3 perltodo - Perl TO-DO List
7 This is a list of wishes for Perl. The most up to date version of this file
8 is at http://perl5.git.perl.org/perl.git/blob_plain/HEAD:/pod/perltodo.pod
10 The tasks we think are smaller or easier are listed first. Anyone is welcome
11 to work on any of these, but it's a good idea to first contact
12 I<perl5-porters@perl.org> to avoid duplication of effort, and to learn from
13 any previous attempts. By all means contact a pumpking privately first if you
16 Whilst patches to make the list shorter are most welcome, ideas to add to
17 the list are also encouraged. Check the perl5-porters archives for past
18 ideas, and any discussion about them. One set of archives may be found at:
20 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
22 What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
23 not, but if your patch is incorporated, then we'll add your name to the
24 F<AUTHORS> file, which ships in the official distribution. How many other
25 programming languages offer you 1 line of immortality?
27 =head1 Tasks that only need Perl knowledge
29 =head2 Migrate t/ from custom TAP generation
31 Many tests below F<t/> still generate TAP by "hand", rather than using library
32 functions. As explained in L<perlhack/Writing a test>, tests in F<t/> are
33 written in a particular way to test that more complex constructions actually
34 work before using them routinely. Hence they don't use C<Test::More>, but
35 instead there is an intentionally simpler library, F<t/test.pl>. However,
36 quite a few tests in F<t/> have not been refactored to use it. Refactoring
37 any of these tests, one at a time, is a useful thing TODO.
39 =head2 Automate perldelta generation
41 The perldelta file accompanying each release summaries the major changes.
42 It's mostly manually generated currently, but some of that could be
43 automated with a bit of perl, specifically the generation of
47 =item Modules and Pragmata
49 =item New Documentation
55 See F<Porting/how_to_write_a_perldelta.pod> for details.
57 =head2 Remove duplication of test setup.
59 Schwern notes, that there's duplication of code - lots and lots of tests have
60 some variation on the big block of C<$Is_Foo> checks. We can safely put this
61 into a file, change it to build an C<%Is> hash and require it. Maybe just put
62 it into F<test.pl>. Throw in the handy tainting subroutines.
64 =head2 POD -E<gt> HTML conversion in the core still sucks
66 Which is crazy given just how simple POD purports to be, and how simple HTML
67 can be. It's not actually I<as> simple as it sounds, particularly with the
68 flexibility POD allows for C<=item>, but it would be good to improve the
69 visual appeal of the HTML generated, and to avoid it having any validation
70 errors. See also L</make HTML install work>, as the layout of installation tree
71 is needed to improve the cross-linking.
73 The addition of C<Pod::Simple> and its related modules may make this task
76 =head2 Make ExtUtils::ParseXS use strict;
78 F<lib/ExtUtils/ParseXS.pm> contains this line
80 # use strict; # One of these days...
82 Simply uncomment it, and fix all the resulting issues :-)
84 The more practical approach, to break the task down into manageable chunks, is
85 to work your way though the code from bottom to top, or if necessary adding
86 extra C<{ ... }> blocks, and turning on strict within them.
88 =head2 Make Schwern poorer
90 We should have tests for everything. When all the core's modules are tested,
91 Schwern has promised to donate to $500 to TPF. We may need volunteers to
92 hold him upside down and shake vigorously in order to actually extract the
95 =head2 Improve the coverage of the core tests
97 Use Devel::Cover to ascertain the core modules's test coverage, then add
98 tests that are currently missing.
102 A full test suite for the B module would be nice.
104 =head2 A decent benchmark
106 C<perlbench> seems impervious to any recent changes made to the perl core. It
107 would be useful to have a reasonable general benchmarking suite that roughly
108 represented what current perl programs do, and measurably reported whether
109 tweaks to the core improve, degrade or don't really affect performance, to
110 guide people attempting to optimise the guts of perl. Gisle would welcome
111 new tests for perlbench.
113 =head2 fix tainting bugs
115 Fix the bugs revealed by running the test suite with the C<-t> switch (via
116 C<make test.taintwarn>).
118 =head2 Dual life everything
120 As part of the "dists" plan, anything that doesn't belong in the smallest perl
121 distribution needs to be dual lifed. Anything else can be too. Figure out what
122 changes would be needed to package that module and its tests up for CPAN, and
123 do so. Test it with older perl releases, and fix the problems you find.
125 To make a minimal perl distribution, it's useful to look at
126 F<t/lib/commonsense.t>.
128 =head2 Move dual-life pod/*.PL into ext
130 Nearly all the dual-life modules have been moved to F<ext>. However, we
131 still need to move F<pod/*.PL> into their respective directories
132 in F<ext/>. They're referenced by (at least) C<plextract> in F<Makefile.SH>
133 and C<utils> in F<win32/Makefile> and F<win32/makefile.ml>, and listed
134 explicitly in F<win32/pod.mak>, F<vms/descrip_mms.template> and F<utils.lst>
136 =head2 POSIX memory footprint
138 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
139 various times worked to cut it down. There is probably still fat to cut out -
140 for example POSIX passes Exporter some very memory hungry data structures.
