This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Resync win32/config_H.* and win32/config.* with config_h.SH
[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
f8e3975a 17A searchable archive of the list is at either:
e8cd7eae
GS
18
19 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
20
f8e3975a
IP
21or
22
23 http://archive.develooper.com/perl5-porters@perl.org/
24
e8cd7eae
GS
25List subscribers (the porters themselves) come in several flavours.
26Some are quiet curious lurkers, who rarely pitch in and instead watch
27the ongoing development to ensure they're forewarned of new changes or
28features in Perl. Some are representatives of vendors, who are there
29to make sure that Perl continues to compile and work on their
30platforms. Some patch any reported bug that they know how to fix,
31some are actively patching their pet area (threads, Win32, the regexp
32engine), while others seem to do nothing but complain. In other
33words, it's your usual mix of technical people.
34
35Over this group of porters presides Larry Wall. He has the final word
f6c51b38 36in what does and does not change in the Perl language. Various
b432a672
AL
37releases of Perl are shepherded by a "pumpking", a porter
38responsible for gathering patches, deciding on a patch-by-patch,
f6c51b38 39feature-by-feature basis what will and will not go into the release.
caf100c0 40For instance, Gurusamy Sarathy was the pumpking for the 5.6 release of
961f29c6 41Perl, and Jarkko Hietaniemi was the pumpking for the 5.8 release, and
1a88dbf8 42Rafael Garcia-Suarez holds the pumpking crown for the 5.10 release.
e8cd7eae
GS
43
44In addition, various people are pumpkings for different things. For
961f29c6
MB
45instance, Andy Dougherty and Jarkko Hietaniemi did a grand job as the
46I<Configure> pumpkin up till the 5.8 release. For the 5.10 release
47H.Merijn Brand took over.
e8cd7eae
GS
48
49Larry sees Perl development along the lines of the US government:
50there's the Legislature (the porters), the Executive branch (the
51pumpkings), and the Supreme Court (Larry). The legislature can
52discuss and submit patches to the executive branch all they like, but
53the executive branch is free to veto them. Rarely, the Supreme Court
54will side with the executive branch over the legislature, or the
55legislature over the executive branch. Mostly, however, the
56legislature and the executive branch are supposed to get along and
57work out their differences without impeachment or court cases.
58
59You might sometimes see reference to Rule 1 and Rule 2. Larry's power
60as Supreme Court is expressed in The Rules:
61
62=over 4
63
64=item 1
65
66Larry is always by definition right about how Perl should behave.
67This means he has final veto power on the core functionality.
68
69=item 2
70
71Larry is allowed to change his mind about any matter at a later date,
72regardless of whether he previously invoked Rule 1.
73
74=back
75
76Got that? Larry is always right, even when he was wrong. It's rare
77to see either Rule exercised, but they are often alluded to.
78
79New features and extensions to the language are contentious, because
80the criteria used by the pumpkings, Larry, and other porters to decide
81which features should be implemented and incorporated are not codified
82in a few small design goals as with some other languages. Instead,
83the heuristics are flexible and often difficult to fathom. Here is
84one person's list, roughly in decreasing order of importance, of
85heuristics that new features have to be weighed against:
86
87=over 4
88
89=item Does concept match the general goals of Perl?
90
91These haven't been written anywhere in stone, but one approximation
92is:
93
94 1. Keep it fast, simple, and useful.
95 2. Keep features/concepts as orthogonal as possible.
96 3. No arbitrary limits (platforms, data sizes, cultures).
97 4. Keep it open and exciting to use/patch/advocate Perl everywhere.
98 5. Either assimilate new technologies, or build bridges to them.
99
100=item Where is the implementation?
101
102All the talk in the world is useless without an implementation. In
103almost every case, the person or people who argue for a new feature
104will be expected to be the ones who implement it. Porters capable
105of coding new features have their own agendas, and are not available
106to implement your (possibly good) idea.
107
108=item Backwards compatibility
109
110It's a cardinal sin to break existing Perl programs. New warnings are
111contentious--some say that a program that emits warnings is not
112broken, while others say it is. Adding keywords has the potential to
113break programs, changing the meaning of existing token sequences or
114functions might break programs.
115
116=item Could it be a module instead?
117
118Perl 5 has extension mechanisms, modules and XS, specifically to avoid
119the need to keep changing the Perl interpreter. You can write modules
120that export functions, you can give those functions prototypes so they
121can be called like built-in functions, you can even write XS code to
122mess with the runtime data structures of the Perl interpreter if you
123want to implement really complicated things. If it can be done in a
124module instead of in the core, it's highly unlikely to be added.
125
126=item Is the feature generic enough?
127
128Is this something that only the submitter wants added to the language,
129or would it be broadly useful? Sometimes, instead of adding a feature
130with a tight focus, the porters might decide to wait until someone
131implements the more generalized feature. For instance, instead of
b432a672 132implementing a "delayed evaluation" feature, the porters are waiting
e8cd7eae
GS
133for a macro system that would permit delayed evaluation and much more.
134
135=item Does it potentially introduce new bugs?
136
137Radical rewrites of large chunks of the Perl interpreter have the
138potential to introduce new bugs. The smaller and more localized the
139change, the better.
140
141=item Does it preclude other desirable features?
142
143A patch is likely to be rejected if it closes off future avenues of
144development. For instance, a patch that placed a true and final
145interpretation on prototypes is likely to be rejected because there
146are still options for the future of prototypes that haven't been
147addressed.
148
149=item Is the implementation robust?
150
151Good patches (tight code, complete, correct) stand more chance of
152going in. Sloppy or incorrect patches might be placed on the back
153burner until the pumpking has time to fix, or might be discarded
154altogether without further notice.
155
156=item Is the implementation generic enough to be portable?
157
158The worst patches make use of a system-specific features. It's highly
159unlikely that nonportable additions to the Perl language will be
160accepted.
161
a936dd3c
NC
162=item Is the implementation tested?
163
164Patches which change behaviour (fixing bugs or introducing new features)
165must include regression tests to verify that everything works as expected.
166Without tests provided by the original author, how can anyone else changing
167perl in the future be sure that they haven't unwittingly broken the behaviour
168the patch implements? And without tests, how can the patch's author be
9d077eaa 169confident that his/her hard work put into the patch won't be accidentally
a936dd3c
NC
170thrown away by someone in the future?
171
e8cd7eae
GS
172=item Is there enough documentation?
173
174Patches without documentation are probably ill-thought out or
175incomplete. Nothing can be added without documentation, so submitting
176a patch for the appropriate manpages as well as the source code is
a936dd3c 177always a good idea.
e8cd7eae
GS
178
179=item Is there another way to do it?
180
b432a672
AL
181Larry said "Although the Perl Slogan is I<There's More Than One Way
182to Do It>, I hesitate to make 10 ways to do something". This is a
e8cd7eae
GS
183tricky heuristic to navigate, though--one man's essential addition is
184another man's pointless cruft.
185
186=item Does it create too much work?
187
188Work for the pumpking, work for Perl programmers, work for module
189authors, ... Perl is supposed to be easy.
190
f6c51b38
GS
191=item Patches speak louder than words
192
193Working code is always preferred to pie-in-the-sky ideas. A patch to
194add a feature stands a much higher chance of making it to the language
195than does a random feature request, no matter how fervently argued the
b432a672 196request might be. This ties into "Will it be useful?", as the fact
f6c51b38
GS
197that someone took the time to make the patch demonstrates a strong
198desire for the feature.
199
e8cd7eae
GS
200=back
201
b432a672
AL
202If you're on the list, you might hear the word "core" bandied
203around. It refers to the standard distribution. "Hacking on the
204core" means you're changing the C source code to the Perl
205interpreter. "A core module" is one that ships with Perl.
e8cd7eae 206
a1f349fd
MB
207=head2 Keeping in sync
208
e8cd7eae 209The source code to the Perl interpreter, in its different versions, is
f224927c
JH
210kept in a repository managed by a revision control system ( which is
211currently the Perforce program, see http://perforce.com/ ). The
e8cd7eae
GS
212pumpkings and a few others have access to the repository to check in
213changes. Periodically the pumpking for the development version of Perl
214will release a new version, so the rest of the porters can see what's
2be4c08b
GS
215changed. The current state of the main trunk of repository, and patches
216that describe the individual changes that have happened since the last
217public release are available at this location:
218
eb8afeef
SP
219 http://public.activestate.com/pub/apc/
220 ftp://public.activestate.com/pub/apc/
2be4c08b 221
0cfb3454
GS
222If you're looking for a particular change, or a change that affected
223a particular set of files, you may find the B<Perl Repository Browser>
224useful:
225
226 http://public.activestate.com/cgi-bin/perlbrowse
227
228You may also want to subscribe to the perl5-changes mailing list to
229receive a copy of each patch that gets submitted to the maintenance
230and development "branches" of the perl repository. See
231http://lists.perl.org/ for subscription information.
232
a1f349fd
MB
233If you are a member of the perl5-porters mailing list, it is a good
234thing to keep in touch with the most recent changes. If not only to
235verify if what you would have posted as a bug report isn't already
236solved in the most recent available perl development branch, also
237known as perl-current, bleading edge perl, bleedperl or bleadperl.
2be4c08b
GS
238
239Needless to say, the source code in perl-current is usually in a perpetual
240state of evolution. You should expect it to be very buggy. Do B<not> use
241it for any purpose other than testing and development.
e8cd7eae 242
3e148164
JH
243Keeping in sync with the most recent branch can be done in several ways,
244but the most convenient and reliable way is using B<rsync>, available at
245ftp://rsync.samba.org/pub/rsync/ . (You can also get the most recent
246branch by FTP.)
a1f349fd
MB
247
248If you choose to keep in sync using rsync, there are two approaches
3e148164 249to doing so:
a1f349fd
MB
250
251=over 4
252
253=item rsync'ing the source tree
254
3e148164 255Presuming you are in the directory where your perl source resides
b432a672 256and you have rsync installed and available, you can "upgrade" to
a1f349fd
MB
257the bleadperl using:
258
eb8afeef 259 # rsync -avz rsync://public.activestate.com/perl-current/ .
a1f349fd
MB
260
261This takes care of updating every single item in the source tree to
262the latest applied patch level, creating files that are new (to your
263distribution) and setting date/time stamps of existing files to
264reflect the bleadperl status.
265
c6d0653e
LC
266Note that this will not delete any files that were in '.' before
267the rsync. Once you are sure that the rsync is running correctly,
268run it with the --delete and the --dry-run options like this:
269
eb8afeef 270 # rsync -avz --delete --dry-run rsync://public.activestate.com/perl-current/ .
c6d0653e
LC
271
272This will I<simulate> an rsync run that also deletes files not
273present in the bleadperl master copy. Observe the results from
274this run closely. If you are sure that the actual run would delete
275no files precious to you, you could remove the '--dry-run' option.
276
a1f349fd
MB
277You can than check what patch was the latest that was applied by
278looking in the file B<.patch>, which will show the number of the
279latest patch.
280
281If you have more than one machine to keep in sync, and not all of
282them have access to the WAN (so you are not able to rsync all the
283source trees to the real source), there are some ways to get around
284this problem.
285
286=over 4
287
288=item Using rsync over the LAN
289
290Set up a local rsync server which makes the rsynced source tree
3e148164 291available to the LAN and sync the other machines against this
a1f349fd
MB
292directory.
293
1577cd80 294From http://rsync.samba.org/README.html :
a1f349fd
MB
295
296 "Rsync uses rsh or ssh for communication. It does not need to be
297 setuid and requires no special privileges for installation. It
3958b146 298 does not require an inetd entry or a daemon. You must, however,
a1f349fd
MB
299 have a working rsh or ssh system. Using ssh is recommended for
300 its security features."
301
302=item Using pushing over the NFS
303
304Having the other systems mounted over the NFS, you can take an
3e148164
JH
305active pushing approach by checking the just updated tree against
306the other not-yet synced trees. An example would be
307
308 #!/usr/bin/perl -w
309
310 use strict;
311 use File::Copy;
312
313 my %MF = map {
314 m/(\S+)/;
315 $1 => [ (stat $1)[2, 7, 9] ]; # mode, size, mtime
316 } `cat MANIFEST`;
317
318 my %remote = map { $_ => "/$_/pro/3gl/CPAN/perl-5.7.1" } qw(host1 host2);
319
320 foreach my $host (keys %remote) {
321 unless (-d $remote{$host}) {
322 print STDERR "Cannot Xsync for host $host\n";
323 next;
324 }
325 foreach my $file (keys %MF) {
326 my $rfile = "$remote{$host}/$file";
327 my ($mode, $size, $mtime) = (stat $rfile)[2, 7, 9];
328 defined $size or ($mode, $size, $mtime) = (0, 0, 0);
329 $size == $MF{$file}[1] && $mtime == $MF{$file}[2] and next;
330 printf "%4s %-34s %8d %9d %8d %9d\n",
331 $host, $file, $MF{$file}[1], $MF{$file}[2], $size, $mtime;
332 unlink $rfile;
333 copy ($file, $rfile);
334 utime time, $MF{$file}[2], $rfile;
335 chmod $MF{$file}[0], $rfile;
336 }
337 }
338
339though this is not perfect. It could be improved with checking
a1f349fd
MB
340file checksums before updating. Not all NFS systems support
341reliable utime support (when used over the NFS).
342
343=back
344
345=item rsync'ing the patches
346
347The source tree is maintained by the pumpking who applies patches to
348the files in the tree. These patches are either created by the
349pumpking himself using C<diff -c> after updating the file manually or
350by applying patches sent in by posters on the perl5-porters list.
351These patches are also saved and rsync'able, so you can apply them
352yourself to the source files.
353
354Presuming you are in a directory where your patches reside, you can
3e148164 355get them in sync with
a1f349fd 356
eb8afeef 357 # rsync -avz rsync://public.activestate.com/perl-current-diffs/ .
a1f349fd
MB
358
359This makes sure the latest available patch is downloaded to your
360patch directory.
361
3e148164 362It's then up to you to apply these patches, using something like
a1f349fd 363
7d3c2c28 364 # last="`cat ../perl-current/.patch`.gz"
eb8afeef 365 # rsync -avz rsync://public.activestate.com/perl-current-diffs/ .
a1f349fd
MB
366 # find . -name '*.gz' -newer $last -exec gzcat {} \; >blead.patch
367 # cd ../perl-current
368 # patch -p1 -N <../perl-current-diffs/blead.patch
369
370or, since this is only a hint towards how it works, use CPAN-patchaperl
371from Andreas K├Ânig to have better control over the patching process.
372
373=back
374
f7e1e956 375=head2 Why rsync the source tree
a1f349fd
MB
376
377=over 4
378
10f58044 379=item It's easier to rsync the source tree
a1f349fd
MB
380
381Since you don't have to apply the patches yourself, you are sure all
382files in the source tree are in the right state.
383
a1f349fd
MB
384=item It's more reliable
385
0cfb3454
GS
386While both the rsync-able source and patch areas are automatically
387updated every few minutes, keep in mind that applying patches may
388sometimes mean careful hand-holding, especially if your version of
389the C<patch> program does not understand how to deal with new files,
390files with 8-bit characters, or files without trailing newlines.
a1f349fd
MB
391
392=back
393
f7e1e956 394=head2 Why rsync the patches
a1f349fd
MB
395
396=over 4
397
10f58044 398=item It's easier to rsync the patches
a1f349fd
MB
399
400If you have more than one machine that you want to keep in track with
3e148164 401bleadperl, it's easier to rsync the patches only once and then apply
a1f349fd
MB
402them to all the source trees on the different machines.
403
404In case you try to keep in pace on 5 different machines, for which
405only one of them has access to the WAN, rsync'ing all the source
3e148164 406trees should than be done 5 times over the NFS. Having
a1f349fd 407rsync'ed the patches only once, I can apply them to all the source
3e148164 408trees automatically. Need you say more ;-)
a1f349fd
MB
409
410=item It's a good reference
411
412If you do not only like to have the most recent development branch,
413but also like to B<fix> bugs, or extend features, you want to dive
414into the sources. If you are a seasoned perl core diver, you don't
415need no manuals, tips, roadmaps, perlguts.pod or other aids to find
416your way around. But if you are a starter, the patches may help you
417in finding where you should start and how to change the bits that
418bug you.
419
420The file B<Changes> is updated on occasions the pumpking sees as his
421own little sync points. On those occasions, he releases a tar-ball of
422the current source tree (i.e. perl@7582.tar.gz), which will be an
423excellent point to start with when choosing to use the 'rsync the
424patches' scheme. Starting with perl@7582, which means a set of source
425files on which the latest applied patch is number 7582, you apply all
f18956b7 426succeeding patches available from then on (7583, 7584, ...).
a1f349fd
MB
427
428You can use the patches later as a kind of search archive.
