4 Consistent formatting of this file is achieved with:
5 perl ./Porting/podtidy pod/perlhack.pod
9 perlhack - How to hack on Perl
13 This document explains how Perl development works. It includes details
14 about the Perl 5 Porters email list, the Perl repository, the Perl
15 bug tracker, patch guidelines, and commentary on Perl development
18 =head1 SUPER QUICK PATCH GUIDE
20 If you just want to submit a single small patch like a pod fix, a test
21 for a bug, comment fixes, etc., it's easy! Here's how:
25 =item * Check out the source repository
27 The perl source is in a git repository. You can clone the repository
28 with the following command:
30 % git clone https://github.com/Perl/perl5.git perl
32 =item * Ensure you're following the latest advice
34 In case the advice in this guide has been updated recently, read the
35 latest version directly from the perl source:
37 % perldoc pod/perlhack.pod
39 =item * Create a branch for your change
41 Create a branch based on blead to commit your change to, which will
42 later be used to send it to the Perl issue tracker.
44 % git checkout -b mychange
46 =item * Make your change
48 Hack, hack, hack. Keep in mind that Perl runs on many different
49 platforms, with different operating systems that have different
50 capabilities, different filesystem organizations, and even different
51 character sets. L<perlhacktips> gives advice on this.
53 =item * Test your change
55 You can run all the tests with the following commands:
57 % ./Configure -des -Dusedevel
60 Keep hacking until the tests pass.
62 =item * Commit your change
64 Committing your work will save the change I<on your local system>:
66 % git commit -a -m 'Commit message goes here'
68 Make sure the commit message describes your change in a single
69 sentence. For example, "Fixed spelling errors in perlhack.pod".
71 =item * Send your change to the Perl issue tracker
73 The next step is to submit your patch to the Perl core ticket system.
75 Create a GitHub fork of the perl5 repository and add it as a remote,
76 if you haven't already, as described in the GitHub documentation at
77 L<https://help.github.com/en/articles/working-with-forks>.
79 % git remote add fork git@github.com:MyUser/perl5.git
81 Then, push your new branch to your fork.
83 % git push -u fork mychange
85 Finally, create a Pull Request on GitHub from your branch to blead as
86 described in the GitHub documentation at
87 L<https://help.github.com/en/articles/creating-a-pull-request-from-a-fork>.
91 The porters appreciate the time you spent helping to make Perl better.
94 =item * Acknowledgement
96 All contributors are credited (by name and email address) in the
97 AUTHORS file, which is part of the perl distribution, as well as the
100 If you don’t want to be included in the AUTHORS file, just let us
101 know. Otherwise we will take your submission of a patch as permission
102 to credit you in the AUTHORS file.
106 The next time you wish to make a patch, you need to start from the
107 latest perl in a pristine state. Check you don't have any local changes
108 or added files in your perl check-out which you wish to keep, then run
113 % git reset --hard origin/blead
120 If you want to report a bug in Perl, or browse existing Perl bugs and
121 patches, use the GitHub issue tracker at
122 L<https://github.com/perl/perl5/issues>.
124 Please check the archive of the perl5-porters list (see below) and/or
125 the bug tracking system before submitting a bug report. Often, you'll
126 find that the bug has been reported already.
128 You can log in to the bug tracking system and comment on existing bug
129 reports. If you have additional information regarding an existing bug,
130 please add it. This will help the porters fix the bug.
132 =head1 PERL 5 PORTERS
134 The perl5-porters (p5p) mailing list is where the Perl standard
135 distribution is maintained and developed. The people who maintain Perl
136 are also referred to as the "Perl 5 Porters", "p5p" or just the
139 A searchable archive of the list is available at
140 L<https://markmail.org/search/?q=perl5-porters>. There is also an archive at
141 L<https://archive.develooper.com/perl5-porters@perl.org/>.
143 =head2 perl-changes mailing list
145 The perl5-changes mailing list receives a copy of each patch that gets
146 submitted to the maintenance and development branches of the perl
147 repository. See L<https://lists.perl.org/list/perl5-changes.html> for
148 subscription and archive information.
152 Many porters are also active on the L<irc://irc.perl.org/#p5p> channel.
153 Feel free to join the channel and ask questions about hacking on the
156 =head1 GETTING THE PERL SOURCE
158 All of Perl's source code is kept centrally in a Git repository at
159 I<github.com>. The repository contains many Perl revisions
160 from Perl 1 onwards and all the revisions from Perforce, the previous
161 version control system.
163 For much more detail on using git with the Perl repository, please see
166 =head2 Read access via Git
168 You will need a copy of Git for your computer. You can fetch a copy of
169 the repository using the git protocol:
171 % git clone git://github.com/Perl/perl5.git perl
173 This clones the repository and makes a local copy in the F<perl>
176 If you cannot use the git protocol for firewall reasons, you can also
179 % git clone https://github.com/Perl/perl5.git perl
181 =head2 Read access via the web
183 You may access the repository over the web. This allows you to browse
184 the tree, see recent commits, subscribe to repository notifications,
185 search for particular commits and more. You may access it at
186 L<https://github.com/Perl/perl5>.
188 =head2 Read access via rsync
190 You can also choose to use rsync to get a copy of the current source
191 tree for the bleadperl branch and all maintenance branches:
193 % rsync -avz rsync://perl5.git.perl.org/perl-current .
