This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Update metaconfig info
[perl5.git] / Porting / pumpkin.pod
CommitLineData
aa689395
PP
1=head1 NAME
2
3Pumpkin - Notes on handling the Perl Patch Pumpkin
4
5=head1 SYNOPSIS
6
7There is no simple synopsis, yet.
8
9=head1 DESCRIPTION
10
11This document attempts to begin to describe some of the
12considerations involved in patching and maintaining perl.
13
14This document is still under construction, and still subject to
15significant changes. Still, I hope parts of it will be useful,
16so I'm releasing it even though it's not done.
17
18For the most part, it's a collection of anecdotal information that
19already assumes some familiarity with the Perl sources. I really need
20an introductory section that describes the organization of the sources
21and all the various auxiliary files that are part of the distribution.
22
23=head1 Where Do I Get Perl Sources and Related Material?
24
25The Comprehensive Perl Archive Network (or CPAN) is the place to go.
26There are many mirrors, but the easiest thing to use is probably
7b5757d1 27http://www.perl.com/CPAN/README.html , which automatically points you to a
aa689395
PP
28mirror site "close" to you.
29
30=head2 Perl5-porters mailing list
31
32The mailing list perl5-porters@perl.org
33is the main group working with the development of perl. If you're
34interested in all the latest developments, you should definitely
35subscribe. The list is high volume, but generally has a
36fairly low noise level.
37
38Subscribe by sending the message (in the body of your letter)
39
40 subscribe perl5-porters
41
42to perl5-porters-request@perl.org .
43
fb73857a
PP
44Archives of the list are held at:
45
46 http://www.rosat.mpe-garching.mpg.de/mailing-lists/perl-porters/
47
aa689395
PP
48=head1 How are Perl Releases Numbered?
49
7b5757d1
AD
50Perl version numbers are floating point numbers, such as 5.004.
51(Observations about the imprecision of floating point numbers for
52representing reality probably have more relevance than you might
53imagine :-) The major version number is 5 and the '004' is the
54patchlevel. (Questions such as whether or not '004' is really a minor
55version number can safely be ignored.:)
56
57The version number is available as the magic variable $],
aa689395
PP
58and can be used in comparisons, e.g.
59
60 print "You've got an old perl\n" if $] < 5.002;
61
aa689395
PP
62You can also require particular version (or later) with
63
64 use 5.002;
65
7b5757d1
AD
66At some point in the future, we may need to decide what to call the
67next big revision. In the .package file used by metaconfig to
68generate Configure, there are two variables that might be relevant:
69$baserev=5.0 and $package=perl5. At various times, I have suggested
70we might change them to $baserev=5.1 and $package=perl5.1 if want
71to signify a fairly major update. Or, we might want to jump to perl6.
72Let's worry about that problem when we get there.
73
aa689395
PP
74=head2 Subversions
75
76In addition, there may be "developer" sub-versions available. These
77are not official releases. They may contain unstable experimental
78features, and are subject to rapid change. Such developer
79sub-versions are numbered with sub-version numbers. For example,
fb73857a
PP
80version 5.003_04 is the 4'th developer version built on top of
815.003. It might include the _01, _02, and _03 changes, but it
82also might not. Sub-versions are allowed to be subversive. (But see
83the next section for recent changes.)
aa689395
PP
84
85These sub-versions can also be used as floating point numbers, so
86you can do things such as
87
7b5757d1 88 print "You've got an unstable perl\n" if $] == 5.00303;
aa689395
PP
89
90You can also require particular version (or later) with
91
7b5757d1 92 use 5.003_03; # the "_" is optional
aa689395
PP
93
94Sub-versions produced by the members of perl5-porters are usually
95available on CPAN in the F<src/5.0/unsupported> directory.
96
7b5757d1
AD
97=head2 Maintenance and Development Subversions
98
99As 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.
102
103The separate bug-fix track is being established to allow us an easy
104way to distribute important bug fixes without waiting for the
105developers to untangle all the other problems in the current
106developer's release.
107
fb73857a
PP
108Trial releases of bug-fix maintenance releases are announced on
109perl5-porters. Trial releases use the new subversion number (to avoid
110testers installing it over the previous release) and include a 'local
111patch' entry in patchlevel.h.
112
7b5757d1
AD
113Watch for announcements of maintenance subversions in
114comp.lang.perl.announce.
115
aa689395
PP
116=head2 Why such a complicated scheme?
117
118Two reasons, really. At least.
119
7b5757d1
AD
120First, we need some way to identify and release collections of patches
121that are known to have new features that need testing and exploration. The
aa689395
PP
122subversion scheme does that nicely while fitting into the
123C<use 5.004;> mold.
124
125Second, since most of the folks who help maintain perl do so on a
126free-time voluntary basis, perl development does not proceed at a
127precise pace, though it always seems to be moving ahead quickly.
128We needed some way to pass around the "patch pumpkin" to allow
129different people chances to work on different aspects of the
130distribution without getting in each other's way. It wouldn't be
131constructive to have multiple people working on incompatible
132implementations of the same idea. Instead what was needed was
133some kind of "baton" or "token" to pass around so everyone knew
134whose turn was next.
135
136=head2 Why is it called the patch pumpkin?
137
138Chip Salzenberg gets credit for that, with a nod to his cow orker,
139David Croy. We had passed around various names (baton, token, hot
140potato) but none caught on. Then, Chip asked:
141
142[begin quote]
143
144 Who has the patch pumpkin?
145
146To explain: David Croy once told me once that at a previous job,
147there was one tape drive and multiple systems that used it for backups.
148But instead of some high-tech exclusion software, they used a low-tech
149method to prevent multiple simultaneous backups: a stuffed pumpkin.
150No one was allowed to make backups unless they had the "backup pumpkin".
151
152[end quote]
153
154The name has stuck.
155
156=head1 Philosophical Issues in Patching Perl
157
158There are no absolute rules, but there are some general guidelines I
159have tried to follow as I apply patches to the perl sources.
160(This section is still under construction.)
161
162=head2 Solve problems as generally as possible
163
7b5757d1
AD
164Never implement a specific restricted solution to a problem when you
165can solve the same problem in a more general, flexible way.
