3 Pumpkin - Notes on handling the Perl Patch Pumpkin And Porting Perl
7 There is no simple synopsis, yet.
11 This document attempts to begin to describe some of the considerations
12 involved in patching, porting, and maintaining perl.
14 This document is still under construction, and still subject to
15 significant changes. Still, I hope parts of it will be useful,
16 so I'm releasing it even though it's not done.
18 For the most part, it's a collection of anecdotal information that
19 already assumes some familiarity with the Perl sources. I really need
20 an introductory section that describes the organization of the sources
21 and all the various auxiliary files that are part of the distribution.
23 =head1 Where Do I Get Perl Sources and Related Material?
25 The Comprehensive Perl Archive Network (or CPAN) is the place to go.
26 There are many mirrors, but the easiest thing to use is probably
27 http://www.cpan.org/README.html , which automatically points you to a
28 mirror site "close" to you.
30 =head2 Perl5-porters mailing list
32 The mailing list perl5-porters@perl.org
33 is the main group working with the development of perl. If you're
34 interested in all the latest developments, you should definitely
35 subscribe. The list is high volume, but generally has a
36 fairly low noise level.
38 Subscribe by sending the message (in the body of your letter)
40 subscribe perl5-porters
42 to perl5-porters-request@perl.org .
44 Archives of the list are held at:
46 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
48 =head1 How are Perl Releases Numbered?
50 Beginning with v5.6.0, even versions will stand for maintenance releases
51 and odd versions for development releases, i.e., v5.6.x for maintenance
52 releases, and v5.7.x for development releases. Before v5.6.0, subversions
53 _01 through _49 were reserved for bug-fix maintenance releases, and
54 subversions _50 through _99 for unstable development versions.
56 For example, in v5.6.1, the revision number is 5, the version is 6,
57 and 1 is the subversion.
59 For compatibility with the older numbering scheme the composite floating
60 point version number continues to be available as the magic variable $],
61 and amounts to C<$revision + $version/1000 + $subversion/100000>. This
62 can still be used in comparisons.
64 print "You've got an old perl\n" if $] < 5.005_03;
66 In addition, the version is also available as a string in $^V.
68 print "You've got a new perl\n" if $^V and $^V ge v5.6.0;
70 You can also require particular version (or later) with:
74 or using the new syntax available only from v5.6 onward:
78 At some point in the future, we may need to decide what to call the
79 next big revision. In the .package file used by metaconfig to
80 generate Configure, there are two variables that might be relevant:
81 $baserev=5 and $package=perl5.
83 Perl releases produced by the members of perl5-porters are usually
84 available on CPAN in the F<src/5.0/maint> and F<src/5.0/devel>
87 =head2 Maintenance and Development Subversions
89 The first rule of maintenance work is "First, do no harm."
91 Trial releases of bug-fix maintenance releases are announced on
92 perl5-porters. Trial releases use the new subversion number (to avoid
93 testers installing it over the previous release) and include a 'local
94 patch' entry in F<patchlevel.h>. The distribution file contains the
95 string C<MAINT_TRIAL> to make clear that the file is not meant for
98 In general, the names of official distribution files for the public
99 always match the regular expression:
101 ^perl\d+\.(\d+)\.\d+(-MAINT_TRIAL_\d+)\.tar\.gz$
103 C<$1> in the pattern is always an even number for maintenance
104 versions, and odd for developer releases.
106 In the past it has been observed that pumpkings tend to invent new
107 naming conventions on the fly. If you are a pumpking, before you
108 invent a new name for any of the three types of perl distributions,
109 please inform the guys from the CPAN who are doing indexing and
110 provide the trees of symlinks and the like. They will have to know
111 I<in advance> what you decide.
113 =head2 Why is it called the patch pumpkin?
115 Chip Salzenberg gets credit for that, with a nod to his cow orker,
116 David Croy. We had passed around various names (baton, token, hot
117 potato) but none caught on. Then, Chip asked:
121 Who has the patch pumpkin?
123 To explain: David Croy once told me once that at a previous job,
124 there was one tape drive and multiple systems that used it for backups.
125 But instead of some high-tech exclusion software, they used a low-tech
126 method to prevent multiple simultaneous backups: a stuffed pumpkin.
127 No one was allowed to make backups unless they had the "backup pumpkin".
133 =head1 Philosophical Issues in Patching and Porting Perl
135 There are no absolute rules, but there are some general guidelines I
136 have tried to follow as I apply patches to the perl sources.
137 (This section is still under construction.)
139 =head2 Solve problems as generally as possible
141 Never implement a specific restricted solution to a problem when you
142 can solve the same problem in a more general, flexible way.
144 For example, for dynamic loading to work on some SVR4 systems, we had
145 to build a shared libperl.so library. In order to build "FAT" binaries
146 on NeXT 4.0 systems, we had to build a special libperl library. Rather
147 than continuing to build a contorted nest of special cases, I
148 generalized the process of building libperl so that NeXT and SVR4 users
149 could still get their work done, but others could build a shared
150 libperl if they wanted to as well.
152 Contain your changes carefully. Assume nothing about other operating
153 systems, not even closely related ones. Your changes must not affect
156 Spy shamelessly on how similar patching or porting issues have been
159 If feasible, try to keep filenames 8.3-compliant to humor those poor
160 souls that get joy from running Perl under such dire limitations.
161 There's a script, F<check83.pl>, for keeping your nose 8.3-clean.
162 In a similar vein, do not create files or directories which differ only
163 in case (upper versus lower).
165 =head2 Seek consensus on major changes
167 If you are making big changes, don't do it in secret. Discuss the
168 ideas in advance on perl5-porters.
170 =head2 Keep the documentation up-to-date
172 If your changes may affect how users use perl, then check to be sure
173 that the documentation is in sync with your changes. Be sure to
174 check all the files F<pod/*.pod> and also the F<INSTALL> document.
176 Consider writing the appropriate documentation first and then
177 implementing your change to correspond to the documentation.
179 =head2 Avoid machine-specific #ifdef's
181 To the extent reasonable, try to avoid machine-specific #ifdef's in
182 the sources. Instead, use feature-specific #ifdef's. The reason is
183 that the machine-specific #ifdef's may not be valid across major
184 releases of the operating system. Further, the feature-specific tests
185 may help out folks on another platform who have the same problem.
