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 common test code for timed bail out
27 Write portable self destruct code for tests to stop them burning CPU in
28 infinite loops. This needs to avoid using alarm, as some of the tests are
29 testing alarm/sleep or timers.
31 =head2 POD -E<gt> HTML conversion in the core still sucks
33 Which is crazy given just how simple POD purports to be, and how simple HTML
34 can be. It's not actually I<as> simple as it sounds, particularly with the
35 flexibility POD allows for C<=item>, but it would be good to improve the
36 visual appeal of the HTML generated, and to avoid it having any validation
37 errors. See also L</make HTML install work>, as the layout of installation tree
38 is needed to improve the cross-linking.
40 The addition of C<Pod::Simple> and its related modules may make this task
43 =head2 Parallel testing
45 (This probably impacts much more than the core: also the Test::Harness
46 and TAP::* modules on CPAN.)
48 The core regression test suite is getting ever more comprehensive, which has
49 the side effect that it takes longer to run. This isn't so good. Investigate
50 whether it would be feasible to give the harness script the B<option> of
51 running sets of tests in parallel. This would be useful for tests in
52 F<t/op/*.t> and F<t/uni/*.t> and maybe some sets of tests in F<lib/>.
60 How does screen layout work when you're running more than one test?
64 How does the caller of test specify how many tests to run in parallel?
68 How do setup/teardown tests identify themselves?
72 Pugs already does parallel testing - can their approach be re-used?
74 =head2 Make Schwern poorer
76 We should have tests for everything. When all the core's modules are tested,
77 Schwern has promised to donate to $500 to TPF. We may need volunteers to
78 hold him upside down and shake vigorously in order to actually extract the
81 =head2 Improve the coverage of the core tests
83 Use Devel::Cover to ascertain the core modules's test coverage, then add
84 tests that are currently missing.
88 A full test suite for the B module would be nice.
90 =head2 Deparse inlined constants
97 will currently deparse as
99 use constant ('PI', 4);
102 because the tokenizer inlines the value of the constant subroutine C<PI>.
103 This allows various compile time optimisations, such as constant folding
104 and dead code elimination. Where these haven't happened (such as the example
105 above) it ought be possible to make B::Deparse work out the name of the
106 original constant, because just enough information survives in the symbol
107 table to do this. Specifically, the same scalar is used for the constant in
108 the optree as is used for the constant subroutine, so by iterating over all
109 symbol tables and generating a mapping of SV address to constant name, it
110 would be possible to provide B::Deparse with this functionality.
112 =head2 A decent benchmark
114 C<perlbench> seems impervious to any recent changes made to the perl core. It
115 would be useful to have a reasonable general benchmarking suite that roughly
116 represented what current perl programs do, and measurably reported whether
117 tweaks to the core improve, degrade or don't really affect performance, to
118 guide people attempting to optimise the guts of perl. Gisle would welcome
119 new tests for perlbench.
121 =head2 fix tainting bugs
123 Fix the bugs revealed by running the test suite with the C<-t> switch (via
124 C<make test.taintwarn>).
126 =head2 Dual life everything
128 As part of the "dists" plan, anything that doesn't belong in the smallest perl
129 distribution needs to be dual lifed. Anything else can be too. Figure out what
130 changes would be needed to package that module and its tests up for CPAN, and
131 do so. Test it with older perl releases, and fix the problems you find.
133 To make a minimal perl distribution, it's useful to look at
134 F<t/lib/commonsense.t>.
136 =head2 Improving C<threads::shared>
138 Investigate whether C<threads::shared> could share aggregates properly with
139 only Perl level changes to shared.pm
141 =head2 POSIX memory footprint
143 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
144 various times worked to cut it down. There is probably still fat to cut out -
145 for example POSIX passes Exporter some very memory hungry data structures.
