This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Add "profile installman" to the TODO.
[perl5.git] / pod / perltodo.pod
CommitLineData
7711098a
GS
1=head1 NAME
2
3perltodo - Perl TO-DO List
4
5=head1 DESCRIPTION
e50bb9a1 6
52960e22
JC
7This is a list of wishes for Perl. The tasks we think are smaller or
8easier are listed first. Anyone is welcome to work on any of these,
9but it's a good idea to first contact I<perl5-porters@perl.org> to
10avoid duplication of effort, and to learn from any previous attempts.
11By all means contact a pumpking privately first if you prefer.
e50bb9a1 12
0bdfc961
NC
13Whilst patches to make the list shorter are most welcome, ideas to add to
14the list are also encouraged. Check the perl5-porters archives for past
15ideas, and any discussion about them. One set of archives may be found at:
e50bb9a1 16
0bdfc961 17 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
938c8732 18
617eabfa
NC
19What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
20not, but if your patch is incorporated, then we'll add your name to the
21F<AUTHORS> file, which ships in the official distribution. How many other
22programming languages offer you 1 line of immortality?
938c8732 23
0bdfc961 24=head1 Tasks that only need Perl knowledge
e50bb9a1 25
5a176cbc
NC
26=head2 Remove duplication of test setup.
27
28Schwern notes, that there's duplication of code - lots and lots of tests have
29some variation on the big block of C<$Is_Foo> checks. We can safely put this
30into a file, change it to build an C<%Is> hash and require it. Maybe just put
31it into F<test.pl>. Throw in the handy tainting subroutines.
32
0bdfc961 33=head2 common test code for timed bail out
e50bb9a1 34
0bdfc961
NC
35Write portable self destruct code for tests to stop them burning CPU in
36infinite loops. This needs to avoid using alarm, as some of the tests are
37testing alarm/sleep or timers.
e50bb9a1 38
87a942b1 39=head2 POD -E<gt> HTML conversion in the core still sucks
e50bb9a1 40
938c8732 41Which is crazy given just how simple POD purports to be, and how simple HTML
adebf063
NC
42can be. It's not actually I<as> simple as it sounds, particularly with the
43flexibility POD allows for C<=item>, but it would be good to improve the
44visual appeal of the HTML generated, and to avoid it having any validation
45errors. See also L</make HTML install work>, as the layout of installation tree
46is needed to improve the cross-linking.
938c8732 47
dc0fb092
SP
48The addition of C<Pod::Simple> and its related modules may make this task
49easier to complete.
50
8537f021
RGS
51=head2 merge checkpods and podchecker
52
53F<pod/checkpods.PL> (and C<make check> in the F<pod/> subdirectory)
54implements a very basic check for pod files, but the errors it discovers
55aren't found by podchecker. Add this check to podchecker, get rid of
56checkpods and have C<make check> use podchecker.
57
aa237293
NC
58=head2 Parallel testing
59
b2e2905c 60(This probably impacts much more than the core: also the Test::Harness
02f21748
RGS
61and TAP::* modules on CPAN.)
62
aa237293
NC
63The core regression test suite is getting ever more comprehensive, which has
64the side effect that it takes longer to run. This isn't so good. Investigate
65whether it would be feasible to give the harness script the B<option> of
66running sets of tests in parallel. This would be useful for tests in
67F<t/op/*.t> and F<t/uni/*.t> and maybe some sets of tests in F<lib/>.
68
69Questions to answer
70
71=over 4
72
73=item 1
74
75How does screen layout work when you're running more than one test?
76
77=item 2
78
79How does the caller of test specify how many tests to run in parallel?
80
81=item 3
82
83How do setup/teardown tests identify themselves?
84
85=back
86
87Pugs already does parallel testing - can their approach be re-used?
88
0bdfc961 89=head2 Make Schwern poorer
e50bb9a1 90
613bd4f7 91We should have tests for everything. When all the core's modules are tested,
0bdfc961
NC
92Schwern has promised to donate to $500 to TPF. We may need volunteers to
93hold him upside down and shake vigorously in order to actually extract the
94cash.
3958b146 95
0bdfc961 96=head2 Improve the coverage of the core tests
e50bb9a1 97
02f21748
RGS
98Use Devel::Cover to ascertain the core modules's test coverage, then add
99tests that are currently missing.
30222c0f 100
0bdfc961 101=head2 test B
e50bb9a1 102
0bdfc961 103A full test suite for the B module would be nice.
e50bb9a1 104
636e63cb
NC
105=head2 Deparse inlined constants
106
107Code such as this
108
109 use constant PI => 4;
110 warn PI
111
112will currently deparse as
113
114 use constant ('PI', 4);
115 warn 4;
116
117because the tokenizer inlines the value of the constant subroutine C<PI>.
118This allows various compile time optimisations, such as constant folding
119and dead code elimination. Where these haven't happened (such as the example
120above) it ought be possible to make B::Deparse work out the name of the
121original constant, because just enough information survives in the symbol
122table to do this. Specifically, the same scalar is used for the constant in
123the optree as is used for the constant subroutine, so by iterating over all
124symbol tables and generating a mapping of SV address to constant name, it
125would be possible to provide B::Deparse with this functionality.
