This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
PADTMP hash pessimisation was fixed with commit 38bb37b9aa18dbe6.
[perl5.git] / Porting / todo.pod
CommitLineData
7711098a
GS
1=head1 NAME
2
c3143508 3todo - Perl TO-DO list
7711098a
GS
4
5=head1 DESCRIPTION
e50bb9a1 6
049aabcb 7This is a list of wishes for Perl. The most up to date version of this file
c3143508 8is at L<http://perl5.git.perl.org/perl.git/blob_plain/HEAD:/Porting/todo.pod>
049aabcb
NC
9
10The tasks we think are smaller or easier are listed first. Anyone is welcome
11to work on any of these, but it's a good idea to first contact
12I<perl5-porters@perl.org> to avoid duplication of effort, and to learn from
13any previous attempts. By all means contact a pumpking privately first if you
14prefer.
e50bb9a1 15
0bdfc961
NC
16Whilst patches to make the list shorter are most welcome, ideas to add to
17the list are also encouraged. Check the perl5-porters archives for past
b4af8972
RB
18ideas, and any discussion about them. One set of archives may be found at
19L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/>
938c8732 20
617eabfa
NC
21What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
22not, but if your patch is incorporated, then we'll add your name to the
23F<AUTHORS> file, which ships in the official distribution. How many other
24programming languages offer you 1 line of immortality?
938c8732 25
0bdfc961 26=head1 Tasks that only need Perl knowledge
e50bb9a1 27
de2b17d8
NC
28=head2 Migrate t/ from custom TAP generation
29
30Many tests below F<t/> still generate TAP by "hand", rather than using library
96090e4f 31functions. As explained in L<perlhack/TESTING>, tests in F<t/> are
de2b17d8
NC
32written in a particular way to test that more complex constructions actually
33work before using them routinely. Hence they don't use C<Test::More>, but
34instead there is an intentionally simpler library, F<t/test.pl>. However,
35quite a few tests in F<t/> have not been refactored to use it. Refactoring
36any of these tests, one at a time, is a useful thing TODO.
37
0d8e5a42
RGS
38The subdirectories F<base>, F<cmd> and F<comp>, that contain the most
39basic tests, should be excluded from this task.
40
0be987a2
NC
41=head2 Automate perldelta generation
42
43The perldelta file accompanying each release summaries the major changes.
44It's mostly manually generated currently, but some of that could be
45automated with a bit of perl, specifically the generation of
46
47=over
48
49=item Modules and Pragmata
50
51=item New Documentation
52
53=item New Tests
54
55=back
56
57See F<Porting/how_to_write_a_perldelta.pod> for details.
58
5a176cbc
NC
59=head2 Remove duplication of test setup.
60
61Schwern notes, that there's duplication of code - lots and lots of tests have
62some variation on the big block of C<$Is_Foo> checks. We can safely put this
63into a file, change it to build an C<%Is> hash and require it. Maybe just put
64it into F<test.pl>. Throw in the handy tainting subroutines.
65
87a942b1 66=head2 POD -E<gt> HTML conversion in the core still sucks
e50bb9a1 67
938c8732 68Which is crazy given just how simple POD purports to be, and how simple HTML
adebf063
NC
69can be. It's not actually I<as> simple as it sounds, particularly with the
70flexibility POD allows for C<=item>, but it would be good to improve the
71visual appeal of the HTML generated, and to avoid it having any validation
72errors. See also L</make HTML install work>, as the layout of installation tree
73is needed to improve the cross-linking.
938c8732 74
dc0fb092
SP
75The addition of C<Pod::Simple> and its related modules may make this task
76easier to complete.
77
0bdfc961 78=head2 Make Schwern poorer
e50bb9a1 79
613bd4f7 80We should have tests for everything. When all the core's modules are tested,
0bdfc961
NC
81Schwern has promised to donate to $500 to TPF. We may need volunteers to
82hold him upside down and shake vigorously in order to actually extract the
83cash.
3958b146 84
0bdfc961 85=head2 Improve the coverage of the core tests
e50bb9a1 86
e1020413 87Use Devel::Cover to ascertain the core modules' test coverage, then add
02f21748 88tests that are currently missing.
30222c0f 89
0bdfc961 90=head2 test B
e50bb9a1 91
0bdfc961 92A full test suite for the B module would be nice.
e50bb9a1 93
0bdfc961 94=head2 A decent benchmark
e50bb9a1 95
617eabfa 96C<perlbench> seems impervious to any recent changes made to the perl core. It
0bdfc961
NC
97would be useful to have a reasonable general benchmarking suite that roughly
98represented what current perl programs do, and measurably reported whether
99tweaks to the core improve, degrade or don't really affect performance, to
100guide people attempting to optimise the guts of perl. Gisle would welcome
101new tests for perlbench.
6168cf99 102
0bdfc961 103=head2 fix tainting bugs
6168cf99 104
0bdfc961
NC
105Fix the bugs revealed by running the test suite with the C<-t> switch (via
106C<make test.taintwarn>).
e50bb9a1 107
0bdfc961 108=head2 Dual life everything
e50bb9a1 109
0bdfc961
NC
110As part of the "dists" plan, anything that doesn't belong in the smallest perl
111distribution needs to be dual lifed. Anything else can be too. Figure out what
112changes would be needed to package that module and its tests up for CPAN, and
113do so. Test it with older perl releases, and fix the problems you find.
e50bb9a1 114
a393eb28
RGS
115To make a minimal perl distribution, it's useful to look at
116F<t/lib/commonsense.t>.
