3 Pumpkin - Notes on handling the Perl Patch Pumpkin
7 There is no simple synopsis, yet.
11 This document attempts to begin to describe some of the
12 considerations involved in patching 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.perl.com/CPAN/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.rosat.mpe-garching.mpg.de/mailing-lists/perl-porters/
48 =head1 How are Perl Releases Numbered?
50 Perl version numbers are floating point numbers, such as 5.004.
51 (Observations about the imprecision of floating point numbers for
52 representing reality probably have more relevance than you might
53 imagine :-) The major version number is 5 and the '004' is the
54 patchlevel. (Questions such as whether or not '004' is really a minor
55 version number can safely be ignored.:)
57 The version number is available as the magic variable $],
58 and can be used in comparisons, e.g.
60 print "You've got an old perl\n" if $] < 5.002;
62 You can also require particular version (or later) with
66 At some point in the future, we may need to decide what to call the
67 next big revision. In the .package file used by metaconfig to
68 generate Configure, there are two variables that might be relevant:
69 $baserev=5.0 and $package=perl5. At various times, I have suggested
70 we might change them to $baserev=5.1 and $package=perl5.1 if want
71 to signify a fairly major update. Or, we might want to jump to perl6.
72 Let's worry about that problem when we get there.
76 In addition, there may be "developer" sub-versions available. These
77 are not official releases. They may contain unstable experimental
78 features, and are subject to rapid change. Such developer
79 sub-versions are numbered with sub-version numbers. For example,
80 version 5.003_04 is the 4'th developer version built on top of
81 5.003. It might include the _01, _02, and _03 changes, but it
82 also might not. Sub-versions are allowed to be subversive. (But see
83 the next section for recent changes.)
85 These sub-versions can also be used as floating point numbers, so
86 you can do things such as
88 print "You've got an unstable perl\n" if $] == 5.00303;
90 You can also require particular version (or later) with
92 use 5.003_03; # the "_" is optional
94 Sub-versions produced by the members of perl5-porters are usually
95 available on CPAN in the F<src/5.0/unsupported> directory.
97 =head2 Maintenance and Development Subversions
99 As an experiment, starting with version 5.004, subversions _01 through
100 _49 will be reserved for bug-fix maintenance releases, and subversions
101 _50 through _99 will be available for unstable development versions.
103 The separate bug-fix track is being established to allow us an easy
104 way to distribute important bug fixes without waiting for the
105 developers to untangle all the other problems in the current
108 Trial releases of bug-fix maintenance releases are announced on
109 perl5-porters. Trial releases use the new subversion number (to avoid
110 testers installing it over the previous release) and include a 'local
111 patch' entry in patchlevel.h.
113 Watch for announcements of maintenance subversions in
114 comp.lang.perl.announce.
116 =head2 Why such a complicated scheme?
118 Two reasons, really. At least.
120 First, we need some way to identify and release collections of patches
121 that are known to have new features that need testing and exploration. The
122 subversion scheme does that nicely while fitting into the
125 Second, since most of the folks who help maintain perl do so on a
126 free-time voluntary basis, perl development does not proceed at a
127 precise pace, though it always seems to be moving ahead quickly.
128 We needed some way to pass around the "patch pumpkin" to allow
129 different people chances to work on different aspects of the
130 distribution without getting in each other's way. It wouldn't be
131 constructive to have multiple people working on incompatible
132 implementations of the same idea. Instead what was needed was
133 some kind of "baton" or "token" to pass around so everyone knew
136 =head2 Why is it called the patch pumpkin?
138 Chip Salzenberg gets credit for that, with a nod to his cow orker,
139 David Croy. We had passed around various names (baton, token, hot
140 potato) but none caught on. Then, Chip asked:
144 Who has the patch pumpkin?
146 To explain: David Croy once told me once that at a previous job,
147 there was one tape drive and multiple systems that used it for backups.
148 But instead of some high-tech exclusion software, they used a low-tech
149 method to prevent multiple simultaneous backups: a stuffed pumpkin.
150 No one was allowed to make backups unless they had the "backup pumpkin".
156 =head1 Philosophical Issues in Patching Perl
158 There are no absolute rules, but there are some general guidelines I
159 have tried to follow as I apply patches to the perl sources.
160 (This section is still under construction.)
162 =head2 Solve problems as generally as possible
164 Never implement a specific restricted solution to a problem when you
165 can solve the same problem in a more general, flexible way.
167 For example, for dynamic loading to work on some SVR4 systems, we had
168 to build a shared libperl.so library. In order to build "FAT" binaries
169 on NeXT 4.0 systems, we had to build a special libperl library. Rather
170 than continuing to build a contorted nest of special cases, I
171 generalized the process of building libperl so that NeXT and SVR4 users
172 could still get their work done, but others could build a shared
173 libperl if they wanted to as well.
175 =head2 Seek consensus on major changes
177 If you are making big changes, don't do it in secret. Discuss the
178 ideas in advance on perl5-porters.
180 =head2 Keep the documentation up-to-date
182 If your changes may affect how users use perl, then check to be sure
183 that the documentation is in sync with your changes. Be sure to
184 check all the files F<pod/*.pod> and also the F<INSTALL> document.
186 Consider writing the appropriate documentation first and then
187 implementing your change to correspond to the documentation.