166
167For example, for dynamic loading to work on some SVR4 systems, we had
168to build a shared libperl.so library. In order to build "FAT" binaries
169on NeXT 4.0 systems, we had to build a special libperl library. Rather
170than continuing to build a contorted nest of special cases, I
171generalized the process of building libperl so that NeXT and SVR4 users
172could still get their work done, but others could build a shared
173libperl if they wanted to as well.
aa689395
PP
174
175=head2 Seek consensus on major changes
176
177If you are making big changes, don't do it in secret. Discuss the
178ideas in advance on perl5-porters.
179
180=head2 Keep the documentation up-to-date
181
182If your changes may affect how users use perl, then check to be sure
183that the documentation is in sync with your changes. Be sure to
184check all the files F<pod/*.pod> and also the F<INSTALL> document.
185
186Consider writing the appropriate documentation first and then
7b5757d1 187implementing your change to correspond to the documentation.
aa689395
PP
188
189=head2 Avoid machine-specific #ifdef's
190
191To the extent reasonable, try to avoid machine-specific #ifdef's in
192the sources. Instead, use feature-specific #ifdef's. The reason is
193that the machine-specific #ifdef's may not be valid across major
194releases of the operating system. Further, the feature-specific tests
195may help out folks on another platform who have the same problem.
196
197=head2 Allow for lots of testing
198
199We should never release a main version without testing it as a
200subversion first.
201
6877a1cf
AD
202=head2 Test popular applications and modules.
203
204We should never release a main version without testing whether or not
205it breaks various popular modules and applications. A partial list of
206such things would include majordomo, metaconfig, apache, Tk, CGI,
207libnet, and libwww, to name just a few. Of course it's quite possible
208that some of those things will be just plain broken and need to be fixed,
209but, in general, we ought to try to avoid breaking widely-installed
210things.
211
7b5757d1 212=head2 Automate generation of derivative files
aa689395
PP
213
214The F<embed.h>, F<keywords.h>, F<opcode.h>, and F<perltoc.pod> files
215are all automatically generated by perl scripts. In general, don't
216patch these directly; patch the data files instead.
217
218F<Configure> and F<config_h.SH> are also automatically generated by
219B<metaconfig>. In general, you should patch the metaconfig units
dfe9444c 220instead of patching these files directly. However, very minor changes to
aa689395 221F<Configure> may be made in between major sync-ups with the metaconfig
dfe9444c
AD
222units, which tends to be complicated operations. But be careful, this
223can quickly spiral out of control. Running metaconfig is not really
224hard.
aa689395 225
a8119d38
AD
226Finally, the sample files in the F<Porting/> subdirectory are
227generated automatically by the script F<U/mksample> included
228with the metaconfig units. See L<"run metaconfig"> below for
229information on obtaining the metaconfig units.
230
aa689395
PP
231=head1 How to Make a Distribution
232
233There really ought to be a 'make dist' target, but there isn't.
234The 'dist' suite of tools also contains a number of tools that I haven't
235learned how to use yet. Some of them may make this all a bit easier.
236
237Here are the steps I go through to prepare a patch & distribution.
238
3e3baf6d
TB
239Lots of it could doubtless be automated but isn't. The Porting/makerel
240(make release) perl script does now help automate some parts of it.
aa689395
PP
241
242=head2 Announce your intentions
243
244First, you should volunteer out loud to take the patch pumpkin. It's
245generally counter-productive to have multiple people working in secret
246on the same thing.
247
248At the same time, announce what you plan to do with the patch pumpkin,
249to allow folks a chance to object or suggest alternatives, or do it for
250you. Naturally, the patch pumpkin holder ought to incorporate various
251bug fixes and documentation improvements that are posted while he or
252she has the pumpkin, but there might also be larger issues at stake.
253
254One of the precepts of the subversion idea is that we shouldn't give
7b5757d1
AD
255the patch pumpkin to anyone unless we have some idea what he or she
256is going to do with it.
aa689395
PP
257
258=head2 refresh pod/perltoc.pod
259
260Presumably, you have done a full C<make> in your working source
261directory. Before you C<make spotless> (if you do), and if you have
262changed any documentation in any module or pod file, change to the
263F<pod> directory and run C<make toc>.
264
3e3baf6d
TB
265=head2 run installhtml to check the validity of the pod files
266
aa689395
PP
267=head2 update patchlevel.h
268
269Don't be shy about using the subversion number, even for a relatively
270modest patch. We've never even come close to using all 99 subversions,
271and it's better to have a distinctive number for your patch. If you
272need feedback on your patch, go ahead and issue it and promise to
273incorporate that feedback quickly (e.g. within 1 week) and send out a
274second patch.
275
276=head2 run metaconfig
277
278If you need to make changes to Configure or config_h.SH, it may be best to
279change the appropriate metaconfig units instead, and regenerate Configure.
280
281 metaconfig -m
282
283will regenerate Configure and config_h.SH. More information on
284obtaining and running metaconfig is in the F<U/README> file that comes
285with Perl's metaconfig units. Perl's metaconfig units should be
286available the same place you found this file. On CPAN, look under my
a8119d38
AD
287directory F<authors/id/ANDYD/> for a file such as
288F<mc_units-5.004_70-01.tar.gz>.
aa689395
PP
289That file should be unpacked in your main perl source directory. It
290contains the files needed to run B<metaconfig> to reproduce Perl's
a8119d38 291Configure script. (Those units are for 5.004_70. There may have been
7b5757d1
AD
292changes since then; please contact me if you want more recent
293versions, and I will try to point you in the right direction.)
aa689395
PP
294
295Alternatively, do consider if the F<*ish.h> files might be a better
296place for your changes.
297
298=head2 MANIFEST
299
300Make sure the MANIFEST is up-to-date. You can use dist's B<manicheck>
301program for this. You can also use
302
3e3baf6d 303 perl -w -MExtUtils::Manifest=fullcheck -e fullcheck
aa689395 304
3e3baf6d
TB
305Both commands will also list extra files in the directory that are not
306listed in MANIFEST.
aa689395 307
bfb7748a 308The MANIFEST is normally sorted.
aa689395
PP
309
310If you are using metaconfig to regenerate Configure, then you should note
311that metaconfig actually uses MANIFEST.new, so you want to be sure
312MANIFEST.new is up-to-date too. I haven't found the MANIFEST/MANIFEST.new
313distinction particularly useful, but that's probably because I still haven't
314learned how to use the full suite of tools in the dist distribution.
