643 7 perl_construct(my_perl);
644 8 PL_perl_destruct_level = 0;
645 9 }
646
647Line 1 is a macro, and its definition is dependent on your operating
648system. Line 3 references C<PL_do_undump>, a global variable - all
649global variables in Perl start with C<PL_>. This tells you whether the
650current running program was created with the C<-u> flag to perl and then
651F<undump>, which means it's going to be false in any sane context.
652
653Line 4 calls a function in F<perl.c> to allocate memory for a Perl
654interpreter. It's quite a simple function, and the guts of it looks like
655this:
656
657 my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));
658
659Here you see an example of Perl's system abstraction, which we'll see
660later: C<PerlMem_malloc> is either your system's C<malloc>, or Perl's
661own C<malloc> as defined in F<malloc.c> if you selected that option at
662configure time.
663
664Next, in line 7, we construct the interpreter; this sets up all the
665special variables that Perl needs, the stacks, and so on.
666
667Now we pass Perl the command line options, and tell it to go:
668
669 exitstatus = perl_parse(my_perl, xs_init, argc, argv, (char **)NULL);
670 if (!exitstatus) {
671 exitstatus = perl_run(my_perl);
672 }
673
674
675C<perl_parse> is actually a wrapper around C<S_parse_body>, as defined
676in F<perl.c>, which processes the command line options, sets up any
677statically linked XS modules, opens the program and calls C<yyparse> to
678parse it.
679
680=item Parsing
681
682The aim of this stage is to take the Perl source, and turn it into an op
683tree. We'll see what one of those looks like later. Strictly speaking,
684there's three things going on here.
685
686C<yyparse>, the parser, lives in F<perly.c>, although you're better off
687reading the original YACC input in F<perly.y>. (Yes, Virginia, there
688B<is> a YACC grammar for Perl!) The job of the parser is to take your
689code and `understand' it, splitting it into sentences, deciding which
690operands go with which operators and so on.
691
692The parser is nobly assisted by the lexer, which chunks up your input
693into tokens, and decides what type of thing each token is: a variable
694name, an operator, a bareword, a subroutine, a core function, and so on.
695The main point of entry to the lexer is C<yylex>, and that and its
696associated routines can be found in F<toke.c>. Perl isn't much like
697other computer languages; it's highly context sensitive at times, it can
698be tricky to work out what sort of token something is, or where a token
699ends. As such, there's a lot of interplay between the tokeniser and the
700parser, which can get pretty frightening if you're not used to it.
701
702As the parser understands a Perl program, it builds up a tree of
703operations for the interpreter to perform during execution. The routines
704which construct and link together the various operations are to be found
705in F<op.c>, and will be examined later.
706
707=item Optimization
708
709Now the parsing stage is complete, and the finished tree represents
710the operations that the Perl interpreter needs to perform to execute our
711program. Next, Perl does a dry run over the tree looking for
712optimisations: constant expressions such as C<3 + 4> will be computed
713now, and the optimizer will also see if any multiple operations can be
714replaced with a single one. For instance, to fetch the variable C<$foo>,
715instead of grabbing the glob C<*foo> and looking at the scalar
716component, the optimizer fiddles the op tree to use a function which
717directly looks up the scalar in question. The main optimizer is C<peep>
718in F<op.c>, and many ops have their own optimizing functions.
719
720=item Running
721
722Now we're finally ready to go: we have compiled Perl byte code, and all
723that's left to do is run it. The actual execution is done by the
724C<runops_standard> function in F<run.c>; more specifically, it's done by
725these three innocent looking lines:
726
727 while ((PL_op = CALL_FPTR(PL_op->op_ppaddr)(aTHX))) {
728 PERL_ASYNC_CHECK();
729 }
730
731You may be more comfortable with the Perl version of that:
732
733 PERL_ASYNC_CHECK() while $Perl::op = &{$Perl::op->{function}};
734
735Well, maybe not. Anyway, each op contains a function pointer, which
736stipulates the function which will actually carry out the operation.
737This function will return the next op in the sequence - this allows for
738things like C<if> which choose the next op dynamically at run time.
739The C<PERL_ASYNC_CHECK> makes sure that things like signals interrupt
740execution if required.
741
742The actual functions called are known as PP code, and they're spread
743between four files: F<pp_hot.c> contains the `hot' code, which is most
744often used and highly optimized, F<pp_sys.c> contains all the
745system-specific functions, F<pp_ctl.c> contains the functions which
746implement control structures (C<if>, C<while> and the like) and F<pp.c>
747contains everything else. These are, if you like, the C code for Perl's
748built-in functions and operators.
749
750=back
751
752=head2 Internal Variable Types
753
754You should by now have had a look at L<perlguts>, which tells you about
755Perl's internal variable types: SVs, HVs, AVs and the rest. If not, do
756that now.
757
758These variables are used not only to represent Perl-space variables, but
759also any constants in the code, as well as some structures completely
760internal to Perl. The symbol table, for instance, is an ordinary Perl
761hash. Your code is represented by an SV as it's read into the parser;
762any program files you call are opened via ordinary Perl filehandles, and
763so on.
764
765The core L<Devel::Peek|Devel::Peek> module lets us examine SVs from a
766Perl program. Let's see, for instance, how Perl treats the constant
767C<"hello">.
768
769 % perl -MDevel::Peek -e 'Dump("hello")'
770 1 SV = PV(0xa041450) at 0xa04ecbc
771 2 REFCNT = 1
772 3 FLAGS = (POK,READONLY,pPOK)
773 4 PV = 0xa0484e0 "hello"\0
774 5 CUR = 5
775 6 LEN = 6
776
777Reading C<Devel::Peek> output takes a bit of practise, so let's go
778through it line by line.
779
780Line 1 tells us we're looking at an SV which lives at C<0xa04ecbc> in
781memory. SVs themselves are very simple structures, but they contain a
782pointer to a more complex structure. In this case, it's a PV, a
783structure which holds a string value, at location C<0xa041450>. Line 2
784is the reference count; there are no other references to this data, so
785it's 1.
786
787Line 3 are the flags for this SV - it's OK to use it as a PV, it's a
788read-only SV (because it's a constant) and the data is a PV internally.
789Next we've got the contents of the string, starting at location
790C<0xa0484e0>.
791
792Line 5 gives us the current length of the string - note that this does
793B<not> include the null terminator. Line 6 is not the length of the
794string, but the length of the currently allocated buffer; as the string
795grows, Perl automatically extends the available storage via a routine
796called C<SvGROW>.
797
798You can get at any of these quantities from C very easily; just add
799C<Sv> to the name of the field shown in the snippet, and you've got a
800macro which will return the value: C<SvCUR(sv)> returns the current
801length of the string, C<SvREFCOUNT(sv)> returns the reference count,
802C<SvPV(sv, len)> returns the string itself with its length, and so on.
803More macros to manipulate these properties can be found in L<perlguts>.
804
805Let's take an example of manipulating a PV, from C<sv_catpvn>, in F<sv.c>
806
807 1 void
808 2 Perl_sv_catpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
809 3 {
810 4 STRLEN tlen;
811 5 char *junk;
812
813 6 junk = SvPV_force(sv, tlen);
814 7 SvGROW(sv, tlen + len + 1);
815 8 if (ptr == junk)
816 9 ptr = SvPVX(sv);
817 10 Move(ptr,SvPVX(sv)+tlen,len,char);
818 11 SvCUR(sv) += len;
819 12 *SvEND(sv) = '\0';
820 13 (void)SvPOK_only_UTF8(sv); /* validate pointer */
821 14 SvTAINT(sv);
822 15 }
823
824This is a function which adds a string, C<ptr>, of length C<len> onto
825the end of the PV stored in C<sv>. The first thing we do in line 6 is
826make sure that the SV B<has> a valid PV, by calling the C<SvPV_force>
827macro to force a PV. As a side effect, C<tlen> gets set to the current
828value of the PV, and the PV itself is returned to C<junk>.
