enable -ansi -pedantic
[p5sagit/p5-mst-13.2.git] / Porting / pumpkin.pod
CommitLineData
aa689395 1=head1 NAME
2
e25f343d 3Pumpkin - Notes on handling the Perl Patch Pumpkin And Porting Perl
aa689395 4
5=head1 SYNOPSIS
6
7There is no simple synopsis, yet.
8
9=head1 DESCRIPTION
10
98dddfbd 11This document attempts to begin to describe some of the considerations
12involved in patching, porting, and maintaining perl.
aa689395 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
a93751fa 27http://www.cpan.org/README.html , which automatically points you to a
aa689395 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 44Archives of the list are held at:
45
46 http://www.rosat.mpe-garching.mpg.de/mailing-lists/perl-porters/
47
aa689395 48=head1 How are Perl Releases Numbered?
49
f5a32c7f 50Beginning with v5.6.0, even versions will stand for maintenance releases
51and odd versions for development releases, i.e., v5.6.x for maintenance
52releases, and v5.7.x for development releases. Before v5.6.0, subversions
53_01 through _49 were reserved for bug-fix maintenance releases, and
54subversions _50 through _99 for unstable development versions.
7b5757d1 55
f5a32c7f 56For example, in v5.6.1, the revision number is 5, the version is 6,
57and 1 is the subversion.
aa689395 58
f5a32c7f 59For compatibility with the older numbering scheme the composite floating
60point version number continues to be available as the magic variable $],
76ba0908 61and amounts to C<$revision + $version/1000 + $subversion/100000>. This
f5a32c7f 62can still be used in comparisons.
aa689395 63
f5a32c7f 64 print "You've got an old perl\n" if $] < 5.005_03;
aa689395 65
f5a32c7f 66In addition, the version is also available as a string in $^V.
aa689395 67
f5a32c7f 68 print "You've got a new perl\n" if $^V and $^V ge v5.6.0;
7b5757d1 69
f5a32c7f 70You can also require particular version (or later) with:
aa689395 71
f5a32c7f 72 use 5.006;
aa689395 73
f5a32c7f 74or using the new syntax available only from v5.6 onward:
aa689395 75
f5a32c7f 76 use v5.6.0;
aa689395 77
f5a32c7f 78At some point in the future, we may need to decide what to call the
79next big revision. In the .package file used by metaconfig to
80generate Configure, there are two variables that might be relevant:
81$baserev=5 and $package=perl5.
aa689395 82
f5a32c7f 83Perl releases produced by the members of perl5-porters are usually
e04b929a 84available on CPAN in the F<src/5.0/maint> and F<src/5.0/devel>
85directories.
aa689395 86
7b5757d1 87=head2 Maintenance and Development Subversions
88
f5a32c7f 89The first rule of maintenance work is "First, do no harm."
7b5757d1 90
fb73857a 91Trial releases of bug-fix maintenance releases are announced on
92perl5-porters. Trial releases use the new subversion number (to avoid
93testers installing it over the previous release) and include a 'local
e04b929a 94patch' entry in patchlevel.h. The distribution file contains the
95string C<MAINT_TRIAL> to make clear that the file is not meant for
96public consumption.
fb73857a 97
e04b929a 98In general, the names of official distribution files for the public
f5a32c7f 99always match the regular expression:
e04b929a 100
f5a32c7f 101 ^perl\d+\.(\d+)\.\d+(-MAINT_TRIAL_\d+)\.tar\.gz$
e04b929a 102
f5a32c7f 103C<$1> in the pattern is always an even number for maintenance
104versions, and odd for developer releases.
e04b929a 105
106In the past it has been observed that pumkings tend to invent new
107naming conventions on the fly. If you are a pumpking, before you
108invent a new name for any of the three types of perl distributions,
109please inform the guys from the CPAN who are doing indexing and
110provide the trees of symlinks and the like. They will have to know
111I<in advance> what you decide.
20f245af 112
aa689395 113=head2 Why is it called the patch pumpkin?
114
115Chip Salzenberg gets credit for that, with a nod to his cow orker,
116David Croy. We had passed around various names (baton, token, hot
117potato) but none caught on. Then, Chip asked:
118
119[begin quote]
120
121 Who has the patch pumpkin?
122
123To explain: David Croy once told me once that at a previous job,
124there was one tape drive and multiple systems that used it for backups.
125But instead of some high-tech exclusion software, they used a low-tech
126method to prevent multiple simultaneous backups: a stuffed pumpkin.
127No one was allowed to make backups unless they had the "backup pumpkin".
128
129[end quote]
130
131The name has stuck.
132
a6968aa6 133=head1 Philosophical Issues in Patching and Porting Perl
aa689395 134
135There are no absolute rules, but there are some general guidelines I
136have tried to follow as I apply patches to the perl sources.
137(This section is still under construction.)
138
139=head2 Solve problems as generally as possible
140
7b5757d1 141Never implement a specific restricted solution to a problem when you
142can solve the same problem in a more general, flexible way.
143
144For example, for dynamic loading to work on some SVR4 systems, we had
145to build a shared libperl.so library. In order to build "FAT" binaries
146on NeXT 4.0 systems, we had to build a special libperl library. Rather
147than continuing to build a contorted nest of special cases, I
148generalized the process of building libperl so that NeXT and SVR4 users
149could still get their work done, but others could build a shared
150libperl if they wanted to as well.
aa689395 151
a6968aa6 152Contain your changes carefully. Assume nothing about other operating
153systems, not even closely related ones. Your changes must not affect
154other platforms.
155
156Spy shamelessly on how similar patching or porting issues have been
157settled elsewhere.
158
159If feasible, try to keep filenames 8.3-compliant to humor those poor
160souls that get joy from running Perl under such dire limitations.
9e371ce5 161There's a script, check83.pl, for keeping your nose 8.3-clean.
a6968aa6 162
aa689395 163=head2 Seek consensus on major changes
164
165If you are making big changes, don't do it in secret. Discuss the
166ideas in advance on perl5-porters.
167
168=head2 Keep the documentation up-to-date
169
170If your changes may affect how users use perl, then check to be sure
171that the documentation is in sync with your changes. Be sure to
172check all the files F<pod/*.pod> and also the F<INSTALL> document.
173
174Consider writing the appropriate documentation first and then
7b5757d1 175implementing your change to correspond to the documentation.
aa689395 176
177=head2 Avoid machine-specific #ifdef's
178
179To the extent reasonable, try to avoid machine-specific #ifdef's in
180the sources. Instead, use feature-specific #ifdef's. The reason is
181that the machine-specific #ifdef's may not be valid across major
182releases of the operating system. Further, the feature-specific tests
183may help out folks on another platform who have the same problem.
184
a6968aa6 185=head2 Machine-specific files
186
98dddfbd 187=over 4
188
189=item source code
190
a6968aa6 191If you have many machine-specific #defines or #includes, consider
192creating an "osish.h" (os2ish.h, vmsish.h, and so on) and including
193that in perl.h. If you have several machine-specific files (function
194emulations, function stubs, build utility wrappers) you may create a
195separate subdirectory (djgpp, win32) and put the files in there.
98dddfbd 196Remember to update C<MANIFEST> when you add files.
a6968aa6 197
ff935051 198If your system supports dynamic loading but none of the existing
98dddfbd 199methods at F<ext/DynaLoader/dl_*.xs> work for you, you must write
200a new one. Study the existing ones to see what kind of interface
201you must supply.
