1PERLHACK(1) Perl Programmers Reference Guide PERLHACK(1)
2
6 perlhack - How to hack on Perl
7
9 This document explains how Perl development works. It includes details
10 about the Perl 5 Porters email list, the Perl repository, the Perlbug
11 bug tracker, patch guidelines, and commentary on Perl development
12 philosophy.
13
15 If you just want to submit a single small patch like a pod fix, a test
16 for a bug, comment fixes, etc., it's easy! Here's how:
17 · Check out the source repository
19 The perl source is in a git repository. You can clone the
21 repository with the following command:
22 % git clone git://perl5.git.perl.org/perl.git perl
24 · Ensure you're following the latest advice
26 In case the advice in this guide has been updated recently, read
28 the latest version directly from the perl source:
29 % perldoc pod/perlhack.pod
31 · Make your change
33 Hack, hack, hack. Keep in mind that Perl runs on many different
35 platforms, with different operating systems that have different
36 capabilities, different filesystem organizations, and even
37 different character sets. perlhacktips gives advice on this.
38 · Test your change
40 You can run all the tests with the following commands:
42 % ./Configure -des -Dusedevel
44 % make test
45 Keep hacking until the tests pass.
47 · Commit your change
49 Committing your work will save the change on your local system:
51 % git commit -a -m 'Commit message goes here'
53 Make sure the commit message describes your change in a single
55 sentence. For example, "Fixed spelling errors in perlhack.pod".
56 · Send your change to perlbug
58 The next step is to submit your patch to the Perl core ticket
60 system via email.
61 If your changes are in a single git commit, run the following
63 commands to generate the patch file and attach it to your bug
64 report:
65 % git format-patch -1
67 % ./perl -Ilib utils/perlbug -p 0001-*.patch
68 The perlbug program will ask you a few questions about your email
70 address and the patch you're submitting. Once you've answered them
71 it will submit your patch via email.
72 If your changes are in multiple commits, generate a patch file for
74 each one and provide them to perlbug's "-p" option separated by
75 commas:
76 % git format-patch -3
78 % ./perl -Ilib utils/perlbug -p 0001-fix1.patch,0002-fix2.patch,\
79 > 0003-fix3.patch
80 When prompted, pick a subject that summarizes your changes.
82 · Thank you
84 The porters appreciate the time you spent helping to make Perl
86 better. Thank you!
87 · Acknowledgement
89 All contributors are credited (by name and email address) in the
91 AUTHORS file, which is part of the perl distribution, as well as
92 the Git commit history.
93 If you donXt want to be included in the AUTHORS file, just let us
95 know. Otherwise we will take your submission of a patch as
96 permission to credit you in the AUTHORS file.
97 · Next time
99 The next time you wish to make a patch, you need to start from the
101 latest perl in a pristine state. Check you don't have any local
102 changes or added files in your perl check-out which you wish to
103 keep, then run these commands:
104 % git pull
106 % git reset --hard origin/blead
107 % git clean -dxf
108
110 If you want to report a bug in Perl, you must use the perlbug command
111 line tool. This tool will ensure that your bug report includes all the
112 relevant system and configuration information.
113 To browse existing Perl bugs and patches, you can use the web interface
115 at <http://rt.perl.org/>.
116 Please check the archive of the perl5-porters list (see below) and/or
118 the bug tracking system before submitting a bug report. Often, you'll
119 find that the bug has been reported already.
120 You can log in to the bug tracking system and comment on existing bug
122 reports. If you have additional information regarding an existing bug,
123 please add it. This will help the porters fix the bug.
124
126 The perl5-porters (p5p) mailing list is where the Perl standard
127 distribution is maintained and developed. The people who maintain Perl
128 are also referred to as the "Perl 5 Porters", "p5p" or just the
129 "porters".
130 A searchable archive of the list is available at
132 <http://markmail.org/search/?q=perl5-porters>. There is also an
133 archive at <http://archive.develooper.com/perl5-porters@perl.org/>.
134 perl-changes mailing list
136 The perl5-changes mailing list receives a copy of each patch that gets
137 submitted to the maintenance and development branches of the perl
138 repository. See <http://lists.perl.org/list/perl5-changes.html> for
139 subscription and archive information.
140 #p5p on IRC
142 Many porters are also active on the <irc://irc.perl.org/#p5p> channel.
143 Feel free to join the channel and ask questions about hacking on the
144 Perl core.
145
147 All of Perl's source code is kept centrally in a Git repository at
148 perl5.git.perl.org. The repository contains many Perl revisions from
149 Perl 1 onwards and all the revisions from Perforce, the previous
150 version control system.
151 For much more detail on using git with the Perl repository, please see
153 perlgit.
154 Read access via Git
156 You will need a copy of Git for your computer. You can fetch a copy of
157 the repository using the git protocol:
158 % git clone git://perl5.git.perl.org/perl.git perl
160 This clones the repository and makes a local copy in the perl
162 directory.