142 =head2 embed.pl/makedef.pl
144 There is a script F<embed.pl> that generates several header files to prefix
145 all of Perl's symbols in a consistent way, to provide some semblance of
146 namespace support in C<C>. Functions are declared in F<embed.fnc>, variables
147 in F<interpvar.h>. Quite a few of the functions and variables
148 are conditionally declared there, using C<#ifdef>. However, F<embed.pl>
149 doesn't understand the C macros, so the rules about which symbols are present
150 when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay.
151 It would be good to teach C<embed.pl> to understand the conditional
152 compilation, and hence remove the duplication, and the mistakes it has caused.
154 =head2 use strict; and AutoLoad
156 Currently if you write
159 use AutoLoader 'AUTOLOAD';
164 print join (' ', No, strict, here), "!\n";
167 then C<use strict;> isn't in force within the autoloaded subroutines. It would
168 be more consistent (and less surprising) to arrange for all lexical pragmas
169 in force at the __END__ block to be in force within each autoloaded subroutine.
171 There's a similar problem with SelfLoader.
173 =head2 profile installman
175 The F<installman> script is slow. All it is doing text processing, which we're
176 told is something Perl is good at. So it would be nice to know what it is doing
177 that is taking so much CPU, and where possible address it.
180 =head1 Tasks that need a little sysadmin-type knowledge
182 Or if you prefer, tasks that you would learn from, and broaden your skills
185 =head2 make HTML install work
187 There is an C<installhtml> target in the Makefile. It's marked as
188 "experimental". It would be good to get this tested, make it work reliably, and
189 remove the "experimental" tag. This would include
195 Checking that cross linking between various parts of the documentation works.
196 In particular that links work between the modules (files with POD in F<lib/>)
197 and the core documentation (files in F<pod/>)
201 Work out how to split C<perlfunc> into chunks, preferably one per function
202 group, preferably with general case code that could be used elsewhere.
203 Challenges here are correctly identifying the groups of functions that go
204 together, and making the right named external cross-links point to the right
205 page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
206 C<endservent>, two or more C<=items> giving the different parameter lists, such
209 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
210 =item substr EXPR,OFFSET,LENGTH
211 =item substr EXPR,OFFSET
213 and different parameter lists having different meanings. (eg C<select>)
217 =head2 compressed man pages
219 Be able to install them. This would probably need a configure test to see how
220 the system does compressed man pages (same directory/different directory?
221 same filename/different filename), as well as tweaking the F<installman> script
222 to compress as necessary.
224 =head2 Add a code coverage target to the Makefile
226 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
227 to do this manually are roughly
233 do a normal C<Configure>, but include Devel::Cover as a module to install
234 (see F<INSTALL> for how to do this)
242 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
246 Process the resulting Devel::Cover database
250 This just give you the coverage of the F<.pm>s. To also get the C level
257 Additionally tell C<Configure> to use the appropriate C compiler flags for
264 (instead of C<make perl>)
268 After running the tests run C<gcov> to generate all the F<.gcov> files.
269 (Including down in the subdirectories of F<ext/>
273 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
274 to get their stats into the cover_db directory.
278 Then process the Devel::Cover database
282 It would be good to add a single switch to C<Configure> to specify that you
283 wanted to perform perl level coverage, and another to specify C level
284 coverage, and have C<Configure> and the F<Makefile> do all the right things
287 =head2 Make Config.pm cope with differences between built and installed perl
289 Quite often vendors ship a perl binary compiled with their (pay-for)
290 compilers. People install a free compiler, such as gcc. To work out how to
291 build extensions, Perl interrogates C<%Config>, so in this situation
292 C<%Config> describes compilers that aren't there, and extension building
293 fails. This forces people into choosing between re-compiling perl themselves
294 using the compiler they have, or only using modules that the vendor ships.
296 It would be good to find a way teach C<Config.pm> about the installation setup,
297 possibly involving probing at install time or later, so that the C<%Config> in
298 a binary distribution better describes the installed machine, when the
299 installed machine differs from the build machine in some significant way.
301 =head2 linker specification files
303 Some platforms mandate that you provide a list of a shared library's external
304 symbols to the linker, so the core already has the infrastructure in place to
305 do this for generating shared perl libraries. My understanding is that the
306 GNU toolchain can accept an optional linker specification file, and restrict
307 visibility just to symbols declared in that file. It would be good to extend
308 F<makedef.pl> to support this format, and to provide a means within
309 C<Configure> to enable it. This would allow Unix users to test that the
310 export list is correct, and to build a perl that does not pollute the global
311 namespace with private symbols.