429
430=over 4
431
432=item Finding a start point
433
434If you want to fix/change the behaviour of function/feature Foo, just
435scan the patches for patches that mention Foo either in the subject,
3e148164 436the comments, or the body of the fix. A good chance the patch shows
a1f349fd
MB
437you the files that are affected by that patch which are very likely
438to be the starting point of your journey into the guts of perl.
439
440=item Finding how to fix a bug
441
442If you've found I<where> the function/feature Foo misbehaves, but you
443don't know how to fix it (but you do know the change you want to
444make), you can, again, peruse the patches for similar changes and
445look how others apply the fix.
446
447=item Finding the source of misbehaviour
448
449When you keep in sync with bleadperl, the pumpking would love to
3958b146 450I<see> that the community efforts really work. So after each of his
a1f349fd
MB
451sync points, you are to 'make test' to check if everything is still
452in working order. If it is, you do 'make ok', which will send an OK
453report to perlbug@perl.org. (If you do not have access to a mailer
3e148164 454from the system you just finished successfully 'make test', you can
a1f349fd
MB
455do 'make okfile', which creates the file C<perl.ok>, which you can
456than take to your favourite mailer and mail yourself).
457
3958b146 458But of course, as always, things will not always lead to a success
a1f349fd
MB
459path, and one or more test do not pass the 'make test'. Before
460sending in a bug report (using 'make nok' or 'make nokfile'), check
461the mailing list if someone else has reported the bug already and if
462so, confirm it by replying to that message. If not, you might want to
463trace the source of that misbehaviour B<before> sending in the bug,
464which will help all the other porters in finding the solution.
465
3e148164
JH
466Here the saved patches come in very handy. You can check the list of
467patches to see which patch changed what file and what change caused
468the misbehaviour. If you note that in the bug report, it saves the
469one trying to solve it, looking for that point.
a1f349fd
MB
470
471=back
472
473If searching the patches is too bothersome, you might consider using
474perl's bugtron to find more information about discussions and
475ramblings on posted bugs.
476
3e148164
JH
477If you want to get the best of both worlds, rsync both the source
478tree for convenience, reliability and ease and rsync the patches
479for reference.
480
52315700
RF
481=back
482
fcc89a64
DR
483=head2 Working with the source
484
485Because you cannot use the Perforce client, you cannot easily generate
486diffs against the repository, nor will merges occur when you update
487via rsync. If you edit a file locally and then rsync against the
488latest source, changes made in the remote copy will I<overwrite> your
489local versions!
490
491The best way to deal with this is to maintain a tree of symlinks to
492the rsync'd source. Then, when you want to edit a file, you remove
493the symlink, copy the real file into the other tree, and edit it. You
494can then diff your edited file against the original to generate a
495patch, and you can safely update the original tree.
496
497Perl's F<Configure> script can generate this tree of symlinks for you.
498The following example assumes that you have used rsync to pull a copy
499of the Perl source into the F<perl-rsync> directory. In the directory
500above that one, you can execute the following commands:
501
502 mkdir perl-dev
503 cd perl-dev
504 ../perl-rsync/Configure -Dmksymlinks -Dusedevel -D"optimize=-g"
505
506This will start the Perl configuration process. After a few prompts,
507you should see something like this:
508
509 Symbolic links are supported.
510
511 Checking how to test for symbolic links...
512 Your builtin 'test -h' may be broken.
513 Trying external '/usr/bin/test -h'.
514 You can test for symbolic links with '/usr/bin/test -h'.
515
516 Creating the symbolic links...
517 (First creating the subdirectories...)
518 (Then creating the symlinks...)
519
520The specifics may vary based on your operating system, of course.
521After you see this, you can abort the F<Configure> script, and you
522will see that the directory you are in has a tree of symlinks to the
523F<perl-rsync> directories and files.
524
525If you plan to do a lot of work with the Perl source, here are some
526Bourne shell script functions that can make your life easier:
527
528 function edit {
529 if [ -L $1 ]; then
530 mv $1 $1.orig
531 cp $1.orig $1
532 vi $1
533 else
534 /bin/vi $1
535 fi
536 }
537
538 function unedit {
539 if [ -L $1.orig ]; then
540 rm $1
541 mv $1.orig $1
542 fi
543 }
544
545Replace "vi" with your favorite flavor of editor.
546
547Here is another function which will quickly generate a patch for the
548files which have been edited in your symlink tree:
549
550 mkpatchorig() {
551 local diffopts
552 for f in `find . -name '*.orig' | sed s,^\./,,`
553 do
554 case `echo $f | sed 's,.orig$,,;s,.*\.,,'` in
555 c) diffopts=-p ;;
556 pod) diffopts='-F^=' ;;
557 *) diffopts= ;;
558 esac
559 diff -du $diffopts $f `echo $f | sed 's,.orig$,,'`
560 done
561 }
562
563This function produces patches which include enough context to make
564your changes obvious. This makes it easier for the Perl pumpking(s)
565to review them when you send them to the perl5-porters list, and that
566means they're more likely to get applied.
567
568This function assumed a GNU diff, and may require some tweaking for
569other diff variants.
52315700 570
3fd28c4e 571=head2 Perlbug administration
52315700 572
3fd28c4e
RF
573There is a single remote administrative interface for modifying bug status,
574category, open issues etc. using the B<RT> I<bugtracker> system, maintained
575by I<Robert Spier>. Become an administrator, and close any bugs you can get
576your sticky mitts on:
52315700 577
3fd28c4e 578 http://rt.perl.org
52315700 579
3fd28c4e 580The bugtracker mechanism for B<perl5> bugs in particular is at:
52315700 581
3fd28c4e 582 http://bugs6.perl.org/perlbug
52315700 583
3fd28c4e 584To email the bug system administrators:
52315700 585
3fd28c4e 586 "perlbug-admin" <perlbug-admin@perl.org>
52315700 587
52315700 588
a1f349fd
MB
589=head2 Submitting patches
590
f7e1e956
MS
591Always submit patches to I<perl5-porters@perl.org>. If you're
592patching a core module and there's an author listed, send the author a
593copy (see L<Patching a core module>). This lets other porters review
594your patch, which catches a surprising number of errors in patches.
595Either use the diff program (available in source code form from
f224927c 596ftp://ftp.gnu.org/pub/gnu/ , or use Johan Vromans' I<makepatch>
f7e1e956
MS
597(available from I<CPAN/authors/id/JV/>). Unified diffs are preferred,
598but context diffs are accepted. Do not send RCS-style diffs or diffs
599without context lines. More information is given in the
600I<Porting/patching.pod> file in the Perl source distribution. Please
824d470b
RGS
601patch against the latest B<development> version. (e.g., even if you're
602fixing a bug in the 5.8 track, patch against the latest B<development>
603version rsynced from rsync://public.activestate.com/perl-current/ )
604
605If changes are accepted, they are applied to the development branch. Then
606the 5.8 pumpking decides which of those patches is to be backported to the
607maint branch. Only patches that survive the heat of the development
f7e1e956
MS
608branch get applied to maintenance versions.
609
824d470b 610Your patch should also update the documentation and test suite. See
f7e1e956 611L<Writing a test>.
e8cd7eae 612
824d470b
RGS
613Patching documentation also follows the same order: if accepted, a patch
614is first applied to B<development>, and if relevant then it's backported
615to B<maintenance>. (With an exception for some patches that document
616behaviour that only appears in the maintenance branch, but which has
617changed in the development version.)
618
e8cd7eae
GS
619To report a bug in Perl, use the program I<perlbug> which comes with
620Perl (if you can't get Perl to work, send mail to the address
f18956b7 621I<perlbug@perl.org> or I<perlbug@perl.com>). Reporting bugs through
e8cd7eae 622I<perlbug> feeds into the automated bug-tracking system, access to
f224927c 623which is provided through the web at http://bugs.perl.org/ . It
e8cd7eae
GS
624often pays to check the archives of the perl5-porters mailing list to
625see whether the bug you're reporting has been reported before, and if
626so whether it was considered a bug. See above for the location of
627the searchable archives.
628
f224927c 629The CPAN testers ( http://testers.cpan.org/ ) are a group of
ba139f7d
JH
630volunteers who test CPAN modules on a variety of platforms. Perl
631Smokers ( http://archives.develooper.com/daily-build@perl.org/ )
632automatically tests Perl source releases on platforms with various
633configurations. Both efforts welcome volunteers.
e8cd7eae 634
e8cd7eae
GS
635It's a good idea to read and lurk for a while before chipping in.
636That way you'll get to see the dynamic of the conversations, learn the
637personalities of the players, and hopefully be better prepared to make
638a useful contribution when do you speak up.
639
640If after all this you still think you want to join the perl5-porters
f6c51b38
GS
641mailing list, send mail to I<perl5-porters-subscribe@perl.org>. To
642unsubscribe, send mail to I<perl5-porters-unsubscribe@perl.org>.
e8cd7eae 643
a422fd2d
SC
644To hack on the Perl guts, you'll need to read the following things:
645
646=over 3
647
648=item L<perlguts>
649
650This is of paramount importance, since it's the documentation of what
651goes where in the Perl source. Read it over a couple of times and it
652might start to make sense - don't worry if it doesn't yet, because the
653best way to study it is to read it in conjunction with poking at Perl
654source, and we'll do that later on.
655
656You might also want to look at Gisle Aas's illustrated perlguts -
657there's no guarantee that this will be absolutely up-to-date with the
658latest documentation in the Perl core, but the fundamentals will be
1577cd80 659right. ( http://gisle.aas.no/perl/illguts/ )
a422fd2d
SC
660
661=item L<perlxstut> and L<perlxs>
662
663A working knowledge of XSUB programming is incredibly useful for core
664hacking; XSUBs use techniques drawn from the PP code, the portion of the
665guts that actually executes a Perl program. It's a lot gentler to learn
666those techniques from simple examples and explanation than from the core
667itself.
668
669=item L<perlapi>
670
671The documentation for the Perl API explains what some of the internal
672functions do, as well as the many macros used in the source.
673
674=item F<Porting/pumpkin.pod>
675
676This is a collection of words of wisdom for a Perl porter; some of it is
677only useful to the pumpkin holder, but most of it applies to anyone
678wanting to go about Perl development.
679
680=item The perl5-porters FAQ
681
7d7d5695
RGS
682This should be available from http://simon-cozens.org/writings/p5p-faq ;
683alternatively, you can get the FAQ emailed to you by sending mail to
684C<perl5-porters-faq@perl.org>. It contains hints on reading perl5-porters,
685information on how perl5-porters works and how Perl development in general
686works.
a422fd2d
SC
687
688=back
689
690=head2 Finding Your Way Around
691
692Perl maintenance can be split into a number of areas, and certain people
693(pumpkins) will have responsibility for each area. These areas sometimes
694correspond to files or directories in the source kit. Among the areas are:
695
696=over 3
697
698=item Core modules
699
700Modules shipped as part of the Perl core live in the F<lib/> and F<ext/>
701subdirectories: F<lib/> is for the pure-Perl modules, and F<ext/>
702contains the core XS modules.
703
f7e1e956
MS
704=item Tests
705
706There are tests for nearly all the modules, built-ins and major bits
707of functionality. Test files all have a .t suffix. Module tests live
708in the F<lib/> and F<ext/> directories next to the module being
709tested. Others live in F<t/>. See L<Writing a test>
710
a422fd2d
SC
711=item Documentation
712
713Documentation maintenance includes looking after everything in the
714F<pod/> directory, (as well as contributing new documentation) and
715the documentation to the modules in core.
716
717=item Configure
718
719The configure process is the way we make Perl portable across the
720myriad of operating systems it supports. Responsibility for the
721configure, build and installation process, as well as the overall
722portability of the core code rests with the configure pumpkin - others
723help out with individual operating systems.
724
725The files involved are the operating system directories, (F<win32/>,
726F<os2/>, F<vms/> and so on) the shell scripts which generate F<config.h>
727and F<Makefile>, as well as the metaconfig files which generate
728F<Configure>. (metaconfig isn't included in the core distribution.)
729
730=item Interpreter
731
732And of course, there's the core of the Perl interpreter itself. Let's
733have a look at that in a little more detail.
734
735=back
736
737Before we leave looking at the layout, though, don't forget that
738F<MANIFEST> contains not only the file names in the Perl distribution,
739but short descriptions of what's in them, too. For an overview of the
740important files, try this:
741
742 perl -lne 'print if /^[^\/]+\.[ch]\s+/' MANIFEST
743
744=head2 Elements of the interpreter
745
746The work of the interpreter has two main stages: compiling the code
747into the internal representation, or bytecode, and then executing it.
748L<perlguts/Compiled code> explains exactly how the compilation stage
749happens.
750
751Here is a short breakdown of perl's operation:
752
753=over 3
754
755=item Startup
756
757The action begins in F<perlmain.c>. (or F<miniperlmain.c> for miniperl)
758This is very high-level code, enough to fit on a single screen, and it
759resembles the code found in L<perlembed>; most of the real action takes
760place in F<perl.c>
761
762First, F<perlmain.c> allocates some memory and constructs a Perl
763interpreter:
764
765 1 PERL_SYS_INIT3(&argc,&argv,&env);
766 2
767 3 if (!PL_do_undump) {
768 4 my_perl = perl_alloc();
769 5 if (!my_perl)
770 6 exit(1);
771 7 perl_construct(my_perl);
772 8 PL_perl_destruct_level = 0;
773 9 }
774
775Line 1 is a macro, and its definition is dependent on your operating
776system. Line 3 references C<PL_do_undump>, a global variable - all
777global variables in Perl start with C<PL_>. This tells you whether the
778current running program was created with the C<-u> flag to perl and then
779F<undump>, which means it's going to be false in any sane context.
780
781Line 4 calls a function in F<perl.c> to allocate memory for a Perl
782interpreter. It's quite a simple function, and the guts of it looks like
783this:
784
785 my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));
786
787Here you see an example of Perl's system abstraction, which we'll see
788later: C<PerlMem_malloc> is either your system's C<malloc>, or Perl's
789own C<malloc> as defined in F<malloc.c> if you selected that option at
790configure time.
791
792Next, in line 7, we construct the interpreter; this sets up all the
793special variables that Perl needs, the stacks, and so on.
794
795Now we pass Perl the command line options, and tell it to go:
796
797 exitstatus = perl_parse(my_perl, xs_init, argc, argv, (char **)NULL);
798 if (!exitstatus) {
799 exitstatus = perl_run(my_perl);
800 }
801
802
803C<perl_parse> is actually a wrapper around C<S_parse_body>, as defined
804in F<perl.c>, which processes the command line options, sets up any
805statically linked XS modules, opens the program and calls C<yyparse> to
806parse it.
807
808=item Parsing
809
810The aim of this stage is to take the Perl source, and turn it into an op
811tree. We'll see what one of those looks like later. Strictly speaking,
812there's three things going on here.
813
814C<yyparse>, the parser, lives in F<perly.c>, although you're better off
815reading the original YACC input in F<perly.y>. (Yes, Virginia, there
816B<is> a YACC grammar for Perl!) The job of the parser is to take your
b432a672 817code and "understand" it, splitting it into sentences, deciding which
a422fd2d
SC
818operands go with which operators and so on.
819
820The parser is nobly assisted by the lexer, which chunks up your input
821into tokens, and decides what type of thing each token is: a variable
822name, an operator, a bareword, a subroutine, a core function, and so on.
823The main point of entry to the lexer is C<yylex>, and that and its
824associated routines can be found in F<toke.c>. Perl isn't much like
825other computer languages; it's highly context sensitive at times, it can
826be tricky to work out what sort of token something is, or where a token
827ends. As such, there's a lot of interplay between the tokeniser and the
828parser, which can get pretty frightening if you're not used to it.
829
830As the parser understands a Perl program, it builds up a tree of
831operations for the interpreter to perform during execution. The routines
832which construct and link together the various operations are to be found
833in F<op.c>, and will be examined later.
834
835=item Optimization
836
837Now the parsing stage is complete, and the finished tree represents
838the operations that the Perl interpreter needs to perform to execute our
839program. Next, Perl does a dry run over the tree looking for
840optimisations: constant expressions such as C<3 + 4> will be computed
841now, and the optimizer will also see if any multiple operations can be
842replaced with a single one. For instance, to fetch the variable C<$foo>,
843instead of grabbing the glob C<*foo> and looking at the scalar
844component, the optimizer fiddles the op tree to use a function which
845directly looks up the scalar in question. The main optimizer is C<peep>
846in F<op.c>, and many ops have their own optimizing functions.