202
203=item build hints
a6968aa6 204
205There are two kinds of hints: hints for building Perl and hints for
206extensions. The former live in the C<hints> subdirectory, the latter
207in C<ext/*/hints> subdirectories.
208
209The top level hints are Bourne-shell scripts that set, modify and
210unset appropriate Configure variables, based on the Configure command
211line options and possibly existing config.sh and Policy.sh files from
212previous Configure runs.
213
76ba0908 214The extension hints are written in Perl (by the time they are used
a6968aa6 215miniperl has been built) and control the building of their respective
216extensions. They can be used to for example manipulate compilation
217and linking flags.
218
98dddfbd 219=item build and installation Makefiles, scripts, and so forth
220
221Sometimes you will also need to tweak the Perl build and installation
222procedure itself, like for example F<Makefile.SH> and F<installperl>.
223Tread very carefully, even more than usual. Contain your changes
224with utmost care.
a6968aa6 225
98dddfbd 226=item test suite
227
228Many of the tests in C<t> subdirectory assume machine-specific things
a6968aa6 229like existence of certain functions, something about filesystem
230semantics, certain external utilities and their error messages. Use
231the C<$^O> and the C<Config> module (which contains the results of the
232Configure run, in effect the C<config.sh> converted to Perl) to either
98dddfbd 233skip (preferably not) or customize (preferable) the tests for your
234platform.
235
236=item modules
237
238Certain standard modules may need updating if your operating system
239sports for example a native filesystem naming. You may want to update
240some or all of the modules File::Basename, File::Spec, File::Path, and
241File::Copy to become aware of your native filesystem syntax and
242peculiarities.
243
b972f109 244Remember to have a $VERSION in the modules. You can use the
245Porting/checkVERSION.pl script for checking this.
246
98dddfbd 247=item documentation
248
249If your operating system comes from outside UNIX you almost certainly
250will have differences in the available operating system functionality
251(missing system calls, different semantics, whatever). Please
252document these at F<pod/perlport.pod>. If your operating system is
253the first B<not> to have a system call also update the list of
254"portability-bewares" at the beginning of F<pod/perlfunc.pod>.
255
256A file called F<README.youros> at the top level that explains things
257like how to install perl at this platform, where to get any possibly
258required additional software, and for example what test suite errors
76ba0908 259to expect, is nice too. Such files are in the process of being written
260in pod format and will eventually be renamed F<INSTALL.youros>.
98dddfbd 261
262You may also want to write a separate F<.pod> file for your operating
263system to tell about existing mailing lists, os-specific modules,
264documentation, whatever. Please name these along the lines of
265F<perl>I<youros>.pod. [unfinished: where to put this file (the pod/
266subdirectory, of course: but more importantly, which/what index files
267should be updated?)]
268
269=back
a6968aa6 270
aa689395 271=head2 Allow for lots of testing
272
273We should never release a main version without testing it as a
274subversion first.
275
6877a1cf 276=head2 Test popular applications and modules.
277
278We should never release a main version without testing whether or not
279it breaks various popular modules and applications. A partial list of
280such things would include majordomo, metaconfig, apache, Tk, CGI,
281libnet, and libwww, to name just a few. Of course it's quite possible
282that some of those things will be just plain broken and need to be fixed,
283but, in general, we ought to try to avoid breaking widely-installed
284things.
285
98dddfbd 286=head2 Automated generation of derivative files
aa689395 287
288The F<embed.h>, F<keywords.h>, F<opcode.h>, and F<perltoc.pod> files
289are all automatically generated by perl scripts. In general, don't
290patch these directly; patch the data files instead.
291
292F<Configure> and F<config_h.SH> are also automatically generated by
293B<metaconfig>. In general, you should patch the metaconfig units
a6968aa6 294instead of patching these files directly. However, very minor changes
295to F<Configure> may be made in between major sync-ups with the
296metaconfig units, which tends to be complicated operations. But be
297careful, this can quickly spiral out of control. Running metaconfig
298is not really hard.
aa689395 299
98dddfbd 300Also F<Makefile> is automatically produced from F<Makefile.SH>.
301In general, look out for all F<*.SH> files.
302
a8119d38 303Finally, the sample files in the F<Porting/> subdirectory are
304generated automatically by the script F<U/mksample> included
305with the metaconfig units. See L<"run metaconfig"> below for
306information on obtaining the metaconfig units.
307
aa689395 308=head1 How to Make a Distribution
309
310There really ought to be a 'make dist' target, but there isn't.
311The 'dist' suite of tools also contains a number of tools that I haven't
312learned how to use yet. Some of them may make this all a bit easier.
313
314Here are the steps I go through to prepare a patch & distribution.
315
3e3baf6d 316Lots of it could doubtless be automated but isn't. The Porting/makerel
317(make release) perl script does now help automate some parts of it.
aa689395 318
319=head2 Announce your intentions
320
321First, you should volunteer out loud to take the patch pumpkin. It's
322generally counter-productive to have multiple people working in secret
323on the same thing.
324
325At the same time, announce what you plan to do with the patch pumpkin,
326to allow folks a chance to object or suggest alternatives, or do it for
327you. Naturally, the patch pumpkin holder ought to incorporate various
328bug fixes and documentation improvements that are posted while he or
329she has the pumpkin, but there might also be larger issues at stake.
330
331One of the precepts of the subversion idea is that we shouldn't give
7b5757d1 332the patch pumpkin to anyone unless we have some idea what he or she
333is going to do with it.
aa689395 334
335=head2 refresh pod/perltoc.pod
336
337Presumably, you have done a full C<make> in your working source
338directory. Before you C<make spotless> (if you do), and if you have
339changed any documentation in any module or pod file, change to the
340F<pod> directory and run C<make toc>.
341
3e3baf6d 342=head2 run installhtml to check the validity of the pod files
343
aa689395 344=head2 update patchlevel.h
345
346Don't be shy about using the subversion number, even for a relatively
347modest patch. We've never even come close to using all 99 subversions,
348and it's better to have a distinctive number for your patch. If you
349need feedback on your patch, go ahead and issue it and promise to
350incorporate that feedback quickly (e.g. within 1 week) and send out a
351second patch.
352
05ff1fbb 353If you update the subversion number, you may need to change the version
354number near the top of the F<Changes> file.
355
aa689395 356=head2 run metaconfig
357
358If you need to make changes to Configure or config_h.SH, it may be best to
359change the appropriate metaconfig units instead, and regenerate Configure.
360
361 metaconfig -m
362
20f245af 363will regenerate Configure and config_h.SH. Much more information
364on obtaining and running metaconfig is in the F<U/README> file
365that comes with Perl's metaconfig units. Perl's metaconfig units
366should be available on CPAN. A set of units that will work with
367perl5.005 is in the file F<mc_units-5.005_00-01.tar.gz> under
a93751fa 368http://www.cpan.org/authors/id/ANDYD/ . The mc_units tar file
20f245af 369should be unpacked in your main perl source directory. Note: those
370units were for use with 5.005. There may have been changes since then.
d562869c 371Check for later versions or contact perl5-porters@perl.org to obtain a
20f245af 372pointer to the current version.
aa689395 373
374Alternatively, do consider if the F<*ish.h> files might be a better
375place for your changes.
