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, 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 =head1 How are Perl Releases Numbered?
46 Perl version numbers are floating point numbers, such as 5.004. The
47 major version number is 5, the minor version is '0', and '03' is the
48 patchlevel. The version number is available as the magic variable $],
49 and can be used in comparisons, e.g.
51 print "You've got an old perl\n" if $] < 5.002;
53 (Observations about the imprecision of floating point numbers for
54 representing reality probably have more relevance than you might
57 You can also require particular version (or later) with
63 In addition, there may be "developer" sub-versions available. These
64 are not official releases. They may contain unstable experimental
65 features, and are subject to rapid change. Such developer
66 sub-versions are numbered with sub-version numbers. For example,
67 version 5.004_04 is the 4'th developer version built on top of
68 5.004. It might include the _01, _02, and _03 changes, but it
69 also might not. Sub-versions are allowed to be subversive.
71 These sub-versions can also be used as floating point numbers, so
72 you can do things such as
74 print "You've got an unstable perl\n" if $] == 5.00403;
76 You can also require particular version (or later) with
78 use 5.004_03; # the "_" is optional
80 Sub-versions produced by the members of perl5-porters are usually
81 available on CPAN in the F<src/5.0/unsupported> directory.
83 =head2 Why such a complicated scheme?
85 Two reasons, really. At least.
87 First, we need some way to identify releases that are known to
88 have new features that need testing and exploration. The
89 subversion scheme does that nicely while fitting into the
92 Second, since most of the folks who help maintain perl do so on a
93 free-time voluntary basis, perl development does not proceed at a
94 precise pace, though it always seems to be moving ahead quickly.
95 We needed some way to pass around the "patch pumpkin" to allow
96 different people chances to work on different aspects of the
97 distribution without getting in each other's way. It wouldn't be
98 constructive to have multiple people working on incompatible
99 implementations of the same idea. Instead what was needed was
100 some kind of "baton" or "token" to pass around so everyone knew
103 =head2 Why is it called the patch pumpkin?
105 Chip Salzenberg gets credit for that, with a nod to his cow orker,
106 David Croy. We had passed around various names (baton, token, hot
107 potato) but none caught on. Then, Chip asked:
111 Who has the patch pumpkin?
113 To explain: David Croy once told me once that at a previous job,
114 there was one tape drive and multiple systems that used it for backups.
115 But instead of some high-tech exclusion software, they used a low-tech
116 method to prevent multiple simultaneous backups: a stuffed pumpkin.
117 No one was allowed to make backups unless they had the "backup pumpkin".
123 =head1 Philosophical Issues in Patching Perl
125 There are no absolute rules, but there are some general guidelines I
126 have tried to follow as I apply patches to the perl sources.
127 (This section is still under construction.)
129 =head2 Solve problems as generally as possible
131 (I still have to think of a good example here.)
133 =head2 Seek consensus on major changes
135 If you are making big changes, don't do it in secret. Discuss the
136 ideas in advance on perl5-porters.
138 =head2 Keep the documentation up-to-date
140 If your changes may affect how users use perl, then check to be sure
141 that the documentation is in sync with your changes. Be sure to
142 check all the files F<pod/*.pod> and also the F<INSTALL> document.
144 Consider writing the appropriate documentation first and then
145 implementing it to correspond to the documentation.
147 =head2 Avoid machine-specific #ifdef's
149 To the extent reasonable, try to avoid machine-specific #ifdef's in
150 the sources. Instead, use feature-specific #ifdef's. The reason is
151 that the machine-specific #ifdef's may not be valid across major
152 releases of the operating system. Further, the feature-specific tests
153 may help out folks on another platform who have the same problem.
155 =head2 Allow for lots of testing
157 We should never release a main version without testing it as a
160 =head2 Automate generatation of derivative files
162 The F<embed.h>, F<keywords.h>, F<opcode.h>, and F<perltoc.pod> files
163 are all automatically generated by perl scripts. In general, don't
164 patch these directly; patch the data files instead.
166 F<Configure> and F<config_h.SH> are also automatically generated by
167 B<metaconfig>. In general, you should patch the metaconfig units
168 instead of patching these files directly. However, minor changes to
169 F<Configure> may be made in between major sync-ups with the metaconfig
170 units, which tends to be complicated operations.