194 % rsync -avz rsync://perl5.git.perl.org/perl-5.12.x .
195 % rsync -avz rsync://perl5.git.perl.org/perl-5.10.x .
196 % rsync -avz rsync://perl5.git.perl.org/perl-5.8.x .
197 % rsync -avz rsync://perl5.git.perl.org/perl-5.6.x .
198 % rsync -avz rsync://perl5.git.perl.org/perl-5.005xx .
200 (Add the C<--delete> option to remove leftover files.)
202 To get a full list of the available sync points:
204 % rsync perl5.git.perl.org::
206 =head2 Write access via git
208 If you have a commit bit, please see L<perlgit> for more details on
213 If you're planning to do more extensive work than a single small fix,
214 we encourage you to read the documentation below. This will help you
215 focus your work and make your patches easier to incorporate into the
218 =head2 Submitting patches
220 If you have a small patch to submit, please submit it via the GitHub
221 Pull Request workflow. You may also send patches to the p5p list.
223 Patches are reviewed and discussed on GitHub or the p5p list. Simple,
224 uncontroversial patches will usually be applied without any discussion.
225 When the patch is applied, the ticket will be updated and you will
228 In other cases, the patch will need more work or discussion.
229 You are encouraged to participate in the discussion and advocate for
230 your patch. Sometimes your patch may get lost in the shuffle. It's
231 appropriate to send a reminder email to p5p if no action has been taken
232 in a month. Please remember that the Perl 5 developers are all
233 volunteers, and be polite.
235 Changes are always applied directly to the main development branch,
236 called "blead". Some patches may be backported to a maintenance
237 branch. If you think your patch is appropriate for the maintenance
238 branch (see L<perlpolicy/MAINTENANCE BRANCHES>), please explain why
241 =head2 Getting your patch accepted
243 If you are submitting a code patch there are several things that you
244 can do to help the Perl 5 Porters accept your patch.
248 Using the GitHub Pull Request workflow, your patch will automatically
249 be available in a suitable format. If you wish to submit a patch to
250 the p5p list for review, make sure to create it appropriately.
252 If you used git to check out the Perl source, then using C<git
253 format-patch> will produce a patch in a style suitable for Perl. The
254 C<format-patch> command produces one patch file for each commit you
255 made. If you prefer to send a single patch for all commits, you can
260 % git diff blead my-branch-name
262 This produces a patch based on the difference between blead and your
263 current branch. It's important to make sure that blead is up to date
264 before producing the diff, that's why we call C<git pull> first.
266 We strongly recommend that you use git if possible. It will make your
267 life easier, and ours as well.
269 However, if you're not using git, you can still produce a suitable
270 patch. You'll need a pristine copy of the Perl source to diff against.
271 The porters prefer unified diffs. Using GNU C<diff>, you can produce a
274 % diff -Npurd perl.pristine perl.mine
276 Make sure that you C<make realclean> in your copy of Perl to remove any
277 build artifacts, or you may get a confusing result.
279 =head3 Commit message
281 As you craft each patch you intend to submit to the Perl core, it's
282 important to write a good commit message. This is especially important
283 if your submission will consist of a series of commits.
285 The first line of the commit message should be a short description
286 without a period. It should be no longer than the subject line of an
287 email, 50 characters being a good rule of thumb.
289 A lot of Git tools (Gitweb, GitHub, git log --pretty=oneline, ...) will
290 only display the first line (cut off at 50 characters) when presenting
293 The commit message should include a description of the problem that the
294 patch corrects or new functionality that the patch adds.
296 As a general rule of thumb, your commit message should help a
297 programmer who knows the Perl core quickly understand what you were
298 trying to do, how you were trying to do it, and why the change matters
305 Your commit message should describe why the change you are making is
306 important. When someone looks at your change in six months or six
307 years, your intent should be clear.
309 If you're deprecating a feature with the intent of later simplifying
310 another bit of code, say so. If you're fixing a performance problem or
311 adding a new feature to support some other bit of the core, mention
316 Your commit message should describe what part of the Perl core you're
317 changing and what you expect your patch to do.
321 While it's not necessary for documentation changes, new tests or
322 trivial patches, it's often worth explaining how your change works.
323 Even if it's clear to you today, it may not be clear to a porter next
328 A commit message isn't intended to take the place of comments in your
329 code. Commit messages should describe the change you made, while code
330 comments should describe the current state of the code.
332 If you've just implemented a new feature, complete with doc, tests and
333 well-commented code, a brief commit message will often suffice. If,
334 however, you've just changed a single character deep in the parser or
335 lexer, you might need to write a small novel to ensure that future
336 readers understand what you did and why you did it.