313 =head2 Cross-compile support
315 Currently C<Configure> understands C<-Dusecrosscompile> option. This option
316 arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
317 assumed then to be copied to TARGET machine and used as a replacement of full
320 This could be done little differently. Namely C<miniperl> should be built for
321 HOST and then full C<perl> with extensions should be compiled for TARGET.
322 This, however, might require extra trickery for %Config: we have one config
323 first for HOST and then another for TARGET. Tools like MakeMaker will be
324 mightily confused. Having around two different types of executables and
325 libraries (HOST and TARGET) makes life interesting for Makefiles and
326 shell (and Perl) scripts. There is $Config{run}, normally empty, which
327 can be used as an execution wrapper. Also note that in some
328 cross-compilation/execution environments the HOST and the TARGET do
329 not see the same filesystem(s), the $Config{run} may need to do some
330 file/directory copying back and forth.
334 Make F<pod/roffitall> be updated by F<pod/buildtoc>.
336 =head2 Split "linker" from "compiler"
338 Right now, Configure probes for two commands, and sets two variables:
342 =item * C<cc> (in F<cc.U>)
344 This variable holds the name of a command to execute a C compiler which
345 can resolve multiple global references that happen to have the same
346 name. Usual values are F<cc> and F<gcc>.
347 Fervent ANSI compilers may be called F<c89>. AIX has F<xlc>.
349 =item * C<ld> (in F<dlsrc.U>)
351 This variable indicates the program to be used to link
352 libraries for dynamic loading. On some systems, it is F<ld>.
353 On ELF systems, it should be C<$cc>. Mostly, we'll try to respect
354 the hint file setting.
358 There is an implicit historical assumption from around Perl5.000alpha
359 something, that C<$cc> is also the correct command for linking object files
360 together to make an executable. This may be true on Unix, but it's not true
361 on other platforms, and there are a maze of work arounds in other places (such
362 as F<Makefile.SH>) to cope with this.
364 Ideally, we should create a new variable to hold the name of the executable
365 linker program, probe for it in F<Configure>, and centralise all the special
366 case logic there or in hints files.
368 A small bikeshed issue remains - what to call it, given that C<$ld> is already
369 taken (arguably for the wrong thing now, but on SunOS 4.1 it is the command
370 for creating dynamically-loadable modules) and C<$link> could be confused with
371 the Unix command line executable of the same name, which does something
372 completely different. Andy Dougherty makes the counter argument "In parrot, I
373 tried to call the command used to link object files and libraries into an
374 executable F<link>, since that's what my vaguely-remembered DOS and VMS
375 experience suggested. I don't think any real confusion has ensued, so it's
376 probably a reasonable name for perl5 to use."
378 "Alas, I've always worried that introducing it would make things worse,
379 since now the module building utilities would have to look for
380 C<$Config{link}> and institute a fall-back plan if it weren't found."
381 Although I can see that as confusing, given that C<$Config{d_link}> is true
382 when (hard) links are available.
384 =head2 Configure Windows using PowerShell
386 Currently, Windows uses hard-coded config files based to build the
387 config.h for compiling Perl. Makefiles are also hard-coded and need to be
388 hand edited prior to building Perl. While this makes it easy to create a perl.exe
389 that works across multiple Windows versions, being able to accurately
390 configure a perl.exe for a specific Windows versions and VS C++ would be
391 a nice enhancement. With PowerShell available on Windows XP and up, this
392 may now be possible. Step 1 might be to investigate whether this is possible
393 and use this to clean up our current makefile situation. Step 2 would be to
394 see if there would be a way to use our existing metaconfig units to configure a
395 Windows Perl or whether we go in a separate direction and make it so. Of
396 course, we all know what step 3 is.
398 =head2 decouple -g and -DDEBUGGING
400 Currently F<Configure> automatically adds C<-DDEBUGGING> to the C compiler
401 flags if it spots C<-g> in the optimiser flags. The pre-processor directive
402 C<DEBUGGING> enables F<perl>'s command line C<-D> options, but in the process
403 makes F<perl> slower. It would be good to disentangle this logic, so that
404 C-level debugging with C<-g> and Perl level debugging with C<-D> can easily
405 be enabled independently.
407 =head1 Tasks that need a little C knowledge
409 These tasks would need a little C knowledge, but don't need any specific
410 background or experience with XS, or how the Perl interpreter works
412 =head2 Weed out needless PERL_UNUSED_ARG
414 The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about
415 unused arguments. Often the arguments can't be removed, as there is an
416 external constraint that determines the prototype of the function, so this
417 approach is valid. However, there are some cases where C<PERL_UNUSED_ARG>
418 could be removed. Specifically
424 The prototypes of (nearly all) static functions can be changed
428 Unused arguments generated by short cut macros are wasteful - the short cut
429 macro used can be changed.
433 =head2 Modernize the order of directories in @INC
435 The way @INC is laid out by default, one cannot upgrade core (dual-life)
436 modules without overwriting files. This causes problems for binary
437 package builders. One possible proposal is laid out in this
439 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>.
443 Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
444 On these systems, it might be the default compilation mode, and there
445 is currently no guarantee that passing no use64bitall option to the
446 Configure process will build a 32bit perl. Implementing -Duse32bit*
447 options would be nice for perl 5.12.
449 =head2 Profile Perl - am I hot or not?
