This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Reindent bash functions properly
[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
902821cc 453report to I<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
62aa7ed0
RGS
531 cp $1.orig $1
532 vi $1
fcc89a64 533 else
62aa7ed0
RGS
534 vi $1
535 fi
fcc89a64
DR
536 }
537
538 function unedit {
539 if [ -L $1.orig ]; then
540 rm $1
62aa7ed0
RGS
541 mv $1.orig $1
542 fi
fcc89a64
DR
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
62aa7ed0
RGS
552 for f in `find . -name '*.orig' | sed s,^\./,,`
553 do
554 case `echo $f | sed 's,.orig$,,;s,.*\.,,'` in
555 c) diffopts=-p ;;
fcc89a64
DR
556 pod) diffopts='-F^=' ;;
557 *) diffopts= ;;
62aa7ed0
RGS
558 esac
559 diff -du $diffopts $f `echo $f | sed 's,.orig$,,'`
560 done
fcc89a64
DR
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
902821cc
RGS
573There is a single remote administrative interface for modifying bug status,
574category, open issues etc. using the B<RT> bugtracker system, maintained
575by Robert Spier. Become an administrator, and close any bugs you can get
3fd28c4e 576your sticky mitts on:
52315700 577
902821cc 578 http://rt.perl.org/rt3/
52315700 579
3fd28c4e 580To email the bug system administrators:
52315700 581
3fd28c4e 582 "perlbug-admin" <perlbug-admin@perl.org>
52315700 583
a1f349fd
MB
584=head2 Submitting patches
585
f7e1e956
MS
586Always submit patches to I<perl5-porters@perl.org>. If you're
587patching a core module and there's an author listed, send the author a
588copy (see L<Patching a core module>). This lets other porters review
589your patch, which catches a surprising number of errors in patches.
590Either use the diff program (available in source code form from
f224927c 591ftp://ftp.gnu.org/pub/gnu/ , or use Johan Vromans' I<makepatch>
f7e1e956
MS
592(available from I<CPAN/authors/id/JV/>). Unified diffs are preferred,
593but context diffs are accepted. Do not send RCS-style diffs or diffs
594without context lines. More information is given in the
595I<Porting/patching.pod> file in the Perl source distribution. Please
824d470b
RGS
596patch against the latest B<development> version. (e.g., even if you're
597fixing a bug in the 5.8 track, patch against the latest B<development>
598version rsynced from rsync://public.activestate.com/perl-current/ )
599
600If changes are accepted, they are applied to the development branch. Then
601the 5.8 pumpking decides which of those patches is to be backported to the
602maint branch. Only patches that survive the heat of the development
f7e1e956
MS
603branch get applied to maintenance versions.
604
2c8694a7
JH
605Your patch should update the documentation and test suite. See
606L<Writing a test>. If you have added or removed files in the distribution,
607edit the MANIFEST file accordingly, sort the MANIFEST file using
608C<make manisort>, and include those changes as part of your patch.
e8cd7eae 609
824d470b
RGS
610Patching documentation also follows the same order: if accepted, a patch
611is first applied to B<development>, and if relevant then it's backported
612to B<maintenance>. (With an exception for some patches that document
613behaviour that only appears in the maintenance branch, but which has
614changed in the development version.)
615
e8cd7eae
GS
616To report a bug in Perl, use the program I<perlbug> which comes with
617Perl (if you can't get Perl to work, send mail to the address
f18956b7 618I<perlbug@perl.org> or I<perlbug@perl.com>). Reporting bugs through
e8cd7eae 619I<perlbug> feeds into the automated bug-tracking system, access to
902821cc 620which is provided through the web at http://rt.perl.org/rt3/ . It
e8cd7eae
GS
621often pays to check the archives of the perl5-porters mailing list to
622see whether the bug you're reporting has been reported before, and if
623so whether it was considered a bug. See above for the location of
624the searchable archives.
625
f224927c 626The CPAN testers ( http://testers.cpan.org/ ) are a group of
ba139f7d 627volunteers who test CPAN modules on a variety of platforms. Perl
902821cc
RGS
628Smokers ( http://www.nntp.perl.org/group/perl.daily-build )
629automatically test Perl source releases on platforms with various
ba139f7d 630configurations. Both efforts welcome volunteers.
e8cd7eae 631
e8cd7eae
GS
632It's a good idea to read and lurk for a while before chipping in.
633That way you'll get to see the dynamic of the conversations, learn the
634personalities of the players, and hopefully be better prepared to make
635a useful contribution when do you speak up.
636
637If after all this you still think you want to join the perl5-porters
f6c51b38
GS
638mailing list, send mail to I<perl5-porters-subscribe@perl.org>. To
639unsubscribe, send mail to I<perl5-porters-unsubscribe@perl.org>.
e8cd7eae 640
a422fd2d
SC
641To hack on the Perl guts, you'll need to read the following things:
642
643=over 3
644
645=item L<perlguts>
646
647This is of paramount importance, since it's the documentation of what
648goes where in the Perl source. Read it over a couple of times and it
649might start to make sense - don't worry if it doesn't yet, because the
650best way to study it is to read it in conjunction with poking at Perl
651source, and we'll do that later on.
652
653You might also want to look at Gisle Aas's illustrated perlguts -
654there's no guarantee that this will be absolutely up-to-date with the
655latest documentation in the Perl core, but the fundamentals will be
1577cd80 656right. ( http://gisle.aas.no/perl/illguts/ )
a422fd2d
SC
657
658=item L<perlxstut> and L<perlxs>
659
660A working knowledge of XSUB programming is incredibly useful for core
661hacking; XSUBs use techniques drawn from the PP code, the portion of the
662guts that actually executes a Perl program. It's a lot gentler to learn
663those techniques from simple examples and explanation than from the core
664itself.
665
666=item L<perlapi>
667
668The documentation for the Perl API explains what some of the internal
669functions do, as well as the many macros used in the source.
670
671=item F<Porting/pumpkin.pod>
672
673This is a collection of words of wisdom for a Perl porter; some of it is
674only useful to the pumpkin holder, but most of it applies to anyone
675wanting to go about Perl development.
676
677=item The perl5-porters FAQ
678
902821cc
RGS
679This should be available from http://dev.perl.org/perl5/docs/p5p-faq.html .
680It contains hints on reading perl5-porters, information on how
681perl5-porters works and how Perl development in general works.
a422fd2d
SC
682
683=back
684
685=head2 Finding Your Way Around
686
687Perl maintenance can be split into a number of areas, and certain people
688(pumpkins) will have responsibility for each area. These areas sometimes
689correspond to files or directories in the source kit. Among the areas are:
690
691=over 3
692
693=item Core modules
694
695Modules shipped as part of the Perl core live in the F<lib/> and F<ext/>
696subdirectories: F<lib/> is for the pure-Perl modules, and F<ext/>
697contains the core XS modules.
698
f7e1e956
MS
699=item Tests
700
701There are tests for nearly all the modules, built-ins and major bits
702of functionality. Test files all have a .t suffix. Module tests live
703in the F<lib/> and F<ext/> directories next to the module being
704tested. Others live in F<t/>. See L<Writing a test>
705
a422fd2d
SC
706=item Documentation
707
708Documentation maintenance includes looking after everything in the
709F<pod/> directory, (as well as contributing new documentation) and
710the documentation to the modules in core.
711
712=item Configure
713
714The configure process is the way we make Perl portable across the
715myriad of operating systems it supports. Responsibility for the
716configure, build and installation process, as well as the overall
717portability of the core code rests with the configure pumpkin - others
718help out with individual operating systems.
719
720The files involved are the operating system directories, (F<win32/>,
721F<os2/>, F<vms/> and so on) the shell scripts which generate F<config.h>
722and F<Makefile>, as well as the metaconfig files which generate
723F<Configure>. (metaconfig isn't included in the core distribution.)
724
725=item Interpreter
726
727And of course, there's the core of the Perl interpreter itself. Let's
728have a look at that in a little more detail.
729
730=back
731
732Before we leave looking at the layout, though, don't forget that
733F<MANIFEST> contains not only the file names in the Perl distribution,
734but short descriptions of what's in them, too. For an overview of the
735important files, try this:
736
737 perl -lne 'print if /^[^\/]+\.[ch]\s+/' MANIFEST
738
739=head2 Elements of the interpreter
740
741The work of the interpreter has two main stages: compiling the code
742into the internal representation, or bytecode, and then executing it.
743L<perlguts/Compiled code> explains exactly how the compilation stage
744happens.
745
746Here is a short breakdown of perl's operation:
747
748=over 3
749
750=item Startup
751
752The action begins in F<perlmain.c>. (or F<miniperlmain.c> for miniperl)
753This is very high-level code, enough to fit on a single screen, and it
754resembles the code found in L<perlembed>; most of the real action takes
755place in F<perl.c>
756
757First, F<perlmain.c> allocates some memory and constructs a Perl
758interpreter:
759
760 1 PERL_SYS_INIT3(&argc,&argv,&env);
761 2
762 3 if (!PL_do_undump) {
763 4 my_perl = perl_alloc();
764 5 if (!my_perl)
765 6 exit(1);
766 7 perl_construct(my_perl);
767 8 PL_perl_destruct_level = 0;
768 9 }
769
770Line 1 is a macro, and its definition is dependent on your operating
771system. Line 3 references C<PL_do_undump>, a global variable - all
772global variables in Perl start with C<PL_>. This tells you whether the
773current running program was created with the C<-u> flag to perl and then
774F<undump>, which means it's going to be false in any sane context.
775
776Line 4 calls a function in F<perl.c> to allocate memory for a Perl
777interpreter. It's quite a simple function, and the guts of it looks like
778this:
779
780 my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));
781
782Here you see an example of Perl's system abstraction, which we'll see
783later: C<PerlMem_malloc> is either your system's C<malloc>, or Perl's
784own C<malloc> as defined in F<malloc.c> if you selected that option at
785configure time.
786
787Next, in line 7, we construct the interpreter; this sets up all the
788special variables that Perl needs, the stacks, and so on.
789
790Now we pass Perl the command line options, and tell it to go:
791
792 exitstatus = perl_parse(my_perl, xs_init, argc, argv, (char **)NULL);
793 if (!exitstatus) {
794 exitstatus = perl_run(my_perl);
795 }
796
797
798C<perl_parse> is actually a wrapper around C<S_parse_body>, as defined
799in F<perl.c>, which processes the command line options, sets up any
800statically linked XS modules, opens the program and calls C<yyparse> to
801parse it.
802
803=item Parsing
804
805The aim of this stage is to take the Perl source, and turn it into an op
806tree. We'll see what one of those looks like later. Strictly speaking,
807there's three things going on here.
808
809C<yyparse>, the parser, lives in F<perly.c>, although you're better off
810reading the original YACC input in F<perly.y>. (Yes, Virginia, there
811B<is> a YACC grammar for Perl!) The job of the parser is to take your
b432a672 812code and "understand" it, splitting it into sentences, deciding which
a422fd2d
SC
813operands go with which operators and so on.
814
815The parser is nobly assisted by the lexer, which chunks up your input
816into tokens, and decides what type of thing each token is: a variable
817name, an operator, a bareword, a subroutine, a core function, and so on.
818The main point of entry to the lexer is C<yylex>, and that and its
819associated routines can be found in F<toke.c>. Perl isn't much like
820other computer languages; it's highly context sensitive at times, it can
821be tricky to work out what sort of token something is, or where a token
822ends. As such, there's a lot of interplay between the tokeniser and the
823parser, which can get pretty frightening if you're not used to it.
824
825As the parser understands a Perl program, it builds up a tree of
826operations for the interpreter to perform during execution. The routines
827which construct and link together the various operations are to be found
828in F<op.c>, and will be examined later.
829
830=item Optimization
831
832Now the parsing stage is complete, and the finished tree represents
833the operations that the Perl interpreter needs to perform to execute our
834program. Next, Perl does a dry run over the tree looking for
835optimisations: constant expressions such as C<3 + 4> will be computed
836now, and the optimizer will also see if any multiple operations can be
837replaced with a single one. For instance, to fetch the variable C<$foo>,
838instead of grabbing the glob C<*foo> and looking at the scalar
839component, the optimizer fiddles the op tree to use a function which
840directly looks up the scalar in question. The main optimizer is C<peep>
841in F<op.c>, and many ops have their own optimizing functions.
842
843=item Running
844
845Now we're finally ready to go: we have compiled Perl byte code, and all
846that's left to do is run it. The actual execution is done by the
847C<runops_standard> function in F<run.c>; more specifically, it's done by
848these three innocent looking lines:
849
850 while ((PL_op = CALL_FPTR(PL_op->op_ppaddr)(aTHX))) {
851 PERL_ASYNC_CHECK();
852 }
853
854You may be more comfortable with the Perl version of that:
855
856 PERL_ASYNC_CHECK() while $Perl::op = &{$Perl::op->{function}};
857
858Well, maybe not. Anyway, each op contains a function pointer, which
859stipulates the function which will actually carry out the operation.
860This function will return the next op in the sequence - this allows for
861things like C<if> which choose the next op dynamically at run time.
862The C<PERL_ASYNC_CHECK> makes sure that things like signals interrupt
863execution if required.
864
865The actual functions called are known as PP code, and they're spread
b432a672 866between four files: F<pp_hot.c> contains the "hot" code, which is most
a422fd2d
SC
867often used and highly optimized, F<pp_sys.c> contains all the
868system-specific functions, F<pp_ctl.c> contains the functions which
869implement control structures (C<if>, C<while> and the like) and F<pp.c>
870contains everything else. These are, if you like, the C code for Perl's
871built-in functions and operators.
872
dfc98234
DM
873Note that each C<pp_> function is expected to return a pointer to the next
874op. Calls to perl subs (and eval blocks) are handled within the same
875runops loop, and do not consume extra space on the C stack. For example,
876C<pp_entersub> and C<pp_entertry> just push a C<CxSUB> or C<CxEVAL> block
877struct onto the context stack which contain the address of the op
878following the sub call or eval. They then return the first op of that sub
879or eval block, and so execution continues of that sub or block. Later, a
880C<pp_leavesub> or C<pp_leavetry> op pops the C<CxSUB> or C<CxEVAL>,
881retrieves the return op from it, and returns it.
882
883=item Exception handing
884
28a5cf3b 885Perl's exception handing (i.e. C<die> etc) is built on top of the low-level
dfc98234 886C<setjmp()>/C<longjmp()> C-library functions. These basically provide a
28a5cf3b 887way to capture the current PC and SP registers and later restore them; i.e.
dfc98234
DM
888a C<longjmp()> continues at the point in code where a previous C<setjmp()>
889was done, with anything further up on the C stack being lost. This is why
890code should always save values using C<SAVE_FOO> rather than in auto
891variables.