376
377=head2 MANIFEST
378
379Make sure the MANIFEST is up-to-date. You can use dist's B<manicheck>
380program for this. You can also use
381
3e3baf6d 382 perl -w -MExtUtils::Manifest=fullcheck -e fullcheck
aa689395 383
3e3baf6d 384Both commands will also list extra files in the directory that are not
385listed in MANIFEST.
aa689395 386
bfb7748a 387The MANIFEST is normally sorted.
aa689395 388
389If you are using metaconfig to regenerate Configure, then you should note
390that metaconfig actually uses MANIFEST.new, so you want to be sure
391MANIFEST.new is up-to-date too. I haven't found the MANIFEST/MANIFEST.new
392distinction particularly useful, but that's probably because I still haven't
393learned how to use the full suite of tools in the dist distribution.
394
395=head2 Check permissions
396
397All the tests in the t/ directory ought to be executable. The
398main makefile used to do a 'chmod t/*/*.t', but that resulted in
399a self-modifying distribution--something some users would strongly
d562869c 400prefer to avoid. The F<t/TEST> script will check for this
401and do the chmod if needed, but the tests still ought to be
402executable.
aa689395 403
404In all, the following files should probably be executable:
405
406 Configure
407 configpm
32fcaa0b 408 configure.gnu
aa689395 409 embed.pl
410 installperl
411 installman
412 keywords.pl
aa689395 413 myconfig
414 opcode.pl
415 perly.fixer
416 t/TEST
417 t/*/*.t
418 *.SH
419 vms/ext/Stdio/test.pl
420 vms/ext/filespec.t
aa689395 421 x2p/*.SH
422
423Other things ought to be readable, at least :-).
424
425Probably, the permissions for the files could be encoded in MANIFEST
426somehow, but I'm reluctant to change MANIFEST itself because that
427could break old scripts that use MANIFEST.
428
429I seem to recall that some SVR3 systems kept some sort of file that listed
430permissions for system files; something like that might be appropriate.
431
432=head2 Run Configure
433
434This will build a config.sh and config.h. You can skip this if you haven't
693762b4 435changed Configure or config_h.SH at all. I use the following command
aa689395 436
693762b4 437 sh Configure -Dprefix=/opt/perl -Doptimize=-O -Dusethreads \
438 -Dcf_by='yourname' \
439 -Dcf_email='yourname@yourhost.yourplace.com' \
440 -Dperladmin='yourname@yourhost.yourplace.com' \
441 -Dmydomain='.yourplace.com' \
442 -Dmyhostname='yourhost' \
443 -des
aa689395 444
693762b4 445=head2 Update Porting/config.sh and Porting/config_H
dfe9444c 446
693762b4 447[XXX
448This section needs revision. We're currently working on easing
449the task of keeping the vms, win32, and plan9 config.sh info
450up-to-date. The plan is to use keep up-to-date 'canned' config.sh
451files in the appropriate subdirectories and then generate 'canned'
452config.h files for vms, win32, etc. from the generic config.sh file.
453This is to ease maintenance. When Configure gets updated, the parts
454sometimes get scrambled around, and the changes in config_H can
455sometimes be very hard to follow. config.sh, on the other hand, can
456safely be sorted, so it's easy to track (typically very small) changes
457to config.sh and then propoagate them to a canned 'config.h' by any
458number of means, including a perl script in win32/ or carrying
459config.sh and config_h.SH to a Unix system and running sh
76ba0908 460config_h.SH.) Vms uses configure.com to generate its own config.sh
461and config.h. If you want to add a new variable to config.sh check
462with vms folk how to add it to configure.com too.
693762b4 463XXX]
464
465The Porting/config.sh and Porting/config_H files are provided to
466help those folks who can't run Configure. It is important to keep
467them up-to-date. If you have changed config_h.SH, those changes must
468be reflected in config_H as well. (The name config_H was chosen to
469distinguish the file from config.h even on case-insensitive file systems.)
470Simply edit the existing config_H file; keep the first few explanatory
471lines and then copy your new config.h below.
aa689395 472
76ba0908 473It may also be necessary to update win32/config.?c, and
aa689395 474plan9/config.plan9, though you should be quite careful in doing so if
475you are not familiar with those systems. You might want to issue your
476patch with a promise to quickly issue a follow-up that handles those
477directories.
478
479=head2 make run_byacc
480
481If you have byacc-1.8.2 (available from CPAN), and if there have been
482changes to F<perly.y>, you can regenerate the F<perly.c> file. The
483run_byacc makefile target does this by running byacc and then applying
484some patches so that byacc dynamically allocates space, rather than
485having fixed limits. This patch is handled by the F<perly.fixer>
486script. Depending on the nature of the changes to F<perly.y>, you may
487or may not have to hand-edit the patch to apply correctly. If you do,
488you should include the edited patch in the new distribution. If you
489have byacc-1.9, the patch won't apply cleanly. Changes to the printf
490output statements mean the patch won't apply cleanly. Long ago I
491started to fix F<perly.fixer> to detect this, but I never completed the
492task.
493
76ba0908 494If C<perly.c> or C<perly.h> changes, make sure you run C<perl vms/vms_yfix.pl>
495to update the corresponding VMS files. This could be taken care of by
496the regen_all target in the Unix Makefile. See also
497L<VMS-specific updates>.
ebb99254 498
aa689395 499Some additional notes from Larry on this:
500
e262e9be 501Don't forget to regenerate perly_c.diff.
aa689395 502
7b5757d1 503 byacc -d perly.y
aa689395 504 mv y.tab.c perly.c
e262e9be 505 patch perly.c <perly_c.diff
aa689395 506 # manually apply any failed hunks
eade9b71 507 diff -c perly.c.orig perly.c >perly_c.diff
aa689395 508
509One chunk of lines that often fails begins with
510
511 #line 29 "perly.y"
512
513and ends one line before
514
515 #define YYERRCODE 256
516
517This only happens when you add or remove a token type. I suppose this
518could be automated, but it doesn't happen very often nowadays.
519
520Larry
521
76ba0908 522=head2 make regen_all
523
524This target takes care of the PERLYVMS, regen_headers, and regen_pods
525targets.
526
aa689395 527=head2 make regen_headers
528
529The F<embed.h>, F<keywords.h>, and F<opcode.h> files are all automatically
530generated by perl scripts. Since the user isn't guaranteed to have a
531working perl, we can't require the user to generate them. Hence you have
532to, if you're making a distribution.
533
534I used to include rules like the following in the makefile:
535
536 # The following three header files are generated automatically
537 # The correct versions should be already supplied with the perl kit,
538 # in case you don't have perl or 'sh' available.
539 # The - is to ignore error return codes in case you have the source
540 # installed read-only or you don't have perl yet.
541 keywords.h: keywords.pl
542 @echo "Don't worry if this fails."
543 - perl keywords.pl
544
545
7b5757d1 546However, I got B<lots> of mail consisting of people worrying because the
aa689395 547command failed. I eventually decided that I would save myself time
548and effort by manually running C<make regen_headers> myself rather
549than answering all the questions and complaints about the failing
550command.
551
76ba0908 552=head2 make regen_pods
553
554Will run `make regen_pods` in the pod directory for indexing.
555
3e3baf6d 556=head2 global.sym, interp.sym and perlio.sym
aa689395 557
558Make sure these files are up-to-date. Read the comments in these
559files and in perl_exp.SH to see what to do.
560
561=head2 Binary compatibility
562
563If you do change F<global.sym> or F<interp.sym>, think carefully about
564what you are doing. To the extent reasonable, we'd like to maintain
76ba0908 565source and binary compatibility with older releases of perl. That way,
aa689395 566extensions built under one version of perl will continue to work with
567new versions of perl.