163 If you cannot use the git protocol for firewall reasons, you can also
165 clone via http, though this is much slower:
166 % git clone http://perl5.git.perl.org/perl.git perl
168 Read access via the web
170 You may access the repository over the web. This allows you to browse
171 the tree, see recent commits, subscribe to RSS feeds for the changes,
172 search for particular commits and more. You may access it at
173 <http://perl5.git.perl.org/perl.git>. A mirror of the repository is
174 found at <https://github.com/Perl/perl5>.
175 Read access via rsync
177 You can also choose to use rsync to get a copy of the current source
178 tree for the bleadperl branch and all maintenance branches:
179 % rsync -avz rsync://perl5.git.perl.org/perl-current .
181 % rsync -avz rsync://perl5.git.perl.org/perl-5.12.x .
182 % rsync -avz rsync://perl5.git.perl.org/perl-5.10.x .
183 % rsync -avz rsync://perl5.git.perl.org/perl-5.8.x .
184 % rsync -avz rsync://perl5.git.perl.org/perl-5.6.x .
185 % rsync -avz rsync://perl5.git.perl.org/perl-5.005xx .
186 (Add the "--delete" option to remove leftover files.)
188 To get a full list of the available sync points:
190 % rsync perl5.git.perl.org::
192 Write access via git
194 If you have a commit bit, please see perlgit for more details on using
195 git.
196
198 If you're planning to do more extensive work than a single small fix,
199 we encourage you to read the documentation below. This will help you
200 focus your work and make your patches easier to incorporate into the
201 Perl source.
202 Submitting patches
204 If you have a small patch to submit, please submit it via perlbug. You
205 can also send email directly to perlbug@perl.org. Please note that
206 messages sent to perlbug may be held in a moderation queue, so you
207 won't receive a response immediately.
208 You'll know your submission has been processed when you receive an
210 email from our ticket tracking system. This email will give you a
211 ticket number. Once your patch has made it to the ticket tracking
212 system, it will also be sent to the perl5-porters@perl.org list.
213 If your patch is related to an already-opened ticket you can also
215 attach your patch to that ticket, without having to use perlbug.
216 Patches are reviewed and discussed on the p5p list. Simple,
218 uncontroversial patches will usually be applied without any discussion.
219 When the patch is applied, the ticket will be updated and you will
220 receive email. In addition, an email will be sent to the p5p list.
221 In other cases, the patch will need more work or discussion. That will
223 happen on the p5p list.
224 You are encouraged to participate in the discussion and advocate for
226 your patch. Sometimes your patch may get lost in the shuffle. It's
227 appropriate to send a reminder email to p5p if no action has been taken
228 in a month. Please remember that the Perl 5 developers are all
229 volunteers, and be polite.
230 Changes are always applied directly to the main development branch,
232 called "blead". Some patches may be backported to a maintenance
233 branch. If you think your patch is appropriate for the maintenance
234 branch (see "MAINTENANCE BRANCHES" in perlpolicy), please explain why
235 when you submit it.
236 Getting your patch accepted
238 If you are submitting a code patch there are several things that you
239 can do to help the Perl 5 Porters accept your patch.
240 Patch style
242 If you used git to check out the Perl source, then using "git
244 format-patch" will produce a patch in a style suitable for Perl. The
245 "format-patch" command produces one patch file for each commit you
246 made. If you prefer to send a single patch for all commits, you can
247 use "git diff".
248 % git checkout blead
250 % git pull
251 % git diff blead my-branch-name
252 This produces a patch based on the difference between blead and your
254 current branch. It's important to make sure that blead is up to date
255 before producing the diff, that's why we call "git pull" first.
256 We strongly recommend that you use git if possible. It will make your
258 life easier, and ours as well.
259 However, if you're not using git, you can still produce a suitable
261 patch. You'll need a pristine copy of the Perl source to diff against.
262 The porters prefer unified diffs. Using GNU "diff", you can produce a
263 diff like this:
264 % diff -Npurd perl.pristine perl.mine
266 Make sure that you "make realclean" in your copy of Perl to remove any
268 build artifacts, or you may get a confusing result.
269 Commit message
271 As you craft each patch you intend to submit to the Perl core, it's
273 important to write a good commit message. This is especially important
274 if your submission will consist of a series of commits.
275 The first line of the commit message should be a short description
277 without a period. It should be no longer than the subject line of an
278 email, 50 characters being a good rule of thumb.
279 A lot of Git tools (Gitweb, GitHub, git log --pretty=oneline, ...) will
281 only display the first line (cut off at 50 characters) when presenting
282 commit summaries.
283 The commit message should include a description of the problem that the
285 patch corrects or new functionality that the patch adds.
286 As a general rule of thumb, your commit message should help a
288 programmer who knows the Perl core quickly understand what you were
289 trying to do, how you were trying to do it, and why the change matters
290 to Perl.
291 · Why
293 Your commit message should describe why the change you are making
295 is important. When someone looks at your change in six months or
296 six years, your intent should be clear.
