3 perltodo - Perl TO-DO List
7 This is a list of wishes for Perl. The tasks we think are smaller or easier
8 are listed first. Anyone is welcome to work on any of these, but it's a good
9 idea to first contact I<perl5-porters@perl.org> to avoid duplication of
10 effort. By all means contact a pumpking privately first if you prefer.
12 Whilst patches to make the list shorter are most welcome, ideas to add to
13 the list are also encouraged. Check the perl5-porters archives for past
14 ideas, and any discussion about them. One set of archives may be found at:
16 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
18 What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
19 not, but if your patch is incorporated, then we'll add your name to the
20 F<AUTHORS> file, which ships in the official distribution. How many other
21 programming languages offer you 1 line of immortality?
23 =head1 Tasks that only need Perl knowledge
25 =head2 merge common code in installperl and installman
27 There are some common subroutines and a common C<BEGIN> block in F<installperl>
28 and F<installman>. These should probably be merged. It would also be good to
29 check for duplication in all the utility scripts supplied in the source
30 tarball. It might be good to move them all to a subdirectory, but this would
31 require careful checking to find all places that call them, and change those
34 =head2 common test code for timed bail out
36 Write portable self destruct code for tests to stop them burning CPU in
37 infinite loops. This needs to avoid using alarm, as some of the tests are
38 testing alarm/sleep or timers.
40 =head2 POD -E<gt> HTML conversion in the core still sucks
42 Which is crazy given just how simple POD purports to be, and how simple HTML
43 can be. It's not actually I<as> simple as it sounds, particularly with the
44 flexibility POD allows for C<=item>, but it would be good to improve the
45 visual appeal of the HTML generated, and to avoid it having any validation
46 errors. See also L</make HTML install work>, as the layout of installation tree
47 is needed to improve the cross-linking.
49 The addition of C<Pod::Simple> and its related modules may make this task
52 =head2 merge checkpods and podchecker
54 F<pod/checkpods.PL> (and C<make check> in the F<pod/> subdirectory)
55 implements a very basic check for pod files, but the errors it discovers
56 aren't found by podchecker. Add this check to podchecker, get rid of
57 checkpods and have C<make check> use podchecker.
59 =head2 perlmodlib.PL rewrite
61 Currently perlmodlib.PL needs to be run from a source directory where perl
62 has been built, or some modules won't be found, and others will be
63 skipped. Make it run from a clean perl source tree (so it's reproducible).
65 =head2 Parallel testing
67 (This probably impacts much more than the core: also the Test::Harness
68 and TAP::* modules on CPAN.)
70 The core regression test suite is getting ever more comprehensive, which has
71 the side effect that it takes longer to run. This isn't so good. Investigate
72 whether it would be feasible to give the harness script the B<option> of
73 running sets of tests in parallel. This would be useful for tests in
74 F<t/op/*.t> and F<t/uni/*.t> and maybe some sets of tests in F<lib/>.
82 How does screen layout work when you're running more than one test?
86 How does the caller of test specify how many tests to run in parallel?
90 How do setup/teardown tests identify themselves?
94 Pugs already does parallel testing - can their approach be re-used?
96 =head2 Make Schwern poorer
98 We should have tests for everything. When all the core's modules are tested,
99 Schwern has promised to donate to $500 to TPF. We may need volunteers to
100 hold him upside down and shake vigorously in order to actually extract the
103 =head2 Improve the coverage of the core tests
105 Use Devel::Cover to ascertain the core modules's test coverage, then add
106 tests that are currently missing.
110 A full test suite for the B module would be nice.
112 =head2 Deparse inlined constants
116 use constant PI => 4;
119 will currently deparse as
121 use constant ('PI', 4);
124 because the tokenizer inlines the value of the constant subroutine C<PI>.
125 This allows various compile time optimisations, such as constant folding
126 and dead code elimination. Where these haven't happened (such as the example
127 above) it ought be possible to make B::Deparse work out the name of the
128 original constant, because just enough information survives in the symbol
129 table to do this. Specifically, the same scalar is used for the constant in
130 the optree as is used for the constant subroutine, so by iterating over all
131 symbol tables and generating a mapping of SV address to constant name, it
132 would be possible to provide B::Deparse with this functionality.