568
569Of course, some incompatible changes may well be necessary. I'm just
570suggesting that we not make any such changes without thinking carefully
571about them first. If possible, we should provide
572backwards-compatibility stubs. There's a lot of XS code out there.
573Let's not force people to keep changing it.
574
575=head2 Changes
576
577Be sure to update the F<Changes> file. Try to include both an overall
578summary as well as detailed descriptions of the changes. Your
3e3baf6d 579audience will include other developers and users, so describe
aa689395 580user-visible changes (if any) in terms they will understand, not in
581code like "initialize foo variable in bar function".
582
583There are differing opinions on whether the detailed descriptions
584ought to go in the Changes file or whether they ought to be available
585separately in the patch file (or both). There is no disagreement that
586detailed descriptions ought to be easily available somewhere.
587
05ff1fbb 588If you update the subversion number in F<patchlevel.h>, you may need
589to change the version number near the top of the F<Changes> file.
590
2a26e2f1 591=head2 Todo
592
593The F<Todo> file contains a roughly-catgorized unordered list of
594aspects of Perl that could use enhancement, features that could be
595added, areas that could be cleaned up, and so on. During your term as
596pumpkin-holder, you will probably address some of these issues, and
597perhaps identify others which, while you decide not to address them
598this time around, may be tackled in the future. Update the file
599reflect the situation as it stands when you hand over the pumpkin.
600
601You might like, early in your pumpkin-holding career, to see if you
602can find champions for partiticular issues on the to-do list: an issue
603owned is an issue more likely to be resolved.
604
c4f23d77 605There are also some more porting-specific L<Todo> items later in this
606file.
607
aa689395 608=head2 OS/2-specific updates
609
610In the os2 directory is F<diff.configure>, a set of OS/2-specific
611diffs against B<Configure>. If you make changes to Configure, you may
612want to consider regenerating this diff file to save trouble for the
613OS/2 maintainer.
614
7b5757d1 615You can also consider the OS/2 diffs as reminders of portability
616things that need to be fixed in Configure.
617
aa689395 618=head2 VMS-specific updates
619
ebb99254 620If you have changed F<perly.y> or F<perly.c>, then you most probably want
76ba0908 621to update F<vms/perly_{h,c}.vms> by running C<perl vms/vms_yfix.pl>, or
622by running `make regen_all` which will run that script for you.
aa689395 623
76ba0908 624The Perl revision number appears as "perl5" in configure.com.
625It is courteous to update that if necessary.
aa689395 626
627=head2 Making the new distribution
628
629Suppose, for example, that you want to make version 5.004_08. Then you can
630do something like the following
631
632 mkdir ../perl5.004_08
633 awk '{print $1}' MANIFEST | cpio -pdm ../perl5.004_08
634 cd ../
635 tar cf perl5.004_08.tar perl5.004_08
636 gzip --best perl5.004_08.tar
637
3e3baf6d 638These steps, with extra checks, are automated by the Porting/makerel
639script.
640
aa689395 641=head2 Making a new patch
642
643I find the F<makepatch> utility quite handy for making patches.
644You can obtain it from any CPAN archive under
a93751fa 645http://www.cpan.org/authors/Johan_Vromans/ . There are a couple
3e3baf6d 646of differences between my version and the standard one. I have mine do
647a
aa689395 648
649 # Print a reassuring "End of Patch" note so people won't
650 # wonder if their mailer truncated patches.
651 print "\n\nEnd of Patch.\n";
652
3e3baf6d 653at the end. That's because I used to get questions from people asking
654if their mail was truncated.
655
656It also writes Index: lines which include the new directory prefix
657(change Index: print, approx line 294 or 310 depending on the version,
658to read: print PATCH ("Index: $newdir$new\n");). That helps patches
659work with more POSIX conformant patch programs.
aa689395 660
661Here's how I generate a new patch. I'll use the hypothetical
6625.004_07 to 5.004_08 patch as an example.
663
664 # unpack perl5.004_07/
665 gzip -d -c perl5.004_07.tar.gz | tar -xof -
666 # unpack perl5.004_08/
667 gzip -d -c perl5.004_08.tar.gz | tar -xof -
668 makepatch perl5.004_07 perl5.004_08 > perl5.004_08.pat
669
670Makepatch will automatically generate appropriate B<rm> commands to remove
671deleted files. Unfortunately, it will not correctly set permissions
672for newly created files, so you may have to do so manually. For example,
673patch 5.003_04 created a new test F<t/op/gv.t> which needs to be executable,
674so at the top of the patch, I inserted the following lines:
675
676 # Make a new test
677 touch t/op/gv.t
678 chmod +x t/opt/gv.t
679
680Now, of course, my patch is now wrong because makepatch didn't know I
681was going to do that command, and it patched against /dev/null.
682
683So, what I do is sort out all such shell commands that need to be in the
684patch (including possible mv-ing of files, if needed) and put that in the
685shell commands at the top of the patch. Next, I delete all the patch parts
686of perl5.004_08.pat, leaving just the shell commands. Then, I do the
687following:
688
7b5757d1 689 cd perl5.004_07
690 sh ../perl5.004_08.pat
aa689395 691 cd ..
7b5757d1 692 makepatch perl5.004_07 perl5.004_08 >> perl5.004_08.pat
aa689395 693
694(Note the append to preserve my shell commands.)
695Now, my patch will line up with what the end users are going to do.
696
697=head2 Testing your patch
698
699It seems obvious, but be sure to test your patch. That is, verify that
700it produces exactly the same thing as your full distribution.
701
7b5757d1 702 rm -rf perl5.004_07
703 gzip -d -c perl5.004_07.tar.gz | tar -xf -
704 cd perl5.004_07
705 sh ../perl5.004_08.pat
706 patch -p1 -N < ../perl5.004_08.pat
aa689395 707 cd ..
7b5757d1 708 gdiff -r perl5.004_07 perl5.004_08
aa689395 709
710where B<gdiff> is GNU diff. Other diff's may also do recursive checking.
711
712=head2 More testing
713
714Again, it's obvious, but you should test your new version as widely as you
715can. You can be sure you'll hear about it quickly if your version doesn't
716work on both ANSI and pre-ANSI compilers, and on common systems such as
717SunOS 4.1.[34], Solaris, and Linux.
718
719If your changes include conditional code, try to test the different
720branches as thoroughly as you can. For example, if your system
721supports dynamic loading, you can also test static loading with
722
723 sh Configure -Uusedl
724
725You can also hand-tweak your config.h to try out different #ifdef
726branches.
727
d2560b70 728=head2 Other tests
729
730=over 4
731
732=item CHECK_FORMAT
733
734To test the correct use of printf-style arguments, C<Configure> with
735S<-Dccflags='-DCHECK_FORMAT -Wformat'> and run C<make>. The compiler
736will produce warning of incorrect use of format arguments. CHECK_FORMAT
737changes perl-defined formats to common formats, so DO NOT USE the executable
738produced by this process.
739
740A more accurate approach is the following commands:
741
b3fe4827 742=over 4
743
744=item *
745
746build miniperl with -DCHECK_FORMAT
747
748 make clean
749 make miniperl OPTIMIZE=-DCHECK_FORMAT >& mini.log
750
751=item *
752
753build a clean miniperl,
754and build everything else from that with -DCHECK_FORMAT
755
d2560b70 756 make clean
b3fe4827 757 make miniperl
436c6dd3 758 make all OPTIMIZE='-DCHECK_FORMAT -Wformat' >& make.log
b3fe4827 759
760=item *
761
762clean up, and print warnings from the log files
763
d2560b70 764 make clean
b3fe4827 765 perl -nwe 'print if /^\S+:/ and not /^make\b/' \
766 mini.log make.log
767
768=back
d2560b70 769
770(-Wformat support by Robin Barker.)