829
b1866b2d 830In line 7, we make sure that the SV will have enough room to accommodate
a422fd2d
SC
831the old string, the new string and the null terminator. If C<LEN> isn't
832big enough, C<SvGROW> will reallocate space for us.
833
834Now, if C<junk> is the same as the string we're trying to add, we can
835grab the string directly from the SV; C<SvPVX> is the address of the PV
836in the SV.
837
838Line 10 does the actual catenation: the C<Move> macro moves a chunk of
839memory around: we move the string C<ptr> to the end of the PV - that's
840the start of the PV plus its current length. We're moving C<len> bytes
841of type C<char>. After doing so, we need to tell Perl we've extended the
842string, by altering C<CUR> to reflect the new length. C<SvEND> is a
843macro which gives us the end of the string, so that needs to be a
844C<"\0">.
845
846Line 13 manipulates the flags; since we've changed the PV, any IV or NV
847values will no longer be valid: if we have C<$a=10; $a.="6";> we don't
848want to use the old IV of 10. C<SvPOK_only_utf8> is a special UTF8-aware
849version of C<SvPOK_only>, a macro which turns off the IOK and NOK flags
850and turns on POK. The final C<SvTAINT> is a macro which launders tainted
851data if taint mode is turned on.
852
853AVs and HVs are more complicated, but SVs are by far the most common
854variable type being thrown around. Having seen something of how we
855manipulate these, let's go on and look at how the op tree is
856constructed.
857
858=head2 Op Trees
859
860First, what is the op tree, anyway? The op tree is the parsed
861representation of your program, as we saw in our section on parsing, and
862it's the sequence of operations that Perl goes through to execute your
863program, as we saw in L</Running>.
864
865An op is a fundamental operation that Perl can perform: all the built-in
866functions and operators are ops, and there are a series of ops which
867deal with concepts the interpreter needs internally - entering and
868leaving a block, ending a statement, fetching a variable, and so on.
869
870The op tree is connected in two ways: you can imagine that there are two
871"routes" through it, two orders in which you can traverse the tree.
872First, parse order reflects how the parser understood the code, and
873secondly, execution order tells perl what order to perform the
874operations in.
875
876The easiest way to examine the op tree is to stop Perl after it has
877finished parsing, and get it to dump out the tree. This is exactly what
878the compiler backends L<B::Terse|B::Terse> and L<B::Debug|B::Debug> do.
879
880Let's have a look at how Perl sees C<$a = $b + $c>:
881
882 % perl -MO=Terse -e '$a=$b+$c'
883 1 LISTOP (0x8179888) leave
884 2 OP (0x81798b0) enter
885 3 COP (0x8179850) nextstate
886 4 BINOP (0x8179828) sassign
887 5 BINOP (0x8179800) add [1]
888 6 UNOP (0x81796e0) null [15]
889 7 SVOP (0x80fafe0) gvsv GV (0x80fa4cc) *b
890 8 UNOP (0x81797e0) null [15]
891 9 SVOP (0x8179700) gvsv GV (0x80efeb0) *c
892 10 UNOP (0x816b4f0) null [15]
893 11 SVOP (0x816dcf0) gvsv GV (0x80fa460) *a
894
895Let's start in the middle, at line 4. This is a BINOP, a binary
896operator, which is at location C<0x8179828>. The specific operator in
897question is C<sassign> - scalar assignment - and you can find the code
898which implements it in the function C<pp_sassign> in F<pp_hot.c>. As a
899binary operator, it has two children: the add operator, providing the
900result of C<$b+$c>, is uppermost on line 5, and the left hand side is on
901line 10.
902
903Line 10 is the null op: this does exactly nothing. What is that doing
904there? If you see the null op, it's a sign that something has been
905optimized away after parsing. As we mentioned in L</Optimization>,
906the optimization stage sometimes converts two operations into one, for
907example when fetching a scalar variable. When this happens, instead of
908rewriting the op tree and cleaning up the dangling pointers, it's easier
909just to replace the redundant operation with the null op. Originally,
910the tree would have looked like this:
911
912 10 SVOP (0x816b4f0) rv2sv [15]
913 11 SVOP (0x816dcf0) gv GV (0x80fa460) *a
914
915That is, fetch the C<a> entry from the main symbol table, and then look
916at the scalar component of it: C<gvsv> (C<pp_gvsv> into F<pp_hot.c>)
917happens to do both these things.
918
919The right hand side, starting at line 5 is similar to what we've just
920seen: we have the C<add> op (C<pp_add> also in F<pp_hot.c>) add together
921two C<gvsv>s.
922
923Now, what's this about?
924
925 1 LISTOP (0x8179888) leave
926 2 OP (0x81798b0) enter
927 3 COP (0x8179850) nextstate
928
929C<enter> and C<leave> are scoping ops, and their job is to perform any
930housekeeping every time you enter and leave a block: lexical variables
931are tidied up, unreferenced variables are destroyed, and so on. Every
932program will have those first three lines: C<leave> is a list, and its
933children are all the statements in the block. Statements are delimited
934by C<nextstate>, so a block is a collection of C<nextstate> ops, with
935the ops to be performed for each statement being the children of
936C<nextstate>. C<enter> is a single op which functions as a marker.
937
938That's how Perl parsed the program, from top to bottom:
939
940 Program
941 |
942 Statement
943 |
944 =
945 / \
946 / \
947 $a +
948 / \
949 $b $c
950
951However, it's impossible to B<perform> the operations in this order:
952you have to find the values of C<$b> and C<$c> before you add them
953together, for instance. So, the other thread that runs through the op
954tree is the execution order: each op has a field C<op_next> which points
955to the next op to be run, so following these pointers tells us how perl
956executes the code. We can traverse the tree in this order using
957the C<exec> option to C<B::Terse>:
958
959 % perl -MO=Terse,exec -e '$a=$b+$c'
960 1 OP (0x8179928) enter
961 2 COP (0x81798c8) nextstate
962 3 SVOP (0x81796c8) gvsv GV (0x80fa4d4) *b
963 4 SVOP (0x8179798) gvsv GV (0x80efeb0) *c
964 5 BINOP (0x8179878) add [1]
965 6 SVOP (0x816dd38) gvsv GV (0x80fa468) *a
966 7 BINOP (0x81798a0) sassign
967 8 LISTOP (0x8179900) leave
968
969This probably makes more sense for a human: enter a block, start a
970statement. Get the values of C<$b> and C<$c>, and add them together.
971Find C<$a>, and assign one to the other. Then leave.
972
973The way Perl builds up these op trees in the parsing process can be
974unravelled by examining F<perly.y>, the YACC grammar. Let's take the
975piece we need to construct the tree for C<$a = $b + $c>
976
977 1 term : term ASSIGNOP term
978 2 { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
979 3 | term ADDOP term
980 4 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
981
982If you're not used to reading BNF grammars, this is how it works: You're
983fed certain things by the tokeniser, which generally end up in upper
984case. Here, C<ADDOP>, is provided when the tokeniser sees C<+> in your
985code. C<ASSIGNOP> is provided when C<=> is used for assigning. These are
986`terminal symbols', because you can't get any simpler than them.