117
0bdfc961 118=head2 POSIX memory footprint
e50bb9a1 119
0bdfc961
NC
120Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
121various times worked to cut it down. There is probably still fat to cut out -
122for example POSIX passes Exporter some very memory hungry data structures.
e50bb9a1 123
eed36644
NC
124=head2 embed.pl/makedef.pl
125
126There is a script F<embed.pl> that generates several header files to prefix
127all of Perl's symbols in a consistent way, to provide some semblance of
128namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
907b3e23 129in F<interpvar.h>. Quite a few of the functions and variables
eed36644
NC
130are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
131doesn't understand the C macros, so the rules about which symbols are present
132when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
133It would be good to teach C<embed.pl> to understand the conditional
134compilation, and hence remove the duplication, and the mistakes it has caused.
e50bb9a1 135
801de10e
NC
136=head2 use strict; and AutoLoad
137
138Currently if you write
139
140 package Whack;
141 use AutoLoader 'AUTOLOAD';
142 use strict;
143 1;
144 __END__
145 sub bloop {
146 print join (' ', No, strict, here), "!\n";
147 }
148
149then C<use strict;> isn't in force within the autoloaded subroutines. It would
150be more consistent (and less surprising) to arrange for all lexical pragmas
151in force at the __END__ block to be in force within each autoloaded subroutine.
152
773b3597
RGS
153There's a similar problem with SelfLoader.
154
91d0cbf6
NC
155=head2 profile installman
156
157The F<installman> script is slow. All it is doing text processing, which we're
158told is something Perl is good at. So it would be nice to know what it is doing
159that is taking so much CPU, and where possible address it.
160
c69ca1d4 161=head2 enable lexical enabling/disabling of individual warnings
a9ed9b74
JV
162
163Currently, warnings can only be enabled or disabled by category. There
164are times when it would be useful to quash a single warning, not a
165whole category.
91d0cbf6 166
85234543
KW
167=head2 document diagnostics
168
169Many diagnostic messages are not currently documented. The list is at the end
170of t/porting/diag.t.
171
0bdfc961 172=head1 Tasks that need a little sysadmin-type knowledge
e50bb9a1 173
0bdfc961
NC
174Or if you prefer, tasks that you would learn from, and broaden your skills
175base...
e50bb9a1 176
cd793d32 177=head2 make HTML install work
e50bb9a1 178
adebf063
NC
179There is an C<installhtml> target in the Makefile. It's marked as
180"experimental". It would be good to get this tested, make it work reliably, and
181remove the "experimental" tag. This would include
182
183=over 4
184
185=item 1
186
187Checking that cross linking between various parts of the documentation works.
188In particular that links work between the modules (files with POD in F<lib/>)
189and the core documentation (files in F<pod/>)
190
191=item 2
192
617eabfa
NC
193Work out how to split C<perlfunc> into chunks, preferably one per function
194group, preferably with general case code that could be used elsewhere.
195Challenges here are correctly identifying the groups of functions that go
196together, and making the right named external cross-links point to the right
197page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
198C<endservent>, two or more C<=items> giving the different parameter lists, such
199as
adebf063
NC
200
201 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
adebf063 202 =item substr EXPR,OFFSET,LENGTH
adebf063
NC
203 =item substr EXPR,OFFSET
204
205and different parameter lists having different meanings. (eg C<select>)
206
207=back
3a89a73c 208
0bdfc961
NC
209=head2 compressed man pages
210
211Be able to install them. This would probably need a configure test to see how
212the system does compressed man pages (same directory/different directory?
213same filename/different filename), as well as tweaking the F<installman> script
214to compress as necessary.
215
30222c0f
NC
216=head2 Add a code coverage target to the Makefile
217
218Make it easy for anyone to run Devel::Cover on the core's tests. The steps
219to do this manually are roughly
220
221=over 4
222
223=item *
224
225do a normal C<Configure>, but include Devel::Cover as a module to install
f11a3063 226(see L<INSTALL> for how to do this)
30222c0f
NC
227
228=item *
229
230 make perl
231
232=item *
233
234 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
235
236=item *
237
238Process the resulting Devel::Cover database
239
240=back
241
242This just give you the coverage of the F<.pm>s. To also get the C level
243coverage you need to
244
245=over 4
246
247=item *
248
249Additionally tell C<Configure> to use the appropriate C compiler flags for
250C<gcov>
251
252=item *
253
254 make perl.gcov
255
256(instead of C<make perl>)
257
258=item *
259
260After running the tests run C<gcov> to generate all the F<.gcov> files.
261(Including down in the subdirectories of F<ext/>
262
263=item *
264
265(From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
266to get their stats into the cover_db directory.
267
268=item *
269
270Then process the Devel::Cover database
271
272=back
273
274It would be good to add a single switch to C<Configure> to specify that you
275wanted to perform perl level coverage, and another to specify C level
276coverage, and have C<Configure> and the F<Makefile> do all the right things
277automatically.
278
02f21748 279=head2 Make Config.pm cope with differences between built and installed perl
0bdfc961
NC
280
281Quite often vendors ship a perl binary compiled with their (pay-for)
282compilers. People install a free compiler, such as gcc. To work out how to
283build extensions, Perl interrogates C<%Config>, so in this situation
284C<%Config> describes compilers that aren't there, and extension building
285fails. This forces people into choosing between re-compiling perl themselves
286using the compiler they have, or only using modules that the vendor ships.
287
288It would be good to find a way teach C<Config.pm> about the installation setup,
289possibly involving probing at install time or later, so that the C<%Config> in
290a binary distribution better describes the installed machine, when the
291installed machine differs from the build machine in some significant way.