172 =head1 How to Make a Distribution
174 There really ought to be a 'make dist' target, but there isn't.
175 The 'dist' suite of tools also contains a number of tools that I haven't
176 learned how to use yet. Some of them may make this all a bit easier.
178 Here are the steps I go through to prepare a patch & distribution.
180 Lots of it could doubtless be automated but isn't.
182 =head2 Announce your intentions
184 First, you should volunteer out loud to take the patch pumpkin. It's
185 generally counter-productive to have multiple people working in secret
188 At the same time, announce what you plan to do with the patch pumpkin,
189 to allow folks a chance to object or suggest alternatives, or do it for
190 you. Naturally, the patch pumpkin holder ought to incorporate various
191 bug fixes and documentation improvements that are posted while he or
192 she has the pumpkin, but there might also be larger issues at stake.
194 One of the precepts of the subversion idea is that we shouldn't give
195 it to anyone unless we have some idea what you're going to do with
198 =head2 refresh pod/perltoc.pod
200 Presumably, you have done a full C<make> in your working source
201 directory. Before you C<make spotless> (if you do), and if you have
202 changed any documentation in any module or pod file, change to the
203 F<pod> directory and run C<make toc>.
205 =head2 update patchlevel.h
207 Don't be shy about using the subversion number, even for a relatively
208 modest patch. We've never even come close to using all 99 subversions,
209 and it's better to have a distinctive number for your patch. If you
210 need feedback on your patch, go ahead and issue it and promise to
211 incorporate that feedback quickly (e.g. within 1 week) and send out a
214 =head2 run metaconfig
216 If you need to make changes to Configure or config_h.SH, it may be best to
217 change the appropriate metaconfig units instead, and regenerate Configure.
221 will regenerate Configure and config_h.SH. More information on
222 obtaining and running metaconfig is in the F<U/README> file that comes
223 with Perl's metaconfig units. Perl's metaconfig units should be
224 available the same place you found this file. On CPAN, look under my
225 directory F<id/ANDYD/> for a file such as F<5.003_07-02.U.tar.gz>.
226 That file should be unpacked in your main perl source directory. It
227 contains the files needed to run B<metaconfig> to reproduce Perl's
230 Alternatively, do consider if the F<*ish.h> files might be a better
231 place for your changes.
235 Make sure the MANIFEST is up-to-date. You can use dist's B<manicheck>
236 program for this. You can also use
238 perl -MExtUtils::Manifest -e fullcheck
240 to do half the job. This will make sure everything listed in MANIFEST
241 is included in the distribution. dist's B<manicheck> command will
242 also list extra files in the directory that are not listed in
245 The MANIFEST is normally sorted, with one exception. Perl includes
246 both a F<Configure> script and a F<configure> script. The
247 F<configure> script is a front-end to the main F<Configure>, but
248 is there to aid folks who use autoconf-generated F<configure> files
249 for other software. The problem is that F<Configure> and F<configure>
250 are the same on case-insensitive file systems, so I deliberately put
251 F<configure> first in the MANIFEST so that the extraction of
252 F<Configure> will overwrite F<configure> and leave you with the
253 correct script. (The F<configure> script must also have write
254 permission for this to work, so it's the only file in the distribution
255 I normally have with write permission.)
257 If you are using metaconfig to regenerate Configure, then you should note
258 that metaconfig actually uses MANIFEST.new, so you want to be sure
259 MANIFEST.new is up-to-date too. I haven't found the MANIFEST/MANIFEST.new
260 distinction particularly useful, but that's probably because I still haven't
261 learned how to use the full suite of tools in the dist distribution.
263 =head2 Check permissions
265 All the tests in the t/ directory ought to be executable. The
266 main makefile used to do a 'chmod t/*/*.t', but that resulted in
267 a self-modifying distribution--something some users would strongly
268 prefer to avoid. Probably, the F<t/TEST> script should check for this
269 and do the chmod if needed, but it doesn't currently.
271 In all, the following files should probably be executable:
287 vms/ext/Stdio/test.pl
292 Other things ought to be readable, at least :-).
294 Probably, the permissions for the files could be encoded in MANIFEST
295 somehow, but I'm reluctant to change MANIFEST itself because that
296 could break old scripts that use MANIFEST.
298 I seem to recall that some SVR3 systems kept some sort of file that listed
299 permissions for system files; something like that might be appropriate.
