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 Parallel testing
69 (This probably impacts much more than the core: also the Test::Harness
70 and TAP::* modules on CPAN.)
72 The core regression test suite is getting ever more comprehensive, which has
73 the side effect that it takes longer to run. This isn't so good. Investigate
74 whether it would be feasible to give the harness script the B<option> of
75 running sets of tests in parallel. This would be useful for tests in
76 F<t/op/*.t> and F<t/uni/*.t> and maybe some sets of tests in F<lib/>.
84 How does screen layout work when you're running more than one test?
88 How does the caller of test specify how many tests to run in parallel?
92 How do setup/teardown tests identify themselves?
96 Pugs already does parallel testing - can their approach be re-used?
98 =head2 Make Schwern poorer
100 We should have tests for everything. When all the core's modules are tested,
101 Schwern has promised to donate to $500 to TPF. We may need volunteers to
102 hold him upside down and shake vigorously in order to actually extract the
105 =head2 Improve the coverage of the core tests
107 Use Devel::Cover to ascertain the core modules's test coverage, then add
108 tests that are currently missing.
112 A full test suite for the B module would be nice.
114 =head2 Deparse inlined constants
118 use constant PI => 4;
121 will currently deparse as
123 use constant ('PI', 4);
126 because the tokenizer inlines the value of the constant subroutine C<PI>.
127 This allows various compile time optimisations, such as constant folding
128 and dead code elimination. Where these haven't happened (such as the example
129 above) it ought be possible to make B::Deparse work out the name of the
130 original constant, because just enough information survives in the symbol
131 table to do this. Specifically, the same scalar is used for the constant in
132 the optree as is used for the constant subroutine, so by iterating over all
133 symbol tables and generating a mapping of SV address to constant name, it
134 would be possible to provide B::Deparse with this functionality.
136 =head2 A decent benchmark
138 C<perlbench> seems impervious to any recent changes made to the perl core. It
139 would be useful to have a reasonable general benchmarking suite that roughly
140 represented what current perl programs do, and measurably reported whether
141 tweaks to the core improve, degrade or don't really affect performance, to
142 guide people attempting to optimise the guts of perl. Gisle would welcome
143 new tests for perlbench.
145 =head2 fix tainting bugs
147 Fix the bugs revealed by running the test suite with the C<-t> switch (via
148 C<make test.taintwarn>).
150 =head2 Dual life everything
152 As part of the "dists" plan, anything that doesn't belong in the smallest perl
153 distribution needs to be dual lifed. Anything else can be too. Figure out what
154 changes would be needed to package that module and its tests up for CPAN, and
155 do so. Test it with older perl releases, and fix the problems you find.
157 To make a minimal perl distribution, it's useful to look at
158 F<t/lib/commonsense.t>.
160 =head2 Bundle dual life modules in ext/
162 For maintenance (and branch merging) reasons, it would be useful to move
163 some architecture-independent dual-life modules from lib/ to ext/, if this
164 has no negative impact on the build of perl itself.
166 However, we need to make sure that they are still installed in
167 architecture-independent directories by C<make install>.
169 =head2 Improving C<threads::shared>
171 Investigate whether C<threads::shared> could share aggregates properly with
172 only Perl level changes to shared.pm
174 =head2 POSIX memory footprint
176 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
177 various times worked to cut it down. There is probably still fat to cut out -
178 for example POSIX passes Exporter some very memory hungry data structures.
180 =head2 embed.pl/makedef.pl
182 There is a script F<embed.pl> that generates several header files to prefix
183 all of Perl's symbols in a consistent way, to provide some semblance of
184 namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
185 in F<interpvar.h>. Quite a few of the functions and variables
186 are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
187 doesn't understand the C macros, so the rules about which symbols are present
188 when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
189 It would be good to teach C<embed.pl> to understand the conditional
190 compilation, and hence remove the duplication, and the mistakes it has caused.
192 =head2 use strict; and AutoLoad
194 Currently if you write
197 use AutoLoader 'AUTOLOAD';
202 print join (' ', No, strict, here), "!\n";
205 then C<use strict;> isn't in force within the autoloaded subroutines. It would
206 be more consistent (and less surprising) to arrange for all lexical pragmas
207 in force at the __END__ block to be in force within each autoloaded subroutine.