338 =head3 Comments, Comments, Comments
340 Be sure to adequately comment your code. While commenting every line
341 is unnecessary, anything that takes advantage of side effects of
342 operators, that creates changes that will be felt outside of the
343 function being patched, or that others may find confusing should be
344 documented. If you are going to err, it is better to err on the side
345 of adding too many comments than too few.
347 The best comments explain I<why> the code does what it does, not I<what
352 In general, please follow the particular style of the code you are
355 In particular, follow these general guidelines for patching Perl
362 4-wide indents for code, 2-wide indents for nested CPP C<#define>s,
363 with 8-wide tabstops.
367 Use spaces for indentation, not tab characters.
369 The codebase is a mixture of tabs and spaces for indentation, and we
370 are moving to spaces only. Converting lines you're patching from 8-wide
371 tabs to spaces will help this migration.
375 Try hard not to exceed 79-columns
383 Uncuddled elses and "K&R" style for indenting control constructs
387 No C++ style (//) comments
391 Mark places that need to be revisited with XXX (and revisit often!)
395 Opening brace lines up with "if" when conditional spans multiple lines;
396 should be at end-of-line otherwise
400 In function definitions, name starts in column 0 (return value-type is on
405 Single space after keywords that are followed by parens, no space
406 between function name and following paren
410 Avoid assignments in conditionals, but if they're unavoidable, use
411 extra paren, e.g. "if (a && (b = c)) ..."
415 "return foo;" rather than "return(foo);"
419 "if (!foo) ..." rather than "if (foo == FALSE) ..." etc.
423 Do not declare variables using "register". It may be counterproductive
424 with modern compilers, and is deprecated in C++, under which the Perl
425 source is regularly compiled.
429 In-line functions that are in headers that are accessible to XS code
430 need to be able to compile without warnings with commonly used extra
431 compilation flags, such as gcc's C<-Wswitch-default> which warns
432 whenever a switch statement does not have a "default" case. The use of
433 these extra flags is to catch potential problems in legal C code, and
434 is often used by Perl aggregators, such as Linux distributors.
440 If your patch changes code (rather than just changing documentation),
441 you should also include one or more test cases which illustrate the bug
442 you're fixing or validate the new functionality you're adding. In
443 general, you should update an existing test file rather than create a
446 Your test suite additions should generally follow these guidelines
447 (courtesy of Gurusamy Sarathy <gsar@activestate.com>):
453 Know what you're testing. Read the docs, and the source.
457 Tend to fail, not succeed.
461 Interpret results strictly.
465 Use unrelated features (this will flush out bizarre interactions).
469 Use non-standard idioms (otherwise you are not testing TIMTOWTDI).
473 Avoid using hardcoded test numbers whenever possible (the EXPECTED/GOT
474 found in t/op/tie.t is much more maintainable, and gives better failure
479 Give meaningful error messages when a test fails.
483 Avoid using qx// and system() unless you are testing for them. If you
484 do use them, make sure that you cover _all_ perl platforms.
488 Unlink any temporary files you create.
492 Promote unforeseen warnings to errors with $SIG{__WARN__}.
496 Be sure to use the libraries and modules shipped with the version being
497 tested, not those that were already installed.
501 Add comments to the code explaining what you are testing for.
505 Make updating the '1..42' string unnecessary. Or make sure that you
510 Test _all_ behaviors of a given operator, library, or function.
512 Test all optional arguments.
514 Test return values in various contexts (boolean, scalar, list, lvalue).
516 Use both global and lexical variables.
518 Don't forget the exceptional, pathological cases.
522 =head2 Patching a core module
524 This works just like patching anything else, with one extra
527 Modules in the F<cpan/> directory of the source tree are maintained
528 outside of the Perl core. When the author updates the module, the
529 updates are simply copied into the core. See that module's
530 documentation or its listing on L<https://metacpan.org/> for more
531 information on reporting bugs and submitting patches.
533 In most cases, patches to modules in F<cpan/> should be sent upstream
534 and should not be applied to the Perl core individually. If a patch to
535 a file in F<cpan/> absolutely cannot wait for the fix to be made
536 upstream, released to CPAN and copied to blead, you must add (or
537 update) a C<CUSTOMIZED> entry in the F<"Porting/Maintainers.pl"> file
538 to flag that a local modification has been made. See
539 F<"Porting/Maintainers.pl"> for more details.
541 In contrast, modules in the F<dist/> directory are maintained in the
544 =head2 Updating perldelta
546 For changes significant enough to warrant a F<pod/perldelta.pod> entry,
547 the porters will greatly appreciate it if you submit a delta entry
548 along with your actual change. Significant changes include, but are
555 Adding, deprecating, or removing core features
559 Adding, deprecating, removing, or upgrading core or dual-life modules
563 Adding new core tests
567 Fixing security issues and user-visible bugs in the core
571 Changes that might break existing code, either on the perl or C level
575 Significant performance improvements
579 Adding, removing, or significantly changing documentation in the
584 Important platform-specific changes
588 Please make sure you add the perldelta entry to the right section
589 within F<pod/perldelta.pod>. More information on how to write good
590 perldelta entries is available in the C<Style> section of
591 F<Porting/how_to_write_a_perldelta.pod>.
593 =head2 What makes for a good patch?