297 If you're deprecating a feature with the intent of later
299 simplifying another bit of code, say so. If you're fixing a
300 performance problem or adding a new feature to support some other
301 bit of the core, mention that.
302 · What
304 Your commit message should describe what part of the Perl core
306 you're changing and what you expect your patch to do.
307 · How
309 While it's not necessary for documentation changes, new tests or
311 trivial patches, it's often worth explaining how your change works.
312 Even if it's clear to you today, it may not be clear to a porter
313 next month or next year.
314 A commit message isn't intended to take the place of comments in your
316 code. Commit messages should describe the change you made, while code
317 comments should describe the current state of the code.
318 If you've just implemented a new feature, complete with doc, tests and
320 well-commented code, a brief commit message will often suffice. If,
321 however, you've just changed a single character deep in the parser or
322 lexer, you might need to write a small novel to ensure that future
323 readers understand what you did and why you did it.
324 Comments, Comments, Comments
326 Be sure to adequately comment your code. While commenting every line
328 is unnecessary, anything that takes advantage of side effects of
329 operators, that creates changes that will be felt outside of the
330 function being patched, or that others may find confusing should be
331 documented. If you are going to err, it is better to err on the side
332 of adding too many comments than too few.
333 The best comments explain why the code does what it does, not what it
335 does.
336 Style
338 In general, please follow the particular style of the code you are
340 patching.
341 In particular, follow these general guidelines for patching Perl
343 sources:
344 · 4-wide indents for code, 2-wide indents for nested CPP "#define"s,
346 with 8-wide tabstops.
347 · Use spaces for indentation, not tab characters.
349 The codebase is a mixture of tabs and spaces for indentation, and
351 we are moving to spaces only. Converting lines you're patching
352 from 8-wide tabs to spaces will help this migration.
353 · Try hard not to exceed 79-columns
355 · ANSI C prototypes
357 · Uncuddled elses and "K&R" style for indenting control constructs
359 · No C++ style (//) comments
361 · Mark places that need to be revisited with XXX (and revisit often!)
363 · Opening brace lines up with "if" when conditional spans multiple
365 lines; should be at end-of-line otherwise
366 · In function definitions, name starts in column 0 (return value-type
368 is on previous line)
369 · Single space after keywords that are followed by parens, no space
371 between function name and following paren
372 · Avoid assignments in conditionals, but if they're unavoidable, use
374 extra paren, e.g. "if (a && (b = c)) ..."
375 · "return foo;" rather than "return(foo);"
377 · "if (!foo) ..." rather than "if (foo == FALSE) ..." etc.
379 · Do not declare variables using "register". It may be
381 counterproductive with modern compilers, and is deprecated in C++,
382 under which the Perl source is regularly compiled.
383 · In-line functions that are in headers that are accessible to XS
385 code need to be able to compile without warnings with commonly used
386 extra compilation flags, such as gcc's "-Wswitch-default" which
387 warns whenever a switch statement does not have a "default" case.
388 The use of these extra flags is to catch potential problems in
389 legal C code, and is often used by Perl aggregators, such as Linux
390 distributors.
391 Test suite
393 If your patch changes code (rather than just changing documentation),
395 you should also include one or more test cases which illustrate the bug
396 you're fixing or validate the new functionality you're adding. In
397 general, you should update an existing test file rather than create a
398 new one.
399 Your test suite additions should generally follow these guidelines
401 (courtesy of Gurusamy Sarathy <gsar@activestate.com>):
402 · Know what you're testing. Read the docs, and the source.
404 · Tend to fail, not succeed.
406 · Interpret results strictly.
408 · Use unrelated features (this will flush out bizarre interactions).
410 · Use non-standard idioms (otherwise you are not testing TIMTOWTDI).
412 · Avoid using hardcoded test numbers whenever possible (the
414 EXPECTED/GOT found in t/op/tie.t is much more maintainable, and
415 gives better failure reports).
416 · Give meaningful error messages when a test fails.
418 · Avoid using qx// and system() unless you are testing for them. If
420 you do use them, make sure that you cover _all_ perl platforms.
421 · Unlink any temporary files you create.
423 · Promote unforeseen warnings to errors with $SIG{__WARN__}.
425 · Be sure to use the libraries and modules shipped with the version
427 being tested, not those that were already installed.
428 · Add comments to the code explaining what you are testing for.
430 · Make updating the '1..42' string unnecessary. Or make sure that
432 you update it.
433 · Test _all_ behaviors of a given operator, library, or function.
435 Test all optional arguments.
437 Test return values in various contexts (boolean, scalar, list,
439 lvalue).
440 Use both global and lexical variables.
442 Don't forget the exceptional, pathological cases.
444 Patching a core module
446 This works just like patching anything else, with one extra
447 consideration.
448 Modules in the cpan/ directory of the source tree are maintained
450 outside of the Perl core. When the author updates the module, the
451 updates are simply copied into the core. See that module's
452 documentation or its listing on <http://search.cpan.org/> for more
453 information on reporting bugs and submitting patches.