209 There's a similar problem with SelfLoader.
211 =head1 Tasks that need a little sysadmin-type knowledge
213 Or if you prefer, tasks that you would learn from, and broaden your skills
216 =head2 make HTML install work
218 There is an C<installhtml> target in the Makefile. It's marked as
219 "experimental". It would be good to get this tested, make it work reliably, and
220 remove the "experimental" tag. This would include
226 Checking that cross linking between various parts of the documentation works.
227 In particular that links work between the modules (files with POD in F<lib/>)
228 and the core documentation (files in F<pod/>)
232 Work out how to split C<perlfunc> into chunks, preferably one per function
233 group, preferably with general case code that could be used elsewhere.
234 Challenges here are correctly identifying the groups of functions that go
235 together, and making the right named external cross-links point to the right
236 page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
237 C<endservent>, two or more C<=items> giving the different parameter lists, such
240 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
241 =item substr EXPR,OFFSET,LENGTH
242 =item substr EXPR,OFFSET
244 and different parameter lists having different meanings. (eg C<select>)
248 =head2 compressed man pages
250 Be able to install them. This would probably need a configure test to see how
251 the system does compressed man pages (same directory/different directory?
252 same filename/different filename), as well as tweaking the F<installman> script
253 to compress as necessary.
255 =head2 Add a code coverage target to the Makefile
257 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
258 to do this manually are roughly
264 do a normal C<Configure>, but include Devel::Cover as a module to install
265 (see F<INSTALL> for how to do this)
273 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
277 Process the resulting Devel::Cover database
281 This just give you the coverage of the F<.pm>s. To also get the C level
288 Additionally tell C<Configure> to use the appropriate C compiler flags for
295 (instead of C<make perl>)
299 After running the tests run C<gcov> to generate all the F<.gcov> files.
300 (Including down in the subdirectories of F<ext/>
304 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
305 to get their stats into the cover_db directory.
309 Then process the Devel::Cover database
313 It would be good to add a single switch to C<Configure> to specify that you
314 wanted to perform perl level coverage, and another to specify C level
315 coverage, and have C<Configure> and the F<Makefile> do all the right things
318 =head2 Make Config.pm cope with differences between built and installed perl
320 Quite often vendors ship a perl binary compiled with their (pay-for)
321 compilers. People install a free compiler, such as gcc. To work out how to
322 build extensions, Perl interrogates C<%Config>, so in this situation
323 C<%Config> describes compilers that aren't there, and extension building
324 fails. This forces people into choosing between re-compiling perl themselves
325 using the compiler they have, or only using modules that the vendor ships.
327 It would be good to find a way teach C<Config.pm> about the installation setup,
328 possibly involving probing at install time or later, so that the C<%Config> in
329 a binary distribution better describes the installed machine, when the
330 installed machine differs from the build machine in some significant way.
332 =head2 linker specification files
334 Some platforms mandate that you provide a list of a shared library's external
335 symbols to the linker, so the core already has the infrastructure in place to
336 do this for generating shared perl libraries. My understanding is that the
337 GNU toolchain can accept an optional linker specification file, and restrict
338 visibility just to symbols declared in that file. It would be good to extend
339 F<makedef.pl> to support this format, and to provide a means within
340 C<Configure> to enable it. This would allow Unix users to test that the
341 export list is correct, and to build a perl that does not pollute the global
342 namespace with private symbols.
344 =head2 Cross-compile support
346 Currently C<Configure> understands C<-Dusecrosscompile> option. This option
347 arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
348 assumed then to be copied to TARGET machine and used as a replacement of full
351 This could be done little differently. Namely C<miniperl> should be built for
352 HOST and then full C<perl> with extensions should be compiled for TARGET.
353 This, however, might require extra trickery for %Config: we have one config
354 first for HOST and then another for TARGET. Tools like MakeMaker will be
355 mightily confused. Having around two different types of executables and
356 libraries (HOST and TARGET) makes life interesting for Makefiles and
357 shell (and Perl) scripts. There is $Config{run}, normally empty, which
358 can be used as an execution wrapper. Also note that in some
359 cross-compilation/execution environments the HOST and the TARGET do
360 not see the same filesystem(s), the $Config{run} may need to do some
361 file/directory copying back and forth.