187 =head2 Machine-specific files
193 If you have many machine-specific #defines or #includes, consider
194 creating an "osish.h" (F<os2ish.h>, F<vmsish.h>, and so on) and including
195 that in F<perl.h>. If you have several machine-specific files (function
196 emulations, function stubs, build utility wrappers) you may create a
197 separate subdirectory (djgpp, win32) and put the files in there.
198 Remember to update C<MANIFEST> when you add files.
200 If your system supports dynamic loading but none of the existing
201 methods at F<ext/DynaLoader/dl_*.xs> work for you, you must write
202 a new one. Study the existing ones to see what kind of interface
207 There are two kinds of hints: hints for building Perl and hints for
208 extensions. The former live in the C<hints> subdirectory, the latter
209 in C<ext/*/hints> subdirectories.
211 The top level hints are Bourne-shell scripts that set, modify and
212 unset appropriate Configure variables, based on the Configure command
213 line options and possibly existing config.sh and Policy.sh files from
214 previous Configure runs.
216 The extension hints are written in Perl (by the time they are used
217 miniperl has been built) and control the building of their respective
218 extensions. They can be used to for example manipulate compilation
221 =item build and installation Makefiles, scripts, and so forth
223 Sometimes you will also need to tweak the Perl build and installation
224 procedure itself, like for example F<Makefile.SH> and F<installperl>.
225 Tread very carefully, even more than usual. Contain your changes
230 Many of the tests in C<t> subdirectory assume machine-specific things
231 like existence of certain functions, something about filesystem
232 semantics, certain external utilities and their error messages. Use
233 the C<$^O> and the C<Config> module (which contains the results of the
234 Configure run, in effect the C<config.sh> converted to Perl) to either
235 skip (preferably not) or customize (preferable) the tests for your
240 Certain standard modules may need updating if your operating system
241 sports for example a native filesystem naming. You may want to update
242 some or all of the modules File::Basename, File::Spec, File::Path, and
243 File::Copy to become aware of your native filesystem syntax and
246 Remember to have a $VERSION in the modules. You can use the
247 F<Porting/checkVERSION.pl> script for checking this.
251 If your operating system comes from outside UNIX you almost certainly
252 will have differences in the available operating system functionality
253 (missing system calls, different semantics, whatever). Please
254 document these at F<pod/perlport.pod>. If your operating system is
255 the first B<not> to have a system call also update the list of
256 "portability-bewares" at the beginning of F<pod/perlfunc.pod>.
258 A file called F<README.youros> at the top level that explains things
259 like how to install perl at this platform, where to get any possibly
260 required additional software, and for example what test suite errors
261 to expect, is nice too. Such files are in the process of being written
262 in pod format and will eventually be renamed F<INSTALL.youros>.
264 You may also want to write a separate F<.pod> file for your operating
265 system to tell about existing mailing lists, os-specific modules,
266 documentation, whatever. Please name these along the lines of
267 F<perl>I<youros>.pod. [unfinished: where to put this file (the pod/
268 subdirectory, of course: but more importantly, which/what index files
273 =head2 Allow for lots of testing
275 We should never release a main version without testing it as a
278 =head2 Test popular applications and modules.
280 We should never release a main version without testing whether or not
281 it breaks various popular modules and applications. A partial list of
282 such things would include majordomo, metaconfig, apache, Tk, CGI,
283 libnet, and libwww, to name just a few. Of course it's quite possible
284 that some of those things will be just plain broken and need to be fixed,
285 but, in general, we ought to try to avoid breaking widely-installed
288 =head2 Automated generation of derivative files
290 The F<embed.h>, F<keywords.h>, F<opcode.h>, F<regcharclass.h>,
291 F<l1_char_class_tab.h>, and F<perltoc.pod> files
292 are all automatically generated by perl scripts. In general, don't
293 patch these directly; patch the data files instead.
295 F<Configure> and F<config_h.SH> are also automatically generated by
296 B<metaconfig>. In general, you should patch the metaconfig units
297 instead of patching these files directly. However, very minor changes
298 to F<Configure> may be made in between major sync-ups with the
299 metaconfig units, which tends to be complicated operations. But be
300 careful, this can quickly spiral out of control. Running metaconfig
303 Also F<Makefile> is automatically produced from F<Makefile.SH>.
304 In general, look out for all F<*.SH> files.
306 Finally, the sample files in the F<Porting/> subdirectory are
307 generated automatically by the script F<U/mksample> included
308 with the metaconfig units. See L<"run metaconfig"> below for
309 information on obtaining the metaconfig units.
311 =head1 How to Make a Distribution
313 This section has now been expanded and moved into its own file,
314 F<Porting/release_managers_guide.pod>.
316 I've kept some of the subsections here for now, as they don't directly
317 relate to building a release any more, but still contain what might be
318 useful information - DAPM 7/2009.
320 =head2 run metaconfig
322 If you need to make changes to Configure or config_h.SH, it may be best to
323 change the appropriate metaconfig units instead, and regenerate Configure.
327 will regenerate F<Configure> and F<config_h.SH>. Much more information
328 on obtaining and running metaconfig is in the F<U/README> file
329 that comes with Perl's metaconfig units.
331 Since metaconfig is hard to change, running correction scripts after
332 this generation is sometimes needed. Configure gained complexity over
333 time, and the order in which config_h.SH is generated can cause havoc
334 when compiling perl. Therefor, you need to run Porting/config_h.pl
335 after that generation. All that and more is described in the README
336 files that come with the metaunits.
338 Perl's metaconfig units should be available on CPAN. A set of units
339 that will work with perl5.9.x is in a file with a name similar to
340 F<mc_units-20070423.tgz> under L<http://www.cpan.org/authors/id/H/HM/HMBRAND/>.
341 The mc_units tar file should be unpacked in your main perl source directory.
342 Note: those units were for use with 5.9.x. There may have been changes since
343 then. Check for later versions or contact perl5-porters@perl.org to obtain a
344 pointer to the current version.
