3 perltodo - Perl TO-DO List
7 This is a list of wishes for Perl. The tasks we think are smaller or
8 easier are listed first. Anyone is welcome to work on any of these,
9 but it's a good idea to first contact I<perl5-porters@perl.org> to
10 avoid duplication of effort, and to learn from any previous attempts.
11 By all means contact a pumpking privately first if you prefer.
13 Whilst patches to make the list shorter are most welcome, ideas to add to
14 the list are also encouraged. Check the perl5-porters archives for past
15 ideas, and any discussion about them. One set of archives may be found at:
17 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
19 What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
20 not, but if your patch is incorporated, then we'll add your name to the
21 F<AUTHORS> file, which ships in the official distribution. How many other
22 programming languages offer you 1 line of immortality?
24 =head1 Tasks that only need Perl knowledge
26 =head2 Remove duplication of test setup.
28 Schwern notes, that there's duplication of code - lots and lots of tests have
29 some variation on the big block of C<$Is_Foo> checks. We can safely put this
30 into a file, change it to build an C<%Is> hash and require it. Maybe just put
31 it into F<test.pl>. Throw in the handy tainting subroutines.
33 =head2 merge common code in installperl and installman
35 There are some common subroutines and a common C<BEGIN> block in F<installperl>
36 and F<installman>. These should probably be merged. It would also be good to
37 check for duplication in all the utility scripts supplied in the source
38 tarball. It might be good to move them all to a subdirectory, but this would
39 require careful checking to find all places that call them, and change those
42 =head2 common test code for timed bail out
44 Write portable self destruct code for tests to stop them burning CPU in
45 infinite loops. This needs to avoid using alarm, as some of the tests are
46 testing alarm/sleep or timers.
48 =head2 POD -E<gt> HTML conversion in the core still sucks
50 Which is crazy given just how simple POD purports to be, and how simple HTML
51 can be. It's not actually I<as> simple as it sounds, particularly with the
52 flexibility POD allows for C<=item>, but it would be good to improve the
53 visual appeal of the HTML generated, and to avoid it having any validation
54 errors. See also L</make HTML install work>, as the layout of installation tree
55 is needed to improve the cross-linking.
57 The addition of C<Pod::Simple> and its related modules may make this task
60 =head2 merge checkpods and podchecker
62 F<pod/checkpods.PL> (and C<make check> in the F<pod/> subdirectory)
63 implements a very basic check for pod files, but the errors it discovers
64 aren't found by podchecker. Add this check to podchecker, get rid of
65 checkpods and have C<make check> use podchecker.
67 =head2 perlmodlib.PL rewrite
69 Currently perlmodlib.PL needs to be run from a source directory where perl
70 has been built, or some modules won't be found, and others will be
71 skipped. Make it run from a clean perl source tree (so it's reproducible).
73 =head2 Parallel testing
75 (This probably impacts much more than the core: also the Test::Harness
76 and TAP::* modules on CPAN.)
78 The core regression test suite is getting ever more comprehensive, which has
79 the side effect that it takes longer to run. This isn't so good. Investigate
80 whether it would be feasible to give the harness script the B<option> of
81 running sets of tests in parallel. This would be useful for tests in
82 F<t/op/*.t> and F<t/uni/*.t> and maybe some sets of tests in F<lib/>.
90 How does screen layout work when you're running more than one test?
94 How does the caller of test specify how many tests to run in parallel?
98 How do setup/teardown tests identify themselves?
102 Pugs already does parallel testing - can their approach be re-used?
104 =head2 Make Schwern poorer
106 We should have tests for everything. When all the core's modules are tested,
107 Schwern has promised to donate to $500 to TPF. We may need volunteers to
108 hold him upside down and shake vigorously in order to actually extract the
111 =head2 Improve the coverage of the core tests
113 Use Devel::Cover to ascertain the core modules's test coverage, then add
114 tests that are currently missing.