595 New features and extensions to the language can be contentious. There
596 is no specific set of criteria which determine what features get added,
597 but here are some questions to consider when developing a patch:
599 =head3 Does the concept match the general goals of Perl?
601 Our goals include, but are not limited to:
607 Keep it fast, simple, and useful.
611 Keep features/concepts as orthogonal as possible.
615 No arbitrary limits (platforms, data sizes, cultures).
619 Keep it open and exciting to use/patch/advocate Perl everywhere.
623 Either assimilate new technologies, or build bridges to them.
627 =head3 Where is the implementation?
629 All the talk in the world is useless without an implementation. In
630 almost every case, the person or people who argue for a new feature
631 will be expected to be the ones who implement it. Porters capable of
632 coding new features have their own agendas, and are not available to
633 implement your (possibly good) idea.
635 =head3 Backwards compatibility
637 It's a cardinal sin to break existing Perl programs. New warnings can
638 be contentious--some say that a program that emits warnings is not
639 broken, while others say it is. Adding keywords has the potential to
640 break programs, changing the meaning of existing token sequences or
641 functions might break programs.
643 The Perl 5 core includes mechanisms to help porters make backwards
644 incompatible changes more compatible such as the L<feature> and
645 L<deprecate> modules. Please use them when appropriate.
647 =head3 Could it be a module instead?
649 Perl 5 has extension mechanisms, modules and XS, specifically to avoid
650 the need to keep changing the Perl interpreter. You can write modules
651 that export functions, you can give those functions prototypes so they
652 can be called like built-in functions, you can even write XS code to
653 mess with the runtime data structures of the Perl interpreter if you
654 want to implement really complicated things.
656 Whenever possible, new features should be prototyped in a CPAN module
657 before they will be considered for the core.
659 =head3 Is the feature generic enough?
661 Is this something that only the submitter wants added to the language,
662 or is it broadly useful? Sometimes, instead of adding a feature with a
663 tight focus, the porters might decide to wait until someone implements
664 the more generalized feature.
666 =head3 Does it potentially introduce new bugs?
668 Radical rewrites of large chunks of the Perl interpreter have the
669 potential to introduce new bugs.
671 =head3 How big is it?
673 The smaller and more localized the change, the better. Similarly, a
674 series of small patches is greatly preferred over a single large patch.
676 =head3 Does it preclude other desirable features?
678 A patch is likely to be rejected if it closes off future avenues of
679 development. For instance, a patch that placed a true and final
680 interpretation on prototypes is likely to be rejected because there are
681 still options for the future of prototypes that haven't been addressed.
683 =head3 Is the implementation robust?
685 Good patches (tight code, complete, correct) stand more chance of going
686 in. Sloppy or incorrect patches might be placed on the back burner
687 until the pumpking has time to fix, or might be discarded altogether
688 without further notice.
690 =head3 Is the implementation generic enough to be portable?
692 The worst patches make use of system-specific features. It's highly
693 unlikely that non-portable additions to the Perl language will be
696 =head3 Is the implementation tested?
698 Patches which change behaviour (fixing bugs or introducing new
699 features) must include regression tests to verify that everything works
702 Without tests provided by the original author, how can anyone else
703 changing perl in the future be sure that they haven't unwittingly
704 broken the behaviour the patch implements? And without tests, how can
705 the patch's author be confident that his/her hard work put into the
706 patch won't be accidentally thrown away by someone in the future?
708 =head3 Is there enough documentation?
710 Patches without documentation are probably ill-thought out or
711 incomplete. No features can be added or changed without documentation,
712 so submitting a patch for the appropriate pod docs as well as the
713 source code is important.
715 =head3 Is there another way to do it?
717 Larry said "Although the Perl Slogan is I<There's More Than One Way to
718 Do It>, I hesitate to make 10 ways to do something". This is a tricky
719 heuristic to navigate, though--one man's essential addition is another
720 man's pointless cruft.
722 =head3 Does it create too much work?
724 Work for the pumpking, work for Perl programmers, work for module
725 authors, ... Perl is supposed to be easy.
727 =head3 Patches speak louder than words
729 Working code is always preferred to pie-in-the-sky ideas. A patch to
730 add a feature stands a much higher chance of making it to the language
731 than does a random feature request, no matter how fervently argued the
732 request might be. This ties into "Will it be useful?", as the fact
733 that someone took the time to make the patch demonstrates a strong
734 desire for the feature.