771
93189314 772=item gcc -ansi -pedantic
773
774Configure -Dgccansipedantic [ -Dcc=gcc ] will enable (via the cflags script,
775not $Config{ccflags}) the gcc strict ANSI C flags -ansi and -pedantic for
776the compilation of the core files on platforms where it knows it can
777do so (like Linux, see cflags.SH for the full list), and on some
778platforms only one (Solaris can do only -pedantic, not -ansi).
779The flag -DPERL_GCC_PEDANTIC also gets added, since gcc does not add
780any internal cpp flag to signify that -pedantic is being used, as it
781does for -ansi (__STRICT_ANSI__).
782
783The -ansi and -pedantic are useful in catching at least the following
784nonportable practices:
785
786=over 4
787
788=item *
789
790gcc-specific extensions
791
792=item *
793
794lvalue casts
795
796=item *
797
798// C++ comments
799
800=item *
801
802enum trailing commas
803
804=back
805
806The -Dgccansipedantic should be used only when cleaning up the code,
807not for production builds, since otherwise gcc cannot inline certain
808things.
809
d2560b70 810=back
811
d33b2eba 812=head1 Running Purify
f5a32c7f 813
814Purify is a commercial tool that is helpful in identifying memory
815overruns, wild pointers, memory leaks and other such badness. Perl
816must be compiled in a specific way for optimal testing with Purify.
817
818Use the following commands to test perl with Purify:
819
820 sh Configure -des -Doptimize=-g -Uusemymalloc -Dusemultiplicity \
821 -Accflags=-DPURIFY
822 setenv PURIFYOPTIONS "-chain-length=25"
823 make all pureperl
824 cd t
825 ln -s ../pureperl perl
365a6279 826 setenv PERL_DESTRUCT_LEVEL 2
f5a32c7f 827 ./perl TEST
828
829Disabling Perl's malloc allows Purify to monitor allocations and leaks
830more closely; using Perl's malloc will make Purify report most leaks
831in the "potential" leaks category. Enabling the multiplicity option
832allows perl to clean up thoroughly when the interpreter shuts down, which
833reduces the number of bogus leak reports from Purify. The -DPURIFY
834enables any Purify-specific debugging code in the sources.
835
836Purify outputs messages in "Viewer" windows by default. If you don't have
837a windowing environment or if you simply want the Purify output to
838unobtrusively go to a log file instead of to the interactive window,
839use the following options instead:
840
841 setenv PURIFYOPTIONS "-chain-length=25 -windows=no -log-file=perl.log \
842 -append-logfile=yes"
843
844The only currently known leaks happen when there are compile-time errors
845within eval or require. (Fixing these is non-trivial, unfortunately, but
846they must be fixed eventually.)
847
aa689395 848=head1 Common Gotcha's
849
850=over 4
851
852=item #elif
853
854The '#elif' preprocessor directive is not understood on all systems.
855Specifically, I know that Pyramids don't understand it. Thus instead of the
856simple
857
858 #if defined(I_FOO)
859 # include <foo.h>
860 #elif defined(I_BAR)
861 # include <bar.h>
862 #else
863 # include <fubar.h>
864 #endif
865
866You have to do the more Byzantine
867
868 #if defined(I_FOO)
869 # include <foo.h>
870 #else
871 # if defined(I_BAR)
872 # include <bar.h>
873 # else
874 # include <fubar.h>
875 # endif
876 #endif
877
878Incidentally, whitespace between the leading '#' and the preprocessor
879command is not guaranteed, but is very portable and you may use it freely.
880I think it makes things a bit more readable, especially once things get
881rather deeply nested. I also think that things should almost never get
882too deeply nested, so it ought to be a moot point :-)
883
884=item Probably Prefer POSIX
885
886It's often the case that you'll need to choose whether to do
887something the BSD-ish way or the POSIX-ish way. It's usually not
888a big problem when the two systems use different names for similar
889functions, such as memcmp() and bcmp(). The perl.h header file
890handles these by appropriate #defines, selecting the POSIX mem*()
891functions if available, but falling back on the b*() functions, if
892need be.
893
894More serious is the case where some brilliant person decided to
895use the same function name but give it a different meaning or
896calling sequence :-). getpgrp() and setpgrp() come to mind.
897These are a real problem on systems that aim for conformance to
898one standard (e.g. POSIX), but still try to support the other way
899of doing things (e.g. BSD). My general advice (still not really
900implemented in the source) is to do something like the following.
901Suppose there are two alternative versions, fooPOSIX() and
902fooBSD().
903
904 #ifdef HAS_FOOPOSIX
905 /* use fooPOSIX(); */
906 #else
907 # ifdef HAS_FOOBSD
908 /* try to emulate fooPOSIX() with fooBSD();
909 perhaps with the following: */
910 # define fooPOSIX fooBSD
911 # else
912 # /* Uh, oh. We have to supply our own. */
913 # define fooPOSIX Perl_fooPOSIX
914 # endif
915 #endif
916
917=item Think positively
918
919If you need to add an #ifdef test, it is usually easier to follow if you
920think positively, e.g.
921
922 #ifdef HAS_NEATO_FEATURE
923 /* use neato feature */
924 #else
925 /* use some fallback mechanism */
926 #endif
927
928rather than the more impenetrable
929
930 #ifndef MISSING_NEATO_FEATURE
931 /* Not missing it, so we must have it, so use it */
932 #else
933 /* Are missing it, so fall back on something else. */
934 #endif
935
936Of course for this toy example, there's not much difference. But when
937the #ifdef's start spanning a couple of screen fulls, and the #else's
938are marked something like
939
940 #else /* !MISSING_NEATO_FEATURE */
941
942I find it easy to get lost.
943
944=item Providing Missing Functions -- Problem
945
946Not all systems have all the neat functions you might want or need, so
947you might decide to be helpful and provide an emulation. This is
948sound in theory and very kind of you, but please be careful about what
949you name the function. Let me use the C<pause()> function as an
950illustration.
951
952Perl5.003 has the following in F<perl.h>
953
954 #ifndef HAS_PAUSE
955 #define pause() sleep((32767<<16)+32767)
956 #endif
957
958Configure sets HAS_PAUSE if the system has the pause() function, so
959this #define only kicks in if the pause() function is missing.
960Nice idea, right?
961
962Unfortunately, some systems apparently have a prototype for pause()
963in F<unistd.h>, but don't actually have the function in the library.
964(Or maybe they do have it in a library we're not using.)
965
966Thus, the compiler sees something like
967
968 extern int pause(void);
969 /* . . . */
970 #define pause() sleep((32767<<16)+32767)
971
972and dies with an error message. (Some compilers don't mind this;
973others apparently do.)
974
975To work around this, 5.003_03 and later have the following in perl.h:
976
977 /* Some unistd.h's give a prototype for pause() even though
978 HAS_PAUSE ends up undefined. This causes the #define
979 below to be rejected by the compiler. Sigh.
980 */
981 #ifdef HAS_PAUSE
982 # define Pause pause
983 #else
984 # define Pause() sleep((32767<<16)+32767)
985 #endif
986
987This works.