118 A full test suite for the B module would be nice.
120 =head2 Deparse inlined constants
124 use constant PI => 4;
127 will currently deparse as
129 use constant ('PI', 4);
132 because the tokenizer inlines the value of the constant subroutine C<PI>.
133 This allows various compile time optimisations, such as constant folding
134 and dead code elimination. Where these haven't happened (such as the example
135 above) it ought be possible to make B::Deparse work out the name of the
136 original constant, because just enough information survives in the symbol
137 table to do this. Specifically, the same scalar is used for the constant in
138 the optree as is used for the constant subroutine, so by iterating over all
139 symbol tables and generating a mapping of SV address to constant name, it
140 would be possible to provide B::Deparse with this functionality.
142 =head2 A decent benchmark
144 C<perlbench> seems impervious to any recent changes made to the perl core. It
145 would be useful to have a reasonable general benchmarking suite that roughly
146 represented what current perl programs do, and measurably reported whether
147 tweaks to the core improve, degrade or don't really affect performance, to
148 guide people attempting to optimise the guts of perl. Gisle would welcome
149 new tests for perlbench.
151 =head2 fix tainting bugs
153 Fix the bugs revealed by running the test suite with the C<-t> switch (via
154 C<make test.taintwarn>).
156 =head2 Dual life everything
158 As part of the "dists" plan, anything that doesn't belong in the smallest perl
159 distribution needs to be dual lifed. Anything else can be too. Figure out what
160 changes would be needed to package that module and its tests up for CPAN, and
161 do so. Test it with older perl releases, and fix the problems you find.
163 To make a minimal perl distribution, it's useful to look at
164 F<t/lib/commonsense.t>.
166 =head2 Improving C<threads::shared>
168 Investigate whether C<threads::shared> could share aggregates properly with
169 only Perl level changes to shared.pm
171 =head2 POSIX memory footprint
173 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
174 various times worked to cut it down. There is probably still fat to cut out -
175 for example POSIX passes Exporter some very memory hungry data structures.
177 =head2 embed.pl/makedef.pl
179 There is a script F<embed.pl> that generates several header files to prefix
180 all of Perl's symbols in a consistent way, to provide some semblance of
181 namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
182 in F<interpvar.h>. Quite a few of the functions and variables
183 are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
184 doesn't understand the C macros, so the rules about which symbols are present
185 when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
186 It would be good to teach C<embed.pl> to understand the conditional
187 compilation, and hence remove the duplication, and the mistakes it has caused.
189 =head2 use strict; and AutoLoad
191 Currently if you write
194 use AutoLoader 'AUTOLOAD';
199 print join (' ', No, strict, here), "!\n";
202 then C<use strict;> isn't in force within the autoloaded subroutines. It would
203 be more consistent (and less surprising) to arrange for all lexical pragmas
204 in force at the __END__ block to be in force within each autoloaded subroutine.
206 There's a similar problem with SelfLoader.
208 =head1 Tasks that need a little sysadmin-type knowledge
210 Or if you prefer, tasks that you would learn from, and broaden your skills
213 =head2 make HTML install work
215 There is an C<installhtml> target in the Makefile. It's marked as
216 "experimental". It would be good to get this tested, make it work reliably, and
217 remove the "experimental" tag. This would include
223 Checking that cross linking between various parts of the documentation works.
224 In particular that links work between the modules (files with POD in F<lib/>)
225 and the core documentation (files in F<pod/>)
229 Work out how to split C<perlfunc> into chunks, preferably one per function
230 group, preferably with general case code that could be used elsewhere.
231 Challenges here are correctly identifying the groups of functions that go
232 together, and making the right named external cross-links point to the right
233 page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
234 C<endservent>, two or more C<=items> giving the different parameter lists, such
237 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
238 =item substr EXPR,OFFSET,LENGTH
239 =item substr EXPR,OFFSET
241 and different parameter lists having different meanings. (eg C<select>)
245 =head2 compressed man pages
247 Be able to install them. This would probably need a configure test to see how
248 the system does compressed man pages (same directory/different directory?