987
988The grammar, lines one and three of the snippet above, tells you how to
989build up more complex forms. These complex forms, `non-terminal symbols'
990are generally placed in lower case. C<term> here is a non-terminal
991symbol, representing a single expression.
992
993The grammar gives you the following rule: you can make the thing on the
994left of the colon if you see all the things on the right in sequence.
995This is called a "reduction", and the aim of parsing is to completely
996reduce the input. There are several different ways you can perform a
997reduction, separated by vertical bars: so, C<term> followed by C<=>
998followed by C<term> makes a C<term>, and C<term> followed by C<+>
999followed by C<term> can also make a C<term>.
1000
1001So, if you see two terms with an C<=> or C<+>, between them, you can
1002turn them into a single expression. When you do this, you execute the
1003code in the block on the next line: if you see C<=>, you'll do the code
1004in line 2. If you see C<+>, you'll do the code in line 4. It's this code
1005which contributes to the op tree.
1006
1007 | term ADDOP term
1008 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
1009
1010What this does is creates a new binary op, and feeds it a number of
1011variables. The variables refer to the tokens: C<$1> is the first token in
1012the input, C<$2> the second, and so on - think regular expression
1013backreferences. C<$$> is the op returned from this reduction. So, we
1014call C<newBINOP> to create a new binary operator. The first parameter to
1015C<newBINOP>, a function in F<op.c>, is the op type. It's an addition
1016operator, so we want the type to be C<ADDOP>. We could specify this
1017directly, but it's right there as the second token in the input, so we
1018use C<$2>. The second parameter is the op's flags: 0 means `nothing
1019special'. Then the things to add: the left and right hand side of our
1020expression, in scalar context.
1021
1022=head2 Stacks
1023
1024When perl executes something like C<addop>, how does it pass on its
1025results to the next op? The answer is, through the use of stacks. Perl
1026has a number of stacks to store things it's currently working on, and
1027we'll look at the three most important ones here.
1028
1029=over 3
1030
1031=item Argument stack
1032
1033Arguments are passed to PP code and returned from PP code using the
1034argument stack, C<ST>. The typical way to handle arguments is to pop
1035them off the stack, deal with them how you wish, and then push the result
1036back onto the stack. This is how, for instance, the cosine operator
1037works:
1038
1039 NV value;
1040 value = POPn;
1041 value = Perl_cos(value);
1042 XPUSHn(value);
1043
1044We'll see a more tricky example of this when we consider Perl's macros
1045below. C<POPn> gives you the NV (floating point value) of the top SV on
1046the stack: the C<$x> in C<cos($x)>. Then we compute the cosine, and push
1047the result back as an NV. The C<X> in C<XPUSHn> means that the stack
1048should be extended if necessary - it can't be necessary here, because we
1049know there's room for one more item on the stack, since we've just
1050removed one! The C<XPUSH*> macros at least guarantee safety.
1051
1052Alternatively, you can fiddle with the stack directly: C<SP> gives you
1053the first element in your portion of the stack, and C<TOP*> gives you
1054the top SV/IV/NV/etc. on the stack. So, for instance, to do unary
1055negation of an integer:
1056
1057 SETi(-TOPi);
1058
1059Just set the integer value of the top stack entry to its negation.
1060
1061Argument stack manipulation in the core is exactly the same as it is in
1062XSUBs - see L<perlxstut>, L<perlxs> and L<perlguts> for a longer
1063description of the macros used in stack manipulation.
1064
1065=item Mark stack
1066
1067I say `your portion of the stack' above because PP code doesn't
1068necessarily get the whole stack to itself: if your function calls
1069another function, you'll only want to expose the arguments aimed for the
1070called function, and not (necessarily) let it get at your own data. The
1071way we do this is to have a `virtual' bottom-of-stack, exposed to each
1072function. The mark stack keeps bookmarks to locations in the argument
1073stack usable by each function. For instance, when dealing with a tied
1074variable, (internally, something with `P' magic) Perl has to call
1075methods for accesses to the tied variables. However, we need to separate
1076the arguments exposed to the method to the argument exposed to the
1077original function - the store or fetch or whatever it may be. Here's how
1078the tied C<push> is implemented; see C<av_push> in F<av.c>:
1079
1080 1 PUSHMARK(SP);
1081 2 EXTEND(SP,2);
1082 3 PUSHs(SvTIED_obj((SV*)av, mg));
1083 4 PUSHs(val);
1084 5 PUTBACK;
1085 6 ENTER;
1086 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1087 8 LEAVE;
1088 9 POPSTACK;
13a2d996 1089
a422fd2d
SC
1090The lines which concern the mark stack are the first, fifth and last
1091lines: they save away, restore and remove the current position of the
1092argument stack.
1093
1094Let's examine the whole implementation, for practice:
1095
1096 1 PUSHMARK(SP);
1097
1098Push the current state of the stack pointer onto the mark stack. This is
1099so that when we've finished adding items to the argument stack, Perl
1100knows how many things we've added recently.
1101
1102 2 EXTEND(SP,2);
1103 3 PUSHs(SvTIED_obj((SV*)av, mg));
1104 4 PUSHs(val);
1105
1106We're going to add two more items onto the argument stack: when you have
1107a tied array, the C<PUSH> subroutine receives the object and the value
1108to be pushed, and that's exactly what we have here - the tied object,
1109retrieved with C<SvTIED_obj>, and the value, the SV C<val>.
1110
1111 5 PUTBACK;
1112
1113Next we tell Perl to make the change to the global stack pointer: C<dSP>
1114only gave us a local copy, not a reference to the global.
1115
1116 6 ENTER;
1117 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1118 8 LEAVE;
1119
1120C<ENTER> and C<LEAVE> localise a block of code - they make sure that all
1121variables are tidied up, everything that has been localised gets
1122its previous value returned, and so on. Think of them as the C<{> and
1123C<}> of a Perl block.
1124
1125To actually do the magic method call, we have to call a subroutine in
1126Perl space: C<call_method> takes care of that, and it's described in
1127L<perlcall>. We call the C<PUSH> method in scalar context, and we're
1128going to discard its return value.
1129
1130 9 POPSTACK;
1131
1132Finally, we remove the value we placed on the mark stack, since we
1133don't need it any more.
1134
1135=item Save stack
1136
1137C doesn't have a concept of local scope, so perl provides one. We've
1138seen that C<ENTER> and C<LEAVE> are used as scoping braces; the save
1139stack implements the C equivalent of, for example:
1140
1141 {
1142 local $foo = 42;
1143 ...
1144 }
1145
1146See L<perlguts/Localising Changes> for how to use the save stack.
1147
1148=back
1149
1150=head2 Millions of Macros
1151
1152One thing you'll notice about the Perl source is that it's full of
1153macros. Some have called the pervasive use of macros the hardest thing
1154to understand, others find it adds to clarity. Let's take an example,
1155the code which implements the addition operator:
1156
1157 1 PP(pp_add)
1158 2 {
39644a26 1159 3 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
a422fd2d
SC
1160 4 {
1161 5 dPOPTOPnnrl_ul;
1162 6 SETn( left + right );
1163 7 RETURN;
1164 8 }
1165 9 }
1166
1167Every line here (apart from the braces, of course) contains a macro. The
1168first line sets up the function declaration as Perl expects for PP code;
1169line 3 sets up variable declarations for the argument stack and the
1170target, the return value of the operation. Finally, it tries to see if
1171the addition operation is overloaded; if so, the appropriate subroutine
1172is called.