738 The core uses the same testing style as the rest of Perl, a simple
739 "ok/not ok" run through Test::Harness, but there are a few special
742 There are three ways to write a test in the core: L<Test::More>,
743 F<t/test.pl> and ad hoc C<print $test ? "ok 42\n" : "not ok 42\n">.
744 The decision of which to use depends on what part of the test suite
745 you're working on. This is a measure to prevent a high-level failure
746 (such as Config.pm breaking) from causing basic functionality tests to
749 The F<t/test.pl> library provides some of the features of
750 L<Test::More>, but avoids loading most modules and uses as few core
751 features as possible.
753 If you write your own test, use the L<Test Anything
754 Protocol|https://testanything.org>.
758 =item * F<t/base>, F<t/comp> and F<t/opbasic>
760 Since we don't know if C<require> works, or even subroutines, use ad hoc
761 tests for these three. Step carefully to avoid using the feature being
762 tested. Tests in F<t/opbasic>, for instance, have been placed there
763 rather than in F<t/op> because they test functionality which
764 F<t/test.pl> presumes has already been demonstrated to work.
766 =item * F<t/cmd>, F<t/run>, F<t/io> and F<t/op>
768 Now that basic require() and subroutines are tested, you can use the
769 F<t/test.pl> library.
771 You can also use certain libraries like Config conditionally, but be
772 sure to skip the test gracefully if it's not there.
774 =item * Everything else
776 Now that the core of Perl is tested, L<Test::More> can and should be
777 used. You can also use the full suite of core modules in the tests.
781 When you say "make test", Perl uses the F<t/TEST> program to run the
782 test suite (except under Win32 where it uses F<t/harness> instead).
783 All tests are run from the F<t/> directory, B<not> the directory which
784 contains the test. This causes some problems with the tests in
785 F<lib/>, so here's some opportunity for some patching.
787 You must be triply conscious of cross-platform concerns. This usually
788 boils down to using L<File::Spec>, avoiding things like C<fork()>
789 and C<system()> unless absolutely necessary, and not assuming that a
790 given character has a particular ordinal value (code point) or that its
791 UTF-8 representation is composed of particular bytes.
793 There are several functions available to specify characters and code
794 points portably in tests. The always-preloaded functions
795 C<utf8::unicode_to_native()> and its inverse
796 C<utf8::native_to_unicode()> take code points and translate
797 appropriately. The file F<t/charset_tools.pl> has several functions
798 that can be useful. It has versions of the previous two functions
799 that take strings as inputs -- not single numeric code points:
800 C<uni_to_native()> and C<native_to_uni()>. If you must look at the
801 individual bytes comprising a UTF-8 encoded string,
802 C<byte_utf8a_to_utf8n()> takes as input a string of those bytes encoded
803 for an ASCII platform, and returns the equivalent string in the native
804 platform. For example, C<byte_utf8a_to_utf8n("\xC2\xA0")> returns the
805 byte sequence on the current platform that form the UTF-8 for C<U+00A0>,
806 since C<"\xC2\xA0"> are the UTF-8 bytes on an ASCII platform for that
807 code point. This function returns C<"\xC2\xA0"> on an ASCII platform, and
808 C<"\x80\x41"> on an EBCDIC 1047 one.
810 But easiest is, if the character is specifiable as a literal, like
811 C<"A"> or C<"%">, to use that; if not so specificable, you can use use
812 C<\N{}> , if the side effects aren't troublesome. Simply specify all
813 your characters in hex, using C<\N{U+ZZ}> instead of C<\xZZ>. C<\N{}>
814 is the Unicode name, and so it
815 always gives you the Unicode character. C<\N{U+41}> is the character
816 whose Unicode code point is C<0x41>, hence is C<'A'> on all platforms.
817 The side effects are:
823 These select Unicode rules. That means that in double-quotish strings,
824 the string is always converted to UTF-8 to force a Unicode
825 interpretation (you can C<utf8::downgrade()> afterwards to convert back
826 to non-UTF8, if possible). In regular expression patterns, the
827 conversion isn't done, but if the character set modifier would
828 otherwise be C</d>, it is changed to C</u>.
832 If you use the form C<\N{I<character name>}>, the L<charnames> module
833 gets automatically loaded. This may not be suitable for the test level
838 If you are testing locales (see L<perllocale>), there are helper
839 functions in F<t/loc_tools.pl> to enable you to see what locales there
840 are on the current platform.
842 =head2 Special C<make test> targets
844 There are various special make targets that can be used to test Perl
845 slightly differently than the standard "test" target. Not all them are
846 expected to give a 100% success rate. Many of them have several
847 aliases, and many of them are not available on certain operating
854 This runs some basic sanity tests on the source tree and helps catch
855 basic errors before you submit a patch.