454 In most cases, patches to modules in cpan/ should be sent upstream and
456 should not be applied to the Perl core individually. If a patch to a
457 file in cpan/ absolutely cannot wait for the fix to be made upstream,
458 released to CPAN and copied to blead, you must add (or update) a
459 "CUSTOMIZED" entry in the "Porting/Maintainers.pl" file to flag that a
460 local modification has been made. See "Porting/Maintainers.pl" for
461 more details.
462 In contrast, modules in the dist/ directory are maintained in the core.
464 Updating perldelta
466 For changes significant enough to warrant a pod/perldelta.pod entry,
467 the porters will greatly appreciate it if you submit a delta entry
468 along with your actual change. Significant changes include, but are
469 not limited to:
470 · Adding, deprecating, or removing core features
472 · Adding, deprecating, removing, or upgrading core or dual-life
474 modules
475 · Adding new core tests
477 · Fixing security issues and user-visible bugs in the core
479 · Changes that might break existing code, either on the perl or C
481 level
482 · Significant performance improvements
484 · Adding, removing, or significantly changing documentation in the
486 pod/ directory
487 · Important platform-specific changes
489 Please make sure you add the perldelta entry to the right section
491 within pod/perldelta.pod. More information on how to write good
492 perldelta entries is available in the "Style" section of
493 Porting/how_to_write_a_perldelta.pod.
494 What makes for a good patch?
496 New features and extensions to the language can be contentious. There
497 is no specific set of criteria which determine what features get added,
498 but here are some questions to consider when developing a patch:
499 Does the concept match the general goals of Perl?
501 Our goals include, but are not limited to:
503 1. Keep it fast, simple, and useful.
505 2. Keep features/concepts as orthogonal as possible.
507 3. No arbitrary limits (platforms, data sizes, cultures).
509 4. Keep it open and exciting to use/patch/advocate Perl everywhere.
511 5. Either assimilate new technologies, or build bridges to them.
513 Where is the implementation?
515 All the talk in the world is useless without an implementation. In
517 almost every case, the person or people who argue for a new feature
518 will be expected to be the ones who implement it. Porters capable of
519 coding new features have their own agendas, and are not available to
520 implement your (possibly good) idea.
521 Backwards compatibility
523 It's a cardinal sin to break existing Perl programs. New warnings can
525 be contentious--some say that a program that emits warnings is not
526 broken, while others say it is. Adding keywords has the potential to
527 break programs, changing the meaning of existing token sequences or
528 functions might break programs.
529 The Perl 5 core includes mechanisms to help porters make backwards
531 incompatible changes more compatible such as the feature and deprecate
532 modules. Please use them when appropriate.
533 Could it be a module instead?
535 Perl 5 has extension mechanisms, modules and XS, specifically to avoid
537 the need to keep changing the Perl interpreter. You can write modules
538 that export functions, you can give those functions prototypes so they
539 can be called like built-in functions, you can even write XS code to
540 mess with the runtime data structures of the Perl interpreter if you
541 want to implement really complicated things.
542 Whenever possible, new features should be prototyped in a CPAN module
544 before they will be considered for the core.
545 Is the feature generic enough?
547 Is this something that only the submitter wants added to the language,
549 or is it broadly useful? Sometimes, instead of adding a feature with a
550 tight focus, the porters might decide to wait until someone implements
551 the more generalized feature.
552 Does it potentially introduce new bugs?
554 Radical rewrites of large chunks of the Perl interpreter have the
556 potential to introduce new bugs.
557 How big is it?
559 The smaller and more localized the change, the better. Similarly, a
561 series of small patches is greatly preferred over a single large patch.
562 Does it preclude other desirable features?
564 A patch is likely to be rejected if it closes off future avenues of
566 development. For instance, a patch that placed a true and final
567 interpretation on prototypes is likely to be rejected because there are
568 still options for the future of prototypes that haven't been addressed.
569 Is the implementation robust?
571 Good patches (tight code, complete, correct) stand more chance of going
573 in. Sloppy or incorrect patches might be placed on the back burner
574 until the pumpking has time to fix, or might be discarded altogether
575 without further notice.
576 Is the implementation generic enough to be portable?
578 The worst patches make use of system-specific features. It's highly
580 unlikely that non-portable additions to the Perl language will be
581 accepted.
582 Is the implementation tested?
584 Patches which change behaviour (fixing bugs or introducing new
586 features) must include regression tests to verify that everything works
587 as expected.
588 Without tests provided by the original author, how can anyone else
590 changing perl in the future be sure that they haven't unwittingly
591 broken the behaviour the patch implements? And without tests, how can
592 the patch's author be confident that his/her hard work put into the
593 patch won't be accidentally thrown away by someone in the future?
594 Is there enough documentation?
596 Patches without documentation are probably ill-thought out or
598 incomplete. No features can be added or changed without documentation,
599 so submitting a patch for the appropriate pod docs as well as the
600 source code is important.