365 Make F<pod/roffitall> be updated by F<pod/buildtoc>.
367 =head1 Tasks that need a little C knowledge
369 These tasks would need a little C knowledge, but don't need any specific
370 background or experience with XS, or how the Perl interpreter works
372 =head2 Weed out needless PERL_UNUSED_ARG
374 The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about
375 unused arguments. Often the arguments can't be removed, as there is an
376 external constraint that determines the prototype of the function, so this
377 approach is valid. However, there are some cases where C<PERL_UNUSED_ARG>
378 could be removed. Specifically
384 The prototypes of (nearly all) static functions can be changed
388 Unused arguments generated by short cut macros are wasteful - the short cut
389 macro used can be changed.
393 =head2 Modernize the order of directories in @INC
395 The way @INC is laid out by default, one cannot upgrade core (dual-life)
396 modules without overwriting files. This causes problems for binary
397 package builders. One possible proposal is laid out in this
399 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>.
403 Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
404 On these systems, it might be the default compilation mode, and there
405 is currently no guarantee that passing no use64bitall option to the
406 Configure process will build a 32bit perl. Implementing -Duse32bit*
407 options would be nice for perl 5.12.
409 =head2 Make it clear from -v if this is the exact official release
411 Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that
412 usually defines one local patch, of the form "MAINT12345" or "RC1". The output
413 of perl -v doesn't report that a perl isn't an official release, and this
414 information can get lost in bugs reports. Because of this, the minor version
415 isn't bumped up until RC time, to minimise the possibility of versions of perl
416 escaping that believe themselves to be newer than they actually are.
418 It would be useful to find an elegant way to have the "this is an interim
419 maintenance release" or "this is a release candidate" in the terse -v output,
420 and have it so that it's easy for the pumpking to remove this just as the
421 release tarball is rolled up. This way the version pulled out of rsync would
422 always say "I'm a development release" and it would be safe to bump the
423 reported minor version as soon as a release ships, which would aid perl
426 This task is really about thinking of an elegant way to arrange the C source
427 such that it's trivial for the Pumpking to flag "this is an official release"
428 when making a tarball, yet leave the default source saying "I'm not the
431 =head2 Profile Perl - am I hot or not?
433 The Perl source code is stable enough that it makes sense to profile it,
434 identify and optimise the hotspots. It would be good to measure the
435 performance of the Perl interpreter using free tools such as cachegrind,
436 gprof, and dtrace, and work to reduce the bottlenecks they reveal.
438 As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
439 the ops that are most commonly used. The idea is that by grouping them, their
440 object code will be adjacent in the executable, so they have a greater chance
441 of already being in the CPU cache (or swapped in) due to being near another op
444 Except that it's not clear if these really are the most commonly used ops. So
445 as part of exercising your skills with coverage and profiling tools you might
446 want to determine what ops I<really> are the most commonly used. And in turn
447 suggest evictions and promotions to achieve a better F<pp_hot.c>.
449 =head2 Allocate OPs from arenas
451 Currently all new OP structures are individually malloc()ed and free()d.
452 All C<malloc> implementations have space overheads, and are now as fast as
453 custom allocates so it would both use less memory and less CPU to allocate
454 the various OP structures from arenas. The SV arena code can probably be
457 Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use
458 Perl_Slab_alloc() to pack optrees into a contiguous block, which is
459 probably superior to the use of OP arenas, esp. from a cache locality
460 standpoint. See L<Profile Perl - am I hot or not?>.
462 =head2 Improve win32/wince.c
464 Currently, numerous functions look virtually, if not completely,
465 identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't
468 =head2 Use secure CRT functions when building with VC8 on Win32
470 Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
471 that they were "unsafe" and introduced differently named secure versions of
472 them as replacements, e.g. instead of writing
474 FILE* f = fopen(__FILE__, "r");
479 errno_t err = fopen_s(&f, __FILE__, "r");
481 Currently, the warnings about these deprecations have been disabled by adding
482 -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
483 warning suppressant and actually make use of the new secure CRT functions.
485 There is also a similar issue with POSIX CRT function names like fileno having
486 been deprecated in favour of ISO C++ conformant names like _fileno. These
487 warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
488 might be nice to do as Microsoft suggest here too, although, unlike the secure
489 functions issue, there is presumably little or no benefit in this case.