292
728f4ecd
NC
293=head2 linker specification files
294
295Some platforms mandate that you provide a list of a shared library's external
296symbols to the linker, so the core already has the infrastructure in place to
297do this for generating shared perl libraries. My understanding is that the
298GNU toolchain can accept an optional linker specification file, and restrict
299visibility just to symbols declared in that file. It would be good to extend
300F<makedef.pl> to support this format, and to provide a means within
301C<Configure> to enable it. This would allow Unix users to test that the
302export list is correct, and to build a perl that does not pollute the global
32d539f5
RU
303namespace with private symbols, and will fail in the same way as msvc or mingw
304builds or when using PERL_DL_NONLAZY=1.
728f4ecd 305
a229ae3b
RGS
306=head2 Cross-compile support
307
308Currently C<Configure> understands C<-Dusecrosscompile> option. This option
309arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
310assumed then to be copied to TARGET machine and used as a replacement of full
311C<perl> executable.
312
d1307786 313This could be done little differently. Namely C<miniperl> should be built for
a229ae3b 314HOST and then full C<perl> with extensions should be compiled for TARGET.
d1307786 315This, however, might require extra trickery for %Config: we have one config
87a942b1
JH
316first for HOST and then another for TARGET. Tools like MakeMaker will be
317mightily confused. Having around two different types of executables and
318libraries (HOST and TARGET) makes life interesting for Makefiles and
319shell (and Perl) scripts. There is $Config{run}, normally empty, which
320can be used as an execution wrapper. Also note that in some
321cross-compilation/execution environments the HOST and the TARGET do
322not see the same filesystem(s), the $Config{run} may need to do some
323file/directory copying back and forth.
0bdfc961 324
98fca0e8
NC
325=head2 Split "linker" from "compiler"
326
327Right now, Configure probes for two commands, and sets two variables:
328
329=over 4
330
b91dd380 331=item * C<cc> (in F<cc.U>)
98fca0e8
NC
332
333This variable holds the name of a command to execute a C compiler which
334can resolve multiple global references that happen to have the same
335name. Usual values are F<cc> and F<gcc>.
336Fervent ANSI compilers may be called F<c89>. AIX has F<xlc>.
337
b91dd380 338=item * C<ld> (in F<dlsrc.U>)
98fca0e8
NC
339
340This variable indicates the program to be used to link
341libraries for dynamic loading. On some systems, it is F<ld>.
342On ELF systems, it should be C<$cc>. Mostly, we'll try to respect
343the hint file setting.
344
345=back
346
8d159ec1
NC
347There is an implicit historical assumption from around Perl5.000alpha
348something, that C<$cc> is also the correct command for linking object files
349together to make an executable. This may be true on Unix, but it's not true
350on other platforms, and there are a maze of work arounds in other places (such
351as F<Makefile.SH>) to cope with this.
98fca0e8
NC
352
353Ideally, we should create a new variable to hold the name of the executable
354linker program, probe for it in F<Configure>, and centralise all the special
355case logic there or in hints files.
356
357A small bikeshed issue remains - what to call it, given that C<$ld> is already
8d159ec1
NC
358taken (arguably for the wrong thing now, but on SunOS 4.1 it is the command
359for creating dynamically-loadable modules) and C<$link> could be confused with
360the Unix command line executable of the same name, which does something
361completely different. Andy Dougherty makes the counter argument "In parrot, I
362tried to call the command used to link object files and libraries into an
363executable F<link>, since that's what my vaguely-remembered DOS and VMS
364experience suggested. I don't think any real confusion has ensued, so it's
365probably a reasonable name for perl5 to use."
98fca0e8
NC
366
367"Alas, I've always worried that introducing it would make things worse,
368since now the module building utilities would have to look for
369C<$Config{link}> and institute a fall-back plan if it weren't found."
8d159ec1
NC
370Although I can see that as confusing, given that C<$Config{d_link}> is true
371when (hard) links are available.
98fca0e8 372
75585ce3
SP
373=head2 Configure Windows using PowerShell
374
375Currently, Windows uses hard-coded config files based to build the
376config.h for compiling Perl. Makefiles are also hard-coded and need to be
377hand edited prior to building Perl. While this makes it easy to create a perl.exe
378that works across multiple Windows versions, being able to accurately
379configure a perl.exe for a specific Windows versions and VS C++ would be
380a nice enhancement. With PowerShell available on Windows XP and up, this
381may now be possible. Step 1 might be to investigate whether this is possible
382and use this to clean up our current makefile situation. Step 2 would be to
383see if there would be a way to use our existing metaconfig units to configure a
384Windows Perl or whether we go in a separate direction and make it so. Of
385course, we all know what step 3 is.
386
0bdfc961
NC
387=head1 Tasks that need a little C knowledge
388
389These tasks would need a little C knowledge, but don't need any specific
390background or experience with XS, or how the Perl interpreter works
391
3d826b29
NC
392=head2 Weed out needless PERL_UNUSED_ARG
393
394The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about
395unused arguments. Often the arguments can't be removed, as there is an
396external constraint that determines the prototype of the function, so this
397approach is valid. However, there are some cases where C<PERL_UNUSED_ARG>
398could be removed. Specifically
399
400=over 4
401
402=item *
403
404The prototypes of (nearly all) static functions can be changed
405
406=item *
407
408Unused arguments generated by short cut macros are wasteful - the short cut
409macro used can be changed.
410
411=back
412
fbf638cb
RGS
413=head2 Modernize the order of directories in @INC
414
415The way @INC is laid out by default, one cannot upgrade core (dual-life)
416modules without overwriting files. This causes problems for binary
3d14fd97
AD
417package builders. One possible proposal is laid out in this
418message:
b4af8972 419L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>
fbf638cb 420
bcbaa2d5
RGS
421=head2 -Duse32bit*
422
423Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
424On these systems, it might be the default compilation mode, and there
425is currently no guarantee that passing no use64bitall option to the
426Configure process will build a 32bit perl. Implementing -Duse32bit*
fd2dadea 427options would be nice for perl 5.14.
bcbaa2d5 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,
c23989d1 465identical in both F<win32/wince.c> and F<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:
b4af8972 507L<http://bugs.activestate.com/show_bug.cgi?id=74552>
038ae9a4
SH
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
3880c8ec
NC
547=head2 Shared arenas
548
549Several SV body structs are now the same size, notably PVMG and PVGV, PVAV and
550PVHV, and PVCV and PVFM. It should be possible to allocate and return same
551sized bodies from the same actual arena, rather than maintaining one arena for
552each. This could save 4-6K per thread, of memory no longer tied up in the
553not-yet-allocated part of an arena.