451 The Perl source code is stable enough that it makes sense to profile it,
452 identify and optimise the hotspots. It would be good to measure the
453 performance of the Perl interpreter using free tools such as cachegrind,
454 gprof, and dtrace, and work to reduce the bottlenecks they reveal.
456 As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
457 the ops that are most commonly used. The idea is that by grouping them, their
458 object code will be adjacent in the executable, so they have a greater chance
459 of already being in the CPU cache (or swapped in) due to being near another op
462 Except that it's not clear if these really are the most commonly used ops. So
463 as part of exercising your skills with coverage and profiling tools you might
464 want to determine what ops I<really> are the most commonly used. And in turn
465 suggest evictions and promotions to achieve a better F<pp_hot.c>.
467 One piece of Perl code that might make a good testbed is F<installman>.
469 =head2 Allocate OPs from arenas
471 Currently all new OP structures are individually malloc()ed and free()d.
472 All C<malloc> implementations have space overheads, and are now as fast as
473 custom allocates so it would both use less memory and less CPU to allocate
474 the various OP structures from arenas. The SV arena code can probably be
477 Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use
478 Perl_Slab_alloc() to pack optrees into a contiguous block, which is
479 probably superior to the use of OP arenas, esp. from a cache locality
480 standpoint. See L<Profile Perl - am I hot or not?>.
482 =head2 Improve win32/wince.c
484 Currently, numerous functions look virtually, if not completely,
485 identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't
488 =head2 Use secure CRT functions when building with VC8 on Win32
490 Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
491 that they were "unsafe" and introduced differently named secure versions of
492 them as replacements, e.g. instead of writing
494 FILE* f = fopen(__FILE__, "r");
499 errno_t err = fopen_s(&f, __FILE__, "r");
501 Currently, the warnings about these deprecations have been disabled by adding
502 -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
503 warning suppressant and actually make use of the new secure CRT functions.
505 There is also a similar issue with POSIX CRT function names like fileno having
506 been deprecated in favour of ISO C++ conformant names like _fileno. These
507 warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
508 might be nice to do as Microsoft suggest here too, although, unlike the secure
509 functions issue, there is presumably little or no benefit in this case.
511 =head2 Fix POSIX::access() and chdir() on Win32
513 These functions currently take no account of DACLs and therefore do not behave
514 correctly in situations where access is restricted by DACLs (as opposed to the
515 read-only attribute).
517 Furthermore, POSIX::access() behaves differently for directories having the
518 read-only attribute set depending on what CRT library is being used. For
519 example, the _access() function in the VC6 and VC7 CRTs (wrongly) claim that
520 such directories are not writable, whereas in fact all directories are writable
521 unless access is denied by DACLs. (In the case of directories, the read-only
522 attribute actually only means that the directory cannot be deleted.) This CRT
523 bug is fixed in the VC8 and VC9 CRTs (but, of course, the directory may still
524 not actually be writable if access is indeed denied by DACLs).
526 For the chdir() issue, see ActiveState bug #74552:
527 http://bugs.activestate.com/show_bug.cgi?id=74552
529 Therefore, DACLs should be checked both for consistency across CRTs and for
532 (Note that perl's -w operator should not be modified to check DACLs. It has
533 been written so that it reflects the state of the read-only attribute, even
534 for directories (whatever CRT is being used), for symmetry with chmod().)
536 =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
538 Maybe create a utility that checks after each libperl.a creation that
539 none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
540 ever creep back to libperl.a.
542 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
544 Note, of course, that this will only tell whether B<your> platform
545 is using those naughty interfaces.
547 =head2 -D_FORTIFY_SOURCE=2, -fstack-protector
549 Recent glibcs support C<-D_FORTIFY_SOURCE=2> and recent gcc
550 (4.1 onwards?) supports C<-fstack-protector>, both of which give
551 protection against various kinds of buffer overflow problems.
552 These should probably be used for compiling Perl whenever available,
553 Configure and/or hints files should be adjusted to probe for the
554 availability of these features and enable them as appropriate.
556 =head2 Arenas for GPs? For MAGIC?
558 C<struct gp> and C<struct magic> are both currently allocated by C<malloc>.
559 It might be a speed or memory saving to change to using arenas. Or it might
560 not. It would need some suitable benchmarking first. In particular, C<GP>s
561 can probably be changed with minimal compatibility impact (probably nothing
562 outside of the core, or even outside of F<gv.c> allocates them), but they
563 probably aren't allocated/deallocated often enough for a speed saving. Whereas
564 C<MAGIC> is allocated/deallocated more often, but in turn, is also something
565 more externally visible, so changing the rules here may bite external code.
569 Several SV body structs are now the same size, notably PVMG and PVGV, PVAV and
570 PVHV, and PVCV and PVFM. It should be possible to allocate and return same
571 sized bodies from the same actual arena, rather than maintaining one arena for
572 each. This could save 4-6K per thread, of memory no longer tied up in the
573 not-yet-allocated part of an arena.