491 =head2 Fix POSIX::access() and chdir() on Win32
493 These functions currently take no account of DACLs and therefore do not behave
494 correctly in situations where access is restricted by DACLs (as opposed to the
495 read-only attribute).
497 Furthermore, POSIX::access() behaves differently for directories having the
498 read-only attribute set depending on what CRT library is being used. For
499 example, the _access() function in the VC6 and VC7 CRTs (wrongly) claim that
500 such directories are not writable, whereas in fact all directories are writable
501 unless access is denied by DACLs. (In the case of directories, the read-only
502 attribute actually only means that the directory cannot be deleted.) This CRT
503 bug is fixed in the VC8 and VC9 CRTs (but, of course, the directory may still
504 not actually be writable if access is indeed denied by DACLs).
506 For the chdir() issue, see ActiveState bug #74552:
507 http://bugs.activestate.com/show_bug.cgi?id=74552
509 Therefore, DACLs should be checked both for consistency across CRTs and for
512 (Note that perl's -w operator should not be modified to check DACLs. It has
513 been written so that it reflects the state of the read-only attribute, even
514 for directories (whatever CRT is being used), for symmetry with chmod().)
516 =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
518 Maybe create a utility that checks after each libperl.a creation that
519 none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
520 ever creep back to libperl.a.
522 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
524 Note, of course, that this will only tell whether B<your> platform
525 is using those naughty interfaces.
527 =head2 -D_FORTIFY_SOURCE=2, -fstack-protector
529 Recent glibcs support C<-D_FORTIFY_SOURCE=2> and recent gcc
530 (4.1 onwards?) supports C<-fstack-protector>, both of which give
531 protection against various kinds of buffer overflow problems.
532 These should probably be used for compiling Perl whenever available,
533 Configure and/or hints files should be adjusted to probe for the
534 availability of these features and enable them as appropriate.
536 =head2 Arenas for GPs? For MAGIC?
538 C<struct gp> and C<struct magic> are both currently allocated by C<malloc>.
539 It might be a speed or memory saving to change to using arenas. Or it might
540 not. It would need some suitable benchmarking first. In particular, C<GP>s
541 can probably be changed with minimal compatibility impact (probably nothing
542 outside of the core, or even outside of F<gv.c> allocates them), but they
543 probably aren't allocated/deallocated often enough for a speed saving. Whereas
544 C<MAGIC> is allocated/deallocated more often, but in turn, is also something
545 more externally visible, so changing the rules here may bite external code.
548 =head1 Tasks that need a knowledge of XS
550 These tasks would need C knowledge, and roughly the level of knowledge of
551 the perl API that comes from writing modules that use XS to interface to
554 =head2 investigate removing int_macro_int from POSIX.xs
556 As a hang over from the original C<constant> implementation, F<POSIX.xs>
557 contains a function C<int_macro_int> which in conjunction with C<AUTOLOAD> is
558 used to wrap the C functions C<WEXITSTATUS>, C<WIFEXITED>, C<WIFSIGNALED>,
559 C<WIFSTOPPED>, C<WSTOPSIG> and C<WTERMSIG>. It's probably worth replacing
560 this complexity with 5 simple direct wrappings of those 5 functions.
562 However, it would be interesting if someone could measure the memory usage
563 before and after, both for the case of C<use POSIX();> and the case of
564 actually calling the Perl space functions.
566 =head2 safely supporting POSIX SA_SIGINFO
568 Some years ago Jarkko supplied patches to provide support for the POSIX
569 SA_SIGINFO feature in Perl, passing the extra data to the Perl signal handler.
571 Unfortunately, it only works with "unsafe" signals, because under safe
572 signals, by the time Perl gets to run the signal handler, the extra
573 information has been lost. Moreover, it's not easy to store it somewhere,
574 as you can't call mutexs, or do anything else fancy, from inside a signal
577 So it strikes me that we could provide safe SA_SIGINFO support
583 Provide global variables for two file descriptors
587 When the first request is made via C<sigaction> for C<SA_SIGINFO>, create a
588 pipe, store the reader in one, the writer in the other
592 In the "safe" signal handler (C<Perl_csighandler()>/C<S_raise_signal()>), if
593 the C<siginfo_t> pointer non-C<NULL>, and the writer file handle is open,
599 serialise signal number, C<struct siginfo_t> (or at least the parts we care
600 about) into a small auto char buff
604 C<write()> that (non-blocking) to the writer fd
610 if it writes 100%, flag the signal in a counter of "signals on the pipe" akin
611 to the current per-signal-number counts
615 if it writes 0%, assume the pipe is full. Flag the data as lost?