303 This will build a config.sh and config.h. You can skip this if you haven't
304 changed Configure or config_h.SH at all.
306 =head2 Update config_H
308 The config_H file is provided to help those folks who can't run Configure.
309 It is important to keep it up-to-date. If you have changed config_h.SH,
310 those changes must be reflected in config_H as well. (The name config_H was
311 chosen to distinguish the file from config.h even on case-insensitive file
312 systems.) Simply edit the existing config_H file; keep the first few
313 explanatory lines and then copy your new config.h below.
315 It may also be necessary to update vms/config.vms and
316 plan9/config.plan9, though you should be quite careful in doing so if
317 you are not familiar with those systems. You might want to issue your
318 patch with a promise to quickly issue a follow-up that handles those
321 =head2 make run_byacc
323 If you have byacc-1.8.2 (available from CPAN), and if there have been
324 changes to F<perly.y>, you can regenerate the F<perly.c> file. The
325 run_byacc makefile target does this by running byacc and then applying
326 some patches so that byacc dynamically allocates space, rather than
327 having fixed limits. This patch is handled by the F<perly.fixer>
328 script. Depending on the nature of the changes to F<perly.y>, you may
329 or may not have to hand-edit the patch to apply correctly. If you do,
330 you should include the edited patch in the new distribution. If you
331 have byacc-1.9, the patch won't apply cleanly. Changes to the printf
332 output statements mean the patch won't apply cleanly. Long ago I
333 started to fix F<perly.fixer> to detect this, but I never completed the
336 Some additional notes from Larry on this:
338 Don't forget to regenerate perly.c.diff.
342 patch perly.c <perly.c.diff
343 # manually apply any failed hunks
344 diff -c2 perly.c.orig perly.c >perly.c.diff
346 One chunk of lines that often fails begins with
350 and ends one line before
352 #define YYERRCODE 256
354 This only happens when you add or remove a token type. I suppose this
355 could be automated, but it doesn't happen very often nowadays.
359 =head2 make regen_headers
361 The F<embed.h>, F<keywords.h>, and F<opcode.h> files are all automatically
362 generated by perl scripts. Since the user isn't guaranteed to have a
363 working perl, we can't require the user to generate them. Hence you have
364 to, if you're making a distribution.
366 I used to include rules like the following in the makefile:
368 # The following three header files are generated automatically
369 # The correct versions should be already supplied with the perl kit,
370 # in case you don't have perl or 'sh' available.
371 # The - is to ignore error return codes in case you have the source
372 # installed read-only or you don't have perl yet.
373 keywords.h: keywords.pl
374 @echo "Don't worry if this fails."
378 However, I got lots of mail consisting of people worrying because the
379 command failed. I eventually decided that I would save myself time
380 and effort by manually running C<make regen_headers> myself rather
381 than answering all the questions and complaints about the failing
384 =head2 global.sym and interp.sym
386 Make sure these files are up-to-date. Read the comments in these
387 files and in perl_exp.SH to see what to do.
389 =head2 Binary compatibility
391 If you do change F<global.sym> or F<interp.sym>, think carefully about
392 what you are doing. To the extent reasonable, we'd like to maintain
393 souce and binary compatibility with older releases of perl. That way,
394 extensions built under one version of perl will continue to work with
395 new versions of perl.
397 Of course, some incompatible changes may well be necessary. I'm just
398 suggesting that we not make any such changes without thinking carefully
399 about them first. If possible, we should provide
400 backwards-compatibility stubs. There's a lot of XS code out there.
401 Let's not force people to keep changing it.
405 Be sure to update the F<Changes> file. Try to include both an overall
406 summary as well as detailed descriptions of the changes. Your
407 audience will include bother developers and users, so describe
408 user-visible changes (if any) in terms they will understand, not in
409 code like "initialize foo variable in bar function".
411 There are differing opinions on whether the detailed descriptions
412 ought to go in the Changes file or whether they ought to be available
413 separately in the patch file (or both). There is no disagreement that
414 detailed descriptions ought to be easily available somewhere.