859 Run F<miniperl> on F<t/base>, F<t/comp>, F<t/cmd>, F<t/run>, F<t/io>,
860 F<t/op>, F<t/uni> and F<t/mro> tests.
862 =item * test.valgrind check.valgrind
864 (Only in Linux) Run all the tests using the memory leak + naughty
865 memory access tool "valgrind". The log files will be named
866 F<testname.valgrind>.
870 Run the test suite with the F<t/harness> controlling program, instead
871 of F<t/TEST>. F<t/harness> is more sophisticated, and uses the
872 L<Test::Harness> module, thus using this test target supposes that perl
873 mostly works. The main advantage for our purposes is that it prints a
874 detailed summary of failed tests at the end. Also, unlike F<t/TEST>,
875 it doesn't redirect stderr to stdout.
877 Note that under Win32 F<t/harness> is always used instead of F<t/TEST>,
878 so there is no special "test_harness" target.
880 Under Win32's "test" target you may use the TEST_SWITCHES and
881 TEST_FILES environment variables to control the behaviour of
882 F<t/harness>. This means you can say
884 nmake test TEST_FILES="op/*.t"
885 nmake test TEST_SWITCHES="-torture" TEST_FILES="op/*.t"
887 =item * test-notty test_notty
889 Sets PERL_SKIP_TTY_TEST to true before running normal test.
893 =head2 Parallel tests
895 The core distribution can now run its regression tests in parallel on
896 Unix-like platforms. Instead of running C<make test>, set C<TEST_JOBS>
897 in your environment to the number of tests to run in parallel, and run
898 C<make test_harness>. On a Bourne-like shell, this can be done as
900 TEST_JOBS=3 make test_harness # Run 3 tests in parallel
902 An environment variable is used, rather than parallel make itself,
903 because L<TAP::Harness> needs to be able to schedule individual
904 non-conflicting test scripts itself, and there is no standard interface
905 to C<make> utilities to interact with their job schedulers.
907 Note that currently some test scripts may fail when run in parallel
908 (most notably F<dist/IO/t/io_dir.t>). If necessary, run just the
909 failing scripts again sequentially and see if the failures go away.
911 =head2 Running tests by hand
913 You can run part of the test suite by hand by using one of the
914 following commands from the F<t/> directory:
916 ./perl -I../lib TEST list-of-.t-files
920 ./perl -I../lib harness list-of-.t-files
922 (If you don't specify test scripts, the whole test suite will be run.)
924 =head2 Using F<t/harness> for testing
926 If you use C<harness> for testing, you have several command line
927 options available to you. The arguments are as follows, and are in the
928 order that they must appear if used together.
930 harness -v -torture -re=pattern LIST OF FILES TO TEST
931 harness -v -torture -re LIST OF PATTERNS TO MATCH
933 If C<LIST OF FILES TO TEST> is omitted, the file list is obtained from
934 the manifest. The file list may include shell wildcards which will be
941 Run the tests under verbose mode so you can see what tests were run,
946 Run the torture tests as well as the normal set.
950 Filter the file list so that all the test files run match PATTERN.
951 Note that this form is distinct from the B<-re LIST OF PATTERNS> form
952 below in that it allows the file list to be provided as well.
954 =item * -re LIST OF PATTERNS
956 Filter the file list so that all the test files run match
957 /(LIST|OF|PATTERNS)/. Note that with this form the patterns are joined
958 by '|' and you cannot supply a list of files, instead the test files
959 are obtained from the MANIFEST.
963 You can run an individual test by a command similar to
965 ./perl -I../lib path/to/foo.t
967 except that the harnesses set up some environment variables that may
968 affect the execution of the test:
974 indicates that we're running this test as part of the perl core test
975 suite. This is useful for modules that have a dual life on CPAN.
977 =item * PERL_DESTRUCT_LEVEL=2
979 is set to 2 if it isn't set already (see
980 L<perlhacktips/PERL_DESTRUCT_LEVEL>).
984 (used only by F<t/TEST>) if set, overrides the path to the perl
985 executable that should be used to run the tests (the default being
988 =item * PERL_SKIP_TTY_TEST
990 if set, tells to skip the tests that need a terminal. It's actually
991 set automatically by the Makefile, but can also be forced artificially
992 by running 'make test_notty'.
996 =head3 Other environment variables that may influence tests
1000 =item * PERL_TEST_Net_Ping
1002 Setting this variable runs all the Net::Ping modules tests, otherwise
1003 some tests that interact with the outside world are skipped. See
1006 =item * PERL_TEST_NOVREXX
1008 Setting this variable skips the vrexx.t tests for OS2::REXX.
1010 =item * PERL_TEST_NUMCONVERTS
1012 This sets a variable in op/numconvert.t.
1014 =item * PERL_TEST_MEMORY
1016 Setting this variable includes the tests in F<t/bigmem/>. This should
1017 be set to the number of gigabytes of memory available for testing, eg.
1018 C<PERL_TEST_MEMORY=4> indicates that tests that require 4GiB of
1019 available memory can be run safely.
1023 See also the documentation for the Test and Test::Harness modules, for
1024 more environment variables that affect testing.