1173
1174Line 5 is another variable declaration - all variable declarations start
1175with C<d> - which pops from the top of the argument stack two NVs (hence
1176C<nn>) and puts them into the variables C<right> and C<left>, hence the
1177C<rl>. These are the two operands to the addition operator. Next, we
1178call C<SETn> to set the NV of the return value to the result of adding
1179the two values. This done, we return - the C<RETURN> macro makes sure
1180that our return value is properly handled, and we pass the next operator
1181to run back to the main run loop.
1182
1183Most of these macros are explained in L<perlapi>, and some of the more
1184important ones are explained in L<perlxs> as well. Pay special attention
1185to L<perlguts/Background and PERL_IMPLICIT_CONTEXT> for information on
1186the C<[pad]THX_?> macros.
1187
1188
1189=head2 Poking at Perl
1190
1191To really poke around with Perl, you'll probably want to build Perl for
1192debugging, like this:
1193
1194 ./Configure -d -D optimize=-g
1195 make
1196
1197C<-g> is a flag to the C compiler to have it produce debugging
1198information which will allow us to step through a running program.
1199F<Configure> will also turn on the C<DEBUGGING> compilation symbol which
1200enables all the internal debugging code in Perl. There are a whole bunch
1201of things you can debug with this: L<perlrun> lists them all, and the
1202best way to find out about them is to play about with them. The most
1203useful options are probably
1204
1205 l Context (loop) stack processing
1206 t Trace execution
1207 o Method and overloading resolution
1208 c String/numeric conversions
1209
1210Some of the functionality of the debugging code can be achieved using XS
1211modules.
13a2d996 1212
a422fd2d
SC
1213 -Dr => use re 'debug'
1214 -Dx => use O 'Debug'
1215
1216=head2 Using a source-level debugger
1217
1218If the debugging output of C<-D> doesn't help you, it's time to step
1219through perl's execution with a source-level debugger.
1220
1221=over 3
1222
1223=item *
1224
1225We'll use C<gdb> for our examples here; the principles will apply to any
1226debugger, but check the manual of the one you're using.
1227
1228=back
1229
1230To fire up the debugger, type
1231
1232 gdb ./perl
1233
1234You'll want to do that in your Perl source tree so the debugger can read
1235the source code. You should see the copyright message, followed by the
1236prompt.
1237
1238 (gdb)
1239
1240C<help> will get you into the documentation, but here are the most
1241useful commands:
1242
1243=over 3
1244
1245=item run [args]
1246
1247Run the program with the given arguments.
1248
1249=item break function_name
1250
1251=item break source.c:xxx
1252
1253Tells the debugger that we'll want to pause execution when we reach
cea6626f 1254either the named function (but see L<perlguts/Internal Functions>!) or the given
a422fd2d
SC
1255line in the named source file.
1256
1257=item step
1258
1259Steps through the program a line at a time.
1260
1261=item next
1262
1263Steps through the program a line at a time, without descending into
1264functions.
1265
1266=item continue
1267
1268Run until the next breakpoint.
1269
1270=item finish
1271
1272Run until the end of the current function, then stop again.
1273
13a2d996 1274=item 'enter'
a422fd2d
SC
1275
1276Just pressing Enter will do the most recent operation again - it's a
1277blessing when stepping through miles of source code.
1278
1279=item print
1280
1281Execute the given C code and print its results. B<WARNING>: Perl makes
1282heavy use of macros, and F<gdb> is not aware of macros. You'll have to
1283substitute them yourself. So, for instance, you can't say
1284
1285 print SvPV_nolen(sv)
1286
1287but you have to say
1288
1289 print Perl_sv_2pv_nolen(sv)
1290
1291You may find it helpful to have a "macro dictionary", which you can
1292produce by saying C<cpp -dM perl.c | sort>. Even then, F<cpp> won't
1293recursively apply the macros for you.
1294
1295=back
1296
1297=head2 Dumping Perl Data Structures
1298
1299One way to get around this macro hell is to use the dumping functions in
1300F<dump.c>; these work a little like an internal
1301L<Devel::Peek|Devel::Peek>, but they also cover OPs and other structures
1302that you can't get at from Perl. Let's take an example. We'll use the
1303C<$a = $b + $c> we used before, but give it a bit of context:
1304C<$b = "6XXXX"; $c = 2.3;>. Where's a good place to stop and poke around?
1305
1306What about C<pp_add>, the function we examined earlier to implement the
1307C<+> operator:
1308
1309 (gdb) break Perl_pp_add
1310 Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.
1311
cea6626f 1312Notice we use C<Perl_pp_add> and not C<pp_add> - see L<perlguts/Internal Functions>.
a422fd2d
SC
1313With the breakpoint in place, we can run our program:
1314
1315 (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'
1316
1317Lots of junk will go past as gdb reads in the relevant source files and
1318libraries, and then:
1319
1320 Breakpoint 1, Perl_pp_add () at pp_hot.c:309
39644a26 1321 309 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
a422fd2d
SC
1322 (gdb) step
1323 311 dPOPTOPnnrl_ul;
1324 (gdb)
1325
1326We looked at this bit of code before, and we said that C<dPOPTOPnnrl_ul>
1327arranges for two C<NV>s to be placed into C<left> and C<right> - let's
1328slightly expand it:
1329
1330 #define dPOPTOPnnrl_ul NV right = POPn; \
1331 SV *leftsv = TOPs; \
1332 NV left = USE_LEFT(leftsv) ? SvNV(leftsv) : 0.0
1333
1334C<POPn> takes the SV from the top of the stack and obtains its NV either
1335directly (if C<SvNOK> is set) or by calling the C<sv_2nv> function.
1336C<TOPs> takes the next SV from the top of the stack - yes, C<POPn> uses
1337C<TOPs> - but doesn't remove it. We then use C<SvNV> to get the NV from
1338C<leftsv> in the same way as before - yes, C<POPn> uses C<SvNV>.
1339
1340Since we don't have an NV for C<$b>, we'll have to use C<sv_2nv> to
1341convert it. If we step again, we'll find ourselves there:
1342
1343 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
1344 1669 if (!sv)
1345 (gdb)
1346
1347We can now use C<Perl_sv_dump> to investigate the SV:
1348
1349 SV = PV(0xa057cc0) at 0xa0675d0
1350 REFCNT = 1
1351 FLAGS = (POK,pPOK)
1352 PV = 0xa06a510 "6XXXX"\0
1353 CUR = 5
1354 LEN = 6
1355 $1 = void
1356
1357We know we're going to get C<6> from this, so let's finish the
1358subroutine:
1359
1360 (gdb) finish
1361 Run till exit from #0 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
1362 0x462669 in Perl_pp_add () at pp_hot.c:311
1363 311 dPOPTOPnnrl_ul;
1364
1365We can also dump out this op: the current op is always stored in
1366C<PL_op>, and we can dump it with C<Perl_op_dump>. This'll give us
1367similar output to L<B::Debug|B::Debug>.