416 =head2 OS/2-specific updates
418 In the os2 directory is F<diff.configure>, a set of OS/2-specific
419 diffs against B<Configure>. If you make changes to Configure, you may
420 want to consider regenerating this diff file to save trouble for the
423 =head2 VMS-specific updates
425 If you have changed F<perly.y>, then you may want to update
426 F<vms/perly_{h,c}.vms> by running C<perl vms/vms_yfix.pl>.
428 The Perl version number appears in several places under F<vms>.
429 It is courteous to update these versions. For example, if you are
430 making 5.004_42, replace "5.00441" with "5.00442".
432 =head2 Making the new distribution
434 Suppose, for example, that you want to make version 5.004_08. Then you can
435 do something like the following
437 mkdir ../perl5.004_08
438 awk '{print $1}' MANIFEST | cpio -pdm ../perl5.004_08
440 tar cf perl5.004_08.tar perl5.004_08
441 gzip --best perl5.004_08.tar
443 =head2 Making a new patch
445 I find the F<makepatch> utility quite handy for making patches.
446 You can obtain it from any CPAN archive under
447 http://www.perl.com/CPAN/authors/Johan_Vromans/. The only
448 difference between my version and the standard one is that I have mine
451 # Print a reassuring "End of Patch" note so people won't
452 # wonder if their mailer truncated patches.
453 print "\n\nEnd of Patch.\n";
455 at the end. That's because I used to get questions from people asking if
456 their mail was truncated.
458 Here's how I generate a new patch. I'll use the hypothetical
459 5.004_07 to 5.004_08 patch as an example.
461 # unpack perl5.004_07/
462 gzip -d -c perl5.004_07.tar.gz | tar -xof -
463 # unpack perl5.004_08/
464 gzip -d -c perl5.004_08.tar.gz | tar -xof -
465 makepatch perl5.004_07 perl5.004_08 > perl5.004_08.pat
467 Makepatch will automatically generate appropriate B<rm> commands to remove
468 deleted files. Unfortunately, it will not correctly set permissions
469 for newly created files, so you may have to do so manually. For example,
470 patch 5.003_04 created a new test F<t/op/gv.t> which needs to be executable,
471 so at the top of the patch, I inserted the following lines:
477 Now, of course, my patch is now wrong because makepatch didn't know I
478 was going to do that command, and it patched against /dev/null.
480 So, what I do is sort out all such shell commands that need to be in the
481 patch (including possible mv-ing of files, if needed) and put that in the
482 shell commands at the top of the patch. Next, I delete all the patch parts
483 of perl5.004_08.pat, leaving just the shell commands. Then, I do the
487 sh ../perl5.003_08.pat
489 makepatch perl5.003_07 perl5.003_08 >> perl5.003_08.pat
491 (Note the append to preserve my shell commands.)
492 Now, my patch will line up with what the end users are going to do.
494 =head2 Testing your patch
496 It seems obvious, but be sure to test your patch. That is, verify that
497 it produces exactly the same thing as your full distribution.
500 gzip -d -c perl5.003_07.tar.gz | tar -xf -
502 sh ../perl5.003_08.pat
503 patch -p1 -N < ../perl5.003_08.pat
505 gdiff -r perl5.003_07 perl5.003_08
507 where B<gdiff> is GNU diff. Other diff's may also do recursive checking.
511 Again, it's obvious, but you should test your new version as widely as you
512 can. You can be sure you'll hear about it quickly if your version doesn't
513 work on both ANSI and pre-ANSI compilers, and on common systems such as
514 SunOS 4.1.[34], Solaris, and Linux.
516 If your changes include conditional code, try to test the different
517 branches as thoroughly as you can. For example, if your system
518 supports dynamic loading, you can also test static loading with
522 You can also hand-tweak your config.h to try out different #ifdef
525 =head1 Common Gotcha's
531 The '#elif' preprocessor directive is not understood on all systems.
532 Specifically, I know that Pyramids don't understand it. Thus instead of the
543 You have to do the more Byzantine
555 Incidentally, whitespace between the leading '#' and the preprocessor
556 command is not guaranteed, but is very portable and you may use it freely.
557 I think it makes things a bit more readable, especially once things get
558 rather deeply nested. I also think that things should almost never get
559 too deeply nested, so it ought to be a moot point :-)
561 =item Probably Prefer POSIX
563 It's often the case that you'll need to choose whether to do
564 something the BSD-ish way or the POSIX-ish way. It's usually not
565 a big problem when the two systems use different names for similar
566 functions, such as memcmp() and bcmp(). The perl.h header file
567 handles these by appropriate #defines, selecting the POSIX mem*()
568 functions if available, but falling back on the b*() functions, if
571 More serious is the case where some brilliant person decided to
572 use the same function name but give it a different meaning or
573 calling sequence :-). getpgrp() and setpgrp() come to mind.