249 same filename/different filename), as well as tweaking the F<installman> script
250 to compress as necessary.
252 =head2 Add a code coverage target to the Makefile
254 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
255 to do this manually are roughly
261 do a normal C<Configure>, but include Devel::Cover as a module to install
262 (see F<INSTALL> for how to do this)
270 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
274 Process the resulting Devel::Cover database
278 This just give you the coverage of the F<.pm>s. To also get the C level
285 Additionally tell C<Configure> to use the appropriate C compiler flags for
292 (instead of C<make perl>)
296 After running the tests run C<gcov> to generate all the F<.gcov> files.
297 (Including down in the subdirectories of F<ext/>
301 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
302 to get their stats into the cover_db directory.
306 Then process the Devel::Cover database
310 It would be good to add a single switch to C<Configure> to specify that you
311 wanted to perform perl level coverage, and another to specify C level
312 coverage, and have C<Configure> and the F<Makefile> do all the right things
315 =head2 Make Config.pm cope with differences between built and installed perl
317 Quite often vendors ship a perl binary compiled with their (pay-for)
318 compilers. People install a free compiler, such as gcc. To work out how to
319 build extensions, Perl interrogates C<%Config>, so in this situation
320 C<%Config> describes compilers that aren't there, and extension building
321 fails. This forces people into choosing between re-compiling perl themselves
322 using the compiler they have, or only using modules that the vendor ships.
324 It would be good to find a way teach C<Config.pm> about the installation setup,
325 possibly involving probing at install time or later, so that the C<%Config> in
326 a binary distribution better describes the installed machine, when the
327 installed machine differs from the build machine in some significant way.
329 =head2 linker specification files
331 Some platforms mandate that you provide a list of a shared library's external
332 symbols to the linker, so the core already has the infrastructure in place to
333 do this for generating shared perl libraries. My understanding is that the
334 GNU toolchain can accept an optional linker specification file, and restrict
335 visibility just to symbols declared in that file. It would be good to extend
336 F<makedef.pl> to support this format, and to provide a means within
337 C<Configure> to enable it. This would allow Unix users to test that the
338 export list is correct, and to build a perl that does not pollute the global
339 namespace with private symbols.
341 =head2 Cross-compile support
343 Currently C<Configure> understands C<-Dusecrosscompile> option. This option
344 arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
345 assumed then to be copied to TARGET machine and used as a replacement of full
348 This could be done little differently. Namely C<miniperl> should be built for
349 HOST and then full C<perl> with extensions should be compiled for TARGET.
350 This, however, might require extra trickery for %Config: we have one config
351 first for HOST and then another for TARGET. Tools like MakeMaker will be
352 mightily confused. Having around two different types of executables and
353 libraries (HOST and TARGET) makes life interesting for Makefiles and
354 shell (and Perl) scripts. There is $Config{run}, normally empty, which
355 can be used as an execution wrapper. Also note that in some
356 cross-compilation/execution environments the HOST and the TARGET do
357 not see the same filesystem(s), the $Config{run} may need to do some
358 file/directory copying back and forth.
362 Make F<pod/roffitall> be updated by F<pod/buildtoc>.
364 =head1 Tasks that need a little C knowledge
366 These tasks would need a little C knowledge, but don't need any specific
367 background or experience with XS, or how the Perl interpreter works
369 =head2 Weed out needless PERL_UNUSED_ARG
371 The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about
372 unused arguments. Often the arguments can't be removed, as there is an
373 external constraint that determines the prototype of the function, so this
374 approach is valid. However, there are some cases where C<PERL_UNUSED_ARG>
375 could be removed. Specifically
381 The prototypes of (nearly all) static functions can be changed
385 Unused arguments generated by short cut macros are wasteful - the short cut
386 macro used can be changed.
390 =head2 Modernize the order of directories in @INC
392 The way @INC is laid out by default, one cannot upgrade core (dual-life)
393 modules without overwriting files. This causes problems for binary
394 package builders. One possible proposal is laid out in this
396 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>.