1368
1369 {
1370 13 TYPE = add ===> 14
1371 TARG = 1
1372 FLAGS = (SCALAR,KIDS)
1373 {
1374 TYPE = null ===> (12)
1375 (was rv2sv)
1376 FLAGS = (SCALAR,KIDS)
1377 {
1378 11 TYPE = gvsv ===> 12
1379 FLAGS = (SCALAR)
1380 GV = main::b
1381 }
1382 }
1383
10f58044 1384# finish this later #
a422fd2d
SC
1385
1386=head2 Patching
1387
1388All right, we've now had a look at how to navigate the Perl sources and
1389some things you'll need to know when fiddling with them. Let's now get
1390on and create a simple patch. Here's something Larry suggested: if a
1391C<U> is the first active format during a C<pack>, (for example,
1392C<pack "U3C8", @stuff>) then the resulting string should be treated as
1393UTF8 encoded.
1394
1395How do we prepare to fix this up? First we locate the code in question -
1396the C<pack> happens at runtime, so it's going to be in one of the F<pp>
1397files. Sure enough, C<pp_pack> is in F<pp.c>. Since we're going to be
1398altering this file, let's copy it to F<pp.c~>.
1399
a6ec74c1
JH
1400[Well, it was in F<pp.c> when this tutorial was written. It has now been
1401split off with C<pp_unpack> to its own file, F<pp_pack.c>]
1402
a422fd2d
SC
1403Now let's look over C<pp_pack>: we take a pattern into C<pat>, and then
1404loop over the pattern, taking each format character in turn into
1405C<datum_type>. Then for each possible format character, we swallow up
1406the other arguments in the pattern (a field width, an asterisk, and so
1407on) and convert the next chunk input into the specified format, adding
1408it onto the output SV C<cat>.
1409
1410How do we know if the C<U> is the first format in the C<pat>? Well, if
1411we have a pointer to the start of C<pat> then, if we see a C<U> we can
1412test whether we're still at the start of the string. So, here's where
1413C<pat> is set up:
1414
1415 STRLEN fromlen;
1416 register char *pat = SvPVx(*++MARK, fromlen);
1417 register char *patend = pat + fromlen;
1418 register I32 len;
1419 I32 datumtype;
1420 SV *fromstr;
1421
1422We'll have another string pointer in there:
1423
1424 STRLEN fromlen;
1425 register char *pat = SvPVx(*++MARK, fromlen);
1426 register char *patend = pat + fromlen;
1427 + char *patcopy;
1428 register I32 len;
1429 I32 datumtype;
1430 SV *fromstr;
1431
1432And just before we start the loop, we'll set C<patcopy> to be the start
1433of C<pat>:
1434
1435 items = SP - MARK;
1436 MARK++;
1437 sv_setpvn(cat, "", 0);
1438 + patcopy = pat;
1439 while (pat < patend) {
1440
1441Now if we see a C<U> which was at the start of the string, we turn on
1442the UTF8 flag for the output SV, C<cat>:
1443
1444 + if (datumtype == 'U' && pat==patcopy+1)
1445 + SvUTF8_on(cat);
1446 if (datumtype == '#') {
1447 while (pat < patend && *pat != '\n')
1448 pat++;
1449
1450Remember that it has to be C<patcopy+1> because the first character of
1451the string is the C<U> which has been swallowed into C<datumtype!>
1452
1453Oops, we forgot one thing: what if there are spaces at the start of the
1454pattern? C<pack(" U*", @stuff)> will have C<U> as the first active
1455character, even though it's not the first thing in the pattern. In this
1456case, we have to advance C<patcopy> along with C<pat> when we see spaces:
1457
1458 if (isSPACE(datumtype))
1459 continue;
1460
1461needs to become
1462
1463 if (isSPACE(datumtype)) {
1464 patcopy++;
1465 continue;
1466 }
1467
1468OK. That's the C part done. Now we must do two additional things before
1469this patch is ready to go: we've changed the behaviour of Perl, and so
1470we must document that change. We must also provide some more regression
1471tests to make sure our patch works and doesn't create a bug somewhere
1472else along the line.
1473
1474The regression tests for each operator live in F<t/op/>, and so we make
1475a copy of F<t/op/pack.t> to F<t/op/pack.t~>. Now we can add our tests
1476to the end. First, we'll test that the C<U> does indeed create Unicode
1477strings:
1478
1479 print 'not ' unless "1.20.300.4000" eq sprintf "%vd", pack("U*",1,20,300,4000);
1480 print "ok $test\n"; $test++;
1481
1482Now we'll test that we got that space-at-the-beginning business right:
1483
1484 print 'not ' unless "1.20.300.4000" eq
1485 sprintf "%vd", pack(" U*",1,20,300,4000);
1486 print "ok $test\n"; $test++;
1487
1488And finally we'll test that we don't make Unicode strings if C<U> is B<not>
1489the first active format:
1490
1491 print 'not ' unless v1.20.300.4000 ne
1492 sprintf "%vd", pack("C0U*",1,20,300,4000);
1493 print "ok $test\n"; $test++;
1494
b1866b2d 1495Mustn't forget to change the number of tests which appears at the top, or
a422fd2d
SC
1496else the automated tester will get confused:
1497
1498 -print "1..156\n";
1499 +print "1..159\n";
1500
1501We now compile up Perl, and run it through the test suite. Our new
1502tests pass, hooray!
1503
1504Finally, the documentation. The job is never done until the paperwork is
1505over, so let's describe the change we've just made. The relevant place
1506is F<pod/perlfunc.pod>; again, we make a copy, and then we'll insert
1507this text in the description of C<pack>:
1508
1509 =item *
1510
1511 If the pattern begins with a C<U>, the resulting string will be treated
1512 as Unicode-encoded. You can force UTF8 encoding on in a string with an
1513 initial C<U0>, and the bytes that follow will be interpreted as Unicode
1514 characters. If you don't want this to happen, you can begin your pattern
1515 with C<C0> (or anything else) to force Perl not to UTF8 encode your
1516 string, and then follow this with a C<U*> somewhere in your pattern.
1517
1518All done. Now let's create the patch. F<Porting/patching.pod> tells us
1519that if we're making major changes, we should copy the entire directory
1520to somewhere safe before we begin fiddling, and then do
13a2d996 1521
a422fd2d
SC
1522 diff -ruN old new > patch
1523
1524However, we know which files we've changed, and we can simply do this:
1525
1526 diff -u pp.c~ pp.c > patch
1527 diff -u t/op/pack.t~ t/op/pack.t >> patch
1528 diff -u pod/perlfunc.pod~ pod/perlfunc.pod >> patch
1529
1530We end up with a patch looking a little like this:
1531
1532 --- pp.c~ Fri Jun 02 04:34:10 2000
1533 +++ pp.c Fri Jun 16 11:37:25 2000
1534 @@ -4375,6 +4375,7 @@
1535 register I32 items;
1536 STRLEN fromlen;
1537 register char *pat = SvPVx(*++MARK, fromlen);
1538 + char *patcopy;
1539 register char *patend = pat + fromlen;
1540 register I32 len;
1541 I32 datumtype;
1542 @@ -4405,6 +4406,7 @@
1543 ...
1544
1545And finally, we submit it, with our rationale, to perl5-porters. Job
1546done!
1547
f7e1e956
MS
1548=head2 Patching a core module
1549
1550This works just like patching anything else, with an extra
1551consideration. Many core modules also live on CPAN. If this is so,
1552patch the CPAN version instead of the core and send the patch off to
1553the module maintainer (with a copy to p5p). This will help the module
1554maintainer keep the CPAN version in sync with the core version without
1555constantly scanning p5p.
1556
acbe17fc
JP
1557=head2 Adding a new function to the core
1558
1559If, as part of a patch to fix a bug, or just because you have an
1560especially good idea, you decide to add a new function to the core,
1561discuss your ideas on p5p well before you start work. It may be that
1562someone else has already attempted to do what you are considering and
1563can give lots of good advice or even provide you with bits of code
1564that they already started (but never finished).