147 =head2 embed.pl/makedef.pl
149 There is a script F<embed.pl> that generates several header files to prefix
150 all of Perl's symbols in a consistent way, to provide some semblance of
151 namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
152 in F<interpvar.h>. Quite a few of the functions and variables
153 are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
154 doesn't understand the C macros, so the rules about which symbols are present
155 when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
156 It would be good to teach C<embed.pl> to understand the conditional
157 compilation, and hence remove the duplication, and the mistakes it has caused.
159 =head2 use strict; and AutoLoad
161 Currently if you write
164 use AutoLoader 'AUTOLOAD';
169 print join (' ', No, strict, here), "!\n";
172 then C<use strict;> isn't in force within the autoloaded subroutines. It would
173 be more consistent (and less surprising) to arrange for all lexical pragmas
174 in force at the __END__ block to be in force within each autoloaded subroutine.
176 There's a similar problem with SelfLoader.
178 =head1 Tasks that need a little sysadmin-type knowledge
180 Or if you prefer, tasks that you would learn from, and broaden your skills
183 =head2 make HTML install work
185 There is an C<installhtml> target in the Makefile. It's marked as
186 "experimental". It would be good to get this tested, make it work reliably, and
187 remove the "experimental" tag. This would include
193 Checking that cross linking between various parts of the documentation works.
194 In particular that links work between the modules (files with POD in F<lib/>)
195 and the core documentation (files in F<pod/>)
199 Work out how to split C<perlfunc> into chunks, preferably one per function
200 group, preferably with general case code that could be used elsewhere.
201 Challenges here are correctly identifying the groups of functions that go
202 together, and making the right named external cross-links point to the right
203 page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
204 C<endservent>, two or more C<=items> giving the different parameter lists, such
207 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
208 =item substr EXPR,OFFSET,LENGTH
209 =item substr EXPR,OFFSET
211 and different parameter lists having different meanings. (eg C<select>)
215 =head2 compressed man pages
217 Be able to install them. This would probably need a configure test to see how
218 the system does compressed man pages (same directory/different directory?
219 same filename/different filename), as well as tweaking the F<installman> script
220 to compress as necessary.
222 =head2 Add a code coverage target to the Makefile
224 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
225 to do this manually are roughly
231 do a normal C<Configure>, but include Devel::Cover as a module to install
232 (see F<INSTALL> for how to do this)
240 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
244 Process the resulting Devel::Cover database
248 This just give you the coverage of the F<.pm>s. To also get the C level
255 Additionally tell C<Configure> to use the appropriate C compiler flags for
262 (instead of C<make perl>)
266 After running the tests run C<gcov> to generate all the F<.gcov> files.
267 (Including down in the subdirectories of F<ext/>
271 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
272 to get their stats into the cover_db directory.
276 Then process the Devel::Cover database
280 It would be good to add a single switch to C<Configure> to specify that you
281 wanted to perform perl level coverage, and another to specify C level
282 coverage, and have C<Configure> and the F<Makefile> do all the right things
285 =head2 Make Config.pm cope with differences between built and installed perl
287 Quite often vendors ship a perl binary compiled with their (pay-for)
288 compilers. People install a free compiler, such as gcc. To work out how to
289 build extensions, Perl interrogates C<%Config>, so in this situation
290 C<%Config> describes compilers that aren't there, and extension building
291 fails. This forces people into choosing between re-compiling perl themselves
292 using the compiler they have, or only using modules that the vendor ships.
294 It would be good to find a way teach C<Config.pm> about the installation setup,
295 possibly involving probing at install time or later, so that the C<%Config> in
296 a binary distribution better describes the installed machine, when the
297 installed machine differs from the build machine in some significant way.
299 =head2 linker specification files
301 Some platforms mandate that you provide a list of a shared library's external
302 symbols to the linker, so the core already has the infrastructure in place to
303 do this for generating shared perl libraries. My understanding is that the
304 GNU toolchain can accept an optional linker specification file, and restrict
305 visibility just to symbols declared in that file. It would be good to extend
306 F<makedef.pl> to support this format, and to provide a means within
307 C<Configure> to enable it. This would allow Unix users to test that the
308 export list is correct, and to build a perl that does not pollute the global
309 namespace with private symbols.