554
8964cfe0 555
6d71adcd
NC
556=head1 Tasks that need a knowledge of XS
557
558These tasks would need C knowledge, and roughly the level of knowledge of
559the perl API that comes from writing modules that use XS to interface to
560C.
561
e851c105
DG
562=head2 Write an XS cookbook
563
564Create pod/perlxscookbook.pod with short, task-focused 'recipes' in XS that
565demonstrate common tasks and good practices. (Some of these might be
566extracted from perlguts.) The target audience should be XS novices, who need
567more examples than perlguts but something less overwhelming than perlapi.
568Recipes should provide "one pretty good way to do it" instead of TIMTOWTDI.
569
5b7d14ff
DG
570Rather than focusing on interfacing Perl to C libraries, such a cookbook
571should probably focus on how to optimize Perl routines by re-writing them
572in XS. This will likely be more motivating to those who mostly work in
573Perl but are looking to take the next step into XS.
574
575Deconstructing and explaining some simpler XS modules could be one way to
576bootstrap a cookbook. (List::Util? Class::XSAccessor? Tree::Ternary_XS?)
577Another option could be deconstructing the implementation of some simpler
578functions in op.c.
579
05fb4e20
NC
580=head2 Allow XSUBs to inline themselves as OPs
581
582For a simple XSUB, often the subroutine dispatch takes more time than the
583XSUB itself. The tokeniser already has the ability to inline constant
584subroutines - it would be good to provide a way to inline other subroutines.
585
586Specifically, simplest approach looks to be to allow an XSUB to provide an
587alternative implementation of itself as a custom OP. A new flag bit in
588C<CvFLAGS()> would signal to the peephole optimiser to take an optree
589such as this:
590
591 b <@> leave[1 ref] vKP/REFC ->(end)
592 1 <0> enter ->2
593 2 <;> nextstate(main 1 -e:1) v:{ ->3
594 a <2> sassign vKS/2 ->b
595 8 <1> entersub[t2] sKS/TARG,1 ->9
596 - <1> ex-list sK ->8
597 3 <0> pushmark s ->4
598 4 <$> const(IV 1) sM ->5
599 6 <1> rv2av[t1] lKM/1 ->7
600 5 <$> gv(*a) s ->6
601 - <1> ex-rv2cv sK ->-
602 7 <$> gv(*x) s/EARLYCV ->8
603 - <1> ex-rv2sv sKRM*/1 ->a
604 9 <$> gvsv(*b) s ->a
605
606perform the symbol table lookup of C<rv2cv> and C<gv(*x)>, locate the
607pointer to the custom OP that provides the direct implementation, and re-
608write the optree something like:
609
610 b <@> leave[1 ref] vKP/REFC ->(end)
611 1 <0> enter ->2
612 2 <;> nextstate(main 1 -e:1) v:{ ->3
613 a <2> sassign vKS/2 ->b
614 7 <1> custom_x -> 8
615 - <1> ex-list sK ->7
616 3 <0> pushmark s ->4
617 4 <$> const(IV 1) sM ->5
618 6 <1> rv2av[t1] lKM/1 ->7
619 5 <$> gv(*a) s ->6
620 - <1> ex-rv2cv sK ->-
621 - <$> ex-gv(*x) s/EARLYCV ->7
622 - <1> ex-rv2sv sKRM*/1 ->a
623 8 <$> gvsv(*b) s ->a
624
625I<i.e.> the C<gv(*)> OP has been nulled and spliced out of the execution
626path, and the C<entersub> OP has been replaced by the custom op.
627
628This approach should provide a measurable speed up to simple XSUBs inside
629tight loops. Initially one would have to write the OP alternative
630implementation by hand, but it's likely that this should be reasonably
631straightforward for the type of XSUB that would benefit the most. Longer
632term, once the run-time implementation is proven, it should be possible to
633progressively update ExtUtils::ParseXS to generate OP implementations for
634some XSUBs.
635
318bf708
NC
636=head2 Remove the use of SVs as temporaries in dump.c
637
638F<dump.c> contains debugging routines to dump out the contains of perl data
639structures, such as C<SV>s, C<AV>s and C<HV>s. Currently, the dumping code
640B<uses> C<SV>s for its temporary buffers, which was a logical initial
641implementation choice, as they provide ready made memory handling.
642
643However, they also lead to a lot of confusion when it happens that what you're
644trying to debug is seen by the code in F<dump.c>, correctly or incorrectly, as
645a temporary scalar it can use for a temporary buffer. It's also not possible
646to dump scalars before the interpreter is properly set up, such as during
647ithreads cloning. It would be good to progressively replace the use of scalars
648as string accumulation buffers with something much simpler, directly allocated
649by C<malloc>. The F<dump.c> code is (or should be) only producing 7 bit
650US-ASCII, so output character sets are not an issue.
651
652Producing and proving an internal simple buffer allocation would make it easier
653to re-write the internals of the PerlIO subsystem to avoid using C<SV>s for
654B<its> buffers, use of which can cause problems similar to those of F<dump.c>,
655at similar times.