988
989The curious reader may wonder why I didn't do the following in
990F<util.c> instead:
991
992 #ifndef HAS_PAUSE
993 void pause()
994 {
995 sleep((32767<<16)+32767);
996 }
997 #endif
998
999That is, since the function is missing, just provide it.
1000Then things would probably be been alright, it would seem.
1001
1002Well, almost. It could be made to work. The problem arises from the
1003conflicting needs of dynamic loading and namespace protection.
1004
1005For dynamic loading to work on AIX (and VMS) we need to provide a list
1006of symbols to be exported. This is done by the script F<perl_exp.SH>,
1007which reads F<global.sym> and F<interp.sym>. Thus, the C<pause>
1008symbol would have to be added to F<global.sym> So far, so good.
1009
1010On the other hand, one of the goals of Perl5 is to make it easy to
1011either extend or embed perl and link it with other libraries. This
1012means we have to be careful to keep the visible namespace "clean".
1013That is, we don't want perl's global variables to conflict with
1014those in the other application library. Although this work is still
1015in progress, the way it is currently done is via the F<embed.h> file.
1016This file is built from the F<global.sym> and F<interp.sym> files,
1017since those files already list the globally visible symbols. If we
1018had added C<pause> to global.sym, then F<embed.h> would contain the
1019line
1020
1021 #define pause Perl_pause
1022
1023and calls to C<pause> in the perl sources would now point to
1024C<Perl_pause>. Now, when B<ld> is run to build the F<perl> executable,
1025it will go looking for C<perl_pause>, which probably won't exist in any
1026of the standard libraries. Thus the build of perl will fail.
1027
1028Those systems where C<HAS_PAUSE> is not defined would be ok, however,
1029since they would get a C<Perl_pause> function in util.c. The rest of
1030the world would be in trouble.
1031
1032And yes, this scenario has happened. On SCO, the function C<chsize>
1033is available. (I think it's in F<-lx>, the Xenix compatibility
1034library.) Since the perl4 days (and possibly before), Perl has
1035included a C<chsize> function that gets called something akin to
1036
1037 #ifndef HAS_CHSIZE
1038 I32 chsize(fd, length)
1039 /* . . . */
1040 #endif
1041
1042When 5.003 added
1043
1044 #define chsize Perl_chsize
1045
1046to F<embed.h>, the compile started failing on SCO systems.
1047
1048The "fix" is to give the function a different name. The one
1049implemented in 5.003_05 isn't optimal, but here's what was done:
1050
1051 #ifdef HAS_CHSIZE
1052 # ifdef my_chsize /* Probably #defined to Perl_my_chsize in embed.h */
1053 # undef my_chsize
1054 # endif
1055 # define my_chsize chsize
1056 #endif
1057
1058My explanatory comment in patch 5.003_05 said:
1059
1060 Undef and then re-define my_chsize from Perl_my_chsize to
1061 just plain chsize if this system HAS_CHSIZE. This probably only
1062 applies to SCO. This shows the perils of having internal
1063 functions with the same name as external library functions :-).
1064
1065Now, we can safely put C<my_chsize> in F<global.sym>, export it, and
1066hide it with F<embed.h>.
1067
1068To be consistent with what I did for C<pause>, I probably should have
1069called the new function C<Chsize>, rather than C<my_chsize>.
1070However, the perl sources are quite inconsistent on this (Consider
1071New, Mymalloc, and Myremalloc, to name just a few.)
1072
1073There is a problem with this fix, however, in that C<Perl_chsize>
1074was available as a F<libperl.a> library function in 5.003, but it
1075isn't available any more (as of 5.003_07). This means that we've
1076broken binary compatibility. This is not good.
1077
1078=item Providing missing functions -- some ideas
1079
1080We currently don't have a standard way of handling such missing
1081function names. Right now, I'm effectively thinking aloud about a
1082solution. Some day, I'll try to formally propose a solution.
1083
1084Part of the problem is that we want to have some functions listed as
1085exported but not have their names mangled by embed.h or possibly
1086conflict with names in standard system headers. We actually already
1087have such a list at the end of F<perl_exp.SH> (though that list is
1088out-of-date):
1089
1090 # extra globals not included above.
1091 cat <<END >> perl.exp
1092 perl_init_ext
1093 perl_init_fold
1094 perl_init_i18nl14n
1095 perl_alloc
1096 perl_construct
1097 perl_destruct
1098 perl_free
1099 perl_parse
1100 perl_run
1101 perl_get_sv
1102 perl_get_av
1103 perl_get_hv
1104 perl_get_cv
1105 perl_call_argv
1106 perl_call_pv
1107 perl_call_method
1108 perl_call_sv
1109 perl_requirepv
1110 safecalloc
1111 safemalloc
1112 saferealloc
1113 safefree
1114
1115This still needs much thought, but I'm inclined to think that one
1116possible solution is to prefix all such functions with C<perl_> in the
1117source and list them along with the other C<perl_*> functions in
1118F<perl_exp.SH>.
1119
1120Thus, for C<chsize>, we'd do something like the following:
1121
1122 /* in perl.h */
1123 #ifdef HAS_CHSIZE
1124 # define perl_chsize chsize
1125 #endif
1126
1127then in some file (e.g. F<util.c> or F<doio.c>) do
1128
1129 #ifndef HAS_CHSIZE
1130 I32 perl_chsize(fd, length)
1131 /* implement the function here . . . */
1132 #endif
1133
1134Alternatively, we could just always use C<chsize> everywhere and move
1135C<chsize> from F<global.sym> to the end of F<perl_exp.SH>. That would
1136probably be fine as long as our C<chsize> function agreed with all the
1137C<chsize> function prototypes in the various systems we'll be using.
1138As long as the prototypes in actual use don't vary that much, this is
1139probably a good alternative. (As a counter-example, note how Configure
1140and perl have to go through hoops to find and use get Malloc_t and
1141Free_t for C<malloc> and C<free>.)
1142
1143At the moment, this latter option is what I tend to prefer.
1144
1145=item All the world's a VAX
1146
1147Sorry, showing my age:-). Still, all the world is not BSD 4.[34],
1148SVR4, or POSIX. Be aware that SVR3-derived systems are still quite
1149common (do you have any idea how many systems run SCO?) If you don't
1150have a bunch of v7 manuals handy, the metaconfig units (by default
1151installed in F</usr/local/lib/dist/U>) are a good resource to look at
1152for portability.
1153
1154=back
1155
1156=head1 Miscellaneous Topics
1157
1158=head2 Autoconf
1159
1160Why does perl use a metaconfig-generated Configure script instead of an
1161autoconf-generated configure script?
1162
1163Metaconfig and autoconf are two tools with very similar purposes.
1164Metaconfig is actually the older of the two, and was originally written
1165by Larry Wall, while autoconf is probably now used in a wider variety of
1166packages. The autoconf info file discusses the history of autoconf and
1167how it came to be. The curious reader is referred there for further
1168information.
1169
1170Overall, both tools are quite good, I think, and the choice of which one
1171to use could be argued either way. In March, 1994, when I was just
1172starting to work on Configure support for Perl5, I considered both
1173autoconf and metaconfig, and eventually decided to use metaconfig for the
1174following reasons:
1175
1176=over 4
1177
1178=item Compatibility with Perl4
1179
1180Perl4 used metaconfig, so many of the #ifdef's were already set up for
1181metaconfig. Of course metaconfig had evolved some since Perl4's days,
1182but not so much that it posed any serious problems.