346 Alternatively, do consider if the F<*ish.h> files or the hint files might be
347 a better place for your changes.
351 If you are using metaconfig to regenerate Configure, then you should note
352 that metaconfig actually uses MANIFEST.new, so you want to be sure
353 MANIFEST.new is up-to-date too. I haven't found the MANIFEST/MANIFEST.new
354 distinction particularly useful, but that's probably because I still haven't
355 learned how to use the full suite of tools in the dist distribution.
360 This will build a config.sh and config.h. You can skip this if you haven't
361 changed Configure or config_h.SH at all. I use the following command
363 sh Configure -Dprefix=/opt/perl -Doptimize=-O -Dusethreads \
365 -Dcf_email='yourname@yourhost.yourplace.com' \
366 -Dperladmin='yourname@yourhost.yourplace.com' \
367 -Dmydomain='.yourplace.com' \
368 -Dmyhostname='yourhost' \
371 =head2 Update Porting/config.sh and Porting/config_H
374 This section needs revision. We're currently working on easing
375 the task of keeping the vms, win32, and plan9 config.sh info
376 up-to-date. The plan is to use keep up-to-date 'canned' config.sh
377 files in the appropriate subdirectories and then generate 'canned'
378 config.h files for vms, win32, etc. from the generic config.sh file.
379 This is to ease maintenance. When Configure gets updated, the parts
380 sometimes get scrambled around, and the changes in config_H can
381 sometimes be very hard to follow. config.sh, on the other hand, can
382 safely be sorted, so it's easy to track (typically very small) changes
383 to config.sh and then propagate them to a canned 'config.h' by any
384 number of means, including a perl script in win32/ or carrying
385 F<config.sh> and F<config_h.SH> to a Unix system and running sh
386 config_h.SH.) Vms uses F<configure.com> to generate its own F<config.sh>
387 and F<config.h>. If you want to add a new variable to F<config.sh> check
388 with vms folk how to add it to configure.com too.
391 The F<Porting/config.sh> and F<Porting/config_H> files are provided to
392 help those folks who can't run Configure. It is important to keep
393 them up-to-date. If you have changed F<config_h.SH>, those changes must
394 be reflected in config_H as well. (The name config_H was chosen to
395 distinguish the file from config.h even on case-insensitive file systems.)
396 Simply edit the existing config_H file; keep the first few explanatory
397 lines and then copy your new config.h below.
399 It may also be necessary to update win32/config.?c, and
400 F<plan9/config.plan9>, though you should be quite careful in doing so if
401 you are not familiar with those systems. You might want to issue your
402 patch with a promise to quickly issue a follow-up that handles those
405 =head2 make regen_perly
407 If F<perly.y> has been edited, it is necessary to run this target to rebuild
408 F<perly.h>, F<perly.act> and F<perly.tab>. In fact this target just runs the Perl
409 script F<regen_perly.pl>. Note that F<perly.c> is I<not> rebuilt; this is just a
410 plain static file now.
412 This target relies on you having Bison installed on your system. Running
413 the target will tell you if you haven't got the right version, and if so,
414 where to get the right one. Or if you prefer, you could hack
415 F<regen_perly.pl> to work with your version of Bison. The important things
416 are that the regexes can still extract out the right chunks of the Bison
417 output into F<perly.act> and F<perly.tab>, and that the contents of those two
418 files, plus F<perly.h>, are functionally equivalent to those produced by the
419 supported version of Bison.
421 Note that in the old days, you had to do C<make run_byacc> instead.
423 =head2 make regen_all
425 This target takes care of the regen_headers target.
426 (It used to also call the regen_pods target, but that has been eliminated.)
428 =head2 make regen_headers
430 The F<embed.h>, F<keywords.h>, and F<opcode.h> files are all automatically
431 generated by perl scripts. Since the user isn't guaranteed to have a
432 working perl, we can't require the user to generate them. Hence you have
433 to, if you're making a distribution.
435 I used to include rules like the following in the makefile:
437 # The following three header files are generated automatically
438 # The correct versions should be already supplied with the perl kit,
439 # in case you don't have perl or 'sh' available.
440 # The - is to ignore error return codes in case you have the source
441 # installed read-only or you don't have perl yet.
442 keywords.h: keywords.pl
443 @echo "Don't worry if this fails."
447 However, I got B<lots> of mail consisting of people worrying because the
448 command failed. I eventually decided that I would save myself time
449 and effort by manually running C<make regen_headers> myself rather
450 than answering all the questions and complaints about the failing
453 =head2 globvar.sym, and perlio.sym
455 Make sure these files are up-to-date. Read the comments in these
456 files and in F<perl_exp.SH> to see what to do.
458 =head2 Binary compatibility
460 If you do change F<embed.fnc> think carefully about
461 what you are doing. To the extent reasonable, we'd like to maintain
462 source and binary compatibility with older releases of perl. That way,
463 extensions built under one version of perl will continue to work with
464 new versions of perl.
466 Of course, some incompatible changes may well be necessary. I'm just
467 suggesting that we not make any such changes without thinking carefully
468 about them first. If possible, we should provide
469 backwards-compatibility stubs. There's a lot of XS code out there.
470 Let's not force people to keep changing it.
474 F<cpan/Devel-PPPort/PPPort.pm> needs to be synchronized to include all
475 new macros added to .h files (normally F<perl.h> and F<XSUB.h>, but others
476 as well). Since chances are that when a new macro is added the
477 committer will forget to update F<PPPort.pm>, it's the best to diff for
478 changes in .h files when making a new release and making sure that
479 F<PPPort.pm> contains them all.
481 The pumpking can delegate the synchronization responsibility to anybody
482 else, but the release process is the only place where we can make sure
483 that no new macros fell through the cracks.
488 The F<pod/perltodo.pod> file contains a roughly-categorized unordered
489 list of aspects of Perl that could use enhancement, features that could
490 be added, areas that could be cleaned up, and so on. During your term
491 as pumpkin-holder, you will probably address some of these issues, and
492 perhaps identify others which, while you decide not to address them this
493 time around, may be tackled in the future. Update the file to reflect
494 the situation as it stands when you hand over the pumpkin.