656
5d96f598
NC
657=head2 safely supporting POSIX SA_SIGINFO
658
659Some years ago Jarkko supplied patches to provide support for the POSIX
660SA_SIGINFO feature in Perl, passing the extra data to the Perl signal handler.
661
662Unfortunately, it only works with "unsafe" signals, because under safe
663signals, by the time Perl gets to run the signal handler, the extra
664information has been lost. Moreover, it's not easy to store it somewhere,
665as you can't call mutexs, or do anything else fancy, from inside a signal
666handler.
667
668So it strikes me that we could provide safe SA_SIGINFO support
669
670=over 4
671
672=item 1
673
674Provide global variables for two file descriptors
675
676=item 2
677
678When the first request is made via C<sigaction> for C<SA_SIGINFO>, create a
679pipe, store the reader in one, the writer in the other
680
681=item 3
682
683In the "safe" signal handler (C<Perl_csighandler()>/C<S_raise_signal()>), if
684the C<siginfo_t> pointer non-C<NULL>, and the writer file handle is open,
685
686=over 8
687
688=item 1
689
690serialise signal number, C<struct siginfo_t> (or at least the parts we care
691about) into a small auto char buff
692
693=item 2
694
695C<write()> that (non-blocking) to the writer fd
696
697=over 12
698
699=item 1
700
701if it writes 100%, flag the signal in a counter of "signals on the pipe" akin
702to the current per-signal-number counts
703
704=item 2
705
706if it writes 0%, assume the pipe is full. Flag the data as lost?
707
708=item 3
709
710if it writes partially, croak a panic, as your OS is broken.
711
712=back
713
714=back
715
716=item 4
717
718in the regular C<PERL_ASYNC_CHECK()> processing, if there are "signals on
719the pipe", read the data out, deserialise, build the Perl structures on
720the stack (code in C<Perl_sighandler()>, the "unsafe" handler), and call as
721usual.
722
723=back
724
725I think that this gets us decent C<SA_SIGINFO> support, without the current risk
726of running Perl code inside the signal handler context. (With all the dangers
727of things like C<malloc> corruption that that currently offers us)
728
729For more information see the thread starting with this message:
b4af8972 730L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-03/msg00305.html>
5d96f598 731
6d71adcd
NC
732=head2 autovivification
733
734Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
735
736This task is incremental - even a little bit of work on it will help.
737
738=head2 Unicode in Filenames
739
740chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
741opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
742system, truncate, unlink, utime, -X. All these could potentially accept
743Unicode filenames either as input or output (and in the case of system
744and qx Unicode in general, as input or output to/from the shell).
745Whether a filesystem - an operating system pair understands Unicode in
746filenames varies.
747
748Known combinations that have some level of understanding include
749Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
750OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
751create Unicode filenames, what forms of Unicode are accepted and used
752(UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
753and so on, varies. Finding the right level of interfacing to Perl
754requires some thought. Remember that an OS does not implicate a
755filesystem.
756
757(The Windows -C command flag "wide API support" has been at least
758temporarily retired in 5.8.1, and the -C has been repurposed, see
759L<perlrun>.)
760
87a942b1
JH
761Most probably the right way to do this would be this:
762L</"Virtualize operating system access">.
763
6d71adcd
NC
764=head2 Unicode in %ENV
765
766Currently the %ENV entries are always byte strings.
87a942b1 767See L</"Virtualize operating system access">.
6d71adcd 768
1f2e7916
JD
769=head2 Unicode and glob()
770
771Currently glob patterns and filenames returned from File::Glob::glob()
87a942b1 772are always byte strings. See L</"Virtualize operating system access">.
1f2e7916 773
6d71adcd
NC
774=head2 use less 'memory'
775
776Investigate trade offs to switch out perl's choices on memory usage.
777Particularly perl should be able to give memory back.
778
779This task is incremental - even a little bit of work on it will help.
780
781=head2 Re-implement C<:unique> in a way that is actually thread-safe
782
783The old implementation made bad assumptions on several levels. A good 90%
784solution might be just to make C<:unique> work to share the string buffer
785of SvPVs. That way large constant strings can be shared between ithreads,
786such as the configuration information in F<Config>.
787
788=head2 Make tainting consistent
789
790Tainting would be easier to use if it didn't take documented shortcuts and
791allow taint to "leak" everywhere within an expression.
792
793=head2 readpipe(LIST)
794
795system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
796running a shell. readpipe() (the function behind qx//) could be similarly
797extended.
798
6d71adcd
NC
799=head2 Audit the code for destruction ordering assumptions
800
801Change 25773 notes
802
803 /* Need to check SvMAGICAL, as during global destruction it may be that
804 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
805 is now part of the linked list of SV heads, rather than pointing to
806 the original body. */
807 /* FIXME - audit the code for other bugs like this one. */
808
809adding the C<SvMAGICAL> check to
810
811 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
812 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
813
814Go through the core and look for similar assumptions that SVs have particular
815types, as all bets are off during global destruction.
816
749904bf
JH
817=head2 Extend PerlIO and PerlIO::Scalar
818
819PerlIO::Scalar doesn't know how to truncate(). Implementing this
820would require extending the PerlIO vtable.
821
822Similarly the PerlIO vtable doesn't know about formats (write()), or
823about stat(), or chmod()/chown(), utime(), or flock().
824
825(For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
826would mean.)
827
828PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
829opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
830readlink().
831
94da6c29
JH
832See also L</"Virtualize operating system access">.
833
3236f110
NC
834=head2 -C on the #! line
835
836It should be possible to make -C work correctly if found on the #! line,
837given that all perl command line options are strict ASCII, and -C changes
838only the interpretation of non-ASCII characters, and not for the script file
839handle. To make it work needs some investigation of the ordering of function
840calls during startup, and (by implication) a bit of tweaking of that order.