601 Is there another way to do it?
603 Larry said "Although the Perl Slogan is There's More Than One Way to Do
605 It, I hesitate to make 10 ways to do something". This is a tricky
606 heuristic to navigate, though--one man's essential addition is another
607 man's pointless cruft.
608 Does it create too much work?
610 Work for the pumpking, work for Perl programmers, work for module
612 authors, ... Perl is supposed to be easy.
613 Patches speak louder than words
615 Working code is always preferred to pie-in-the-sky ideas. A patch to
617 add a feature stands a much higher chance of making it to the language
618 than does a random feature request, no matter how fervently argued the
619 request might be. This ties into "Will it be useful?", as the fact
620 that someone took the time to make the patch demonstrates a strong
621 desire for the feature.
622
624 The core uses the same testing style as the rest of Perl, a simple
625 "ok/not ok" run through Test::Harness, but there are a few special
626 considerations.
627 There are three ways to write a test in the core: Test::More, t/test.pl
629 and ad hoc "print $test ? "ok 42\n" : "not ok 42\n"". The decision of
630 which to use depends on what part of the test suite you're working on.
631 This is a measure to prevent a high-level failure (such as Config.pm
632 breaking) from causing basic functionality tests to fail.
633 The t/test.pl library provides some of the features of Test::More, but
635 avoids loading most modules and uses as few core features as possible.
636 If you write your own test, use the Test Anything Protocol
638 <http://testanything.org>.
639 · t/base, t/comp and t/opbasic
641 Since we don't know if "require" works, or even subroutines, use ad
643 hoc tests for these three. Step carefully to avoid using the
644 feature being tested. Tests in t/opbasic, for instance, have been
645 placed there rather than in t/op because they test functionality
646 which t/test.pl presumes has already been demonstrated to work.
647 · t/cmd, t/run, t/io and t/op
649 Now that basic require() and subroutines are tested, you can use
651 the t/test.pl library.
652 You can also use certain libraries like Config conditionally, but
654 be sure to skip the test gracefully if it's not there.
655 · Everything else
657 Now that the core of Perl is tested, Test::More can and should be
659 used. You can also use the full suite of core modules in the
660 tests.
661 When you say "make test", Perl uses the t/TEST program to run the test
663 suite (except under Win32 where it uses t/harness instead). All tests
664 are run from the t/ directory, not the directory which contains the
665 test. This causes some problems with the tests in lib/, so here's some
666 opportunity for some patching.
667 You must be triply conscious of cross-platform concerns. This usually
669 boils down to using File::Spec, avoiding things like "fork()" and
670 "system()" unless absolutely necessary, and not assuming that a given
671 character has a particular ordinal value (code point) or that its UTF-8
672 representation is composed of particular bytes.
673 There are several functions available to specify characters and code
675 points portably in tests. The always-preloaded functions
676 "utf8::unicode_to_native()" and its inverse "utf8::native_to_unicode()"
677 take code points and translate appropriately. The file
678 t/charset_tools.pl has several functions that can be useful. It has
679 versions of the previous two functions that take strings as inputs --
680 not single numeric code points: "uni_to_native()" and
681 "native_to_uni()". If you must look at the individual bytes comprising
682 a UTF-8 encoded string, "byte_utf8a_to_utf8n()" takes as input a string
683 of those bytes encoded for an ASCII platform, and returns the
684 equivalent string in the native platform. For example,
685 "byte_utf8a_to_utf8n("\xC2\xA0")" returns the byte sequence on the
686 current platform that form the UTF-8 for "U+00A0", since "\xC2\xA0" are
687 the UTF-8 bytes on an ASCII platform for that code point. This
688 function returns "\xC2\xA0" on an ASCII platform, and "\x80\x41" on an
689 EBCDIC 1047 one.
690 But easiest is, if the character is specifiable as a literal, like "A"
692 or "%", to use that; if not so specificable, you can use use "\N{}" ,
693 if the side effects aren't troublesome. Simply specify all your
694 characters in hex, using "\N{U+ZZ}" instead of "\xZZ". "\N{}" is the
695 Unicode name, and so it always gives you the Unicode character.
696 "\N{U+41}" is the character whose Unicode code point is 0x41, hence is
697 'A' on all platforms. The side effects are:
698 · These select Unicode rules. That means that in double-quotish
700 strings, the string is always converted to UTF-8 to force a Unicode
701 interpretation (you can "utf8::downgrade()" afterwards to convert
702 back to non-UTF8, if possible). In regular expression patterns,
703 the conversion isn't done, but if the character set modifier would
704 otherwise be "/d", it is changed to "/u".
705 · If you use the form "\N{character name}", the charnames module gets
707 automatically loaded. This may not be suitable for the test level
708 you are doing.
709 If you are testing locales (see perllocale), there are helper functions
711 in t/loc_tools.pl to enable you to see what locales there are on the
712 current platform.
713 Special "make test" targets
715 There are various special make targets that can be used to test Perl
716 slightly differently than the standard "test" target. Not all them are
717 expected to give a 100% success rate. Many of them have several
718 aliases, and many of them are not available on certain operating
719 systems.