496 You might like, early in your pumpkin-holding career, to see if you
497 can find champions for particular issues on the to-do list: an issue
498 owned is an issue more likely to be resolved.
500 There are also some more porting-specific L</Todo> items later in this
503 =head2 OS/2-specific updates
505 In the os2 directory is F<diff.configure>, a set of OS/2-specific
506 diffs against B<Configure>. If you make changes to Configure, you may
507 want to consider regenerating this diff file to save trouble for the
510 You can also consider the OS/2 diffs as reminders of portability
511 things that need to be fixed in Configure.
513 =head2 VMS-specific updates
515 The Perl revision number appears as "perl5" in F<configure.com>.
516 It is courteous to update that if necessary.
519 =head2 Making a new patch
521 I find the F<makepatch> utility quite handy for making patches.
522 You can obtain it from any CPAN archive under
523 L<http://www.cpan.org/authors/Johan_Vromans/>. There are a couple
524 of differences between my version and the standard one. I have mine do
527 # Print a reassuring "End of Patch" note so people won't
528 # wonder if their mailer truncated patches.
529 print "\n\nEnd of Patch.\n";
531 at the end. That's because I used to get questions from people asking
532 if their mail was truncated.
534 It also writes Index: lines which include the new directory prefix
535 (change Index: print, approx line 294 or 310 depending on the version,
536 to read: print PATCH ("Index: $newdir$new\n");). That helps patches
537 work with more POSIX conformant patch programs.
539 Here's how I generate a new patch. I'll use the hypothetical
540 5.004_07 to 5.004_08 patch as an example.
542 # unpack perl5.004_07/
543 gzip -d -c perl5.004_07.tar.gz | tar -xof -
544 # unpack perl5.004_08/
545 gzip -d -c perl5.004_08.tar.gz | tar -xof -
546 makepatch perl5.004_07 perl5.004_08 > perl5.004_08.pat
548 Makepatch will automatically generate appropriate B<rm> commands to remove
549 deleted files. Unfortunately, it will not correctly set permissions
550 for newly created files, so you may have to do so manually. For example,
551 patch 5.003_04 created a new test F<t/op/gv.t> which needs to be executable,
552 so at the top of the patch, I inserted the following lines:
558 Now, of course, my patch is now wrong because makepatch didn't know I
559 was going to do that command, and it patched against /dev/null.
561 So, what I do is sort out all such shell commands that need to be in the
562 patch (including possible mv-ing of files, if needed) and put that in the
563 shell commands at the top of the patch. Next, I delete all the patch parts
564 of perl5.004_08.pat, leaving just the shell commands. Then, I do the
568 sh ../perl5.004_08.pat
570 makepatch perl5.004_07 perl5.004_08 >> perl5.004_08.pat
572 (Note the append to preserve my shell commands.)
573 Now, my patch will line up with what the end users are going to do.
575 =head2 Testing your patch
577 It seems obvious, but be sure to test your patch. That is, verify that
578 it produces exactly the same thing as your full distribution.
581 gzip -d -c perl5.004_07.tar.gz | tar -xf -
583 sh ../perl5.004_08.pat
584 patch -p1 -N < ../perl5.004_08.pat
586 gdiff -r perl5.004_07 perl5.004_08
588 where B<gdiff> is GNU diff. Other diff's may also do recursive checking.
592 Again, it's obvious, but you should test your new version as widely as you
593 can. You can be sure you'll hear about it quickly if your version doesn't
594 work on both ANSI and pre-ANSI compilers, and on common systems such as
595 SunOS 4.1.[34], Solaris, and Linux.
597 If your changes include conditional code, try to test the different
598 branches as thoroughly as you can. For example, if your system
599 supports dynamic loading, you can also test static loading with
603 You can also hand-tweak your config.h to try out different #ifdef
610 =item gcc -ansi -pedantic
612 Configure -Dgccansipedantic [ -Dcc=gcc ] will enable (via the cflags script,
613 not $Config{ccflags}) the gcc strict ANSI C flags -ansi and -pedantic for
614 the compilation of the core files on platforms where it knows it can
615 do so (like Linux, see cflags.SH for the full list), and on some
616 platforms only one (Solaris can do only -pedantic, not -ansi).
617 The flag -DPERL_GCC_PEDANTIC also gets added, since gcc does not add
618 any internal cpp flag to signify that -pedantic is being used, as it
619 does for -ansi (__STRICT_ANSI__).
621 Note that the -ansi and -pedantic are enabled only for version 3 (and
622 later) of gcc, since even gcc version 2.95.4 finds lots of seemingly
623 false "value computed not used" errors from Perl.
625 The -ansi and -pedantic are useful in catching at least the following
626 nonportable practices:
632 gcc-specific extensions
648 The -Dgccansipedantic should be used only when cleaning up the code,
649 not for production builds, since otherwise gcc cannot inline certain
654 =head1 Running Purify
656 Purify is a commercial tool that is helpful in identifying memory
657 overruns, wild pointers, memory leaks and other such badness. Perl
658 must be compiled in a specific way for optimal testing with Purify.
660 Use the following commands to test perl with Purify:
662 sh Configure -des -Doptimize=-g -Uusemymalloc -Dusemultiplicity \
664 setenv PURIFYOPTIONS "-chain-length=25"
667 ln -s ../pureperl perl
668 setenv PERL_DESTRUCT_LEVEL 2
671 Disabling Perl's malloc allows Purify to monitor allocations and leaks
672 more closely; using Perl's malloc will make Purify report most leaks
673 in the "potential" leaks category. Enabling the multiplicity option
674 allows perl to clean up thoroughly when the interpreter shuts down, which
675 reduces the number of bogus leak reports from Purify. The -DPURIFY
676 enables any Purify-specific debugging code in the sources.
678 Purify outputs messages in "Viewer" windows by default. If you don't have
679 a windowing environment or if you simply want the Purify output to
680 unobtrusively go to a log file instead of to the interactive window,
681 use the following options instead:
683 setenv PURIFYOPTIONS "-chain-length=25 -windows=no -log-file=perl.log \
686 The only currently known leaks happen when there are compile-time errors
687 within eval or require. (Fixing these is non-trivial, unfortunately, but
688 they must be fixed eventually.)