576 =head1 Tasks that need a knowledge of XS
578 These tasks would need C knowledge, and roughly the level of knowledge of
579 the perl API that comes from writing modules that use XS to interface to
582 =head2 Write an XS cookbook
584 Create pod/perlxscookbook.pod with short, task-focused 'recipes' in XS that
585 demonstrate common tasks and good practices. (Some of these might be
586 extracted from perlguts.) The target audience should be XS novices, who need
587 more examples than perlguts but something less overwhelming than perlapi.
588 Recipes should provide "one pretty good way to do it" instead of TIMTOWTDI.
590 Rather than focusing on interfacing Perl to C libraries, such a cookbook
591 should probably focus on how to optimize Perl routines by re-writing them
592 in XS. This will likely be more motivating to those who mostly work in
593 Perl but are looking to take the next step into XS.
595 Deconstructing and explaining some simpler XS modules could be one way to
596 bootstrap a cookbook. (List::Util? Class::XSAccessor? Tree::Ternary_XS?)
597 Another option could be deconstructing the implementation of some simpler
600 =head2 Remove the use of SVs as temporaries in dump.c
602 F<dump.c> contains debugging routines to dump out the contains of perl data
603 structures, such as C<SV>s, C<AV>s and C<HV>s. Currently, the dumping code
604 B<uses> C<SV>s for its temporary buffers, which was a logical initial
605 implementation choice, as they provide ready made memory handling.
607 However, they also lead to a lot of confusion when it happens that what you're
608 trying to debug is seen by the code in F<dump.c>, correctly or incorrectly, as
609 a temporary scalar it can use for a temporary buffer. It's also not possible
610 to dump scalars before the interpreter is properly set up, such as during
611 ithreads cloning. It would be good to progressively replace the use of scalars
612 as string accumulation buffers with something much simpler, directly allocated
613 by C<malloc>. The F<dump.c> code is (or should be) only producing 7 bit
614 US-ASCII, so output character sets are not an issue.
616 Producing and proving an internal simple buffer allocation would make it easier
617 to re-write the internals of the PerlIO subsystem to avoid using C<SV>s for
618 B<its> buffers, use of which can cause problems similar to those of F<dump.c>,
621 =head2 safely supporting POSIX SA_SIGINFO
623 Some years ago Jarkko supplied patches to provide support for the POSIX
624 SA_SIGINFO feature in Perl, passing the extra data to the Perl signal handler.
626 Unfortunately, it only works with "unsafe" signals, because under safe
627 signals, by the time Perl gets to run the signal handler, the extra
628 information has been lost. Moreover, it's not easy to store it somewhere,
629 as you can't call mutexs, or do anything else fancy, from inside a signal
632 So it strikes me that we could provide safe SA_SIGINFO support
638 Provide global variables for two file descriptors
642 When the first request is made via C<sigaction> for C<SA_SIGINFO>, create a
643 pipe, store the reader in one, the writer in the other
647 In the "safe" signal handler (C<Perl_csighandler()>/C<S_raise_signal()>), if
648 the C<siginfo_t> pointer non-C<NULL>, and the writer file handle is open,
654 serialise signal number, C<struct siginfo_t> (or at least the parts we care
655 about) into a small auto char buff
659 C<write()> that (non-blocking) to the writer fd
665 if it writes 100%, flag the signal in a counter of "signals on the pipe" akin
666 to the current per-signal-number counts
670 if it writes 0%, assume the pipe is full. Flag the data as lost?
674 if it writes partially, croak a panic, as your OS is broken.
682 in the regular C<PERL_ASYNC_CHECK()> processing, if there are "signals on
683 the pipe", read the data out, deserialise, build the Perl structures on
684 the stack (code in C<Perl_sighandler()>, the "unsafe" handler), and call as
689 I think that this gets us decent C<SA_SIGINFO> support, without the current risk
690 of running Perl code inside the signal handler context. (With all the dangers
691 of things like C<malloc> corruption that that currently offers us)
693 For more information see the thread starting with this message:
694 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-03/msg00305.html
696 =head2 autovivification
698 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
700 This task is incremental - even a little bit of work on it will help.
702 =head2 Unicode in Filenames
704 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
705 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
706 system, truncate, unlink, utime, -X. All these could potentially accept
707 Unicode filenames either as input or output (and in the case of system
708 and qx Unicode in general, as input or output to/from the shell).
709 Whether a filesystem - an operating system pair understands Unicode in
712 Known combinations that have some level of understanding include
713 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
714 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
715 create Unicode filenames, what forms of Unicode are accepted and used
716 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
717 and so on, varies. Finding the right level of interfacing to Perl
718 requires some thought. Remember that an OS does not implicate a
721 (The Windows -C command flag "wide API support" has been at least
722 temporarily retired in 5.8.1, and the -C has been repurposed, see
725 Most probably the right way to do this would be this:
726 L</"Virtualize operating system access">.
728 =head2 Unicode in %ENV
730 Currently the %ENV entries are always byte strings.
731 See L</"Virtualize operating system access">.
733 =head2 Unicode and glob()
735 Currently glob patterns and filenames returned from File::Glob::glob()
736 are always byte strings. See L</"Virtualize operating system access">.