126
0bdfc961 127=head2 A decent benchmark
e50bb9a1 128
617eabfa 129C<perlbench> seems impervious to any recent changes made to the perl core. It
0bdfc961
NC
130would be useful to have a reasonable general benchmarking suite that roughly
131represented what current perl programs do, and measurably reported whether
132tweaks to the core improve, degrade or don't really affect performance, to
133guide people attempting to optimise the guts of perl. Gisle would welcome
134new tests for perlbench.
6168cf99 135
0bdfc961 136=head2 fix tainting bugs
6168cf99 137
0bdfc961
NC
138Fix the bugs revealed by running the test suite with the C<-t> switch (via
139C<make test.taintwarn>).
e50bb9a1 140
0bdfc961 141=head2 Dual life everything
e50bb9a1 142
0bdfc961
NC
143As part of the "dists" plan, anything that doesn't belong in the smallest perl
144distribution needs to be dual lifed. Anything else can be too. Figure out what
145changes would be needed to package that module and its tests up for CPAN, and
146do so. Test it with older perl releases, and fix the problems you find.
e50bb9a1 147
a393eb28
RGS
148To make a minimal perl distribution, it's useful to look at
149F<t/lib/commonsense.t>.
150
c2aba5b8
RGS
151=head2 Bundle dual life modules in ext/
152
153For maintenance (and branch merging) reasons, it would be useful to move
154some architecture-independent dual-life modules from lib/ to ext/, if this
155has no negative impact on the build of perl itself.
156
157However, we need to make sure that they are still installed in
158architecture-independent directories by C<make install>.
159
0bdfc961 160=head2 Improving C<threads::shared>
722d2a37 161
0bdfc961
NC
162Investigate whether C<threads::shared> could share aggregates properly with
163only Perl level changes to shared.pm
722d2a37 164
0bdfc961 165=head2 POSIX memory footprint
e50bb9a1 166
0bdfc961
NC
167Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
168various times worked to cut it down. There is probably still fat to cut out -
169for example POSIX passes Exporter some very memory hungry data structures.
e50bb9a1 170
eed36644
NC
171=head2 embed.pl/makedef.pl
172
173There is a script F<embed.pl> that generates several header files to prefix
174all of Perl's symbols in a consistent way, to provide some semblance of
175namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
907b3e23 176in F<interpvar.h>. Quite a few of the functions and variables
eed36644
NC
177are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
178doesn't understand the C macros, so the rules about which symbols are present
179when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
180It would be good to teach C<embed.pl> to understand the conditional
181compilation, and hence remove the duplication, and the mistakes it has caused.
e50bb9a1 182
801de10e
NC
183=head2 use strict; and AutoLoad
184
185Currently if you write
186
187 package Whack;
188 use AutoLoader 'AUTOLOAD';
189 use strict;
190 1;
191 __END__
192 sub bloop {
193 print join (' ', No, strict, here), "!\n";
194 }
195
196then C<use strict;> isn't in force within the autoloaded subroutines. It would
197be more consistent (and less surprising) to arrange for all lexical pragmas
198in force at the __END__ block to be in force within each autoloaded subroutine.
199
773b3597
RGS
200There's a similar problem with SelfLoader.
201
91d0cbf6
NC
202=head2 profile installman
203
204The F<installman> script is slow. All it is doing text processing, which we're
205told is something Perl is good at. So it would be nice to know what it is doing
206that is taking so much CPU, and where possible address it.
207
208
0bdfc961 209=head1 Tasks that need a little sysadmin-type knowledge
e50bb9a1 210
0bdfc961
NC
211Or if you prefer, tasks that you would learn from, and broaden your skills
212base...
e50bb9a1 213
cd793d32 214=head2 make HTML install work
e50bb9a1 215
adebf063
NC
216There is an C<installhtml> target in the Makefile. It's marked as
217"experimental". It would be good to get this tested, make it work reliably, and
218remove the "experimental" tag. This would include
219
220=over 4
221
222=item 1
223
224Checking that cross linking between various parts of the documentation works.
225In particular that links work between the modules (files with POD in F<lib/>)
226and the core documentation (files in F<pod/>)
227
228=item 2
229
617eabfa
NC
230Work out how to split C<perlfunc> into chunks, preferably one per function
231group, preferably with general case code that could be used elsewhere.
232Challenges here are correctly identifying the groups of functions that go
233together, and making the right named external cross-links point to the right
234page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
235C<endservent>, two or more C<=items> giving the different parameter lists, such
236as
adebf063
NC
237
238 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
adebf063 239 =item substr EXPR,OFFSET,LENGTH
adebf063
NC
240 =item substr EXPR,OFFSET
241
242and different parameter lists having different meanings. (eg C<select>)
243
244=back
3a89a73c 245
0bdfc961
NC
246=head2 compressed man pages
247
248Be able to install them. This would probably need a configure test to see how
249the system does compressed man pages (same directory/different directory?
250same filename/different filename), as well as tweaking the F<installman> script
251to compress as necessary.
252
30222c0f
NC
253=head2 Add a code coverage target to the Makefile
254
255Make it easy for anyone to run Devel::Cover on the core's tests. The steps
256to do this manually are roughly
257
258=over 4
259
260=item *
261
262do a normal C<Configure>, but include Devel::Cover as a module to install
263(see F<INSTALL> for how to do this)
264
265=item *
266
267 make perl
268
269=item *
270
271 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
272
273=item *
274
275Process the resulting Devel::Cover database
276
277=back
278
279This just give you the coverage of the F<.pm>s. To also get the C level
280coverage you need to
281
282=over 4
283
284=item *
285
286Additionally tell C<Configure> to use the appropriate C compiler flags for
287C<gcov>
288
289=item *
290
291 make perl.gcov
292
293(instead of C<make perl>)
294
295=item *
296
297After running the tests run C<gcov> to generate all the F<.gcov> files.