892
893The perl core wraps C<setjmp()> etc in the macros C<JMPENV_PUSH> and
894C<JMPENV_JUMP>. The basic rule of perl exceptions is that C<exit>, and
895C<die> (in the absence of C<eval>) perform a C<JMPENV_JUMP(2)>, while
896C<die> within C<eval> does a C<JMPENV_JUMP(3)>.
897
898At entry points to perl, such as C<perl_parse()>, C<perl_run()> and
899C<call_sv(cv, G_EVAL)> each does a C<JMPENV_PUSH>, then enter a runops
900loop or whatever, and handle possible exception returns. For a 2 return,
901final cleanup is performed, such as popping stacks and calling C<CHECK> or
902C<END> blocks. Amongst other things, this is how scope cleanup still
903occurs during an C<exit>.
904
905If a C<die> can find a C<CxEVAL> block on the context stack, then the
906stack is popped to that level and the return op in that block is assigned
907to C<PL_restartop>; then a C<JMPENV_JUMP(3)> is performed. This normally
908passes control back to the guard. In the case of C<perl_run> and
909C<call_sv>, a non-null C<PL_restartop> triggers re-entry to the runops
910loop. The is the normal way that C<die> or C<croak> is handled within an
911C<eval>.
912
913Sometimes ops are executed within an inner runops loop, such as tie, sort
914or overload code. In this case, something like
915
916 sub FETCH { eval { die } }
917
918would cause a longjmp right back to the guard in C<perl_run>, popping both
919runops loops, which is clearly incorrect. One way to avoid this is for the
920tie code to do a C<JMPENV_PUSH> before executing C<FETCH> in the inner
921runops loop, but for efficiency reasons, perl in fact just sets a flag,
922using C<CATCH_SET(TRUE)>. The C<pp_require>, C<pp_entereval> and
923C<pp_entertry> ops check this flag, and if true, they call C<docatch>,
924which does a C<JMPENV_PUSH> and starts a new runops level to execute the
925code, rather than doing it on the current loop.
926
927As a further optimisation, on exit from the eval block in the C<FETCH>,
928execution of the code following the block is still carried on in the inner
929loop. When an exception is raised, C<docatch> compares the C<JMPENV>
930level of the C<CxEVAL> with C<PL_top_env> and if they differ, just
931re-throws the exception. In this way any inner loops get popped.
932
933Here's an example.
934
935 1: eval { tie @a, 'A' };
936 2: sub A::TIEARRAY {
937 3: eval { die };
938 4: die;
939 5: }
940
941To run this code, C<perl_run> is called, which does a C<JMPENV_PUSH> then
942enters a runops loop. This loop executes the eval and tie ops on line 1,
943with the eval pushing a C<CxEVAL> onto the context stack.
944
945The C<pp_tie> does a C<CATCH_SET(TRUE)>, then starts a second runops loop
946to execute the body of C<TIEARRAY>. When it executes the entertry op on
947line 3, C<CATCH_GET> is true, so C<pp_entertry> calls C<docatch> which
948does a C<JMPENV_PUSH> and starts a third runops loop, which then executes
949the die op. At this point the C call stack looks like this:
950
951 Perl_pp_die
952 Perl_runops # third loop
953 S_docatch_body
954 S_docatch
955 Perl_pp_entertry
956 Perl_runops # second loop
957 S_call_body
958 Perl_call_sv
959 Perl_pp_tie
960 Perl_runops # first loop
961 S_run_body
962 perl_run
963 main
964
965and the context and data stacks, as shown by C<-Dstv>, look like:
966
967 STACK 0: MAIN
968 CX 0: BLOCK =>
969 CX 1: EVAL => AV() PV("A"\0)
970 retop=leave
971 STACK 1: MAGIC
972 CX 0: SUB =>
973 retop=(null)
974 CX 1: EVAL => *
975 retop=nextstate
976
977The die pops the first C<CxEVAL> off the context stack, sets
978C<PL_restartop> from it, does a C<JMPENV_JUMP(3)>, and control returns to
979the top C<docatch>. This then starts another third-level runops level,
980which executes the nextstate, pushmark and die ops on line 4. At the point
981that the second C<pp_die> is called, the C call stack looks exactly like
982that above, even though we are no longer within an inner eval; this is
983because of the optimization mentioned earlier. However, the context stack
984now looks like this, ie with the top CxEVAL popped:
985
986 STACK 0: MAIN
987 CX 0: BLOCK =>
988 CX 1: EVAL => AV() PV("A"\0)
989 retop=leave
990 STACK 1: MAGIC
991 CX 0: SUB =>
992 retop=(null)
993
994The die on line 4 pops the context stack back down to the CxEVAL, leaving
995it as:
996
997 STACK 0: MAIN
998 CX 0: BLOCK =>
999
1000As usual, C<PL_restartop> is extracted from the C<CxEVAL>, and a
1001C<JMPENV_JUMP(3)> done, which pops the C stack back to the docatch:
1002
1003 S_docatch
1004 Perl_pp_entertry
1005 Perl_runops # second loop
1006 S_call_body
1007 Perl_call_sv
1008 Perl_pp_tie
1009 Perl_runops # first loop
1010 S_run_body
1011 perl_run
1012 main
1013
1014In this case, because the C<JMPENV> level recorded in the C<CxEVAL>
1015differs from the current one, C<docatch> just does a C<JMPENV_JUMP(3)>
1016and the C stack unwinds to:
1017
1018 perl_run
1019 main
1020
1021Because C<PL_restartop> is non-null, C<run_body> starts a new runops loop
1022and execution continues.
1023
a422fd2d
SC
1024=back
1025
1026=head2 Internal Variable Types
1027
1028You should by now have had a look at L<perlguts>, which tells you about
1029Perl's internal variable types: SVs, HVs, AVs and the rest. If not, do
1030that now.
1031
1032These variables are used not only to represent Perl-space variables, but
1033also any constants in the code, as well as some structures completely
1034internal to Perl. The symbol table, for instance, is an ordinary Perl
1035hash. Your code is represented by an SV as it's read into the parser;
1036any program files you call are opened via ordinary Perl filehandles, and
1037so on.
1038
1039The core L<Devel::Peek|Devel::Peek> module lets us examine SVs from a
1040Perl program. Let's see, for instance, how Perl treats the constant
1041C<"hello">.
1042
1043 % perl -MDevel::Peek -e 'Dump("hello")'
1044 1 SV = PV(0xa041450) at 0xa04ecbc
1045 2 REFCNT = 1
1046 3 FLAGS = (POK,READONLY,pPOK)
1047 4 PV = 0xa0484e0 "hello"\0
1048 5 CUR = 5
1049 6 LEN = 6
1050
1051Reading C<Devel::Peek> output takes a bit of practise, so let's go
1052through it line by line.
1053
1054Line 1 tells us we're looking at an SV which lives at C<0xa04ecbc> in
1055memory. SVs themselves are very simple structures, but they contain a
1056pointer to a more complex structure. In this case, it's a PV, a
1057structure which holds a string value, at location C<0xa041450>. Line 2
1058is the reference count; there are no other references to this data, so
1059it's 1.
1060
1061Line 3 are the flags for this SV - it's OK to use it as a PV, it's a
1062read-only SV (because it's a constant) and the data is a PV internally.
1063Next we've got the contents of the string, starting at location
1064C<0xa0484e0>.
1065
1066Line 5 gives us the current length of the string - note that this does
1067B<not> include the null terminator. Line 6 is not the length of the
1068string, but the length of the currently allocated buffer; as the string
1069grows, Perl automatically extends the available storage via a routine
1070called C<SvGROW>.
1071
1072You can get at any of these quantities from C very easily; just add
1073C<Sv> to the name of the field shown in the snippet, and you've got a
1074macro which will return the value: C<SvCUR(sv)> returns the current
1075length of the string, C<SvREFCOUNT(sv)> returns the reference count,
1076C<SvPV(sv, len)> returns the string itself with its length, and so on.
1077More macros to manipulate these properties can be found in L<perlguts>.
1078
1079Let's take an example of manipulating a PV, from C<sv_catpvn>, in F<sv.c>
1080
1081 1 void
1082 2 Perl_sv_catpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
1083 3 {
1084 4 STRLEN tlen;
1085 5 char *junk;
1086
1087 6 junk = SvPV_force(sv, tlen);
1088 7 SvGROW(sv, tlen + len + 1);
1089 8 if (ptr == junk)
1090 9 ptr = SvPVX(sv);
1091 10 Move(ptr,SvPVX(sv)+tlen,len,char);
1092 11 SvCUR(sv) += len;
1093 12 *SvEND(sv) = '\0';
1094 13 (void)SvPOK_only_UTF8(sv); /* validate pointer */
1095 14 SvTAINT(sv);
1096 15 }
1097
1098This is a function which adds a string, C<ptr>, of length C<len> onto
1099the end of the PV stored in C<sv>. The first thing we do in line 6 is
1100make sure that the SV B<has> a valid PV, by calling the C<SvPV_force>
1101macro to force a PV. As a side effect, C<tlen> gets set to the current
1102value of the PV, and the PV itself is returned to C<junk>.
1103
b1866b2d 1104In line 7, we make sure that the SV will have enough room to accommodate
a422fd2d
SC
1105the old string, the new string and the null terminator. If C<LEN> isn't
1106big enough, C<SvGROW> will reallocate space for us.
1107
1108Now, if C<junk> is the same as the string we're trying to add, we can
1109grab the string directly from the SV; C<SvPVX> is the address of the PV
1110in the SV.
1111
1112Line 10 does the actual catenation: the C<Move> macro moves a chunk of
1113memory around: we move the string C<ptr> to the end of the PV - that's
1114the start of the PV plus its current length. We're moving C<len> bytes
1115of type C<char>. After doing so, we need to tell Perl we've extended the
1116string, by altering C<CUR> to reflect the new length. C<SvEND> is a
1117macro which gives us the end of the string, so that needs to be a
1118C<"\0">.
1119
1120Line 13 manipulates the flags; since we've changed the PV, any IV or NV
1121values will no longer be valid: if we have C<$a=10; $a.="6";> we don't
1e54db1a 1122want to use the old IV of 10. C<SvPOK_only_utf8> is a special UTF-8-aware
a422fd2d
SC
1123version of C<SvPOK_only>, a macro which turns off the IOK and NOK flags
1124and turns on POK. The final C<SvTAINT> is a macro which launders tainted
1125data if taint mode is turned on.
1126
1127AVs and HVs are more complicated, but SVs are by far the most common
1128variable type being thrown around. Having seen something of how we
1129manipulate these, let's go on and look at how the op tree is
1130constructed.
1131
1132=head2 Op Trees
1133
1134First, what is the op tree, anyway? The op tree is the parsed
1135representation of your program, as we saw in our section on parsing, and
1136it's the sequence of operations that Perl goes through to execute your
1137program, as we saw in L</Running>.
1138
1139An op is a fundamental operation that Perl can perform: all the built-in
1140functions and operators are ops, and there are a series of ops which
1141deal with concepts the interpreter needs internally - entering and
1142leaving a block, ending a statement, fetching a variable, and so on.
1143
1144The op tree is connected in two ways: you can imagine that there are two
1145"routes" through it, two orders in which you can traverse the tree.
1146First, parse order reflects how the parser understood the code, and
1147secondly, execution order tells perl what order to perform the
1148operations in.
1149
1150The easiest way to examine the op tree is to stop Perl after it has
1151finished parsing, and get it to dump out the tree. This is exactly what
7d7d5695
RGS
1152the compiler backends L<B::Terse|B::Terse>, L<B::Concise|B::Concise>
1153and L<B::Debug|B::Debug> do.
a422fd2d
SC
1154
1155Let's have a look at how Perl sees C<$a = $b + $c>:
1156
1157 % perl -MO=Terse -e '$a=$b+$c'
1158 1 LISTOP (0x8179888) leave
1159 2 OP (0x81798b0) enter
1160 3 COP (0x8179850) nextstate
1161 4 BINOP (0x8179828) sassign
1162 5 BINOP (0x8179800) add [1]
1163 6 UNOP (0x81796e0) null [15]
1164 7 SVOP (0x80fafe0) gvsv GV (0x80fa4cc) *b
1165 8 UNOP (0x81797e0) null [15]
1166 9 SVOP (0x8179700) gvsv GV (0x80efeb0) *c
1167 10 UNOP (0x816b4f0) null [15]
1168 11 SVOP (0x816dcf0) gvsv GV (0x80fa460) *a
1169
1170Let's start in the middle, at line 4. This is a BINOP, a binary
1171operator, which is at location C<0x8179828>. The specific operator in
1172question is C<sassign> - scalar assignment - and you can find the code
1173which implements it in the function C<pp_sassign> in F<pp_hot.c>. As a
1174binary operator, it has two children: the add operator, providing the
1175result of C<$b+$c>, is uppermost on line 5, and the left hand side is on
1176line 10.
1177
1178Line 10 is the null op: this does exactly nothing. What is that doing
1179there? If you see the null op, it's a sign that something has been
1180optimized away after parsing. As we mentioned in L</Optimization>,
1181the optimization stage sometimes converts two operations into one, for
1182example when fetching a scalar variable. When this happens, instead of
1183rewriting the op tree and cleaning up the dangling pointers, it's easier
1184just to replace the redundant operation with the null op. Originally,
1185the tree would have looked like this:
1186
1187 10 SVOP (0x816b4f0) rv2sv [15]
1188 11 SVOP (0x816dcf0) gv GV (0x80fa460) *a
1189
1190That is, fetch the C<a> entry from the main symbol table, and then look
1191at the scalar component of it: C<gvsv> (C<pp_gvsv> into F<pp_hot.c>)
1192happens to do both these things.
1193
1194The right hand side, starting at line 5 is similar to what we've just
1195seen: we have the C<add> op (C<pp_add> also in F<pp_hot.c>) add together
1196two C<gvsv>s.
1197
1198Now, what's this about?
1199
1200 1 LISTOP (0x8179888) leave
1201 2 OP (0x81798b0) enter
1202 3 COP (0x8179850) nextstate
1203
1204C<enter> and C<leave> are scoping ops, and their job is to perform any
1205housekeeping every time you enter and leave a block: lexical variables
1206are tidied up, unreferenced variables are destroyed, and so on. Every
1207program will have those first three lines: C<leave> is a list, and its
1208children are all the statements in the block. Statements are delimited
1209by C<nextstate>, so a block is a collection of C<nextstate> ops, with
1210the ops to be performed for each statement being the children of
1211C<nextstate>. C<enter> is a single op which functions as a marker.