1565
1566You have to follow all of the advice given above for patching. It is
1567extremely important to test any addition thoroughly and add new tests
1568to explore all boundary conditions that your new function is expected
1569to handle. If your new function is used only by one module (e.g. toke),
1570then it should probably be named S_your_function (for static); on the
1571other hand, if you expect it to accessable from other functions in
1572Perl, you should name it Perl_your_function. See L<perlguts/Internal Functions>
1573for more details.
1574
1575The location of any new code is also an important consideration. Don't
1576just create a new top level .c file and put your code there; you would
1577have to make changes to Configure (so the Makefile is created properly),
1578as well as possibly lots of include files. This is strictly pumpking
1579business.
1580
1581It is better to add your function to one of the existing top level
1582source code files, but your choice is complicated by the nature of
1583the Perl distribution. Only the files that are marked as compiled
1584static are located in the perl executable. Everything else is located
1585in the shared library (or DLL if you are running under WIN32). So,
1586for example, if a function was only used by functions located in
1587toke.c, then your code can go in toke.c. If, however, you want to call
1588the function from universal.c, then you should put your code in another
1589location, for example util.c.
1590
1591In addition to writing your c-code, you will need to create an
1592appropriate entry in embed.pl describing your function, then run
1593'make regen_headers' to create the entries in the numerous header
1594files that perl needs to compile correctly. See L<perlguts/Internal Functions>
1595for information on the various options that you can set in embed.pl.
1596You will forget to do this a few (or many) times and you will get
1597warnings during the compilation phase. Make sure that you mention
1598this when you post your patch to P5P; the pumpking needs to know this.
1599
1600When you write your new code, please be conscious of existing code
1601conventions used in the perl source files. See <perlstyle> for
1602details. Although most of the guidelines discussed seem to focus on
1603Perl code, rather than c, they all apply (except when they don't ;).
1604See also I<Porting/patching.pod> file in the Perl source distribution
1605for lots of details about both formatting and submitting patches of
1606your changes.
1607
1608Lastly, TEST TEST TEST TEST TEST any code before posting to p5p.
1609Test on as many platforms as you can find. Test as many perl
1610Configure options as you can (e.g. MULTIPLICITY). If you have
1611profiling or memory tools, see L<EXTERNAL TOOLS FOR DEBUGGING PERL>
1612below for how to use them to futher test your code. Remember that
1613most of the people on P5P are doing this on their own time and
1614don't have the time to debug your code.
f7e1e956
MS
1615
1616=head2 Writing a test
1617
1618Every module and built-in function has an associated test file (or
1619should...). If you add or change functionality, you have to write a
1620test. If you fix a bug, you have to write a test so that bug never
1621comes back. If you alter the docs, it would be nice to test what the
1622new documentation says.
1623
1624In short, if you submit a patch you probably also have to patch the
1625tests.
1626
1627For modules, the test file is right next to the module itself.
1628F<lib/strict.t> tests F<lib/strict.pm>. This is a recent innovation,
1629so there are some snags (and it would be wonderful for you to brush
1630them out), but it basically works that way. Everything else lives in
1631F<t/>.
1632
1633=over 3
1634
1635=item F<t/base/>
1636
1637Testing of the absolute basic functionality of Perl. Things like
1638C<if>, basic file reads and writes, simple regexes, etc. These are
1639run first in the test suite and if any of them fail, something is
1640I<really> broken.
1641
1642=item F<t/cmd/>
1643
1644These test the basic control structures, C<if/else>, C<while>,
1645subroutines, etc...
1646
1647=item F<t/comp/>
1648
1649Tests basic issues of how Perl parses and compiles itself.
1650
1651=item F<t/io/>
1652
1653Tests for built-in IO functions, including command line arguments.
1654
1655=item F<t/lib/>
1656
1657The old home for the module tests, you shouldn't put anything new in
1658here. There are still some bits and pieces hanging around in here
1659that need to be moved. Perhaps you could move them? Thanks!
1660
1661=item F<t/op/>
1662
1663Tests for perl's built in functions that don't fit into any of the
1664other directories.
1665
1666=item F<t/pod/>
1667
1668Tests for POD directives. There are still some tests for the Pod
1669modules hanging around in here that need to be moved out into F<lib/>.
1670
1671=item F<t/run/>
1672
1673Testing features of how perl actually runs, including exit codes and
1674handling of PERL* environment variables.
1675
1676=back
1677
1678The core uses the same testing style as the rest of Perl, a simple
1679"ok/not ok" run through Test::Harness, but there are a few special
1680considerations.
1681
1682For most libraries and extensions, you'll want to use the Test::More
1683library rather than rolling your own test functions. If a module test
1684doesn't use Test::More, consider rewriting it so it does. For the
23acf682 1685rest it's best to use a simple C<print "ok $test_num\n"> style to avoid
f7e1e956
MS
1686broken core functionality from causing the whole test to collapse.
1687
1688When you say "make test" Perl uses the F<t/TEST> program to run the
1689test suite. All tests are run from the F<t/> directory, B<not> the
1690directory which contains the test. This causes some problems with the
1691tests in F<lib/>, so here's some opportunity for some patching.
1692
1693You must be triply conscious of cross-platform concerns. This usually
1694boils down to using File::Spec and avoiding things like C<fork()> and
1695C<system()> unless absolutely necessary.
1696
1697
902b9dbf
MF
1698=head1 EXTERNAL TOOLS FOR DEBUGGING PERL
1699
1700Sometimes it helps to use external tools while debugging and
1701testing Perl. This section tries to guide you through using
1702some common testing and debugging tools with Perl. This is
1703meant as a guide to interfacing these tools with Perl, not
1704as any kind of guide to the use of the tools themselves.
1705
1706=head2 Rational Software's Purify
1707
1708Purify is a commercial tool that is helpful in identifying
1709memory overruns, wild pointers, memory leaks and other such
1710badness. Perl must be compiled in a specific way for
1711optimal testing with Purify. Purify is available under
1712Windows NT, Solaris, HP-UX, SGI, and Siemens Unix.
1713
1714The only currently known leaks happen when there are
1715compile-time errors within eval or require. (Fixing these
1716is non-trivial, unfortunately, but they must be fixed
1717eventually.)
1718
1719=head2 Purify on Unix
1720
1721On Unix, Purify creates a new Perl binary. To get the most
1722benefit out of Purify, you should create the perl to Purify
1723using:
1724
1725 sh Configure -Accflags=-DPURIFY -Doptimize='-g' \
1726 -Uusemymalloc -Dusemultiplicity
1727
1728where these arguments mean:
1729
1730=over 4
1731
1732=item -Accflags=-DPURIFY
1733
1734Disables Perl's arena memory allocation functions, as well as
1735forcing use of memory allocation functions derived from the
1736system malloc.
1737
1738=item -Doptimize='-g'
1739
1740Adds debugging information so that you see the exact source
1741statements where the problem occurs. Without this flag, all
1742you will see is the source filename of where the error occurred.
1743
1744=item -Uusemymalloc
1745
1746Disable Perl's malloc so that Purify can more closely monitor
1747allocations and leaks. Using Perl's malloc will make Purify
1748report most leaks in the "potential" leaks category.
1749
1750=item -Dusemultiplicity
1751
1752Enabling the multiplicity option allows perl to clean up
1753thoroughly when the interpreter shuts down, which reduces the
1754number of bogus leak reports from Purify.