1183
1184=item Metaconfig worked for me
1185
d1be9408 1186My system at the time was Interactive 2.2, an SVR3.2/386 derivative that
aa689395 1187also had some POSIX support. Metaconfig-generated Configure scripts
1188worked fine for me on that system. On the other hand, autoconf-generated
1189scripts usually didn't. (They did come quite close, though, in some
1190cases.) At the time, I actually fetched a large number of GNU packages
1191and checked. Not a single one configured and compiled correctly
1192out-of-the-box with the system's cc compiler.
1193
1194=item Configure can be interactive
1195
1196With both autoconf and metaconfig, if the script works, everything is
1197fine. However, one of my main problems with autoconf-generated scripts
1198was that if it guessed wrong about something, it could be B<very> hard to
1199go back and fix it. For example, autoconf always insisted on passing the
1200-Xp flag to cc (to turn on POSIX behavior), even when that wasn't what I
1201wanted or needed for that package. There was no way short of editing the
1202configure script to turn this off. You couldn't just edit the resulting
1203Makefile at the end because the -Xp flag influenced a number of other
1204configure tests.
1205
1206Metaconfig's Configure scripts, on the other hand, can be interactive.
1207Thus if Configure is guessing things incorrectly, you can go back and fix
1208them. This isn't as important now as it was when we were actively
1209developing Configure support for new features such as dynamic loading,
1210but it's still useful occasionally.
1211
1212=item GPL
1213
1214At the time, autoconf-generated scripts were covered under the GNU Public
1215License, and hence weren't suitable for inclusion with Perl, which has a
1216different licensing policy. (Autoconf's licensing has since changed.)
1217
1218=item Modularity
1219
1220Metaconfig builds up Configure from a collection of discrete pieces
1221called "units". You can override the standard behavior by supplying your
1222own unit. With autoconf, you have to patch the standard files instead.
1223I find the metaconfig "unit" method easier to work with. Others
1224may find metaconfig's units clumsy to work with.
1225
1226=back
1227
aa689395 1228=head2 Why isn't there a directory to override Perl's library?
1229
1230Mainly because no one's gotten around to making one. Note that
1231"making one" involves changing perl.c, Configure, config_h.SH (and
1232associated files, see above), and I<documenting> it all in the
1233INSTALL file.
1234
1235Apparently, most folks who want to override one of the standard library
1236files simply do it by overwriting the standard library files.
1237
1238=head2 APPLLIB
1239
1240In the perl.c sources, you'll find an undocumented APPLLIB_EXP
1241variable, sort of like PRIVLIB_EXP and ARCHLIB_EXP (which are
1242documented in config_h.SH). Here's what APPLLIB_EXP is for, from
1243a mail message from Larry:
1244
1245 The main intent of APPLLIB_EXP is for folks who want to send out a
1246 version of Perl embedded in their product. They would set the symbol
1247 to be the name of the library containing the files needed to run or to
1248 support their particular application. This works at the "override"
1249 level to make sure they get their own versions of any library code that
1250 they absolutely must have configuration control over.
1251
1252 As such, I don't see any conflict with a sysadmin using it for a
1253 override-ish sort of thing, when installing a generic Perl. It should
1254 probably have been named something to do with overriding though. Since
1255 it's undocumented we could still change it... :-)
1256
1257Given that it's already there, you can use it to override
1258distribution modules. If you do
1259
1260 sh Configure -Dccflags='-DAPPLLIB_EXP=/my/override'
1261
1262then perl.c will put /my/override ahead of ARCHLIB and PRIVLIB.
1263
c4f23d77 1264=head2 Shared libperl.so location
1265
1266Why isn't the shared libperl.so installed in /usr/lib/ along
1267with "all the other" shared libraries? Instead, it is installed
1268in $archlib, which is typically something like
1269
1270 /usr/local/lib/perl5/archname/5.00404
1271
1272and is architecture- and version-specific.
1273
1274The basic reason why a shared libperl.so gets put in $archlib is so that
1275you can have more than one version of perl on the system at the same time,
1276and have each refer to its own libperl.so.
1277
1278Three examples might help. All of these work now; none would work if you
1279put libperl.so in /usr/lib.
1280
1281=over
1282
1283=item 1.
1284
1285Suppose you want to have both threaded and non-threaded perl versions
1286around. Configure will name both perl libraries "libperl.so" (so that
1287you can link to them with -lperl). The perl binaries tell them apart
1288by having looking in the appropriate $archlib directories.
1289
1290=item 2.
1291
1292Suppose you have perl5.004_04 installed and you want to try to compile
1293it again, perhaps with different options or after applying a patch.
1294If you already have libperl.so installed in /usr/lib/, then it may be
1295either difficult or impossible to get ld.so to find the new libperl.so
1296that you're trying to build. If, instead, libperl.so is tucked away in
1297$archlib, then you can always just change $archlib in the current perl
1298you're trying to build so that ld.so won't find your old libperl.so.
1299(The INSTALL file suggests you do this when building a debugging perl.)
1300
1301=item 3.
1302
1303The shared perl library is not a "well-behaved" shared library with
1304proper major and minor version numbers, so you can't necessarily
1305have perl5.004_04 and perl5.004_05 installed simultaneously. Suppose
1306perl5.004_04 were to install /usr/lib/libperl.so.4.4, and perl5.004_05
1307were to install /usr/lib/libperl.so.4.5. Now, when you try to run
1308perl5.004_04, ld.so might try to load libperl.so.4.5, since it has
1309the right "major version" number. If this works at all, it almost
1310certainly defeats the reason for keeping perl5.004_04 around. Worse,
1311with development subversions, you certaily can't guarantee that
1312libperl.so.4.4 and libperl.so.4.55 will be compatible.
1313
1314Anyway, all this leads to quite obscure failures that are sure to drive
1315casual users crazy. Even experienced users will get confused :-). Upon
1316reflection, I'd say leave libperl.so in $archlib.
1317
2032ff04 1318=item 4.
1319
1320Indentation style: over the years Perl has become a mishmash of
1321various indentation styles, but the original "Larry style" can
1322probably be restored with (GNU) indent somewhat like this:
1323
1324 indent -kr -nce -psl -sc
1325
55c0ed8c 1326A more ambitious solution would also specify a list of Perl specific
1327types with -TSV -TAV -THV .. -TMAGIC -TPerlIO ... but that list would
1328be quite ungainly. Also note that GNU indent also doesn't do aligning
1329of consecutive assignments, which would truly wreck the layout in
1330places like sv.c:Perl_sv_upgrade() or sv.c:Perl_clone_using().
1331Similarly nicely aligned &&s, ||s and ==s would not be respected.
2032ff04 1332
c4f23d77 1333=back
1334
aa689395 1335=head1 Upload Your Work to CPAN
1336
1337You can upload your work to CPAN if you have a CPAN id. Check out
a93751fa 1338http://www.cpan.org/modules/04pause.html for information on
aa689395 1339_PAUSE_, the Perl Author's Upload Server.
1340
1341I typically upload both the patch file, e.g. F<perl5.004_08.pat.gz>
1342and the full tar file, e.g. F<perl5.004_08.tar.gz>.
1343
1344If you want your patch to appear in the F<src/5.0/unsupported>
1345directory on CPAN, send e-mail to the CPAN master librarian. (Check
a93751fa 1346out http://www.cpan.org/CPAN.html ).
aa689395 1347
1348=head1 Help Save the World
1349
1350You should definitely announce your patch on the perl5-porters list.