690 =head1 Common Gotchas
694 =item Probably Prefer POSIX
696 It's often the case that you'll need to choose whether to do
697 something the BSD-ish way or the POSIX-ish way. It's usually not
698 a big problem when the two systems use different names for similar
699 functions, such as memcmp() and bcmp(). The perl.h header file
700 handles these by appropriate #defines, selecting the POSIX mem*()
701 functions if available, but falling back on the b*() functions, if
704 More serious is the case where some brilliant person decided to
705 use the same function name but give it a different meaning or
706 calling sequence :-). getpgrp() and setpgrp() come to mind.
707 These are a real problem on systems that aim for conformance to
708 one standard (e.g. POSIX), but still try to support the other way
709 of doing things (e.g. BSD). My general advice (still not really
710 implemented in the source) is to do something like the following.
711 Suppose there are two alternative versions, fooPOSIX() and
715 /* use fooPOSIX(); */
718 /* try to emulate fooPOSIX() with fooBSD();
719 perhaps with the following: */
720 # define fooPOSIX fooBSD
722 # /* Uh, oh. We have to supply our own. */
723 # define fooPOSIX Perl_fooPOSIX
727 =item Think positively
729 If you need to add an #ifdef test, it is usually easier to follow if you
730 think positively, e.g.
732 #ifdef HAS_NEATO_FEATURE
733 /* use neato feature */
735 /* use some fallback mechanism */
738 rather than the more impenetrable
740 #ifndef MISSING_NEATO_FEATURE
741 /* Not missing it, so we must have it, so use it */
743 /* Are missing it, so fall back on something else. */
746 Of course for this toy example, there's not much difference. But when
747 the #ifdef's start spanning a couple of screen fulls, and the #else's
748 are marked something like
750 #else /* !MISSING_NEATO_FEATURE */
752 I find it easy to get lost.
754 =item Providing Missing Functions -- Problem
756 Not all systems have all the neat functions you might want or need, so
757 you might decide to be helpful and provide an emulation. This is
758 sound in theory and very kind of you, but please be careful about what
759 you name the function. Let me use the C<pause()> function as an
762 Perl5.003 has the following in F<perl.h>
765 #define pause() sleep((32767<<16)+32767)
768 Configure sets HAS_PAUSE if the system has the pause() function, so
769 this #define only kicks in if the pause() function is missing.
772 Unfortunately, some systems apparently have a prototype for pause()
773 in F<unistd.h>, but don't actually have the function in the library.
774 (Or maybe they do have it in a library we're not using.)
776 Thus, the compiler sees something like
778 extern int pause(void);
780 #define pause() sleep((32767<<16)+32767)
782 and dies with an error message. (Some compilers don't mind this;
783 others apparently do.)
785 To work around this, 5.003_03 and later have the following in perl.h:
787 /* Some unistd.h's give a prototype for pause() even though
788 HAS_PAUSE ends up undefined. This causes the #define
789 below to be rejected by the compiler. Sigh.
794 # define Pause() sleep((32767<<16)+32767)
799 The curious reader may wonder why I didn't do the following in
805 sleep((32767<<16)+32767);
809 That is, since the function is missing, just provide it.
810 Then things would probably be been alright, it would seem.
812 Well, almost. It could be made to work. The problem arises from the
813 conflicting needs of dynamic loading and namespace protection.
815 For dynamic loading to work on AIX (and VMS) we need to provide a list
816 of symbols to be exported. This is done by the script F<perl_exp.SH>,
817 which reads F<embed.fnc>. Thus, the C<pause>
818 symbol would have to be added to F<embed.fnc> So far, so good.
820 On the other hand, one of the goals of Perl5 is to make it easy to
821 either extend or embed perl and link it with other libraries. This
822 means we have to be careful to keep the visible namespace "clean".
823 That is, we don't want perl's global variables to conflict with
824 those in the other application library. Although this work is still
825 in progress, the way it is currently done is via the F<embed.h> file.
826 This file is built from the F<embed.fnc> file,
827 since those files already list the globally visible symbols. If we
828 had added C<pause> to F<embed.fnc>, then F<embed.h> would contain the
831 #define pause Perl_pause
833 and calls to C<pause> in the perl sources would now point to
834 C<Perl_pause>. Now, when B<ld> is run to build the F<perl> executable,
835 it will go looking for C<perl_pause>, which probably won't exist in any
836 of the standard libraries. Thus the build of perl will fail.
838 Those systems where C<HAS_PAUSE> is not defined would be ok, however,
839 since they would get a C<Perl_pause> function in util.c. The rest of
840 the world would be in trouble.
842 And yes, this scenario has happened. On SCO, the function C<chsize>
843 is available. (I think it's in F<-lx>, the Xenix compatibility
844 library.) Since the perl4 days (and possibly before), Perl has
845 included a C<chsize> function that gets called something akin to
848 I32 chsize(fd, length)
854 #define chsize Perl_chsize
856 to F<embed.h>, the compile started failing on SCO systems.
858 The "fix" is to give the function a different name. The one
859 implemented in 5.003_05 isn't optimal, but here's what was done:
862 # ifdef my_chsize /* Probably #defined to Perl_my_chsize in embed.h */
865 # define my_chsize chsize
868 My explanatory comment in patch 5.003_05 said:
870 Undef and then re-define my_chsize from Perl_my_chsize to
871 just plain chsize if this system HAS_CHSIZE. This probably only
872 applies to SCO. This shows the perils of having internal
873 functions with the same name as external library functions :-).
875 Now, we can safely put C<my_chsize> in C<embed.fnc>, export it, and
876 hide it with F<embed.h>.
878 To be consistent with what I did for C<pause>, I probably should have
879 called the new function C<Chsize>, rather than C<my_chsize>.
880 However, the perl sources are quite inconsistent on this (Consider
881 New, Mymalloc, and Myremalloc, to name just a few.)
883 There is a problem with this fix, however, in that C<Perl_chsize>
884 was available as a F<libperl.a> library function in 5.003, but it
885 isn't available any more (as of 5.003_07). This means that we've
886 broken binary compatibility. This is not good.