738 =head2 Unicode and lc/uc operators
740 Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on
741 what the internal encoding of their argument is. That should not be the
742 case. Maybe add a pragma to switch behaviour.
744 =head2 use less 'memory'
746 Investigate trade offs to switch out perl's choices on memory usage.
747 Particularly perl should be able to give memory back.
749 This task is incremental - even a little bit of work on it will help.
751 =head2 Re-implement C<:unique> in a way that is actually thread-safe
753 The old implementation made bad assumptions on several levels. A good 90%
754 solution might be just to make C<:unique> work to share the string buffer
755 of SvPVs. That way large constant strings can be shared between ithreads,
756 such as the configuration information in F<Config>.
758 =head2 Make tainting consistent
760 Tainting would be easier to use if it didn't take documented shortcuts and
761 allow taint to "leak" everywhere within an expression.
763 =head2 readpipe(LIST)
765 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
766 running a shell. readpipe() (the function behind qx//) could be similarly
769 =head2 Audit the code for destruction ordering assumptions
773 /* Need to check SvMAGICAL, as during global destruction it may be that
774 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
775 is now part of the linked list of SV heads, rather than pointing to
776 the original body. */
777 /* FIXME - audit the code for other bugs like this one. */
779 adding the C<SvMAGICAL> check to
781 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
782 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
784 Go through the core and look for similar assumptions that SVs have particular
785 types, as all bets are off during global destruction.
787 =head2 Extend PerlIO and PerlIO::Scalar
789 PerlIO::Scalar doesn't know how to truncate(). Implementing this
790 would require extending the PerlIO vtable.
792 Similarly the PerlIO vtable doesn't know about formats (write()), or
793 about stat(), or chmod()/chown(), utime(), or flock().
795 (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
798 PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
799 opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
802 See also L</"Virtualize operating system access">.
804 =head2 -C on the #! line
806 It should be possible to make -C work correctly if found on the #! line,
807 given that all perl command line options are strict ASCII, and -C changes
808 only the interpretation of non-ASCII characters, and not for the script file
809 handle. To make it work needs some investigation of the ordering of function
810 calls during startup, and (by implication) a bit of tweaking of that order.
812 =head2 Organize error messages
814 Perl's diagnostics (error messages, see L<perldiag>) could use
815 reorganizing and formalizing so that each error message has its
816 stable-for-all-eternity unique id, categorized by severity, type, and
817 subsystem. (The error messages would be listed in a datafile outside
818 of the Perl source code, and the source code would only refer to the
819 messages by the id.) This clean-up and regularizing should apply
820 for all croak() messages.
822 This would enable all sorts of things: easier translation/localization
823 of the messages (though please do keep in mind the caveats of
824 L<Locale::Maketext> about too straightforward approaches to
825 translation), filtering by severity, and instead of grepping for a
826 particular error message one could look for a stable error id. (Of
827 course, changing the error messages by default would break all the
828 existing software depending on some particular error message...)
830 This kind of functionality is known as I<message catalogs>. Look for
831 inspiration for example in the catgets() system, possibly even use it
832 if available-- but B<only> if available, all platforms will B<not>
835 For the really pure at heart, consider extending this item to cover
836 also the warning messages (see L<perllexwarn>, C<warnings.pl>).
838 =head1 Tasks that need a knowledge of the interpreter
840 These tasks would need C knowledge, and knowledge of how the interpreter works,
841 or a willingness to learn.
843 =head2 truncate() prototype
845 The prototype of truncate() is currently C<$$>. It should probably
846 be C<*$> instead. (This is changed in F<opcode.pl>)
848 =head2 decapsulation of smart match argument
850 Currently C<$foo ~~ $object> will die with the message "Smart matching a
851 non-overloaded object breaks encapsulation". It would be nice to allow
852 to bypass this by using explictly the syntax C<$foo ~~ %$object> or
855 =head2 error reporting of [$a ; $b]
857 Using C<;> inside brackets is a syntax error, and we don't propose to change
858 that by giving it any meaning. However, it's not reported very helpfully:
860 $ perl -e '$a = [$b; $c];'
861 syntax error at -e line 1, near "$b;"
862 syntax error at -e line 1, near "$c]"
863 Execution of -e aborted due to compilation errors.
865 It should be possible to hook into the tokeniser or the lexer, so that when a
866 C<;> is parsed where it is not legal as a statement terminator (ie inside
867 C<{}> used as a hashref, C<[]> or C<()>) it issues an error something like
868 I<';' isn't legal inside an expression - if you need multiple statements use a
869 do {...} block>. See the thread starting at
870 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-09/msg00573.html
872 =head2 lexicals used only once
876 $ perl -we '$pie = 42'
877 Name "main::pie" used only once: possible typo at -e line 1.
881 $ perl -we 'my $pie = 42'
883 Logically all lexicals used only once should warn, if the user asks for
884 warnings. An unworked RT ticket (#5087) has been open for almost seven
885 years for this discrepancy.