847
848=item Running
849
850Now we're finally ready to go: we have compiled Perl byte code, and all
851that's left to do is run it. The actual execution is done by the
852C<runops_standard> function in F<run.c>; more specifically, it's done by
853these three innocent looking lines:
854
855 while ((PL_op = CALL_FPTR(PL_op->op_ppaddr)(aTHX))) {
856 PERL_ASYNC_CHECK();
857 }
858
859You may be more comfortable with the Perl version of that:
860
861 PERL_ASYNC_CHECK() while $Perl::op = &{$Perl::op->{function}};
862
863Well, maybe not. Anyway, each op contains a function pointer, which
864stipulates the function which will actually carry out the operation.
865This function will return the next op in the sequence - this allows for
866things like C<if> which choose the next op dynamically at run time.
867The C<PERL_ASYNC_CHECK> makes sure that things like signals interrupt
868execution if required.
869
870The actual functions called are known as PP code, and they're spread
b432a672 871between four files: F<pp_hot.c> contains the "hot" code, which is most
a422fd2d
SC
872often used and highly optimized, F<pp_sys.c> contains all the
873system-specific functions, F<pp_ctl.c> contains the functions which
874implement control structures (C<if>, C<while> and the like) and F<pp.c>
875contains everything else. These are, if you like, the C code for Perl's
876built-in functions and operators.
877
dfc98234
DM
878Note that each C<pp_> function is expected to return a pointer to the next
879op. Calls to perl subs (and eval blocks) are handled within the same
880runops loop, and do not consume extra space on the C stack. For example,
881C<pp_entersub> and C<pp_entertry> just push a C<CxSUB> or C<CxEVAL> block
882struct onto the context stack which contain the address of the op
883following the sub call or eval. They then return the first op of that sub
884or eval block, and so execution continues of that sub or block. Later, a
885C<pp_leavesub> or C<pp_leavetry> op pops the C<CxSUB> or C<CxEVAL>,
886retrieves the return op from it, and returns it.
887
888=item Exception handing
889
28a5cf3b 890Perl's exception handing (i.e. C<die> etc) is built on top of the low-level
dfc98234 891C<setjmp()>/C<longjmp()> C-library functions. These basically provide a
28a5cf3b 892way to capture the current PC and SP registers and later restore them; i.e.
dfc98234
DM
893a C<longjmp()> continues at the point in code where a previous C<setjmp()>
894was done, with anything further up on the C stack being lost. This is why
895code should always save values using C<SAVE_FOO> rather than in auto
896variables.
897
898The perl core wraps C<setjmp()> etc in the macros C<JMPENV_PUSH> and
899C<JMPENV_JUMP>. The basic rule of perl exceptions is that C<exit>, and
900C<die> (in the absence of C<eval>) perform a C<JMPENV_JUMP(2)>, while
901C<die> within C<eval> does a C<JMPENV_JUMP(3)>.
902
903At entry points to perl, such as C<perl_parse()>, C<perl_run()> and
904C<call_sv(cv, G_EVAL)> each does a C<JMPENV_PUSH>, then enter a runops
905loop or whatever, and handle possible exception returns. For a 2 return,
906final cleanup is performed, such as popping stacks and calling C<CHECK> or
907C<END> blocks. Amongst other things, this is how scope cleanup still
908occurs during an C<exit>.
909
910If a C<die> can find a C<CxEVAL> block on the context stack, then the
911stack is popped to that level and the return op in that block is assigned
912to C<PL_restartop>; then a C<JMPENV_JUMP(3)> is performed. This normally
913passes control back to the guard. In the case of C<perl_run> and
914C<call_sv>, a non-null C<PL_restartop> triggers re-entry to the runops
915loop. The is the normal way that C<die> or C<croak> is handled within an
916C<eval>.
917
918Sometimes ops are executed within an inner runops loop, such as tie, sort
919or overload code. In this case, something like
920
921 sub FETCH { eval { die } }
922
923would cause a longjmp right back to the guard in C<perl_run>, popping both
924runops loops, which is clearly incorrect. One way to avoid this is for the
925tie code to do a C<JMPENV_PUSH> before executing C<FETCH> in the inner
926runops loop, but for efficiency reasons, perl in fact just sets a flag,
927using C<CATCH_SET(TRUE)>. The C<pp_require>, C<pp_entereval> and
928C<pp_entertry> ops check this flag, and if true, they call C<docatch>,
929which does a C<JMPENV_PUSH> and starts a new runops level to execute the
930code, rather than doing it on the current loop.
931
932As a further optimisation, on exit from the eval block in the C<FETCH>,
933execution of the code following the block is still carried on in the inner
934loop. When an exception is raised, C<docatch> compares the C<JMPENV>
935level of the C<CxEVAL> with C<PL_top_env> and if they differ, just
936re-throws the exception. In this way any inner loops get popped.
937
938Here's an example.
939
940 1: eval { tie @a, 'A' };
941 2: sub A::TIEARRAY {
942 3: eval { die };
943 4: die;
944 5: }
945
946To run this code, C<perl_run> is called, which does a C<JMPENV_PUSH> then
947enters a runops loop. This loop executes the eval and tie ops on line 1,
948with the eval pushing a C<CxEVAL> onto the context stack.
949
950The C<pp_tie> does a C<CATCH_SET(TRUE)>, then starts a second runops loop
951to execute the body of C<TIEARRAY>. When it executes the entertry op on
952line 3, C<CATCH_GET> is true, so C<pp_entertry> calls C<docatch> which
953does a C<JMPENV_PUSH> and starts a third runops loop, which then executes
954the die op. At this point the C call stack looks like this:
955
956 Perl_pp_die
957 Perl_runops # third loop
958 S_docatch_body
959 S_docatch
960 Perl_pp_entertry
961 Perl_runops # second loop
962 S_call_body
963 Perl_call_sv
964 Perl_pp_tie
965 Perl_runops # first loop
966 S_run_body
967 perl_run
968 main
969
970and the context and data stacks, as shown by C<-Dstv>, look like:
971
972 STACK 0: MAIN
973 CX 0: BLOCK =>
974 CX 1: EVAL => AV() PV("A"\0)
975 retop=leave
976 STACK 1: MAGIC
977 CX 0: SUB =>
978 retop=(null)
979 CX 1: EVAL => *
980 retop=nextstate
981
982The die pops the first C<CxEVAL> off the context stack, sets
983C<PL_restartop> from it, does a C<JMPENV_JUMP(3)>, and control returns to
984the top C<docatch>. This then starts another third-level runops level,
985which executes the nextstate, pushmark and die ops on line 4. At the point
986that the second C<pp_die> is called, the C call stack looks exactly like
987that above, even though we are no longer within an inner eval; this is
988because of the optimization mentioned earlier. However, the context stack
989now looks like this, ie with the top CxEVAL popped:
990
991 STACK 0: MAIN
992 CX 0: BLOCK =>
993 CX 1: EVAL => AV() PV("A"\0)
994 retop=leave
995 STACK 1: MAGIC
996 CX 0: SUB =>
997 retop=(null)
998
999The die on line 4 pops the context stack back down to the CxEVAL, leaving
1000it as:
1001
1002 STACK 0: MAIN
1003 CX 0: BLOCK =>
1004
1005As usual, C<PL_restartop> is extracted from the C<CxEVAL>, and a
1006C<JMPENV_JUMP(3)> done, which pops the C stack back to the docatch:
1007
1008 S_docatch
1009 Perl_pp_entertry
1010 Perl_runops # second loop
1011 S_call_body
1012 Perl_call_sv
1013 Perl_pp_tie
1014 Perl_runops # first loop
1015 S_run_body
1016 perl_run
1017 main
1018
1019In this case, because the C<JMPENV> level recorded in the C<CxEVAL>
1020differs from the current one, C<docatch> just does a C<JMPENV_JUMP(3)>
1021and the C stack unwinds to:
1022
1023 perl_run
1024 main
1025
1026Because C<PL_restartop> is non-null, C<run_body> starts a new runops loop
1027and execution continues.
1028
a422fd2d
SC
1029=back
1030
1031=head2 Internal Variable Types
1032
1033You should by now have had a look at L<perlguts>, which tells you about
1034Perl's internal variable types: SVs, HVs, AVs and the rest. If not, do
1035that now.
1036
1037These variables are used not only to represent Perl-space variables, but
1038also any constants in the code, as well as some structures completely
1039internal to Perl. The symbol table, for instance, is an ordinary Perl
1040hash. Your code is represented by an SV as it's read into the parser;
1041any program files you call are opened via ordinary Perl filehandles, and
1042so on.
1043
1044The core L<Devel::Peek|Devel::Peek> module lets us examine SVs from a
1045Perl program. Let's see, for instance, how Perl treats the constant
1046C<"hello">.
1047
1048 % perl -MDevel::Peek -e 'Dump("hello")'
1049 1 SV = PV(0xa041450) at 0xa04ecbc
1050 2 REFCNT = 1
1051 3 FLAGS = (POK,READONLY,pPOK)
1052 4 PV = 0xa0484e0 "hello"\0
1053 5 CUR = 5
1054 6 LEN = 6
1055
1056Reading C<Devel::Peek> output takes a bit of practise, so let's go
1057through it line by line.
1058
1059Line 1 tells us we're looking at an SV which lives at C<0xa04ecbc> in
1060memory. SVs themselves are very simple structures, but they contain a
1061pointer to a more complex structure. In this case, it's a PV, a
1062structure which holds a string value, at location C<0xa041450>. Line 2
1063is the reference count; there are no other references to this data, so
1064it's 1.
1065
1066Line 3 are the flags for this SV - it's OK to use it as a PV, it's a
1067read-only SV (because it's a constant) and the data is a PV internally.
1068Next we've got the contents of the string, starting at location
1069C<0xa0484e0>.
1070
1071Line 5 gives us the current length of the string - note that this does
1072B<not> include the null terminator. Line 6 is not the length of the
1073string, but the length of the currently allocated buffer; as the string
1074grows, Perl automatically extends the available storage via a routine
1075called C<SvGROW>.
1076
1077You can get at any of these quantities from C very easily; just add
1078C<Sv> to the name of the field shown in the snippet, and you've got a
1079macro which will return the value: C<SvCUR(sv)> returns the current
1080length of the string, C<SvREFCOUNT(sv)> returns the reference count,
1081C<SvPV(sv, len)> returns the string itself with its length, and so on.
1082More macros to manipulate these properties can be found in L<perlguts>.
1083
1084Let's take an example of manipulating a PV, from C<sv_catpvn>, in F<sv.c>
1085
1086 1 void
1087 2 Perl_sv_catpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
1088 3 {
1089 4 STRLEN tlen;
1090 5 char *junk;
1091
1092 6 junk = SvPV_force(sv, tlen);
1093 7 SvGROW(sv, tlen + len + 1);
1094 8 if (ptr == junk)
1095 9 ptr = SvPVX(sv);
1096 10 Move(ptr,SvPVX(sv)+tlen,len,char);
1097 11 SvCUR(sv) += len;
1098 12 *SvEND(sv) = '\0';
1099 13 (void)SvPOK_only_UTF8(sv); /* validate pointer */
1100 14 SvTAINT(sv);
1101 15 }
1102
1103This is a function which adds a string, C<ptr>, of length C<len> onto
1104the end of the PV stored in C<sv>. The first thing we do in line 6 is
1105make sure that the SV B<has> a valid PV, by calling the C<SvPV_force>
1106macro to force a PV. As a side effect, C<tlen> gets set to the current
1107value of the PV, and the PV itself is returned to C<junk>.
1108
b1866b2d 1109In line 7, we make sure that the SV will have enough room to accommodate
a422fd2d
SC
1110the old string, the new string and the null terminator. If C<LEN> isn't
1111big enough, C<SvGROW> will reallocate space for us.
1112
1113Now, if C<junk> is the same as the string we're trying to add, we can
1114grab the string directly from the SV; C<SvPVX> is the address of the PV
1115in the SV.
1116
1117Line 10 does the actual catenation: the C<Move> macro moves a chunk of
1118memory around: we move the string C<ptr> to the end of the PV - that's
1119the start of the PV plus its current length. We're moving C<len> bytes
1120of type C<char>. After doing so, we need to tell Perl we've extended the
1121string, by altering C<CUR> to reflect the new length. C<SvEND> is a
1122macro which gives us the end of the string, so that needs to be a
1123C<"\0">.
1124
1125Line 13 manipulates the flags; since we've changed the PV, any IV or NV
1126values will no longer be valid: if we have C<$a=10; $a.="6";> we don't
1e54db1a 1127want to use the old IV of 10. C<SvPOK_only_utf8> is a special UTF-8-aware
a422fd2d
SC
1128version of C<SvPOK_only>, a macro which turns off the IOK and NOK flags
1129and turns on POK. The final C<SvTAINT> is a macro which launders tainted
1130data if taint mode is turned on.
1131
1132AVs and HVs are more complicated, but SVs are by far the most common
1133variable type being thrown around. Having seen something of how we
1134manipulate these, let's go on and look at how the op tree is
1135constructed.
1136
1137=head2 Op Trees
1138
1139First, what is the op tree, anyway? The op tree is the parsed
1140representation of your program, as we saw in our section on parsing, and
1141it's the sequence of operations that Perl goes through to execute your
1142program, as we saw in L</Running>.
1143
1144An op is a fundamental operation that Perl can perform: all the built-in
1145functions and operators are ops, and there are a series of ops which
1146deal with concepts the interpreter needs internally - entering and
1147leaving a block, ending a statement, fetching a variable, and so on.
1148
1149The op tree is connected in two ways: you can imagine that there are two
1150"routes" through it, two orders in which you can traverse the tree.
1151First, parse order reflects how the parser understood the code, and
1152secondly, execution order tells perl what order to perform the
1153operations in.
1154
1155The easiest way to examine the op tree is to stop Perl after it has
1156finished parsing, and get it to dump out the tree. This is exactly what
7d7d5695
RGS
1157the compiler backends L<B::Terse|B::Terse>, L<B::Concise|B::Concise>
1158and L<B::Debug|B::Debug> do.
a422fd2d
SC
1159
1160Let's have a look at how Perl sees C<$a = $b + $c>:
1161
1162 % perl -MO=Terse -e '$a=$b+$c'
1163 1 LISTOP (0x8179888) leave
1164 2 OP (0x81798b0) enter
1165 3 COP (0x8179850) nextstate
1166 4 BINOP (0x8179828) sassign
1167 5 BINOP (0x8179800) add [1]
1168 6 UNOP (0x81796e0) null [15]
1169 7 SVOP (0x80fafe0) gvsv GV (0x80fa4cc) *b
1170 8 UNOP (0x81797e0) null [15]
1171 9 SVOP (0x8179700) gvsv GV (0x80efeb0) *c
1172 10 UNOP (0x816b4f0) null [15]
1173 11 SVOP (0x816dcf0) gvsv GV (0x80fa460) *a
1174
1175Let's start in the middle, at line 4. This is a BINOP, a binary
1176operator, which is at location C<0x8179828>. The specific operator in
1177question is C<sassign> - scalar assignment - and you can find the code
1178which implements it in the function C<pp_sassign> in F<pp_hot.c>. As a
1179binary operator, it has two children: the add operator, providing the
1180result of C<$b+$c>, is uppermost on line 5, and the left hand side is on
1181line 10.
1182
1183Line 10 is the null op: this does exactly nothing. What is that doing
1184there? If you see the null op, it's a sign that something has been
1185optimized away after parsing. As we mentioned in L</Optimization>,
1186the optimization stage sometimes converts two operations into one, for
1187example when fetching a scalar variable. When this happens, instead of
1188rewriting the op tree and cleaning up the dangling pointers, it's easier
1189just to replace the redundant operation with the null op. Originally,
1190the tree would have looked like this:
1191
1192 10 SVOP (0x816b4f0) rv2sv [15]
1193 11 SVOP (0x816dcf0) gv GV (0x80fa460) *a
1194
1195That is, fetch the C<a> entry from the main symbol table, and then look
1196at the scalar component of it: C<gvsv> (C<pp_gvsv> into F<pp_hot.c>)
1197happens to do both these things.
1198
1199The right hand side, starting at line 5 is similar to what we've just
1200seen: we have the C<add> op (C<pp_add> also in F<pp_hot.c>) add together
1201two C<gvsv>s.
1202
1203Now, what's this about?