311 =head2 Cross-compile support
313 Currently C<Configure> understands C<-Dusecrosscompile> option. This option
314 arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
315 assumed then to be copied to TARGET machine and used as a replacement of full
318 This could be done little differently. Namely C<miniperl> should be built for
319 HOST and then full C<perl> with extensions should be compiled for TARGET.
320 This, however, might require extra trickery for %Config: we have one config
321 first for HOST and then another for TARGET. Tools like MakeMaker will be
322 mightily confused. Having around two different types of executables and
323 libraries (HOST and TARGET) makes life interesting for Makefiles and
324 shell (and Perl) scripts. There is $Config{run}, normally empty, which
325 can be used as an execution wrapper. Also note that in some
326 cross-compilation/execution environments the HOST and the TARGET do
327 not see the same filesystem(s), the $Config{run} may need to do some
328 file/directory copying back and forth.
330 =head1 Tasks that need a little C knowledge
332 These tasks would need a little C knowledge, but don't need any specific
333 background or experience with XS, or how the Perl interpreter works
335 =head2 Modernize the order of directories in @INC
337 The way @INC is laid out by default, one cannot upgrade core (dual-life)
338 modules without overwriting files. This causes problems for binary
339 package builders. One possible proposal is laid out in this
341 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>.
345 Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
346 On these systems, it might be the default compilation mode, and there
347 is currently no guarantee that passing no use64bitall option to the
348 Configure process will build a 32bit perl. Implementing -Duse32bit*
349 options would be nice for perl 5.12.
351 =head2 Make it clear from -v if this is the exact official release
353 Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that
354 usually defines one local patch, of the form "MAINT12345" or "RC1". The output
355 of perl -v doesn't report that a perl isn't an official release, and this
356 information can get lost in bugs reports. Because of this, the minor version
357 isn't bumped up until RC time, to minimise the possibility of versions of perl
358 escaping that believe themselves to be newer than they actually are.
360 It would be useful to find an elegant way to have the "this is an interim
361 maintenance release" or "this is a release candidate" in the terse -v output,
362 and have it so that it's easy for the pumpking to remove this just as the
363 release tarball is rolled up. This way the version pulled out of rsync would
364 always say "I'm a development release" and it would be safe to bump the
365 reported minor version as soon as a release ships, which would aid perl
368 This task is really about thinking of an elegant way to arrange the C source
369 such that it's trivial for the Pumpking to flag "this is an official release"
370 when making a tarball, yet leave the default source saying "I'm not the
373 =head2 Profile Perl - am I hot or not?
375 The Perl source code is stable enough that it makes sense to profile it,
376 identify and optimise the hotspots. It would be good to measure the
377 performance of the Perl interpreter using free tools such as cachegrind,
378 gprof, and dtrace, and work to reduce the bottlenecks they reveal.
380 As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
381 the ops that are most commonly used. The idea is that by grouping them, their
382 object code will be adjacent in the executable, so they have a greater chance
383 of already being in the CPU cache (or swapped in) due to being near another op
386 Except that it's not clear if these really are the most commonly used ops. So
387 as part of exercising your skills with coverage and profiling tools you might
388 want to determine what ops I<really> are the most commonly used. And in turn
389 suggest evictions and promotions to achieve a better F<pp_hot.c>.
391 =head2 Allocate OPs from arenas
393 Currently all new OP structures are individually malloc()ed and free()d.
394 All C<malloc> implementations have space overheads, and are now as fast as
395 custom allocates so it would both use less memory and less CPU to allocate
396 the various OP structures from arenas. The SV arena code can probably be
399 Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use
400 Perl_Slab_alloc() to pack optrees into a contiguous block, which is
401 probably superior to the use of OP arenas, esp. from a cache locality
402 standpoint. See L<Profile Perl - am I hot or not?>.