720 · test_porting
722 This runs some basic sanity tests on the source tree and helps
724 catch basic errors before you submit a patch.
725 · minitest
727 Run miniperl on t/base, t/comp, t/cmd, t/run, t/io, t/op, t/uni and
729 t/mro tests.
730 · test.valgrind check.valgrind
732 (Only in Linux) Run all the tests using the memory leak + naughty
734 memory access tool "valgrind". The log files will be named
735 testname.valgrind.
736 · test_harness
738 Run the test suite with the t/harness controlling program, instead
740 of t/TEST. t/harness is more sophisticated, and uses the
741 Test::Harness module, thus using this test target supposes that
742 perl mostly works. The main advantage for our purposes is that it
743 prints a detailed summary of failed tests at the end. Also, unlike
744 t/TEST, it doesn't redirect stderr to stdout.
745 Note that under Win32 t/harness is always used instead of t/TEST,
747 so there is no special "test_harness" target.
748 Under Win32's "test" target you may use the TEST_SWITCHES and
750 TEST_FILES environment variables to control the behaviour of
751 t/harness. This means you can say
752 nmake test TEST_FILES="op/*.t"
754 nmake test TEST_SWITCHES="-torture" TEST_FILES="op/*.t"
755 · test-notty test_notty
757 Sets PERL_SKIP_TTY_TEST to true before running normal test.
759 Parallel tests
761 The core distribution can now run its regression tests in parallel on
762 Unix-like platforms. Instead of running "make test", set "TEST_JOBS"
763 in your environment to the number of tests to run in parallel, and run
764 "make test_harness". On a Bourne-like shell, this can be done as
765 TEST_JOBS=3 make test_harness # Run 3 tests in parallel
767 An environment variable is used, rather than parallel make itself,
769 because TAP::Harness needs to be able to schedule individual non-
770 conflicting test scripts itself, and there is no standard interface to
771 "make" utilities to interact with their job schedulers.
772 Note that currently some test scripts may fail when run in parallel
774 (most notably dist/IO/t/io_dir.t). If necessary, run just the failing
775 scripts again sequentially and see if the failures go away.
776 Running tests by hand
778 You can run part of the test suite by hand by using one of the
779 following commands from the t/ directory:
780 ./perl -I../lib TEST list-of-.t-files
782 or
784 ./perl -I../lib harness list-of-.t-files
786 (If you don't specify test scripts, the whole test suite will be run.)
788 Using t/harness for testing
790 If you use "harness" for testing, you have several command line options
791 available to you. The arguments are as follows, and are in the order
792 that they must appear if used together.
793 harness -v -torture -re=pattern LIST OF FILES TO TEST
795 harness -v -torture -re LIST OF PATTERNS TO MATCH
796 If "LIST OF FILES TO TEST" is omitted, the file list is obtained from
798 the manifest. The file list may include shell wildcards which will be
799 expanded out.
800 · -v
802 Run the tests under verbose mode so you can see what tests were
804 run, and debug output.
805 · -torture
807 Run the torture tests as well as the normal set.
809 · -re=PATTERN
811 Filter the file list so that all the test files run match PATTERN.
813 Note that this form is distinct from the -re LIST OF PATTERNS form
814 below in that it allows the file list to be provided as well.
815 · -re LIST OF PATTERNS
817 Filter the file list so that all the test files run match
819 /(LIST|OF|PATTERNS)/. Note that with this form the patterns are
820 joined by '|' and you cannot supply a list of files, instead the
821 test files are obtained from the MANIFEST.
822 You can run an individual test by a command similar to
824 ./perl -I../lib path/to/foo.t
826 except that the harnesses set up some environment variables that may
828 affect the execution of the test:
829 · PERL_CORE=1
831 indicates that we're running this test as part of the perl core
833 test suite. This is useful for modules that have a dual life on
834 CPAN.
835 · PERL_DESTRUCT_LEVEL=2
837 is set to 2 if it isn't set already (see "PERL_DESTRUCT_LEVEL" in
839 perlhacktips).
840 · PERL
842 (used only by t/TEST) if set, overrides the path to the perl
844 executable that should be used to run the tests (the default being
845 ./perl).
846 · PERL_SKIP_TTY_TEST
848 if set, tells to skip the tests that need a terminal. It's
850 actually set automatically by the Makefile, but can also be forced
851 artificially by running 'make test_notty'.
852 Other environment variables that may influence tests
854 · PERL_TEST_Net_Ping
856 Setting this variable runs all the Net::Ping modules tests,
858 otherwise some tests that interact with the outside world are
859 skipped. See perl58delta.
860 · PERL_TEST_NOVREXX
862 Setting this variable skips the vrexx.t tests for OS2::REXX.
864 · PERL_TEST_NUMCONVERTS
866 This sets a variable in op/numconvert.t.
868 · PERL_TEST_MEMORY
870 Setting this variable includes the tests in t/bigmem/. This should
872 be set to the number of gigabytes of memory available for testing,
873 eg. "PERL_TEST_MEMORY=4" indicates that tests that require 4GiB of
874 available memory can be run safely.