400 Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
401 On these systems, it might be the default compilation mode, and there
402 is currently no guarantee that passing no use64bitall option to the
403 Configure process will build a 32bit perl. Implementing -Duse32bit*
404 options would be nice for perl 5.12.
406 =head2 Make it clear from -v if this is the exact official release
408 Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that
409 usually defines one local patch, of the form "MAINT12345" or "RC1". The output
410 of perl -v doesn't report that a perl isn't an official release, and this
411 information can get lost in bugs reports. Because of this, the minor version
412 isn't bumped up until RC time, to minimise the possibility of versions of perl
413 escaping that believe themselves to be newer than they actually are.
415 It would be useful to find an elegant way to have the "this is an interim
416 maintenance release" or "this is a release candidate" in the terse -v output,
417 and have it so that it's easy for the pumpking to remove this just as the
418 release tarball is rolled up. This way the version pulled out of rsync would
419 always say "I'm a development release" and it would be safe to bump the
420 reported minor version as soon as a release ships, which would aid perl
423 This task is really about thinking of an elegant way to arrange the C source
424 such that it's trivial for the Pumpking to flag "this is an official release"
425 when making a tarball, yet leave the default source saying "I'm not the
428 =head2 Profile Perl - am I hot or not?
430 The Perl source code is stable enough that it makes sense to profile it,
431 identify and optimise the hotspots. It would be good to measure the
432 performance of the Perl interpreter using free tools such as cachegrind,
433 gprof, and dtrace, and work to reduce the bottlenecks they reveal.
435 As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
436 the ops that are most commonly used. The idea is that by grouping them, their
437 object code will be adjacent in the executable, so they have a greater chance
438 of already being in the CPU cache (or swapped in) due to being near another op
441 Except that it's not clear if these really are the most commonly used ops. So
442 as part of exercising your skills with coverage and profiling tools you might
443 want to determine what ops I<really> are the most commonly used. And in turn
444 suggest evictions and promotions to achieve a better F<pp_hot.c>.
446 =head2 Allocate OPs from arenas
448 Currently all new OP structures are individually malloc()ed and free()d.
449 All C<malloc> implementations have space overheads, and are now as fast as
450 custom allocates so it would both use less memory and less CPU to allocate
451 the various OP structures from arenas. The SV arena code can probably be
454 Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use
455 Perl_Slab_alloc() to pack optrees into a contiguous block, which is
456 probably superior to the use of OP arenas, esp. from a cache locality
457 standpoint. See L<Profile Perl - am I hot or not?>.
459 =head2 Improve win32/wince.c
461 Currently, numerous functions look virtually, if not completely,
462 identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't
465 =head2 Use secure CRT functions when building with VC8 on Win32
467 Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
468 that they were "unsafe" and introduced differently named secure versions of
469 them as replacements, e.g. instead of writing
471 FILE* f = fopen(__FILE__, "r");
476 errno_t err = fopen_s(&f, __FILE__, "r");
478 Currently, the warnings about these deprecations have been disabled by adding
479 -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
480 warning suppressant and actually make use of the new secure CRT functions.
482 There is also a similar issue with POSIX CRT function names like fileno having
483 been deprecated in favour of ISO C++ conformant names like _fileno. These
484 warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
485 might be nice to do as Microsoft suggest here too, although, unlike the secure
486 functions issue, there is presumably little or no benefit in this case.
488 =head2 Fix POSIX::access() and chdir() on Win32
490 These functions currently take no account of DACLs and therefore do not behave
491 correctly in situations where access is restricted by DACLs (as opposed to the
492 read-only attribute).
494 Furthermore, POSIX::access() behaves differently for directories having the
495 read-only attribute set depending on what CRT library is being used. For
496 example, the _access() function in the VC6 and VC7 CRTs (wrongly) claim that
497 such directories are not writable, whereas in fact all directories are writable
498 unless access is denied by DACLs. (In the case of directories, the read-only
499 attribute actually only means that the directory cannot be deleted.) This CRT
500 bug is fixed in the VC8 and VC9 CRTs (but, of course, the directory may still
501 not actually be writable if access is indeed denied by DACLs).