404 =head2 Improve win32/wince.c
406 Currently, numerous functions look virtually, if not completely,
407 identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't
410 =head2 Use secure CRT functions when building with VC8 on Win32
412 Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
413 that they were "unsafe" and introduced differently named secure versions of
414 them as replacements, e.g. instead of writing
416 FILE* f = fopen(__FILE__, "r");
421 errno_t err = fopen_s(&f, __FILE__, "r");
423 Currently, the warnings about these deprecations have been disabled by adding
424 -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
425 warning suppressant and actually make use of the new secure CRT functions.
427 There is also a similar issue with POSIX CRT function names like fileno having
428 been deprecated in favour of ISO C++ conformant names like _fileno. These
429 warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
430 might be nice to do as Microsoft suggest here too, although, unlike the secure
431 functions issue, there is presumably little or no benefit in this case.
433 =head2 __FUNCTION__ for MSVC-pre-7.0
435 Jarkko notes that one can things morally equivalent to C<__FUNCTION__>
436 (or C<__func__>) even in MSVC-pre-7.0, contrary to popular belief.
437 See L<http://www.codeproject.com/debug/extendedtrace.asp> if you feel like
438 making C<PERL_MEM_LOG> more useful on Win32.
440 =head1 Tasks that need a knowledge of XS
442 These tasks would need C knowledge, and roughly the level of knowledge of
443 the perl API that comes from writing modules that use XS to interface to
446 =head2 autovivification
448 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
450 This task is incremental - even a little bit of work on it will help.
452 =head2 Unicode in Filenames
454 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
455 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
456 system, truncate, unlink, utime, -X. All these could potentially accept
457 Unicode filenames either as input or output (and in the case of system
458 and qx Unicode in general, as input or output to/from the shell).
459 Whether a filesystem - an operating system pair understands Unicode in
462 Known combinations that have some level of understanding include
463 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
464 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
465 create Unicode filenames, what forms of Unicode are accepted and used
466 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
467 and so on, varies. Finding the right level of interfacing to Perl
468 requires some thought. Remember that an OS does not implicate a
471 (The Windows -C command flag "wide API support" has been at least
472 temporarily retired in 5.8.1, and the -C has been repurposed, see
475 Most probably the right way to do this would be this:
476 L</"Virtualize operating system access">.
478 =head2 Unicode in %ENV
480 Currently the %ENV entries are always byte strings.
481 See L</"Virtualize operating system access">.
483 =head2 Unicode and glob()
485 Currently glob patterns and filenames returned from File::Glob::glob()
486 are always byte strings. See L</"Virtualize operating system access">.
488 =head2 Unicode and lc/uc operators
490 Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on
491 what the internal encoding of their argument is. That should not be the
492 case. Maybe add a pragma to switch behaviour.
494 =head2 use less 'memory'
496 Investigate trade offs to switch out perl's choices on memory usage.
497 Particularly perl should be able to give memory back.
499 This task is incremental - even a little bit of work on it will help.
501 =head2 Re-implement C<:unique> in a way that is actually thread-safe
503 The old implementation made bad assumptions on several levels. A good 90%
504 solution might be just to make C<:unique> work to share the string buffer
505 of SvPVs. That way large constant strings can be shared between ithreads,
506 such as the configuration information in F<Config>.
508 =head2 Make tainting consistent
510 Tainting would be easier to use if it didn't take documented shortcuts and
511 allow taint to "leak" everywhere within an expression.
513 =head2 readpipe(LIST)
515 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
516 running a shell. readpipe() (the function behind qx//) could be similarly
519 =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
521 Maybe create a utility that checks after each libperl.a creation that
522 none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
523 ever creep back to libperl.a.
525 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
527 Note, of course, that this will only tell whether B<your> platform
528 is using those naughty interfaces.