298(Including down in the subdirectories of F<ext/>
299
300=item *
301
302(From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
303to get their stats into the cover_db directory.
304
305=item *
306
307Then process the Devel::Cover database
308
309=back
310
311It would be good to add a single switch to C<Configure> to specify that you
312wanted to perform perl level coverage, and another to specify C level
313coverage, and have C<Configure> and the F<Makefile> do all the right things
314automatically.
315
02f21748 316=head2 Make Config.pm cope with differences between built and installed perl
0bdfc961
NC
317
318Quite often vendors ship a perl binary compiled with their (pay-for)
319compilers. People install a free compiler, such as gcc. To work out how to
320build extensions, Perl interrogates C<%Config>, so in this situation
321C<%Config> describes compilers that aren't there, and extension building
322fails. This forces people into choosing between re-compiling perl themselves
323using the compiler they have, or only using modules that the vendor ships.
324
325It would be good to find a way teach C<Config.pm> about the installation setup,
326possibly involving probing at install time or later, so that the C<%Config> in
327a binary distribution better describes the installed machine, when the
328installed machine differs from the build machine in some significant way.
329
728f4ecd
NC
330=head2 linker specification files
331
332Some platforms mandate that you provide a list of a shared library's external
333symbols to the linker, so the core already has the infrastructure in place to
334do this for generating shared perl libraries. My understanding is that the
335GNU toolchain can accept an optional linker specification file, and restrict
336visibility just to symbols declared in that file. It would be good to extend
337F<makedef.pl> to support this format, and to provide a means within
338C<Configure> to enable it. This would allow Unix users to test that the
339export list is correct, and to build a perl that does not pollute the global
340namespace with private symbols.
341
a229ae3b
RGS
342=head2 Cross-compile support
343
344Currently C<Configure> understands C<-Dusecrosscompile> option. This option
345arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
346assumed then to be copied to TARGET machine and used as a replacement of full
347C<perl> executable.
348
d1307786 349This could be done little differently. Namely C<miniperl> should be built for
a229ae3b 350HOST and then full C<perl> with extensions should be compiled for TARGET.
d1307786 351This, however, might require extra trickery for %Config: we have one config
87a942b1
JH
352first for HOST and then another for TARGET. Tools like MakeMaker will be
353mightily confused. Having around two different types of executables and
354libraries (HOST and TARGET) makes life interesting for Makefiles and
355shell (and Perl) scripts. There is $Config{run}, normally empty, which
356can be used as an execution wrapper. Also note that in some
357cross-compilation/execution environments the HOST and the TARGET do
358not see the same filesystem(s), the $Config{run} may need to do some
359file/directory copying back and forth.
0bdfc961 360
8537f021
RGS
361=head2 roffitall
362
363Make F<pod/roffitall> be updated by F<pod/buildtoc>.
364
0bdfc961
NC
365=head1 Tasks that need a little C knowledge
366
367These tasks would need a little C knowledge, but don't need any specific
368background or experience with XS, or how the Perl interpreter works
369
3d826b29
NC
370=head2 Weed out needless PERL_UNUSED_ARG
371
372The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about
373unused arguments. Often the arguments can't be removed, as there is an
374external constraint that determines the prototype of the function, so this
375approach is valid. However, there are some cases where C<PERL_UNUSED_ARG>
376could be removed. Specifically
377
378=over 4
379
380=item *
381
382The prototypes of (nearly all) static functions can be changed
383
384=item *
385
386Unused arguments generated by short cut macros are wasteful - the short cut
387macro used can be changed.
388
389=back
390
fbf638cb
RGS
391=head2 Modernize the order of directories in @INC
392
393The way @INC is laid out by default, one cannot upgrade core (dual-life)
394modules without overwriting files. This causes problems for binary
3d14fd97
AD
395package builders. One possible proposal is laid out in this
396message:
397L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>.
fbf638cb 398
bcbaa2d5
RGS
399=head2 -Duse32bit*
400
401Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
402On these systems, it might be the default compilation mode, and there
403is currently no guarantee that passing no use64bitall option to the
404Configure process will build a 32bit perl. Implementing -Duse32bit*
405options would be nice for perl 5.12.
406
0bdfc961 407=head2 Make it clear from -v if this is the exact official release
89007cb3 408
617eabfa
NC
409Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that
410usually defines one local patch, of the form "MAINT12345" or "RC1". The output
411of perl -v doesn't report that a perl isn't an official release, and this
89007cb3 412information can get lost in bugs reports. Because of this, the minor version
fa11829f 413isn't bumped up until RC time, to minimise the possibility of versions of perl
89007cb3
NC
414escaping that believe themselves to be newer than they actually are.
415
416It would be useful to find an elegant way to have the "this is an interim
417maintenance release" or "this is a release candidate" in the terse -v output,
418and have it so that it's easy for the pumpking to remove this just as the
419release tarball is rolled up. This way the version pulled out of rsync would
420always say "I'm a development release" and it would be safe to bump the
421reported minor version as soon as a release ships, which would aid perl
422developers.