503 For the chdir() issue, see ActiveState bug #74552:
504 http://bugs.activestate.com/show_bug.cgi?id=74552
506 Therefore, DACLs should be checked both for consistency across CRTs and for
509 (Note that perl's -w operator should not be modified to check DACLs. It has
510 been written so that it reflects the state of the read-only attribute, even
511 for directories (whatever CRT is being used), for symmetry with chmod().)
513 =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
515 Maybe create a utility that checks after each libperl.a creation that
516 none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
517 ever creep back to libperl.a.
519 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
521 Note, of course, that this will only tell whether B<your> platform
522 is using those naughty interfaces.
524 =head2 -D_FORTIFY_SOURCE=2, -fstack-protector
526 Recent glibcs support C<-D_FORTIFY_SOURCE=2> and recent gcc
527 (4.1 onwards?) supports C<-fstack-protector>, both of which give
528 protection against various kinds of buffer overflow problems.
529 These should probably be used for compiling Perl whenever available,
530 Configure and/or hints files should be adjusted to probe for the
531 availability of these features and enable them as appropriate.
533 =head2 Arenas for GPs? For MAGIC?
535 C<struct gp> and C<struct magic> are both currently allocated by C<malloc>.
536 It might be a speed or memory saving to change to using arenas. Or it might
537 not. It would need some suitable benchmarking first. In particular, C<GP>s
538 can probably be changed with minimal compatibility impact (probably nothing
539 outside of the core, or even outside of F<gv.c> allocates them), but they
540 probably aren't allocated/deallocated often enough for a speed saving. Whereas
541 C<MAGIC> is allocated/deallocated more often, but in turn, is also something
542 more externally visible, so changing the rules here may bite external code.
545 =head1 Tasks that need a knowledge of XS
547 These tasks would need C knowledge, and roughly the level of knowledge of
548 the perl API that comes from writing modules that use XS to interface to
551 =head2 autovivification
553 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
555 This task is incremental - even a little bit of work on it will help.
557 =head2 Unicode in Filenames
559 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
560 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
561 system, truncate, unlink, utime, -X. All these could potentially accept
562 Unicode filenames either as input or output (and in the case of system
563 and qx Unicode in general, as input or output to/from the shell).
564 Whether a filesystem - an operating system pair understands Unicode in
567 Known combinations that have some level of understanding include
568 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
569 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
570 create Unicode filenames, what forms of Unicode are accepted and used
571 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
572 and so on, varies. Finding the right level of interfacing to Perl
573 requires some thought. Remember that an OS does not implicate a
576 (The Windows -C command flag "wide API support" has been at least
577 temporarily retired in 5.8.1, and the -C has been repurposed, see
580 Most probably the right way to do this would be this:
581 L</"Virtualize operating system access">.
583 =head2 Unicode in %ENV
585 Currently the %ENV entries are always byte strings.
586 See L</"Virtualize operating system access">.
588 =head2 Unicode and glob()
590 Currently glob patterns and filenames returned from File::Glob::glob()
591 are always byte strings. See L</"Virtualize operating system access">.
593 =head2 Unicode and lc/uc operators
595 Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on
596 what the internal encoding of their argument is. That should not be the
597 case. Maybe add a pragma to switch behaviour.
599 =head2 use less 'memory'
601 Investigate trade offs to switch out perl's choices on memory usage.
602 Particularly perl should be able to give memory back.
604 This task is incremental - even a little bit of work on it will help.
606 =head2 Re-implement C<:unique> in a way that is actually thread-safe
608 The old implementation made bad assumptions on several levels. A good 90%
609 solution might be just to make C<:unique> work to share the string buffer
610 of SvPVs. That way large constant strings can be shared between ithreads,
611 such as the configuration information in F<Config>.
613 =head2 Make tainting consistent
615 Tainting would be easier to use if it didn't take documented shortcuts and
616 allow taint to "leak" everywhere within an expression.