619 if it writes partially, croak a panic, as your OS is broken.
627 in the regular C<PERL_ASYNC_CHECK()> processing, if there are "signals on
628 the pipe", read the data out, deserialise, build the Perl structures on
629 the stack (code in C<Perl_sighandler()>, the "unsafe" handler), and call as
634 I think that this gets us decent C<SA_SIGINFO> support, without the current risk
635 of running Perl code inside the signal handler context. (With all the dangers
636 of things like C<malloc> corruption that that currently offers us)
638 For more information see the thread starting with this message:
639 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-03/msg00305.html
641 =head2 autovivification
643 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
645 This task is incremental - even a little bit of work on it will help.
647 =head2 Unicode in Filenames
649 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
650 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
651 system, truncate, unlink, utime, -X. All these could potentially accept
652 Unicode filenames either as input or output (and in the case of system
653 and qx Unicode in general, as input or output to/from the shell).
654 Whether a filesystem - an operating system pair understands Unicode in
657 Known combinations that have some level of understanding include
658 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
659 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
660 create Unicode filenames, what forms of Unicode are accepted and used
661 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
662 and so on, varies. Finding the right level of interfacing to Perl
663 requires some thought. Remember that an OS does not implicate a
666 (The Windows -C command flag "wide API support" has been at least
667 temporarily retired in 5.8.1, and the -C has been repurposed, see
670 Most probably the right way to do this would be this:
671 L</"Virtualize operating system access">.
673 =head2 Unicode in %ENV
675 Currently the %ENV entries are always byte strings.
676 See L</"Virtualize operating system access">.
678 =head2 Unicode and glob()
680 Currently glob patterns and filenames returned from File::Glob::glob()
681 are always byte strings. See L</"Virtualize operating system access">.
683 =head2 Unicode and lc/uc operators
685 Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on
686 what the internal encoding of their argument is. That should not be the
687 case. Maybe add a pragma to switch behaviour.
689 =head2 use less 'memory'
691 Investigate trade offs to switch out perl's choices on memory usage.
692 Particularly perl should be able to give memory back.
694 This task is incremental - even a little bit of work on it will help.
696 =head2 Re-implement C<:unique> in a way that is actually thread-safe
698 The old implementation made bad assumptions on several levels. A good 90%
699 solution might be just to make C<:unique> work to share the string buffer
700 of SvPVs. That way large constant strings can be shared between ithreads,
701 such as the configuration information in F<Config>.
703 =head2 Make tainting consistent
705 Tainting would be easier to use if it didn't take documented shortcuts and
706 allow taint to "leak" everywhere within an expression.
708 =head2 readpipe(LIST)
710 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
711 running a shell. readpipe() (the function behind qx//) could be similarly
714 =head2 Audit the code for destruction ordering assumptions
718 /* Need to check SvMAGICAL, as during global destruction it may be that
719 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
720 is now part of the linked list of SV heads, rather than pointing to
721 the original body. */
722 /* FIXME - audit the code for other bugs like this one. */
724 adding the C<SvMAGICAL> check to
726 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
727 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
729 Go through the core and look for similar assumptions that SVs have particular
730 types, as all bets are off during global destruction.
732 =head2 Extend PerlIO and PerlIO::Scalar
734 PerlIO::Scalar doesn't know how to truncate(). Implementing this
735 would require extending the PerlIO vtable.
737 Similarly the PerlIO vtable doesn't know about formats (write()), or
738 about stat(), or chmod()/chown(), utime(), or flock().
740 (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
743 PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
744 opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
747 See also L</"Virtualize operating system access">.
749 =head2 -C on the #! line
751 It should be possible to make -C work correctly if found on the #! line,
752 given that all perl command line options are strict ASCII, and -C changes
753 only the interpretation of non-ASCII characters, and not for the script file
754 handle. To make it work needs some investigation of the ordering of function
755 calls during startup, and (by implication) a bit of tweaking of that order.