423
0bdfc961
NC
424This task is really about thinking of an elegant way to arrange the C source
425such that it's trivial for the Pumpking to flag "this is an official release"
426when making a tarball, yet leave the default source saying "I'm not the
427official release".
428
fee0a0f7 429=head2 Profile Perl - am I hot or not?
62403a3c 430
fee0a0f7
NC
431The Perl source code is stable enough that it makes sense to profile it,
432identify and optimise the hotspots. It would be good to measure the
433performance of the Perl interpreter using free tools such as cachegrind,
434gprof, and dtrace, and work to reduce the bottlenecks they reveal.
435
436As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
437the ops that are most commonly used. The idea is that by grouping them, their
438object code will be adjacent in the executable, so they have a greater chance
439of already being in the CPU cache (or swapped in) due to being near another op
440already in use.
62403a3c
NC
441
442Except that it's not clear if these really are the most commonly used ops. So
fee0a0f7
NC
443as part of exercising your skills with coverage and profiling tools you might
444want to determine what ops I<really> are the most commonly used. And in turn
445suggest evictions and promotions to achieve a better F<pp_hot.c>.
62403a3c 446
91d0cbf6
NC
447One piece of Perl code that might make a good testbed is F<installman>.
448
98fed0ad
NC
449=head2 Allocate OPs from arenas
450
451Currently all new OP structures are individually malloc()ed and free()d.
452All C<malloc> implementations have space overheads, and are now as fast as
453custom allocates so it would both use less memory and less CPU to allocate
454the various OP structures from arenas. The SV arena code can probably be
455re-used for this.
456
539f2c54
JC
457Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use
458Perl_Slab_alloc() to pack optrees into a contiguous block, which is
459probably superior to the use of OP arenas, esp. from a cache locality
460standpoint. See L<Profile Perl - am I hot or not?>.
461
a229ae3b 462=head2 Improve win32/wince.c
0bdfc961 463
a229ae3b 464Currently, numerous functions look virtually, if not completely,
02f21748 465identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't
6d71adcd
NC
466be good.
467
c5b31784
SH
468=head2 Use secure CRT functions when building with VC8 on Win32
469
470Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
471that they were "unsafe" and introduced differently named secure versions of
472them as replacements, e.g. instead of writing
473
474 FILE* f = fopen(__FILE__, "r");
475
476one should now write
477
478 FILE* f;
479 errno_t err = fopen_s(&f, __FILE__, "r");
480
481Currently, the warnings about these deprecations have been disabled by adding
482-D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
483warning suppressant and actually make use of the new secure CRT functions.
484
485There is also a similar issue with POSIX CRT function names like fileno having
486been deprecated in favour of ISO C++ conformant names like _fileno. These
26a6faa8 487warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
c5b31784
SH
488might be nice to do as Microsoft suggest here too, although, unlike the secure
489functions issue, there is presumably little or no benefit in this case.
490
038ae9a4
SH
491=head2 Fix POSIX::access() and chdir() on Win32
492
493These functions currently take no account of DACLs and therefore do not behave
494correctly in situations where access is restricted by DACLs (as opposed to the
495read-only attribute).
496
497Furthermore, POSIX::access() behaves differently for directories having the
498read-only attribute set depending on what CRT library is being used. For
499example, the _access() function in the VC6 and VC7 CRTs (wrongly) claim that
500such directories are not writable, whereas in fact all directories are writable
501unless access is denied by DACLs. (In the case of directories, the read-only
502attribute actually only means that the directory cannot be deleted.) This CRT
503bug is fixed in the VC8 and VC9 CRTs (but, of course, the directory may still
504not actually be writable if access is indeed denied by DACLs).
505
506For the chdir() issue, see ActiveState bug #74552:
507http://bugs.activestate.com/show_bug.cgi?id=74552
508
509Therefore, DACLs should be checked both for consistency across CRTs and for
510the correct answer.
511
512(Note that perl's -w operator should not be modified to check DACLs. It has
513been written so that it reflects the state of the read-only attribute, even
514for directories (whatever CRT is being used), for symmetry with chmod().)
515
16815324
NC
516=head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
517
518Maybe create a utility that checks after each libperl.a creation that
519none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
520ever creep back to libperl.a.
521
522 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
523
524Note, of course, that this will only tell whether B<your> platform
525is using those naughty interfaces.
526
de96509d
JH
527=head2 -D_FORTIFY_SOURCE=2, -fstack-protector
528
529Recent glibcs support C<-D_FORTIFY_SOURCE=2> and recent gcc
530(4.1 onwards?) supports C<-fstack-protector>, both of which give
531protection against various kinds of buffer overflow problems.
532These should probably be used for compiling Perl whenever available,
533Configure and/or hints files should be adjusted to probe for the
534availability of these features and enable them as appropriate.
16815324 535
8964cfe0
NC
536=head2 Arenas for GPs? For MAGIC?
537
538C<struct gp> and C<struct magic> are both currently allocated by C<malloc>.
539It might be a speed or memory saving to change to using arenas. Or it might
540not. It would need some suitable benchmarking first. In particular, C<GP>s
541can probably be changed with minimal compatibility impact (probably nothing
542outside of the core, or even outside of F<gv.c> allocates them), but they
543probably aren't allocated/deallocated often enough for a speed saving. Whereas
544C<MAGIC> is allocated/deallocated more often, but in turn, is also something
545more externally visible, so changing the rules here may bite external code.