189 =head2 Avoid machine-specific #ifdef's
191 To the extent reasonable, try to avoid machine-specific #ifdef's in
192 the sources. Instead, use feature-specific #ifdef's. The reason is
193 that the machine-specific #ifdef's may not be valid across major
194 releases of the operating system. Further, the feature-specific tests
195 may help out folks on another platform who have the same problem.
197 =head2 Allow for lots of testing
199 We should never release a main version without testing it as a
202 =head2 Test popular applications and modules.
204 We should never release a main version without testing whether or not
205 it breaks various popular modules and applications. A partial list of
206 such things would include majordomo, metaconfig, apache, Tk, CGI,
207 libnet, and libwww, to name just a few. Of course it's quite possible
208 that some of those things will be just plain broken and need to be fixed,
209 but, in general, we ought to try to avoid breaking widely-installed
212 =head2 Automate generation of derivative files
214 The F<embed.h>, F<keywords.h>, F<opcode.h>, and F<perltoc.pod> files
215 are all automatically generated by perl scripts. In general, don't
216 patch these directly; patch the data files instead.
218 F<Configure> and F<config_h.SH> are also automatically generated by
219 B<metaconfig>. In general, you should patch the metaconfig units
220 instead of patching these files directly. However, very minor changes to
221 F<Configure> may be made in between major sync-ups with the metaconfig
222 units, which tends to be complicated operations. But be careful, this
223 can quickly spiral out of control. Running metaconfig is not really
226 =head1 How to Make a Distribution
228 There really ought to be a 'make dist' target, but there isn't.
229 The 'dist' suite of tools also contains a number of tools that I haven't
230 learned how to use yet. Some of them may make this all a bit easier.
232 Here are the steps I go through to prepare a patch & distribution.
234 Lots of it could doubtless be automated but isn't. The Porting/makerel
235 (make release) perl script does now help automate some parts of it.
237 =head2 Announce your intentions
239 First, you should volunteer out loud to take the patch pumpkin. It's
240 generally counter-productive to have multiple people working in secret
243 At the same time, announce what you plan to do with the patch pumpkin,
244 to allow folks a chance to object or suggest alternatives, or do it for
245 you. Naturally, the patch pumpkin holder ought to incorporate various
246 bug fixes and documentation improvements that are posted while he or
247 she has the pumpkin, but there might also be larger issues at stake.
249 One of the precepts of the subversion idea is that we shouldn't give
250 the patch pumpkin to anyone unless we have some idea what he or she
251 is going to do with it.
253 =head2 refresh pod/perltoc.pod
255 Presumably, you have done a full C<make> in your working source
256 directory. Before you C<make spotless> (if you do), and if you have
257 changed any documentation in any module or pod file, change to the
258 F<pod> directory and run C<make toc>.
260 =head2 run installhtml to check the validity of the pod files
262 =head2 update patchlevel.h
264 Don't be shy about using the subversion number, even for a relatively
265 modest patch. We've never even come close to using all 99 subversions,
266 and it's better to have a distinctive number for your patch. If you
267 need feedback on your patch, go ahead and issue it and promise to
268 incorporate that feedback quickly (e.g. within 1 week) and send out a
271 =head2 run metaconfig
273 If you need to make changes to Configure or config_h.SH, it may be best to
274 change the appropriate metaconfig units instead, and regenerate Configure.
278 will regenerate Configure and config_h.SH. More information on
279 obtaining and running metaconfig is in the F<U/README> file that comes
280 with Perl's metaconfig units. Perl's metaconfig units should be
281 available the same place you found this file. On CPAN, look under my
282 directory F<authors/id/ANDYD/> for a file such as F<5.003_07-02.U.tar.gz>.
283 That file should be unpacked in your main perl source directory. It
284 contains the files needed to run B<metaconfig> to reproduce Perl's
285 Configure script. (Those units are for 5.003_07. There have been
286 changes since then; please contact me if you want more recent
287 versions, and I will try to point you in the right direction.)
289 Alternatively, do consider if the F<*ish.h> files might be a better
290 place for your changes.
294 Make sure the MANIFEST is up-to-date. You can use dist's B<manicheck>
295 program for this. You can also use
297 perl -w -MExtUtils::Manifest=fullcheck -e fullcheck
299 Both commands will also list extra files in the directory that are not
302 The MANIFEST is normally sorted, with one exception. Perl includes
303 both a F<Configure> script and a F<configure> script. The
304 F<configure> script is a front-end to the main F<Configure>, but
305 is there to aid folks who use autoconf-generated F<configure> files
306 for other software. The problem is that F<Configure> and F<configure>
307 are the same on case-insensitive file systems, so I deliberately put
308 F<configure> first in the MANIFEST so that the extraction of
309 F<Configure> will overwrite F<configure> and leave you with the
310 correct script. (The F<configure> script must also have write
311 permission for this to work, so it's the only file in the distribution
312 I normally have with write permission.)
314 If you are using metaconfig to regenerate Configure, then you should note
315 that metaconfig actually uses MANIFEST.new, so you want to be sure
316 MANIFEST.new is up-to-date too. I haven't found the MANIFEST/MANIFEST.new
317 distinction particularly useful, but that's probably because I still haven't
318 learned how to use the full suite of tools in the dist distribution.
320 =head2 Check permissions
322 All the tests in the t/ directory ought to be executable. The
323 main makefile used to do a 'chmod t/*/*.t', but that resulted in
324 a self-modifying distribution--something some users would strongly
325 prefer to avoid. Probably, the F<t/TEST> script should check for this
326 and do the chmod if needed, but it doesn't currently.