1204
1205 1 LISTOP (0x8179888) leave
1206 2 OP (0x81798b0) enter
1207 3 COP (0x8179850) nextstate
1208
1209C<enter> and C<leave> are scoping ops, and their job is to perform any
1210housekeeping every time you enter and leave a block: lexical variables
1211are tidied up, unreferenced variables are destroyed, and so on. Every
1212program will have those first three lines: C<leave> is a list, and its
1213children are all the statements in the block. Statements are delimited
1214by C<nextstate>, so a block is a collection of C<nextstate> ops, with
1215the ops to be performed for each statement being the children of
1216C<nextstate>. C<enter> is a single op which functions as a marker.
1217
1218That's how Perl parsed the program, from top to bottom:
1219
1220 Program
1221 |
1222 Statement
1223 |
1224 =
1225 / \
1226 / \
1227 $a +
1228 / \
1229 $b $c
1230
1231However, it's impossible to B<perform> the operations in this order:
1232you have to find the values of C<$b> and C<$c> before you add them
1233together, for instance. So, the other thread that runs through the op
1234tree is the execution order: each op has a field C<op_next> which points
1235to the next op to be run, so following these pointers tells us how perl
1236executes the code. We can traverse the tree in this order using
1237the C<exec> option to C<B::Terse>:
1238
1239 % perl -MO=Terse,exec -e '$a=$b+$c'
1240 1 OP (0x8179928) enter
1241 2 COP (0x81798c8) nextstate
1242 3 SVOP (0x81796c8) gvsv GV (0x80fa4d4) *b
1243 4 SVOP (0x8179798) gvsv GV (0x80efeb0) *c
1244 5 BINOP (0x8179878) add [1]
1245 6 SVOP (0x816dd38) gvsv GV (0x80fa468) *a
1246 7 BINOP (0x81798a0) sassign
1247 8 LISTOP (0x8179900) leave
1248
1249This probably makes more sense for a human: enter a block, start a
1250statement. Get the values of C<$b> and C<$c>, and add them together.
1251Find C<$a>, and assign one to the other. Then leave.
1252
1253The way Perl builds up these op trees in the parsing process can be
1254unravelled by examining F<perly.y>, the YACC grammar. Let's take the
1255piece we need to construct the tree for C<$a = $b + $c>
1256
1257 1 term : term ASSIGNOP term
1258 2 { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
1259 3 | term ADDOP term
1260 4 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
1261
1262If you're not used to reading BNF grammars, this is how it works: You're
1263fed certain things by the tokeniser, which generally end up in upper
1264case. Here, C<ADDOP>, is provided when the tokeniser sees C<+> in your
1265code. C<ASSIGNOP> is provided when C<=> is used for assigning. These are
b432a672 1266"terminal symbols", because you can't get any simpler than them.
a422fd2d
SC
1267
1268The grammar, lines one and three of the snippet above, tells you how to
b432a672 1269build up more complex forms. These complex forms, "non-terminal symbols"
a422fd2d
SC
1270are generally placed in lower case. C<term> here is a non-terminal
1271symbol, representing a single expression.
1272
1273The grammar gives you the following rule: you can make the thing on the
1274left of the colon if you see all the things on the right in sequence.
1275This is called a "reduction", and the aim of parsing is to completely
1276reduce the input. There are several different ways you can perform a
1277reduction, separated by vertical bars: so, C<term> followed by C<=>
1278followed by C<term> makes a C<term>, and C<term> followed by C<+>
1279followed by C<term> can also make a C<term>.
1280
1281So, if you see two terms with an C<=> or C<+>, between them, you can
1282turn them into a single expression. When you do this, you execute the
1283code in the block on the next line: if you see C<=>, you'll do the code
1284in line 2. If you see C<+>, you'll do the code in line 4. It's this code
1285which contributes to the op tree.
1286
1287 | term ADDOP term
1288 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
1289
1290What this does is creates a new binary op, and feeds it a number of
1291variables. The variables refer to the tokens: C<$1> is the first token in
1292the input, C<$2> the second, and so on - think regular expression
1293backreferences. C<$$> is the op returned from this reduction. So, we
1294call C<newBINOP> to create a new binary operator. The first parameter to
1295C<newBINOP>, a function in F<op.c>, is the op type. It's an addition
1296operator, so we want the type to be C<ADDOP>. We could specify this
1297directly, but it's right there as the second token in the input, so we
b432a672
AL
1298use C<$2>. The second parameter is the op's flags: 0 means "nothing
1299special". Then the things to add: the left and right hand side of our
a422fd2d
SC
1300expression, in scalar context.
1301
1302=head2 Stacks
1303
1304When perl executes something like C<addop>, how does it pass on its
1305results to the next op? The answer is, through the use of stacks. Perl
1306has a number of stacks to store things it's currently working on, and
1307we'll look at the three most important ones here.
1308
1309=over 3
1310
1311=item Argument stack
1312
1313Arguments are passed to PP code and returned from PP code using the
1314argument stack, C<ST>. The typical way to handle arguments is to pop
1315them off the stack, deal with them how you wish, and then push the result
1316back onto the stack. This is how, for instance, the cosine operator
1317works:
1318
1319 NV value;
1320 value = POPn;
1321 value = Perl_cos(value);
1322 XPUSHn(value);
1323
1324We'll see a more tricky example of this when we consider Perl's macros
1325below. C<POPn> gives you the NV (floating point value) of the top SV on
1326the stack: the C<$x> in C<cos($x)>. Then we compute the cosine, and push
1327the result back as an NV. The C<X> in C<XPUSHn> means that the stack
1328should be extended if necessary - it can't be necessary here, because we
1329know there's room for one more item on the stack, since we've just
1330removed one! The C<XPUSH*> macros at least guarantee safety.
1331
1332Alternatively, you can fiddle with the stack directly: C<SP> gives you
1333the first element in your portion of the stack, and C<TOP*> gives you
1334the top SV/IV/NV/etc. on the stack. So, for instance, to do unary
1335negation of an integer:
1336
1337 SETi(-TOPi);
1338
1339Just set the integer value of the top stack entry to its negation.
1340
1341Argument stack manipulation in the core is exactly the same as it is in
1342XSUBs - see L<perlxstut>, L<perlxs> and L<perlguts> for a longer
1343description of the macros used in stack manipulation.
1344
1345=item Mark stack
1346
b432a672 1347I say "your portion of the stack" above because PP code doesn't
a422fd2d
SC
1348necessarily get the whole stack to itself: if your function calls
1349another function, you'll only want to expose the arguments aimed for the
1350called function, and not (necessarily) let it get at your own data. The
b432a672 1351way we do this is to have a "virtual" bottom-of-stack, exposed to each
a422fd2d
SC
1352function. The mark stack keeps bookmarks to locations in the argument
1353stack usable by each function. For instance, when dealing with a tied
b432a672 1354variable, (internally, something with "P" magic) Perl has to call
a422fd2d
SC
1355methods for accesses to the tied variables. However, we need to separate
1356the arguments exposed to the method to the argument exposed to the
ed233832
DM
1357original function - the store or fetch or whatever it may be. Here's
1358roughly how the tied C<push> is implemented; see C<av_push> in F<av.c>:
a422fd2d
SC
1359
1360 1 PUSHMARK(SP);
1361 2 EXTEND(SP,2);
1362 3 PUSHs(SvTIED_obj((SV*)av, mg));
1363 4 PUSHs(val);
1364 5 PUTBACK;
1365 6 ENTER;
1366 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1367 8 LEAVE;
13a2d996 1368
a422fd2d
SC
1369Let's examine the whole implementation, for practice:
1370
1371 1 PUSHMARK(SP);
1372
1373Push the current state of the stack pointer onto the mark stack. This is
1374so that when we've finished adding items to the argument stack, Perl
1375knows how many things we've added recently.
1376
1377 2 EXTEND(SP,2);
1378 3 PUSHs(SvTIED_obj((SV*)av, mg));
1379 4 PUSHs(val);
1380
1381We're going to add two more items onto the argument stack: when you have
1382a tied array, the C<PUSH> subroutine receives the object and the value
1383to be pushed, and that's exactly what we have here - the tied object,
1384retrieved with C<SvTIED_obj>, and the value, the SV C<val>.
1385
1386 5 PUTBACK;
1387
e89a6d4e
JD
1388Next we tell Perl to update the global stack pointer from our internal
1389variable: C<dSP> only gave us a local copy, not a reference to the global.
a422fd2d
SC
1390
1391 6 ENTER;
1392 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1393 8 LEAVE;
1394
1395C<ENTER> and C<LEAVE> localise a block of code - they make sure that all
1396variables are tidied up, everything that has been localised gets
1397its previous value returned, and so on. Think of them as the C<{> and
1398C<}> of a Perl block.
1399
1400To actually do the magic method call, we have to call a subroutine in
1401Perl space: C<call_method> takes care of that, and it's described in
1402L<perlcall>. We call the C<PUSH> method in scalar context, and we're
e89a6d4e
JD
1403going to discard its return value. The call_method() function
1404removes the top element of the mark stack, so there is nothing for
1405the caller to clean up.
a422fd2d 1406
a422fd2d
SC
1407=item Save stack
1408
1409C doesn't have a concept of local scope, so perl provides one. We've
1410seen that C<ENTER> and C<LEAVE> are used as scoping braces; the save
1411stack implements the C equivalent of, for example:
1412
1413 {
1414 local $foo = 42;
1415 ...
1416 }
1417
1418See L<perlguts/Localising Changes> for how to use the save stack.
1419
1420=back
1421
1422=head2 Millions of Macros
1423
1424One thing you'll notice about the Perl source is that it's full of
1425macros. Some have called the pervasive use of macros the hardest thing
1426to understand, others find it adds to clarity. Let's take an example,
1427the code which implements the addition operator:
1428
1429 1 PP(pp_add)
1430 2 {
39644a26 1431 3 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
a422fd2d
SC
1432 4 {
1433 5 dPOPTOPnnrl_ul;
1434 6 SETn( left + right );
1435 7 RETURN;
1436 8 }
1437 9 }
1438
1439Every line here (apart from the braces, of course) contains a macro. The
1440first line sets up the function declaration as Perl expects for PP code;
1441line 3 sets up variable declarations for the argument stack and the
1442target, the return value of the operation. Finally, it tries to see if
1443the addition operation is overloaded; if so, the appropriate subroutine
1444is called.
1445
1446Line 5 is another variable declaration - all variable declarations start
1447with C<d> - which pops from the top of the argument stack two NVs (hence
1448C<nn>) and puts them into the variables C<right> and C<left>, hence the
1449C<rl>. These are the two operands to the addition operator. Next, we
1450call C<SETn> to set the NV of the return value to the result of adding
1451the two values. This done, we return - the C<RETURN> macro makes sure
1452that our return value is properly handled, and we pass the next operator
1453to run back to the main run loop.
1454
1455Most of these macros are explained in L<perlapi>, and some of the more
1456important ones are explained in L<perlxs> as well. Pay special attention
1457to L<perlguts/Background and PERL_IMPLICIT_CONTEXT> for information on
1458the C<[pad]THX_?> macros.
1459
52d59bef
JH
1460=head2 The .i Targets
1461
1462You can expand the macros in a F<foo.c> file by saying
1463
1464 make foo.i
1465
1466which will expand the macros using cpp. Don't be scared by the results.
1467
955fec6b
JH
1468=head1 SOURCE CODE STATIC ANALYSIS
1469
1470Various tools exist for analysing C source code B<statically>, as
1471opposed to B<dynamically>, that is, without executing the code.
1472It is possible to detect resource leaks, undefined behaviour, type
1473mismatches, portability problems, code paths that would cause illegal
1474memory accesses, and other similar problems by just parsing the C code
1475and looking at the resulting graph, what does it tell about the
1476execution and data flows. As a matter of fact, this is exactly
1477how C compilers know to give warnings about dubious code.
1478
1479=head2 lint, splint
1480
1481The good old C code quality inspector, C<lint>, is available in
1482several platforms, but please be aware that there are several
1483different implementations of it by different vendors, which means that
1484the flags are not identical across different platforms.
1485
1486There is a lint variant called C<splint> (Secure Programming Lint)
1487available from http://www.splint.org/ that should compile on any
1488Unix-like platform.
1489
1490There are C<lint> and <splint> targets in Makefile, but you may have
1491to diddle with the flags (see above).
1492
1493=head2 Coverity
1494
1495Coverity (http://www.coverity.com/) is a product similar to lint and
1496as a testbed for their product they periodically check several open
1497source projects, and they give out accounts to open source developers
1498to the defect databases.
1499
1500=head2 cpd (cut-and-paste detector)
1501
1502The cpd tool detects cut-and-paste coding. If one instance of the
1503cut-and-pasted code changes, all the other spots should probably be
1504changed, too. Therefore such code should probably be turned into a
1505subroutine or a macro.
1506
1507cpd (http://pmd.sourceforge.net/cpd.html) is part of the pmd project
1508(http://pmd.sourceforge.net/). pmd was originally written for static
1509analysis of Java code, but later the cpd part of it was extended to
1510parse also C and C++.
1511
1512Download the pmd-X.y.jar from the SourceForge site, and then run
1513it on source code thusly:
1514
1515 java -cp pmd-X.Y.jar net.sourceforge.pmd.cpd.CPD --minimum-tokens 100 --files /some/where/src --language c > cpd.txt
1516
1517You may run into memory limits, in which case you should use the -Xmx option:
1518
1519 java -Xmx512M ...
1520
1521=head2 gcc warnings
1522
1523Though much can be written about the inconsistency and coverage
1524problems of gcc warnings (like C<-Wall> not meaning "all the
1525warnings", or some common portability problems not being covered by
1526C<-Wall>, or C<-ansi> and C<-pedantic> both being a poorly defined
1527collection of warnings, and so forth), gcc is still a useful tool in
1528keeping our coding nose clean.
1529
1530The C<-Wall> is by default on.
1531
a8e98a71
JH
1532The C<-ansi> (and its sidekick, C<-pedantic>) would be nice to be on
1533always, but unfortunately they are not safe on all platforms, they can
1534for example cause fatal conflicts with the system headers (Solaris
1535being a prime example). If Configure C<-Dgccansipedantic> is used,
1536the C<cflags> frontend selects C<-ansi -pedantic> for the platforms
1537where they are known to be safe.
955fec6b
JH
1538
1539Starting from Perl 5.9.4 the following extra flags are added:
1540
1541=over 4
1542
1543=item *
1544
1545C<-Wendif-labels>
1546
1547=item *
1548
1549C<-Wextra>
1550
1551=item *
1552
1553C<-Wdeclaration-after-statement>
1554
1555=back
1556
1557The following flags would be nice to have but they would first need
1558their own Stygian stablemaster:
1559
1560=over 4
1561
1562=item *
1563
1564C<-Wpointer-arith>
1565
1566=item *
1567
1568C<-Wshadow>
1569
1570=item *
1571
1572C<-Wstrict-prototypes>
1573
955fec6b
JH
1574=back
1575
1576The C<-Wtraditional> is another example of the annoying tendency of
1577gcc to bundle a lot of warnings under one switch -- it would be
1578impossible to deploy in practice because it would complain a lot -- but
1579it does contain some warnings that would be beneficial to have available
1580on their own, such as the warning about string constants inside macros
1581containing the macro arguments: this behaved differently pre-ANSI
1582than it does in ANSI, and some C compilers are still in transition,
1583AIX being an example.
1584
1585=head2 Warnings of other C compilers
1586
1587Other C compilers (yes, there B<are> other C compilers than gcc) often
1588have their "strict ANSI" or "strict ANSI with some portability extensions"
1589modes on, like for example the Sun Workshop has its C<-Xa> mode on
1590(though implicitly), or the DEC (these days, HP...) has its C<-std1>
1591mode on.
1592
1593=head2 DEBUGGING
1594
1595You can compile a special debugging version of Perl, which allows you
1596to use the C<-D> option of Perl to tell more about what Perl is doing.
1597But sometimes there is no alternative than to dive in with a debugger,
1598either to see the stack trace of a core dump (very useful in a bug
1599report), or trying to figure out what went wrong before the core dump
1600happened, or how did we end up having wrong or unexpected results.
1601
a422fd2d
SC
1602=head2 Poking at Perl
1603
1604To really poke around with Perl, you'll probably want to build Perl for
1605debugging, like this:
1606
1607 ./Configure -d -D optimize=-g
1608 make
1609
1610C<-g> is a flag to the C compiler to have it produce debugging
955fec6b
JH
1611information which will allow us to step through a running program,
1612and to see in which C function we are at (without the debugging
1613information we might see only the numerical addresses of the functions,
1614which is not very helpful).
1615
a422fd2d
SC
1616F<Configure> will also turn on the C<DEBUGGING> compilation symbol which
1617enables all the internal debugging code in Perl. There are a whole bunch
1618of things you can debug with this: L<perlrun> lists them all, and the
1619best way to find out about them is to play about with them. The most
1620useful options are probably
1621
1622 l Context (loop) stack processing
1623 t Trace execution
1624 o Method and overloading resolution
1625 c String/numeric conversions
1626
1627Some of the functionality of the debugging code can be achieved using XS
1628modules.