1026 =head2 Performance testing
1028 The file F<t/perf/benchmarks> contains snippets of perl code which are
1029 intended to be benchmarked across a range of perls by the
1030 F<Porting/bench.pl> tool. If you fix or enhance a performance issue, you
1031 may want to add a representative code sample to the file, then run
1032 F<bench.pl> against the previous and current perls to see what difference
1033 it has made, and whether anything else has slowed down as a consequence.
1035 The file F<t/perf/opcount.t> is designed to test whether a particular
1036 code snippet has been compiled into an optree containing specified
1037 numbers of particular op types. This is good for testing whether
1038 optimisations which alter ops, such as converting an C<aelem> op into an
1039 C<aelemfast> op, are really doing that.
1041 The files F<t/perf/speed.t> and F<t/re/speed.t> are designed to test
1042 things that run thousands of times slower if a particular optimisation
1043 is broken (for example, the utf8 length cache on long utf8 strings).
1044 Add a test that will take a fraction of a second normally, and minutes
1045 otherwise, causing the test file to time out on failure.
1047 =head2 Building perl at older commits
1049 In the course of hacking on the Perl core distribution, you may have occasion
1050 to configure, build and test perl at an old commit. Sometimes C<make> will
1051 fail during this process. If that happens, you may be able to salvage the
1052 situation by using the Devel::PatchPerl library from CPAN (not included in the
1053 core) to bring the source code at that commit to a buildable state.
1055 Here's a real world example, taken from work done to resolve
1056 L<perl #10118|https://github.com/Perl/perl5/issues/10118>.
1057 Use of F<Porting/bisect.pl> had identified commit
1058 C<ba77e4cc9d1ceebf472c9c5c18b2377ee47062e6> as the commit in which a bug was
1059 corrected. To confirm, a P5P developer wanted to configure and build perl at
1060 commit C<ba77e4c^> (presumably "bad") and then at C<ba77e4c> (presumably
1061 "good"). Normal configuration and build was attempted:
1063 $ sh ./Configure -des -Dusedevel
1066 C<make>, however, failed with output (excerpted) like this:
1068 cc -fstack-protector -L/usr/local/lib -o miniperl \
1069 gv.o toke.o perly.o pad.o regcomp.o dump.o util.o \
1070 mg.o reentr.o mro.o hv.o av.o run.o pp_hot.o sv.o \
1071 pp.o scope.o pp_ctl.o pp_sys.o doop.o doio.o regexec.o \
1072 utf8.o taint.o deb.o universal.o globals.o perlio.o \
1073 perlapi.o numeric.o mathoms.o locale.o pp_pack.o pp_sort.o \
1074 miniperlmain.o opmini.o perlmini.o
1075 pp.o: In function `Perl_pp_pow':
1076 pp.c:(.text+0x2db9): undefined reference to `pow'
1078 collect2: error: ld returned 1 exit status
1079 makefile:348: recipe for target 'miniperl' failed
1080 make: *** [miniperl] Error 1
1082 Another P5P contributor recommended installation and use of Devel::PatchPerl
1083 for this situation, first to determine the version of perl at the commit in
1084 question, then to patch the source code at that point to facilitate a build.
1086 $ perl -MDevel::PatchPerl -e \
1087 'print Devel::PatchPerl->determine_version("/path/to/sourcecode"), "\n";'
1089 $ perl -MDevel::PatchPerl -e \
1090 'Devel::PatchPerl->patch_source("5.11.1", "/path/to/sourcecode");'
1092 Once the source was patched, C<./Configure> and C<make test_prep> were called
1093 and completed successfully, enabling confirmation of the findings in RT
1096 =head1 MORE READING FOR GUTS HACKERS
1098 To hack on the Perl guts, you'll need to read the following things:
1102 =item * L<perlsource>
1104 An overview of the Perl source tree. This will help you find the files
1107 =item * L<perlinterp>
1109 An overview of the Perl interpreter source code and some details on how
1110 Perl does what it does.
1112 =item * L<perlhacktut>
1114 This document walks through the creation of a small patch to Perl's C
1115 code. If you're just getting started with Perl core hacking, this will
1116 help you understand how it works.
1118 =item * L<perlhacktips>
1120 More details on hacking the Perl core. This document focuses on lower
1121 level details such as how to write tests, compilation issues,
1122 portability, debugging, etc.
1124 If you plan on doing serious C hacking, make sure to read this.
1128 This is of paramount importance, since it's the documentation of what
1129 goes where in the Perl source. Read it over a couple of times and it
1130 might start to make sense - don't worry if it doesn't yet, because the
1131 best way to study it is to read it in conjunction with poking at Perl
1132 source, and we'll do that later on.
1134 Gisle Aas's "illustrated perlguts", also known as I<illguts>, has very
1137 L<https://search.cpan.org/dist/illguts/>
1139 =item * L<perlxstut> and L<perlxs>
1141 A working knowledge of XSUB programming is incredibly useful for core
1142 hacking; XSUBs use techniques drawn from the PP code, the portion of
1143 the guts that actually executes a Perl program. It's a lot gentler to
1144 learn those techniques from simple examples and explanation than from
1149 The documentation for the Perl API explains what some of the internal
1150 functions do, as well as the many macros used in the source.