328 In all, the following files should probably be executable:
343 vms/ext/Stdio/test.pl
347 Other things ought to be readable, at least :-).
349 Probably, the permissions for the files could be encoded in MANIFEST
350 somehow, but I'm reluctant to change MANIFEST itself because that
351 could break old scripts that use MANIFEST.
353 I seem to recall that some SVR3 systems kept some sort of file that listed
354 permissions for system files; something like that might be appropriate.
358 This will build a config.sh and config.h. You can skip this if you haven't
359 changed Configure or config_h.SH at all. I use the following command
361 sh Configure -Dprefix=/opt/perl -Doptimize=-O -Dusethreads \
363 -Dcf_email='yourname@yourhost.yourplace.com' \
364 -Dperladmin='yourname@yourhost.yourplace.com' \
365 -Dmydomain='.yourplace.com' \
366 -Dmyhostname='yourhost' \
369 =head2 Update Porting/config.sh and Porting/config_H
372 This section needs revision. We're currently working on easing
373 the task of keeping the vms, win32, and plan9 config.sh info
374 up-to-date. The plan is to use keep up-to-date 'canned' config.sh
375 files in the appropriate subdirectories and then generate 'canned'
376 config.h files for vms, win32, etc. from the generic config.sh file.
377 This is to ease maintenance. When Configure gets updated, the parts
378 sometimes get scrambled around, and the changes in config_H can
379 sometimes be very hard to follow. config.sh, on the other hand, can
380 safely be sorted, so it's easy to track (typically very small) changes
381 to config.sh and then propoagate them to a canned 'config.h' by any
382 number of means, including a perl script in win32/ or carrying
383 config.sh and config_h.SH to a Unix system and running sh
387 The Porting/config.sh and Porting/config_H files are provided to
388 help those folks who can't run Configure. It is important to keep
389 them up-to-date. If you have changed config_h.SH, those changes must
390 be reflected in config_H as well. (The name config_H was chosen to
391 distinguish the file from config.h even on case-insensitive file systems.)
392 Simply edit the existing config_H file; keep the first few explanatory
393 lines and then copy your new config.h below.
395 It may also be necessary to update vms/config.vms and
396 plan9/config.plan9, though you should be quite careful in doing so if
397 you are not familiar with those systems. You might want to issue your
398 patch with a promise to quickly issue a follow-up that handles those
401 =head2 make run_byacc
403 If you have byacc-1.8.2 (available from CPAN), and if there have been
404 changes to F<perly.y>, you can regenerate the F<perly.c> file. The
405 run_byacc makefile target does this by running byacc and then applying
406 some patches so that byacc dynamically allocates space, rather than
407 having fixed limits. This patch is handled by the F<perly.fixer>
408 script. Depending on the nature of the changes to F<perly.y>, you may
409 or may not have to hand-edit the patch to apply correctly. If you do,
410 you should include the edited patch in the new distribution. If you
411 have byacc-1.9, the patch won't apply cleanly. Changes to the printf
412 output statements mean the patch won't apply cleanly. Long ago I
413 started to fix F<perly.fixer> to detect this, but I never completed the
416 Some additional notes from Larry on this:
418 Don't forget to regenerate perly.c.diff.
422 patch perly.c <perly.c.diff
423 # manually apply any failed hunks
424 diff -c2 perly.c.orig perly.c >perly.c.diff
426 One chunk of lines that often fails begins with
430 and ends one line before
432 #define YYERRCODE 256
434 This only happens when you add or remove a token type. I suppose this
435 could be automated, but it doesn't happen very often nowadays.
439 =head2 make regen_headers
441 The F<embed.h>, F<keywords.h>, and F<opcode.h> files are all automatically
442 generated by perl scripts. Since the user isn't guaranteed to have a
443 working perl, we can't require the user to generate them. Hence you have
444 to, if you're making a distribution.
446 I used to include rules like the following in the makefile:
448 # The following three header files are generated automatically
449 # The correct versions should be already supplied with the perl kit,
450 # in case you don't have perl or 'sh' available.
451 # The - is to ignore error return codes in case you have the source
452 # installed read-only or you don't have perl yet.
453 keywords.h: keywords.pl
454 @echo "Don't worry if this fails."
458 However, I got B<lots> of mail consisting of people worrying because the
459 command failed. I eventually decided that I would save myself time
460 and effort by manually running C<make regen_headers> myself rather
461 than answering all the questions and complaints about the failing
464 =head2 global.sym, interp.sym and perlio.sym
466 Make sure these files are up-to-date. Read the comments in these
467 files and in perl_exp.SH to see what to do.
469 =head2 Binary compatibility
471 If you do change F<global.sym> or F<interp.sym>, think carefully about
472 what you are doing. To the extent reasonable, we'd like to maintain
473 souce and binary compatibility with older releases of perl. That way,
474 extensions built under one version of perl will continue to work with
475 new versions of perl.
477 Of course, some incompatible changes may well be necessary. I'm just
478 suggesting that we not make any such changes without thinking carefully
479 about them first. If possible, we should provide
480 backwards-compatibility stubs. There's a lot of XS code out there.