134 =head2 A decent benchmark
136 C<perlbench> seems impervious to any recent changes made to the perl core. It
137 would be useful to have a reasonable general benchmarking suite that roughly
138 represented what current perl programs do, and measurably reported whether
139 tweaks to the core improve, degrade or don't really affect performance, to
140 guide people attempting to optimise the guts of perl. Gisle would welcome
141 new tests for perlbench.
143 =head2 fix tainting bugs
145 Fix the bugs revealed by running the test suite with the C<-t> switch (via
146 C<make test.taintwarn>).
148 =head2 Dual life everything
150 As part of the "dists" plan, anything that doesn't belong in the smallest perl
151 distribution needs to be dual lifed. Anything else can be too. Figure out what
152 changes would be needed to package that module and its tests up for CPAN, and
153 do so. Test it with older perl releases, and fix the problems you find.
155 To make a minimal perl distribution, it's useful to look at
156 F<t/lib/commonsense.t>.
158 =head2 Improving C<threads::shared>
160 Investigate whether C<threads::shared> could share aggregates properly with
161 only Perl level changes to shared.pm
163 =head2 POSIX memory footprint
165 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
166 various times worked to cut it down. There is probably still fat to cut out -
167 for example POSIX passes Exporter some very memory hungry data structures.
169 =head2 embed.pl/makedef.pl
171 There is a script F<embed.pl> that generates several header files to prefix
172 all of Perl's symbols in a consistent way, to provide some semblance of
173 namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
174 in F<interpvar.h>. Quite a few of the functions and variables
175 are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
176 doesn't understand the C macros, so the rules about which symbols are present
177 when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
178 It would be good to teach C<embed.pl> to understand the conditional
179 compilation, and hence remove the duplication, and the mistakes it has caused.
181 =head2 use strict; and AutoLoad
183 Currently if you write
186 use AutoLoader 'AUTOLOAD';
191 print join (' ', No, strict, here), "!\n";
194 then C<use strict;> isn't in force within the autoloaded subroutines. It would
195 be more consistent (and less surprising) to arrange for all lexical pragmas
196 in force at the __END__ block to be in force within each autoloaded subroutine.
198 There's a similar problem with SelfLoader.
200 =head1 Tasks that need a little sysadmin-type knowledge
202 Or if you prefer, tasks that you would learn from, and broaden your skills
205 =head2 make HTML install work
207 There is an C<installhtml> target in the Makefile. It's marked as
208 "experimental". It would be good to get this tested, make it work reliably, and
209 remove the "experimental" tag. This would include
215 Checking that cross linking between various parts of the documentation works.
216 In particular that links work between the modules (files with POD in F<lib/>)
217 and the core documentation (files in F<pod/>)
221 Work out how to split C<perlfunc> into chunks, preferably one per function
222 group, preferably with general case code that could be used elsewhere.
223 Challenges here are correctly identifying the groups of functions that go
224 together, and making the right named external cross-links point to the right
225 page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
226 C<endservent>, two or more C<=items> giving the different parameter lists, such
229 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
230 =item substr EXPR,OFFSET,LENGTH
231 =item substr EXPR,OFFSET
233 and different parameter lists having different meanings. (eg C<select>)
237 =head2 compressed man pages
239 Be able to install them. This would probably need a configure test to see how
240 the system does compressed man pages (same directory/different directory?
241 same filename/different filename), as well as tweaking the F<installman> script
242 to compress as necessary.
244 =head2 Add a code coverage target to the Makefile
246 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
247 to do this manually are roughly
253 do a normal C<Configure>, but include Devel::Cover as a module to install
254 (see F<INSTALL> for how to do this)
262 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
266 Process the resulting Devel::Cover database
270 This just give you the coverage of the F<.pm>s. To also get the C level
277 Additionally tell C<Configure> to use the appropriate C compiler flags for
284 (instead of C<make perl>)
288 After running the tests run C<gcov> to generate all the F<.gcov> files.
289 (Including down in the subdirectories of F<ext/>
293 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
294 to get their stats into the cover_db directory.
298 Then process the Devel::Cover database
302 It would be good to add a single switch to C<Configure> to specify that you
303 wanted to perform perl level coverage, and another to specify C level
304 coverage, and have C<Configure> and the F<Makefile> do all the right things
307 =head2 Make Config.pm cope with differences between built and installed perl
309 Quite often vendors ship a perl binary compiled with their (pay-for)
310 compilers. People install a free compiler, such as gcc. To work out how to
311 build extensions, Perl interrogates C<%Config>, so in this situation
312 C<%Config> describes compilers that aren't there, and extension building
313 fails. This forces people into choosing between re-compiling perl themselves
314 using the compiler they have, or only using modules that the vendor ships.