1212
1213That's how Perl parsed the program, from top to bottom:
1214
1215 Program
1216 |
1217 Statement
1218 |
1219 =
1220 / \
1221 / \
1222 $a +
1223 / \
1224 $b $c
1225
1226However, it's impossible to B<perform> the operations in this order:
1227you have to find the values of C<$b> and C<$c> before you add them
1228together, for instance. So, the other thread that runs through the op
1229tree is the execution order: each op has a field C<op_next> which points
1230to the next op to be run, so following these pointers tells us how perl
1231executes the code. We can traverse the tree in this order using
1232the C<exec> option to C<B::Terse>:
1233
1234 % perl -MO=Terse,exec -e '$a=$b+$c'
1235 1 OP (0x8179928) enter
1236 2 COP (0x81798c8) nextstate
1237 3 SVOP (0x81796c8) gvsv GV (0x80fa4d4) *b
1238 4 SVOP (0x8179798) gvsv GV (0x80efeb0) *c
1239 5 BINOP (0x8179878) add [1]
1240 6 SVOP (0x816dd38) gvsv GV (0x80fa468) *a
1241 7 BINOP (0x81798a0) sassign
1242 8 LISTOP (0x8179900) leave
1243
1244This probably makes more sense for a human: enter a block, start a
1245statement. Get the values of C<$b> and C<$c>, and add them together.
1246Find C<$a>, and assign one to the other. Then leave.
1247
1248The way Perl builds up these op trees in the parsing process can be
1249unravelled by examining F<perly.y>, the YACC grammar. Let's take the
1250piece we need to construct the tree for C<$a = $b + $c>
1251
1252 1 term : term ASSIGNOP term
1253 2 { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
1254 3 | term ADDOP term
1255 4 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
1256
1257If you're not used to reading BNF grammars, this is how it works: You're
1258fed certain things by the tokeniser, which generally end up in upper
1259case. Here, C<ADDOP>, is provided when the tokeniser sees C<+> in your
1260code. C<ASSIGNOP> is provided when C<=> is used for assigning. These are
b432a672 1261"terminal symbols", because you can't get any simpler than them.
a422fd2d
SC
1262
1263The grammar, lines one and three of the snippet above, tells you how to
b432a672 1264build up more complex forms. These complex forms, "non-terminal symbols"
a422fd2d
SC
1265are generally placed in lower case. C<term> here is a non-terminal
1266symbol, representing a single expression.
1267
1268The grammar gives you the following rule: you can make the thing on the
1269left of the colon if you see all the things on the right in sequence.
1270This is called a "reduction", and the aim of parsing is to completely
1271reduce the input. There are several different ways you can perform a
1272reduction, separated by vertical bars: so, C<term> followed by C<=>
1273followed by C<term> makes a C<term>, and C<term> followed by C<+>
1274followed by C<term> can also make a C<term>.
1275
1276So, if you see two terms with an C<=> or C<+>, between them, you can
1277turn them into a single expression. When you do this, you execute the
1278code in the block on the next line: if you see C<=>, you'll do the code
1279in line 2. If you see C<+>, you'll do the code in line 4. It's this code
1280which contributes to the op tree.
1281
1282 | term ADDOP term
1283 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
1284
1285What this does is creates a new binary op, and feeds it a number of
1286variables. The variables refer to the tokens: C<$1> is the first token in
1287the input, C<$2> the second, and so on - think regular expression
1288backreferences. C<$$> is the op returned from this reduction. So, we
1289call C<newBINOP> to create a new binary operator. The first parameter to
1290C<newBINOP>, a function in F<op.c>, is the op type. It's an addition
1291operator, so we want the type to be C<ADDOP>. We could specify this
1292directly, but it's right there as the second token in the input, so we
b432a672
AL
1293use C<$2>. The second parameter is the op's flags: 0 means "nothing
1294special". Then the things to add: the left and right hand side of our
a422fd2d
SC
1295expression, in scalar context.
1296
1297=head2 Stacks
1298
1299When perl executes something like C<addop>, how does it pass on its
1300results to the next op? The answer is, through the use of stacks. Perl
1301has a number of stacks to store things it's currently working on, and
1302we'll look at the three most important ones here.
1303
1304=over 3
1305
1306=item Argument stack
1307
1308Arguments are passed to PP code and returned from PP code using the
1309argument stack, C<ST>. The typical way to handle arguments is to pop
1310them off the stack, deal with them how you wish, and then push the result
1311back onto the stack. This is how, for instance, the cosine operator
1312works:
1313
1314 NV value;
1315 value = POPn;
1316 value = Perl_cos(value);
1317 XPUSHn(value);
1318
1319We'll see a more tricky example of this when we consider Perl's macros
1320below. C<POPn> gives you the NV (floating point value) of the top SV on
1321the stack: the C<$x> in C<cos($x)>. Then we compute the cosine, and push
1322the result back as an NV. The C<X> in C<XPUSHn> means that the stack
1323should be extended if necessary - it can't be necessary here, because we
1324know there's room for one more item on the stack, since we've just
1325removed one! The C<XPUSH*> macros at least guarantee safety.
1326
1327Alternatively, you can fiddle with the stack directly: C<SP> gives you
1328the first element in your portion of the stack, and C<TOP*> gives you
1329the top SV/IV/NV/etc. on the stack. So, for instance, to do unary
1330negation of an integer:
1331
1332 SETi(-TOPi);
1333
1334Just set the integer value of the top stack entry to its negation.
1335
1336Argument stack manipulation in the core is exactly the same as it is in
1337XSUBs - see L<perlxstut>, L<perlxs> and L<perlguts> for a longer
1338description of the macros used in stack manipulation.
1339
1340=item Mark stack
1341
b432a672 1342I say "your portion of the stack" above because PP code doesn't
a422fd2d
SC
1343necessarily get the whole stack to itself: if your function calls
1344another function, you'll only want to expose the arguments aimed for the
1345called function, and not (necessarily) let it get at your own data. The
b432a672 1346way we do this is to have a "virtual" bottom-of-stack, exposed to each
a422fd2d
SC
1347function. The mark stack keeps bookmarks to locations in the argument
1348stack usable by each function. For instance, when dealing with a tied
b432a672 1349variable, (internally, something with "P" magic) Perl has to call
a422fd2d
SC
1350methods for accesses to the tied variables. However, we need to separate
1351the arguments exposed to the method to the argument exposed to the
ed233832
DM
1352original function - the store or fetch or whatever it may be. Here's
1353roughly how the tied C<push> is implemented; see C<av_push> in F<av.c>:
a422fd2d
SC
1354
1355 1 PUSHMARK(SP);
1356 2 EXTEND(SP,2);
1357 3 PUSHs(SvTIED_obj((SV*)av, mg));
1358 4 PUSHs(val);
1359 5 PUTBACK;
1360 6 ENTER;
1361 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1362 8 LEAVE;
13a2d996 1363
a422fd2d
SC
1364Let's examine the whole implementation, for practice:
1365
1366 1 PUSHMARK(SP);
1367
1368Push the current state of the stack pointer onto the mark stack. This is
1369so that when we've finished adding items to the argument stack, Perl
1370knows how many things we've added recently.
1371
1372 2 EXTEND(SP,2);
1373 3 PUSHs(SvTIED_obj((SV*)av, mg));
1374 4 PUSHs(val);
1375
1376We're going to add two more items onto the argument stack: when you have
1377a tied array, the C<PUSH> subroutine receives the object and the value
1378to be pushed, and that's exactly what we have here - the tied object,
1379retrieved with C<SvTIED_obj>, and the value, the SV C<val>.
1380
1381 5 PUTBACK;
1382
e89a6d4e
JD
1383Next we tell Perl to update the global stack pointer from our internal
1384variable: C<dSP> only gave us a local copy, not a reference to the global.
a422fd2d
SC
1385
1386 6 ENTER;
1387 7 call_method("PUSH", G_SCALAR|G_DISCARD);
1388 8 LEAVE;
1389
1390C<ENTER> and C<LEAVE> localise a block of code - they make sure that all
1391variables are tidied up, everything that has been localised gets
1392its previous value returned, and so on. Think of them as the C<{> and
1393C<}> of a Perl block.
1394
1395To actually do the magic method call, we have to call a subroutine in
1396Perl space: C<call_method> takes care of that, and it's described in
1397L<perlcall>. We call the C<PUSH> method in scalar context, and we're
e89a6d4e
JD
1398going to discard its return value. The call_method() function
1399removes the top element of the mark stack, so there is nothing for
1400the caller to clean up.
a422fd2d 1401
a422fd2d
SC
1402=item Save stack
1403
1404C doesn't have a concept of local scope, so perl provides one. We've
1405seen that C<ENTER> and C<LEAVE> are used as scoping braces; the save
1406stack implements the C equivalent of, for example:
1407
1408 {
1409 local $foo = 42;
1410 ...
1411 }
1412
1413See L<perlguts/Localising Changes> for how to use the save stack.
1414
1415=back
1416
1417=head2 Millions of Macros
1418
1419One thing you'll notice about the Perl source is that it's full of
1420macros. Some have called the pervasive use of macros the hardest thing
1421to understand, others find it adds to clarity. Let's take an example,
1422the code which implements the addition operator:
1423
1424 1 PP(pp_add)
1425 2 {
39644a26 1426 3 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
a422fd2d
SC
1427 4 {
1428 5 dPOPTOPnnrl_ul;
1429 6 SETn( left + right );
1430 7 RETURN;
1431 8 }
1432 9 }
1433
1434Every line here (apart from the braces, of course) contains a macro. The
1435first line sets up the function declaration as Perl expects for PP code;
1436line 3 sets up variable declarations for the argument stack and the
1437target, the return value of the operation. Finally, it tries to see if
1438the addition operation is overloaded; if so, the appropriate subroutine
1439is called.
1440
1441Line 5 is another variable declaration - all variable declarations start
1442with C<d> - which pops from the top of the argument stack two NVs (hence
1443C<nn>) and puts them into the variables C<right> and C<left>, hence the
1444C<rl>. These are the two operands to the addition operator. Next, we
1445call C<SETn> to set the NV of the return value to the result of adding
1446the two values. This done, we return - the C<RETURN> macro makes sure
1447that our return value is properly handled, and we pass the next operator
1448to run back to the main run loop.
1449
1450Most of these macros are explained in L<perlapi>, and some of the more
1451important ones are explained in L<perlxs> as well. Pay special attention
1452to L<perlguts/Background and PERL_IMPLICIT_CONTEXT> for information on
1453the C<[pad]THX_?> macros.
1454
52d59bef
JH
1455=head2 The .i Targets
1456
1457You can expand the macros in a F<foo.c> file by saying
1458
1459 make foo.i
1460
1461which will expand the macros using cpp. Don't be scared by the results.
1462
955fec6b
JH
1463=head1 SOURCE CODE STATIC ANALYSIS
1464
1465Various tools exist for analysing C source code B<statically>, as
1466opposed to B<dynamically>, that is, without executing the code.
1467It is possible to detect resource leaks, undefined behaviour, type
1468mismatches, portability problems, code paths that would cause illegal
1469memory accesses, and other similar problems by just parsing the C code
1470and looking at the resulting graph, what does it tell about the
1471execution and data flows. As a matter of fact, this is exactly
1472how C compilers know to give warnings about dubious code.
1473
1474=head2 lint, splint
1475
1476The good old C code quality inspector, C<lint>, is available in
1477several platforms, but please be aware that there are several
1478different implementations of it by different vendors, which means that
1479the flags are not identical across different platforms.
1480
1481There is a lint variant called C<splint> (Secure Programming Lint)
1482available from http://www.splint.org/ that should compile on any
1483Unix-like platform.
1484
1485There are C<lint> and <splint> targets in Makefile, but you may have
1486to diddle with the flags (see above).
1487
1488=head2 Coverity
1489
1490Coverity (http://www.coverity.com/) is a product similar to lint and
1491as a testbed for their product they periodically check several open
1492source projects, and they give out accounts to open source developers
1493to the defect databases.
1494
1495=head2 cpd (cut-and-paste detector)
1496
1497The cpd tool detects cut-and-paste coding. If one instance of the
1498cut-and-pasted code changes, all the other spots should probably be
1499changed, too. Therefore such code should probably be turned into a
1500subroutine or a macro.
1501
1502cpd (http://pmd.sourceforge.net/cpd.html) is part of the pmd project
1503(http://pmd.sourceforge.net/). pmd was originally written for static
1504analysis of Java code, but later the cpd part of it was extended to
1505parse also C and C++.
1506
a52aaefa
RGS
1507Download the pmd-bin-X.Y.zip () from the SourceForge site, extract the
1508pmd-X.Y.jar from it, and then run that on source code thusly:
955fec6b
JH
1509
1510 java -cp pmd-X.Y.jar net.sourceforge.pmd.cpd.CPD --minimum-tokens 100 --files /some/where/src --language c > cpd.txt
1511
1512You may run into memory limits, in which case you should use the -Xmx option:
1513
1514 java -Xmx512M ...
1515
1516=head2 gcc warnings
1517
1518Though much can be written about the inconsistency and coverage
1519problems of gcc warnings (like C<-Wall> not meaning "all the
1520warnings", or some common portability problems not being covered by
1521C<-Wall>, or C<-ansi> and C<-pedantic> both being a poorly defined
1522collection of warnings, and so forth), gcc is still a useful tool in
1523keeping our coding nose clean.
1524
1525The C<-Wall> is by default on.
1526
a8e98a71
JH
1527The C<-ansi> (and its sidekick, C<-pedantic>) would be nice to be on
1528always, but unfortunately they are not safe on all platforms, they can
1529for example cause fatal conflicts with the system headers (Solaris
1530being a prime example). If Configure C<-Dgccansipedantic> is used,
1531the C<cflags> frontend selects C<-ansi -pedantic> for the platforms
1532where they are known to be safe.
955fec6b
JH
1533
1534Starting from Perl 5.9.4 the following extra flags are added:
1535
1536=over 4
1537
1538=item *
1539
1540C<-Wendif-labels>
1541
1542=item *
1543
1544C<-Wextra>
1545
1546=item *
1547
1548C<-Wdeclaration-after-statement>
1549
1550=back
1551
1552The following flags would be nice to have but they would first need
1553their own Stygian stablemaster:
1554
1555=over 4
1556
1557=item *
1558
1559C<-Wpointer-arith>
1560
1561=item *
1562
1563C<-Wshadow>
1564
1565=item *
1566
1567C<-Wstrict-prototypes>
1568
955fec6b
JH
1569=back
1570
1571The C<-Wtraditional> is another example of the annoying tendency of
1572gcc to bundle a lot of warnings under one switch -- it would be
1573impossible to deploy in practice because it would complain a lot -- but
1574it does contain some warnings that would be beneficial to have available
1575on their own, such as the warning about string constants inside macros
1576containing the macro arguments: this behaved differently pre-ANSI
1577than it does in ANSI, and some C compilers are still in transition,
1578AIX being an example.