757 =head2 Organize error messages
759 Perl's diagnostics (error messages, see L<perldiag>) could use
760 reorganizing and formalizing so that each error message has its
761 stable-for-all-eternity unique id, categorized by severity, type, and
762 subsystem. (The error messages would be listed in a datafile outside
763 of the Perl source code, and the source code would only refer to the
764 messages by the id.) This clean-up and regularizing should apply
765 for all croak() messages.
767 This would enable all sorts of things: easier translation/localization
768 of the messages (though please do keep in mind the caveats of
769 L<Locale::Maketext> about too straightforward approaches to
770 translation), filtering by severity, and instead of grepping for a
771 particular error message one could look for a stable error id. (Of
772 course, changing the error messages by default would break all the
773 existing software depending on some particular error message...)
775 This kind of functionality is known as I<message catalogs>. Look for
776 inspiration for example in the catgets() system, possibly even use it
777 if available-- but B<only> if available, all platforms will B<not>
780 For the really pure at heart, consider extending this item to cover
781 also the warning messages (see L<perllexwarn>, C<warnings.pl>).
783 =head1 Tasks that need a knowledge of the interpreter
785 These tasks would need C knowledge, and knowledge of how the interpreter works,
786 or a willingness to learn.
788 =head2 lexicals used only once
792 $ perl -we '$pie = 42'
793 Name "main::pie" used only once: possible typo at -e line 1.
797 $ perl -we 'my $pie = 42'
799 Logically all lexicals used only once should warn, if the user asks for
800 warnings. An unworked RT ticket (#5087) has been open for almost seven
801 years for this discrepancy.
805 The handling of Unicode is unclean in many places. For example, the regexp
806 engine matches in Unicode semantics whenever the string or the pattern is
807 flagged as UTF-8, but that should not be dependent on an internal storage
808 detail of the string. Likewise, case folding behaviour is dependent on the
809 UTF8 internal flag being on or off.
811 =head2 Properly Unicode safe tokeniser and pads.
813 The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
814 variable names are stored in stashes as raw bytes, without the utf-8 flag
815 set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
816 tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
817 source filters. All this could be fixed.
819 =head2 state variable initialization in list context
821 Currently this is illegal:
823 state ($a, $b) = foo();
825 In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different
826 semantics, which is tricky to implement in Perl 5 as currently they produce
827 the same opcode trees. The Perl 6 design is firm, so it would be good to
828 implement the necessary code in Perl 5. There are comments in
829 C<Perl_newASSIGNOP()> that show the code paths taken by various assignment
830 constructions involving state variables.
832 =head2 Implement $value ~~ 0 .. $range
834 It would be nice to extend the syntax of the C<~~> operator to also
835 understand numeric (and maybe alphanumeric) ranges.
837 =head2 A does() built-in
839 Like ref(), only useful. It would call the C<DOES> method on objects; it
840 would also tell whether something can be dereferenced as an
841 array/hash/etc., or used as a regexp, etc.
842 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
844 =head2 Tied filehandles and write() don't mix
846 There is no method on tied filehandles to allow them to be called back by
849 =head2 Attach/detach debugger from running program
851 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
852 program if you pass the process ID. It would be good to do this with the Perl
853 debugger on a running Perl program, although I'm not sure how it would be
854 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
856 =head2 Optimize away empty destructors
858 Defining an empty DESTROY method might be useful (notably in
859 AUTOLOAD-enabled classes), but it's still a bit expensive to call. That
860 could probably be optimized.
862 =head2 LVALUE functions for lists
864 The old perltodo notes that lvalue functions don't work for list or hash
865 slices. This would be good to fix.
867 =head2 LVALUE functions in the debugger
869 The old perltodo notes that lvalue functions don't work in the debugger. This
870 would be good to fix.
872 =head2 regexp optimiser optional
874 The regexp optimiser is not optional. It should configurable to be, to allow
875 its performance to be measured, and its bugs to be easily demonstrated.