316 It would be good to find a way teach C<Config.pm> about the installation setup,
317 possibly involving probing at install time or later, so that the C<%Config> in
318 a binary distribution better describes the installed machine, when the
319 installed machine differs from the build machine in some significant way.
321 =head2 linker specification files
323 Some platforms mandate that you provide a list of a shared library's external
324 symbols to the linker, so the core already has the infrastructure in place to
325 do this for generating shared perl libraries. My understanding is that the
326 GNU toolchain can accept an optional linker specification file, and restrict
327 visibility just to symbols declared in that file. It would be good to extend
328 F<makedef.pl> to support this format, and to provide a means within
329 C<Configure> to enable it. This would allow Unix users to test that the
330 export list is correct, and to build a perl that does not pollute the global
331 namespace with private symbols.
333 =head2 Cross-compile support
335 Currently C<Configure> understands C<-Dusecrosscompile> option. This option
336 arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
337 assumed then to be copied to TARGET machine and used as a replacement of full
340 This could be done little differently. Namely C<miniperl> should be built for
341 HOST and then full C<perl> with extensions should be compiled for TARGET.
342 This, however, might require extra trickery for %Config: we have one config
343 first for HOST and then another for TARGET. Tools like MakeMaker will be
344 mightily confused. Having around two different types of executables and
345 libraries (HOST and TARGET) makes life interesting for Makefiles and
346 shell (and Perl) scripts. There is $Config{run}, normally empty, which
347 can be used as an execution wrapper. Also note that in some
348 cross-compilation/execution environments the HOST and the TARGET do
349 not see the same filesystem(s), the $Config{run} may need to do some
350 file/directory copying back and forth.
354 Make F<pod/roffitall> be updated by F<pod/buildtoc>.
356 =head1 Tasks that need a little C knowledge
358 These tasks would need a little C knowledge, but don't need any specific
359 background or experience with XS, or how the Perl interpreter works
361 =head2 Exterminate PL_na!
363 C<PL_na> festers still in the darkest corners of various typemap files.
364 It needs to be exterminated, replaced by a local variable of type C<STRLEN>.
366 =head2 Modernize the order of directories in @INC
368 The way @INC is laid out by default, one cannot upgrade core (dual-life)
369 modules without overwriting files. This causes problems for binary
370 package builders. One possible proposal is laid out in this
372 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>.
376 Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
377 On these systems, it might be the default compilation mode, and there
378 is currently no guarantee that passing no use64bitall option to the
379 Configure process will build a 32bit perl. Implementing -Duse32bit*
380 options would be nice for perl 5.12.
382 =head2 Make it clear from -v if this is the exact official release
384 Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that
385 usually defines one local patch, of the form "MAINT12345" or "RC1". The output
386 of perl -v doesn't report that a perl isn't an official release, and this
387 information can get lost in bugs reports. Because of this, the minor version
388 isn't bumped up until RC time, to minimise the possibility of versions of perl
389 escaping that believe themselves to be newer than they actually are.
391 It would be useful to find an elegant way to have the "this is an interim
392 maintenance release" or "this is a release candidate" in the terse -v output,
393 and have it so that it's easy for the pumpking to remove this just as the
394 release tarball is rolled up. This way the version pulled out of rsync would
395 always say "I'm a development release" and it would be safe to bump the
396 reported minor version as soon as a release ships, which would aid perl
399 This task is really about thinking of an elegant way to arrange the C source
400 such that it's trivial for the Pumpking to flag "this is an official release"
401 when making a tarball, yet leave the default source saying "I'm not the
404 =head2 Profile Perl - am I hot or not?
406 The Perl source code is stable enough that it makes sense to profile it,
407 identify and optimise the hotspots. It would be good to measure the
408 performance of the Perl interpreter using free tools such as cachegrind,
409 gprof, and dtrace, and work to reduce the bottlenecks they reveal.