315
316=head2 Check permissions
317
318All the tests in the t/ directory ought to be executable. The
319main makefile used to do a 'chmod t/*/*.t', but that resulted in
320a self-modifying distribution--something some users would strongly
321prefer to avoid. Probably, the F<t/TEST> script should check for this
322and do the chmod if needed, but it doesn't currently.
323
324In all, the following files should probably be executable:
325
326 Configure
327 configpm
32fcaa0b 328 configure.gnu
aa689395
PP
329 embed.pl
330 installperl
331 installman
332 keywords.pl
aa689395
PP
333 myconfig
334 opcode.pl
335 perly.fixer
336 t/TEST
337 t/*/*.t
338 *.SH
339 vms/ext/Stdio/test.pl
340 vms/ext/filespec.t
aa689395
PP
341 x2p/*.SH
342
343Other things ought to be readable, at least :-).
344
345Probably, the permissions for the files could be encoded in MANIFEST
346somehow, but I'm reluctant to change MANIFEST itself because that
347could break old scripts that use MANIFEST.
348
349I seem to recall that some SVR3 systems kept some sort of file that listed
350permissions for system files; something like that might be appropriate.
351
352=head2 Run Configure
353
354This will build a config.sh and config.h. You can skip this if you haven't
693762b4 355changed Configure or config_h.SH at all. I use the following command
aa689395 356
693762b4
AD
357 sh Configure -Dprefix=/opt/perl -Doptimize=-O -Dusethreads \
358 -Dcf_by='yourname' \
359 -Dcf_email='yourname@yourhost.yourplace.com' \
360 -Dperladmin='yourname@yourhost.yourplace.com' \
361 -Dmydomain='.yourplace.com' \
362 -Dmyhostname='yourhost' \
363 -des
aa689395 364
693762b4 365=head2 Update Porting/config.sh and Porting/config_H
dfe9444c 366
693762b4
AD
367[XXX
368This section needs revision. We're currently working on easing
369the task of keeping the vms, win32, and plan9 config.sh info
370up-to-date. The plan is to use keep up-to-date 'canned' config.sh
371files in the appropriate subdirectories and then generate 'canned'
372config.h files for vms, win32, etc. from the generic config.sh file.
373This is to ease maintenance. When Configure gets updated, the parts
374sometimes get scrambled around, and the changes in config_H can
375sometimes be very hard to follow. config.sh, on the other hand, can
376safely be sorted, so it's easy to track (typically very small) changes
377to config.sh and then propoagate them to a canned 'config.h' by any
378number of means, including a perl script in win32/ or carrying
379config.sh and config_h.SH to a Unix system and running sh
380config_h.SH.)
381XXX]
382
383The Porting/config.sh and Porting/config_H files are provided to
384help those folks who can't run Configure. It is important to keep
385them up-to-date. If you have changed config_h.SH, those changes must
386be reflected in config_H as well. (The name config_H was chosen to
387distinguish the file from config.h even on case-insensitive file systems.)
388Simply edit the existing config_H file; keep the first few explanatory
389lines and then copy your new config.h below.
aa689395
PP
390
391It may also be necessary to update vms/config.vms and
392plan9/config.plan9, though you should be quite careful in doing so if
393you are not familiar with those systems. You might want to issue your
394patch with a promise to quickly issue a follow-up that handles those
395directories.
396
397=head2 make run_byacc
398
399If you have byacc-1.8.2 (available from CPAN), and if there have been
400changes to F<perly.y>, you can regenerate the F<perly.c> file. The
401run_byacc makefile target does this by running byacc and then applying
402some patches so that byacc dynamically allocates space, rather than
403having fixed limits. This patch is handled by the F<perly.fixer>
404script. Depending on the nature of the changes to F<perly.y>, you may
405or may not have to hand-edit the patch to apply correctly. If you do,
406you should include the edited patch in the new distribution. If you
407have byacc-1.9, the patch won't apply cleanly. Changes to the printf
408output statements mean the patch won't apply cleanly. Long ago I
409started to fix F<perly.fixer> to detect this, but I never completed the
410task.
411
412Some additional notes from Larry on this:
413
414Don't forget to regenerate perly.c.diff.
415
7b5757d1 416 byacc -d perly.y
aa689395
PP
417 mv y.tab.c perly.c
418 patch perly.c <perly.c.diff
419 # manually apply any failed hunks
420 diff -c2 perly.c.orig perly.c >perly.c.diff
421
422One chunk of lines that often fails begins with
423
424 #line 29 "perly.y"
425
426and ends one line before
427
428 #define YYERRCODE 256
429
430This only happens when you add or remove a token type. I suppose this
431could be automated, but it doesn't happen very often nowadays.
432
433Larry
434
435=head2 make regen_headers
436
437The F<embed.h>, F<keywords.h>, and F<opcode.h> files are all automatically
438generated by perl scripts. Since the user isn't guaranteed to have a
439working perl, we can't require the user to generate them. Hence you have
440to, if you're making a distribution.
441
442I used to include rules like the following in the makefile:
443
444 # The following three header files are generated automatically
445 # The correct versions should be already supplied with the perl kit,
446 # in case you don't have perl or 'sh' available.
447 # The - is to ignore error return codes in case you have the source
448 # installed read-only or you don't have perl yet.
449 keywords.h: keywords.pl
450 @echo "Don't worry if this fails."
451 - perl keywords.pl
452
453
7b5757d1 454However, I got B<lots> of mail consisting of people worrying because the
aa689395
PP
455command failed. I eventually decided that I would save myself time
456and effort by manually running C<make regen_headers> myself rather
457than answering all the questions and complaints about the failing
458command.
459
3e3baf6d 460=head2 global.sym, interp.sym and perlio.sym
aa689395
PP
461
462Make sure these files are up-to-date. Read the comments in these
463files and in perl_exp.SH to see what to do.
464
465=head2 Binary compatibility
466
467If you do change F<global.sym> or F<interp.sym>, think carefully about
468what you are doing. To the extent reasonable, we'd like to maintain
469souce and binary compatibility with older releases of perl. That way,
470extensions built under one version of perl will continue to work with
471new versions of perl.