574 These are a real problem on systems that aim for conformance to
575 one standard (e.g. POSIX), but still try to support the other way
576 of doing things (e.g. BSD). My general advice (still not really
577 implemented in the source) is to do something like the following.
578 Suppose there are two alternative versions, fooPOSIX() and
582 /* use fooPOSIX(); */
585 /* try to emulate fooPOSIX() with fooBSD();
586 perhaps with the following: */
587 # define fooPOSIX fooBSD
589 # /* Uh, oh. We have to supply our own. */
590 # define fooPOSIX Perl_fooPOSIX
594 =item Think positively
596 If you need to add an #ifdef test, it is usually easier to follow if you
597 think positively, e.g.
599 #ifdef HAS_NEATO_FEATURE
600 /* use neato feature */
602 /* use some fallback mechanism */
605 rather than the more impenetrable
607 #ifndef MISSING_NEATO_FEATURE
608 /* Not missing it, so we must have it, so use it */
610 /* Are missing it, so fall back on something else. */
613 Of course for this toy example, there's not much difference. But when
614 the #ifdef's start spanning a couple of screen fulls, and the #else's
615 are marked something like
617 #else /* !MISSING_NEATO_FEATURE */
619 I find it easy to get lost.
621 =item Providing Missing Functions -- Problem
623 Not all systems have all the neat functions you might want or need, so
624 you might decide to be helpful and provide an emulation. This is
625 sound in theory and very kind of you, but please be careful about what
626 you name the function. Let me use the C<pause()> function as an
629 Perl5.003 has the following in F<perl.h>
632 #define pause() sleep((32767<<16)+32767)
635 Configure sets HAS_PAUSE if the system has the pause() function, so
636 this #define only kicks in if the pause() function is missing.
639 Unfortunately, some systems apparently have a prototype for pause()
640 in F<unistd.h>, but don't actually have the function in the library.
641 (Or maybe they do have it in a library we're not using.)
643 Thus, the compiler sees something like
645 extern int pause(void);
647 #define pause() sleep((32767<<16)+32767)
649 and dies with an error message. (Some compilers don't mind this;
650 others apparently do.)
652 To work around this, 5.003_03 and later have the following in perl.h:
654 /* Some unistd.h's give a prototype for pause() even though
655 HAS_PAUSE ends up undefined. This causes the #define
656 below to be rejected by the compiler. Sigh.
661 # define Pause() sleep((32767<<16)+32767)
666 The curious reader may wonder why I didn't do the following in
672 sleep((32767<<16)+32767);
676 That is, since the function is missing, just provide it.
677 Then things would probably be been alright, it would seem.
679 Well, almost. It could be made to work. The problem arises from the
680 conflicting needs of dynamic loading and namespace protection.
682 For dynamic loading to work on AIX (and VMS) we need to provide a list
683 of symbols to be exported. This is done by the script F<perl_exp.SH>,
684 which reads F<global.sym> and F<interp.sym>. Thus, the C<pause>
685 symbol would have to be added to F<global.sym> So far, so good.
687 On the other hand, one of the goals of Perl5 is to make it easy to
688 either extend or embed perl and link it with other libraries. This
689 means we have to be careful to keep the visible namespace "clean".
690 That is, we don't want perl's global variables to conflict with
691 those in the other application library. Although this work is still
692 in progress, the way it is currently done is via the F<embed.h> file.
693 This file is built from the F<global.sym> and F<interp.sym> files,
694 since those files already list the globally visible symbols. If we
695 had added C<pause> to global.sym, then F<embed.h> would contain the
698 #define pause Perl_pause
700 and calls to C<pause> in the perl sources would now point to
701 C<Perl_pause>. Now, when B<ld> is run to build the F<perl> executable,
702 it will go looking for C<perl_pause>, which probably won't exist in any
703 of the standard libraries. Thus the build of perl will fail.
705 Those systems where C<HAS_PAUSE> is not defined would be ok, however,
706 since they would get a C<Perl_pause> function in util.c. The rest of
707 the world would be in trouble.