481 Let's not force people to keep changing it.
485 Be sure to update the F<Changes> file. Try to include both an overall
486 summary as well as detailed descriptions of the changes. Your
487 audience will include other developers and users, so describe
488 user-visible changes (if any) in terms they will understand, not in
489 code like "initialize foo variable in bar function".
491 There are differing opinions on whether the detailed descriptions
492 ought to go in the Changes file or whether they ought to be available
493 separately in the patch file (or both). There is no disagreement that
494 detailed descriptions ought to be easily available somewhere.
498 The F<Todo> file contains a roughly-catgorized unordered list of
499 aspects of Perl that could use enhancement, features that could be
500 added, areas that could be cleaned up, and so on. During your term as
501 pumpkin-holder, you will probably address some of these issues, and
502 perhaps identify others which, while you decide not to address them
503 this time around, may be tackled in the future. Update the file
504 reflect the situation as it stands when you hand over the pumpkin.
506 You might like, early in your pumpkin-holding career, to see if you
507 can find champions for partiticular issues on the to-do list: an issue
508 owned is an issue more likely to be resolved.
510 There are also some more porting-specific L<Todo> items later in this
513 =head2 OS/2-specific updates
515 In the os2 directory is F<diff.configure>, a set of OS/2-specific
516 diffs against B<Configure>. If you make changes to Configure, you may
517 want to consider regenerating this diff file to save trouble for the
520 You can also consider the OS/2 diffs as reminders of portability
521 things that need to be fixed in Configure.
523 =head2 VMS-specific updates
525 If you have changed F<perly.y>, then you may want to update
526 F<vms/perly_{h,c}.vms> by running C<perl vms/vms_yfix.pl>.
528 The Perl version number appears in several places under F<vms>.
529 It is courteous to update these versions. For example, if you are
530 making 5.004_42, replace "5.00441" with "5.00442".
532 =head2 Making the new distribution
534 Suppose, for example, that you want to make version 5.004_08. Then you can
535 do something like the following
537 mkdir ../perl5.004_08
538 awk '{print $1}' MANIFEST | cpio -pdm ../perl5.004_08
540 tar cf perl5.004_08.tar perl5.004_08
541 gzip --best perl5.004_08.tar
543 These steps, with extra checks, are automated by the Porting/makerel
546 =head2 Making a new patch
548 I find the F<makepatch> utility quite handy for making patches.
549 You can obtain it from any CPAN archive under
550 http://www.perl.com/CPAN/authors/Johan_Vromans/ . There are a couple
551 of differences between my version and the standard one. I have mine do
554 # Print a reassuring "End of Patch" note so people won't
555 # wonder if their mailer truncated patches.
556 print "\n\nEnd of Patch.\n";
558 at the end. That's because I used to get questions from people asking
559 if their mail was truncated.
561 It also writes Index: lines which include the new directory prefix
562 (change Index: print, approx line 294 or 310 depending on the version,
563 to read: print PATCH ("Index: $newdir$new\n");). That helps patches
564 work with more POSIX conformant patch programs.
566 Here's how I generate a new patch. I'll use the hypothetical
567 5.004_07 to 5.004_08 patch as an example.
569 # unpack perl5.004_07/
570 gzip -d -c perl5.004_07.tar.gz | tar -xof -
571 # unpack perl5.004_08/
572 gzip -d -c perl5.004_08.tar.gz | tar -xof -
573 makepatch perl5.004_07 perl5.004_08 > perl5.004_08.pat
575 Makepatch will automatically generate appropriate B<rm> commands to remove
576 deleted files. Unfortunately, it will not correctly set permissions
577 for newly created files, so you may have to do so manually. For example,
578 patch 5.003_04 created a new test F<t/op/gv.t> which needs to be executable,
579 so at the top of the patch, I inserted the following lines:
585 Now, of course, my patch is now wrong because makepatch didn't know I
586 was going to do that command, and it patched against /dev/null.
588 So, what I do is sort out all such shell commands that need to be in the
589 patch (including possible mv-ing of files, if needed) and put that in the
590 shell commands at the top of the patch. Next, I delete all the patch parts
591 of perl5.004_08.pat, leaving just the shell commands. Then, I do the
595 sh ../perl5.004_08.pat
597 makepatch perl5.004_07 perl5.004_08 >> perl5.004_08.pat
599 (Note the append to preserve my shell commands.)
600 Now, my patch will line up with what the end users are going to do.
602 =head2 Testing your patch
604 It seems obvious, but be sure to test your patch. That is, verify that
605 it produces exactly the same thing as your full distribution.
608 gzip -d -c perl5.004_07.tar.gz | tar -xf -
610 sh ../perl5.004_08.pat
611 patch -p1 -N < ../perl5.004_08.pat
613 gdiff -r perl5.004_07 perl5.004_08
615 where B<gdiff> is GNU diff. Other diff's may also do recursive checking.
619 Again, it's obvious, but you should test your new version as widely as you
620 can. You can be sure you'll hear about it quickly if your version doesn't
621 work on both ANSI and pre-ANSI compilers, and on common systems such as
622 SunOS 4.1.[34], Solaris, and Linux.
624 If your changes include conditional code, try to test the different
625 branches as thoroughly as you can. For example, if your system
626 supports dynamic loading, you can also test static loading with
630 You can also hand-tweak your config.h to try out different #ifdef
633 =head1 Common Gotcha's
639 The '#elif' preprocessor directive is not understood on all systems.
640 Specifically, I know that Pyramids don't understand it. Thus instead of the
651 You have to do the more Byzantine
663 Incidentally, whitespace between the leading '#' and the preprocessor
664 command is not guaranteed, but is very portable and you may use it freely.