546
547
6d71adcd
NC
548=head1 Tasks that need a knowledge of XS
549
550These tasks would need C knowledge, and roughly the level of knowledge of
551the perl API that comes from writing modules that use XS to interface to
552C.
553
839c1407
NC
554=head2 investigate removing int_macro_int from POSIX.xs
555
556As a hang over from the original C<constant> implementation, F<POSIX.xs>
557contains a function C<int_macro_int> which in conjunction with C<AUTOLOAD> is
558used to wrap the C functions C<WEXITSTATUS>, C<WIFEXITED>, C<WIFSIGNALED>,
559C<WIFSTOPPED>, C<WSTOPSIG> and C<WTERMSIG>. It's probably worth replacing
560this complexity with 5 simple direct wrappings of those 5 functions.
561
562However, it would be interesting if someone could measure the memory usage
563before and after, both for the case of C<use POSIX();> and the case of
564actually calling the Perl space functions.
565
5d96f598
NC
566=head2 safely supporting POSIX SA_SIGINFO
567
568Some years ago Jarkko supplied patches to provide support for the POSIX
569SA_SIGINFO feature in Perl, passing the extra data to the Perl signal handler.
570
571Unfortunately, it only works with "unsafe" signals, because under safe
572signals, by the time Perl gets to run the signal handler, the extra
573information has been lost. Moreover, it's not easy to store it somewhere,
574as you can't call mutexs, or do anything else fancy, from inside a signal
575handler.
576
577So it strikes me that we could provide safe SA_SIGINFO support
578
579=over 4
580
581=item 1
582
583Provide global variables for two file descriptors
584
585=item 2
586
587When the first request is made via C<sigaction> for C<SA_SIGINFO>, create a
588pipe, store the reader in one, the writer in the other
589
590=item 3
591
592In the "safe" signal handler (C<Perl_csighandler()>/C<S_raise_signal()>), if
593the C<siginfo_t> pointer non-C<NULL>, and the writer file handle is open,
594
595=over 8
596
597=item 1
598
599serialise signal number, C<struct siginfo_t> (or at least the parts we care
600about) into a small auto char buff
601
602=item 2
603
604C<write()> that (non-blocking) to the writer fd
605
606=over 12
607
608=item 1
609
610if it writes 100%, flag the signal in a counter of "signals on the pipe" akin
611to the current per-signal-number counts
612
613=item 2
614
615if it writes 0%, assume the pipe is full. Flag the data as lost?
616
617=item 3
618
619if it writes partially, croak a panic, as your OS is broken.
620
621=back
622
623=back
624
625=item 4
626
627in the regular C<PERL_ASYNC_CHECK()> processing, if there are "signals on
628the pipe", read the data out, deserialise, build the Perl structures on
629the stack (code in C<Perl_sighandler()>, the "unsafe" handler), and call as
630usual.
631
632=back
633
634I think that this gets us decent C<SA_SIGINFO> support, without the current risk
635of running Perl code inside the signal handler context. (With all the dangers
636of things like C<malloc> corruption that that currently offers us)
637
638For more information see the thread starting with this message:
639http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-03/msg00305.html
640
6d71adcd
NC
641=head2 autovivification
642
643Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
644
645This task is incremental - even a little bit of work on it will help.
646
647=head2 Unicode in Filenames
648
649chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
650opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
651system, truncate, unlink, utime, -X. All these could potentially accept
652Unicode filenames either as input or output (and in the case of system
653and qx Unicode in general, as input or output to/from the shell).
654Whether a filesystem - an operating system pair understands Unicode in
655filenames varies.
656
657Known combinations that have some level of understanding include
658Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
659OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
660create Unicode filenames, what forms of Unicode are accepted and used
661(UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
662and so on, varies. Finding the right level of interfacing to Perl
663requires some thought. Remember that an OS does not implicate a
664filesystem.
665
666(The Windows -C command flag "wide API support" has been at least
667temporarily retired in 5.8.1, and the -C has been repurposed, see
668L<perlrun>.)
669
87a942b1
JH
670Most probably the right way to do this would be this:
671L</"Virtualize operating system access">.
672
6d71adcd
NC
673=head2 Unicode in %ENV
674
675Currently the %ENV entries are always byte strings.
87a942b1 676See L</"Virtualize operating system access">.
6d71adcd 677
1f2e7916
JD
678=head2 Unicode and glob()
679
680Currently glob patterns and filenames returned from File::Glob::glob()
87a942b1 681are always byte strings. See L</"Virtualize operating system access">.
1f2e7916 682
dbb0c492
RGS
683=head2 Unicode and lc/uc operators
684
685Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on
686what the internal encoding of their argument is. That should not be the
687case. Maybe add a pragma to switch behaviour.
688
6d71adcd
NC
689=head2 use less 'memory'
690
691Investigate trade offs to switch out perl's choices on memory usage.
692Particularly perl should be able to give memory back.
693
694This task is incremental - even a little bit of work on it will help.
695
696=head2 Re-implement C<:unique> in a way that is actually thread-safe
697
698The old implementation made bad assumptions on several levels. A good 90%
699solution might be just to make C<:unique> work to share the string buffer
700of SvPVs. That way large constant strings can be shared between ithreads,
701such as the configuration information in F<Config>.