888 =item Providing missing functions -- some ideas
890 We currently don't have a standard way of handling such missing
891 function names. Right now, I'm effectively thinking aloud about a
892 solution. Some day, I'll try to formally propose a solution.
894 Part of the problem is that we want to have some functions listed as
895 exported but not have their names mangled by embed.h or possibly
896 conflict with names in standard system headers. We actually already
897 have such a list at the end of F<perl_exp.SH> (though that list is
900 # extra globals not included above.
901 cat <<END >> perl.exp
925 This still needs much thought, but I'm inclined to think that one
926 possible solution is to prefix all such functions with C<perl_> in the
927 source and list them along with the other C<perl_*> functions in
930 Thus, for C<chsize>, we'd do something like the following:
934 # define perl_chsize chsize
937 then in some file (e.g. F<util.c> or F<doio.c>) do
940 I32 perl_chsize(fd, length)
941 /* implement the function here . . . */
944 Alternatively, we could just always use C<chsize> everywhere and move
945 C<chsize> from F<embed.fnc> to the end of F<perl_exp.SH>. That would
946 probably be fine as long as our C<chsize> function agreed with all the
947 C<chsize> function prototypes in the various systems we'll be using.
948 As long as the prototypes in actual use don't vary that much, this is
949 probably a good alternative. (As a counter-example, note how Configure
950 and perl have to go through hoops to find and use get Malloc_t and
951 Free_t for C<malloc> and C<free>.)
953 At the moment, this latter option is what I tend to prefer.
955 =item All the world's a VAX
957 Sorry, showing my age:-). Still, all the world is not BSD 4.[34],
958 SVR4, or POSIX. Be aware that SVR3-derived systems are still quite
959 common (do you have any idea how many systems run SCO?) If you don't
960 have a bunch of v7 manuals handy, the metaconfig units (by default
961 installed in F</usr/local/lib/dist/U>) are a good resource to look at
966 =head1 Miscellaneous Topics
970 Why does perl use a metaconfig-generated Configure script instead of an
971 autoconf-generated configure script?
973 Metaconfig and autoconf are two tools with very similar purposes.
974 Metaconfig is actually the older of the two, and was originally written
975 by Larry Wall, while autoconf is probably now used in a wider variety of
976 packages. The autoconf info file discusses the history of autoconf and
977 how it came to be. The curious reader is referred there for further
980 Overall, both tools are quite good, I think, and the choice of which one
981 to use could be argued either way. In March, 1994, when I was just
982 starting to work on Configure support for Perl5, I considered both
983 autoconf and metaconfig, and eventually decided to use metaconfig for the
988 =item Compatibility with Perl4
990 Perl4 used metaconfig, so many of the #ifdef's were already set up for
991 metaconfig. Of course metaconfig had evolved some since Perl4's days,
992 but not so much that it posed any serious problems.
994 =item Metaconfig worked for me
996 My system at the time was Interactive 2.2, an SVR3.2/386 derivative that
997 also had some POSIX support. Metaconfig-generated Configure scripts
998 worked fine for me on that system. On the other hand, autoconf-generated
999 scripts usually didn't. (They did come quite close, though, in some
1000 cases.) At the time, I actually fetched a large number of GNU packages
1001 and checked. Not a single one configured and compiled correctly
1002 out-of-the-box with the system's cc compiler.
1004 =item Configure can be interactive
1006 With both autoconf and metaconfig, if the script works, everything is
1007 fine. However, one of my main problems with autoconf-generated scripts
1008 was that if it guessed wrong about something, it could be B<very> hard to
1009 go back and fix it. For example, autoconf always insisted on passing the
1010 -Xp flag to cc (to turn on POSIX behavior), even when that wasn't what I
1011 wanted or needed for that package. There was no way short of editing the
1012 configure script to turn this off. You couldn't just edit the resulting
1013 Makefile at the end because the -Xp flag influenced a number of other
1016 Metaconfig's Configure scripts, on the other hand, can be interactive.
1017 Thus if Configure is guessing things incorrectly, you can go back and fix
1018 them. This isn't as important now as it was when we were actively
1019 developing Configure support for new features such as dynamic loading,
1020 but it's still useful occasionally.
1024 At the time, autoconf-generated scripts were covered under the GNU Public
1025 License, and hence weren't suitable for inclusion with Perl, which has a
1026 different licensing policy. (Autoconf's licensing has since changed.)
1030 Metaconfig builds up Configure from a collection of discrete pieces
1031 called "units". You can override the standard behavior by supplying your
1032 own unit. With autoconf, you have to patch the standard files instead.
1033 I find the metaconfig "unit" method easier to work with. Others
1034 may find metaconfig's units clumsy to work with.
1038 =head2 Why isn't there a directory to override Perl's library?
1040 Mainly because no one's gotten around to making one. Note that
1041 "making one" involves changing perl.c, Configure, config_h.SH (and
1042 associated files, see above), and I<documenting> it all in the
1045 Apparently, most folks who want to override one of the standard library
1046 files simply do it by overwriting the standard library files.
1050 In the perl.c sources, you'll find an undocumented APPLLIB_EXP
1051 variable, sort of like PRIVLIB_EXP and ARCHLIB_EXP (which are
1052 documented in config_h.SH). Here's what APPLLIB_EXP is for, from
1053 a mail message from Larry:
1055 The main intent of APPLLIB_EXP is for folks who want to send out a
1056 version of Perl embedded in their product. They would set the symbol
1057 to be the name of the library containing the files needed to run or to
1058 support their particular application. This works at the "override"
1059 level to make sure they get their own versions of any library code that
1060 they absolutely must have configuration control over.
1062 As such, I don't see any conflict with a sysadmin using it for a
1063 override-ish sort of thing, when installing a generic Perl. It should
1064 probably have been named something to do with overriding though. Since
1065 it's undocumented we could still change it... :-)
1067 Given that it's already there, you can use it to override distribution modules.
1068 One way to do that is to add
1070 ccflags="$ccflags -DAPPLLIB_EXP=\"/my/override\""
1072 to your config.over file. (You have to be particularly careful to get the
1073 double quotes in. APPLLIB_EXP must be a valid C string. It might
1074 actually be easier to just #define it yourself in perl.c.)