665 I think it makes things a bit more readable, especially once things get
666 rather deeply nested. I also think that things should almost never get
667 too deeply nested, so it ought to be a moot point :-)
669 =item Probably Prefer POSIX
671 It's often the case that you'll need to choose whether to do
672 something the BSD-ish way or the POSIX-ish way. It's usually not
673 a big problem when the two systems use different names for similar
674 functions, such as memcmp() and bcmp(). The perl.h header file
675 handles these by appropriate #defines, selecting the POSIX mem*()
676 functions if available, but falling back on the b*() functions, if
679 More serious is the case where some brilliant person decided to
680 use the same function name but give it a different meaning or
681 calling sequence :-). getpgrp() and setpgrp() come to mind.
682 These are a real problem on systems that aim for conformance to
683 one standard (e.g. POSIX), but still try to support the other way
684 of doing things (e.g. BSD). My general advice (still not really
685 implemented in the source) is to do something like the following.
686 Suppose there are two alternative versions, fooPOSIX() and
690 /* use fooPOSIX(); */
693 /* try to emulate fooPOSIX() with fooBSD();
694 perhaps with the following: */
695 # define fooPOSIX fooBSD
697 # /* Uh, oh. We have to supply our own. */
698 # define fooPOSIX Perl_fooPOSIX
702 =item Think positively
704 If you need to add an #ifdef test, it is usually easier to follow if you
705 think positively, e.g.
707 #ifdef HAS_NEATO_FEATURE
708 /* use neato feature */
710 /* use some fallback mechanism */
713 rather than the more impenetrable
715 #ifndef MISSING_NEATO_FEATURE
716 /* Not missing it, so we must have it, so use it */
718 /* Are missing it, so fall back on something else. */
721 Of course for this toy example, there's not much difference. But when
722 the #ifdef's start spanning a couple of screen fulls, and the #else's
723 are marked something like
725 #else /* !MISSING_NEATO_FEATURE */
727 I find it easy to get lost.
729 =item Providing Missing Functions -- Problem
731 Not all systems have all the neat functions you might want or need, so
732 you might decide to be helpful and provide an emulation. This is
733 sound in theory and very kind of you, but please be careful about what
734 you name the function. Let me use the C<pause()> function as an
737 Perl5.003 has the following in F<perl.h>
740 #define pause() sleep((32767<<16)+32767)
743 Configure sets HAS_PAUSE if the system has the pause() function, so
744 this #define only kicks in if the pause() function is missing.
747 Unfortunately, some systems apparently have a prototype for pause()
748 in F<unistd.h>, but don't actually have the function in the library.
749 (Or maybe they do have it in a library we're not using.)
751 Thus, the compiler sees something like
753 extern int pause(void);
755 #define pause() sleep((32767<<16)+32767)
757 and dies with an error message. (Some compilers don't mind this;
758 others apparently do.)
760 To work around this, 5.003_03 and later have the following in perl.h:
762 /* Some unistd.h's give a prototype for pause() even though
763 HAS_PAUSE ends up undefined. This causes the #define
764 below to be rejected by the compiler. Sigh.
769 # define Pause() sleep((32767<<16)+32767)
774 The curious reader may wonder why I didn't do the following in
780 sleep((32767<<16)+32767);
784 That is, since the function is missing, just provide it.
785 Then things would probably be been alright, it would seem.
787 Well, almost. It could be made to work. The problem arises from the
788 conflicting needs of dynamic loading and namespace protection.
790 For dynamic loading to work on AIX (and VMS) we need to provide a list
791 of symbols to be exported. This is done by the script F<perl_exp.SH>,
792 which reads F<global.sym> and F<interp.sym>. Thus, the C<pause>
793 symbol would have to be added to F<global.sym> So far, so good.
795 On the other hand, one of the goals of Perl5 is to make it easy to
796 either extend or embed perl and link it with other libraries. This
797 means we have to be careful to keep the visible namespace "clean".
798 That is, we don't want perl's global variables to conflict with
799 those in the other application library. Although this work is still
800 in progress, the way it is currently done is via the F<embed.h> file.
801 This file is built from the F<global.sym> and F<interp.sym> files,
802 since those files already list the globally visible symbols. If we
803 had added C<pause> to global.sym, then F<embed.h> would contain the
806 #define pause Perl_pause
808 and calls to C<pause> in the perl sources would now point to
809 C<Perl_pause>. Now, when B<ld> is run to build the F<perl> executable,
810 it will go looking for C<perl_pause>, which probably won't exist in any
811 of the standard libraries. Thus the build of perl will fail.
813 Those systems where C<HAS_PAUSE> is not defined would be ok, however,
814 since they would get a C<Perl_pause> function in util.c. The rest of
815 the world would be in trouble.
817 And yes, this scenario has happened. On SCO, the function C<chsize>
818 is available. (I think it's in F<-lx>, the Xenix compatibility
819 library.) Since the perl4 days (and possibly before), Perl has
820 included a C<chsize> function that gets called something akin to
823 I32 chsize(fd, length)
829 #define chsize Perl_chsize
831 to F<embed.h>, the compile started failing on SCO systems.