1755
1756=back
1757
1758Once you've compiled a perl suitable for Purify'ing, then you
1759can just:
1760
1761 make pureperl
1762
1763which creates a binary named 'pureperl' that has been Purify'ed.
1764This binary is used in place of the standard 'perl' binary
1765when you want to debug Perl memory problems.
1766
1767As an example, to show any memory leaks produced during the
1768standard Perl testset you would create and run the Purify'ed
1769perl as:
1770
1771 make pureperl
1772 cd t
1773 ../pureperl -I../lib harness
1774
1775which would run Perl on test.pl and report any memory problems.
1776
1777Purify outputs messages in "Viewer" windows by default. If
1778you don't have a windowing environment or if you simply
1779want the Purify output to unobtrusively go to a log file
1780instead of to the interactive window, use these following
1781options to output to the log file "perl.log":
1782
1783 setenv PURIFYOPTIONS "-chain-length=25 -windows=no \
1784 -log-file=perl.log -append-logfile=yes"
1785
1786If you plan to use the "Viewer" windows, then you only need this option:
1787
1788 setenv PURIFYOPTIONS "-chain-length=25"
1789
1790=head2 Purify on NT
1791
1792Purify on Windows NT instruments the Perl binary 'perl.exe'
1793on the fly. There are several options in the makefile you
1794should change to get the most use out of Purify:
1795
1796=over 4
1797
1798=item DEFINES
1799
1800You should add -DPURIFY to the DEFINES line so the DEFINES
1801line looks something like:
1802
1803 DEFINES = -DWIN32 -D_CONSOLE -DNO_STRICT $(CRYPT_FLAG) -DPURIFY=1
1804
1805to disable Perl's arena memory allocation functions, as
1806well as to force use of memory allocation functions derived
1807from the system malloc.
1808
1809=item USE_MULTI = define
1810
1811Enabling the multiplicity option allows perl to clean up
1812thoroughly when the interpreter shuts down, which reduces the
1813number of bogus leak reports from Purify.
1814
1815=item #PERL_MALLOC = define
1816
1817Disable Perl's malloc so that Purify can more closely monitor
1818allocations and leaks. Using Perl's malloc will make Purify
1819report most leaks in the "potential" leaks category.
1820
1821=item CFG = Debug
1822
1823Adds debugging information so that you see the exact source
1824statements where the problem occurs. Without this flag, all
1825you will see is the source filename of where the error occurred.
1826
1827=back
1828
1829As an example, to show any memory leaks produced during the
1830standard Perl testset you would create and run Purify as:
1831
1832 cd win32
1833 make
1834 cd ../t
1835 purify ../perl -I../lib harness
1836
1837which would instrument Perl in memory, run Perl on test.pl,
1838then finally report any memory problems.
1839
09187cb1
JH
1840=head2 Compaq's/Digital's Third Degree
1841
1842Third Degree is a tool for memory leak detection and memory access checks.
1843It is one of the many tools in the ATOM toolkit. The toolkit is only
1844available on Tru64 (formerly known as Digital UNIX formerly known as
1845DEC OSF/1).
1846
1847When building Perl, you must first run Configure with -Doptimize=-g
1848and -Uusemymalloc flags, after that you can use the make targets
51a35ef1
JH
1849"perl.third" and "test.third". (What is required is that Perl must be
1850compiled using the C<-g> flag, you may need to re-Configure.)
09187cb1 1851
64cea5fd 1852The short story is that with "atom" you can instrument the Perl
83f0ef60 1853executable to create a new executable called F<perl.third>. When the
4ae3d70a 1854instrumented executable is run, it creates a log of dubious memory
83f0ef60 1855traffic in file called F<perl.3log>. See the manual pages of atom and
4ae3d70a
JH
1856third for more information. The most extensive Third Degree
1857documentation is available in the Compaq "Tru64 UNIX Programmer's
1858Guide", chapter "Debugging Programs with Third Degree".
64cea5fd 1859
83f0ef60 1860The "test.third" leaves a lot of files named F<perl.3log.*> in the t/
64cea5fd
JH
1861subdirectory. There is a problem with these files: Third Degree is so
1862effective that it finds problems also in the system libraries.
83f0ef60
JH
1863Therefore there are certain types of errors that you should ignore in
1864your debugging. Errors with stack traces matching
64cea5fd
JH
1865
1866 __actual_atof|__catgets|_doprnt|__exc_|__exec|_findio|__localtime|setlocale|__sia_|__strxfrm
1867
1868(all in libc.so) are known to be non-serious. You can also
1869ignore the combinations
1870
1871 Perl_gv_fetchfile() calling strcpy()
1872 S_doopen_pmc() calling strcmp()
1873
1874causing "rih" (reading invalid heap) errors.
1875
1876There are also leaks that for given certain definition of a leak,
1877aren't. See L</PERL_DESTRUCT_LEVEL> for more information.
1878
1879=head2 PERL_DESTRUCT_LEVEL
1880
1881If you want to run any of the tests yourself manually using the
1882pureperl or perl.third executables, please note that by default
1883perl B<does not> explicitly cleanup all the memory it has allocated
1884(such as global memory arenas) but instead lets the exit() of
1885the whole program "take care" of such allocations, also known
1886as "global destruction of objects".
1887
1888There is a way to tell perl to do complete cleanup: set the
1889environment variable PERL_DESTRUCT_LEVEL to a non-zero value.
1890The t/TEST wrapper does set this to 2, and this is what you
1891need to do too, if you don't want to see the "global leaks":
1892
1893 PERL_DESTRUCT_LEVEL=2 ./perl.third t/foo/bar.t
09187cb1 1894
51a35ef1
JH
1895=head2 Profiling
1896
1897Depending on your platform there are various of profiling Perl.
1898
1899There are two commonly used techniques of profiling executables:
10f58044 1900I<statistical time-sampling> and I<basic-block counting>.
51a35ef1
JH
1901
1902The first method takes periodically samples of the CPU program
1903counter, and since the program counter can be correlated with the code
1904generated for functions, we get a statistical view of in which
1905functions the program is spending its time. The caveats are that very
1906small/fast functions have lower probability of showing up in the
1907profile, and that periodically interrupting the program (this is
1908usually done rather frequently, in the scale of milliseconds) imposes
1909an additional overhead that may skew the results. The first problem
1910can be alleviated by running the code for longer (in general this is a
1911good idea for profiling), the second problem is usually kept in guard
1912by the profiling tools themselves.
1913
10f58044 1914The second method divides up the generated code into I<basic blocks>.
51a35ef1
JH
1915Basic blocks are sections of code that are entered only in the
1916beginning and exited only at the end. For example, a conditional jump
1917starts a basic block. Basic block profiling usually works by
10f58044 1918I<instrumenting> the code by adding I<enter basic block #nnnn>
51a35ef1
JH
1919book-keeping code to the generated code. During the execution of the
1920code the basic block counters are then updated appropriately. The
1921caveat is that the added extra code can skew the results: again, the
1922profiling tools usually try to factor their own effects out of the
1923results.
1924
83f0ef60
JH
1925=head2 Gprof Profiling
1926
51a35ef1
JH
1927gprof is a profiling tool available in many UNIX platforms,
1928it uses F<statistical time-sampling>.
83f0ef60
JH
1929
1930You can build a profiled version of perl called "perl.gprof" by
51a35ef1
JH
1931invoking the make target "perl.gprof" (What is required is that Perl
1932must be compiled using the C<-pg> flag, you may need to re-Configure).