1579
1580=head2 Warnings of other C compilers
1581
1582Other C compilers (yes, there B<are> other C compilers than gcc) often
1583have their "strict ANSI" or "strict ANSI with some portability extensions"
1584modes on, like for example the Sun Workshop has its C<-Xa> mode on
1585(though implicitly), or the DEC (these days, HP...) has its C<-std1>
1586mode on.
1587
1588=head2 DEBUGGING
1589
1590You can compile a special debugging version of Perl, which allows you
1591to use the C<-D> option of Perl to tell more about what Perl is doing.
1592But sometimes there is no alternative than to dive in with a debugger,
1593either to see the stack trace of a core dump (very useful in a bug
1594report), or trying to figure out what went wrong before the core dump
1595happened, or how did we end up having wrong or unexpected results.
1596
a422fd2d
SC
1597=head2 Poking at Perl
1598
1599To really poke around with Perl, you'll probably want to build Perl for
1600debugging, like this:
1601
1602 ./Configure -d -D optimize=-g
1603 make
1604
1605C<-g> is a flag to the C compiler to have it produce debugging
955fec6b
JH
1606information which will allow us to step through a running program,
1607and to see in which C function we are at (without the debugging
1608information we might see only the numerical addresses of the functions,
1609which is not very helpful).
1610
a422fd2d
SC
1611F<Configure> will also turn on the C<DEBUGGING> compilation symbol which
1612enables all the internal debugging code in Perl. There are a whole bunch
1613of things you can debug with this: L<perlrun> lists them all, and the
1614best way to find out about them is to play about with them. The most
1615useful options are probably
1616
1617 l Context (loop) stack processing
1618 t Trace execution
1619 o Method and overloading resolution
1620 c String/numeric conversions
1621
1622Some of the functionality of the debugging code can be achieved using XS
1623modules.
13a2d996 1624
a422fd2d
SC
1625 -Dr => use re 'debug'
1626 -Dx => use O 'Debug'
1627
1628=head2 Using a source-level debugger
1629
1630If the debugging output of C<-D> doesn't help you, it's time to step
1631through perl's execution with a source-level debugger.
1632
1633=over 3
1634
1635=item *
1636
955fec6b
JH
1637We'll use C<gdb> for our examples here; the principles will apply to
1638any debugger (many vendors call their debugger C<dbx>), but check the
1639manual of the one you're using.
a422fd2d
SC
1640
1641=back
1642
1643To fire up the debugger, type
1644
1645 gdb ./perl
1646
955fec6b
JH
1647Or if you have a core dump:
1648
1649 gdb ./perl core
1650
a422fd2d
SC
1651You'll want to do that in your Perl source tree so the debugger can read
1652the source code. You should see the copyright message, followed by the
1653prompt.
1654
1655 (gdb)
1656
1657C<help> will get you into the documentation, but here are the most
1658useful commands:
1659
1660=over 3
1661
1662=item run [args]
1663
1664Run the program with the given arguments.
1665
1666=item break function_name
1667
1668=item break source.c:xxx
1669
1670Tells the debugger that we'll want to pause execution when we reach
cea6626f 1671either the named function (but see L<perlguts/Internal Functions>!) or the given
a422fd2d
SC
1672line in the named source file.
1673
1674=item step
1675
1676Steps through the program a line at a time.
1677
1678=item next
1679
1680Steps through the program a line at a time, without descending into
1681functions.
1682
1683=item continue
1684
1685Run until the next breakpoint.
1686
1687=item finish
1688
1689Run until the end of the current function, then stop again.
1690
13a2d996 1691=item 'enter'
a422fd2d
SC
1692
1693Just pressing Enter will do the most recent operation again - it's a
1694blessing when stepping through miles of source code.
1695
1696=item print
1697
1698Execute the given C code and print its results. B<WARNING>: Perl makes
52d59bef
JH
1699heavy use of macros, and F<gdb> does not necessarily support macros
1700(see later L</"gdb macro support">). You'll have to substitute them
1701yourself, or to invoke cpp on the source code files
1702(see L</"The .i Targets">)
1703So, for instance, you can't say
a422fd2d
SC
1704
1705 print SvPV_nolen(sv)
1706
1707but you have to say
1708
1709 print Perl_sv_2pv_nolen(sv)
1710
ffc145e8
RK
1711=back
1712
a422fd2d
SC
1713You may find it helpful to have a "macro dictionary", which you can
1714produce by saying C<cpp -dM perl.c | sort>. Even then, F<cpp> won't
52d59bef
JH
1715recursively apply those macros for you.
1716
1717=head2 gdb macro support
a422fd2d 1718
52d59bef 1719Recent versions of F<gdb> have fairly good macro support, but
ea031e66
RGS
1720in order to use it you'll need to compile perl with macro definitions
1721included in the debugging information. Using F<gcc> version 3.1, this
1722means configuring with C<-Doptimize=-g3>. Other compilers might use a
1723different switch (if they support debugging macros at all).
1724
a422fd2d
SC
1725=head2 Dumping Perl Data Structures
1726
1727One way to get around this macro hell is to use the dumping functions in
1728F<dump.c>; these work a little like an internal
1729L<Devel::Peek|Devel::Peek>, but they also cover OPs and other structures
1730that you can't get at from Perl. Let's take an example. We'll use the
1731C<$a = $b + $c> we used before, but give it a bit of context:
1732C<$b = "6XXXX"; $c = 2.3;>. Where's a good place to stop and poke around?
1733
1734What about C<pp_add>, the function we examined earlier to implement the
1735C<+> operator:
1736
1737 (gdb) break Perl_pp_add
1738 Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.
1739
cea6626f 1740Notice we use C<Perl_pp_add> and not C<pp_add> - see L<perlguts/Internal Functions>.
a422fd2d
SC
1741With the breakpoint in place, we can run our program:
1742
1743 (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'
1744
1745Lots of junk will go past as gdb reads in the relevant source files and
1746libraries, and then:
1747
1748 Breakpoint 1, Perl_pp_add () at pp_hot.c:309
39644a26 1749 309 dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
a422fd2d
SC
1750 (gdb) step
1751 311 dPOPTOPnnrl_ul;
1752 (gdb)
1753
1754We looked at this bit of code before, and we said that C<dPOPTOPnnrl_ul>
1755arranges for two C<NV>s to be placed into C<left> and C<right> - let's
1756slightly expand it:
1757
1758 #define dPOPTOPnnrl_ul NV right = POPn; \
1759 SV *leftsv = TOPs; \
1760 NV left = USE_LEFT(leftsv) ? SvNV(leftsv) : 0.0
1761
1762C<POPn> takes the SV from the top of the stack and obtains its NV either
1763directly (if C<SvNOK> is set) or by calling the C<sv_2nv> function.
1764C<TOPs> takes the next SV from the top of the stack - yes, C<POPn> uses
1765C<TOPs> - but doesn't remove it. We then use C<SvNV> to get the NV from
1766C<leftsv> in the same way as before - yes, C<POPn> uses C<SvNV>.
1767
1768Since we don't have an NV for C<$b>, we'll have to use C<sv_2nv> to
1769convert it. If we step again, we'll find ourselves there:
1770
1771 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
1772 1669 if (!sv)
1773 (gdb)
1774
1775We can now use C<Perl_sv_dump> to investigate the SV:
1776
1777 SV = PV(0xa057cc0) at 0xa0675d0
1778 REFCNT = 1
1779 FLAGS = (POK,pPOK)
1780 PV = 0xa06a510 "6XXXX"\0
1781 CUR = 5
1782 LEN = 6
1783 $1 = void
1784
1785We know we're going to get C<6> from this, so let's finish the
1786subroutine:
1787
1788 (gdb) finish
1789 Run till exit from #0 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
1790 0x462669 in Perl_pp_add () at pp_hot.c:311
1791 311 dPOPTOPnnrl_ul;
1792
1793We can also dump out this op: the current op is always stored in
1794C<PL_op>, and we can dump it with C<Perl_op_dump>. This'll give us
1795similar output to L<B::Debug|B::Debug>.
1796
1797 {
1798 13 TYPE = add ===> 14
1799 TARG = 1
1800 FLAGS = (SCALAR,KIDS)
1801 {
1802 TYPE = null ===> (12)
1803 (was rv2sv)
1804 FLAGS = (SCALAR,KIDS)
1805 {
1806 11 TYPE = gvsv ===> 12
1807 FLAGS = (SCALAR)
1808 GV = main::b
1809 }
1810 }
1811
10f58044 1812# finish this later #
a422fd2d
SC
1813
1814=head2 Patching
1815
1816All right, we've now had a look at how to navigate the Perl sources and
1817some things you'll need to know when fiddling with them. Let's now get
1818on and create a simple patch. Here's something Larry suggested: if a
1819C<U> is the first active format during a C<pack>, (for example,
1820C<pack "U3C8", @stuff>) then the resulting string should be treated as
1e54db1a 1821UTF-8 encoded.
a422fd2d
SC
1822
1823How do we prepare to fix this up? First we locate the code in question -
1824the C<pack> happens at runtime, so it's going to be in one of the F<pp>
1825files. Sure enough, C<pp_pack> is in F<pp.c>. Since we're going to be
1826altering this file, let's copy it to F<pp.c~>.
1827
a6ec74c1
JH
1828[Well, it was in F<pp.c> when this tutorial was written. It has now been
1829split off with C<pp_unpack> to its own file, F<pp_pack.c>]
1830
a422fd2d
SC
1831Now let's look over C<pp_pack>: we take a pattern into C<pat>, and then
1832loop over the pattern, taking each format character in turn into
1833C<datum_type>. Then for each possible format character, we swallow up
1834the other arguments in the pattern (a field width, an asterisk, and so
1835on) and convert the next chunk input into the specified format, adding
1836it onto the output SV C<cat>.
1837
1838How do we know if the C<U> is the first format in the C<pat>? Well, if
1839we have a pointer to the start of C<pat> then, if we see a C<U> we can
1840test whether we're still at the start of the string. So, here's where
1841C<pat> is set up:
1842
1843 STRLEN fromlen;
1844 register char *pat = SvPVx(*++MARK, fromlen);
1845 register char *patend = pat + fromlen;
1846 register I32 len;
1847 I32 datumtype;
1848 SV *fromstr;
1849
1850We'll have another string pointer in there:
1851
1852 STRLEN fromlen;
1853 register char *pat = SvPVx(*++MARK, fromlen);
1854 register char *patend = pat + fromlen;
1855 + char *patcopy;
1856 register I32 len;
1857 I32 datumtype;
1858 SV *fromstr;
1859
1860And just before we start the loop, we'll set C<patcopy> to be the start
1861of C<pat>:
1862
1863 items = SP - MARK;
1864 MARK++;
1865 sv_setpvn(cat, "", 0);
1866 + patcopy = pat;
1867 while (pat < patend) {
1868
1869Now if we see a C<U> which was at the start of the string, we turn on
1e54db1a 1870the C<UTF8> flag for the output SV, C<cat>:
a422fd2d
SC
1871
1872 + if (datumtype == 'U' && pat==patcopy+1)
1873 + SvUTF8_on(cat);
1874 if (datumtype == '#') {
1875 while (pat < patend && *pat != '\n')
1876 pat++;
1877
1878Remember that it has to be C<patcopy+1> because the first character of
1879the string is the C<U> which has been swallowed into C<datumtype!>
1880
1881Oops, we forgot one thing: what if there are spaces at the start of the
1882pattern? C<pack(" U*", @stuff)> will have C<U> as the first active
1883character, even though it's not the first thing in the pattern. In this
1884case, we have to advance C<patcopy> along with C<pat> when we see spaces:
1885
1886 if (isSPACE(datumtype))
1887 continue;
1888
1889needs to become
1890
1891 if (isSPACE(datumtype)) {
1892 patcopy++;
1893 continue;
1894 }
1895
1896OK. That's the C part done. Now we must do two additional things before
1897this patch is ready to go: we've changed the behaviour of Perl, and so
1898we must document that change. We must also provide some more regression
1899tests to make sure our patch works and doesn't create a bug somewhere
1900else along the line.
1901
b23b8711
MS
1902The regression tests for each operator live in F<t/op/>, and so we
1903make a copy of F<t/op/pack.t> to F<t/op/pack.t~>. Now we can add our
1904tests to the end. First, we'll test that the C<U> does indeed create
1905Unicode strings.
1906
1907t/op/pack.t has a sensible ok() function, but if it didn't we could
35c336e6 1908use the one from t/test.pl.
b23b8711 1909
35c336e6
MS
1910 require './test.pl';
1911 plan( tests => 159 );
b23b8711
MS
1912
1913so instead of this:
a422fd2d
SC
1914
1915 print 'not ' unless "1.20.300.4000" eq sprintf "%vd", pack("U*",1,20,300,4000);
1916 print "ok $test\n"; $test++;
1917
35c336e6
MS
1918we can write the more sensible (see L<Test::More> for a full
1919explanation of is() and other testing functions).
b23b8711 1920
35c336e6 1921 is( "1.20.300.4000", sprintf "%vd", pack("U*",1,20,300,4000),
812f5127 1922 "U* produces unicode" );
b23b8711 1923
a422fd2d
SC
1924Now we'll test that we got that space-at-the-beginning business right:
1925
35c336e6 1926 is( "1.20.300.4000", sprintf "%vd", pack(" U*",1,20,300,4000),
812f5127 1927 " with spaces at the beginning" );
a422fd2d
SC
1928
1929And finally we'll test that we don't make Unicode strings if C<U> is B<not>
1930the first active format:
1931
35c336e6 1932 isnt( v1.20.300.4000, sprintf "%vd", pack("C0U*",1,20,300,4000),
812f5127 1933 "U* not first isn't unicode" );
a422fd2d 1934
35c336e6
MS
1935Mustn't forget to change the number of tests which appears at the top,
1936or else the automated tester will get confused. This will either look
1937like this:
a422fd2d 1938
35c336e6
MS
1939 print "1..156\n";
1940
1941or this:
1942
1943 plan( tests => 156 );
a422fd2d
SC
1944
1945We now compile up Perl, and run it through the test suite. Our new
1946tests pass, hooray!
1947
1948Finally, the documentation. The job is never done until the paperwork is
1949over, so let's describe the change we've just made. The relevant place
1950is F<pod/perlfunc.pod>; again, we make a copy, and then we'll insert
1951this text in the description of C<pack>:
1952
1953 =item *
1954
1955 If the pattern begins with a C<U>, the resulting string will be treated
1e54db1a
JH
1956 as UTF-8-encoded Unicode. You can force UTF-8 encoding on in a string
1957 with an initial C<U0>, and the bytes that follow will be interpreted as
1958 Unicode characters. If you don't want this to happen, you can begin your
1959 pattern with C<C0> (or anything else) to force Perl not to UTF-8 encode your
a422fd2d
SC
1960 string, and then follow this with a C<U*> somewhere in your pattern.