833 The "fix" is to give the function a different name. The one
834 implemented in 5.003_05 isn't optimal, but here's what was done:
837 # ifdef my_chsize /* Probably #defined to Perl_my_chsize in embed.h */
840 # define my_chsize chsize
843 My explanatory comment in patch 5.003_05 said:
845 Undef and then re-define my_chsize from Perl_my_chsize to
846 just plain chsize if this system HAS_CHSIZE. This probably only
847 applies to SCO. This shows the perils of having internal
848 functions with the same name as external library functions :-).
850 Now, we can safely put C<my_chsize> in F<global.sym>, export it, and
851 hide it with F<embed.h>.
853 To be consistent with what I did for C<pause>, I probably should have
854 called the new function C<Chsize>, rather than C<my_chsize>.
855 However, the perl sources are quite inconsistent on this (Consider
856 New, Mymalloc, and Myremalloc, to name just a few.)
858 There is a problem with this fix, however, in that C<Perl_chsize>
859 was available as a F<libperl.a> library function in 5.003, but it
860 isn't available any more (as of 5.003_07). This means that we've
861 broken binary compatibility. This is not good.
863 =item Providing missing functions -- some ideas
865 We currently don't have a standard way of handling such missing
866 function names. Right now, I'm effectively thinking aloud about a
867 solution. Some day, I'll try to formally propose a solution.
869 Part of the problem is that we want to have some functions listed as
870 exported but not have their names mangled by embed.h or possibly
871 conflict with names in standard system headers. We actually already
872 have such a list at the end of F<perl_exp.SH> (though that list is
875 # extra globals not included above.
876 cat <<END >> perl.exp
900 This still needs much thought, but I'm inclined to think that one
901 possible solution is to prefix all such functions with C<perl_> in the
902 source and list them along with the other C<perl_*> functions in
905 Thus, for C<chsize>, we'd do something like the following:
909 # define perl_chsize chsize
912 then in some file (e.g. F<util.c> or F<doio.c>) do
915 I32 perl_chsize(fd, length)
916 /* implement the function here . . . */
919 Alternatively, we could just always use C<chsize> everywhere and move
920 C<chsize> from F<global.sym> to the end of F<perl_exp.SH>. That would
921 probably be fine as long as our C<chsize> function agreed with all the
922 C<chsize> function prototypes in the various systems we'll be using.
923 As long as the prototypes in actual use don't vary that much, this is
924 probably a good alternative. (As a counter-example, note how Configure
925 and perl have to go through hoops to find and use get Malloc_t and
926 Free_t for C<malloc> and C<free>.)
928 At the moment, this latter option is what I tend to prefer.
930 =item All the world's a VAX
932 Sorry, showing my age:-). Still, all the world is not BSD 4.[34],
933 SVR4, or POSIX. Be aware that SVR3-derived systems are still quite
934 common (do you have any idea how many systems run SCO?) If you don't
935 have a bunch of v7 manuals handy, the metaconfig units (by default
936 installed in F</usr/local/lib/dist/U>) are a good resource to look at
941 =head1 Miscellaneous Topics
945 Why does perl use a metaconfig-generated Configure script instead of an
946 autoconf-generated configure script?
948 Metaconfig and autoconf are two tools with very similar purposes.
949 Metaconfig is actually the older of the two, and was originally written
950 by Larry Wall, while autoconf is probably now used in a wider variety of
951 packages. The autoconf info file discusses the history of autoconf and
952 how it came to be. The curious reader is referred there for further
955 Overall, both tools are quite good, I think, and the choice of which one
956 to use could be argued either way. In March, 1994, when I was just
957 starting to work on Configure support for Perl5, I considered both
958 autoconf and metaconfig, and eventually decided to use metaconfig for the
963 =item Compatibility with Perl4
965 Perl4 used metaconfig, so many of the #ifdef's were already set up for
966 metaconfig. Of course metaconfig had evolved some since Perl4's days,
967 but not so much that it posed any serious problems.
969 =item Metaconfig worked for me
971 My system at the time was Interactive 2.2, a SVR3.2/386 derivative that
972 also had some POSIX support. Metaconfig-generated Configure scripts
973 worked fine for me on that system. On the other hand, autoconf-generated
974 scripts usually didn't. (They did come quite close, though, in some
975 cases.) At the time, I actually fetched a large number of GNU packages
976 and checked. Not a single one configured and compiled correctly
977 out-of-the-box with the system's cc compiler.
979 =item Configure can be interactive
981 With both autoconf and metaconfig, if the script works, everything is
982 fine. However, one of my main problems with autoconf-generated scripts
983 was that if it guessed wrong about something, it could be B<very> hard to
984 go back and fix it. For example, autoconf always insisted on passing the
985 -Xp flag to cc (to turn on POSIX behavior), even when that wasn't what I
986 wanted or needed for that package. There was no way short of editing the
987 configure script to turn this off. You couldn't just edit the resulting
988 Makefile at the end because the -Xp flag influenced a number of other
991 Metaconfig's Configure scripts, on the other hand, can be interactive.
992 Thus if Configure is guessing things incorrectly, you can go back and fix
993 them. This isn't as important now as it was when we were actively
994 developing Configure support for new features such as dynamic loading,
995 but it's still useful occasionally.
999 At the time, autoconf-generated scripts were covered under the GNU Public
1000 License, and hence weren't suitable for inclusion with Perl, which has a
1001 different licensing policy. (Autoconf's licensing has since changed.)