411 As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
412 the ops that are most commonly used. The idea is that by grouping them, their
413 object code will be adjacent in the executable, so they have a greater chance
414 of already being in the CPU cache (or swapped in) due to being near another op
417 Except that it's not clear if these really are the most commonly used ops. So
418 as part of exercising your skills with coverage and profiling tools you might
419 want to determine what ops I<really> are the most commonly used. And in turn
420 suggest evictions and promotions to achieve a better F<pp_hot.c>.
422 =head2 Allocate OPs from arenas
424 Currently all new OP structures are individually malloc()ed and free()d.
425 All C<malloc> implementations have space overheads, and are now as fast as
426 custom allocates so it would both use less memory and less CPU to allocate
427 the various OP structures from arenas. The SV arena code can probably be
430 Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use
431 Perl_Slab_alloc() to pack optrees into a contiguous block, which is
432 probably superior to the use of OP arenas, esp. from a cache locality
433 standpoint. See L<Profile Perl - am I hot or not?>.
435 =head2 Improve win32/wince.c
437 Currently, numerous functions look virtually, if not completely,
438 identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't
441 =head2 Use secure CRT functions when building with VC8 on Win32
443 Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
444 that they were "unsafe" and introduced differently named secure versions of
445 them as replacements, e.g. instead of writing
447 FILE* f = fopen(__FILE__, "r");
452 errno_t err = fopen_s(&f, __FILE__, "r");
454 Currently, the warnings about these deprecations have been disabled by adding
455 -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
456 warning suppressant and actually make use of the new secure CRT functions.
458 There is also a similar issue with POSIX CRT function names like fileno having
459 been deprecated in favour of ISO C++ conformant names like _fileno. These
460 warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
461 might be nice to do as Microsoft suggest here too, although, unlike the secure
462 functions issue, there is presumably little or no benefit in this case.
464 =head2 __FUNCTION__ for MSVC-pre-7.0
466 Jarkko notes that one can things morally equivalent to C<__FUNCTION__>
467 (or C<__func__>) even in MSVC-pre-7.0, contrary to popular belief.
468 See L<http://www.codeproject.com/debug/extendedtrace.asp> if you feel like
469 making C<PERL_MEM_LOG> more useful on Win32.
471 =head1 Tasks that need a knowledge of XS
473 These tasks would need C knowledge, and roughly the level of knowledge of
474 the perl API that comes from writing modules that use XS to interface to
477 =head2 autovivification
479 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
481 This task is incremental - even a little bit of work on it will help.
483 =head2 Unicode in Filenames
485 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
486 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
487 system, truncate, unlink, utime, -X. All these could potentially accept
488 Unicode filenames either as input or output (and in the case of system
489 and qx Unicode in general, as input or output to/from the shell).
490 Whether a filesystem - an operating system pair understands Unicode in
493 Known combinations that have some level of understanding include
494 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
495 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
496 create Unicode filenames, what forms of Unicode are accepted and used
497 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
498 and so on, varies. Finding the right level of interfacing to Perl
499 requires some thought. Remember that an OS does not implicate a
502 (The Windows -C command flag "wide API support" has been at least
503 temporarily retired in 5.8.1, and the -C has been repurposed, see
506 Most probably the right way to do this would be this:
507 L</"Virtualize operating system access">.
509 =head2 Unicode in %ENV
511 Currently the %ENV entries are always byte strings.
512 See L</"Virtualize operating system access">.
514 =head2 Unicode and glob()
516 Currently glob patterns and filenames returned from File::Glob::glob()
517 are always byte strings. See L</"Virtualize operating system access">.
519 =head2 Unicode and lc/uc operators
521 Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on
522 what the internal encoding of their argument is. That should not be the
523 case. Maybe add a pragma to switch behaviour.
525 =head2 use less 'memory'
527 Investigate trade offs to switch out perl's choices on memory usage.
528 Particularly perl should be able to give memory back.
530 This task is incremental - even a little bit of work on it will help.
532 =head2 Re-implement C<:unique> in a way that is actually thread-safe
534 The old implementation made bad assumptions on several levels. A good 90%
535 solution might be just to make C<:unique> work to share the string buffer
536 of SvPVs. That way large constant strings can be shared between ithreads,
537 such as the configuration information in F<Config>.
539 =head2 Make tainting consistent
541 Tainting would be easier to use if it didn't take documented shortcuts and
542 allow taint to "leak" everywhere within an expression.