530 =head2 Audit the code for destruction ordering assumptions
534 /* Need to check SvMAGICAL, as during global destruction it may be that
535 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
536 is now part of the linked list of SV heads, rather than pointing to
537 the original body. */
538 /* FIXME - audit the code for other bugs like this one. */
540 adding the C<SvMAGICAL> check to
542 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
543 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
545 Go through the core and look for similar assumptions that SVs have particular
546 types, as all bets are off during global destruction.
548 =head2 Extend PerlIO and PerlIO::Scalar
550 PerlIO::Scalar doesn't know how to truncate(). Implementing this
551 would require extending the PerlIO vtable.
553 Similarly the PerlIO vtable doesn't know about formats (write()), or
554 about stat(), or chmod()/chown(), utime(), or flock().
556 (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
559 PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
560 opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
563 See also L</"Virtualize operating system access">.
565 =head2 -C on the #! line
567 It should be possible to make -C work correctly if found on the #! line,
568 given that all perl command line options are strict ASCII, and -C changes
569 only the interpretation of non-ASCII characters, and not for the script file
570 handle. To make it work needs some investigation of the ordering of function
571 calls during startup, and (by implication) a bit of tweaking of that order.
574 =head1 Tasks that need a knowledge of the interpreter
576 These tasks would need C knowledge, and knowledge of how the interpreter works,
577 or a willingness to learn.
579 =head2 state variable initialization in list context
581 Currently this is illegal:
583 state ($a, $b) = foo();
585 The current Perl 6 design is that C<state ($a) = foo();> and
586 C<(state $a) = foo();> have different semantics, which is tricky to implement
587 in Perl 5 as currently the produce the same opcode trees. It would be useful
588 to clarify that the Perl 6 design is firm, and then implement the necessary
589 code in Perl 5. There are comments in C<Perl_newASSIGNOP()> that show the
590 code paths taken by various assignment constructions involving state variables.
592 =head2 Implement $value ~~ 0 .. $range
594 It would be nice to extend the syntax of the C<~~> operator to also
595 understand numeric (and maybe alphanumeric) ranges.
597 =head2 A does() built-in
599 Like ref(), only useful. It would call the C<DOES> method on objects; it
600 would also tell whether something can be dereferenced as an
601 array/hash/etc., or used as a regexp, etc.
602 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
604 =head2 Tied filehandles and write() don't mix
606 There is no method on tied filehandles to allow them to be called back by
609 =head2 Attach/detach debugger from running program
611 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
612 program if you pass the process ID. It would be good to do this with the Perl
613 debugger on a running Perl program, although I'm not sure how it would be
614 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
616 =head2 Optimize away empty destructors
618 Defining an empty DESTROY method might be useful (notably in
619 AUTOLOAD-enabled classes), but it's still a bit expensive to call. That
620 could probably be optimized.
622 =head2 LVALUE functions for lists
624 The old perltodo notes that lvalue functions don't work for list or hash
625 slices. This would be good to fix.
627 =head2 LVALUE functions in the debugger
629 The old perltodo notes that lvalue functions don't work in the debugger. This
630 would be good to fix.
632 =head2 regexp optimiser optional
634 The regexp optimiser is not optional. It should configurable to be, to allow
635 its performance to be measured, and its bugs to be easily demonstrated.
637 =head2 delete &function
639 Allow to delete functions. One can already undef them, but they're still
642 =head2 C</w> regex modifier
644 That flag would enable to match whole words, and also to interpolate
645 arrays as alternations. With it, C</P/w> would be roughly equivalent to:
647 do { local $"='|'; /\b(?:P)\b/ }
649 See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
652 =head2 optional optimizer
654 Make the peephole optimizer optional. Currently it performs two tasks as
655 it walks the optree - genuine peephole optimisations, and necessary fixups of
656 ops. It would be good to find an efficient way to switch out the
657 optimisations whilst keeping the fixups.