472
473Of course, some incompatible changes may well be necessary. I'm just
474suggesting that we not make any such changes without thinking carefully
475about them first. If possible, we should provide
476backwards-compatibility stubs. There's a lot of XS code out there.
477Let's not force people to keep changing it.
478
479=head2 Changes
480
481Be sure to update the F<Changes> file. Try to include both an overall
482summary as well as detailed descriptions of the changes. Your
3e3baf6d 483audience will include other developers and users, so describe
aa689395
PP
484user-visible changes (if any) in terms they will understand, not in
485code like "initialize foo variable in bar function".
486
487There are differing opinions on whether the detailed descriptions
488ought to go in the Changes file or whether they ought to be available
489separately in the patch file (or both). There is no disagreement that
490detailed descriptions ought to be easily available somewhere.
491
2a26e2f1
DD
492=head2 Todo
493
494The F<Todo> file contains a roughly-catgorized unordered list of
495aspects of Perl that could use enhancement, features that could be
496added, areas that could be cleaned up, and so on. During your term as
497pumpkin-holder, you will probably address some of these issues, and
498perhaps identify others which, while you decide not to address them
499this time around, may be tackled in the future. Update the file
500reflect the situation as it stands when you hand over the pumpkin.
501
502You might like, early in your pumpkin-holding career, to see if you
503can find champions for partiticular issues on the to-do list: an issue
504owned is an issue more likely to be resolved.
505
c4f23d77
AD
506There are also some more porting-specific L<Todo> items later in this
507file.
508
aa689395
PP
509=head2 OS/2-specific updates
510
511In the os2 directory is F<diff.configure>, a set of OS/2-specific
512diffs against B<Configure>. If you make changes to Configure, you may
513want to consider regenerating this diff file to save trouble for the
514OS/2 maintainer.
515
7b5757d1
AD
516You can also consider the OS/2 diffs as reminders of portability
517things that need to be fixed in Configure.
518
aa689395
PP
519=head2 VMS-specific updates
520
521If you have changed F<perly.y>, then you may want to update
522F<vms/perly_{h,c}.vms> by running C<perl vms/vms_yfix.pl>.
523
524The Perl version number appears in several places under F<vms>.
525It is courteous to update these versions. For example, if you are
526making 5.004_42, replace "5.00441" with "5.00442".
527
528=head2 Making the new distribution
529
530Suppose, for example, that you want to make version 5.004_08. Then you can
531do something like the following
532
533 mkdir ../perl5.004_08
534 awk '{print $1}' MANIFEST | cpio -pdm ../perl5.004_08
535 cd ../
536 tar cf perl5.004_08.tar perl5.004_08
537 gzip --best perl5.004_08.tar
538
3e3baf6d
TB
539These steps, with extra checks, are automated by the Porting/makerel
540script.
541
aa689395
PP
542=head2 Making a new patch
543
544I find the F<makepatch> utility quite handy for making patches.
545You can obtain it from any CPAN archive under
3e3baf6d
TB
546http://www.perl.com/CPAN/authors/Johan_Vromans/ . There are a couple
547of differences between my version and the standard one. I have mine do
548a
aa689395
PP
549
550 # Print a reassuring "End of Patch" note so people won't
551 # wonder if their mailer truncated patches.
552 print "\n\nEnd of Patch.\n";
553
3e3baf6d
TB
554at the end. That's because I used to get questions from people asking
555if their mail was truncated.
556
557It also writes Index: lines which include the new directory prefix
558(change Index: print, approx line 294 or 310 depending on the version,
559to read: print PATCH ("Index: $newdir$new\n");). That helps patches
560work with more POSIX conformant patch programs.
aa689395
PP
561
562Here's how I generate a new patch. I'll use the hypothetical
5635.004_07 to 5.004_08 patch as an example.
564
565 # unpack perl5.004_07/
566 gzip -d -c perl5.004_07.tar.gz | tar -xof -
567 # unpack perl5.004_08/
568 gzip -d -c perl5.004_08.tar.gz | tar -xof -
569 makepatch perl5.004_07 perl5.004_08 > perl5.004_08.pat
570
571Makepatch will automatically generate appropriate B<rm> commands to remove
572deleted files. Unfortunately, it will not correctly set permissions
573for newly created files, so you may have to do so manually. For example,
574patch 5.003_04 created a new test F<t/op/gv.t> which needs to be executable,
575so at the top of the patch, I inserted the following lines:
576
577 # Make a new test
578 touch t/op/gv.t
579 chmod +x t/opt/gv.t
580
581Now, of course, my patch is now wrong because makepatch didn't know I
582was going to do that command, and it patched against /dev/null.
583
584So, what I do is sort out all such shell commands that need to be in the
585patch (including possible mv-ing of files, if needed) and put that in the
586shell commands at the top of the patch. Next, I delete all the patch parts
587of perl5.004_08.pat, leaving just the shell commands. Then, I do the
588following:
589
7b5757d1
AD
590 cd perl5.004_07
591 sh ../perl5.004_08.pat
aa689395 592 cd ..
7b5757d1 593 makepatch perl5.004_07 perl5.004_08 >> perl5.004_08.pat
aa689395
PP
594
595(Note the append to preserve my shell commands.)
596Now, my patch will line up with what the end users are going to do.
597
598=head2 Testing your patch
599
600It seems obvious, but be sure to test your patch. That is, verify that
601it produces exactly the same thing as your full distribution.
602
7b5757d1
AD
603 rm -rf perl5.004_07
604 gzip -d -c perl5.004_07.tar.gz | tar -xf -
605 cd perl5.004_07
606 sh ../perl5.004_08.pat
607 patch -p1 -N < ../perl5.004_08.pat
aa689395 608 cd ..
7b5757d1 609 gdiff -r perl5.004_07 perl5.004_08
aa689395
PP
610
611where B<gdiff> is GNU diff. Other diff's may also do recursive checking.
612
613=head2 More testing
614
615Again, it's obvious, but you should test your new version as widely as you
616can. You can be sure you'll hear about it quickly if your version doesn't
617work on both ANSI and pre-ANSI compilers, and on common systems such as
618SunOS 4.1.[34], Solaris, and Linux.
619
620If your changes include conditional code, try to test the different
621branches as thoroughly as you can. For example, if your system
622supports dynamic loading, you can also test static loading with
623
624 sh Configure -Uusedl
625
626You can also hand-tweak your config.h to try out different #ifdef
627branches.