889 The handling of Unicode is unclean in many places. For example, the regexp
890 engine matches in Unicode semantics whenever the string or the pattern is
891 flagged as UTF-8, but that should not be dependent on an internal storage
892 detail of the string. Likewise, case folding behaviour is dependent on the
893 UTF8 internal flag being on or off.
895 =head2 Properly Unicode safe tokeniser and pads.
897 The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack -
898 variable names are stored in stashes as raw bytes, without the utf-8 flag
899 set. The pad API only takes a C<char *> pointer, so that's all bytes too. The
900 tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from
901 source filters. All this could be fixed.
903 =head2 state variable initialization in list context
905 Currently this is illegal:
907 state ($a, $b) = foo();
909 In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different
910 semantics, which is tricky to implement in Perl 5 as currently they produce
911 the same opcode trees. The Perl 6 design is firm, so it would be good to
912 implement the necessary code in Perl 5. There are comments in
913 C<Perl_newASSIGNOP()> that show the code paths taken by various assignment
914 constructions involving state variables.
916 =head2 Implement $value ~~ 0 .. $range
918 It would be nice to extend the syntax of the C<~~> operator to also
919 understand numeric (and maybe alphanumeric) ranges.
921 =head2 A does() built-in
923 Like ref(), only useful. It would call the C<DOES> method on objects; it
924 would also tell whether something can be dereferenced as an
925 array/hash/etc., or used as a regexp, etc.
926 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
928 =head2 Tied filehandles and write() don't mix
930 There is no method on tied filehandles to allow them to be called back by
933 =head2 Propagate compilation hints to the debugger
935 Currently a debugger started with -dE on the command-line doesn't see the
936 features enabled by -E. More generally hints (C<$^H> and C<%^H>) aren't
937 propagated to the debugger. Probably it would be a good thing to propagate
938 hints from the innermost non-C<DB::> scope: this would make code eval'ed
939 in the debugger see the features (and strictures, etc.) currently in
942 =head2 Attach/detach debugger from running program
944 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
945 program if you pass the process ID. It would be good to do this with the Perl
946 debugger on a running Perl program, although I'm not sure how it would be
947 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
949 =head2 LVALUE functions for lists
951 The old perltodo notes that lvalue functions don't work for list or hash
952 slices. This would be good to fix.
954 =head2 regexp optimiser optional
956 The regexp optimiser is not optional. It should configurable to be, to allow
957 its performance to be measured, and its bugs to be easily demonstrated.
959 =head2 delete &function
961 Allow to delete functions. One can already undef them, but they're still
964 =head2 C</w> regex modifier
966 That flag would enable to match whole words, and also to interpolate
967 arrays as alternations. With it, C</P/w> would be roughly equivalent to:
969 do { local $"='|'; /\b(?:P)\b/ }
971 See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
974 =head2 optional optimizer
976 Make the peephole optimizer optional. Currently it performs two tasks as
977 it walks the optree - genuine peephole optimisations, and necessary fixups of
978 ops. It would be good to find an efficient way to switch out the
979 optimisations whilst keeping the fixups.
981 =head2 You WANT *how* many
983 Currently contexts are void, scalar and list. split has a special mechanism in
984 place to pass in the number of return values wanted. It would be useful to
985 have a general mechanism for this, backwards compatible and little speed hit.
986 This would allow proposals such as short circuiting sort to be implemented
989 =head2 lexical aliases
991 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
993 =head2 entersub XS vs Perl
995 At the moment pp_entersub is huge, and has code to deal with entering both
996 perl and XS subroutines. Subroutine implementations rarely change between
997 perl and XS at run time, so investigate using 2 ops to enter subs (one for
998 XS, one for perl) and swap between if a sub is redefined.
1002 Self-ties are currently illegal because they caused too many segfaults. Maybe
1003 the causes of these could be tracked down and self-ties on all types
1006 =head2 Optimize away @_
1008 The old perltodo notes "Look at the "reification" code in C<av.c>".
1010 =head2 Virtualize operating system access
1012 Implement a set of "vtables" that virtualizes operating system access
1013 (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
1014 least these interfaces should take SVs as "name" arguments instead of
1015 bare char pointers; probably the most flexible and extensible way
1016 would be for the Perl-facing interfaces to accept HVs. The system
1017 needs to be per-operating-system and per-file-system
1018 hookable/filterable, preferably both from XS and Perl level
1019 (L<perlport/"Files and Filesystems"> is good reading at this point,
1020 in fact, all of L<perlport> is.)
1022 This has actually already been implemented (but only for Win32),
1023 take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
1024 variants go through a set of "vtables" for operating system access,
1025 non-Win32 systems currently go straight for the POSIX/UNIX-style
1026 system/library call. Similar system as for Win32 should be
1027 implemented for all platforms. The existing Win32 implementation
1028 probably does not need to survive alongside this proposed new
1029 implementation, the approaches could be merged.
1031 What would this give us? One often-asked-for feature this would
1032 enable is using Unicode for filenames, and other "names" like %ENV,
1033 usernames, hostnames, and so forth.