877 =head2 delete &function
879 Allow to delete functions. One can already undef them, but they're still
882 =head2 C</w> regex modifier
884 That flag would enable to match whole words, and also to interpolate
885 arrays as alternations. With it, C</P/w> would be roughly equivalent to:
887 do { local $"='|'; /\b(?:P)\b/ }
889 See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
892 =head2 optional optimizer
894 Make the peephole optimizer optional. Currently it performs two tasks as
895 it walks the optree - genuine peephole optimisations, and necessary fixups of
896 ops. It would be good to find an efficient way to switch out the
897 optimisations whilst keeping the fixups.
899 =head2 You WANT *how* many
901 Currently contexts are void, scalar and list. split has a special mechanism in
902 place to pass in the number of return values wanted. It would be useful to
903 have a general mechanism for this, backwards compatible and little speed hit.
904 This would allow proposals such as short circuiting sort to be implemented
907 =head2 lexical aliases
909 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
911 =head2 entersub XS vs Perl
913 At the moment pp_entersub is huge, and has code to deal with entering both
914 perl and XS subroutines. Subroutine implementations rarely change between
915 perl and XS at run time, so investigate using 2 ops to enter subs (one for
916 XS, one for perl) and swap between if a sub is redefined.
920 Self-ties are currently illegal because they caused too many segfaults. Maybe
921 the causes of these could be tracked down and self-ties on all types
924 =head2 Optimize away @_
926 The old perltodo notes "Look at the "reification" code in C<av.c>".
928 =head2 The yada yada yada operators
930 Perl 6's Synopsis 3 says:
932 I<The ... operator is the "yada, yada, yada" list operator, which is used as
933 the body in function prototypes. It complains bitterly (by calling fail)
934 if it is ever executed. Variant ??? calls warn, and !!! calls die.>
936 Those would be nice to add to Perl 5. That could be done without new ops.
938 =head2 Virtualize operating system access
940 Implement a set of "vtables" that virtualizes operating system access
941 (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
942 least these interfaces should take SVs as "name" arguments instead of
943 bare char pointers; probably the most flexible and extensible way
944 would be for the Perl-facing interfaces to accept HVs. The system
945 needs to be per-operating-system and per-file-system
946 hookable/filterable, preferably both from XS and Perl level
947 (L<perlport/"Files and Filesystems"> is good reading at this point,
948 in fact, all of L<perlport> is.)
950 This has actually already been implemented (but only for Win32),
951 take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
952 variants go through a set of "vtables" for operating system access,
953 non-Win32 systems currently go straight for the POSIX/UNIX-style
954 system/library call. Similar system as for Win32 should be
955 implemented for all platforms. The existing Win32 implementation
956 probably does not need to survive alongside this proposed new
957 implementation, the approaches could be merged.
959 What would this give us? One often-asked-for feature this would
960 enable is using Unicode for filenames, and other "names" like %ENV,
961 usernames, hostnames, and so forth.
962 (See L<perlunicode/"When Unicode Does Not Happen">.)
964 But this kind of virtualization would also allow for things like
965 virtual filesystems, virtual networks, and "sandboxes" (though as long
966 as dynamic loading of random object code is allowed, not very safe
967 sandboxes since external code of course know not of Perl's vtables).
968 An example of a smaller "sandbox" is that this feature can be used to
969 implement per-thread working directories: Win32 already does this.
971 See also L</"Extend PerlIO and PerlIO::Scalar">.
973 =head2 Investigate PADTMP hash pessimisation
975 The peephole optimier converts constants used for hash key lookups to shared
976 hash key scalars. Under ithreads, something is undoing this work.
977 See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html
979 =head2 Store the current pad in the OP slab allocator
982 I hope that I got that "current pad" part correct
984 Currently we leak ops in various cases of parse failure. I suggested that we
985 could solve this by always using the op slab allocator, and walking it to
986 free ops. Dave comments that as some ops are already freed during optree
987 creation one would have to mark which ops are freed, and not double free them
988 when walking the slab. He notes that one problem with this is that for some ops
989 you have to know which pad was current at the time of allocation, which does
990 change. I suggested storing a pointer to the current pad in the memory allocated
991 for the slab, and swapping to a new slab each time the pad changes. Dave thinks
992 that this would work.