628
629=head1 Common Gotcha's
630
631=over 4
632
633=item #elif
634
635The '#elif' preprocessor directive is not understood on all systems.
636Specifically, I know that Pyramids don't understand it. Thus instead of the
637simple
638
639 #if defined(I_FOO)
640 # include <foo.h>
641 #elif defined(I_BAR)
642 # include <bar.h>
643 #else
644 # include <fubar.h>
645 #endif
646
647You have to do the more Byzantine
648
649 #if defined(I_FOO)
650 # include <foo.h>
651 #else
652 # if defined(I_BAR)
653 # include <bar.h>
654 # else
655 # include <fubar.h>
656 # endif
657 #endif
658
659Incidentally, whitespace between the leading '#' and the preprocessor
660command is not guaranteed, but is very portable and you may use it freely.
661I think it makes things a bit more readable, especially once things get
662rather deeply nested. I also think that things should almost never get
663too deeply nested, so it ought to be a moot point :-)
664
665=item Probably Prefer POSIX
666
667It's often the case that you'll need to choose whether to do
668something the BSD-ish way or the POSIX-ish way. It's usually not
669a big problem when the two systems use different names for similar
670functions, such as memcmp() and bcmp(). The perl.h header file
671handles these by appropriate #defines, selecting the POSIX mem*()
672functions if available, but falling back on the b*() functions, if
673need be.
674
675More serious is the case where some brilliant person decided to
676use the same function name but give it a different meaning or
677calling sequence :-). getpgrp() and setpgrp() come to mind.
678These are a real problem on systems that aim for conformance to
679one standard (e.g. POSIX), but still try to support the other way
680of doing things (e.g. BSD). My general advice (still not really
681implemented in the source) is to do something like the following.
682Suppose there are two alternative versions, fooPOSIX() and
683fooBSD().
684
685 #ifdef HAS_FOOPOSIX
686 /* use fooPOSIX(); */
687 #else
688 # ifdef HAS_FOOBSD
689 /* try to emulate fooPOSIX() with fooBSD();
690 perhaps with the following: */
691 # define fooPOSIX fooBSD
692 # else
693 # /* Uh, oh. We have to supply our own. */
694 # define fooPOSIX Perl_fooPOSIX
695 # endif
696 #endif
697
698=item Think positively
699
700If you need to add an #ifdef test, it is usually easier to follow if you
701think positively, e.g.
702
703 #ifdef HAS_NEATO_FEATURE
704 /* use neato feature */
705 #else
706 /* use some fallback mechanism */
707 #endif
708
709rather than the more impenetrable
710
711 #ifndef MISSING_NEATO_FEATURE
712 /* Not missing it, so we must have it, so use it */
713 #else
714 /* Are missing it, so fall back on something else. */
715 #endif
716
717Of course for this toy example, there's not much difference. But when
718the #ifdef's start spanning a couple of screen fulls, and the #else's
719are marked something like
720
721 #else /* !MISSING_NEATO_FEATURE */
722
723I find it easy to get lost.
724
725=item Providing Missing Functions -- Problem
726
727Not all systems have all the neat functions you might want or need, so
728you might decide to be helpful and provide an emulation. This is
729sound in theory and very kind of you, but please be careful about what
730you name the function. Let me use the C<pause()> function as an
731illustration.
732
733Perl5.003 has the following in F<perl.h>
734
735 #ifndef HAS_PAUSE
736 #define pause() sleep((32767<<16)+32767)
737 #endif
738
739Configure sets HAS_PAUSE if the system has the pause() function, so
740this #define only kicks in if the pause() function is missing.
741Nice idea, right?
742
743Unfortunately, some systems apparently have a prototype for pause()
744in F<unistd.h>, but don't actually have the function in the library.
745(Or maybe they do have it in a library we're not using.)
746
747Thus, the compiler sees something like
748
749 extern int pause(void);
750 /* . . . */
751 #define pause() sleep((32767<<16)+32767)
752
753and dies with an error message. (Some compilers don't mind this;
754others apparently do.)
755
756To work around this, 5.003_03 and later have the following in perl.h:
757
758 /* Some unistd.h's give a prototype for pause() even though
759 HAS_PAUSE ends up undefined. This causes the #define
760 below to be rejected by the compiler. Sigh.
761 */
762 #ifdef HAS_PAUSE
763 # define Pause pause
764 #else
765 # define Pause() sleep((32767<<16)+32767)
766 #endif
767
768This works.
769
770The curious reader may wonder why I didn't do the following in
771F<util.c> instead:
772
773 #ifndef HAS_PAUSE
774 void pause()
775 {
776 sleep((32767<<16)+32767);
777 }
778 #endif
779
780That is, since the function is missing, just provide it.
781Then things would probably be been alright, it would seem.
782
783Well, almost. It could be made to work. The problem arises from the
784conflicting needs of dynamic loading and namespace protection.
785
786For dynamic loading to work on AIX (and VMS) we need to provide a list
787of symbols to be exported. This is done by the script F<perl_exp.SH>,
788which reads F<global.sym> and F<interp.sym>. Thus, the C<pause>
789symbol would have to be added to F<global.sym> So far, so good.
790
791On the other hand, one of the goals of Perl5 is to make it easy to
792either extend or embed perl and link it with other libraries. This
793means we have to be careful to keep the visible namespace "clean".
794That is, we don't want perl's global variables to conflict with
795those in the other application library. Although this work is still
796in progress, the way it is currently done is via the F<embed.h> file.
797This file is built from the F<global.sym> and F<interp.sym> files,
798since those files already list the globally visible symbols. If we
799had added C<pause> to global.sym, then F<embed.h> would contain the
800line
801
802 #define pause Perl_pause
803
804and calls to C<pause> in the perl sources would now point to
805C<Perl_pause>. Now, when B<ld> is run to build the F<perl> executable,
806it will go looking for C<perl_pause>, which probably won't exist in any
807of the standard libraries. Thus the build of perl will fail.
808
809Those systems where C<HAS_PAUSE> is not defined would be ok, however,
810since they would get a C<Perl_pause> function in util.c. The rest of
811the world would be in trouble.