841
d6c1e11f
JH
842=head2 Organize error messages
843
844Perl's diagnostics (error messages, see L<perldiag>) could use
a8d0aeb9 845reorganizing and formalizing so that each error message has its
d6c1e11f
JH
846stable-for-all-eternity unique id, categorized by severity, type, and
847subsystem. (The error messages would be listed in a datafile outside
c4bd451b
CB
848of the Perl source code, and the source code would only refer to the
849messages by the id.) This clean-up and regularizing should apply
d6c1e11f
JH
850for all croak() messages.
851
852This would enable all sorts of things: easier translation/localization
853of the messages (though please do keep in mind the caveats of
854L<Locale::Maketext> about too straightforward approaches to
855translation), filtering by severity, and instead of grepping for a
856particular error message one could look for a stable error id. (Of
857course, changing the error messages by default would break all the
858existing software depending on some particular error message...)
859
860This kind of functionality is known as I<message catalogs>. Look for
861inspiration for example in the catgets() system, possibly even use it
862if available-- but B<only> if available, all platforms will B<not>
de96509d 863have catgets().
d6c1e11f
JH
864
865For the really pure at heart, consider extending this item to cover
866also the warning messages (see L<perllexwarn>, C<warnings.pl>).
3236f110 867
0bdfc961 868=head1 Tasks that need a knowledge of the interpreter
3298bd4d 869
0bdfc961
NC
870These tasks would need C knowledge, and knowledge of how the interpreter works,
871or a willingness to learn.
3298bd4d 872
10517af5
JD
873=head2 forbid labels with keyword names
874
875Currently C<goto keyword> "computes" the label value:
876
877 $ perl -e 'goto print'
878 Can't find label 1 at -e line 1.
879
343c8006
JD
880It is controversial if the right way to avoid the confusion is to forbid
881labels with keyword names, or if it would be better to always treat
882bareword expressions after a "goto" as a label and never as a keyword.
10517af5 883
de6375e3
RGS
884=head2 truncate() prototype
885
886The prototype of truncate() is currently C<$$>. It should probably
887be C<*$> instead. (This is changed in F<opcode.pl>)
888
2d0587d8
RGS
889=head2 decapsulation of smart match argument
890
891Currently C<$foo ~~ $object> will die with the message "Smart matching a
892non-overloaded object breaks encapsulation". It would be nice to allow
c69ca1d4 893to bypass this by using explicitly the syntax C<$foo ~~ %$object> or
2d0587d8
RGS
894C<$foo ~~ @$object>.
895
565590b5
NC
896=head2 error reporting of [$a ; $b]
897
898Using C<;> inside brackets is a syntax error, and we don't propose to change
899that by giving it any meaning. However, it's not reported very helpfully:
900
901 $ perl -e '$a = [$b; $c];'
902 syntax error at -e line 1, near "$b;"
903 syntax error at -e line 1, near "$c]"
904 Execution of -e aborted due to compilation errors.
905
906It should be possible to hook into the tokeniser or the lexer, so that when a
907C<;> is parsed where it is not legal as a statement terminator (ie inside
908C<{}> used as a hashref, C<[]> or C<()>) it issues an error something like
909I<';' isn't legal inside an expression - if you need multiple statements use a
910do {...} block>. See the thread starting at
b4af8972 911L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-09/msg00573.html>
565590b5 912
718140ec
NC
913=head2 lexicals used only once
914
915This warns:
916
917 $ perl -we '$pie = 42'
918 Name "main::pie" used only once: possible typo at -e line 1.
919
920This does not:
921
922 $ perl -we 'my $pie = 42'
923
924Logically all lexicals used only once should warn, if the user asks for
d6f4ea2e
SP
925warnings. An unworked RT ticket (#5087) has been open for almost seven
926years for this discrepancy.
718140ec 927
a3d15f9a
RGS
928=head2 UTF-8 revamp
929
85c006b6
KW
930The handling of Unicode is unclean in many places. In the regex engine
931there are especially many problems. The swash data structure could be
932replaced my something better. Inversion lists and maps are likely
933candidates. The whole Unicode database could be placed in-core for a
934huge speed-up. Only minimal work was done on the optimizer when utf8
935was added, with the result that the synthetic start class often will
936fail to narrow down the possible choices when given non-Latin1 input.
a3d15f9a
RGS
937
938=head2 Properly Unicode safe tokeniser and pads.
939
940The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
941variable names are stored in stashes as raw bytes, without the utf-8 flag
942set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
943tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
944source filters. All this could be fixed.
945
636e63cb
NC
946=head2 state variable initialization in list context
947
948Currently this is illegal:
949
950 state ($a, $b) = foo();
951
a2874905 952In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different
a8d0aeb9 953semantics, which is tricky to implement in Perl 5 as currently they produce
a2874905 954the same opcode trees. The Perl 6 design is firm, so it would be good to
a8d0aeb9 955implement the necessary code in Perl 5. There are comments in
a2874905
NC
956C<Perl_newASSIGNOP()> that show the code paths taken by various assignment
957constructions involving state variables.
636e63cb 958
4fedb12c
RGS
959=head2 Implement $value ~~ 0 .. $range
960
961It would be nice to extend the syntax of the C<~~> operator to also
962understand numeric (and maybe alphanumeric) ranges.
a393eb28
RGS
963
964=head2 A does() built-in
965
966Like ref(), only useful. It would call the C<DOES> method on objects; it
967would also tell whether something can be dereferenced as an
968array/hash/etc., or used as a regexp, etc.
969L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
970
971=head2 Tied filehandles and write() don't mix
972
973There is no method on tied filehandles to allow them to be called back by
974formats.