709 And yes, this scenario has happened. On SCO, the function C<chsize>
710 is available. (I think it's in F<-lx>, the Xenix compatibility
711 library.) Since the perl4 days (and possibly before), Perl has
712 included a C<chsize> function that gets called something akin to
715 I32 chsize(fd, length)
721 #define chsize Perl_chsize
723 to F<embed.h>, the compile started failing on SCO systems.
725 The "fix" is to give the function a different name. The one
726 implemented in 5.003_05 isn't optimal, but here's what was done:
729 # ifdef my_chsize /* Probably #defined to Perl_my_chsize in embed.h */
732 # define my_chsize chsize
735 My explanatory comment in patch 5.003_05 said:
737 Undef and then re-define my_chsize from Perl_my_chsize to
738 just plain chsize if this system HAS_CHSIZE. This probably only
739 applies to SCO. This shows the perils of having internal
740 functions with the same name as external library functions :-).
742 Now, we can safely put C<my_chsize> in F<global.sym>, export it, and
743 hide it with F<embed.h>.
745 To be consistent with what I did for C<pause>, I probably should have
746 called the new function C<Chsize>, rather than C<my_chsize>.
747 However, the perl sources are quite inconsistent on this (Consider
748 New, Mymalloc, and Myremalloc, to name just a few.)
750 There is a problem with this fix, however, in that C<Perl_chsize>
751 was available as a F<libperl.a> library function in 5.003, but it
752 isn't available any more (as of 5.003_07). This means that we've
753 broken binary compatibility. This is not good.
755 =item Providing missing functions -- some ideas
757 We currently don't have a standard way of handling such missing
758 function names. Right now, I'm effectively thinking aloud about a
759 solution. Some day, I'll try to formally propose a solution.
761 Part of the problem is that we want to have some functions listed as
762 exported but not have their names mangled by embed.h or possibly
763 conflict with names in standard system headers. We actually already
764 have such a list at the end of F<perl_exp.SH> (though that list is
767 # extra globals not included above.
768 cat <<END >> perl.exp
792 This still needs much thought, but I'm inclined to think that one
793 possible solution is to prefix all such functions with C<perl_> in the
794 source and list them along with the other C<perl_*> functions in
797 Thus, for C<chsize>, we'd do something like the following:
801 # define perl_chsize chsize
804 then in some file (e.g. F<util.c> or F<doio.c>) do
807 I32 perl_chsize(fd, length)
808 /* implement the function here . . . */
811 Alternatively, we could just always use C<chsize> everywhere and move
812 C<chsize> from F<global.sym> to the end of F<perl_exp.SH>. That would
813 probably be fine as long as our C<chsize> function agreed with all the
814 C<chsize> function prototypes in the various systems we'll be using.
815 As long as the prototypes in actual use don't vary that much, this is
816 probably a good alternative. (As a counter-example, note how Configure
817 and perl have to go through hoops to find and use get Malloc_t and
818 Free_t for C<malloc> and C<free>.)
820 At the moment, this latter option is what I tend to prefer.
822 =item All the world's a VAX
824 Sorry, showing my age:-). Still, all the world is not BSD 4.[34],
825 SVR4, or POSIX. Be aware that SVR3-derived systems are still quite
826 common (do you have any idea how many systems run SCO?) If you don't
827 have a bunch of v7 manuals handy, the metaconfig units (by default
828 installed in F</usr/local/lib/dist/U>) are a good resource to look at
833 =head1 Miscellaneous Topics
837 Why does perl use a metaconfig-generated Configure script instead of an
838 autoconf-generated configure script?
840 Metaconfig and autoconf are two tools with very similar purposes.
841 Metaconfig is actually the older of the two, and was originally written
842 by Larry Wall, while autoconf is probably now used in a wider variety of
843 packages. The autoconf info file discusses the history of autoconf and
844 how it came to be. The curious reader is referred there for further
847 Overall, both tools are quite good, I think, and the choice of which one
848 to use could be argued either way. In March, 1994, when I was just
849 starting to work on Configure support for Perl5, I considered both
850 autoconf and metaconfig, and eventually decided to use metaconfig for the
855 =item Compatibility with Perl4
857 Perl4 used metaconfig, so many of the #ifdef's were already set up for
858 metaconfig. Of course metaconfig had evolved some since Perl4's days,
859 but not so much that it posed any serious problems.