812
813And yes, this scenario has happened. On SCO, the function C<chsize>
814is available. (I think it's in F<-lx>, the Xenix compatibility
815library.) Since the perl4 days (and possibly before), Perl has
816included a C<chsize> function that gets called something akin to
817
818 #ifndef HAS_CHSIZE
819 I32 chsize(fd, length)
820 /* . . . */
821 #endif
822
823When 5.003 added
824
825 #define chsize Perl_chsize
826
827to F<embed.h>, the compile started failing on SCO systems.
828
829The "fix" is to give the function a different name. The one
830implemented in 5.003_05 isn't optimal, but here's what was done:
831
832 #ifdef HAS_CHSIZE
833 # ifdef my_chsize /* Probably #defined to Perl_my_chsize in embed.h */
834 # undef my_chsize
835 # endif
836 # define my_chsize chsize
837 #endif
838
839My explanatory comment in patch 5.003_05 said:
840
841 Undef and then re-define my_chsize from Perl_my_chsize to
842 just plain chsize if this system HAS_CHSIZE. This probably only
843 applies to SCO. This shows the perils of having internal
844 functions with the same name as external library functions :-).
845
846Now, we can safely put C<my_chsize> in F<global.sym>, export it, and
847hide it with F<embed.h>.
848
849To be consistent with what I did for C<pause>, I probably should have
850called the new function C<Chsize>, rather than C<my_chsize>.
851However, the perl sources are quite inconsistent on this (Consider
852New, Mymalloc, and Myremalloc, to name just a few.)
853
854There is a problem with this fix, however, in that C<Perl_chsize>
855was available as a F<libperl.a> library function in 5.003, but it
856isn't available any more (as of 5.003_07). This means that we've
857broken binary compatibility. This is not good.
858
859=item Providing missing functions -- some ideas
860
861We currently don't have a standard way of handling such missing
862function names. Right now, I'm effectively thinking aloud about a
863solution. Some day, I'll try to formally propose a solution.
864
865Part of the problem is that we want to have some functions listed as
866exported but not have their names mangled by embed.h or possibly
867conflict with names in standard system headers. We actually already
868have such a list at the end of F<perl_exp.SH> (though that list is
869out-of-date):
870
871 # extra globals not included above.
872 cat <<END >> perl.exp
873 perl_init_ext
874 perl_init_fold
875 perl_init_i18nl14n
876 perl_alloc
877 perl_construct
878 perl_destruct
879 perl_free
880 perl_parse
881 perl_run
882 perl_get_sv
883 perl_get_av
884 perl_get_hv
885 perl_get_cv
886 perl_call_argv
887 perl_call_pv
888 perl_call_method
889 perl_call_sv
890 perl_requirepv
891 safecalloc
892 safemalloc
893 saferealloc
894 safefree
895
896This still needs much thought, but I'm inclined to think that one
897possible solution is to prefix all such functions with C<perl_> in the
898source and list them along with the other C<perl_*> functions in
899F<perl_exp.SH>.
900
901Thus, for C<chsize>, we'd do something like the following:
902
903 /* in perl.h */
904 #ifdef HAS_CHSIZE
905 # define perl_chsize chsize
906 #endif
907
908then in some file (e.g. F<util.c> or F<doio.c>) do
909
910 #ifndef HAS_CHSIZE
911 I32 perl_chsize(fd, length)
912 /* implement the function here . . . */
913 #endif
914
915Alternatively, we could just always use C<chsize> everywhere and move
916C<chsize> from F<global.sym> to the end of F<perl_exp.SH>. That would
917probably be fine as long as our C<chsize> function agreed with all the
918C<chsize> function prototypes in the various systems we'll be using.
919As long as the prototypes in actual use don't vary that much, this is
920probably a good alternative. (As a counter-example, note how Configure
921and perl have to go through hoops to find and use get Malloc_t and
922Free_t for C<malloc> and C<free>.)
923
924At the moment, this latter option is what I tend to prefer.
925
926=item All the world's a VAX
927
928Sorry, showing my age:-). Still, all the world is not BSD 4.[34],
929SVR4, or POSIX. Be aware that SVR3-derived systems are still quite
930common (do you have any idea how many systems run SCO?) If you don't
931have a bunch of v7 manuals handy, the metaconfig units (by default
932installed in F</usr/local/lib/dist/U>) are a good resource to look at
933for portability.
934
935=back
936
937=head1 Miscellaneous Topics
938
939=head2 Autoconf
940
941Why does perl use a metaconfig-generated Configure script instead of an
942autoconf-generated configure script?
943
944Metaconfig and autoconf are two tools with very similar purposes.
945Metaconfig is actually the older of the two, and was originally written
946by Larry Wall, while autoconf is probably now used in a wider variety of
947packages. The autoconf info file discusses the history of autoconf and
948how it came to be. The curious reader is referred there for further
949information.
950
951Overall, both tools are quite good, I think, and the choice of which one
952to use could be argued either way. In March, 1994, when I was just
953starting to work on Configure support for Perl5, I considered both
954autoconf and metaconfig, and eventually decided to use metaconfig for the
955following reasons:
956
957=over 4
958
959=item Compatibility with Perl4
960
961Perl4 used metaconfig, so many of the #ifdef's were already set up for
962metaconfig. Of course metaconfig had evolved some since Perl4's days,
963but not so much that it posed any serious problems.
964
965=item Metaconfig worked for me
966
967My system at the time was Interactive 2.2, a SVR3.2/386 derivative that
968also had some POSIX support. Metaconfig-generated Configure scripts
969worked fine for me on that system. On the other hand, autoconf-generated
970scripts usually didn't. (They did come quite close, though, in some
971cases.) At the time, I actually fetched a large number of GNU packages
972and checked. Not a single one configured and compiled correctly
973out-of-the-box with the system's cc compiler.
974
975=item Configure can be interactive
976
977With both autoconf and metaconfig, if the script works, everything is
978fine. However, one of my main problems with autoconf-generated scripts
979was that if it guessed wrong about something, it could be B<very> hard to
980go back and fix it. For example, autoconf always insisted on passing the
981-Xp flag to cc (to turn on POSIX behavior), even when that wasn't what I
982wanted or needed for that package. There was no way short of editing the
983configure script to turn this off. You couldn't just edit the resulting
984Makefile at the end because the -Xp flag influenced a number of other
985configure tests.