13a2d996 1629
a422fd2d
SC
1630 -Dr => use re 'debug'
1631 -Dx => use O 'Debug'
1632
1633=head2 Using a source-level debugger
1634
1635If the debugging output of C<-D> doesn't help you, it's time to step
1636through perl's execution with a source-level debugger.
1637
1638=over 3
1639
1640=item *
1641
955fec6b
JH
1642We'll use C<gdb> for our examples here; the principles will apply to
1643any debugger (many vendors call their debugger C<dbx>), but check the
1644manual of the one you're using.
a422fd2d
SC
1645
1646=back
1647
1648To fire up the debugger, type
1649
1650 gdb ./perl
1651
955fec6b
JH
1652Or if you have a core dump:
1653
1654 gdb ./perl core
1655
a422fd2d
SC
1656You'll want to do that in your Perl source tree so the debugger can read
1657the source code. You should see the copyright message, followed by the
1658prompt.
1659
1660 (gdb)
1661
1662C<help> will get you into the documentation, but here are the most
1663useful commands:
1664
1665=over 3
1666
1667=item run [args]
1668
1669Run the program with the given arguments.
1670
1671=item break function_name
1672
1673=item break source.c:xxx
1674
1675Tells the debugger that we'll want to pause execution when we reach
cea6626f 1676either the named function (but see L<perlguts/Internal Functions>!) or the given
a422fd2d
SC
1677line in the named source file.
1678
1679=item step
1680
1681Steps through the program a line at a time.
1682
1683=item next
1684
1685Steps through the program a line at a time, without descending into
1686functions.
1687
1688=item continue
1689
1690Run until the next breakpoint.
1691
1692=item finish
1693
1694Run until the end of the current function, then stop again.
1695
13a2d996 1696=item 'enter'
a422fd2d
SC
1697
1698Just pressing Enter will do the most recent operation again - it's a
1699blessing when stepping through miles of source code.
1700
1701=item print
1702
1703Execute the given C code and print its results. B<WARNING>: Perl makes
52d59bef
JH
1704heavy use of macros, and F<gdb> does not necessarily support macros
1705(see later L</"gdb macro support">). You'll have to substitute them
1706yourself, or to invoke cpp on the source code files
1707(see L</"The .i Targets">)
1708So, for instance, you can't say
a422fd2d
SC
1709
1710 print SvPV_nolen(sv)
1711
1712but you have to say
1713
1714 print Perl_sv_2pv_nolen(sv)
1715
ffc145e8
RK
1716=back
1717
a422fd2d
SC
1718You may find it helpful to have a "macro dictionary", which you can
1719produce by saying C<cpp -dM perl.c | sort>. Even then, F<cpp> won't
52d59bef
JH
1720recursively apply those macros for you.
1721
1722=head2 gdb macro support
a422fd2d 1723
52d59bef 1724Recent versions of F<gdb> have fairly good macro support, but
ea031e66
RGS
1725in order to use it you'll need to compile perl with macro definitions
1726included in the debugging information. Using F<gcc> version 3.1, this
1727means configuring with C<-Doptimize=-g3>. Other compilers might use a
1728different switch (if they support debugging macros at all).
1729
a422fd2d
SC
1730=head2 Dumping Perl Data Structures
1731
1732One way to get around this macro hell is to use the dumping functions in
1733F<dump.c>; these work a little like an internal
1734L<Devel::Peek|Devel::Peek>, but they also cover OPs and other structures
1735that you can't get at from Perl. Let's take an example. We'll use the
1736C<$a = $b + $c> we used before, but give it a bit of context:
1737C<$b = "6XXXX"; $c = 2.3;>. Where's a good place to stop and poke around?
1738
1739What about C<pp_add>, the function we examined earlier to implement the
1740C<+> operator:
1741
1742 (gdb) break Perl_pp_add
1743 Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.
1744
cea6626f 1745Notice we use C<Perl_pp_add> and not C<pp_add> - see L<perlguts/Internal Functions>.
a422fd2d
SC
1746With the breakpoint in place, we can run our program:
1747
1748 (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'
1749
1750Lots of junk will go past as gdb reads in the relevant source files and
1751libraries, and then:
1752
1753 Breakpoint 1, Perl_pp_add () at pp_hot.c:309
39644a26 1754 309 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
a422fd2d
SC
1755 (gdb) step
1756 311 dPOPTOPnnrl_ul;
1757 (gdb)
1758
1759We looked at this bit of code before, and we said that C<dPOPTOPnnrl_ul>
1760arranges for two C<NV>s to be placed into C<left> and C<right> - let's
1761slightly expand it:
1762
1763 #define dPOPTOPnnrl_ul NV right = POPn; \
1764 SV *leftsv = TOPs; \
1765 NV left = USE_LEFT(leftsv) ? SvNV(leftsv) : 0.0
1766
1767C<POPn> takes the SV from the top of the stack and obtains its NV either
1768directly (if C<SvNOK> is set) or by calling the C<sv_2nv> function.
1769C<TOPs> takes the next SV from the top of the stack - yes, C<POPn> uses
1770C<TOPs> - but doesn't remove it. We then use C<SvNV> to get the NV from
1771C<leftsv> in the same way as before - yes, C<POPn> uses C<SvNV>.
1772
1773Since we don't have an NV for C<$b>, we'll have to use C<sv_2nv> to
1774convert it. If we step again, we'll find ourselves there:
1775
1776 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
1777 1669 if (!sv)
1778 (gdb)
1779
1780We can now use C<Perl_sv_dump> to investigate the SV:
1781
1782 SV = PV(0xa057cc0) at 0xa0675d0
1783 REFCNT = 1
1784 FLAGS = (POK,pPOK)
1785 PV = 0xa06a510 "6XXXX"\0
1786 CUR = 5
1787 LEN = 6
1788 $1 = void
1789
1790We know we're going to get C<6> from this, so let's finish the
1791subroutine:
1792
1793 (gdb) finish
1794 Run till exit from #0 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
1795 0x462669 in Perl_pp_add () at pp_hot.c:311
1796 311 dPOPTOPnnrl_ul;
1797
1798We can also dump out this op: the current op is always stored in
1799C<PL_op>, and we can dump it with C<Perl_op_dump>. This'll give us
1800similar output to L<B::Debug|B::Debug>.
1801
1802 {
1803 13 TYPE = add ===> 14
1804 TARG = 1
1805 FLAGS = (SCALAR,KIDS)
1806 {
1807 TYPE = null ===> (12)
1808 (was rv2sv)
1809 FLAGS = (SCALAR,KIDS)
1810 {
1811 11 TYPE = gvsv ===> 12
1812 FLAGS = (SCALAR)
1813 GV = main::b
1814 }
1815 }
1816
10f58044 1817# finish this later #
a422fd2d
SC
1818
1819=head2 Patching
1820
1821All right, we've now had a look at how to navigate the Perl sources and
1822some things you'll need to know when fiddling with them. Let's now get
1823on and create a simple patch. Here's something Larry suggested: if a
1824C<U> is the first active format during a C<pack>, (for example,
1825C<pack "U3C8", @stuff>) then the resulting string should be treated as
1e54db1a 1826UTF-8 encoded.
a422fd2d
SC
1827
1828How do we prepare to fix this up? First we locate the code in question -
1829the C<pack> happens at runtime, so it's going to be in one of the F<pp>
1830files. Sure enough, C<pp_pack> is in F<pp.c>. Since we're going to be
1831altering this file, let's copy it to F<pp.c~>.
1832
a6ec74c1
JH
1833[Well, it was in F<pp.c> when this tutorial was written. It has now been
1834split off with C<pp_unpack> to its own file, F<pp_pack.c>]
1835
a422fd2d
SC
1836Now let's look over C<pp_pack>: we take a pattern into C<pat>, and then
1837loop over the pattern, taking each format character in turn into
1838C<datum_type>. Then for each possible format character, we swallow up
1839the other arguments in the pattern (a field width, an asterisk, and so
1840on) and convert the next chunk input into the specified format, adding
1841it onto the output SV C<cat>.
1842
1843How do we know if the C<U> is the first format in the C<pat>? Well, if
1844we have a pointer to the start of C<pat> then, if we see a C<U> we can
1845test whether we're still at the start of the string. So, here's where
1846C<pat> is set up:
1847
1848 STRLEN fromlen;
1849 register char *pat = SvPVx(*++MARK, fromlen);
1850 register char *patend = pat + fromlen;
1851 register I32 len;
1852 I32 datumtype;
1853 SV *fromstr;
1854
1855We'll have another string pointer in there:
1856
1857 STRLEN fromlen;
1858 register char *pat = SvPVx(*++MARK, fromlen);
1859 register char *patend = pat + fromlen;
1860 + char *patcopy;
1861 register I32 len;
1862 I32 datumtype;
1863 SV *fromstr;
1864
1865And just before we start the loop, we'll set C<patcopy> to be the start
1866of C<pat>:
1867
1868 items = SP - MARK;
1869 MARK++;
1870 sv_setpvn(cat, "", 0);
1871 + patcopy = pat;
1872 while (pat < patend) {
1873
1874Now if we see a C<U> which was at the start of the string, we turn on
1e54db1a 1875the C<UTF8> flag for the output SV, C<cat>:
a422fd2d
SC
1876
1877 + if (datumtype == 'U' && pat==patcopy+1)
1878 + SvUTF8_on(cat);
1879 if (datumtype == '#') {
1880 while (pat < patend && *pat != '\n')
1881 pat++;
1882
1883Remember that it has to be C<patcopy+1> because the first character of
1884the string is the C<U> which has been swallowed into C<datumtype!>
1885
1886Oops, we forgot one thing: what if there are spaces at the start of the
1887pattern? C<pack(" U*", @stuff)> will have C<U> as the first active
1888character, even though it's not the first thing in the pattern. In this
1889case, we have to advance C<patcopy> along with C<pat> when we see spaces:
1890
1891 if (isSPACE(datumtype))
1892 continue;
1893
1894needs to become
1895
1896 if (isSPACE(datumtype)) {
1897 patcopy++;
1898 continue;
1899 }
1900
1901OK. That's the C part done. Now we must do two additional things before
1902this patch is ready to go: we've changed the behaviour of Perl, and so
1903we must document that change. We must also provide some more regression
1904tests to make sure our patch works and doesn't create a bug somewhere
1905else along the line.
1906
b23b8711
MS
1907The regression tests for each operator live in F<t/op/>, and so we
1908make a copy of F<t/op/pack.t> to F<t/op/pack.t~>. Now we can add our
1909tests to the end. First, we'll test that the C<U> does indeed create
1910Unicode strings.
1911
1912t/op/pack.t has a sensible ok() function, but if it didn't we could
35c336e6 1913use the one from t/test.pl.
b23b8711 1914
35c336e6
MS
1915 require './test.pl';
1916 plan( tests => 159 );
b23b8711
MS
1917
1918so instead of this:
a422fd2d
SC
1919
1920 print 'not ' unless "1.20.300.4000" eq sprintf "%vd", pack("U*",1,20,300,4000);
1921 print "ok $test\n"; $test++;
1922
35c336e6
MS
1923we can write the more sensible (see L<Test::More> for a full
1924explanation of is() and other testing functions).
b23b8711 1925
35c336e6 1926 is( "1.20.300.4000", sprintf "%vd", pack("U*",1,20,300,4000),
812f5127 1927 "U* produces unicode" );
b23b8711 1928
a422fd2d
SC
1929Now we'll test that we got that space-at-the-beginning business right:
1930
35c336e6 1931 is( "1.20.300.4000", sprintf "%vd", pack(" U*",1,20,300,4000),
812f5127 1932 " with spaces at the beginning" );
a422fd2d
SC
1933
1934And finally we'll test that we don't make Unicode strings if C<U> is B<not>
1935the first active format:
1936
35c336e6 1937 isnt( v1.20.300.4000, sprintf "%vd", pack("C0U*",1,20,300,4000),
812f5127 1938 "U* not first isn't unicode" );
a422fd2d 1939
35c336e6
MS
1940Mustn't forget to change the number of tests which appears at the top,
1941or else the automated tester will get confused. This will either look
1942like this:
a422fd2d 1943
35c336e6
MS
1944 print "1..156\n";
1945
1946or this:
1947
1948 plan( tests => 156 );
a422fd2d
SC
1949
1950We now compile up Perl, and run it through the test suite. Our new
1951tests pass, hooray!
1952
1953Finally, the documentation. The job is never done until the paperwork is
1954over, so let's describe the change we've just made. The relevant place
1955is F<pod/perlfunc.pod>; again, we make a copy, and then we'll insert
1956this text in the description of C<pack>:
1957
1958 =item *
1959
1960 If the pattern begins with a C<U>, the resulting string will be treated
1e54db1a
JH
1961 as UTF-8-encoded Unicode. You can force UTF-8 encoding on in a string
1962 with an initial C<U0>, and the bytes that follow will be interpreted as
1963 Unicode characters. If you don't want this to happen, you can begin your
1964 pattern with C<C0> (or anything else) to force Perl not to UTF-8 encode your
a422fd2d
SC
1965 string, and then follow this with a C<U*> somewhere in your pattern.
1966
1967All done. Now let's create the patch. F<Porting/patching.pod> tells us
1968that if we're making major changes, we should copy the entire directory
1969to somewhere safe before we begin fiddling, and then do
13a2d996 1970
a422fd2d
SC
1971 diff -ruN old new > patch
1972
1973However, we know which files we've changed, and we can simply do this:
1974
1975 diff -u pp.c~ pp.c > patch
1976 diff -u t/op/pack.t~ t/op/pack.t >> patch
1977 diff -u pod/perlfunc.pod~ pod/perlfunc.pod >> patch
1978
1979We end up with a patch looking a little like this:
1980
1981 --- pp.c~ Fri Jun 02 04:34:10 2000
1982 +++ pp.c Fri Jun 16 11:37:25 2000
1983 @@ -4375,6 +4375,7 @@
1984 register I32 items;
1985 STRLEN fromlen;
1986 register char *pat = SvPVx(*++MARK, fromlen);
1987 + char *patcopy;
1988 register char *patend = pat + fromlen;
1989 register I32 len;
1990 I32 datumtype;
1991 @@ -4405,6 +4406,7 @@
1992 ...
1993
1994And finally, we submit it, with our rationale, to perl5-porters. Job
1995done!
1996
f7e1e956
MS
1997=head2 Patching a core module
1998
1999This works just like patching anything else, with an extra
2000consideration. Many core modules also live on CPAN. If this is so,
2001patch the CPAN version instead of the core and send the patch off to
2002the module maintainer (with a copy to p5p). This will help the module
2003maintainer keep the CPAN version in sync with the core version without
2004constantly scanning p5p.
2005
db300100
RGS
2006The list of maintainers of core modules is usefully documented in
2007F<Porting/Maintainers.pl>.
2008
acbe17fc
JP
2009=head2 Adding a new function to the core
2010
2011If, as part of a patch to fix a bug, or just because you have an
2012especially good idea, you decide to add a new function to the core,
2013discuss your ideas on p5p well before you start work. It may be that
2014someone else has already attempted to do what you are considering and
2015can give lots of good advice or even provide you with bits of code
2016that they already started (but never finished).
2017
2018You have to follow all of the advice given above for patching. It is
2019extremely important to test any addition thoroughly and add new tests
2020to explore all boundary conditions that your new function is expected
2021to handle. If your new function is used only by one module (e.g. toke),
2022then it should probably be named S_your_function (for static); on the
210b36aa 2023other hand, if you expect it to accessible from other functions in
acbe17fc
JP
2024Perl, you should name it Perl_your_function. See L<perlguts/Internal Functions>
2025for more details.
2026
2027The location of any new code is also an important consideration. Don't
2028just create a new top level .c file and put your code there; you would
2029have to make changes to Configure (so the Makefile is created properly),
2030as well as possibly lots of include files. This is strictly pumpking
2031business.
2032
2033It is better to add your function to one of the existing top level
2034source code files, but your choice is complicated by the nature of
2035the Perl distribution. Only the files that are marked as compiled
2036static are located in the perl executable. Everything else is located
2037in the shared library (or DLL if you are running under WIN32). So,
2038for example, if a function was only used by functions located in
2039toke.c, then your code can go in toke.c. If, however, you want to call
2040the function from universal.c, then you should put your code in another
2041location, for example util.c.