4fedb12c 975
53967bb9
RGS
976=head2 Propagate compilation hints to the debugger
977
978Currently a debugger started with -dE on the command-line doesn't see the
979features enabled by -E. More generally hints (C<$^H> and C<%^H>) aren't
980propagated to the debugger. Probably it would be a good thing to propagate
981hints from the innermost non-C<DB::> scope: this would make code eval'ed
982in the debugger see the features (and strictures, etc.) currently in
983scope.
984
d10fc472 985=head2 Attach/detach debugger from running program
1626a787 986
cd793d32
NC
987The old perltodo notes "With C<gdb>, you can attach the debugger to a running
988program if you pass the process ID. It would be good to do this with the Perl
0bdfc961
NC
989debugger on a running Perl program, although I'm not sure how it would be
990done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
1626a787 991
0bdfc961
NC
992=head2 LVALUE functions for lists
993
994The old perltodo notes that lvalue functions don't work for list or hash
995slices. This would be good to fix.
996
0bdfc961
NC
997=head2 regexp optimiser optional
998
999The regexp optimiser is not optional. It should configurable to be, to allow
1000its performance to be measured, and its bugs to be easily demonstrated.
1001
ef36c6a7
RGS
1002=head2 C</w> regex modifier
1003
1004That flag would enable to match whole words, and also to interpolate
1005arrays as alternations. With it, C</P/w> would be roughly equivalent to:
1006
1007 do { local $"='|'; /\b(?:P)\b/ }
1008
b4af8972
RB
1009See
1010L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
ef36c6a7
RGS
1011for the discussion.
1012
0bdfc961
NC
1013=head2 optional optimizer
1014
1015Make the peephole optimizer optional. Currently it performs two tasks as
1016it walks the optree - genuine peephole optimisations, and necessary fixups of
1017ops. It would be good to find an efficient way to switch out the
1018optimisations whilst keeping the fixups.
1019
1020=head2 You WANT *how* many
1021
1022Currently contexts are void, scalar and list. split has a special mechanism in
1023place to pass in the number of return values wanted. It would be useful to
1024have a general mechanism for this, backwards compatible and little speed hit.
1025This would allow proposals such as short circuiting sort to be implemented
1026as a module on CPAN.
1027
1028=head2 lexical aliases
1029
1030Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
1031
1032=head2 entersub XS vs Perl
1033
1034At the moment pp_entersub is huge, and has code to deal with entering both
1035perl and XS subroutines. Subroutine implementations rarely change between
1036perl and XS at run time, so investigate using 2 ops to enter subs (one for
1037XS, one for perl) and swap between if a sub is redefined.
2810d901 1038
de535794 1039=head2 Self-ties
2810d901 1040
de535794 1041Self-ties are currently illegal because they caused too many segfaults. Maybe
a8d0aeb9 1042the causes of these could be tracked down and self-ties on all types
de535794 1043reinstated.
0bdfc961
NC
1044
1045=head2 Optimize away @_
1046
1047The old perltodo notes "Look at the "reification" code in C<av.c>".
1048
87a942b1
JH
1049=head2 Virtualize operating system access
1050
1051Implement a set of "vtables" that virtualizes operating system access
1052(open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
1053least these interfaces should take SVs as "name" arguments instead of
1054bare char pointers; probably the most flexible and extensible way
e1a3d5d1
JH
1055would be for the Perl-facing interfaces to accept HVs. The system
1056needs to be per-operating-system and per-file-system
1057hookable/filterable, preferably both from XS and Perl level
87a942b1
JH
1058(L<perlport/"Files and Filesystems"> is good reading at this point,
1059in fact, all of L<perlport> is.)
1060
e1a3d5d1
JH
1061This has actually already been implemented (but only for Win32),
1062take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
1063variants go through a set of "vtables" for operating system access,
e1020413 1064non-Win32 systems currently go straight for the POSIX/Unix-style
e1a3d5d1
JH
1065system/library call. Similar system as for Win32 should be
1066implemented for all platforms. The existing Win32 implementation
1067probably does not need to survive alongside this proposed new
1068implementation, the approaches could be merged.
87a942b1
JH
1069
1070What would this give us? One often-asked-for feature this would
94da6c29
JH
1071enable is using Unicode for filenames, and other "names" like %ENV,
1072usernames, hostnames, and so forth.
1073(See L<perlunicode/"When Unicode Does Not Happen">.)
1074
1075But this kind of virtualization would also allow for things like
1076virtual filesystems, virtual networks, and "sandboxes" (though as long
1077as dynamic loading of random object code is allowed, not very safe
1078sandboxes since external code of course know not of Perl's vtables).
1079An example of a smaller "sandbox" is that this feature can be used to
1080implement per-thread working directories: Win32 already does this.
1081
1082See also L</"Extend PerlIO and PerlIO::Scalar">.
87a942b1 1083
057163d7
NC
1084=head2 Store the current pad in the OP slab allocator
1085
1086=for clarification
1087I hope that I got that "current pad" part correct
1088
1089Currently we leak ops in various cases of parse failure. I suggested that we
1090could solve this by always using the op slab allocator, and walking it to
1091free ops. Dave comments that as some ops are already freed during optree
1092creation one would have to mark which ops are freed, and not double free them
1093when walking the slab. He notes that one problem with this is that for some ops
1094you have to know which pad was current at the time of allocation, which does
1095change. I suggested storing a pointer to the current pad in the memory allocated
1096for the slab, and swapping to a new slab each time the pad changes. Dave thinks
1097that this would work.
1098
52960e22
JC
1099=head2 repack the optree
1100
1101Repacking the optree after execution order is determined could allow
057163d7
NC
1102removal of NULL ops, and optimal ordering of OPs with respect to cache-line
1103filling. The slab allocator could be reused for this purpose. I think that
1104the best way to do this is to make it an optional step just before the
1105completed optree is attached to anything else, and to use the slab allocator
1106unchanged, so that freeing ops is identical whether or not this step runs.