618 =head2 readpipe(LIST)
620 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
621 running a shell. readpipe() (the function behind qx//) could be similarly
624 =head2 Audit the code for destruction ordering assumptions
628 /* Need to check SvMAGICAL, as during global destruction it may be that
629 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
630 is now part of the linked list of SV heads, rather than pointing to
631 the original body. */
632 /* FIXME - audit the code for other bugs like this one. */
634 adding the C<SvMAGICAL> check to
636 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
637 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
639 Go through the core and look for similar assumptions that SVs have particular
640 types, as all bets are off during global destruction.
642 =head2 Extend PerlIO and PerlIO::Scalar
644 PerlIO::Scalar doesn't know how to truncate(). Implementing this
645 would require extending the PerlIO vtable.
647 Similarly the PerlIO vtable doesn't know about formats (write()), or
648 about stat(), or chmod()/chown(), utime(), or flock().
650 (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
653 PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
654 opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
657 See also L</"Virtualize operating system access">.
659 =head2 -C on the #! line
661 It should be possible to make -C work correctly if found on the #! line,
662 given that all perl command line options are strict ASCII, and -C changes
663 only the interpretation of non-ASCII characters, and not for the script file
664 handle. To make it work needs some investigation of the ordering of function
665 calls during startup, and (by implication) a bit of tweaking of that order.
667 =head2 Propagate const outwards from Perl_moreswitches()
669 Change 32057 changed the parameter and return value of C<Perl_moreswitches()>
670 from <char *> to <const char *>. It should now be possible to propagate
671 const-correctness outwards to C<S_parse_body()>, C<Perl_moreswitches()>
674 =head2 Duplicate logic in S_method_common() and Perl_gv_fetchmethod_autoload()
676 A comment in C<S_method_common> notes
678 /* This code tries to figure out just what went wrong with
679 gv_fetchmethod. It therefore needs to duplicate a lot of
680 the internals of that function. We can't move it inside
681 Perl_gv_fetchmethod_autoload(), however, since that would
682 cause UNIVERSAL->can("NoSuchPackage::foo") to croak, and we
686 If C<Perl_gv_fetchmethod_autoload> gets rewritten to take (more) flag bits,
687 then it ought to be possible to move the logic from C<S_method_common> to
688 the "right" place. When making this change it would probably be good to also
689 pass in at least the method name length, if not also pre-computed hash values
690 when known. (I'm contemplating a plan to pre-compute hash values for common
691 fixed strings such as C<ISA> and pass them in to functions.)
693 =head2 Organize error messages
695 Perl's diagnostics (error messages, see L<perldiag>) could use
696 reorganizing and formalizing so that each error message has its
697 stable-for-all-eternity unique id, categorized by severity, type, and
698 subsystem. (The error messages would be listed in a datafile outside
699 of the Perl source code, and the source code would only refer to the
700 messages by the id.) This clean-up and regularizing should apply
701 for all croak() messages.
703 This would enable all sorts of things: easier translation/localization
704 of the messages (though please do keep in mind the caveats of
705 L<Locale::Maketext> about too straightforward approaches to
706 translation), filtering by severity, and instead of grepping for a
707 particular error message one could look for a stable error id. (Of
708 course, changing the error messages by default would break all the
709 existing software depending on some particular error message...)
711 This kind of functionality is known as I<message catalogs>. Look for
712 inspiration for example in the catgets() system, possibly even use it
713 if available-- but B<only> if available, all platforms will B<not>
716 For the really pure at heart, consider extending this item to cover
717 also the warning messages (see L<perllexwarn>, C<warnings.pl>).
719 =head1 Tasks that need a knowledge of the interpreter
721 These tasks would need C knowledge, and knowledge of how the interpreter works,
722 or a willingness to learn.
724 =head2 lexicals used only once
728 $ perl -we '$pie = 42'
729 Name "main::pie" used only once: possible typo at -e line 1.
733 $ perl -we 'my $pie = 42'
735 Logically all lexicals used only once should warn, if the user asks for
736 warnings. An unworked RT ticket (#5087) has been open for almost seven
737 years for this discrepancy.