1005 Metaconfig builds up Configure from a collection of discrete pieces
1006 called "units". You can override the standard behavior by supplying your
1007 own unit. With autoconf, you have to patch the standard files instead.
1008 I find the metaconfig "unit" method easier to work with. Others
1009 may find metaconfig's units clumsy to work with.
1013 =head2 @INC search order
1015 By default, the list of perl library directories in @INC is the
1023 Specifically, on my Solaris/x86 system, I run
1024 B<sh Configure -Dprefix=/opt/perl> and I have the following
1027 /opt/perl/lib/i86pc-solaris/5.00307
1029 /opt/perl/lib/site_perl/i86pc-solaris
1030 /opt/perl/lib/site_perl
1032 That is, perl's directories come first, followed by the site-specific
1035 The site libraries come second to support the usage of extensions
1036 across perl versions. Read the relevant section in F<INSTALL> for
1037 more information. If we ever make $sitearch version-specific, this
1038 topic could be revisited.
1040 =head2 Why isn't there a directory to override Perl's library?
1042 Mainly because no one's gotten around to making one. Note that
1043 "making one" involves changing perl.c, Configure, config_h.SH (and
1044 associated files, see above), and I<documenting> it all in the
1047 Apparently, most folks who want to override one of the standard library
1048 files simply do it by overwriting the standard library files.
1052 In the perl.c sources, you'll find an undocumented APPLLIB_EXP
1053 variable, sort of like PRIVLIB_EXP and ARCHLIB_EXP (which are
1054 documented in config_h.SH). Here's what APPLLIB_EXP is for, from
1055 a mail message from Larry:
1057 The main intent of APPLLIB_EXP is for folks who want to send out a
1058 version of Perl embedded in their product. They would set the symbol
1059 to be the name of the library containing the files needed to run or to
1060 support their particular application. This works at the "override"
1061 level to make sure they get their own versions of any library code that
1062 they absolutely must have configuration control over.
1064 As such, I don't see any conflict with a sysadmin using it for a
1065 override-ish sort of thing, when installing a generic Perl. It should
1066 probably have been named something to do with overriding though. Since
1067 it's undocumented we could still change it... :-)
1069 Given that it's already there, you can use it to override
1070 distribution modules. If you do
1072 sh Configure -Dccflags='-DAPPLLIB_EXP=/my/override'
1074 then perl.c will put /my/override ahead of ARCHLIB and PRIVLIB.
1076 =head2 Shared libperl.so location
1078 Why isn't the shared libperl.so installed in /usr/lib/ along
1079 with "all the other" shared libraries? Instead, it is installed
1080 in $archlib, which is typically something like
1082 /usr/local/lib/perl5/archname/5.00404
1084 and is architecture- and version-specific.
1086 The basic reason why a shared libperl.so gets put in $archlib is so that
1087 you can have more than one version of perl on the system at the same time,
1088 and have each refer to its own libperl.so.
1090 Three examples might help. All of these work now; none would work if you
1091 put libperl.so in /usr/lib.
1097 Suppose you want to have both threaded and non-threaded perl versions
1098 around. Configure will name both perl libraries "libperl.so" (so that
1099 you can link to them with -lperl). The perl binaries tell them apart
1100 by having looking in the appropriate $archlib directories.
1104 Suppose you have perl5.004_04 installed and you want to try to compile
1105 it again, perhaps with different options or after applying a patch.
1106 If you already have libperl.so installed in /usr/lib/, then it may be
1107 either difficult or impossible to get ld.so to find the new libperl.so
1108 that you're trying to build. If, instead, libperl.so is tucked away in
1109 $archlib, then you can always just change $archlib in the current perl
1110 you're trying to build so that ld.so won't find your old libperl.so.
1111 (The INSTALL file suggests you do this when building a debugging perl.)
1115 The shared perl library is not a "well-behaved" shared library with
1116 proper major and minor version numbers, so you can't necessarily
1117 have perl5.004_04 and perl5.004_05 installed simultaneously. Suppose
1118 perl5.004_04 were to install /usr/lib/libperl.so.4.4, and perl5.004_05
1119 were to install /usr/lib/libperl.so.4.5. Now, when you try to run
1120 perl5.004_04, ld.so might try to load libperl.so.4.5, since it has
1121 the right "major version" number. If this works at all, it almost
1122 certainly defeats the reason for keeping perl5.004_04 around. Worse,
1123 with development subversions, you certaily can't guarantee that
1124 libperl.so.4.4 and libperl.so.4.55 will be compatible.
1126 Anyway, all this leads to quite obscure failures that are sure to drive
1127 casual users crazy. Even experienced users will get confused :-). Upon
1128 reflection, I'd say leave libperl.so in $archlib.
1132 =head1 Upload Your Work to CPAN
1134 You can upload your work to CPAN if you have a CPAN id. Check out
1135 http://www.perl.com/CPAN/modules/04pause.html for information on
1136 _PAUSE_, the Perl Author's Upload Server.
1138 I typically upload both the patch file, e.g. F<perl5.004_08.pat.gz>
1139 and the full tar file, e.g. F<perl5.004_08.tar.gz>.
1141 If you want your patch to appear in the F<src/5.0/unsupported>
1142 directory on CPAN, send e-mail to the CPAN master librarian. (Check
1143 out http://www.perl.com/CPAN/CPAN.html ).