544 =head2 readpipe(LIST)
546 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
547 running a shell. readpipe() (the function behind qx//) could be similarly
550 =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
552 Maybe create a utility that checks after each libperl.a creation that
553 none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
554 ever creep back to libperl.a.
556 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
558 Note, of course, that this will only tell whether B<your> platform
559 is using those naughty interfaces.
561 =head2 Audit the code for destruction ordering assumptions
565 /* Need to check SvMAGICAL, as during global destruction it may be that
566 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
567 is now part of the linked list of SV heads, rather than pointing to
568 the original body. */
569 /* FIXME - audit the code for other bugs like this one. */
571 adding the C<SvMAGICAL> check to
573 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
574 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
576 Go through the core and look for similar assumptions that SVs have particular
577 types, as all bets are off during global destruction.
579 =head2 Extend PerlIO and PerlIO::Scalar
581 PerlIO::Scalar doesn't know how to truncate(). Implementing this
582 would require extending the PerlIO vtable.
584 Similarly the PerlIO vtable doesn't know about formats (write()), or
585 about stat(), or chmod()/chown(), utime(), or flock().
587 (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
590 PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
591 opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
594 See also L</"Virtualize operating system access">.
596 =head2 -C on the #! line
598 It should be possible to make -C work correctly if found on the #! line,
599 given that all perl command line options are strict ASCII, and -C changes
600 only the interpretation of non-ASCII characters, and not for the script file
601 handle. To make it work needs some investigation of the ordering of function
602 calls during startup, and (by implication) a bit of tweaking of that order.
604 =head2 Propagate const outwards from Perl_moreswitches()
606 Change 32057 changed the parameter and return value of C<Perl_moreswitches()>
607 from <char *> to <const char *>. It should now be possible to propagate
608 const-correctness outwards to C<S_parse_body()>, C<Perl_moreswitches()>
612 =head1 Tasks that need a knowledge of the interpreter
614 These tasks would need C knowledge, and knowledge of how the interpreter works,
615 or a willingness to learn.
617 =head2 state variable initialization in list context
619 Currently this is illegal:
621 state ($a, $b) = foo();
623 The current Perl 6 design is that C<state ($a) = foo();> and
624 C<(state $a) = foo();> have different semantics, which is tricky to implement
625 in Perl 5 as currently the produce the same opcode trees. It would be useful
626 to clarify that the Perl 6 design is firm, and then implement the necessary
627 code in Perl 5. There are comments in C<Perl_newASSIGNOP()> that show the
628 code paths taken by various assignment constructions involving state variables.
630 =head2 Implement $value ~~ 0 .. $range
632 It would be nice to extend the syntax of the C<~~> operator to also
633 understand numeric (and maybe alphanumeric) ranges.
635 =head2 A does() built-in
637 Like ref(), only useful. It would call the C<DOES> method on objects; it
638 would also tell whether something can be dereferenced as an
639 array/hash/etc., or used as a regexp, etc.
640 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
642 =head2 Tied filehandles and write() don't mix
644 There is no method on tied filehandles to allow them to be called back by
647 =head2 Attach/detach debugger from running program
649 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
650 program if you pass the process ID. It would be good to do this with the Perl
651 debugger on a running Perl program, although I'm not sure how it would be
652 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
654 =head2 Optimize away empty destructors
656 Defining an empty DESTROY method might be useful (notably in
657 AUTOLOAD-enabled classes), but it's still a bit expensive to call. That
658 could probably be optimized.
660 =head2 LVALUE functions for lists
662 The old perltodo notes that lvalue functions don't work for list or hash
663 slices. This would be good to fix.
665 =head2 LVALUE functions in the debugger
667 The old perltodo notes that lvalue functions don't work in the debugger. This
668 would be good to fix.
670 =head2 regexp optimiser optional
672 The regexp optimiser is not optional. It should configurable to be, to allow
673 its performance to be measured, and its bugs to be easily demonstrated.
675 =head2 delete &function
677 Allow to delete functions. One can already undef them, but they're still
680 =head2 C</w> regex modifier
682 That flag would enable to match whole words, and also to interpolate
683 arrays as alternations. With it, C</P/w> would be roughly equivalent to:
685 do { local $"='|'; /\b(?:P)\b/ }
687 See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
690 =head2 optional optimizer
692 Make the peephole optimizer optional. Currently it performs two tasks as
693 it walks the optree - genuine peephole optimisations, and necessary fixups of
694 ops. It would be good to find an efficient way to switch out the
695 optimisations whilst keeping the fixups.