702
703=head2 Make tainting consistent
704
705Tainting would be easier to use if it didn't take documented shortcuts and
706allow taint to "leak" everywhere within an expression.
707
708=head2 readpipe(LIST)
709
710system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
711running a shell. readpipe() (the function behind qx//) could be similarly
712extended.
713
6d71adcd
NC
714=head2 Audit the code for destruction ordering assumptions
715
716Change 25773 notes
717
718 /* Need to check SvMAGICAL, as during global destruction it may be that
719 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
720 is now part of the linked list of SV heads, rather than pointing to
721 the original body. */
722 /* FIXME - audit the code for other bugs like this one. */
723
724adding the C<SvMAGICAL> check to
725
726 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
727 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
728
729Go through the core and look for similar assumptions that SVs have particular
730types, as all bets are off during global destruction.
731
749904bf
JH
732=head2 Extend PerlIO and PerlIO::Scalar
733
734PerlIO::Scalar doesn't know how to truncate(). Implementing this
735would require extending the PerlIO vtable.
736
737Similarly the PerlIO vtable doesn't know about formats (write()), or
738about stat(), or chmod()/chown(), utime(), or flock().
739
740(For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
741would mean.)
742
743PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
744opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
745readlink().
746
94da6c29
JH
747See also L</"Virtualize operating system access">.
748
3236f110
NC
749=head2 -C on the #! line
750
751It should be possible to make -C work correctly if found on the #! line,
752given that all perl command line options are strict ASCII, and -C changes
753only the interpretation of non-ASCII characters, and not for the script file
754handle. To make it work needs some investigation of the ordering of function
755calls during startup, and (by implication) a bit of tweaking of that order.
756
d6c1e11f
JH
757=head2 Organize error messages
758
759Perl's diagnostics (error messages, see L<perldiag>) could use
a8d0aeb9 760reorganizing and formalizing so that each error message has its
d6c1e11f
JH
761stable-for-all-eternity unique id, categorized by severity, type, and
762subsystem. (The error messages would be listed in a datafile outside
c4bd451b
CB
763of the Perl source code, and the source code would only refer to the
764messages by the id.) This clean-up and regularizing should apply
d6c1e11f
JH
765for all croak() messages.
766
767This would enable all sorts of things: easier translation/localization
768of the messages (though please do keep in mind the caveats of
769L<Locale::Maketext> about too straightforward approaches to
770translation), filtering by severity, and instead of grepping for a
771particular error message one could look for a stable error id. (Of
772course, changing the error messages by default would break all the
773existing software depending on some particular error message...)
774
775This kind of functionality is known as I<message catalogs>. Look for
776inspiration for example in the catgets() system, possibly even use it
777if available-- but B<only> if available, all platforms will B<not>
de96509d 778have catgets().
d6c1e11f
JH
779
780For the really pure at heart, consider extending this item to cover
781also the warning messages (see L<perllexwarn>, C<warnings.pl>).
3236f110 782
0bdfc961 783=head1 Tasks that need a knowledge of the interpreter
3298bd4d 784
0bdfc961
NC
785These tasks would need C knowledge, and knowledge of how the interpreter works,
786or a willingness to learn.
3298bd4d 787
718140ec
NC
788=head2 lexicals used only once
789
790This warns:
791
792 $ perl -we '$pie = 42'
793 Name "main::pie" used only once: possible typo at -e line 1.
794
795This does not:
796
797 $ perl -we 'my $pie = 42'
798
799Logically all lexicals used only once should warn, if the user asks for
d6f4ea2e
SP
800warnings. An unworked RT ticket (#5087) has been open for almost seven
801years for this discrepancy.
718140ec 802
a3d15f9a
RGS
803=head2 UTF-8 revamp
804
805The handling of Unicode is unclean in many places. For example, the regexp
806engine matches in Unicode semantics whenever the string or the pattern is
807flagged as UTF-8, but that should not be dependent on an internal storage
808detail of the string. Likewise, case folding behaviour is dependent on the
809UTF8 internal flag being on or off.
810
811=head2 Properly Unicode safe tokeniser and pads.
812
813The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
814variable names are stored in stashes as raw bytes, without the utf-8 flag
815set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
816tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
817source filters. All this could be fixed.
818
636e63cb
NC
819=head2 state variable initialization in list context
820
821Currently this is illegal:
822
823 state ($a, $b) = foo();
824
a2874905 825In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different
a8d0aeb9 826semantics, which is tricky to implement in Perl 5 as currently they produce
a2874905 827the same opcode trees. The Perl 6 design is firm, so it would be good to
a8d0aeb9 828implement the necessary code in Perl 5. There are comments in
a2874905
NC
829C<Perl_newASSIGNOP()> that show the code paths taken by various assignment
830constructions involving state variables.
636e63cb 831
4fedb12c
RGS
832=head2 Implement $value ~~ 0 .. $range
833
834It would be nice to extend the syntax of the C<~~> operator to also
835understand numeric (and maybe alphanumeric) ranges.
a393eb28
RGS
836
837=head2 A does() built-in
838
839Like ref(), only useful. It would call the C<DOES> method on objects; it
840would also tell whether something can be dereferenced as an
841array/hash/etc., or used as a regexp, etc.
842L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
843
844=head2 Tied filehandles and write() don't mix
845
846There is no method on tied filehandles to allow them to be called back by
847formats.