994 =head2 repack the optree
996 Repacking the optree after execution order is determined could allow
997 removal of NULL ops, and optimal ordering of OPs with respect to cache-line
998 filling. The slab allocator could be reused for this purpose. I think that
999 the best way to do this is to make it an optional step just before the
1000 completed optree is attached to anything else, and to use the slab allocator
1001 unchanged, so that freeing ops is identical whether or not this step runs.
1002 Note that the slab allocator allocates ops downwards in memory, so one would
1003 have to actually "allocate" the ops in reverse-execution order to get them
1004 contiguous in memory in execution order.
1006 See http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html
1008 Note that running this copy, and then freeing all the old location ops would
1009 cause their slabs to be freed, which would eliminate possible memory wastage if
1010 the previous suggestion is implemented, and we swap slabs more frequently.
1012 =head2 eliminate incorrect line numbers in warnings
1020 } elsif ($undef == 0) {
1023 used to produce this output:
1025 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1026 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1028 where the line of the second warning was misreported - it should be line 5.
1029 Rafael fixed this - the problem arose because there was no nextstate OP
1030 between the execution of the C<if> and the C<elsif>, hence C<PL_curcop> still
1031 reports that the currently executing line is line 4. The solution was to inject
1032 a nextstate OPs for each C<elsif>, although it turned out that the nextstate
1033 OP needed to be a nulled OP, rather than a live nextstate OP, else other line
1034 numbers became misreported. (Jenga!)
1036 The problem is more general than C<elsif> (although the C<elsif> case is the
1037 most common and the most confusing). Ideally this code
1047 would produce this output
1049 Use of uninitialized value $undef in addition (+) at wrong.pl line 4.
1050 Use of uninitialized value $undef in addition (+) at wrong.pl line 7.
1052 (rather than lines 4 and 5), but this would seem to require every OP to carry
1053 (at least) line number information.
1055 What might work is to have an optional line number in memory just before the
1056 BASEOP structure, with a flag bit in the op to say whether it's present.
1057 Initially during compile every OP would carry its line number. Then add a late
1058 pass to the optimiser (potentially combined with L</repack the optree>) which
1059 looks at the two ops on every edge of the graph of the execution path. If
1060 the line number changes, flags the destination OP with this information.
1061 Once all paths are traced, replace every op with the flag with a
1062 nextstate-light op (that just updates C<PL_curcop>), which in turn then passes
1063 control on to the true op. All ops would then be replaced by variants that
1064 do not store the line number. (Which, logically, why it would work best in
1065 conjunction with L</repack the optree>, as that is already copying/reallocating
1068 (Although I should note that we're not certain that doing this for the general
1071 =head2 optimize tail-calls
1073 Tail-calls present an opportunity for broadly applicable optimization;
1074 anywhere that C<< return foo(...) >> is called, the outer return can
1075 be replaced by a goto, and foo will return directly to the outer
1076 caller, saving (conservatively) 25% of perl's call&return cost, which
1077 is relatively higher than in C. The scheme language is known to do
1078 this heavily. B::Concise provides good insight into where this
1079 optimization is possible, ie anywhere entersub,leavesub op-sequence
1082 perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)'
1084 Bottom line on this is probably a new pp_tailcall function which
1085 combines the code in pp_entersub, pp_leavesub. This should probably
1086 be done 1st in XS, and using B::Generate to patch the new OP into the
1091 Tasks that will get your name mentioned in the description of the "Highlights
1094 =head2 make ithreads more robust
1096 Generally make ithreads more robust. See also L</iCOW>
1098 This task is incremental - even a little bit of work on it will help, and
1099 will be greatly appreciated.
1101 One bit would be to write the missing code in sv.c:Perl_dirp_dup.
1103 Fix Perl_sv_dup, et al so that threads can return objects.
1107 Sarathy and Arthur have a proposal for an improved Copy On Write which
1108 specifically will be able to COW new ithreads. If this can be implemented
1109 it would be a good thing.
1111 =head2 (?{...}) closures in regexps
1113 Fix (or rewrite) the implementation of the C</(?{...})/> closures.
1115 =head2 A re-entrant regexp engine
1117 This will allow the use of a regex from inside (?{ }), (??{ }) and
1118 (?(?{ })|) constructs.
1120 =head2 Add class set operations to regexp engine
1122 Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
1124 demerphq has this on his todo list, but right at the bottom.