741 The handling of Unicode is unclean in many places. For example, the regexp
742 engine matches in Unicode semantics whenever the string or the pattern is
743 flagged as UTF-8, but that should not be dependent on an internal storage
744 detail of the string. Likewise, case folding behaviour is dependent on the
745 UTF8 internal flag being on or off.
747 =head2 Properly Unicode safe tokeniser and pads.
749 The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
750 variable names are stored in stashes as raw bytes, without the utf-8 flag
751 set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
752 tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
753 source filters. All this could be fixed.
755 =head2 state variable initialization in list context
757 Currently this is illegal:
759 state ($a, $b) = foo();
761 In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different
762 semantics, which is tricky to implement in Perl 5 as currently they produce
763 the same opcode trees. The Perl 6 design is firm, so it would be good to
764 implement the necessary code in Perl 5. There are comments in
765 C<Perl_newASSIGNOP()> that show the code paths taken by various assignment
766 constructions involving state variables.
768 =head2 Implement $value ~~ 0 .. $range
770 It would be nice to extend the syntax of the C<~~> operator to also
771 understand numeric (and maybe alphanumeric) ranges.
773 =head2 A does() built-in
775 Like ref(), only useful. It would call the C<DOES> method on objects; it
776 would also tell whether something can be dereferenced as an
777 array/hash/etc., or used as a regexp, etc.
778 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
780 =head2 Tied filehandles and write() don't mix
782 There is no method on tied filehandles to allow them to be called back by
785 =head2 Attach/detach debugger from running program
787 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
788 program if you pass the process ID. It would be good to do this with the Perl
789 debugger on a running Perl program, although I'm not sure how it would be
790 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
792 =head2 Optimize away empty destructors
794 Defining an empty DESTROY method might be useful (notably in
795 AUTOLOAD-enabled classes), but it's still a bit expensive to call. That
796 could probably be optimized.
798 =head2 LVALUE functions for lists
800 The old perltodo notes that lvalue functions don't work for list or hash
801 slices. This would be good to fix.
803 =head2 LVALUE functions in the debugger
805 The old perltodo notes that lvalue functions don't work in the debugger. This
806 would be good to fix.
808 =head2 regexp optimiser optional
810 The regexp optimiser is not optional. It should configurable to be, to allow
811 its performance to be measured, and its bugs to be easily demonstrated.
813 =head2 delete &function
815 Allow to delete functions. One can already undef them, but they're still
818 =head2 C</w> regex modifier
820 That flag would enable to match whole words, and also to interpolate
821 arrays as alternations. With it, C</P/w> would be roughly equivalent to:
823 do { local $"='|'; /\b(?:P)\b/ }
825 See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
828 =head2 optional optimizer
830 Make the peephole optimizer optional. Currently it performs two tasks as
831 it walks the optree - genuine peephole optimisations, and necessary fixups of
832 ops. It would be good to find an efficient way to switch out the
833 optimisations whilst keeping the fixups.
835 =head2 You WANT *how* many
837 Currently contexts are void, scalar and list. split has a special mechanism in
838 place to pass in the number of return values wanted. It would be useful to
839 have a general mechanism for this, backwards compatible and little speed hit.