875 See also the documentation for the Test and Test::Harness modules, for
877 more environment variables that affect testing.
878 Performance testing
880 The file t/perf/benchmarks contains snippets of perl code which are
881 intended to be benchmarked across a range of perls by the
882 Porting/bench.pl tool. If you fix or enhance a performance issue, you
883 may want to add a representative code sample to the file, then run
884 bench.pl against the previous and current perls to see what difference
885 it has made, and whether anything else has slowed down as a
886 consequence.
887 The file t/perf/opcount.t is designed to test whether a particular code
889 snippet has been compiled into an optree containing specified numbers
890 of particular op types. This is good for testing whether optimisations
891 which alter ops, such as converting an "aelem" op into an "aelemfast"
892 op, are really doing that.
893 The files t/perf/speed.t and t/re/speed.t are designed to test things
895 that run thousands of times slower if a particular optimisation is
896 broken (for example, the utf8 length cache on long utf8 strings). Add
897 a test that will take a fraction of a second normally, and minutes
898 otherwise, causing the test file to time out on failure.
899 Building perl at older commits
901 In the course of hacking on the Perl core distribution, you may have
902 occasion to configure, build and test perl at an old commit. Sometimes
903 "make" will fail during this process. If that happens, you may be able
904 to salvage the situation by using the Devel::PatchPerl library from
905 CPAN (not included in the core) to bring the source code at that commit
906 to a buildable state.
907 Here's a real world example, taken from work done to resolve perl
909 #72414 <https://rt.perl.org/Ticket/Display.html?id=72414>. Use of
910 Porting/bisect.pl had identified commit
911 "ba77e4cc9d1ceebf472c9c5c18b2377ee47062e6" as the commit in which a bug
912 was corrected. To confirm, a P5P developer wanted to configure and
913 build perl at commit "ba77e4c^" (presumably "bad") and then at
914 "ba77e4c" (presumably "good"). Normal configuration and build was
915 attempted:
916 $ sh ./Configure -des -Dusedevel
918 $ make test_prep
919 "make", however, failed with output (excerpted) like this:
921 cc -fstack-protector -L/usr/local/lib -o miniperl \
923 gv.o toke.o perly.o pad.o regcomp.o dump.o util.o \
924 mg.o reentr.o mro.o hv.o av.o run.o pp_hot.o sv.o \
925 pp.o scope.o pp_ctl.o pp_sys.o doop.o doio.o regexec.o \
926 utf8.o taint.o deb.o universal.o globals.o perlio.o \
927 perlapi.o numeric.o mathoms.o locale.o pp_pack.o pp_sort.o \
928 miniperlmain.o opmini.o perlmini.o
929 pp.o: In function `Perl_pp_pow':
930 pp.c:(.text+0x2db9): undefined reference to `pow'
931 ...
932 collect2: error: ld returned 1 exit status
933 makefile:348: recipe for target 'miniperl' failed
934 make: *** [miniperl] Error 1
935 Another P5P contributor recommended installation and use of
937 Devel::PatchPerl for this situation, first to determine the version of
938 perl at the commit in question, then to patch the source code at that
939 point to facilitate a build.
940 $ perl -MDevel::PatchPerl -e \
942 'print Devel::PatchPerl->determine_version("/path/to/sourcecode"), "\n";'
943 5.11.1
944 $ perl -MDevel::PatchPerl -e \
945 'Devel::PatchPerl->patch_source("5.11.1", "/path/to/sourcecode");'
946 Once the source was patched, "./Configure" and "make test_prep" were
948 called and completed successfully, enabling confirmation of the
949 findings in RT #72414.
950
952 To hack on the Perl guts, you'll need to read the following things:
953 · perlsource
955 An overview of the Perl source tree. This will help you find the
957 files you're looking for.
958 · perlinterp
960 An overview of the Perl interpreter source code and some details on
962 how Perl does what it does.
963 · perlhacktut
965 This document walks through the creation of a small patch to Perl's
967 C code. If you're just getting started with Perl core hacking,
968 this will help you understand how it works.
969 · perlhacktips
971 More details on hacking the Perl core. This document focuses on
973 lower level details such as how to write tests, compilation issues,
974 portability, debugging, etc.
975 If you plan on doing serious C hacking, make sure to read this.
977 · perlguts
979 This is of paramount importance, since it's the documentation of
981 what goes where in the Perl source. Read it over a couple of times
982 and it might start to make sense - don't worry if it doesn't yet,
983 because the best way to study it is to read it in conjunction with
984 poking at Perl source, and we'll do that later on.
985 Gisle Aas's "illustrated perlguts", also known as illguts, has very
987 helpful pictures:
988 <http://search.cpan.org/dist/illguts/>
990 · perlxstut and perlxs
992 A working knowledge of XSUB programming is incredibly useful for
994 core hacking; XSUBs use techniques drawn from the PP code, the
995 portion of the guts that actually executes a Perl program. It's a
996 lot gentler to learn those techniques from simple examples and
997 explanation than from the core itself.