1076 Then perl.c will put /my/override ahead of ARCHLIB and PRIVLIB. Perl will
1077 also search architecture-specific and version-specific subdirectories of
1080 =head2 Shared libperl.so location
1082 Why isn't the shared libperl.so installed in /usr/lib/ along
1083 with "all the other" shared libraries? Instead, it is installed
1084 in $archlib, which is typically something like
1086 /usr/local/lib/perl5/archname/5.00404
1088 and is architecture- and version-specific.
1090 The basic reason why a shared libperl.so gets put in $archlib is so that
1091 you can have more than one version of perl on the system at the same time,
1092 and have each refer to its own libperl.so.
1094 Three examples might help. All of these work now; none would work if you
1095 put libperl.so in /usr/lib.
1101 Suppose you want to have both threaded and non-threaded perl versions
1102 around. Configure will name both perl libraries "libperl.so" (so that
1103 you can link to them with -lperl). The perl binaries tell them apart
1104 by having looking in the appropriate $archlib directories.
1108 Suppose you have perl5.004_04 installed and you want to try to compile
1109 it again, perhaps with different options or after applying a patch.
1110 If you already have libperl.so installed in /usr/lib/, then it may be
1111 either difficult or impossible to get ld.so to find the new libperl.so
1112 that you're trying to build. If, instead, libperl.so is tucked away in
1113 $archlib, then you can always just change $archlib in the current perl
1114 you're trying to build so that ld.so won't find your old libperl.so.
1115 (The INSTALL file suggests you do this when building a debugging perl.)
1119 The shared perl library is not a "well-behaved" shared library with
1120 proper major and minor version numbers, so you can't necessarily
1121 have perl5.004_04 and perl5.004_05 installed simultaneously. Suppose
1122 perl5.004_04 were to install /usr/lib/libperl.so.4.4, and perl5.004_05
1123 were to install /usr/lib/libperl.so.4.5. Now, when you try to run
1124 perl5.004_04, ld.so might try to load libperl.so.4.5, since it has
1125 the right "major version" number. If this works at all, it almost
1126 certainly defeats the reason for keeping perl5.004_04 around. Worse,
1127 with development subversions, you certainly can't guarantee that
1128 libperl.so.4.4 and libperl.so.4.55 will be compatible.
1130 Anyway, all this leads to quite obscure failures that are sure to drive
1131 casual users crazy. Even experienced users will get confused :-). Upon
1132 reflection, I'd say leave libperl.so in $archlib.
1136 =head2 Indentation style
1138 Over the years Perl has become a mishmash of
1139 various indentation styles, but the original "Larry style" can
1140 probably be restored with (GNU) indent somewhat like this:
1142 indent -kr -nce -psl -sc
1144 A more ambitious solution would also specify a list of Perl specific
1145 types with -TSV -TAV -THV .. -TMAGIC -TPerlIO ... but that list would
1146 be quite ungainly. Also note that GNU indent also doesn't do aligning
1147 of consecutive assignments, which would truly wreck the layout in
1148 places like sv.c:Perl_sv_upgrade() or sv.c:Perl_clone_using().
1149 Similarly nicely aligned &&s, ||s and ==s would not be respected.
1151 =head1 Upload Your Work to CPAN
1153 You can upload your work to CPAN if you have a CPAN id. Check out
1154 L<http://www.cpan.org/modules/04pause.html> for information on
1155 _PAUSE_, the Perl Author's Upload Server.
1157 I typically upload both the patch file, e.g. F<perl5.004_08.pat.gz>
1158 and the full tar file, e.g. F<perl5.004_08.tar.gz>.
1160 If you want your patch to appear in the F<src/5.0/unsupported>
1161 directory on CPAN, send e-mail to the CPAN master librarian. (Check
1162 out http://www.cpan.org/CPAN.html ).
1164 =head1 Help Save the World
1166 You should definitely announce your patch on the perl5-porters list.
1167 You should also consider announcing your patch on
1168 comp.lang.perl.announce, though you should make it quite clear that a
1169 subversion is not a production release, and be prepared to deal with
1170 people who will not read your disclaimer.
1174 Here, in no particular order, are some Configure and build-related
1175 items that merit consideration. This list isn't exhaustive, it's just
1176 what I came up with off the top of my head.
1178 =head2 Adding missing library functions to Perl
1180 The perl Configure script automatically determines which headers and
1181 functions you have available on your system and arranges for them to be
1182 included in the compilation and linking process. Occasionally, when porting
1183 perl to an operating system for the first time, you may find that the
1184 operating system is missing a key function. While perl may still build
1185 without this function, no perl program will be able to reference the missing
1186 function. You may be able to write the missing function yourself, or you
1187 may be able to find the missing function in the distribution files for
1188 another software package. In this case, you need to instruct the perl
1189 configure-and-build process to use your function. Perform these steps.
1195 Code and test the function you wish to add. Test it carefully; you will
1196 have a much easier time debugging your code independently than when it is a
1201 Here is an implementation of the POSIX truncate function for an operating
1202 system (VOS) that does not supply one, but which does supply the ftruncate()
1205 /* Beginning of modification history */
1206 /* Written 02-01-02 by Nick Ing-Simmons (nick@ing-simmons.net) */
1207 /* End of modification history */
1209 /* VOS doesn't supply a truncate function, so we build one up
1210 from the available POSIX functions. */
1213 #include <sys/types.h>
1217 truncate(const char *path, off_t len)
1219 int fd = open(path,O_WRONLY);
1222 code = ftruncate(fd,len);
1228 Place this file into a subdirectory that has the same name as the operating
1229 system. This file is named perl/vos/vos.c
1233 If your operating system has a hints file (in perl/hints/XXX.sh for an
1234 operating system named XXX), then start with it. If your operating system
1235 has no hints file, then create one. You can use a hints file for a similar
1236 operating system, if one exists, as a template.