4fedb12c 848
d10fc472 849=head2 Attach/detach debugger from running program
1626a787 850
cd793d32
NC
851The old perltodo notes "With C<gdb>, you can attach the debugger to a running
852program if you pass the process ID. It would be good to do this with the Perl
0bdfc961
NC
853debugger on a running Perl program, although I'm not sure how it would be
854done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
1626a787 855
a8cb5b9e
RGS
856=head2 Optimize away empty destructors
857
858Defining an empty DESTROY method might be useful (notably in
859AUTOLOAD-enabled classes), but it's still a bit expensive to call. That
860could probably be optimized.
861
0bdfc961
NC
862=head2 LVALUE functions for lists
863
864The old perltodo notes that lvalue functions don't work for list or hash
865slices. This would be good to fix.
866
867=head2 LVALUE functions in the debugger
868
869The old perltodo notes that lvalue functions don't work in the debugger. This
870would be good to fix.
871
0bdfc961
NC
872=head2 regexp optimiser optional
873
874The regexp optimiser is not optional. It should configurable to be, to allow
875its performance to be measured, and its bugs to be easily demonstrated.
876
02f21748
RGS
877=head2 delete &function
878
879Allow to delete functions. One can already undef them, but they're still
880in the stash.
881
ef36c6a7
RGS
882=head2 C</w> regex modifier
883
884That flag would enable to match whole words, and also to interpolate
885arrays as alternations. With it, C</P/w> would be roughly equivalent to:
886
887 do { local $"='|'; /\b(?:P)\b/ }
888
889See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
890for the discussion.
891
0bdfc961
NC
892=head2 optional optimizer
893
894Make the peephole optimizer optional. Currently it performs two tasks as
895it walks the optree - genuine peephole optimisations, and necessary fixups of
896ops. It would be good to find an efficient way to switch out the
897optimisations whilst keeping the fixups.
898
899=head2 You WANT *how* many
900
901Currently contexts are void, scalar and list. split has a special mechanism in
902place to pass in the number of return values wanted. It would be useful to
903have a general mechanism for this, backwards compatible and little speed hit.
904This would allow proposals such as short circuiting sort to be implemented
905as a module on CPAN.
906
907=head2 lexical aliases
908
909Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
910
911=head2 entersub XS vs Perl
912
913At the moment pp_entersub is huge, and has code to deal with entering both
914perl and XS subroutines. Subroutine implementations rarely change between
915perl and XS at run time, so investigate using 2 ops to enter subs (one for
916XS, one for perl) and swap between if a sub is redefined.
2810d901 917
de535794 918=head2 Self-ties
2810d901 919
de535794 920Self-ties are currently illegal because they caused too many segfaults. Maybe
a8d0aeb9 921the causes of these could be tracked down and self-ties on all types
de535794 922reinstated.
0bdfc961
NC
923
924=head2 Optimize away @_
925
926The old perltodo notes "Look at the "reification" code in C<av.c>".
927
87a942b1
JH
928=head2 Virtualize operating system access
929
930Implement a set of "vtables" that virtualizes operating system access
931(open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
932least these interfaces should take SVs as "name" arguments instead of
933bare char pointers; probably the most flexible and extensible way
e1a3d5d1
JH
934would be for the Perl-facing interfaces to accept HVs. The system
935needs to be per-operating-system and per-file-system
936hookable/filterable, preferably both from XS and Perl level
87a942b1
JH
937(L<perlport/"Files and Filesystems"> is good reading at this point,
938in fact, all of L<perlport> is.)
939
e1a3d5d1
JH
940This has actually already been implemented (but only for Win32),
941take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
942variants go through a set of "vtables" for operating system access,
943non-Win32 systems currently go straight for the POSIX/UNIX-style
944system/library call. Similar system as for Win32 should be
945implemented for all platforms. The existing Win32 implementation
946probably does not need to survive alongside this proposed new
947implementation, the approaches could be merged.
87a942b1
JH
948
949What would this give us? One often-asked-for feature this would
94da6c29
JH
950enable is using Unicode for filenames, and other "names" like %ENV,
951usernames, hostnames, and so forth.
952(See L<perlunicode/"When Unicode Does Not Happen">.)
953
954But this kind of virtualization would also allow for things like
955virtual filesystems, virtual networks, and "sandboxes" (though as long
956as dynamic loading of random object code is allowed, not very safe
957sandboxes since external code of course know not of Perl's vtables).
958An example of a smaller "sandbox" is that this feature can be used to
959implement per-thread working directories: Win32 already does this.
960
961See also L</"Extend PerlIO and PerlIO::Scalar">.
87a942b1 962
ac6197af
NC
963=head2 Investigate PADTMP hash pessimisation
964
965The peephole optimier converts constants used for hash key lookups to shared
057163d7 966hash key scalars. Under ithreads, something is undoing this work.
ac6197af
NC
967See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html
968
057163d7
NC
969=head2 Store the current pad in the OP slab allocator
970
971=for clarification
972I hope that I got that "current pad" part correct
973
974Currently we leak ops in various cases of parse failure. I suggested that we
975could solve this by always using the op slab allocator, and walking it to
976free ops. Dave comments that as some ops are already freed during optree
977creation one would have to mark which ops are freed, and not double free them
978when walking the slab. He notes that one problem with this is that for some ops
979you have to know which pad was current at the time of allocation, which does
980change. I suggested storing a pointer to the current pad in the memory allocated
981for the slab, and swapping to a new slab each time the pad changes. Dave thinks
982that this would work.