1152 =item * F<Porting/pumpkin.pod>
1154 This is a collection of words of wisdom for a Perl porter; some of it
1155 is only useful to the pumpkin holder, but most of it applies to anyone
1156 wanting to go about Perl development.
1160 =head1 CPAN TESTERS AND PERL SMOKERS
1162 The CPAN testers ( L<http://cpantesters.org/> ) are a group of volunteers
1163 who test CPAN modules on a variety of platforms.
1165 Perl Smokers ( L<https://www.nntp.perl.org/group/perl.daily-build/> and
1166 L<https://www.nntp.perl.org/group/perl.daily-build.reports/> )
1167 automatically test Perl source releases on platforms with various
1170 Both efforts welcome volunteers. In order to get involved in smoke
1171 testing of the perl itself visit
1172 L<https://metacpan.org/release/Test-Smoke>. In order to start smoke
1173 testing CPAN modules visit
1174 L<https://metacpan.org/release/CPANPLUS-YACSmoke> or
1175 L<https://metacpan.org/release/minismokebox> or
1176 L<https://metacpan.org/release/CPAN-Reporter>.
1180 If you've read all the documentation in the document and the ones
1181 listed above, you're more than ready to hack on Perl.
1183 Here's some more recommendations
1189 Subscribe to perl5-porters, follow the patches and try and understand
1190 them; don't be afraid to ask if there's a portion you're not clear on -
1191 who knows, you may unearth a bug in the patch...
1195 Do read the README associated with your operating system, e.g.
1196 README.aix on the IBM AIX OS. Don't hesitate to supply patches to that
1197 README if you find anything missing or changed over a new OS release.
1201 Find an area of Perl that seems interesting to you, and see if you can
1202 work out how it works. Scan through the source, and step over it in
1203 the debugger. Play, poke, investigate, fiddle! You'll probably get to
1204 understand not just your chosen area but a much wider range of
1205 F<perl>'s activity as well, and probably sooner than you'd think.
1209 =head2 "The Road goes ever on and on, down from the door where it began."
1211 If you can do these things, you've started on the long road to Perl
1212 porting. Thanks for wanting to help make Perl better - and happy
1215 =head2 Metaphoric Quotations
1217 If you recognized the quote about the Road above, you're in luck.
1219 Most software projects begin each file with a literal description of
1220 each file's purpose. Perl instead begins each with a literary allusion
1221 to that file's purpose.
1223 Like chapters in many books, all top-level Perl source files (along
1224 with a few others here and there) begin with an epigrammatic
1225 inscription that alludes, indirectly and metaphorically, to the
1226 material you're about to read.
1228 Quotations are taken from writings of J.R.R. Tolkien pertaining to his
1229 Legendarium, almost always from I<The Lord of the Rings>. Chapters and
1230 page numbers are given using the following editions:
1236 I<The Hobbit>, by J.R.R. Tolkien. The hardcover, 70th-anniversary
1237 edition of 2007 was used, published in the UK by Harper Collins
1238 Publishers and in the US by the Houghton Mifflin Company.
1242 I<The Lord of the Rings>, by J.R.R. Tolkien. The hardcover,
1243 50th-anniversary edition of 2004 was used, published in the UK by
1244 Harper Collins Publishers and in the US by the Houghton Mifflin
1249 I<The Lays of Beleriand>, by J.R.R. Tolkien and published posthumously
1250 by his son and literary executor, C.J.R. Tolkien, being the 3rd of the
1251 12 volumes in Christopher's mammoth I<History of Middle Earth>. Page
1252 numbers derive from the hardcover edition, first published in 1983 by
1253 George Allen & Unwin; no page numbers changed for the special 3-volume
1254 omnibus edition of 2002 or the various trade-paper editions, all again
1255 now by Harper Collins or Houghton Mifflin.
1259 Other JRRT books fair game for quotes would thus include I<The
1260 Adventures of Tom Bombadil>, I<The Silmarillion>, I<Unfinished Tales>,
1261 and I<The Tale of the Children of Hurin>, all but the first
1262 posthumously assembled by CJRT. But I<The Lord of the Rings> itself is
1263 perfectly fine and probably best to quote from, provided you can find a
1264 suitable quote there.
1266 So if you were to supply a new, complete, top-level source file to add
1267 to Perl, you should conform to this peculiar practice by yourself
1268 selecting an appropriate quotation from Tolkien, retaining the original
1269 spelling and punctuation and using the same format the rest of the
1270 quotes are in. Indirect and oblique is just fine; remember, it's a
1271 metaphor, so being meta is, after all, what it's for.
1275 This document was originally written by Nathan Torkington, and is
1276 maintained by the perl5-porters mailing list.