986
987Metaconfig's Configure scripts, on the other hand, can be interactive.
988Thus if Configure is guessing things incorrectly, you can go back and fix
989them. This isn't as important now as it was when we were actively
990developing Configure support for new features such as dynamic loading,
991but it's still useful occasionally.
992
993=item GPL
994
995At the time, autoconf-generated scripts were covered under the GNU Public
996License, and hence weren't suitable for inclusion with Perl, which has a
997different licensing policy. (Autoconf's licensing has since changed.)
998
999=item Modularity
1000
1001Metaconfig builds up Configure from a collection of discrete pieces
1002called "units". You can override the standard behavior by supplying your
1003own unit. With autoconf, you have to patch the standard files instead.
1004I find the metaconfig "unit" method easier to work with. Others
1005may find metaconfig's units clumsy to work with.
1006
1007=back
1008
1009=head2 @INC search order
1010
1011By default, the list of perl library directories in @INC is the
1012following:
1013
1014 $archlib
1015 $privlib
1016 $sitearch
1017 $sitelib
1018
1019Specifically, on my Solaris/x86 system, I run
1020B<sh Configure -Dprefix=/opt/perl> and I have the following
1021directories:
1022
1023 /opt/perl/lib/i86pc-solaris/5.00307
1024 /opt/perl/lib
1025 /opt/perl/lib/site_perl/i86pc-solaris
1026 /opt/perl/lib/site_perl
1027
1028That is, perl's directories come first, followed by the site-specific
1029directories.
1030
1031The site libraries come second to support the usage of extensions
1032across perl versions. Read the relevant section in F<INSTALL> for
1033more information. If we ever make $sitearch version-specific, this
1034topic could be revisited.
1035
1036=head2 Why isn't there a directory to override Perl's library?
1037
1038Mainly because no one's gotten around to making one. Note that
1039"making one" involves changing perl.c, Configure, config_h.SH (and
1040associated files, see above), and I<documenting> it all in the
1041INSTALL file.
1042
1043Apparently, most folks who want to override one of the standard library
1044files simply do it by overwriting the standard library files.
1045
1046=head2 APPLLIB
1047
1048In the perl.c sources, you'll find an undocumented APPLLIB_EXP
1049variable, sort of like PRIVLIB_EXP and ARCHLIB_EXP (which are
1050documented in config_h.SH). Here's what APPLLIB_EXP is for, from
1051a mail message from Larry:
1052
1053 The main intent of APPLLIB_EXP is for folks who want to send out a
1054 version of Perl embedded in their product. They would set the symbol
1055 to be the name of the library containing the files needed to run or to
1056 support their particular application. This works at the "override"
1057 level to make sure they get their own versions of any library code that
1058 they absolutely must have configuration control over.
1059
1060 As such, I don't see any conflict with a sysadmin using it for a
1061 override-ish sort of thing, when installing a generic Perl. It should
1062 probably have been named something to do with overriding though. Since
1063 it's undocumented we could still change it... :-)
1064
1065Given that it's already there, you can use it to override
1066distribution modules. If you do
1067
1068 sh Configure -Dccflags='-DAPPLLIB_EXP=/my/override'
1069
1070then perl.c will put /my/override ahead of ARCHLIB and PRIVLIB.
1071
c4f23d77
AD
1072=head2 Shared libperl.so location
1073
1074Why isn't the shared libperl.so installed in /usr/lib/ along
1075with "all the other" shared libraries? Instead, it is installed
1076in $archlib, which is typically something like
1077
1078 /usr/local/lib/perl5/archname/5.00404
1079
1080and is architecture- and version-specific.
1081
1082The basic reason why a shared libperl.so gets put in $archlib is so that
1083you can have more than one version of perl on the system at the same time,
1084and have each refer to its own libperl.so.
1085
1086Three examples might help. All of these work now; none would work if you
1087put libperl.so in /usr/lib.
1088
1089=over
1090
1091=item 1.
1092
1093Suppose you want to have both threaded and non-threaded perl versions
1094around. Configure will name both perl libraries "libperl.so" (so that
1095you can link to them with -lperl). The perl binaries tell them apart
1096by having looking in the appropriate $archlib directories.
1097
1098=item 2.
1099
1100Suppose you have perl5.004_04 installed and you want to try to compile
1101it again, perhaps with different options or after applying a patch.
1102If you already have libperl.so installed in /usr/lib/, then it may be
1103either difficult or impossible to get ld.so to find the new libperl.so
1104that you're trying to build. If, instead, libperl.so is tucked away in
1105$archlib, then you can always just change $archlib in the current perl
1106you're trying to build so that ld.so won't find your old libperl.so.
1107(The INSTALL file suggests you do this when building a debugging perl.)
1108
1109=item 3.
1110
1111The shared perl library is not a "well-behaved" shared library with
1112proper major and minor version numbers, so you can't necessarily
1113have perl5.004_04 and perl5.004_05 installed simultaneously. Suppose
1114perl5.004_04 were to install /usr/lib/libperl.so.4.4, and perl5.004_05
1115were to install /usr/lib/libperl.so.4.5. Now, when you try to run
1116perl5.004_04, ld.so might try to load libperl.so.4.5, since it has
1117the right "major version" number. If this works at all, it almost
1118certainly defeats the reason for keeping perl5.004_04 around. Worse,
1119with development subversions, you certaily can't guarantee that
1120libperl.so.4.4 and libperl.so.4.55 will be compatible.
1121
1122Anyway, all this leads to quite obscure failures that are sure to drive
1123casual users crazy. Even experienced users will get confused :-). Upon
1124reflection, I'd say leave libperl.so in $archlib.
1125
1126=back
1127
aa689395
PP
1128=head1 Upload Your Work to CPAN
1129
1130You can upload your work to CPAN if you have a CPAN id. Check out
1131http://www.perl.com/CPAN/modules/04pause.html for information on
1132_PAUSE_, the Perl Author's Upload Server.
1133
1134I typically upload both the patch file, e.g. F<perl5.004_08.pat.gz>
1135and the full tar file, e.g. F<perl5.004_08.tar.gz>.
1136
1137If you want your patch to appear in the F<src/5.0/unsupported>
1138directory on CPAN, send e-mail to the CPAN master librarian. (Check
7b5757d1 1139out http://www.perl.com/CPAN/CPAN.html ).
aa689395
PP
1140
1141=head1 Help Save the World
1142
1143You should definitely announce your patch on the perl5-porters list.