2042
2043In addition to writing your c-code, you will need to create an
2044appropriate entry in embed.pl describing your function, then run
2045'make regen_headers' to create the entries in the numerous header
2046files that perl needs to compile correctly. See L<perlguts/Internal Functions>
2047for information on the various options that you can set in embed.pl.
2048You will forget to do this a few (or many) times and you will get
2049warnings during the compilation phase. Make sure that you mention
2050this when you post your patch to P5P; the pumpking needs to know this.
2051
2052When you write your new code, please be conscious of existing code
884bad00 2053conventions used in the perl source files. See L<perlstyle> for
acbe17fc
JP
2054details. Although most of the guidelines discussed seem to focus on
2055Perl code, rather than c, they all apply (except when they don't ;).
2056See also I<Porting/patching.pod> file in the Perl source distribution
2057for lots of details about both formatting and submitting patches of
2058your changes.
2059
2060Lastly, TEST TEST TEST TEST TEST any code before posting to p5p.
2061Test on as many platforms as you can find. Test as many perl
2062Configure options as you can (e.g. MULTIPLICITY). If you have
2063profiling or memory tools, see L<EXTERNAL TOOLS FOR DEBUGGING PERL>
210b36aa 2064below for how to use them to further test your code. Remember that
acbe17fc
JP
2065most of the people on P5P are doing this on their own time and
2066don't have the time to debug your code.
f7e1e956
MS
2067
2068=head2 Writing a test
2069
2070Every module and built-in function has an associated test file (or
2071should...). If you add or change functionality, you have to write a
2072test. If you fix a bug, you have to write a test so that bug never
2073comes back. If you alter the docs, it would be nice to test what the
2074new documentation says.
2075
2076In short, if you submit a patch you probably also have to patch the
2077tests.
2078
2079For modules, the test file is right next to the module itself.
2080F<lib/strict.t> tests F<lib/strict.pm>. This is a recent innovation,
2081so there are some snags (and it would be wonderful for you to brush
2082them out), but it basically works that way. Everything else lives in
2083F<t/>.
2084
2085=over 3
2086
2087=item F<t/base/>
2088
2089Testing of the absolute basic functionality of Perl. Things like
2090C<if>, basic file reads and writes, simple regexes, etc. These are
2091run first in the test suite and if any of them fail, something is
2092I<really> broken.
2093
2094=item F<t/cmd/>
2095
2096These test the basic control structures, C<if/else>, C<while>,
35c336e6 2097subroutines, etc.
f7e1e956
MS
2098
2099=item F<t/comp/>
2100
2101Tests basic issues of how Perl parses and compiles itself.
2102
2103=item F<t/io/>
2104
2105Tests for built-in IO functions, including command line arguments.
2106
2107=item F<t/lib/>
2108
2109The old home for the module tests, you shouldn't put anything new in
2110here. There are still some bits and pieces hanging around in here
2111that need to be moved. Perhaps you could move them? Thanks!
2112
2113=item F<t/op/>
2114
2115Tests for perl's built in functions that don't fit into any of the
2116other directories.
2117
2118=item F<t/pod/>
2119
2120Tests for POD directives. There are still some tests for the Pod
2121modules hanging around in here that need to be moved out into F<lib/>.
2122
2123=item F<t/run/>
2124
2125Testing features of how perl actually runs, including exit codes and
2126handling of PERL* environment variables.
2127
244d9cb7
RGS
2128=item F<t/uni/>
2129
2130Tests for the core support of Unicode.
2131
2132=item F<t/win32/>
2133
2134Windows-specific tests.
2135
2136=item F<t/x2p>
2137
2138A test suite for the s2p converter.
2139
f7e1e956
MS
2140=back
2141
2142The core uses the same testing style as the rest of Perl, a simple
2143"ok/not ok" run through Test::Harness, but there are a few special
2144considerations.
2145
35c336e6
MS
2146There are three ways to write a test in the core. Test::More,
2147t/test.pl and ad hoc C<print $test ? "ok 42\n" : "not ok 42\n">. The
2148decision of which to use depends on what part of the test suite you're
2149working on. This is a measure to prevent a high-level failure (such
2150as Config.pm breaking) from causing basic functionality tests to fail.
2151
2152=over 4
2153
2154=item t/base t/comp
2155
2156Since we don't know if require works, or even subroutines, use ad hoc
2157tests for these two. Step carefully to avoid using the feature being
2158tested.
2159
2160=item t/cmd t/run t/io t/op
2161
2162Now that basic require() and subroutines are tested, you can use the
2163t/test.pl library which emulates the important features of Test::More
2164while using a minimum of core features.
2165
2166You can also conditionally use certain libraries like Config, but be
2167sure to skip the test gracefully if it's not there.
2168
2169=item t/lib ext lib
2170
2171Now that the core of Perl is tested, Test::More can be used. You can
2172also use the full suite of core modules in the tests.
2173
2174=back
f7e1e956
MS
2175
2176When you say "make test" Perl uses the F<t/TEST> program to run the
7205a85d
YO
2177test suite (except under Win32 where it uses F<t/harness> instead.)
2178All tests are run from the F<t/> directory, B<not> the directory
2179which contains the test. This causes some problems with the tests
2180in F<lib/>, so here's some opportunity for some patching.
f7e1e956
MS
2181
2182You must be triply conscious of cross-platform concerns. This usually
2183boils down to using File::Spec and avoiding things like C<fork()> and
2184C<system()> unless absolutely necessary.
2185
e018f8be
JH
2186=head2 Special Make Test Targets
2187
2188There are various special make targets that can be used to test Perl
2189slightly differently than the standard "test" target. Not all them
2190are expected to give a 100% success rate. Many of them have several
7205a85d
YO
2191aliases, and many of them are not available on certain operating
2192systems.
e018f8be
JH
2193
2194=over 4
2195
2196=item coretest
2197
7d7d5695 2198Run F<perl> on all core tests (F<t/*> and F<lib/[a-z]*> pragma tests).
e018f8be 2199
7205a85d
YO
2200(Not available on Win32)
2201
e018f8be
JH
2202=item test.deparse
2203
b26492ee
RGS
2204Run all the tests through B::Deparse. Not all tests will succeed.
2205
7205a85d
YO
2206(Not available on Win32)
2207
b26492ee
RGS
2208=item test.taintwarn
2209
2210Run all tests with the B<-t> command-line switch. Not all tests
2211are expected to succeed (until they're specifically fixed, of course).
e018f8be 2212
7205a85d
YO
2213(Not available on Win32)
2214
e018f8be
JH
2215=item minitest
2216
2217Run F<miniperl> on F<t/base>, F<t/comp>, F<t/cmd>, F<t/run>, F<t/io>,
2218F<t/op>, and F<t/uni> tests.
2219
7a834142
JH
2220=item test.valgrind check.valgrind utest.valgrind ucheck.valgrind
2221
2222(Only in Linux) Run all the tests using the memory leak + naughty
2223memory access tool "valgrind". The log files will be named
2224F<testname.valgrind>.
2225
e018f8be
JH
2226=item test.third check.third utest.third ucheck.third
2227
2228(Only in Tru64) Run all the tests using the memory leak + naughty
2229memory access tool "Third Degree". The log files will be named
60a57c1c 2230F<perl.3log.testname>.
e018f8be
JH
2231
2232=item test.torture torturetest
2233
2234Run all the usual tests and some extra tests. As of Perl 5.8.0 the
244d9cb7 2235only extra tests are Abigail's JAPHs, F<t/japh/abigail.t>.
e018f8be
JH
2236
2237You can also run the torture test with F<t/harness> by giving
2238C<-torture> argument to F<t/harness>.
2239
2240=item utest ucheck test.utf8 check.utf8
2241
2242Run all the tests with -Mutf8. Not all tests will succeed.
2243
7205a85d
YO
2244(Not available on Win32)
2245
cc0710ff
RGS
2246=item minitest.utf16 test.utf16
2247
2248Runs the tests with UTF-16 encoded scripts, encoded with different
2249versions of this encoding.
2250
2251C<make utest.utf16> runs the test suite with a combination of C<-utf8> and
2252C<-utf16> arguments to F<t/TEST>.
2253
7205a85d
YO
2254(Not available on Win32)
2255
244d9cb7
RGS
2256=item test_harness
2257
2258Run the test suite with the F<t/harness> controlling program, instead of
2259F<t/TEST>. F<t/harness> is more sophisticated, and uses the
2260L<Test::Harness> module, thus using this test target supposes that perl
2261mostly works. The main advantage for our purposes is that it prints a
00bf5cd9
RGS
2262detailed summary of failed tests at the end. Also, unlike F<t/TEST>, it
2263doesn't redirect stderr to stdout.
244d9cb7 2264
7205a85d
YO
2265Note that under Win32 F<t/harness> is always used instead of F<t/TEST>, so
2266there is no special "test_harness" target.
2267
2268Under Win32's "test" target you may use the TEST_SWITCHES and TEST_FILES
2269environment variables to control the behaviour of F<t/harness>. This means
2270you can say
2271
2272 nmake test TEST_FILES="op/*.t"
2273 nmake test TEST_SWITCHES="-torture" TEST_FILES="op/*.t"
2274
2275=item test-notty test_notty
2276
2277Sets PERL_SKIP_TTY_TEST to true before running normal test.
2278
244d9cb7
RGS
2279=back
2280
2281=head2 Running tests by hand
2282
2283You can run part of the test suite by hand by using one the following
2284commands from the F<t/> directory :
2285
2286 ./perl -I../lib TEST list-of-.t-files
2287
2288or
2289
2290 ./perl -I../lib harness list-of-.t-files
2291
2292(if you don't specify test scripts, the whole test suite will be run.)
2293
7205a85d
YO
2294=head3 Using t/harness for testing
2295
2296If you use C<harness> for testing you have several command line options
2297available to you. The arguments are as follows, and are in the order
2298that they must appear if used together.
2299
2300 harness -v -torture -re=pattern LIST OF FILES TO TEST
2301 harness -v -torture -re LIST OF PATTERNS TO MATCH
2302
2303If C<LIST OF FILES TO TEST> is omitted the file list is obtained from
2304the manifest. The file list may include shell wildcards which will be
2305expanded out.
2306
2307=over 4
2308
2309=item -v
2310
2311Run the tests under verbose mode so you can see what tests were run,
2312and debug outbut.
2313
2314=item -torture
2315
2316Run the torture tests as well as the normal set.
2317
2318=item -re=PATTERN
2319
2320Filter the file list so that all the test files run match PATTERN.
2321Note that this form is distinct from the B<-re LIST OF PATTERNS> form below
2322in that it allows the file list to be provided as well.
2323
2324=item -re LIST OF PATTERNS
2325
2326Filter the file list so that all the test files run match
2327/(LIST|OF|PATTERNS)/. Note that with this form the patterns
2328are joined by '|' and you cannot supply a list of files, instead
2329the test files are obtained from the MANIFEST.
2330
2331=back
2332
244d9cb7
RGS
2333You can run an individual test by a command similar to
2334
2335 ./perl -I../lib patho/to/foo.t
2336
2337except that the harnesses set up some environment variables that may
2338affect the execution of the test :
2339
2340=over 4
2341
2342=item PERL_CORE=1
2343
2344indicates that we're running this test part of the perl core test suite.
2345This is useful for modules that have a dual life on CPAN.
2346
2347=item PERL_DESTRUCT_LEVEL=2
2348
2349is set to 2 if it isn't set already (see L</PERL_DESTRUCT_LEVEL>)
2350
2351=item PERL
2352
2353(used only by F<t/TEST>) if set, overrides the path to the perl executable
2354that should be used to run the tests (the default being F<./perl>).
2355
2356=item PERL_SKIP_TTY_TEST
2357
2358if set, tells to skip the tests that need a terminal. It's actually set
2359automatically by the Makefile, but can also be forced artificially by
2360running 'make test_notty'.
2361
e018f8be 2362=back
f7e1e956 2363
d7889f52
JH
2364=head2 Common problems when patching Perl source code
2365
2366Perl source plays by ANSI C89 rules: no C99 (or C++) extensions. In
2367some cases we have to take pre-ANSI requirements into consideration.
2368You don't care about some particular platform having broken Perl?
2369I hear there is still a strong demand for J2EE programmers.
2370
2371=head2 Perl environment problems
2372
2373=over 4
2374
2375=item *
2376
2377Not compiling with threading
2378
2379Compiling with threading (-Duseithreads) completely rewrites
2380the function prototypes of Perl. You better try your changes
0bec6c03 2381with that. Related to this is the difference between "Perl_-less"
d7889f52
JH
2382and "Perl_-ly" APIs, for example:
2383
2384 Perl_sv_setiv(aTHX_ ...);
2385 sv_setiv(...);
2386
ee9468a2
RGS
2387The first one explicitly passes in the context, which is needed for e.g.
2388threaded builds. The second one does that implicitly; do not get them
2389mixed. If you are not passing in a aTHX_, you will need to do a dTHX as
2390the first thing in the function.
d7889f52
JH
2391
2392See L<perlguts/"How multiple interpreters and concurrency are supported">
2393for further discussion about context.
2394
2395=item *
2396
2397Not compiling with -DDEBUGGING
2398
2399The DEBUGGING define exposes more code to the compiler,
0bec6c03 2400therefore more ways for things to go wrong. You should try it.
d7889f52
JH
2401
2402=item *
2403
ee9468a2
RGS
2404Introducing (non-read-only) globals
2405
2406Do not introduce any modifiable globals, truly global or file static.
bc028b6b
JH
2407They are bad form and complicate multithreading and other forms of
2408concurrency. The right way is to introduce them as new interpreter
2409variables, see F<intrpvar.h> (at the very end for binary compatibility).
ee9468a2
RGS
2410
2411Introducing read-only (const) globals is okay, as long as you verify
2412with e.g. C<nm libperl.a|egrep -v ' [TURtr] '> (if your C<nm> has
2413BSD-style output) that the data you added really is read-only.
2414(If it is, it shouldn't show up in the output of that command.)
2415
2416If you want to have static strings, make them constant:
2417
2418 static const char etc[] = "...";
2419
bc028b6b 2420If you want to have arrays of constant strings, note carefully
ee9468a2
RGS
2421the right combination of C<const>s:
2422
2423 static const char * const yippee[] =
2424 {"hi", "ho", "silver"};
2425
bc028b6b
JH
2426There is a way to completely hide any modifiable globals (they are all
2427moved to heap), the compilation setting C<-DPERL_GLOBAL_STRUCT_PRIVATE>.
2428It is not normally used, but can be used for testing, read more
2429about it in L<perlhack>.
2430
ee9468a2
RGS
2431=item *
2432
d7889f52
JH
2433Not exporting your new function
2434
2435Some platforms (Win32, AIX, VMS, OS/2, to name a few) require any
2436function that is part of the public API (the shared Perl library)
2437to be explicitly marked as exported. See the discussion about
2438F<embed.pl> in L<perlguts>.
2439
2440=item *
2441
2442Exporting your new function
2443
2444The new shiny result of either genuine new functionality or your
2445arduous refactoring is now ready and correctly exported. So what
2446could possibly be wrong?
2447
2448Maybe simply that your function did not need to be exported in the
2449first place. Perl has a long and not so glorious history of exporting
2450functions that it should not have.
2451
2452If the function is used only inside one source code file, make it
2453static. See the discussion about F<embed.pl> in L<perlguts>.
2454
2455If the function is used across several files, but intended only for
2456Perl's internal use (and this should be the common case), do not
2457export it to the public API. See the discussion about F<embed.pl>
2458in L<perlguts>.
2459
2460=back
2461
5b38b9cd 2462=head2 Portability problems
d7889f52
JH
2463
2464The following are common causes of compilation and/or execution
2465failures, not common to Perl as such. The C FAQ is good bedtime
0bec6c03
JH
2466reading. Please test your changes with as many C compilers and
2467platforms as possible -- we will, anyway, and it's nice to save
2468oneself from public embarrassment.
2469
9aaf14db
RGS
2470If using gcc, you can add the C<-std=c89> option which will hopefully
2471catch most of these unportabilities. (However it might also catch
2472incompatibilities in your system's header files.)
d1307786 2473
a8e98a71
JH
2474Use the Configure C<-Dgccansipedantic> flag to enable the gcc
2475C<-ansi -pedantic> flags which enforce stricter ANSI rules.
2476
0bec6c03
JH
2477Also study L<perlport> carefully to avoid any bad assumptions
2478about the operating system, filesystem, and so forth.