998 · perlapi
1000 The documentation for the Perl API explains what some of the
1002 internal functions do, as well as the many macros used in the
1003 source.
1004 · Porting/pumpkin.pod
1006 This is a collection of words of wisdom for a Perl porter; some of
1008 it is only useful to the pumpkin holder, but most of it applies to
1009 anyone wanting to go about Perl development.
1010
1012 The CPAN testers ( <http://testers.cpan.org/> ) are a group of
1013 volunteers who test CPAN modules on a variety of platforms.
1014 Perl Smokers ( <http://www.nntp.perl.org/group/perl.daily-build/> and
1016 <http://www.nntp.perl.org/group/perl.daily-build.reports/> )
1017 automatically test Perl source releases on platforms with various
1018 configurations.
1019 Both efforts welcome volunteers. In order to get involved in smoke
1021 testing of the perl itself visit
1022 <http://search.cpan.org/dist/Test-Smoke/>. In order to start smoke
1023 testing CPAN modules visit
1024 <http://search.cpan.org/dist/CPANPLUS-YACSmoke/> or
1025 <http://search.cpan.org/dist/minismokebox/> or
1026 <http://search.cpan.org/dist/CPAN-Reporter/>.
1027
1029 If you've read all the documentation in the document and the ones
1030 listed above, you're more than ready to hack on Perl.
1031 Here's some more recommendations
1033 · Subscribe to perl5-porters, follow the patches and try and
1035 understand them; don't be afraid to ask if there's a portion you're
1036 not clear on - who knows, you may unearth a bug in the patch...
1037 · Do read the README associated with your operating system, e.g.
1039 README.aix on the IBM AIX OS. Don't hesitate to supply patches to
1040 that README if you find anything missing or changed over a new OS
1041 release.
1042 · Find an area of Perl that seems interesting to you, and see if you
1044 can work out how it works. Scan through the source, and step over
1045 it in the debugger. Play, poke, investigate, fiddle! You'll
1046 probably get to understand not just your chosen area but a much
1047 wider range of perl's activity as well, and probably sooner than
1048 you'd think.
1049 "The Road goes ever on and on, down from the door where it began."
1051 If you can do these things, you've started on the long road to Perl
1052 porting. Thanks for wanting to help make Perl better - and happy
1053 hacking!
1054 Metaphoric Quotations
1056 If you recognized the quote about the Road above, you're in luck.
1057 Most software projects begin each file with a literal description of
1059 each file's purpose. Perl instead begins each with a literary allusion
1060 to that file's purpose.
1061 Like chapters in many books, all top-level Perl source files (along
1063 with a few others here and there) begin with an epigrammatic
1064 inscription that alludes, indirectly and metaphorically, to the
1065 material you're about to read.
1066 Quotations are taken from writings of J.R.R. Tolkien pertaining to his
1068 Legendarium, almost always from The Lord of the Rings. Chapters and
1069 page numbers are given using the following editions:
1070 · The Hobbit, by J.R.R. Tolkien. The hardcover, 70th-anniversary
1072 edition of 2007 was used, published in the UK by Harper Collins
1073 Publishers and in the US by the Houghton Mifflin Company.
1074 · The Lord of the Rings, by J.R.R. Tolkien. The hardcover,
1076 50th-anniversary edition of 2004 was used, published in the UK by
1077 Harper Collins Publishers and in the US by the Houghton Mifflin
1078 Company.
1079 · The Lays of Beleriand, by J.R.R. Tolkien and published posthumously
1081 by his son and literary executor, C.J.R. Tolkien, being the 3rd of
1082 the 12 volumes in Christopher's mammoth History of Middle Earth.
1083 Page numbers derive from the hardcover edition, first published in
1084 1983 by George Allen & Unwin; no page numbers changed for the
1085 special 3-volume omnibus edition of 2002 or the various trade-paper
1086 editions, all again now by Harper Collins or Houghton Mifflin.
1087 Other JRRT books fair game for quotes would thus include The Adventures
1089 of Tom Bombadil, The Silmarillion, Unfinished Tales, and The Tale of
1090 the Children of Hurin, all but the first posthumously assembled by
1091 CJRT. But The Lord of the Rings itself is perfectly fine and probably
1092 best to quote from, provided you can find a suitable quote there.
1093 So if you were to supply a new, complete, top-level source file to add
1095 to Perl, you should conform to this peculiar practice by yourself
1096 selecting an appropriate quotation from Tolkien, retaining the original
1097 spelling and punctuation and using the same format the rest of the
1098 quotes are in. Indirect and oblique is just fine; remember, it's a
1099 metaphor, so being meta is, after all, what it's for.
1100
1102 This document was originally written by Nathan Torkington, and is
1103 maintained by the perl5-porters mailing list.
1104perl v5.28.2 2018-11-01 PERLHACK(1)