1961
1962All done. Now let's create the patch. F<Porting/patching.pod> tells us
1963that if we're making major changes, we should copy the entire directory
1964to somewhere safe before we begin fiddling, and then do
13a2d996 1965
a422fd2d
SC
1966 diff -ruN old new > patch
1967
1968However, we know which files we've changed, and we can simply do this:
1969
1970 diff -u pp.c~ pp.c > patch
1971 diff -u t/op/pack.t~ t/op/pack.t >> patch
1972 diff -u pod/perlfunc.pod~ pod/perlfunc.pod >> patch
1973
1974We end up with a patch looking a little like this:
1975
1976 --- pp.c~ Fri Jun 02 04:34:10 2000
1977 +++ pp.c Fri Jun 16 11:37:25 2000
1978 @@ -4375,6 +4375,7 @@
1979 register I32 items;
1980 STRLEN fromlen;
1981 register char *pat = SvPVx(*++MARK, fromlen);
1982 + char *patcopy;
1983 register char *patend = pat + fromlen;
1984 register I32 len;
1985 I32 datumtype;
1986 @@ -4405,6 +4406,7 @@
1987 ...
1988
1989And finally, we submit it, with our rationale, to perl5-porters. Job
1990done!
1991
f7e1e956
MS
1992=head2 Patching a core module
1993
1994This works just like patching anything else, with an extra
1995consideration. Many core modules also live on CPAN. If this is so,
1996patch the CPAN version instead of the core and send the patch off to
1997the module maintainer (with a copy to p5p). This will help the module
1998maintainer keep the CPAN version in sync with the core version without
1999constantly scanning p5p.
2000
db300100
RGS
2001The list of maintainers of core modules is usefully documented in
2002F<Porting/Maintainers.pl>.
2003
acbe17fc
JP
2004=head2 Adding a new function to the core
2005
2006If, as part of a patch to fix a bug, or just because you have an
2007especially good idea, you decide to add a new function to the core,
2008discuss your ideas on p5p well before you start work. It may be that
2009someone else has already attempted to do what you are considering and
2010can give lots of good advice or even provide you with bits of code
2011that they already started (but never finished).
2012
2013You have to follow all of the advice given above for patching. It is
2014extremely important to test any addition thoroughly and add new tests
2015to explore all boundary conditions that your new function is expected
2016to handle. If your new function is used only by one module (e.g. toke),
2017then it should probably be named S_your_function (for static); on the
210b36aa 2018other hand, if you expect it to accessible from other functions in
acbe17fc
JP
2019Perl, you should name it Perl_your_function. See L<perlguts/Internal Functions>
2020for more details.
2021
2022The location of any new code is also an important consideration. Don't
2023just create a new top level .c file and put your code there; you would
2024have to make changes to Configure (so the Makefile is created properly),
2025as well as possibly lots of include files. This is strictly pumpking
2026business.
2027
2028It is better to add your function to one of the existing top level
2029source code files, but your choice is complicated by the nature of
2030the Perl distribution. Only the files that are marked as compiled
2031static are located in the perl executable. Everything else is located
2032in the shared library (or DLL if you are running under WIN32). So,
2033for example, if a function was only used by functions located in
2034toke.c, then your code can go in toke.c. If, however, you want to call
2035the function from universal.c, then you should put your code in another
2036location, for example util.c.
2037
2038In addition to writing your c-code, you will need to create an
2039appropriate entry in embed.pl describing your function, then run
2040'make regen_headers' to create the entries in the numerous header
2041files that perl needs to compile correctly. See L<perlguts/Internal Functions>
2042for information on the various options that you can set in embed.pl.
2043You will forget to do this a few (or many) times and you will get
2044warnings during the compilation phase. Make sure that you mention
2045this when you post your patch to P5P; the pumpking needs to know this.
2046
2047When you write your new code, please be conscious of existing code
884bad00 2048conventions used in the perl source files. See L<perlstyle> for
acbe17fc
JP
2049details. Although most of the guidelines discussed seem to focus on
2050Perl code, rather than c, they all apply (except when they don't ;).
2051See also I<Porting/patching.pod> file in the Perl source distribution
2052for lots of details about both formatting and submitting patches of
2053your changes.
2054
2055Lastly, TEST TEST TEST TEST TEST any code before posting to p5p.
2056Test on as many platforms as you can find. Test as many perl
2057Configure options as you can (e.g. MULTIPLICITY). If you have
2058profiling or memory tools, see L<EXTERNAL TOOLS FOR DEBUGGING PERL>
210b36aa 2059below for how to use them to further test your code. Remember that
acbe17fc
JP
2060most of the people on P5P are doing this on their own time and
2061don't have the time to debug your code.
f7e1e956
MS
2062
2063=head2 Writing a test
2064
2065Every module and built-in function has an associated test file (or
2066should...). If you add or change functionality, you have to write a
2067test. If you fix a bug, you have to write a test so that bug never
2068comes back. If you alter the docs, it would be nice to test what the
2069new documentation says.
2070
2071In short, if you submit a patch you probably also have to patch the
2072tests.
2073
2074For modules, the test file is right next to the module itself.
2075F<lib/strict.t> tests F<lib/strict.pm>. This is a recent innovation,
2076so there are some snags (and it would be wonderful for you to brush
2077them out), but it basically works that way. Everything else lives in
2078F<t/>.
2079
2080=over 3
2081
2082=item F<t/base/>
2083
2084Testing of the absolute basic functionality of Perl. Things like
2085C<if>, basic file reads and writes, simple regexes, etc. These are
2086run first in the test suite and if any of them fail, something is
2087I<really> broken.
2088
2089=item F<t/cmd/>
2090
2091These test the basic control structures, C<if/else>, C<while>,
35c336e6 2092subroutines, etc.
f7e1e956
MS
2093
2094=item F<t/comp/>
2095
2096Tests basic issues of how Perl parses and compiles itself.
2097
2098=item F<t/io/>
2099
2100Tests for built-in IO functions, including command line arguments.
2101
2102=item F<t/lib/>
2103
2104The old home for the module tests, you shouldn't put anything new in
2105here. There are still some bits and pieces hanging around in here
2106that need to be moved. Perhaps you could move them? Thanks!
2107
2108=item F<t/op/>
2109
2110Tests for perl's built in functions that don't fit into any of the
2111other directories.
2112
2113=item F<t/pod/>
2114
2115Tests for POD directives. There are still some tests for the Pod
2116modules hanging around in here that need to be moved out into F<lib/>.
2117
2118=item F<t/run/>
2119
2120Testing features of how perl actually runs, including exit codes and
2121handling of PERL* environment variables.
2122
244d9cb7
RGS
2123=item F<t/uni/>
2124
2125Tests for the core support of Unicode.
2126
2127=item F<t/win32/>
2128
2129Windows-specific tests.
2130
2131=item F<t/x2p>
2132
2133A test suite for the s2p converter.
2134
f7e1e956
MS
2135=back
2136
2137The core uses the same testing style as the rest of Perl, a simple
2138"ok/not ok" run through Test::Harness, but there are a few special
2139considerations.
2140
35c336e6
MS
2141There are three ways to write a test in the core. Test::More,
2142t/test.pl and ad hoc C<print $test ? "ok 42\n" : "not ok 42\n">. The
2143decision of which to use depends on what part of the test suite you're
2144working on. This is a measure to prevent a high-level failure (such
2145as Config.pm breaking) from causing basic functionality tests to fail.
2146
2147=over 4
2148
2149=item t/base t/comp
2150
2151Since we don't know if require works, or even subroutines, use ad hoc
2152tests for these two. Step carefully to avoid using the feature being
2153tested.
2154
2155=item t/cmd t/run t/io t/op
2156
2157Now that basic require() and subroutines are tested, you can use the
2158t/test.pl library which emulates the important features of Test::More
2159while using a minimum of core features.
2160
2161You can also conditionally use certain libraries like Config, but be
2162sure to skip the test gracefully if it's not there.
2163
2164=item t/lib ext lib
2165
2166Now that the core of Perl is tested, Test::More can be used. You can
2167also use the full suite of core modules in the tests.
2168
2169=back
f7e1e956
MS
2170
2171When you say "make test" Perl uses the F<t/TEST> program to run the
7205a85d
YO
2172test suite (except under Win32 where it uses F<t/harness> instead.)
2173All tests are run from the F<t/> directory, B<not> the directory
2174which contains the test. This causes some problems with the tests
2175in F<lib/>, so here's some opportunity for some patching.
f7e1e956
MS
2176
2177You must be triply conscious of cross-platform concerns. This usually
2178boils down to using File::Spec and avoiding things like C<fork()> and
2179C<system()> unless absolutely necessary.
2180
e018f8be
JH
2181=head2 Special Make Test Targets
2182
2183There are various special make targets that can be used to test Perl
2184slightly differently than the standard "test" target. Not all them
2185are expected to give a 100% success rate. Many of them have several
7205a85d
YO
2186aliases, and many of them are not available on certain operating
2187systems.
e018f8be
JH
2188
2189=over 4
2190
2191=item coretest
2192
7d7d5695 2193Run F<perl> on all core tests (F<t/*> and F<lib/[a-z]*> pragma tests).
e018f8be 2194
7205a85d
YO
2195(Not available on Win32)
2196
e018f8be
JH
2197=item test.deparse
2198
b26492ee
RGS
2199Run all the tests through B::Deparse. Not all tests will succeed.
2200
7205a85d
YO
2201(Not available on Win32)
2202
b26492ee
RGS
2203=item test.taintwarn
2204
2205Run all tests with the B<-t> command-line switch. Not all tests
2206are expected to succeed (until they're specifically fixed, of course).
e018f8be 2207
7205a85d
YO
2208(Not available on Win32)
2209
e018f8be
JH
2210=item minitest
2211
2212Run F<miniperl> on F<t/base>, F<t/comp>, F<t/cmd>, F<t/run>, F<t/io>,
2213F<t/op>, and F<t/uni> tests.
2214
7a834142
JH
2215=item test.valgrind check.valgrind utest.valgrind ucheck.valgrind
2216
2217(Only in Linux) Run all the tests using the memory leak + naughty
2218memory access tool "valgrind". The log files will be named
2219F<testname.valgrind>.
2220
e018f8be
JH
2221=item test.third check.third utest.third ucheck.third
2222
2223(Only in Tru64) Run all the tests using the memory leak + naughty
2224memory access tool "Third Degree". The log files will be named
60a57c1c 2225F<perl.3log.testname>.
e018f8be
JH
2226
2227=item test.torture torturetest
2228
2229Run all the usual tests and some extra tests. As of Perl 5.8.0 the
244d9cb7 2230only extra tests are Abigail's JAPHs, F<t/japh/abigail.t>.
e018f8be
JH
2231
2232You can also run the torture test with F<t/harness> by giving
2233C<-torture> argument to F<t/harness>.
2234
2235=item utest ucheck test.utf8 check.utf8
2236
2237Run all the tests with -Mutf8. Not all tests will succeed.
2238
7205a85d
YO
2239(Not available on Win32)
2240
cc0710ff
RGS
2241=item minitest.utf16 test.utf16
2242
2243Runs the tests with UTF-16 encoded scripts, encoded with different
2244versions of this encoding.
2245
2246C<make utest.utf16> runs the test suite with a combination of C<-utf8> and
2247C<-utf16> arguments to F<t/TEST>.
2248
7205a85d
YO
2249(Not available on Win32)
2250
244d9cb7
RGS
2251=item test_harness
2252
2253Run the test suite with the F<t/harness> controlling program, instead of
2254F<t/TEST>. F<t/harness> is more sophisticated, and uses the
2255L<Test::Harness> module, thus using this test target supposes that perl
2256mostly works. The main advantage for our purposes is that it prints a
00bf5cd9
RGS
2257detailed summary of failed tests at the end. Also, unlike F<t/TEST>, it
2258doesn't redirect stderr to stdout.
244d9cb7 2259
7205a85d
YO
2260Note that under Win32 F<t/harness> is always used instead of F<t/TEST>, so
2261there is no special "test_harness" target.
2262
2263Under Win32's "test" target you may use the TEST_SWITCHES and TEST_FILES
2264environment variables to control the behaviour of F<t/harness>. This means
2265you can say
2266
2267 nmake test TEST_FILES="op/*.t"
2268 nmake test TEST_SWITCHES="-torture" TEST_FILES="op/*.t"
2269
2270=item test-notty test_notty
2271
2272Sets PERL_SKIP_TTY_TEST to true before running normal test.
2273
244d9cb7
RGS
2274=back
2275
2276=head2 Running tests by hand
2277
2278You can run part of the test suite by hand by using one the following
2279commands from the F<t/> directory :
2280
2281 ./perl -I../lib TEST list-of-.t-files
2282
2283or
2284
2285 ./perl -I../lib harness list-of-.t-files
2286
2287(if you don't specify test scripts, the whole test suite will be run.)
2288
7205a85d
YO
2289=head3 Using t/harness for testing
2290
2291If you use C<harness> for testing you have several command line options
2292available to you. The arguments are as follows, and are in the order
2293that they must appear if used together.
2294
2295 harness -v -torture -re=pattern LIST OF FILES TO TEST
2296 harness -v -torture -re LIST OF PATTERNS TO MATCH
2297
2298If C<LIST OF FILES TO TEST> is omitted the file list is obtained from
2299the manifest. The file list may include shell wildcards which will be
2300expanded out.
2301
2302=over 4
2303
2304=item -v
2305
2306Run the tests under verbose mode so you can see what tests were run,
2307and debug outbut.
2308
2309=item -torture
2310
2311Run the torture tests as well as the normal set.
2312
2313=item -re=PATTERN
2314
2315Filter the file list so that all the test files run match PATTERN.
2316Note that this form is distinct from the B<-re LIST OF PATTERNS> form below
2317in that it allows the file list to be provided as well.
2318
2319=item -re LIST OF PATTERNS
2320
2321Filter the file list so that all the test files run match
2322/(LIST|OF|PATTERNS)/. Note that with this form the patterns
2323are joined by '|' and you cannot supply a list of files, instead
2324the test files are obtained from the MANIFEST.
2325
2326=back
2327
244d9cb7
RGS
2328You can run an individual test by a command similar to
2329
2330 ./perl -I../lib patho/to/foo.t
2331
2332except that the harnesses set up some environment variables that may
2333affect the execution of the test :
2334
2335=over 4
2336
2337=item PERL_CORE=1
2338
2339indicates that we're running this test part of the perl core test suite.
2340This is useful for modules that have a dual life on CPAN.
2341
2342=item PERL_DESTRUCT_LEVEL=2
2343
2344is set to 2 if it isn't set already (see L</PERL_DESTRUCT_LEVEL>)
2345
2346=item PERL
2347
2348(used only by F<t/TEST>) if set, overrides the path to the perl executable
2349that should be used to run the tests (the default being F<./perl>).