659 =head2 You WANT *how* many
661 Currently contexts are void, scalar and list. split has a special mechanism in
662 place to pass in the number of return values wanted. It would be useful to
663 have a general mechanism for this, backwards compatible and little speed hit.
664 This would allow proposals such as short circuiting sort to be implemented
667 =head2 lexical aliases
669 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
671 =head2 entersub XS vs Perl
673 At the moment pp_entersub is huge, and has code to deal with entering both
674 perl and XS subroutines. Subroutine implementations rarely change between
675 perl and XS at run time, so investigate using 2 ops to enter subs (one for
676 XS, one for perl) and swap between if a sub is redefined.
680 self ties are currently illegal because they caused too many segfaults. Maybe
681 the causes of these could be tracked down and self-ties on all types re-
684 =head2 Optimize away @_
686 The old perltodo notes "Look at the "reification" code in C<av.c>".
688 =head2 Properly Unicode safe tokeniser and pads.
690 The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
691 variable names are stored in stashes as raw bytes, without the utf-8 flag
692 set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
693 tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
694 source filters. All this could be fixed.
696 =head2 The yada yada yada operators
698 Perl 6's Synopsis 3 says:
700 I<The ... operator is the "yada, yada, yada" list operator, which is used as
701 the body in function prototypes. It complains bitterly (by calling fail)
702 if it is ever executed. Variant ??? calls warn, and !!! calls die.>
704 Those would be nice to add to Perl 5. That could be done without new ops.
706 =head2 Virtualize operating system access
708 Implement a set of "vtables" that virtualizes operating system access
709 (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
710 least these interfaces should take SVs as "name" arguments instead of
711 bare char pointers; probably the most flexible and extensible way
712 would be for the Perl-facing interfaces to accept HVs. The system
713 needs to be per-operating-system and per-file-system
714 hookable/filterable, preferably both from XS and Perl level
715 (L<perlport/"Files and Filesystems"> is good reading at this point,
716 in fact, all of L<perlport> is.)
718 This has actually already been implemented (but only for Win32),
719 take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
720 variants go through a set of "vtables" for operating system access,
721 non-Win32 systems currently go straight for the POSIX/UNIX-style
722 system/library call. Similar system as for Win32 should be
723 implemented for all platforms. The existing Win32 implementation
724 probably does not need to survive alongside this proposed new
725 implementation, the approaches could be merged.
727 What would this give us? One often-asked-for feature this would
728 enable is using Unicode for filenames, and other "names" like %ENV,
729 usernames, hostnames, and so forth.
730 (See L<perlunicode/"When Unicode Does Not Happen">.)
732 But this kind of virtualization would also allow for things like
733 virtual filesystems, virtual networks, and "sandboxes" (though as long
734 as dynamic loading of random object code is allowed, not very safe
735 sandboxes since external code of course know not of Perl's vtables).
736 An example of a smaller "sandbox" is that this feature can be used to
737 implement per-thread working directories: Win32 already does this.
739 See also L</"Extend PerlIO and PerlIO::Scalar">.
743 Tasks that will get your name mentioned in the description of the "Highlights
746 =head2 make ithreads more robust
748 Generally make ithreads more robust. See also L</iCOW>
750 This task is incremental - even a little bit of work on it will help, and
751 will be greatly appreciated.
753 One bit would be to write the missing code in sv.c:Perl_dirp_dup.
755 Fix Perl_sv_dup, et al so that threads can return objects.
759 Sarathy and Arthur have a proposal for an improved Copy On Write which
760 specifically will be able to COW new ithreads. If this can be implemented
761 it would be a good thing.
763 =head2 (?{...}) closures in regexps
765 Fix (or rewrite) the implementation of the C</(?{...})/> closures.
767 =head2 A re-entrant regexp engine
769 This will allow the use of a regex from inside (?{ }), (??{ }) and
770 (?(?{ })|) constructs.
772 =head2 Add class set operations to regexp engine
774 Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
776 demerphq has this on his todo list, but right at the bottom.