1144You should also consider announcing your patch on
1145comp.lang.perl.announce, though you should make it quite clear that a
1146subversion is not a production release, and be prepared to deal with
1147people who will not read your disclaimer.
1148
1149=head1 Todo
1150
1151Here, in no particular order, are some Configure and build-related
1152items that merit consideration. This list isn't exhaustive, it's just
1153what I came up with off the top of my head.
1154
1155=head2 Good ideas waiting for round tuits
1156
1157=over 4
1158
1159=item installprefix
1160
1161I think we ought to support
1162
1163 Configure -Dinstallprefix=/blah/blah
1164
1165Currently, we support B<-Dprefix=/blah/blah>, but the changing the install
1166location has to be handled by something like the F<config.over> trick
1167described in F<INSTALL>. AFS users also are treated specially.
1168We should probably duplicate the metaconfig prefix stuff for an
1169install prefix.
1170
c4f23d77 1171=item Configure -Dsrc=/blah/blah
aa689395
PP
1172
1173We should be able to emulate B<configure --srcdir>. Tom Tromey
1174tromey@creche.cygnus.com has submitted some patches to
c4f23d77
AD
1175the dist-users mailing list along these lines. They have been folded
1176back into the main distribution, but various parts of the perl
1177Configure/build/install process still assume src='.'.
aa689395
PP
1178
1179=item Hint file fixes
1180
1181Various hint files work around Configure problems. We ought to fix
1182Configure so that most of them aren't needed.
1183
1184=item Hint file information
1185
1186Some of the hint file information (particularly dynamic loading stuff)
1187ought to be fed back into the main metaconfig distribution.
1188
c4f23d77
AD
1189=item Catch GNU Libc "Stub" functions
1190
1191Some functions (such as lchown()) are present in libc, but are
1192unimplmented. That is, they always fail and set errno=ENOSYS.
1193
1194Thomas Bushnell provided the following sample code and the explanation
1195that follows:
1196
1197 /* System header to define __stub macros and hopefully few prototypes,
1198 which can conflict with char FOO(); below. */
1199 #include <assert.h>
1200 /* Override any gcc2 internal prototype to avoid an error. */
1201 /* We use char because int might match the return type of a gcc2
1202 builtin and then its argument prototype would still apply. */
1203 char FOO();
1204
1205 int main() {
1206
1207 /* The GNU C library defines this for functions which it implements
1208 to always fail with ENOSYS. Some functions are actually named
1209 something starting with __ and the normal name is an alias. */
1210 #if defined (__stub_FOO) || defined (__stub___FOO)
1211 choke me
1212 #else
1213 FOO();
1214 #endif
1215
1216 ; return 0; }
1217
1218The choice of <assert.h> is essentially arbitrary. The GNU libc
1219macros are found in <gnu/stubs.h>. You can include that file instead
1220of <assert.h> (which itself includes <gnu/stubs.h>) if you test for
1221its existence first. <assert.h> is assumed to exist on every system,
1222which is why it's used here. Any GNU libc header file will include
1223the stubs macros. If either __stub_NAME or __stub___NAME is defined,
1224then the function doesn't actually exist. Tests using <assert.h> work
1225on every system around.
1226
1227The declaration of FOO is there to override builtin prototypes for
1228ANSI C functions.
1229
aa689395
PP
1230=back
1231
1232=head2 Probably good ideas waiting for round tuits
1233
1234=over 4
1235
1236=item GNU configure --options
1237
1238I've received sensible suggestions for --exec_prefix and other
1239GNU configure --options. It's not always obvious exactly what is
1240intended, but this merits investigation.
1241
1242=item make clean
1243
1244Currently, B<make clean> isn't all that useful, though
1245B<make realclean> and B<make distclean> are. This needs a bit of
1246thought and documentation before it gets cleaned up.
1247
1248=item Try gcc if cc fails
1249
1250Currently, we just give up.
1251
1252=item bypassing safe*alloc wrappers
1253
1254On some systems, it may be safe to call the system malloc directly
1255without going through the util.c safe* layers. (Such systems would
1256accept free(0), for example.) This might be a time-saver for systems
1257that already have a good malloc. (Recent Linux libc's apparently have
1258a nice malloc that is well-tuned for the system.)
1259
1260=back
1261
1262=head2 Vague possibilities
1263
1264=over 4
1265
aa689395
PP
1266=item MacPerl
1267
3e3baf6d 1268Get some of the Macintosh stuff folded back into the main distribution.
aa689395
PP
1269
1270=item gconvert replacement
1271
1272Maybe include a replacement function that doesn't lose data in rare
1273cases of coercion between string and numerical values.
1274
aa689395
PP
1275=item Improve makedepend
1276
1277The current makedepend process is clunky and annoyingly slow, but it
1278works for most folks. Alas, it assumes that there is a filename
1279$firstmakefile that the B<make> command will try to use before it uses
1280F<Makefile>. Such may not be the case for all B<make> commands,
1281particularly those on non-Unix systems.
1282
1283Probably some variant of the BSD F<.depend> file will be useful.
1284We ought to check how other packages do this, if they do it at all.
1285We could probably pre-generate the dependencies (with the exception of
1286malloc.o, which could probably be determined at F<Makefile.SH>
1287extraction time.
1288
1289=item GNU Makefile standard targets
1290
1291GNU software generally has standardized Makefile targets. Unless we
1292have good reason to do otherwise, I see no reason not to support them.
1293
1294=item File locking
1295
1296Somehow, straighten out, document, and implement lockf(), flock(),
1297and/or fcntl() file locking. It's a mess.
1298
1299=back
1300
fb73857a 1301=head1 AUTHORS
aa689395 1302
fb73857a
PP
1303Original author: Andy Dougherty doughera@lafcol.lafayette.edu .
1304Additions by Chip Salzenberg chip@perl.com and
1305Tim Bunce Tim.Bunce@ig.co.uk .
aa689395
PP
1306
1307All opinions expressed herein are those of the authorZ<>(s).
1308
1309=head1 LAST MODIFIED
1310
a8119d38 1311$Id: pumpkin.pod,v 1.19 1998/07/06 20:48:55 doughera Released $