2350
2351=item PERL_SKIP_TTY_TEST
2352
2353if set, tells to skip the tests that need a terminal. It's actually set
2354automatically by the Makefile, but can also be forced artificially by
2355running 'make test_notty'.
2356
e018f8be 2357=back
f7e1e956 2358
d7889f52
JH
2359=head2 Common problems when patching Perl source code
2360
2361Perl source plays by ANSI C89 rules: no C99 (or C++) extensions. In
2362some cases we have to take pre-ANSI requirements into consideration.
2363You don't care about some particular platform having broken Perl?
2364I hear there is still a strong demand for J2EE programmers.
2365
2366=head2 Perl environment problems
2367
2368=over 4
2369
2370=item *
2371
2372Not compiling with threading
2373
2374Compiling with threading (-Duseithreads) completely rewrites
2375the function prototypes of Perl. You better try your changes
0bec6c03 2376with that. Related to this is the difference between "Perl_-less"
d7889f52
JH
2377and "Perl_-ly" APIs, for example:
2378
2379 Perl_sv_setiv(aTHX_ ...);
2380 sv_setiv(...);
2381
ee9468a2
RGS
2382The first one explicitly passes in the context, which is needed for e.g.
2383threaded builds. The second one does that implicitly; do not get them
def4ed7d
JH
2384mixed. If you are not passing in a aTHX_, you will need to do a dTHX
2385(or a dVAR) as the first thing in the function.
d7889f52
JH
2386
2387See L<perlguts/"How multiple interpreters and concurrency are supported">
2388for further discussion about context.
2389
2390=item *
2391
2392Not compiling with -DDEBUGGING
2393
2394The DEBUGGING define exposes more code to the compiler,
0bec6c03 2395therefore more ways for things to go wrong. You should try it.
d7889f52
JH
2396
2397=item *
2398
ee9468a2
RGS
2399Introducing (non-read-only) globals
2400
2401Do not introduce any modifiable globals, truly global or file static.
bc028b6b
JH
2402They are bad form and complicate multithreading and other forms of
2403concurrency. The right way is to introduce them as new interpreter
2404variables, see F<intrpvar.h> (at the very end for binary compatibility).
ee9468a2
RGS
2405
2406Introducing read-only (const) globals is okay, as long as you verify
2407with e.g. C<nm libperl.a|egrep -v ' [TURtr] '> (if your C<nm> has
2408BSD-style output) that the data you added really is read-only.
2409(If it is, it shouldn't show up in the output of that command.)
2410
2411If you want to have static strings, make them constant:
2412
2413 static const char etc[] = "...";
2414
bc028b6b 2415If you want to have arrays of constant strings, note carefully
ee9468a2
RGS
2416the right combination of C<const>s:
2417
2418 static const char * const yippee[] =
2419 {"hi", "ho", "silver"};
2420
bc028b6b
JH
2421There is a way to completely hide any modifiable globals (they are all
2422moved to heap), the compilation setting C<-DPERL_GLOBAL_STRUCT_PRIVATE>.
2423It is not normally used, but can be used for testing, read more
def4ed7d 2424about it in L<perlguts/"Background and PERL_IMPLICIT_CONTEXT">.
bc028b6b 2425
ee9468a2
RGS
2426=item *
2427
d7889f52
JH
2428Not exporting your new function
2429
2430Some platforms (Win32, AIX, VMS, OS/2, to name a few) require any
2431function that is part of the public API (the shared Perl library)
2432to be explicitly marked as exported. See the discussion about
2433F<embed.pl> in L<perlguts>.
2434
2435=item *
2436
2437Exporting your new function
2438
2439The new shiny result of either genuine new functionality or your
2440arduous refactoring is now ready and correctly exported. So what
def4ed7d 2441could possibly go wrong?
d7889f52
JH
2442
2443Maybe simply that your function did not need to be exported in the
2444first place. Perl has a long and not so glorious history of exporting
2445functions that it should not have.
2446
2447If the function is used only inside one source code file, make it
2448static. See the discussion about F<embed.pl> in L<perlguts>.
2449
2450If the function is used across several files, but intended only for
2451Perl's internal use (and this should be the common case), do not
2452export it to the public API. See the discussion about F<embed.pl>
2453in L<perlguts>.
2454
2455=back
2456
5b38b9cd 2457=head2 Portability problems
d7889f52
JH
2458
2459The following are common causes of compilation and/or execution
2460failures, not common to Perl as such. The C FAQ is good bedtime
0bec6c03
JH
2461reading. Please test your changes with as many C compilers and
2462platforms as possible -- we will, anyway, and it's nice to save
2463oneself from public embarrassment.
2464
9aaf14db
RGS
2465If using gcc, you can add the C<-std=c89> option which will hopefully
2466catch most of these unportabilities. (However it might also catch
2467incompatibilities in your system's header files.)
d1307786 2468
a8e98a71
JH
2469Use the Configure C<-Dgccansipedantic> flag to enable the gcc
2470C<-ansi -pedantic> flags which enforce stricter ANSI rules.
2471
def4ed7d
JH
2472If using the C<gcc -Wall> note that not all the possible warnings
2473(like C<-Wunitialized>) are given unless you also compile with C<-O>.
2474
2475Note that if using gcc, starting from Perl 5.9.5 the Perl core source
2476code files (the ones at the top level of the source code distribution,
2477but not e.g. the extensions under ext/) are automatically compiled
2478with as many as possible of the C<-std=c89>, C<-ansi>, C<-pedantic>,
2479and a selection of C<-W> flags (see cflags.SH).
27565cb6 2480
0bec6c03 2481Also study L<perlport> carefully to avoid any bad assumptions
def4ed7d 2482about the operating system, filesystems, and so forth.
0bec6c03 2483
606fd33d 2484You may once in a while try a "make microperl" to see whether we
63796a85 2485can still compile Perl with just the bare minimum of interfaces.
606fd33d 2486(See README.micro.)
ee9468a2 2487
0bec6c03 2488Do not assume an operating system indicates a certain compiler.
d7889f52
JH
2489
2490=over 4
2491
2492=item *
2493
2494Casting pointers to integers or casting integers to pointers
2495
2496 void castaway(U8* p)
2497 {
2498 IV i = p;
2499
2500or
2501
2502 void castaway(U8* p)
2503 {
2504 IV i = (IV)p;
2505
ee9468a2 2506Both are bad, and broken, and unportable. Use the PTR2IV()
d7889f52
JH
2507macro that does it right. (Likewise, there are PTR2UV(), PTR2NV(),
2508INT2PTR(), and NUM2PTR().)
2509
2510=item *
2511
0bec6c03
JH
2512Casting between data function pointers and data pointers
2513
d7889f52
JH
2514Technically speaking casting between function pointers and data
2515pointers is unportable and undefined, but practically speaking
2516it seems to work, but you should use the FPTR2DPTR() and DPTR2FPTR()
0bec6c03 2517macros. Sometimes you can also play games with unions.
d7889f52
JH
2518
2519=item *
2520
2521Assuming sizeof(int) == sizeof(long)
2522
2523There are platforms where longs are 64 bits, and platforms where ints
2524are 64 bits, and while we are out to shock you, even platforms where
2525shorts are 64 bits. This is all legal according to the C standard.
2526(In other words, "long long" is not a portable way to specify 64 bits,
2527and "long long" is not even guaranteed to be any wider than "long".)
63796a85
JH
2528
2529Instead, use the definitions IV, UV, IVSIZE, I32SIZE, and so forth.
2530Avoid things like I32 because they are B<not> guaranteed to be
2531I<exactly> 32 bits, they are I<at least> 32 bits, nor are they
2532guaranteed to be B<int> or B<long>. If you really explicitly need
253364-bit variables, use I64 and U64, but only if guarded by HAS_QUAD.
d7889f52
JH
2534
2535=item *
2536
2537Assuming one can dereference any type of pointer for any type of data
2538
2539 char *p = ...;
def4ed7d 2540 long pony = *p; /* BAD */
d7889f52
JH
2541
2542Many platforms, quite rightly so, will give you a core dump instead
2543of a pony if the p happens not be correctly aligned.
2544
2545=item *
2546
2547Lvalue casts
2548
def4ed7d 2549 (int)*p = ...; /* BAD */
d7889f52
JH
2550
2551Simply not portable. Get your lvalue to be of the right type,
27565cb6
JH
2552or maybe use temporary variables, or dirty tricks with unions.
2553
2554=item *
2555
606fd33d
JH
2556Assume B<anything> about structs (especially the ones you
2557don't control, like the ones coming from the system headers)
27565cb6
JH
2558
2559=over 8
2560
2561=item *
2562
2563That a certain field exists in a struct
2564
2565=item *
2566
902821cc 2567That no other fields exist besides the ones you know of
27565cb6
JH
2568
2569=item *
2570
606fd33d 2571That a field is of certain signedness, sizeof, or type
27565cb6
JH
2572
2573=item *
2574
2575That the fields are in a certain order
2576
606fd33d
JH
2577=over 8
2578
27565cb6
JH
2579=item *
2580
606fd33d
JH
2581While C guarantees the ordering specified in the struct definition,
2582between different platforms the definitions might differ
2583
2584=back
27565cb6
JH
2585
2586=item *
2587
606fd33d
JH
2588That the sizeof(struct) or the alignments are the same everywhere
2589
2590=over 8
27565cb6
JH
2591
2592=item *
2593
606fd33d
JH
2594There might be padding bytes between the fields to align the fields -
2595the bytes can be anything
2596
2597=item *
2598
2599Structs are required to be aligned to the maximum alignment required
2600by the fields - which for native types is for usually equivalent to
2601sizeof() of the field
2602
2603=back
27565cb6
JH
2604
2605=back
d7889f52
JH
2606
2607=item *
2608
0bec6c03
JH
2609Mixing #define and #ifdef
2610
2611 #define BURGLE(x) ... \
def4ed7d 2612 #ifdef BURGLE_OLD_STYLE /* BAD */
0bec6c03
JH
2613 ... do it the old way ... \
2614 #else
2615 ... do it the new way ... \
2616 #endif
2617
ee9468a2
RGS
2618You cannot portably "stack" cpp directives. For example in the above
2619you need two separate BURGLE() #defines, one for each #ifdef branch.
2620
2621=item *
2622
2623Adding stuff after #endif or #else
2624
2625 #ifdef SNOSH
2626 ...
def4ed7d 2627 #else !SNOSH /* BAD */
ee9468a2 2628 ...
def4ed7d 2629 #endif SNOSH /* BAD */
ee9468a2 2630
def4ed7d
JH
2631The #endif and #else cannot portably have anything non-comment after
2632them. If you want to document what is going (which is a good idea
2633especially if the branches are long), use (C) comments:
ee9468a2
RGS
2634
2635 #ifdef SNOSH
2636 ...
2637 #else /* !SNOSH */
2638 ...
2639 #endif /* SNOSH */
2640
2641The gcc option C<-Wendif-labels> warns about the bad variant
2642(by default on starting from Perl 5.9.4).
0bec6c03
JH
2643
2644=item *
2645
27565cb6
JH
2646Having a comma after the last element of an enum list
2647
2648 enum color {
2649 CERULEAN,
2650 CHARTREUSE,
def4ed7d 2651 CINNABAR, /* BAD */
27565cb6
JH
2652 };
2653
2654is not portable. Leave out the last comma.
2655
2656Also note that whether enums are implicitly morphable to ints
2657varies between compilers, you might need to (int).
2658
2659=item *
2660
d7889f52
JH
2661Using //-comments
2662
def4ed7d 2663 // This function bamfoodles the zorklator. /* BAD */
d7889f52
JH
2664
2665That is C99 or C++. Perl is C89. Using the //-comments is silently
0bec6c03
JH
2666allowed by many C compilers but cranking up the ANSI C89 strictness
2667(which we like to do) causes the compilation to fail.
d7889f52
JH
2668
2669=item *
2670
2671Mixing declarations and code
2672
2673 void zorklator()
2674 {
2675 int n = 3;
def4ed7d 2676 set_zorkmids(n); /* BAD */
d7889f52
JH
2677 int q = 4;
2678
0bec6c03
JH
2679That is C99 or C++. Some C compilers allow that, but you shouldn't.
2680
63796a85
JH
2681The gcc option C<-Wdeclaration-after-statements> scans for such problems
2682(by default on starting from Perl 5.9.4).
2683
0bec6c03
JH
2684=item *
2685
2686Introducing variables inside for()
2687
def4ed7d 2688 for(int i = ...; ...; ...) { /* BAD */
0bec6c03
JH
2689
2690That is C99 or C++. While it would indeed be awfully nice to have that
2691also in C89, to limit the scope of the loop variable, alas, we cannot.
d7889f52
JH
2692
2693=item *
2694
2695Mixing signed char pointers with unsigned char pointers
2696
2697 int foo(char *s) { ... }
2698 ...
2699 unsigned char *t = ...; /* Or U8* t = ... */
def4ed7d 2700 foo(t); /* BAD */
d7889f52
JH
2701
2702While this is legal practice, it is certainly dubious, and downright
2703fatal in at least one platform: for example VMS cc considers this a
def4ed7d
JH
2704fatal error. One cause for people often making this mistake is that a
2705"naked char" and therefore dereferencing a "naked char pointer" have
2706an undefined signedness: it depends on the compiler and the flags of
2707the compiler and the underlying platform whether the result is signed
2708or unsigned. For this very same reason using a 'char' as an array
2709index is bad.
d7889f52
JH
2710
2711=item *
2712
2713Macros that have string constants and their arguments as substrings of
2714the string constants
2715
def4ed7d 2716 #define FOO(n) printf("number = %d\n", n) /* BAD */
d7889f52
JH
2717 FOO(10);
2718
2719Pre-ANSI semantics for that was equivalent to
2720
2721 printf("10umber = %d\10");
2722
0bec6c03
JH
2723which is probably not what you were expecting. Unfortunately at least
2724one reasonably common and modern C compiler does "real backward
63796a85 2725compatibility" here, in AIX that is what still happens even though the
0bec6c03
JH
2726rest of the AIX compiler is very happily C89.
2727
2728=item *
2729
ee9468a2
RGS
2730Using printf formats for non-basic C types
2731
2732 IV i = ...;
def4ed7d 2733 printf("i = %d\n", i); /* BAD */
ee9468a2
RGS
2734
2735While this might by accident work in some platform (where IV happens
2736to be an C<int>), in general it cannot. IV might be something larger.
2737Even worse the situation is with more specific types (defined by Perl's
2738configuration step in F<config.h>):
2739
2740 Uid_t who = ...;
def4ed7d 2741 printf("who = %d\n", who); /* BAD */
ee9468a2
RGS
2742
2743The problem here is that Uid_t might be not only not C<int>-wide
2744but it might also be unsigned, in which case large uids would be
2745printed as negative values.