861 =item Metaconfig worked for me
863 My system at the time was Interactive 2.2, a SVR3.2/386 derivative that
864 also had some POSIX support. Metaconfig-generated Configure scripts
865 worked fine for me on that system. On the other hand, autoconf-generated
866 scripts usually didn't. (They did come quite close, though, in some
867 cases.) At the time, I actually fetched a large number of GNU packages
868 and checked. Not a single one configured and compiled correctly
869 out-of-the-box with the system's cc compiler.
871 =item Configure can be interactive
873 With both autoconf and metaconfig, if the script works, everything is
874 fine. However, one of my main problems with autoconf-generated scripts
875 was that if it guessed wrong about something, it could be B<very> hard to
876 go back and fix it. For example, autoconf always insisted on passing the
877 -Xp flag to cc (to turn on POSIX behavior), even when that wasn't what I
878 wanted or needed for that package. There was no way short of editing the
879 configure script to turn this off. You couldn't just edit the resulting
880 Makefile at the end because the -Xp flag influenced a number of other
883 Metaconfig's Configure scripts, on the other hand, can be interactive.
884 Thus if Configure is guessing things incorrectly, you can go back and fix
885 them. This isn't as important now as it was when we were actively
886 developing Configure support for new features such as dynamic loading,
887 but it's still useful occasionally.
891 At the time, autoconf-generated scripts were covered under the GNU Public
892 License, and hence weren't suitable for inclusion with Perl, which has a
893 different licensing policy. (Autoconf's licensing has since changed.)
897 Metaconfig builds up Configure from a collection of discrete pieces
898 called "units". You can override the standard behavior by supplying your
899 own unit. With autoconf, you have to patch the standard files instead.
900 I find the metaconfig "unit" method easier to work with. Others
901 may find metaconfig's units clumsy to work with.
905 =head2 @INC search order
907 By default, the list of perl library directories in @INC is the
915 Specifically, on my Solaris/x86 system, I run
916 B<sh Configure -Dprefix=/opt/perl> and I have the following
919 /opt/perl/lib/i86pc-solaris/5.00307
921 /opt/perl/lib/site_perl/i86pc-solaris
922 /opt/perl/lib/site_perl
924 That is, perl's directories come first, followed by the site-specific
927 The site libraries come second to support the usage of extensions
928 across perl versions. Read the relevant section in F<INSTALL> for
929 more information. If we ever make $sitearch version-specific, this
930 topic could be revisited.
932 =head2 Why isn't there a directory to override Perl's library?
934 Mainly because no one's gotten around to making one. Note that
935 "making one" involves changing perl.c, Configure, config_h.SH (and
936 associated files, see above), and I<documenting> it all in the
939 Apparently, most folks who want to override one of the standard library
940 files simply do it by overwriting the standard library files.
944 In the perl.c sources, you'll find an undocumented APPLLIB_EXP
945 variable, sort of like PRIVLIB_EXP and ARCHLIB_EXP (which are
946 documented in config_h.SH). Here's what APPLLIB_EXP is for, from
947 a mail message from Larry:
949 The main intent of APPLLIB_EXP is for folks who want to send out a
950 version of Perl embedded in their product. They would set the symbol
951 to be the name of the library containing the files needed to run or to
952 support their particular application. This works at the "override"
953 level to make sure they get their own versions of any library code that
954 they absolutely must have configuration control over.
956 As such, I don't see any conflict with a sysadmin using it for a
957 override-ish sort of thing, when installing a generic Perl. It should
958 probably have been named something to do with overriding though. Since
959 it's undocumented we could still change it... :-)
961 Given that it's already there, you can use it to override
962 distribution modules. If you do
964 sh Configure -Dccflags='-DAPPLLIB_EXP=/my/override'
966 then perl.c will put /my/override ahead of ARCHLIB and PRIVLIB.
968 =head1 Upload Your Work to CPAN
970 You can upload your work to CPAN if you have a CPAN id. Check out
971 http://www.perl.com/CPAN/modules/04pause.html for information on
972 _PAUSE_, the Perl Author's Upload Server.
974 I typically upload both the patch file, e.g. F<perl5.004_08.pat.gz>
975 and the full tar file, e.g. F<perl5.004_08.tar.gz>.