1933Running the profiled version of Perl will create an output file called
1934F<gmon.out> is created which contains the profiling data collected
1935during the execution.
83f0ef60
JH
1936
1937The gprof tool can then display the collected data in various ways.
1938Usually gprof understands the following options:
1939
1940=over 4
1941
1942=item -a
1943
1944Suppress statically defined functions from the profile.
1945
1946=item -b
1947
1948Suppress the verbose descriptions in the profile.
1949
1950=item -e routine
1951
1952Exclude the given routine and its descendants from the profile.
1953
1954=item -f routine
1955
1956Display only the given routine and its descendants in the profile.
1957
1958=item -s
1959
1960Generate a summary file called F<gmon.sum> which then may be given
1961to subsequent gprof runs to accumulate data over several runs.
1962
1963=item -z
1964
1965Display routines that have zero usage.
1966
1967=back
1968
1969For more detailed explanation of the available commands and output
1970formats, see your own local documentation of gprof.
1971
51a35ef1
JH
1972=head2 GCC gcov Profiling
1973
10f58044 1974Starting from GCC 3.0 I<basic block profiling> is officially available
51a35ef1
JH
1975for the GNU CC.
1976
1977You can build a profiled version of perl called F<perl.gcov> by
1978invoking the make target "perl.gcov" (what is required that Perl must
1979be compiled using gcc with the flags C<-fprofile-arcs
1980-ftest-coverage>, you may need to re-Configure).
1981
1982Running the profiled version of Perl will cause profile output to be
1983generated. For each source file an accompanying ".da" file will be
1984created.
1985
1986To display the results you use the "gcov" utility (which should
1987be installed if you have gcc 3.0 or newer installed). F<gcov> is
1988run on source code files, like this
1989
1990 gcov sv.c
1991
1992which will cause F<sv.c.gcov> to be created. The F<.gcov> files
1993contain the source code annotated with relative frequencies of
1994execution indicated by "#" markers.
1995
1996Useful options of F<gcov> include C<-b> which will summarise the
1997basic block, branch, and function call coverage, and C<-c> which
1998instead of relative frequencies will use the actual counts. For
1999more information on the use of F<gcov> and basic block profiling
2000with gcc, see the latest GNU CC manual, as of GCC 3.0 see
2001
2002 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc.html
2003
2004and its section titled "8. gcov: a Test Coverage Program"
2005
2006 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc_8.html#SEC132
2007
4ae3d70a
JH
2008=head2 Pixie Profiling
2009
51a35ef1
JH
2010Pixie is a profiling tool available on IRIX and Tru64 (aka Digital
2011UNIX aka DEC OSF/1) platforms. Pixie does its profiling using
10f58044 2012I<basic-block counting>.
4ae3d70a 2013
83f0ef60 2014You can build a profiled version of perl called F<perl.pixie> by
51a35ef1
JH
2015invoking the make target "perl.pixie" (what is required is that Perl
2016must be compiled using the C<-g> flag, you may need to re-Configure).
2017
2018In Tru64 a file called F<perl.Addrs> will also be silently created,
2019this file contains the addresses of the basic blocks. Running the
2020profiled version of Perl will create a new file called "perl.Counts"
2021which contains the counts for the basic block for that particular
2022program execution.
4ae3d70a 2023
51a35ef1 2024To display the results you use the F<prof> utility. The exact
4ae3d70a
JH
2025incantation depends on your operating system, "prof perl.Counts" in
2026IRIX, and "prof -pixie -all -L. perl" in Tru64.
2027
6c41479b
JH
2028In IRIX the following prof options are available:
2029
2030=over 4
2031
2032=item -h
2033
2034Reports the most heavily used lines in descending order of use.
6e36760b 2035Useful for finding the hotspot lines.
6c41479b
JH
2036
2037=item -l
2038
2039Groups lines by procedure, with procedures sorted in descending order of use.
2040Within a procedure, lines are listed in source order.
6e36760b 2041Useful for finding the hotspots of procedures.
6c41479b
JH
2042
2043=back
2044
2045In Tru64 the following options are available:
2046
2047=over 4
2048
3958b146 2049=item -p[rocedures]
6c41479b 2050
3958b146 2051Procedures sorted in descending order by the number of cycles executed
6e36760b 2052in each procedure. Useful for finding the hotspot procedures.
6c41479b
JH
2053(This is the default option.)
2054
24000d2f 2055=item -h[eavy]
6c41479b 2056
6e36760b
JH
2057Lines sorted in descending order by the number of cycles executed in
2058each line. Useful for finding the hotspot lines.
6c41479b 2059
24000d2f 2060=item -i[nvocations]
6c41479b 2061
6e36760b
JH
2062The called procedures are sorted in descending order by number of calls
2063made to the procedures. Useful for finding the most used procedures.
6c41479b 2064
24000d2f 2065=item -l[ines]
6c41479b
JH
2066
2067Grouped by procedure, sorted by cycles executed per procedure.
6e36760b 2068Useful for finding the hotspots of procedures.
6c41479b
JH
2069
2070=item -testcoverage
2071
2072The compiler emitted code for these lines, but the code was unexecuted.
2073
24000d2f 2074=item -z[ero]
6c41479b
JH
2075
2076Unexecuted procedures.
2077
aa500c9e 2078=back
6c41479b
JH
2079
2080For further information, see your system's manual pages for pixie and prof.
4ae3d70a 2081
a422fd2d
SC
2082=head2 CONCLUSION
2083
2084We've had a brief look around the Perl source, an overview of the stages
2085F<perl> goes through when it's running your code, and how to use a
902b9dbf
MF
2086debugger to poke at the Perl guts. We took a very simple problem and
2087demonstrated how to solve it fully - with documentation, regression
2088tests, and finally a patch for submission to p5p. Finally, we talked
2089about how to use external tools to debug and test Perl.
a422fd2d
SC
2090
2091I'd now suggest you read over those references again, and then, as soon
2092as possible, get your hands dirty. The best way to learn is by doing,
2093so:
2094
2095=over 3
2096
2097=item *
2098
2099Subscribe to perl5-porters, follow the patches and try and understand
2100them; don't be afraid to ask if there's a portion you're not clear on -
2101who knows, you may unearth a bug in the patch...
2102
2103=item *
2104
2105Keep up to date with the bleeding edge Perl distributions and get
2106familiar with the changes. Try and get an idea of what areas people are
2107working on and the changes they're making.
2108
2109=item *
2110
3e148164 2111Do read the README associated with your operating system, e.g. README.aix
a1f349fd
MB
2112on the IBM AIX OS. Don't hesitate to supply patches to that README if
2113you find anything missing or changed over a new OS release.
2114
2115=item *
2116
a422fd2d
SC
2117Find an area of Perl that seems interesting to you, and see if you can
2118work out how it works. Scan through the source, and step over it in the
2119debugger. Play, poke, investigate, fiddle! You'll probably get to
2120understand not just your chosen area but a much wider range of F<perl>'s
2121activity as well, and probably sooner than you'd think.
2122
2123=back
2124
2125=over 3
2126
2127=item I<The Road goes ever on and on, down from the door where it began.>
2128
2129=back
2130
2131If you can do these things, you've started on the long road to Perl porting.
2132Thanks for wanting to help make Perl better - and happy hacking!
2133
e8cd7eae
GS
2134=head1 AUTHOR
2135
2136This document was written by Nathan Torkington, and is maintained by
2137the perl5-porters mailing list.
2138