2746
2747There is no simple solution to this because of printf()'s limited
2748intelligence, but for many types the right format is available as
2749with either 'f' or '_f' suffix, for example:
2750
2751 IVdf /* IV in decimal */
2752 UVxf /* UV is hexadecimal */
2753
2754 printf("i = %"IVdf"\n", i); /* The IVdf is a string constant. */
2755
2756 Uid_t_f /* Uid_t in decimal */
2757
2758 printf("who = %"Uid_t_f"\n", who);
2759
63796a85
JH
2760Or you can try casting to a "wide enough" type:
2761
2762 printf("i = %"IVdf"\n", (IV)something_very_small_and_signed);
2763
2764Also remember that the C<%p> format really does require a void pointer:
2765
2766 U8* p = ...;
2767 printf("p = %p\n", (void*)p);
2768
ee9468a2
RGS
2769The gcc option C<-Wformat> scans for such problems.
2770
2771=item *
2772
0bec6c03
JH
2773Blindly using variadic macros
2774
63796a85
JH
2775gcc has had them for a while with its own syntax, and C99 brought
2776them with a standardized syntax. Don't use the former, and use
2777the latter only if the HAS_C99_VARIADIC_MACROS is defined.
0bec6c03
JH
2778
2779=item *
2780
2781Blindly passing va_list
2782
2783Not all platforms support passing va_list to further varargs (stdarg)
2784functions. The right thing to do is to copy the va_list using the
2785Perl_va_copy() if the NEED_VA_COPY is defined.
d7889f52 2786
ee9468a2
RGS
2787=item *
2788
e5afc1ae 2789Using gcc statement expressions
63796a85 2790
def4ed7d 2791 val = ({...;...;...}); /* BAD */
63796a85 2792
def4ed7d
JH
2793While a nice extension, it's not portable. The Perl code does
2794admittedly use them if available to gain some extra speed
2795(essentially as a funky form of inlining), but you shouldn't.
63796a85
JH
2796
2797=item *
2798
2799Binding together several statements
2800
2801Use the macros STMT_START and STMT_END.
2802
2803 STMT_START {
2804 ...
2805 } STMT_END
2806
2807=item *
2808
ee9468a2
RGS
2809Testing for operating systems or versions when should be testing for features
2810
def4ed7d 2811 #ifdef __FOONIX__ /* BAD */
ee9468a2
RGS
2812 foo = quux();
2813 #endif
2814
2815Unless you know with 100% certainty that quux() is only ever available
2816for the "Foonix" operating system B<and> that is available B<and>
2817correctly working for B<all> past, present, B<and> future versions of
2818"Foonix", the above is very wrong. This is more correct (though still
2819not perfect, because the below is a compile-time check):
2820
2821 #ifdef HAS_QUUX
2822 foo = quux();
2823 #endif
2824
def4ed7d 2825How does the HAS_QUUX become defined where it needs to be? Well, if
ee9468a2
RGS
2826Foonix happens to be UNIXy enought to be able to run the Configure
2827script, and Configure has been taught about detecting and testing
2828quux(), the HAS_QUUX will be correctly defined. In other platforms,
2829the corresponding configuration step will hopefully do the same.
2830
2831In a pinch, if you cannot wait for Configure to be educated,
2832or if you have a good hunch of where quux() might be available,
2833you can temporarily try the following:
2834
2835 #if (defined(__FOONIX__) || defined(__BARNIX__))
2836 # define HAS_QUUX
2837 #endif
2838
2839 ...
2840
2841 #ifdef HAS_QUUX
2842 foo = quux();
2843 #endif
2844
2845But in any case, try to keep the features and operating systems separate.
2846
d7889f52
JH
2847=back
2848
ad7244db
JH
2849=head2 Problematic System Interfaces
2850
2851=over 4
2852
2853=item *
2854
2855malloc(0), realloc(0), calloc(0, 0) are nonportable. To be portable
2856allocate at least one byte. (In general you should rarely need to
2857work at this low level, but instead use the various malloc wrappers.)
2858
2859=item *
2860
2861snprintf() - the return type is unportable. Use my_snprintf() instead.
2862
2863=back
2864
d7889f52
JH
2865=head2 Security problems
2866
2867Last but not least, here are various tips for safer coding.
2868
2869=over 4
2870
2871=item *
2872
2873Do not use gets()
2874
2875Or we will publicly ridicule you. Seriously.
2876
2877=item *
2878
d1307786 2879Do not use strcpy() or strcat() or strncpy() or strncat()
d7889f52 2880
d1307786
JH
2881Use my_strlcpy() and my_strlcat() instead: they either use the native
2882implementation, or Perl's own implementation (borrowed from the public
2883domain implementation of INN).
d7889f52
JH
2884
2885=item *
2886
2887Do not use sprintf() or vsprintf()
2888
0bec6c03 2889If you really want just plain byte strings, use my_snprintf()
64d9b66b 2890and my_vsnprintf() instead, which will try to use snprintf() and
0bec6c03
JH
2891vsnprintf() if those safer APIs are available. If you want something
2892fancier than a plain byte string, use SVs and Perl_sv_catpvf().
d7889f52
JH
2893
2894=back
2895
902b9dbf
MF
2896=head1 EXTERNAL TOOLS FOR DEBUGGING PERL
2897
2898Sometimes it helps to use external tools while debugging and
2899testing Perl. This section tries to guide you through using
2900some common testing and debugging tools with Perl. This is
2901meant as a guide to interfacing these tools with Perl, not
2902as any kind of guide to the use of the tools themselves.
2903
a958818a
JH
2904B<NOTE 1>: Running under memory debuggers such as Purify, valgrind, or
2905Third Degree greatly slows down the execution: seconds become minutes,
2906minutes become hours. For example as of Perl 5.8.1, the
2907ext/Encode/t/Unicode.t takes extraordinarily long to complete under
2908e.g. Purify, Third Degree, and valgrind. Under valgrind it takes more
2909than six hours, even on a snappy computer-- the said test must be
2910doing something that is quite unfriendly for memory debuggers. If you
2911don't feel like waiting, that you can simply kill away the perl
2912process.
2913
2914B<NOTE 2>: To minimize the number of memory leak false alarms (see
2915L</PERL_DESTRUCT_LEVEL> for more information), you have to have
2916environment variable PERL_DESTRUCT_LEVEL set to 2. The F<TEST>
2917and harness scripts do that automatically. But if you are running
2918some of the tests manually-- for csh-like shells:
2919
2920 setenv PERL_DESTRUCT_LEVEL 2
2921
2922and for Bourne-type shells:
2923
2924 PERL_DESTRUCT_LEVEL=2
2925 export PERL_DESTRUCT_LEVEL
2926
2927or in UNIXy environments you can also use the C<env> command:
2928
2929 env PERL_DESTRUCT_LEVEL=2 valgrind ./perl -Ilib ...
a1b65709 2930
37c0adeb
JH
2931B<NOTE 3>: There are known memory leaks when there are compile-time
2932errors within eval or require, seeing C<S_doeval> in the call stack
2933is a good sign of these. Fixing these leaks is non-trivial,
2934unfortunately, but they must be fixed eventually.
2935
f50e5b73
MH
2936B<NOTE 4>: L<DynaLoader> will not clean up after itself completely
2937unless Perl is built with the Configure option
2938C<-Accflags=-DDL_UNLOAD_ALL_AT_EXIT>.
2939
902b9dbf
MF
2940=head2 Rational Software's Purify
2941
2942Purify is a commercial tool that is helpful in identifying
2943memory overruns, wild pointers, memory leaks and other such
2944badness. Perl must be compiled in a specific way for
2945optimal testing with Purify. Purify is available under
2946Windows NT, Solaris, HP-UX, SGI, and Siemens Unix.
2947
902b9dbf
MF
2948=head2 Purify on Unix
2949
2950On Unix, Purify creates a new Perl binary. To get the most
2951benefit out of Purify, you should create the perl to Purify
2952using:
2953
2954 sh Configure -Accflags=-DPURIFY -Doptimize='-g' \
2955 -Uusemymalloc -Dusemultiplicity
2956
2957where these arguments mean:
2958
2959=over 4
2960
2961=item -Accflags=-DPURIFY
2962
2963Disables Perl's arena memory allocation functions, as well as
2964forcing use of memory allocation functions derived from the
2965system malloc.
2966
2967=item -Doptimize='-g'
2968
2969Adds debugging information so that you see the exact source
2970statements where the problem occurs. Without this flag, all
2971you will see is the source filename of where the error occurred.
2972
2973=item -Uusemymalloc
2974
2975Disable Perl's malloc so that Purify can more closely monitor
2976allocations and leaks. Using Perl's malloc will make Purify
2977report most leaks in the "potential" leaks category.
2978
2979=item -Dusemultiplicity
2980
2981Enabling the multiplicity option allows perl to clean up
2982thoroughly when the interpreter shuts down, which reduces the
2983number of bogus leak reports from Purify.
2984
2985=back
2986
2987Once you've compiled a perl suitable for Purify'ing, then you
2988can just:
2989
2990 make pureperl
2991
2992which creates a binary named 'pureperl' that has been Purify'ed.
2993This binary is used in place of the standard 'perl' binary
2994when you want to debug Perl memory problems.
2995
2996As an example, to show any memory leaks produced during the
2997standard Perl testset you would create and run the Purify'ed
2998perl as:
2999
3000 make pureperl
3001 cd t
3002 ../pureperl -I../lib harness
3003
3004which would run Perl on test.pl and report any memory problems.
3005
3006Purify outputs messages in "Viewer" windows by default. If
3007you don't have a windowing environment or if you simply
3008want the Purify output to unobtrusively go to a log file
3009instead of to the interactive window, use these following
3010options to output to the log file "perl.log":
3011
3012 setenv PURIFYOPTIONS "-chain-length=25 -windows=no \
3013 -log-file=perl.log -append-logfile=yes"
3014
3015If you plan to use the "Viewer" windows, then you only need this option:
3016
3017 setenv PURIFYOPTIONS "-chain-length=25"
3018
c406981e
JH
3019In Bourne-type shells:
3020
98631ff8
JL
3021 PURIFYOPTIONS="..."
3022 export PURIFYOPTIONS
c406981e
JH
3023
3024or if you have the "env" utility:
3025
98631ff8 3026 env PURIFYOPTIONS="..." ../pureperl ...
c406981e 3027
902b9dbf
MF
3028=head2 Purify on NT
3029
3030Purify on Windows NT instruments the Perl binary 'perl.exe'
3031on the fly. There are several options in the makefile you
3032should change to get the most use out of Purify:
3033
3034=over 4
3035
3036=item DEFINES
3037
3038You should add -DPURIFY to the DEFINES line so the DEFINES
3039line looks something like:
3040
3041 DEFINES = -DWIN32 -D_CONSOLE -DNO_STRICT $(CRYPT_FLAG) -DPURIFY=1
3042
3043to disable Perl's arena memory allocation functions, as
3044well as to force use of memory allocation functions derived
3045from the system malloc.
3046
3047=item USE_MULTI = define
3048
3049Enabling the multiplicity option allows perl to clean up
3050thoroughly when the interpreter shuts down, which reduces the
3051number of bogus leak reports from Purify.
3052
3053=item #PERL_MALLOC = define
3054
3055Disable Perl's malloc so that Purify can more closely monitor
3056allocations and leaks. Using Perl's malloc will make Purify
3057report most leaks in the "potential" leaks category.
3058
3059=item CFG = Debug
3060
3061Adds debugging information so that you see the exact source
3062statements where the problem occurs. Without this flag, all
3063you will see is the source filename of where the error occurred.
3064
3065=back
3066
3067As an example, to show any memory leaks produced during the
3068standard Perl testset you would create and run Purify as:
3069
3070 cd win32
3071 make
3072 cd ../t
3073 purify ../perl -I../lib harness
3074
3075which would instrument Perl in memory, run Perl on test.pl,
3076then finally report any memory problems.
3077
7a834142
JH
3078=head2 valgrind
3079
3080The excellent valgrind tool can be used to find out both memory leaks
3081and illegal memory accesses. As of August 2003 it unfortunately works
3082only on x86 (ELF) Linux. The special "test.valgrind" target can be used
d44161bf
MHM
3083to run the tests under valgrind. Found errors and memory leaks are
3084logged in files named F<test.valgrind>.
3085
3086As system libraries (most notably glibc) are also triggering errors,
3087valgrind allows to suppress such errors using suppression files. The
3088default suppression file that comes with valgrind already catches a lot
3089of them. Some additional suppressions are defined in F<t/perl.supp>.
7a834142
JH
3090
3091To get valgrind and for more information see
3092
3093 http://developer.kde.org/~sewardj/
3094
f134cc4e 3095=head2 Compaq's/Digital's/HP's Third Degree
09187cb1
JH
3096
3097Third Degree is a tool for memory leak detection and memory access checks.
3098It is one of the many tools in the ATOM toolkit. The toolkit is only
3099available on Tru64 (formerly known as Digital UNIX formerly known as
3100DEC OSF/1).
3101
3102When building Perl, you must first run Configure with -Doptimize=-g
3103and -Uusemymalloc flags, after that you can use the make targets
51a35ef1
JH
3104"perl.third" and "test.third". (What is required is that Perl must be
3105compiled using the C<-g> flag, you may need to re-Configure.)
09187cb1 3106
64cea5fd 3107The short story is that with "atom" you can instrument the Perl
83f0ef60 3108executable to create a new executable called F<perl.third>. When the
4ae3d70a 3109instrumented executable is run, it creates a log of dubious memory
83f0ef60 3110traffic in file called F<perl.3log>. See the manual pages of atom and
4ae3d70a
JH
3111third for more information. The most extensive Third Degree
3112documentation is available in the Compaq "Tru64 UNIX Programmer's
3113Guide", chapter "Debugging Programs with Third Degree".
64cea5fd 3114
9c54ecba 3115The "test.third" leaves a lot of files named F<foo_bar.3log> in the t/
64cea5fd
JH
3116subdirectory. There is a problem with these files: Third Degree is so
3117effective that it finds problems also in the system libraries.
9c54ecba
JH
3118Therefore you should used the Porting/thirdclean script to cleanup
3119the F<*.3log> files.
64cea5fd
JH
3120
3121There are also leaks that for given certain definition of a leak,
3122aren't. See L</PERL_DESTRUCT_LEVEL> for more information.
3123
3124=head2 PERL_DESTRUCT_LEVEL
3125
a958818a
JH
3126If you want to run any of the tests yourself manually using e.g.
3127valgrind, or the pureperl or perl.third executables, please note that
3128by default perl B<does not> explicitly cleanup all the memory it has
3129allocated (such as global memory arenas) but instead lets the exit()
3130of the whole program "take care" of such allocations, also known as
3131"global destruction of objects".