840 This would allow proposals such as short circuiting sort to be implemented
843 =head2 lexical aliases
845 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
847 =head2 entersub XS vs Perl
849 At the moment pp_entersub is huge, and has code to deal with entering both
850 perl and XS subroutines. Subroutine implementations rarely change between
851 perl and XS at run time, so investigate using 2 ops to enter subs (one for
852 XS, one for perl) and swap between if a sub is redefined.
856 Self-ties are currently illegal because they caused too many segfaults. Maybe
857 the causes of these could be tracked down and self-ties on all types
860 =head2 Optimize away @_
862 The old perltodo notes "Look at the "reification" code in C<av.c>".
864 =head2 The yada yada yada operators
866 Perl 6's Synopsis 3 says:
868 I<The ... operator is the "yada, yada, yada" list operator, which is used as
869 the body in function prototypes. It complains bitterly (by calling fail)
870 if it is ever executed. Variant ??? calls warn, and !!! calls die.>
872 Those would be nice to add to Perl 5. That could be done without new ops.
874 =head2 Virtualize operating system access
876 Implement a set of "vtables" that virtualizes operating system access
877 (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
878 least these interfaces should take SVs as "name" arguments instead of
879 bare char pointers; probably the most flexible and extensible way
880 would be for the Perl-facing interfaces to accept HVs. The system
881 needs to be per-operating-system and per-file-system
882 hookable/filterable, preferably both from XS and Perl level
883 (L<perlport/"Files and Filesystems"> is good reading at this point,
884 in fact, all of L<perlport> is.)
886 This has actually already been implemented (but only for Win32),
887 take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
888 variants go through a set of "vtables" for operating system access,
889 non-Win32 systems currently go straight for the POSIX/UNIX-style
890 system/library call. Similar system as for Win32 should be
891 implemented for all platforms. The existing Win32 implementation
892 probably does not need to survive alongside this proposed new
893 implementation, the approaches could be merged.
895 What would this give us? One often-asked-for feature this would
896 enable is using Unicode for filenames, and other "names" like %ENV,
897 usernames, hostnames, and so forth.
898 (See L<perlunicode/"When Unicode Does Not Happen">.)
900 But this kind of virtualization would also allow for things like
901 virtual filesystems, virtual networks, and "sandboxes" (though as long
902 as dynamic loading of random object code is allowed, not very safe
903 sandboxes since external code of course know not of Perl's vtables).
904 An example of a smaller "sandbox" is that this feature can be used to
905 implement per-thread working directories: Win32 already does this.
907 See also L</"Extend PerlIO and PerlIO::Scalar">.
909 =head2 Investigate PADTMP hash pessimisation
911 The peephole optimier converts constants used for hash key lookups to shared
912 hash key scalars. Under ithreads, something is undoing this work. See
913 See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html
915 =head2 repack the optree
917 Repacking the optree after execution order is determined could allow
918 removal of NULL ops, and optimal ordering of OPs wrt cache-line
919 filling. The slab allocator could be reused for this purpose.
921 http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html
923 =head2 optimize tail-calls
925 Tail-calls present an opportunity for broadly applicable optimization;
926 anywhere that C<< return foo(...) >> is called, the outer return can
927 be replaced by a goto, and foo will return directly to the outer
928 caller, saving (conservatively) 25% of perl's call&return cost, which
929 is relatively higher than in C. The scheme language is known to do
930 this heavily. B::Concise provides good insight into where this
931 optimization is possible, ie anywhere entersub,leavesub op-sequence
934 perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)'
936 Bottom line on this is probably a new pp_tailcall function which
937 combines the code in pp_entersub, pp_leavesub. This should probably
938 be done 1st in XS, and using B::Generate to patch the new OP into the
943 Tasks that will get your name mentioned in the description of the "Highlights
946 =head2 make ithreads more robust
948 Generally make ithreads more robust. See also L</iCOW>
950 This task is incremental - even a little bit of work on it will help, and
951 will be greatly appreciated.
953 One bit would be to write the missing code in sv.c:Perl_dirp_dup.
955 Fix Perl_sv_dup, et al so that threads can return objects.
959 Sarathy and Arthur have a proposal for an improved Copy On Write which
960 specifically will be able to COW new ithreads. If this can be implemented
961 it would be a good thing.
963 =head2 (?{...}) closures in regexps
965 Fix (or rewrite) the implementation of the C</(?{...})/> closures.
967 =head2 A re-entrant regexp engine
969 This will allow the use of a regex from inside (?{ }), (??{ }) and
970 (?(?{ })|) constructs.
972 =head2 Add class set operations to regexp engine
974 Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
976 demerphq has this on his todo list, but right at the bottom.