983
52960e22
JC
984=head2 repack the optree
985
986Repacking the optree after execution order is determined could allow
057163d7
NC
987removal of NULL ops, and optimal ordering of OPs with respect to cache-line
988filling. The slab allocator could be reused for this purpose. I think that
989the best way to do this is to make it an optional step just before the
990completed optree is attached to anything else, and to use the slab allocator
991unchanged, so that freeing ops is identical whether or not this step runs.
992Note that the slab allocator allocates ops downwards in memory, so one would
993have to actually "allocate" the ops in reverse-execution order to get them
994contiguous in memory in execution order.
995
996See http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html
997
998Note that running this copy, and then freeing all the old location ops would
999cause their slabs to be freed, which would eliminate possible memory wastage if
1000the previous suggestion is implemented, and we swap slabs more frequently.
52960e22 1001
12e06b6f
NC
1002=head2 eliminate incorrect line numbers in warnings
1003
1004This code
1005
1006 use warnings;
1007 my $undef;
1008
1009 if ($undef == 3) {
1010 } elsif ($undef == 0) {
1011 }
1012
18a16cc5 1013used to produce this output:
12e06b6f
NC
1014
1015 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1016 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1017
18a16cc5
NC
1018where the line of the second warning was misreported - it should be line 5.
1019Rafael fixed this - the problem arose because there was no nextstate OP
1020between the execution of the C<if> and the C<elsif>, hence C<PL_curcop> still
1021reports that the currently executing line is line 4. The solution was to inject
1022a nextstate OPs for each C<elsif>, although it turned out that the nextstate
1023OP needed to be a nulled OP, rather than a live nextstate OP, else other line
1024numbers became misreported. (Jenga!)
12e06b6f
NC
1025
1026The problem is more general than C<elsif> (although the C<elsif> case is the
1027most common and the most confusing). Ideally this code
1028
1029 use warnings;
1030 my $undef;
1031
1032 my $a = $undef + 1;
1033 my $b
1034 = $undef
1035 + 1;
1036
1037would produce this output
1038
1039 Use of uninitialized value $undef in addition (+) at wrong.pl line 4.
1040 Use of uninitialized value $undef in addition (+) at wrong.pl line 7.
1041
1042(rather than lines 4 and 5), but this would seem to require every OP to carry
1043(at least) line number information.
1044
1045What might work is to have an optional line number in memory just before the
1046BASEOP structure, with a flag bit in the op to say whether it's present.
1047Initially during compile every OP would carry its line number. Then add a late
1048pass to the optimiser (potentially combined with L</repack the optree>) which
1049looks at the two ops on every edge of the graph of the execution path. If
1050the line number changes, flags the destination OP with this information.
1051Once all paths are traced, replace every op with the flag with a
1052nextstate-light op (that just updates C<PL_curcop>), which in turn then passes
1053control on to the true op. All ops would then be replaced by variants that
1054do not store the line number. (Which, logically, why it would work best in
1055conjunction with L</repack the optree>, as that is already copying/reallocating
1056all the OPs)
1057
18a16cc5
NC
1058(Although I should note that we're not certain that doing this for the general
1059case is worth it)
1060
52960e22
JC
1061=head2 optimize tail-calls
1062
1063Tail-calls present an opportunity for broadly applicable optimization;
1064anywhere that C<< return foo(...) >> is called, the outer return can
1065be replaced by a goto, and foo will return directly to the outer
1066caller, saving (conservatively) 25% of perl's call&return cost, which
1067is relatively higher than in C. The scheme language is known to do
1068this heavily. B::Concise provides good insight into where this
1069optimization is possible, ie anywhere entersub,leavesub op-sequence
1070occurs.
1071
1072 perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)'
1073
1074Bottom line on this is probably a new pp_tailcall function which
1075combines the code in pp_entersub, pp_leavesub. This should probably
1076be done 1st in XS, and using B::Generate to patch the new OP into the
1077optrees.
1078
0bdfc961
NC
1079=head1 Big projects
1080
1081Tasks that will get your name mentioned in the description of the "Highlights
87a942b1 1082of 5.12"
0bdfc961
NC
1083
1084=head2 make ithreads more robust
1085
4e577f8b 1086Generally make ithreads more robust. See also L</iCOW>
0bdfc961
NC
1087
1088This task is incremental - even a little bit of work on it will help, and
1089will be greatly appreciated.
1090
6c047da7
YST
1091One bit would be to write the missing code in sv.c:Perl_dirp_dup.
1092
59c7f7d5
RGS
1093Fix Perl_sv_dup, et al so that threads can return objects.
1094
0bdfc961
NC
1095=head2 iCOW
1096
1097Sarathy and Arthur have a proposal for an improved Copy On Write which
1098specifically will be able to COW new ithreads. If this can be implemented
1099it would be a good thing.
1100
1101=head2 (?{...}) closures in regexps
1102
1103Fix (or rewrite) the implementation of the C</(?{...})/> closures.
1104
1105=head2 A re-entrant regexp engine
1106
1107This will allow the use of a regex from inside (?{ }), (??{ }) and
1108(?(?{ })|) constructs.
6bda09f9 1109
6bda09f9
YO
1110=head2 Add class set operations to regexp engine
1111
1112Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
1113
1114demerphq has this on his todo list, but right at the bottom.