1107Note that the slab allocator allocates ops downwards in memory, so one would
1108have to actually "allocate" the ops in reverse-execution order to get them
1109contiguous in memory in execution order.
1110
b4af8972
RB
1111See
1112L<http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html>
057163d7
NC
1113
1114Note that running this copy, and then freeing all the old location ops would
1115cause their slabs to be freed, which would eliminate possible memory wastage if
1116the previous suggestion is implemented, and we swap slabs more frequently.
52960e22 1117
12e06b6f
NC
1118=head2 eliminate incorrect line numbers in warnings
1119
1120This code
1121
1122 use warnings;
1123 my $undef;
1124
1125 if ($undef == 3) {
1126 } elsif ($undef == 0) {
1127 }
1128
18a16cc5 1129used to produce this output:
12e06b6f
NC
1130
1131 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1132 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1133
18a16cc5
NC
1134where the line of the second warning was misreported - it should be line 5.
1135Rafael fixed this - the problem arose because there was no nextstate OP
1136between the execution of the C<if> and the C<elsif>, hence C<PL_curcop> still
1137reports that the currently executing line is line 4. The solution was to inject
1138a nextstate OPs for each C<elsif>, although it turned out that the nextstate
1139OP needed to be a nulled OP, rather than a live nextstate OP, else other line
1140numbers became misreported. (Jenga!)
12e06b6f
NC
1141
1142The problem is more general than C<elsif> (although the C<elsif> case is the
1143most common and the most confusing). Ideally this code
1144
1145 use warnings;
1146 my $undef;
1147
1148 my $a = $undef + 1;
1149 my $b
1150 = $undef
1151 + 1;
1152
1153would produce this output
1154
1155 Use of uninitialized value $undef in addition (+) at wrong.pl line 4.
1156 Use of uninitialized value $undef in addition (+) at wrong.pl line 7.
1157
1158(rather than lines 4 and 5), but this would seem to require every OP to carry
1159(at least) line number information.
1160
1161What might work is to have an optional line number in memory just before the
1162BASEOP structure, with a flag bit in the op to say whether it's present.
1163Initially during compile every OP would carry its line number. Then add a late
1164pass to the optimiser (potentially combined with L</repack the optree>) which
1165looks at the two ops on every edge of the graph of the execution path. If
1166the line number changes, flags the destination OP with this information.
1167Once all paths are traced, replace every op with the flag with a
1168nextstate-light op (that just updates C<PL_curcop>), which in turn then passes
1169control on to the true op. All ops would then be replaced by variants that
1170do not store the line number. (Which, logically, why it would work best in
1171conjunction with L</repack the optree>, as that is already copying/reallocating
1172all the OPs)
1173
18a16cc5
NC
1174(Although I should note that we're not certain that doing this for the general
1175case is worth it)
1176
52960e22
JC
1177=head2 optimize tail-calls
1178
1179Tail-calls present an opportunity for broadly applicable optimization;
1180anywhere that C<< return foo(...) >> is called, the outer return can
1181be replaced by a goto, and foo will return directly to the outer
1182caller, saving (conservatively) 25% of perl's call&return cost, which
1183is relatively higher than in C. The scheme language is known to do
1184this heavily. B::Concise provides good insight into where this
1185optimization is possible, ie anywhere entersub,leavesub op-sequence
1186occurs.
1187
1188 perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)'
1189
1190Bottom line on this is probably a new pp_tailcall function which
1191combines the code in pp_entersub, pp_leavesub. This should probably
1192be done 1st in XS, and using B::Generate to patch the new OP into the
1193optrees.
1194
0c397127
KW
1195=head2 Add C<00dddd>
1196
1197It has been proposed that octal constants be specifiable through the syntax
1198C<0oddddd>, parallel to the existing construct to specify hex constants
1199C<0xddddd>
1200
0bdfc961
NC
1201=head1 Big projects
1202
1203Tasks that will get your name mentioned in the description of the "Highlights
fd2dadea 1204of 5.14"
0bdfc961
NC
1205
1206=head2 make ithreads more robust
1207
4e577f8b 1208Generally make ithreads more robust. See also L</iCOW>
0bdfc961
NC
1209
1210This task is incremental - even a little bit of work on it will help, and
1211will be greatly appreciated.
1212
07577ec1
FC
1213One bit would be to determine how to clone directory handles on systems
1214without a C<fchdir> function (in sv.c:Perl_dirp_dup).
6c047da7 1215
59c7f7d5
RGS
1216Fix Perl_sv_dup, et al so that threads can return objects.
1217
0bdfc961
NC
1218=head2 iCOW
1219
1220Sarathy and Arthur have a proposal for an improved Copy On Write which
1221specifically will be able to COW new ithreads. If this can be implemented
1222it would be a good thing.
1223
1224=head2 (?{...}) closures in regexps
1225
1226Fix (or rewrite) the implementation of the C</(?{...})/> closures.
1227
6bda09f9
YO
1228=head2 Add class set operations to regexp engine
1229
1230Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
1231
1232demerphq has this on his todo list, but right at the bottom.
44a7a252
JV
1233
1234
1235=head1 Tasks for microperl
1236
1237
1238[ Each and every one of these may be obsolete, but they were listed
1239 in the old Todo.micro file]
1240
1241
1242=head2 make creating uconfig.sh automatic
1243
1244=head2 make creating Makefile.micro automatic
1245
1246=head2 do away with fork/exec/wait?
1247
1248(system, popen should be enough?)
1249
1250=head2 some of the uconfig.sh really needs to be probed (using cc) in buildtime:
1251
1252(uConfigure? :-) native datatype widths and endianness come to mind
1253