1351You should also consider announcing your patch on
1352comp.lang.perl.announce, though you should make it quite clear that a
1353subversion is not a production release, and be prepared to deal with
1354people who will not read your disclaimer.
1355
1356=head1 Todo
1357
1358Here, in no particular order, are some Configure and build-related
1359items that merit consideration. This list isn't exhaustive, it's just
1360what I came up with off the top of my head.
1361
e25f343d 1362=head2 Adding missing library functions to Perl
1363
1364The perl Configure script automatically determines which headers and
1365functions you have available on your system and arranges for them to be
1366included in the compilation and linking process. Occasionally, when porting
1367perl to an operating system for the first time, you may find that the
1368operating system is missing a key function. While perl may still build
1369without this function, no perl program will be able to reference the missing
1370function. You may be able to write the missing function yourself, or you
1371may be able to find the missing function in the distribution files for
1372another software package. In this case, you need to instruct the perl
1373configure-and-build process to use your function. Perform these steps.
1374
1375=over 3
1376
1377=item *
1378
2ecb232b 1379Code and test the function you wish to add. Test it carefully; you will
e25f343d 1380have a much easier time debugging your code independently than when it is a
1381part of perl.
1382
1383=item *
1384
1385Here is an implementation of the POSIX truncate function for an operating
1386system (VOS) that does not supply one, but which does supply the ftruncate()
1387function.
1388
1389 /* Beginning of modification history */
1390 /* Written 02-01-02 by Nick Ing-Simmons (nick@ing-simmons.net) */
1391 /* End of modification history */
1392
1393 /* VOS doesn't supply a truncate function, so we build one up
1394 from the available POSIX functions. */
1395
1396 #include <fcntl.h>
1397 #include <sys/types.h>
1398 #include <unistd.h>
1399
1400 int
1401 truncate(const char *path, off_t len)
1402 {
1403 int fd = open(path,O_WRONLY);
1404 int code = -1;
1405 if (fd >= 0) {
1406 code = ftruncate(fd,len);
1407 close(fd);
1408 }
1409 return code;
1410 }
1411
1412Place this file into a subdirectory that has the same name as the operating
1413system. This file is named perl/vos/vos.c
1414
1415=item *
1416
1417If your operating system has a hints file (in perl/hints/XXX.sh for an
1418operating system named XXX), then start with it. If your operating system
1419has no hints file, then create one. You can use a hints file for a similar
1420operating system, if one exists, as a template.
1421
1422=item *
1423
1424Add lines like the following to your hints file. The first line
1425(d_truncate="define") instructs Configure that the truncate() function
1426exists. The second line (archobjs="vos.o") instructs the makefiles that the
1427perl executable depends on the existence of a file named "vos.o". (Make
1428will automatically look for "vos.c" and compile it with the same options as
1429the perl source code). The final line ("test -h...") adds a symbolic link
1430to the top-level directory so that make can find vos.c. Of course, you
1431should use your own operating system name for the source file of extensions,
1432not "vos.c".
1433
1434 # VOS does not have truncate() but we supply one in vos.c
1435 d_truncate="define"
1436 archobjs="vos.o"
1437
1438 # Help gmake find vos.c
1439 test -h vos.c || ln -s vos/vos.c vos.c
1440
1441The hints file is a series of shell commands that are run in the top-level
1442directory (the "perl" directory). Thus, these commands are simply executed
1443by Configure at an appropriate place during its execution.
1444
1445=item *
1446
1447At this point, you can run the Configure script and rebuild perl. Carefully
1448test the newly-built perl to ensure that normal paths, and error paths,
1449behave as you expect.
1450
1451=back
1452
aa689395 1453=head2 Good ideas waiting for round tuits
1454
1455=over 4
1456
c4f23d77 1457=item Configure -Dsrc=/blah/blah
aa689395 1458
1459We should be able to emulate B<configure --srcdir>. Tom Tromey
1460tromey@creche.cygnus.com has submitted some patches to
c4f23d77 1461the dist-users mailing list along these lines. They have been folded
1462back into the main distribution, but various parts of the perl
1463Configure/build/install process still assume src='.'.
aa689395 1464
1465=item Hint file fixes
1466
1467Various hint files work around Configure problems. We ought to fix
1468Configure so that most of them aren't needed.
1469
1470=item Hint file information
1471
1472Some of the hint file information (particularly dynamic loading stuff)
1473ought to be fed back into the main metaconfig distribution.
1474
1475=back
1476
1477=head2 Probably good ideas waiting for round tuits
1478
1479=over 4
1480
1481=item GNU configure --options
1482
1483I've received sensible suggestions for --exec_prefix and other
1484GNU configure --options. It's not always obvious exactly what is
1485intended, but this merits investigation.
1486
1487=item make clean
1488
1489Currently, B<make clean> isn't all that useful, though
1490B<make realclean> and B<make distclean> are. This needs a bit of
1491thought and documentation before it gets cleaned up.
1492
1493=item Try gcc if cc fails
1494
1495Currently, we just give up.
1496
1497=item bypassing safe*alloc wrappers
1498
1499On some systems, it may be safe to call the system malloc directly
1500without going through the util.c safe* layers. (Such systems would
1501accept free(0), for example.) This might be a time-saver for systems
1502that already have a good malloc. (Recent Linux libc's apparently have
1503a nice malloc that is well-tuned for the system.)
1504
1505=back
1506
1507=head2 Vague possibilities
1508
1509=over 4
1510
aa689395 1511=item MacPerl
1512
3e3baf6d 1513Get some of the Macintosh stuff folded back into the main distribution.
aa689395 1514
1515=item gconvert replacement
1516
1517Maybe include a replacement function that doesn't lose data in rare
1518cases of coercion between string and numerical values.
1519
aa689395 1520=item Improve makedepend
1521
1522The current makedepend process is clunky and annoyingly slow, but it
1523works for most folks. Alas, it assumes that there is a filename
1524$firstmakefile that the B<make> command will try to use before it uses
1525F<Makefile>. Such may not be the case for all B<make> commands,
1526particularly those on non-Unix systems.
1527
1528Probably some variant of the BSD F<.depend> file will be useful.
1529We ought to check how other packages do this, if they do it at all.
1530We could probably pre-generate the dependencies (with the exception of
1531malloc.o, which could probably be determined at F<Makefile.SH>
1532extraction time.
1533
1534=item GNU Makefile standard targets
1535
1536GNU software generally has standardized Makefile targets. Unless we
1537have good reason to do otherwise, I see no reason not to support them.
1538
1539=item File locking
1540
1541Somehow, straighten out, document, and implement lockf(), flock(),
76ba0908 1542and/or fcntl() file locking. It's a mess. See $d_fcntl_can_lock
1543in recent config.sh files though.
aa689395 1544
1545=back
1546
fb73857a 1547=head1 AUTHORS
aa689395 1548
fb73857a 1549Original author: Andy Dougherty doughera@lafcol.lafayette.edu .
1550Additions by Chip Salzenberg chip@perl.com and
1551Tim Bunce Tim.Bunce@ig.co.uk .
aa689395 1552
1553All opinions expressed herein are those of the authorZ<>(s).
1554
1555=head1 LAST MODIFIED
1556
ff935051 1557$Id: pumpkin.pod,v 1.23 2000/01/13 19:45:13 doughera Released $