64cea5fd
JH
3132
3133There is a way to tell perl to do complete cleanup: set the
3134environment variable PERL_DESTRUCT_LEVEL to a non-zero value.
3135The t/TEST wrapper does set this to 2, and this is what you
3136need to do too, if you don't want to see the "global leaks":
1f56d61a 3137For example, for "third-degreed" Perl:
64cea5fd 3138
1f56d61a 3139 env PERL_DESTRUCT_LEVEL=2 ./perl.third -Ilib t/foo/bar.t
09187cb1 3140
414f2397
RGS
3141(Note: the mod_perl apache module uses also this environment variable
3142for its own purposes and extended its semantics. Refer to the mod_perl
287a822c
RGS
3143documentation for more information. Also, spawned threads do the
3144equivalent of setting this variable to the value 1.)
5a6c59ef
DM
3145
3146If, at the end of a run you get the message I<N scalars leaked>, you can
fd0854ff
DM
3147recompile with C<-DDEBUG_LEAKING_SCALARS>, which will cause the addresses
3148of all those leaked SVs to be dumped along with details as to where each
3149SV was originally allocated. This information is also displayed by
3150Devel::Peek. Note that the extra details recorded with each SV increases
3151memory usage, so it shouldn't be used in production environments. It also
3152converts C<new_SV()> from a macro into a real function, so you can use
3153your favourite debugger to discover where those pesky SVs were allocated.
414f2397 3154
46c6c7e2
JH
3155=head2 PERL_MEM_LOG
3156
9f653bb5 3157If compiled with C<-DPERL_MEM_LOG>, all Newx() and Renew() allocations
46c6c7e2
JH
3158and Safefree() in the Perl core go through logging functions, which is
3159handy for breakpoint setting. If also compiled with C<-DPERL_MEM_LOG_STDERR>,
e352bcff
JH
3160the allocations and frees are logged to STDERR (or more precisely, to the
3161file descriptor 2) in these logging functions, with the calling source code
3162file and line number (and C function name, if supported by the C compiler).
3163
3164This logging is somewhat similar to C<-Dm> but independent of C<-DDEBUGGING>,
3165and at a higher level (the C<-Dm> is directly at the point of C<malloc()>,
a015a77e 3166while the C<PERL_MEM_LOG> is at the level of C<New()>).
46c6c7e2 3167
51a35ef1
JH
3168=head2 Profiling
3169
3170Depending on your platform there are various of profiling Perl.
3171
3172There are two commonly used techniques of profiling executables:
10f58044 3173I<statistical time-sampling> and I<basic-block counting>.
51a35ef1
JH
3174
3175The first method takes periodically samples of the CPU program
3176counter, and since the program counter can be correlated with the code
3177generated for functions, we get a statistical view of in which
3178functions the program is spending its time. The caveats are that very
3179small/fast functions have lower probability of showing up in the
3180profile, and that periodically interrupting the program (this is
3181usually done rather frequently, in the scale of milliseconds) imposes
3182an additional overhead that may skew the results. The first problem
3183can be alleviated by running the code for longer (in general this is a
3184good idea for profiling), the second problem is usually kept in guard
3185by the profiling tools themselves.
3186
10f58044 3187The second method divides up the generated code into I<basic blocks>.
51a35ef1
JH
3188Basic blocks are sections of code that are entered only in the
3189beginning and exited only at the end. For example, a conditional jump
3190starts a basic block. Basic block profiling usually works by
10f58044 3191I<instrumenting> the code by adding I<enter basic block #nnnn>
51a35ef1
JH
3192book-keeping code to the generated code. During the execution of the
3193code the basic block counters are then updated appropriately. The
3194caveat is that the added extra code can skew the results: again, the
3195profiling tools usually try to factor their own effects out of the
3196results.
3197
83f0ef60
JH
3198=head2 Gprof Profiling
3199
51a35ef1
JH
3200gprof is a profiling tool available in many UNIX platforms,
3201it uses F<statistical time-sampling>.
83f0ef60
JH
3202
3203You can build a profiled version of perl called "perl.gprof" by
51a35ef1
JH
3204invoking the make target "perl.gprof" (What is required is that Perl
3205must be compiled using the C<-pg> flag, you may need to re-Configure).
3206Running the profiled version of Perl will create an output file called
3207F<gmon.out> is created which contains the profiling data collected
3208during the execution.
83f0ef60
JH
3209
3210The gprof tool can then display the collected data in various ways.
3211Usually gprof understands the following options:
3212
3213=over 4
3214
3215=item -a
3216
3217Suppress statically defined functions from the profile.
3218
3219=item -b
3220
3221Suppress the verbose descriptions in the profile.
3222
3223=item -e routine
3224
3225Exclude the given routine and its descendants from the profile.
3226
3227=item -f routine
3228
3229Display only the given routine and its descendants in the profile.
3230
3231=item -s
3232
3233Generate a summary file called F<gmon.sum> which then may be given
3234to subsequent gprof runs to accumulate data over several runs.
3235
3236=item -z
3237
3238Display routines that have zero usage.
3239
3240=back
3241
3242For more detailed explanation of the available commands and output
3243formats, see your own local documentation of gprof.
3244
51a35ef1
JH
3245=head2 GCC gcov Profiling
3246
10f58044 3247Starting from GCC 3.0 I<basic block profiling> is officially available
51a35ef1
JH
3248for the GNU CC.
3249
3250You can build a profiled version of perl called F<perl.gcov> by
3251invoking the make target "perl.gcov" (what is required that Perl must
3252be compiled using gcc with the flags C<-fprofile-arcs
3253-ftest-coverage>, you may need to re-Configure).
3254
3255Running the profiled version of Perl will cause profile output to be
3256generated. For each source file an accompanying ".da" file will be
3257created.
3258
3259To display the results you use the "gcov" utility (which should
3260be installed if you have gcc 3.0 or newer installed). F<gcov> is
3261run on source code files, like this
3262
3263 gcov sv.c
3264
3265which will cause F<sv.c.gcov> to be created. The F<.gcov> files
3266contain the source code annotated with relative frequencies of
3267execution indicated by "#" markers.
3268
3269Useful options of F<gcov> include C<-b> which will summarise the
3270basic block, branch, and function call coverage, and C<-c> which
3271instead of relative frequencies will use the actual counts. For
3272more information on the use of F<gcov> and basic block profiling
3273with gcc, see the latest GNU CC manual, as of GCC 3.0 see
3274
3275 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc.html
3276
3277and its section titled "8. gcov: a Test Coverage Program"
3278
3279 http://gcc.gnu.org/onlinedocs/gcc-3.0/gcc_8.html#SEC132
3280
4ae3d70a
JH
3281=head2 Pixie Profiling
3282
51a35ef1
JH
3283Pixie is a profiling tool available on IRIX and Tru64 (aka Digital
3284UNIX aka DEC OSF/1) platforms. Pixie does its profiling using
10f58044 3285I<basic-block counting>.
4ae3d70a 3286
83f0ef60 3287You can build a profiled version of perl called F<perl.pixie> by
51a35ef1
JH
3288invoking the make target "perl.pixie" (what is required is that Perl
3289must be compiled using the C<-g> flag, you may need to re-Configure).
3290
3291In Tru64 a file called F<perl.Addrs> will also be silently created,
3292this file contains the addresses of the basic blocks. Running the
3293profiled version of Perl will create a new file called "perl.Counts"
3294which contains the counts for the basic block for that particular
3295program execution.
4ae3d70a 3296
51a35ef1 3297To display the results you use the F<prof> utility. The exact
4ae3d70a
JH
3298incantation depends on your operating system, "prof perl.Counts" in
3299IRIX, and "prof -pixie -all -L. perl" in Tru64.
3300
6c41479b
JH
3301In IRIX the following prof options are available:
3302
3303=over 4
3304
3305=item -h
3306
3307Reports the most heavily used lines in descending order of use.
6e36760b 3308Useful for finding the hotspot lines.
6c41479b
JH
3309
3310=item -l
3311
3312Groups lines by procedure, with procedures sorted in descending order of use.
3313Within a procedure, lines are listed in source order.
6e36760b 3314Useful for finding the hotspots of procedures.
6c41479b
JH
3315
3316=back
3317
3318In Tru64 the following options are available:
3319
3320=over 4
3321
3958b146 3322=item -p[rocedures]
6c41479b 3323
3958b146 3324Procedures sorted in descending order by the number of cycles executed
6e36760b 3325in each procedure. Useful for finding the hotspot procedures.
6c41479b
JH
3326(This is the default option.)
3327
24000d2f 3328=item -h[eavy]
6c41479b 3329
6e36760b
JH
3330Lines sorted in descending order by the number of cycles executed in
3331each line. Useful for finding the hotspot lines.
6c41479b 3332
24000d2f 3333=item -i[nvocations]
6c41479b 3334
6e36760b
JH
3335The called procedures are sorted in descending order by number of calls
3336made to the procedures. Useful for finding the most used procedures.
6c41479b 3337
24000d2f 3338=item -l[ines]
6c41479b
JH
3339
3340Grouped by procedure, sorted by cycles executed per procedure.
6e36760b 3341Useful for finding the hotspots of procedures.
6c41479b
JH
3342
3343=item -testcoverage
3344
3345The compiler emitted code for these lines, but the code was unexecuted.
3346
24000d2f 3347=item -z[ero]
6c41479b
JH
3348
3349Unexecuted procedures.
3350
aa500c9e 3351=back
6c41479b
JH
3352
3353For further information, see your system's manual pages for pixie and prof.
4ae3d70a 3354
b8ddf6b3
SB
3355=head2 Miscellaneous tricks
3356
3357=over 4
3358
3359=item *
3360
cc177e1a 3361Those debugging perl with the DDD frontend over gdb may find the
b8ddf6b3
SB
3362following useful:
3363
3364You can extend the data conversion shortcuts menu, so for example you
3365can display an SV's IV value with one click, without doing any typing.
3366To do that simply edit ~/.ddd/init file and add after:
3367
3368 ! Display shortcuts.
3369 Ddd*gdbDisplayShortcuts: \
3370 /t () // Convert to Bin\n\
3371 /d () // Convert to Dec\n\
3372 /x () // Convert to Hex\n\
3373 /o () // Convert to Oct(\n\
3374
3375the following two lines:
3376
3377 ((XPV*) (())->sv_any )->xpv_pv // 2pvx\n\
3378 ((XPVIV*) (())->sv_any )->xiv_iv // 2ivx
3379
3380so now you can do ivx and pvx lookups or you can plug there the
3381sv_peek "conversion":
3382
3383 Perl_sv_peek(my_perl, (SV*)()) // sv_peek
3384
3385(The my_perl is for threaded builds.)
3386Just remember that every line, but the last one, should end with \n\
3387
3388Alternatively edit the init file interactively via:
33893rd mouse button -> New Display -> Edit Menu
3390
3391Note: you can define up to 20 conversion shortcuts in the gdb
3392section.
3393
9965345d
JH
3394=item *
3395
7e337ee0
JH
3396If you see in a debugger a memory area mysteriously full of 0xABABABAB
3397or 0xEFEFEFEF, you may be seeing the effect of the Poison() macros,
3398see L<perlclib>.
9965345d 3399
f1fac472
NC
3400=item *
3401
3402Under ithreads the optree is read only. If you want to enforce this, to check
3403for write accesses from buggy code, compile with C<-DPL_OP_SLAB_ALLOC> to
3404enable the OP slab allocator and C<-DPERL_DEBUG_READONLY_OPS> to enable code
3405that allocates op memory via C<mmap>, and sets it read-only at run time.
3406Any write access to an op results in a C<SIGBUS> and abort.
3407
3408This code is intended for development only, and may not be portable even to
3409all Unix variants. Also, it is an 80% solution, in that it isn't able to make
3410all ops read only. Specifically it
3411
3412=over
3413
3414=item 1
3415
3416Only sets read-only on all slabs of ops at C<CHECK> time, hence ops allocated
3417later via C<require> or C<eval> will be re-write
3418
3419=item 2
3420
3421Turns an entire slab of ops read-write if the refcount of any op in the slab
3422needs to be decreased.
3423
3424=item 3
3425
3426Turns an entire slab of ops read-write if any op from the slab is freed.
3427
b8ddf6b3
SB
3428=back
3429
f1fac472
NC
3430It's not possible to turn the slabs to read-only after an action requiring
3431read-write access, as either can happen during op tree building time, so
3432there may still be legitimate write access.
3433
3434However, as an 80% solution it is still effective, as currently it catches
3435a write access during the generation of F<Config.pm>, which means that we
3436can't yet build F<perl> with this enabled.
3437
3438=back
3439
3440
955fec6b 3441=head1 CONCLUSION
a422fd2d 3442
955fec6b
JH
3443We've had a brief look around the Perl source, how to maintain quality
3444of the source code, an overview of the stages F<perl> goes through
3445when it's running your code, how to use debuggers to poke at the Perl
3446guts, and finally how to analyse the execution of Perl. We took a very
3447simple problem and demonstrated how to solve it fully - with
3448documentation, regression tests, and finally a patch for submission to
3449p5p. Finally, we talked about how to use external tools to debug and
3450test Perl.
a422fd2d
SC
3451
3452I'd now suggest you read over those references again, and then, as soon
3453as possible, get your hands dirty. The best way to learn is by doing,
3454so:
3455
3456=over 3
3457
3458=item *
3459
3460Subscribe to perl5-porters, follow the patches and try and understand
3461them; don't be afraid to ask if there's a portion you're not clear on -
3462who knows, you may unearth a bug in the patch...
3463
3464=item *
3465
3466Keep up to date with the bleeding edge Perl distributions and get
3467familiar with the changes. Try and get an idea of what areas people are
3468working on and the changes they're making.
3469
3470=item *
3471
3e148164 3472Do read the README associated with your operating system, e.g. README.aix
a1f349fd
MB
3473on the IBM AIX OS. Don't hesitate to supply patches to that README if
3474you find anything missing or changed over a new OS release.
3475
3476=item *
3477
a422fd2d
SC
3478Find an area of Perl that seems interesting to you, and see if you can
3479work out how it works. Scan through the source, and step over it in the
3480debugger. Play, poke, investigate, fiddle! You'll probably get to
3481understand not just your chosen area but a much wider range of F<perl>'s
3482activity as well, and probably sooner than you'd think.
3483
3484=back
3485
3486=over 3
3487
3488=item I<The Road goes ever on and on, down from the door where it began.>
3489
3490=back
3491
64d9b66b 3492If you can do these things, you've started on the long road to Perl porting.
a422fd2d
SC
3493Thanks for wanting to help make Perl better - and happy hacking!
3494
e8cd7eae
GS
3495=head1 AUTHOR
3496
3497This document was written by Nathan Torkington, and is maintained by
3498the perl5-porters mailing list.