2479
63796a85
JH
2480You may once in a while try a "make miniperl" to see whether we
2481can still compile Perl with just the bare minimum of interfaces.
ee9468a2 2482
0bec6c03 2483Do not assume an operating system indicates a certain compiler.
d7889f52
JH
2484
2485=over 4
2486
2487=item *
2488
2489Casting pointers to integers or casting integers to pointers
2490
2491 void castaway(U8* p)
2492 {
2493 IV i = p;
2494
2495or
2496
2497 void castaway(U8* p)
2498 {
2499 IV i = (IV)p;
2500
ee9468a2 2501Both are bad, and broken, and unportable. Use the PTR2IV()
d7889f52
JH
2502macro that does it right. (Likewise, there are PTR2UV(), PTR2NV(),
2503INT2PTR(), and NUM2PTR().)
2504
2505=item *
2506
0bec6c03
JH
2507Casting between data function pointers and data pointers
2508
d7889f52
JH
2509Technically speaking casting between function pointers and data
2510pointers is unportable and undefined, but practically speaking
2511it seems to work, but you should use the FPTR2DPTR() and DPTR2FPTR()
0bec6c03 2512macros. Sometimes you can also play games with unions.
d7889f52
JH
2513
2514=item *
2515
2516Assuming sizeof(int) == sizeof(long)
2517
2518There are platforms where longs are 64 bits, and platforms where ints
2519are 64 bits, and while we are out to shock you, even platforms where
2520shorts are 64 bits. This is all legal according to the C standard.
2521(In other words, "long long" is not a portable way to specify 64 bits,
2522and "long long" is not even guaranteed to be any wider than "long".)
63796a85
JH
2523
2524Instead, use the definitions IV, UV, IVSIZE, I32SIZE, and so forth.
2525Avoid things like I32 because they are B<not> guaranteed to be
2526I<exactly> 32 bits, they are I<at least> 32 bits, nor are they
2527guaranteed to be B<int> or B<long>. If you really explicitly need
252864-bit variables, use I64 and U64, but only if guarded by HAS_QUAD.
d7889f52
JH
2529
2530=item *
2531
2532Assuming one can dereference any type of pointer for any type of data
2533
2534 char *p = ...;
2535 long pony = *p;
2536
2537Many platforms, quite rightly so, will give you a core dump instead
2538of a pony if the p happens not be correctly aligned.
2539
2540=item *
2541
2542Lvalue casts
2543
2544 (int)*p = ...;
2545
2546Simply not portable. Get your lvalue to be of the right type,
2547or maybe use temporary variables.
2548
2549=item *
2550
0bec6c03
JH
2551Mixing #define and #ifdef
2552
2553 #define BURGLE(x) ... \
2554 #ifdef BURGLE_OLD_STYLE
2555 ... do it the old way ... \
2556 #else
2557 ... do it the new way ... \
2558 #endif
2559
ee9468a2
RGS
2560You cannot portably "stack" cpp directives. For example in the above
2561you need two separate BURGLE() #defines, one for each #ifdef branch.
2562
2563=item *
2564
2565Adding stuff after #endif or #else
2566
2567 #ifdef SNOSH
2568 ...
2569 #else !SNOSH
2570 ...
2571 #endif SNOSH
2572
2573The #endif and #else cannot portably have anything after them. If you
2574want to document what is going (which is a good idea especially if the
2575branches are long), use (C) comments:
2576
2577 #ifdef SNOSH
2578 ...
2579 #else /* !SNOSH */
2580 ...
2581 #endif /* SNOSH */
2582
2583The gcc option C<-Wendif-labels> warns about the bad variant
2584(by default on starting from Perl 5.9.4).
0bec6c03
JH
2585
2586=item *
2587
d7889f52
JH
2588Using //-comments
2589
2590 // This function bamfoodles the zorklator.
2591
2592That is C99 or C++. Perl is C89. Using the //-comments is silently
0bec6c03
JH
2593allowed by many C compilers but cranking up the ANSI C89 strictness
2594(which we like to do) causes the compilation to fail.
d7889f52
JH
2595
2596=item *
2597
2598Mixing declarations and code
2599
2600 void zorklator()
2601 {
2602 int n = 3;
2603 set_zorkmids(n);
2604 int q = 4;
2605
0bec6c03
JH
2606That is C99 or C++. Some C compilers allow that, but you shouldn't.
2607
63796a85
JH
2608The gcc option C<-Wdeclaration-after-statements> scans for such problems
2609(by default on starting from Perl 5.9.4).
2610
0bec6c03
JH
2611=item *
2612
2613Introducing variables inside for()
2614
2615 for(int i = ...; ...; ...)
2616
2617That is C99 or C++. While it would indeed be awfully nice to have that
2618also in C89, to limit the scope of the loop variable, alas, we cannot.
d7889f52
JH
2619
2620=item *
2621
2622Mixing signed char pointers with unsigned char pointers
2623
2624 int foo(char *s) { ... }
2625 ...
2626 unsigned char *t = ...; /* Or U8* t = ... */
2627 foo(t);
2628
2629While this is legal practice, it is certainly dubious, and downright
2630fatal in at least one platform: for example VMS cc considers this a
63796a85
JH
2631fatal error. One cause for people often making this mistake is that
2632a "naked char" and therefore dereferencing a "naked char pointer" have
0bec6c03 2633an undefined signedness: it depends on the compiler and the platform
63796a85
JH
2634whether the result is signed or unsigned. For this very same reason
2635using a 'char' as an array index is bad.
d7889f52
JH
2636
2637=item *
2638
2639Macros that have string constants and their arguments as substrings of
2640the string constants
2641
2642 #define FOO(n) printf("number = %d\n", n)
2643 FOO(10);
2644
2645Pre-ANSI semantics for that was equivalent to
2646
2647 printf("10umber = %d\10");
2648
0bec6c03
JH
2649which is probably not what you were expecting. Unfortunately at least
2650one reasonably common and modern C compiler does "real backward
63796a85 2651compatibility" here, in AIX that is what still happens even though the
0bec6c03
JH
2652rest of the AIX compiler is very happily C89.
2653
2654=item *
2655
ee9468a2
RGS
2656Using printf formats for non-basic C types
2657
2658 IV i = ...;
2659 printf("i = %d\n", i);
2660
2661While this might by accident work in some platform (where IV happens
2662to be an C<int>), in general it cannot. IV might be something larger.
2663Even worse the situation is with more specific types (defined by Perl's
2664configuration step in F<config.h>):
2665
2666 Uid_t who = ...;
2667 printf("who = %d\n", who);
2668
2669The problem here is that Uid_t might be not only not C<int>-wide
2670but it might also be unsigned, in which case large uids would be
2671printed as negative values.
2672
2673There is no simple solution to this because of printf()'s limited
2674intelligence, but for many types the right format is available as
2675with either 'f' or '_f' suffix, for example:
2676
2677 IVdf /* IV in decimal */
2678 UVxf /* UV is hexadecimal */
2679
2680 printf("i = %"IVdf"\n", i); /* The IVdf is a string constant. */
2681
2682 Uid_t_f /* Uid_t in decimal */
2683
2684 printf("who = %"Uid_t_f"\n", who);
2685
63796a85
JH
2686Or you can try casting to a "wide enough" type:
2687
2688 printf("i = %"IVdf"\n", (IV)something_very_small_and_signed);
2689
2690Also remember that the C<%p> format really does require a void pointer:
2691
2692 U8* p = ...;
2693 printf("p = %p\n", (void*)p);
2694
ee9468a2
RGS
2695The gcc option C<-Wformat> scans for such problems.
2696
2697=item *
2698
0bec6c03
JH
2699Blindly using variadic macros
2700
63796a85
JH
2701gcc has had them for a while with its own syntax, and C99 brought
2702them with a standardized syntax. Don't use the former, and use
2703the latter only if the HAS_C99_VARIADIC_MACROS is defined.
0bec6c03
JH
2704
2705=item *
2706
2707Blindly passing va_list
2708
2709Not all platforms support passing va_list to further varargs (stdarg)
2710functions. The right thing to do is to copy the va_list using the
2711Perl_va_copy() if the NEED_VA_COPY is defined.
d7889f52 2712
ee9468a2
RGS
2713=item *
2714
63796a85
JH
2715Using gcc brace groups
2716
2717 val = ({...;...;...});
2718
2719While a nice extension, it's not portable.
2720
2721=item *
2722
2723Binding together several statements
2724
2725Use the macros STMT_START and STMT_END.
2726
2727 STMT_START {
2728 ...
2729 } STMT_END
2730
2731=item *
2732
ee9468a2
RGS
2733Testing for operating systems or versions when should be testing for features
2734
2735 #ifdef __FOONIX__
2736 foo = quux();
2737 #endif
2738
2739Unless you know with 100% certainty that quux() is only ever available
2740for the "Foonix" operating system B<and> that is available B<and>
2741correctly working for B<all> past, present, B<and> future versions of
2742"Foonix", the above is very wrong. This is more correct (though still
2743not perfect, because the below is a compile-time check):
2744
2745 #ifdef HAS_QUUX
2746 foo = quux();
2747 #endif
2748
2749How does the HAS_QUUX become defined where it needs to be? Well, if
2750Foonix happens to be UNIXy enought to be able to run the Configure
2751script, and Configure has been taught about detecting and testing
2752quux(), the HAS_QUUX will be correctly defined. In other platforms,
2753the corresponding configuration step will hopefully do the same.
2754
2755In a pinch, if you cannot wait for Configure to be educated,
2756or if you have a good hunch of where quux() might be available,
2757you can temporarily try the following:
2758
2759 #if (defined(__FOONIX__) || defined(__BARNIX__))
2760 # define HAS_QUUX
2761 #endif
2762
2763 ...
2764
2765 #ifdef HAS_QUUX
2766 foo = quux();
2767 #endif
2768
2769But in any case, try to keep the features and operating systems separate.
2770
d7889f52
JH
2771=back
2772
2773=head2 Security problems
2774
2775Last but not least, here are various tips for safer coding.
2776
2777=over 4
2778
2779=item *
2780
2781Do not use gets()
2782
2783Or we will publicly ridicule you. Seriously.
2784
2785=item *
2786
d1307786 2787Do not use strcpy() or strcat() or strncpy() or strncat()
d7889f52 2788
d1307786
JH
2789Use my_strlcpy() and my_strlcat() instead: they either use the native
2790implementation, or Perl's own implementation (borrowed from the public
2791domain implementation of INN).
d7889f52
JH
2792
2793=item *
2794
2795Do not use sprintf() or vsprintf()
2796
0bec6c03 2797If you really want just plain byte strings, use my_snprintf()
64d9b66b 2798and my_vsnprintf() instead, which will try to use snprintf() and
0bec6c03
JH
2799vsnprintf() if those safer APIs are available. If you want something
2800fancier than a plain byte string, use SVs and Perl_sv_catpvf().
d7889f52
JH
2801
2802=back
2803
902b9dbf
MF
2804=head1 EXTERNAL TOOLS FOR DEBUGGING PERL
2805
2806Sometimes it helps to use external tools while debugging and
2807testing Perl. This section tries to guide you through using
2808some common testing and debugging tools with Perl. This is
2809meant as a guide to interfacing these tools with Perl, not
2810as any kind of guide to the use of the tools themselves.
2811
a958818a
JH
2812B<NOTE 1>: Running under memory debuggers such as Purify, valgrind, or
2813Third Degree greatly slows down the execution: seconds become minutes,
2814minutes become hours. For example as of Perl 5.8.1, the
2815ext/Encode/t/Unicode.t takes extraordinarily long to complete under
2816e.g. Purify, Third Degree, and valgrind. Under valgrind it takes more
2817than six hours, even on a snappy computer-- the said test must be
2818doing something that is quite unfriendly for memory debuggers. If you
2819don't feel like waiting, that you can simply kill away the perl
2820process.
2821
2822B<NOTE 2>: To minimize the number of memory leak false alarms (see
2823L</PERL_DESTRUCT_LEVEL> for more information), you have to have
2824environment variable PERL_DESTRUCT_LEVEL set to 2. The F<TEST>
2825and harness scripts do that automatically. But if you are running
2826some of the tests manually-- for csh-like shells:
2827
2828 setenv PERL_DESTRUCT_LEVEL 2
2829
2830and for Bourne-type shells:
2831
2832 PERL_DESTRUCT_LEVEL=2
2833 export PERL_DESTRUCT_LEVEL
2834
2835or in UNIXy environments you can also use the C<env> command:
2836
2837 env PERL_DESTRUCT_LEVEL=2 valgrind ./perl -Ilib ...
a1b65709 2838
37c0adeb
JH
2839B<NOTE 3>: There are known memory leaks when there are compile-time
2840errors within eval or require, seeing C<S_doeval> in the call stack
2841is a good sign of these. Fixing these leaks is non-trivial,
2842unfortunately, but they must be fixed eventually.
2843
902b9dbf
MF
2844=head2 Rational Software's Purify
2845
2846Purify is a commercial tool that is helpful in identifying
2847memory overruns, wild pointers, memory leaks and other such
2848badness. Perl must be compiled in a specific way for
2849optimal testing with Purify. Purify is available under
2850Windows NT, Solaris, HP-UX, SGI, and Siemens Unix.
2851
902b9dbf
MF
2852=head2 Purify on Unix
2853
2854On Unix, Purify creates a new Perl binary. To get the most
2855benefit out of Purify, you should create the perl to Purify
2856using:
2857
2858 sh Configure -Accflags=-DPURIFY -Doptimize='-g' \
2859 -Uusemymalloc -Dusemultiplicity
2860
2861where these arguments mean:
2862
2863=over 4
2864
2865=item -Accflags=-DPURIFY
2866
2867Disables Perl's arena memory allocation functions, as well as
2868forcing use of memory allocation functions derived from the
2869system malloc.
2870
2871=item -Doptimize='-g'
2872
2873Adds debugging information so that you see the exact source
2874statements where the problem occurs. Without this flag, all
2875you will see is the source filename of where the error occurred.
2876
2877=item -Uusemymalloc
2878
2879Disable Perl's malloc so that Purify can more closely monitor
2880allocations and leaks. Using Perl's malloc will make Purify
2881report most leaks in the "potential" leaks category.
2882
2883=item -Dusemultiplicity
2884
2885Enabling the multiplicity option allows perl to clean up
2886thoroughly when the interpreter shuts down, which reduces the
2887number of bogus leak reports from Purify.
2888
2889=back
2890
2891Once you've compiled a perl suitable for Purify'ing, then you
2892can just:
2893
2894 make pureperl
2895
2896which creates a binary named 'pureperl' that has been Purify'ed.
2897This binary is used in place of the standard 'perl' binary
2898when you want to debug Perl memory problems.
2899
2900As an example, to show any memory leaks produced during the
2901standard Perl testset you would create and run the Purify'ed
2902perl as:
2903
2904 make pureperl
2905 cd t
2906 ../pureperl -I../lib harness
2907
2908which would run Perl on test.pl and report any memory problems.
2909
2910Purify outputs messages in "Viewer" windows by default. If
2911you don't have a windowing environment or if you simply
2912want the Purify output to unobtrusively go to a log file
2913instead of to the interactive window, use these following
2914options to output to the log file "perl.log":
2915
2916 setenv PURIFYOPTIONS "-chain-length=25 -windows=no \
2917 -log-file=perl.log -append-logfile=yes"
2918
2919If you plan to use the "Viewer" windows, then you only need this option:
2920
2921 setenv PURIFYOPTIONS "-chain-length=25"
2922
c406981e
JH
2923In Bourne-type shells:
2924
98631ff8
JL
2925 PURIFYOPTIONS="..."
2926 export PURIFYOPTIONS
c406981e
JH
2927
2928or if you have the "env" utility:
2929
98631ff8 2930 env PURIFYOPTIONS="..." ../pureperl ...
c406981e 2931
902b9dbf
MF
2932=head2 Purify on NT
2933
2934Purify on Windows NT instruments the Perl binary 'perl.exe'
2935on the fly. There are several options in the makefile you
2936should change to get the most use out of Purify:
2937
2938=over 4
2939
2940=item DEFINES
2941
2942You should add -DPURIFY to the DEFINES line so the DEFINES
2943line looks something like:
2944
2945 DEFINES = -DWIN32 -D_CONSOLE -DNO_STRICT $(CRYPT_FLAG) -DPURIFY=1
2946
2947to disable Perl's arena memory allocation functions, as
2948well as to force use of memory allocation functions derived
2949from the system malloc.
2950
2951=item USE_MULTI = define
2952
2953Enabling the multiplicity option allows perl to clean up
2954thoroughly when the interpreter shuts down, which reduces the
2955number of bogus leak reports from Purify.
2956
2957=item #PERL_MALLOC = define
2958
2959Disable Perl's malloc so that Purify can more closely monitor
2960allocations and leaks. Using Perl's malloc will make Purify
2961report most leaks in the "potential" leaks category.