1240 Add lines like the following to your hints file. The first line
1241 (d_truncate="define") instructs Configure that the truncate() function
1242 exists. The second line (archobjs="vos.o") instructs the makefiles that the
1243 perl executable depends on the existence of a file named "vos.o". (Make
1244 will automatically look for "vos.c" and compile it with the same options as
1245 the perl source code). The final line ("test -h...") adds a symbolic link
1246 to the top-level directory so that make can find vos.c. Of course, you
1247 should use your own operating system name for the source file of extensions,
1250 # VOS does not have truncate() but we supply one in vos.c
1254 # Help gmake find vos.c
1255 test -h vos.c || ln -s vos/vos.c vos.c
1257 The hints file is a series of shell commands that are run in the top-level
1258 directory (the "perl" directory). Thus, these commands are simply executed
1259 by Configure at an appropriate place during its execution.
1263 At this point, you can run the Configure script and rebuild perl. Carefully
1264 test the newly-built perl to ensure that normal paths, and error paths,
1265 behave as you expect.
1269 =head2 Good ideas waiting for round tuits
1273 =item Configure -Dsrc=/blah/blah
1275 We should be able to emulate B<configure --srcdir>. Tom Tromey
1276 tromey@creche.cygnus.com has submitted some patches to
1277 the dist-users mailing list along these lines. They have been folded
1278 back into the main distribution, but various parts of the perl
1279 Configure/build/install process still assume src='.'.
1281 =item Hint file fixes
1283 Various hint files work around Configure problems. We ought to fix
1284 Configure so that most of them aren't needed.
1286 =item Hint file information
1288 Some of the hint file information (particularly dynamic loading stuff)
1289 ought to be fed back into the main metaconfig distribution.
1293 =head2 Probably good ideas waiting for round tuits
1297 =item GNU configure --options
1299 I've received sensible suggestions for --exec_prefix and other
1300 GNU configure --options. It's not always obvious exactly what is
1301 intended, but this merits investigation.
1305 Currently, B<make clean> isn't all that useful, though
1306 B<make realclean> and B<make distclean> are. This needs a bit of
1307 thought and documentation before it gets cleaned up.
1309 =item Try gcc if cc fails
1311 Currently, we just give up.
1313 =item bypassing safe*alloc wrappers
1315 On some systems, it may be safe to call the system malloc directly
1316 without going through the util.c safe* layers. (Such systems would
1317 accept free(0), for example.) This might be a time-saver for systems
1318 that already have a good malloc. (Recent Linux libc's apparently have
1319 a nice malloc that is well-tuned for the system.)
1323 =head2 Vague possibilities
1327 =item gconvert replacement
1329 Maybe include a replacement function that doesn't lose data in rare
1330 cases of coercion between string and numerical values.
1332 =item Improve makedepend
1334 The current makedepend process is clunky and annoyingly slow, but it
1335 works for most folks. Alas, it assumes that there is a filename
1336 $firstmakefile that the B<make> command will try to use before it uses
1337 F<Makefile>. Such may not be the case for all B<make> commands,
1338 particularly those on non-Unix systems.
1340 Probably some variant of the BSD F<.depend> file will be useful.
1341 We ought to check how other packages do this, if they do it at all.
1342 We could probably pre-generate the dependencies (with the exception of
1343 malloc.o, which could probably be determined at F<Makefile.SH>
1346 =item GNU Makefile standard targets
1348 GNU software generally has standardized Makefile targets. Unless we
1349 have good reason to do otherwise, I see no reason not to support them.
1353 Somehow, straighten out, document, and implement lockf(), flock(),
1354 and/or fcntl() file locking. It's a mess. See $d_fcntl_can_lock
1355 in recent config.sh files though.
1359 =head2 Copyright Issues
1361 The following is based on the consensus of a couple of IPR lawyers,
1362 but it is of course not a legally binding statement, just a common
1369 Tacking on copyright statements is unnecessary to begin with because
1370 of the Berne convention. But assuming you want to go ahead...
1374 The right form of a copyright statement is
1376 Copyright (C) Year, Year, ... by Someone
1378 The (C) is not required everywhere but it doesn't hurt and in certain
1379 jurisdictions it is required, so let's leave it in. (Yes, it's true
1380 that in some jurisdictions the "(C)" is not legally binding, one should
1381 use the true ringed-C. But we don't have that character available for
1382 Perl's source code.)
1384 The years must be listed out separately. Year-Year is not correct.
1385 Only the years when the piece has changed 'significantly' may be added.
1389 One cannot give away one's copyright trivially. One can give one's
1390 copyright away by using public domain, but even that requires a little
1391 bit more than just saying 'this is in public domain'. (What it
1392 exactly requires depends on your jurisdiction.) But barring public
1393 domain, one cannot "transfer" one's copyright to another person or
1394 entity. In the context of software, it means that contributors cannot
1395 give away their copyright or "transfer" it to the "owner" of the software.
1397 Also remember that in many cases if you are employed by someone,
1398 your work may be copyrighted to your employer, even when you are
1399 contributing on your own time (this all depends on too many things
1400 to list here). But the bottom line is that you definitely can't give
1401 away a copyright you may not even have.
1403 What is possible, however, is that the software can simply state
1405 Copyright (C) Year, Year, ... by Someone and others
1407 and then list the "others" somewhere in the distribution.
1408 And this is exactly what Perl does. (The "somewhere" is
1409 AUTHORS and the Changes* files.)
1413 Split files, merged files, and generated files are problematic.
1414 The rule of thumb: in split files, copy the copyright years of
1415 the original file to all the new files; in merged files make
1416 an union of the copyright years of all the old files; in generated
1417 files propagate the copyright years of the generating file(s).
1421 The files of Perl source code distribution do carry a lot of
1422 copyrights, by various people. (There are many copyrights embedded in
1423 perl.c, for example.) The most straightforward thing for pumpkings to
1424 do is to simply update Larry's copyrights at the beginning of the
1425 *.[hcy], x2p/*.[hcy], *.pl, and README files, and leave all other
1426 copyrights alone. Doing more than that requires quite a bit of tracking.
1432 Original author: Andy Dougherty doughera@lafayette.edu .
1433 Additions by Chip Salzenberg chip@perl.com and
1434 Tim Bunce Tim.Bunce@ig.co.uk .
1436 All opinions expressed herein are those of the authorZ<>(s).
1438 =head1 LAST MODIFIED
1440 2009-07-08-01 Jesse Vincent