1145 =head1 Help Save the World
1147 You should definitely announce your patch on the perl5-porters list.
1148 You should also consider announcing your patch on
1149 comp.lang.perl.announce, though you should make it quite clear that a
1150 subversion is not a production release, and be prepared to deal with
1151 people who will not read your disclaimer.
1155 Here, in no particular order, are some Configure and build-related
1156 items that merit consideration. This list isn't exhaustive, it's just
1157 what I came up with off the top of my head.
1159 =head2 Good ideas waiting for round tuits
1165 I think we ought to support
1167 Configure -Dinstallprefix=/blah/blah
1169 Currently, we support B<-Dprefix=/blah/blah>, but the changing the install
1170 location has to be handled by something like the F<config.over> trick
1171 described in F<INSTALL>. AFS users also are treated specially.
1172 We should probably duplicate the metaconfig prefix stuff for an
1175 =item Configure -Dsrc=/blah/blah
1177 We should be able to emulate B<configure --srcdir>. Tom Tromey
1178 tromey@creche.cygnus.com has submitted some patches to
1179 the dist-users mailing list along these lines. They have been folded
1180 back into the main distribution, but various parts of the perl
1181 Configure/build/install process still assume src='.'.
1183 =item Hint file fixes
1185 Various hint files work around Configure problems. We ought to fix
1186 Configure so that most of them aren't needed.
1188 =item Hint file information
1190 Some of the hint file information (particularly dynamic loading stuff)
1191 ought to be fed back into the main metaconfig distribution.
1193 =item Catch GNU Libc "Stub" functions
1195 Some functions (such as lchown()) are present in libc, but are
1196 unimplmented. That is, they always fail and set errno=ENOSYS.
1198 Thomas Bushnell provided the following sample code and the explanation
1201 /* System header to define __stub macros and hopefully few prototypes,
1202 which can conflict with char FOO(); below. */
1204 /* Override any gcc2 internal prototype to avoid an error. */
1205 /* We use char because int might match the return type of a gcc2
1206 builtin and then its argument prototype would still apply. */
1211 /* The GNU C library defines this for functions which it implements
1212 to always fail with ENOSYS. Some functions are actually named
1213 something starting with __ and the normal name is an alias. */
1214 #if defined (__stub_FOO) || defined (__stub___FOO)
1222 The choice of <assert.h> is essentially arbitrary. The GNU libc
1223 macros are found in <gnu/stubs.h>. You can include that file instead
1224 of <assert.h> (which itself includes <gnu/stubs.h>) if you test for
1225 its existence first. <assert.h> is assumed to exist on every system,
1226 which is why it's used here. Any GNU libc header file will include
1227 the stubs macros. If either __stub_NAME or __stub___NAME is defined,
1228 then the function doesn't actually exist. Tests using <assert.h> work
1229 on every system around.
1231 The declaration of FOO is there to override builtin prototypes for
1236 =head2 Probably good ideas waiting for round tuits
1240 =item GNU configure --options
1242 I've received sensible suggestions for --exec_prefix and other
1243 GNU configure --options. It's not always obvious exactly what is
1244 intended, but this merits investigation.
1248 Currently, B<make clean> isn't all that useful, though
1249 B<make realclean> and B<make distclean> are. This needs a bit of
1250 thought and documentation before it gets cleaned up.
1252 =item Try gcc if cc fails
1254 Currently, we just give up.
1256 =item bypassing safe*alloc wrappers
1258 On some systems, it may be safe to call the system malloc directly
1259 without going through the util.c safe* layers. (Such systems would
1260 accept free(0), for example.) This might be a time-saver for systems
1261 that already have a good malloc. (Recent Linux libc's apparently have
1262 a nice malloc that is well-tuned for the system.)
1266 =head2 Vague possibilities
1272 Get some of the Macintosh stuff folded back into the main distribution.
1274 =item gconvert replacement
1276 Maybe include a replacement function that doesn't lose data in rare
1277 cases of coercion between string and numerical values.
1279 =item Improve makedepend
1281 The current makedepend process is clunky and annoyingly slow, but it
1282 works for most folks. Alas, it assumes that there is a filename
1283 $firstmakefile that the B<make> command will try to use before it uses
1284 F<Makefile>. Such may not be the case for all B<make> commands,
1285 particularly those on non-Unix systems.
1287 Probably some variant of the BSD F<.depend> file will be useful.
1288 We ought to check how other packages do this, if they do it at all.
1289 We could probably pre-generate the dependencies (with the exception of
1290 malloc.o, which could probably be determined at F<Makefile.SH>
1293 =item GNU Makefile standard targets
1295 GNU software generally has standardized Makefile targets. Unless we
1296 have good reason to do otherwise, I see no reason not to support them.
1300 Somehow, straighten out, document, and implement lockf(), flock(),
1301 and/or fcntl() file locking. It's a mess.
1307 Original author: Andy Dougherty doughera@lafcol.lafayette.edu .
1308 Additions by Chip Salzenberg chip@perl.com and
1309 Tim Bunce Tim.Bunce@ig.co.uk .
1311 All opinions expressed herein are those of the authorZ<>(s).
1313 =head1 LAST MODIFIED
1315 $Id: pumpkin.pod,v 1.15 1998/04/23 17:03:48 doughera Released $