2962
2963=item CFG = Debug
2964
2965Adds debugging information so that you see the exact source
2966statements where the problem occurs. Without this flag, all
2967you will see is the source filename of where the error occurred.
2968
2969=back
2970
2971As an example, to show any memory leaks produced during the
2972standard Perl testset you would create and run Purify as:
2973
2974 cd win32
2975 make
2976 cd ../t
2977 purify ../perl -I../lib harness
2978
2979which would instrument Perl in memory, run Perl on test.pl,
2980then finally report any memory problems.
2981
7a834142
JH
2982=head2 valgrind
2983
2984The excellent valgrind tool can be used to find out both memory leaks
2985and illegal memory accesses. As of August 2003 it unfortunately works
2986only on x86 (ELF) Linux. The special "test.valgrind" target can be used
d44161bf
MHM
2987to run the tests under valgrind. Found errors and memory leaks are
2988logged in files named F<test.valgrind>.
2989
2990As system libraries (most notably glibc) are also triggering errors,
2991valgrind allows to suppress such errors using suppression files. The
2992default suppression file that comes with valgrind already catches a lot
2993of them. Some additional suppressions are defined in F<t/perl.supp>.
7a834142
JH
2994
2995To get valgrind and for more information see
2996
2997 http://developer.kde.org/~sewardj/
2998
f134cc4e 2999=head2 Compaq's/Digital's/HP's Third Degree
09187cb1
JH
3000
3001Third Degree is a tool for memory leak detection and memory access checks.
3002It is one of the many tools in the ATOM toolkit. The toolkit is only
3003available on Tru64 (formerly known as Digital UNIX formerly known as
3004DEC OSF/1).
3005
3006When building Perl, you must first run Configure with -Doptimize=-g
3007and -Uusemymalloc flags, after that you can use the make targets
51a35ef1
JH
3008"perl.third" and "test.third". (What is required is that Perl must be
3009compiled using the C<-g> flag, you may need to re-Configure.)
09187cb1 3010
64cea5fd 3011The short story is that with "atom" you can instrument the Perl
83f0ef60 3012executable to create a new executable called F<perl.third>. When the
4ae3d70a 3013instrumented executable is run, it creates a log of dubious memory
83f0ef60 3014traffic in file called F<perl.3log>. See the manual pages of atom and
4ae3d70a
JH
3015third for more information. The most extensive Third Degree
3016documentation is available in the Compaq "Tru64 UNIX Programmer's
3017Guide", chapter "Debugging Programs with Third Degree".
64cea5fd 3018
9c54ecba 3019The "test.third" leaves a lot of files named F<foo_bar.3log> in the t/
64cea5fd
JH
3020subdirectory. There is a problem with these files: Third Degree is so
3021effective that it finds problems also in the system libraries.
9c54ecba
JH
3022Therefore you should used the Porting/thirdclean script to cleanup
3023the F<*.3log> files.
64cea5fd
JH
3024
3025There are also leaks that for given certain definition of a leak,
3026aren't. See L</PERL_DESTRUCT_LEVEL> for more information.
3027
3028=head2 PERL_DESTRUCT_LEVEL
3029
a958818a
JH
3030If you want to run any of the tests yourself manually using e.g.
3031valgrind, or the pureperl or perl.third executables, please note that
3032by default perl B<does not> explicitly cleanup all the memory it has
3033allocated (such as global memory arenas) but instead lets the exit()
3034of the whole program "take care" of such allocations, also known as
3035"global destruction of objects".
64cea5fd
JH
3036
3037There is a way to tell perl to do complete cleanup: set the
3038environment variable PERL_DESTRUCT_LEVEL to a non-zero value.
3039The t/TEST wrapper does set this to 2, and this is what you
3040need to do too, if you don't want to see the "global leaks":
1f56d61a 3041For example, for "third-degreed" Perl:
64cea5fd 3042
1f56d61a 3043 env PERL_DESTRUCT_LEVEL=2 ./perl.third -Ilib t/foo/bar.t
09187cb1 3044
414f2397
RGS
3045(Note: the mod_perl apache module uses also this environment variable
3046for its own purposes and extended its semantics. Refer to the mod_perl
287a822c
RGS
3047documentation for more information. Also, spawned threads do the
3048equivalent of setting this variable to the value 1.)
5a6c59ef
DM
3049
3050If, at the end of a run you get the message I<N scalars leaked>, you can
fd0854ff
DM
3051recompile with C<-DDEBUG_LEAKING_SCALARS>, which will cause the addresses
3052of all those leaked SVs to be dumped along with details as to where each
3053SV was originally allocated. This information is also displayed by
3054Devel::Peek. Note that the extra details recorded with each SV increases
3055memory usage, so it shouldn't be used in production environments. It also
3056converts C<new_SV()> from a macro into a real function, so you can use
3057your favourite debugger to discover where those pesky SVs were allocated.
414f2397 3058
46c6c7e2
JH
3059=head2 PERL_MEM_LOG
3060
9f653bb5 3061If compiled with C<-DPERL_MEM_LOG>, all Newx() and Renew() allocations
46c6c7e2
JH
3062and Safefree() in the Perl core go through logging functions, which is
3063handy for breakpoint setting. If also compiled with C<-DPERL_MEM_LOG_STDERR>,
e352bcff
JH
3064the allocations and frees are logged to STDERR (or more precisely, to the
3065file descriptor 2) in these logging functions, with the calling source code
3066file and line number (and C function name, if supported by the C compiler).
3067
3068This logging is somewhat similar to C<-Dm> but independent of C<-DDEBUGGING>,
3069and at a higher level (the C<-Dm> is directly at the point of C<malloc()>,
a015a77e 3070while the C<PERL_MEM_LOG> is at the level of C<New()>).
46c6c7e2 3071
51a35ef1
JH
3072=head2 Profiling
3073
3074Depending on your platform there are various of profiling Perl.
3075
3076There are two commonly used techniques of profiling executables:
10f58044 3077I<statistical time-sampling> and I<basic-block counting>.
51a35ef1
JH
3078
3079The first method takes periodically samples of the CPU program
3080counter, and since the program counter can be correlated with the code
3081generated for functions, we get a statistical view of in which
3082functions the program is spending its time. The caveats are that very
3083small/fast functions have lower probability of showing up in the
3084profile, and that periodically interrupting the program (this is
3085usually done rather frequently, in the scale of milliseconds) imposes
3086an additional overhead that may skew the results. The first problem
3087can be alleviated by running the code for longer (in general this is a
3088good idea for profiling), the second problem is usually kept in guard
3089by the profiling tools themselves.
3090
10f58044 3091The second method divides up the generated code into I<basic blocks>.
51a35ef1
JH
3092Basic blocks are sections of code that are entered only in the
3093beginning and exited only at the end. For example, a conditional jump
3094starts a basic block. Basic block profiling usually works by
10f58044 3095I<instrumenting> the code by adding I<enter basic block #nnnn>
51a35ef1
JH
3096book-keeping code to the generated code. During the execution of the
3097code the basic block counters are then updated appropriately. The
3098caveat is that the added extra code can skew the results: again, the
3099profiling tools usually try to factor their own effects out of the
3100results.
3101
83f0ef60
JH
3102=head2 Gprof Profiling
3103
51a35ef1
JH
3104gprof is a profiling tool available in many UNIX platforms,
3105it uses F<statistical time-sampling>.
83f0ef60
JH
3106
3107You can build a profiled version of perl called "perl.gprof" by
51a35ef1
JH
3108invoking the make target "perl.gprof" (What is required is that Perl
3109must be compiled using the C<-pg> flag, you may need to re-Configure).
3110Running the profiled version of Perl will create an output file called
3111F<gmon.out> is created which contains the profiling data collected
3112during the execution.
83f0ef60
JH
3113
3114The gprof tool can then display the collected data in various ways.
3115Usually gprof understands the following options:
3116
3117=over 4
3118
3119=item -a
3120
3121Suppress statically defined functions from the profile.
3122
3123=item -b
3124
3125Suppress the verbose descriptions in the profile.
3126
3127=item -e routine
3128
3129Exclude the given routine and its descendants from the profile.
3130
3131=item -f routine
3132
3133Display only the given routine and its descendants in the profile.
3134
3135=item -s
3136
3137Generate a summary file called F<gmon.sum> which then may be given
3138to subsequent gprof runs to accumulate data over several runs.
3139
3140=item -z
3141
3142Display routines that have zero usage.
3143
3144=back
3145
3146For more detailed explanation of the available commands and output
3147formats, see your own local documentation of gprof.
3148
51a35ef1
JH
3149=head2 GCC gcov Profiling
3150
10f58044 3151Starting from GCC 3.0 I<basic block profiling> is officially available
51a35ef1
JH
3152for the GNU CC.
3153
3154You can build a profiled version of perl called F<perl.gcov> by
3155invoking the make target "perl.gcov" (what is required that Perl must
3156be compiled using gcc with the flags C<-fprofile-arcs
3157-ftest-coverage>, you may need to re-Configure).
3158
3159Running the profiled version of Perl will cause profile output to be
3160generated. For each source file an accompanying ".da" file will be
3161created.
3162
3163To display the results you use the "gcov" utility (which should
3164be installed if you have gcc 3.0 or newer installed). F<gcov> is
3165run on source code files, like this
3166
3167 gcov sv.c
3168
3169which will cause F<sv.c.gcov> to be created. The F<.gcov> files
3170contain the source code annotated with relative frequencies of
3171execution indicated by "#" markers.
3172
3173Useful options of F<gcov> include C<-b> which will summarise the
3174basic block, branch, and function call coverage, and C<-c> which
3175instead of relative frequencies will use the actual counts. For
3176more information on the use of F<gcov> and basic block profiling
3177with gcc, see the latest GNU CC manual, as of GCC 3.0 see
3178
3179 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc.html
3180
3181and its section titled "8. gcov: a Test Coverage Program"
3182
3183 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc_8.html#SEC132
3184
4ae3d70a
JH
3185=head2 Pixie Profiling
3186
51a35ef1
JH
3187Pixie is a profiling tool available on IRIX and Tru64 (aka Digital
3188UNIX aka DEC OSF/1) platforms. Pixie does its profiling using
10f58044 3189I<basic-block counting>.
4ae3d70a 3190
83f0ef60 3191You can build a profiled version of perl called F<perl.pixie> by
51a35ef1
JH
3192invoking the make target "perl.pixie" (what is required is that Perl
3193must be compiled using the C<-g> flag, you may need to re-Configure).
3194
3195In Tru64 a file called F<perl.Addrs> will also be silently created,
3196this file contains the addresses of the basic blocks. Running the
3197profiled version of Perl will create a new file called "perl.Counts"
3198which contains the counts for the basic block for that particular
3199program execution.
4ae3d70a 3200
51a35ef1 3201To display the results you use the F<prof> utility. The exact
4ae3d70a
JH
3202incantation depends on your operating system, "prof perl.Counts" in
3203IRIX, and "prof -pixie -all -L. perl" in Tru64.
3204
6c41479b
JH
3205In IRIX the following prof options are available:
3206
3207=over 4
3208
3209=item -h
3210
3211Reports the most heavily used lines in descending order of use.
6e36760b 3212Useful for finding the hotspot lines.
6c41479b
JH
3213
3214=item -l
3215
3216Groups lines by procedure, with procedures sorted in descending order of use.
3217Within a procedure, lines are listed in source order.
6e36760b 3218Useful for finding the hotspots of procedures.
6c41479b
JH
3219
3220=back
3221
3222In Tru64 the following options are available:
3223
3224=over 4
3225
3958b146 3226=item -p[rocedures]
6c41479b 3227
3958b146 3228Procedures sorted in descending order by the number of cycles executed
6e36760b 3229in each procedure. Useful for finding the hotspot procedures.
6c41479b
JH
3230(This is the default option.)
3231
24000d2f 3232=item -h[eavy]
6c41479b 3233
6e36760b
JH
3234Lines sorted in descending order by the number of cycles executed in
3235each line. Useful for finding the hotspot lines.
6c41479b 3236
24000d2f 3237=item -i[nvocations]
6c41479b 3238
6e36760b
JH
3239The called procedures are sorted in descending order by number of calls
3240made to the procedures. Useful for finding the most used procedures.
6c41479b 3241
24000d2f 3242=item -l[ines]
6c41479b
JH
3243
3244Grouped by procedure, sorted by cycles executed per procedure.
6e36760b 3245Useful for finding the hotspots of procedures.
6c41479b
JH
3246
3247=item -testcoverage
3248
3249The compiler emitted code for these lines, but the code was unexecuted.
3250
24000d2f 3251=item -z[ero]
6c41479b
JH
3252
3253Unexecuted procedures.
3254
aa500c9e 3255=back
6c41479b
JH
3256
3257For further information, see your system's manual pages for pixie and prof.
4ae3d70a 3258
b8ddf6b3
SB
3259=head2 Miscellaneous tricks
3260
3261=over 4
3262
3263=item *
3264
cc177e1a 3265Those debugging perl with the DDD frontend over gdb may find the
b8ddf6b3
SB
3266following useful:
3267
3268You can extend the data conversion shortcuts menu, so for example you
3269can display an SV's IV value with one click, without doing any typing.
3270To do that simply edit ~/.ddd/init file and add after:
3271
3272 ! Display shortcuts.
3273 Ddd*gdbDisplayShortcuts: \
3274 /t () // Convert to Bin\n\
3275 /d () // Convert to Dec\n\
3276 /x () // Convert to Hex\n\
3277 /o () // Convert to Oct(\n\
3278
3279the following two lines:
3280
3281 ((XPV*) (())->sv_any )->xpv_pv // 2pvx\n\
3282 ((XPVIV*) (())->sv_any )->xiv_iv // 2ivx
3283
3284so now you can do ivx and pvx lookups or you can plug there the
3285sv_peek "conversion":
3286
3287 Perl_sv_peek(my_perl, (SV*)()) // sv_peek
3288
3289(The my_perl is for threaded builds.)
3290Just remember that every line, but the last one, should end with \n\
3291
3292Alternatively edit the init file interactively via:
32933rd mouse button -> New Display -> Edit Menu
3294
3295Note: you can define up to 20 conversion shortcuts in the gdb
3296section.
3297
9965345d
JH
3298=item *
3299
7e337ee0
JH
3300If you see in a debugger a memory area mysteriously full of 0xABABABAB
3301or 0xEFEFEFEF, you may be seeing the effect of the Poison() macros,
3302see L<perlclib>.
9965345d 3303
b8ddf6b3
SB
3304=back
3305
955fec6b 3306=head1 CONCLUSION
a422fd2d 3307
955fec6b
JH
3308We've had a brief look around the Perl source, how to maintain quality
3309of the source code, an overview of the stages F<perl> goes through
3310when it's running your code, how to use debuggers to poke at the Perl
3311guts, and finally how to analyse the execution of Perl. We took a very
3312simple problem and demonstrated how to solve it fully - with
3313documentation, regression tests, and finally a patch for submission to
3314p5p. Finally, we talked about how to use external tools to debug and
3315test Perl.
a422fd2d
SC
3316
3317I'd now suggest you read over those references again, and then, as soon
3318as possible, get your hands dirty. The best way to learn is by doing,
3319so:
3320
3321=over 3
3322
3323=item *
3324
3325Subscribe to perl5-porters, follow the patches and try and understand
3326them; don't be afraid to ask if there's a portion you're not clear on -
3327who knows, you may unearth a bug in the patch...
3328
3329=item *
3330
3331Keep up to date with the bleeding edge Perl distributions and get
3332familiar with the changes. Try and get an idea of what areas people are
3333working on and the changes they're making.
3334
3335=item *
3336
3e148164 3337Do read the README associated with your operating system, e.g. README.aix
a1f349fd
MB
3338on the IBM AIX OS. Don't hesitate to supply patches to that README if
3339you find anything missing or changed over a new OS release.
3340
3341=item *
3342
a422fd2d
SC
3343Find an area of Perl that seems interesting to you, and see if you can
3344work out how it works. Scan through the source, and step over it in the
3345debugger. Play, poke, investigate, fiddle! You'll probably get to
3346understand not just your chosen area but a much wider range of F<perl>'s
3347activity as well, and probably sooner than you'd think.
3348
3349=back
3350
3351=over 3
3352
3353=item I<The Road goes ever on and on, down from the door where it began.>
3354
3355=back
3356
64d9b66b 3357If you can do these things, you've started on the long road to Perl porting.
a422fd2d
SC
3358Thanks for wanting to help make Perl better - and happy hacking!
3359
e8cd7eae
GS
3360=head1 AUTHOR
3361
3362This document was written by Nathan Torkington, and is maintained by
3363the perl5-porters mailing list.