697 =head2 You WANT *how* many
699 Currently contexts are void, scalar and list. split has a special mechanism in
700 place to pass in the number of return values wanted. It would be useful to
701 have a general mechanism for this, backwards compatible and little speed hit.
702 This would allow proposals such as short circuiting sort to be implemented
705 =head2 lexical aliases
707 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
709 =head2 entersub XS vs Perl
711 At the moment pp_entersub is huge, and has code to deal with entering both
712 perl and XS subroutines. Subroutine implementations rarely change between
713 perl and XS at run time, so investigate using 2 ops to enter subs (one for
714 XS, one for perl) and swap between if a sub is redefined.
718 self ties are currently illegal because they caused too many segfaults. Maybe
719 the causes of these could be tracked down and self-ties on all types re-
722 =head2 Optimize away @_
724 The old perltodo notes "Look at the "reification" code in C<av.c>".
726 =head2 Properly Unicode safe tokeniser and pads.
728 The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
729 variable names are stored in stashes as raw bytes, without the utf-8 flag
730 set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
731 tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
732 source filters. All this could be fixed.
734 =head2 The yada yada yada operators
736 Perl 6's Synopsis 3 says:
738 I<The ... operator is the "yada, yada, yada" list operator, which is used as
739 the body in function prototypes. It complains bitterly (by calling fail)
740 if it is ever executed. Variant ??? calls warn, and !!! calls die.>
742 Those would be nice to add to Perl 5. That could be done without new ops.
744 =head2 Virtualize operating system access
746 Implement a set of "vtables" that virtualizes operating system access
747 (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
748 least these interfaces should take SVs as "name" arguments instead of
749 bare char pointers; probably the most flexible and extensible way
750 would be for the Perl-facing interfaces to accept HVs. The system
751 needs to be per-operating-system and per-file-system
752 hookable/filterable, preferably both from XS and Perl level
753 (L<perlport/"Files and Filesystems"> is good reading at this point,
754 in fact, all of L<perlport> is.)
756 This has actually already been implemented (but only for Win32),
757 take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
758 variants go through a set of "vtables" for operating system access,
759 non-Win32 systems currently go straight for the POSIX/UNIX-style
760 system/library call. Similar system as for Win32 should be
761 implemented for all platforms. The existing Win32 implementation
762 probably does not need to survive alongside this proposed new
763 implementation, the approaches could be merged.
765 What would this give us? One often-asked-for feature this would
766 enable is using Unicode for filenames, and other "names" like %ENV,
767 usernames, hostnames, and so forth.
768 (See L<perlunicode/"When Unicode Does Not Happen">.)
770 But this kind of virtualization would also allow for things like
771 virtual filesystems, virtual networks, and "sandboxes" (though as long
772 as dynamic loading of random object code is allowed, not very safe
773 sandboxes since external code of course know not of Perl's vtables).
774 An example of a smaller "sandbox" is that this feature can be used to
775 implement per-thread working directories: Win32 already does this.
777 See also L</"Extend PerlIO and PerlIO::Scalar">.
779 =head2 Investigate PADTMP hash pessimisation
781 The peephole optimier converts constants used for hash key lookups to shared
782 hash key scalars. Under ithreads, something is undoing this work. See
783 See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html
787 Tasks that will get your name mentioned in the description of the "Highlights
790 =head2 make ithreads more robust
792 Generally make ithreads more robust. See also L</iCOW>
794 This task is incremental - even a little bit of work on it will help, and
795 will be greatly appreciated.
797 One bit would be to write the missing code in sv.c:Perl_dirp_dup.
799 Fix Perl_sv_dup, et al so that threads can return objects.
803 Sarathy and Arthur have a proposal for an improved Copy On Write which
804 specifically will be able to COW new ithreads. If this can be implemented
805 it would be a good thing.
807 =head2 (?{...}) closures in regexps
809 Fix (or rewrite) the implementation of the C</(?{...})/> closures.
811 =head2 A re-entrant regexp engine
813 This will allow the use of a regex from inside (?{ }), (??{ }) and
814 (?(?{ })|) constructs.
816 =head2 Add class set operations to regexp engine
818 Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
820 demerphq has this on his todo list, but right at the bottom.