977 If you want your patch to appear in the F<src/5.0/unsupported>
978 directory on CPAN, send e-mail to the CPAN master librarian. (Check
979 out http://www.perl.com/CPAN/CPAN.html).
981 =head1 Help Save the World
983 You should definitely announce your patch on the perl5-porters list.
984 You should also consider announcing your patch on
985 comp.lang.perl.announce, though you should make it quite clear that a
986 subversion is not a production release, and be prepared to deal with
987 people who will not read your disclaimer.
991 Here, in no particular order, are some Configure and build-related
992 items that merit consideration. This list isn't exhaustive, it's just
993 what I came up with off the top of my head.
995 =head2 Good ideas waiting for round tuits
1001 I think we ought to support
1003 Configure -Dinstallprefix=/blah/blah
1005 Currently, we support B<-Dprefix=/blah/blah>, but the changing the install
1006 location has to be handled by something like the F<config.over> trick
1007 described in F<INSTALL>. AFS users also are treated specially.
1008 We should probably duplicate the metaconfig prefix stuff for an
1011 =item Configure -Dsrcdir=/blah/blah
1013 We should be able to emulate B<configure --srcdir>. Tom Tromey
1014 tromey@creche.cygnus.com has submitted some patches to
1015 the dist-users mailing list along these lines. Eventually, they ought
1016 to get folded back into the main distribution.
1018 =item Hint file fixes
1020 Various hint files work around Configure problems. We ought to fix
1021 Configure so that most of them aren't needed.
1023 =item Hint file information
1025 Some of the hint file information (particularly dynamic loading stuff)
1026 ought to be fed back into the main metaconfig distribution.
1030 =head2 Probably good ideas waiting for round tuits
1034 =item GNU configure --options
1036 I've received sensible suggestions for --exec_prefix and other
1037 GNU configure --options. It's not always obvious exactly what is
1038 intended, but this merits investigation.
1042 Currently, B<make clean> isn't all that useful, though
1043 B<make realclean> and B<make distclean> are. This needs a bit of
1044 thought and documentation before it gets cleaned up.
1046 =item Try gcc if cc fails
1048 Currently, we just give up.
1050 =item bypassing safe*alloc wrappers
1052 On some systems, it may be safe to call the system malloc directly
1053 without going through the util.c safe* layers. (Such systems would
1054 accept free(0), for example.) This might be a time-saver for systems
1055 that already have a good malloc. (Recent Linux libc's apparently have
1056 a nice malloc that is well-tuned for the system.)
1060 =head2 Vague possibilities
1064 =item Win95, WinNT, and Win32 support
1066 We need to get something into the distribution for 32-bit Windows.
1067 I'm tired of all the private e-mail questions I get, and I'm saddened
1068 that so many folks keep trying to reinvent the same wheel.
1072 Get some of the Macintosh stuff folded back into the main
1075 =item gconvert replacement
1077 Maybe include a replacement function that doesn't lose data in rare
1078 cases of coercion between string and numerical values.
1082 Can we support C<long long> on systems where C<long long> is larger
1083 than what we've been using for C<IV>? What if you can't C<sprintf>
1086 =item Improve makedepend
1088 The current makedepend process is clunky and annoyingly slow, but it
1089 works for most folks. Alas, it assumes that there is a filename
1090 $firstmakefile that the B<make> command will try to use before it uses
1091 F<Makefile>. Such may not be the case for all B<make> commands,
1092 particularly those on non-Unix systems.
1094 Probably some variant of the BSD F<.depend> file will be useful.
1095 We ought to check how other packages do this, if they do it at all.
1096 We could probably pre-generate the dependencies (with the exception of
1097 malloc.o, which could probably be determined at F<Makefile.SH>
1100 =item GNU Makefile standard targets
1102 GNU software generally has standardized Makefile targets. Unless we
1103 have good reason to do otherwise, I see no reason not to support them.
1107 Somehow, straighten out, document, and implement lockf(), flock(),
1108 and/or fcntl() file locking. It's a mess.
1114 Andy Dougherty <doughera@lafcol.lafayette.edu>.
1116 Additions by Chip Salzenberg <chip@atlantic.net>.
1118 All opinions expressed herein are those of the authorZ<>(s).
1120 =head1 LAST MODIFIED
1122 $Id: pumpkin.pod,v 1.8 1997/02/18 18:19:20 chip Released $