1034 (See L<perlunicode/"When Unicode Does Not Happen">.)
1036 But this kind of virtualization would also allow for things like
1037 virtual filesystems, virtual networks, and "sandboxes" (though as long
1038 as dynamic loading of random object code is allowed, not very safe
1039 sandboxes since external code of course know not of Perl's vtables).
1040 An example of a smaller "sandbox" is that this feature can be used to
1041 implement per-thread working directories: Win32 already does this.
1043 See also L</"Extend PerlIO and PerlIO::Scalar">.
1045 =head2 Investigate PADTMP hash pessimisation
1047 The peephole optimiser converts constants used for hash key lookups to shared
1048 hash key scalars. Under ithreads, something is undoing this work.
1049 See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html
1051 =head2 Store the current pad in the OP slab allocator
1054 I hope that I got that "current pad" part correct
1056 Currently we leak ops in various cases of parse failure. I suggested that we
1057 could solve this by always using the op slab allocator, and walking it to
1058 free ops. Dave comments that as some ops are already freed during optree
1059 creation one would have to mark which ops are freed, and not double free them
1060 when walking the slab. He notes that one problem with this is that for some ops
1061 you have to know which pad was current at the time of allocation, which does
1062 change. I suggested storing a pointer to the current pad in the memory allocated
1063 for the slab, and swapping to a new slab each time the pad changes. Dave thinks
1064 that this would work.
1066 =head2 repack the optree
1068 Repacking the optree after execution order is determined could allow
1069 removal of NULL ops, and optimal ordering of OPs with respect to cache-line
1070 filling. The slab allocator could be reused for this purpose. I think that
1071 the best way to do this is to make it an optional step just before the
1072 completed optree is attached to anything else, and to use the slab allocator
1073 unchanged, so that freeing ops is identical whether or not this step runs.
1074 Note that the slab allocator allocates ops downwards in memory, so one would
1075 have to actually "allocate" the ops in reverse-execution order to get them
1076 contiguous in memory in execution order.
1078 See http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html
1080 Note that running this copy, and then freeing all the old location ops would
1081 cause their slabs to be freed, which would eliminate possible memory wastage if
1082 the previous suggestion is implemented, and we swap slabs more frequently.
1084 =head2 eliminate incorrect line numbers in warnings
1092 } elsif ($undef == 0) {
1095 used to produce this output:
1097 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1098 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1100 where the line of the second warning was misreported - it should be line 5.
1101 Rafael fixed this - the problem arose because there was no nextstate OP
1102 between the execution of the C<if> and the C<elsif>, hence C<PL_curcop> still
1103 reports that the currently executing line is line 4. The solution was to inject
1104 a nextstate OPs for each C<elsif>, although it turned out that the nextstate
1105 OP needed to be a nulled OP, rather than a live nextstate OP, else other line
1106 numbers became misreported. (Jenga!)
1108 The problem is more general than C<elsif> (although the C<elsif> case is the
1109 most common and the most confusing). Ideally this code
1119 would produce this output
1121 Use of uninitialized value $undef in addition (+) at wrong.pl line 4.
1122 Use of uninitialized value $undef in addition (+) at wrong.pl line 7.
1124 (rather than lines 4 and 5), but this would seem to require every OP to carry
1125 (at least) line number information.
1127 What might work is to have an optional line number in memory just before the
1128 BASEOP structure, with a flag bit in the op to say whether it's present.
1129 Initially during compile every OP would carry its line number. Then add a late
1130 pass to the optimiser (potentially combined with L</repack the optree>) which
1131 looks at the two ops on every edge of the graph of the execution path. If
1132 the line number changes, flags the destination OP with this information.
1133 Once all paths are traced, replace every op with the flag with a
1134 nextstate-light op (that just updates C<PL_curcop>), which in turn then passes
1135 control on to the true op. All ops would then be replaced by variants that
1136 do not store the line number. (Which, logically, why it would work best in
1137 conjunction with L</repack the optree>, as that is already copying/reallocating
1140 (Although I should note that we're not certain that doing this for the general
1143 =head2 optimize tail-calls
1145 Tail-calls present an opportunity for broadly applicable optimization;
1146 anywhere that C<< return foo(...) >> is called, the outer return can
1147 be replaced by a goto, and foo will return directly to the outer
1148 caller, saving (conservatively) 25% of perl's call&return cost, which
1149 is relatively higher than in C. The scheme language is known to do
1150 this heavily. B::Concise provides good insight into where this
1151 optimization is possible, ie anywhere entersub,leavesub op-sequence
1154 perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)'
1156 Bottom line on this is probably a new pp_tailcall function which
1157 combines the code in pp_entersub, pp_leavesub. This should probably
1158 be done 1st in XS, and using B::Generate to patch the new OP into the
1163 Tasks that will get your name mentioned in the description of the "Highlights
1166 =head2 make ithreads more robust
1168 Generally make ithreads more robust. See also L</iCOW>
1170 This task is incremental - even a little bit of work on it will help, and
1171 will be greatly appreciated.
1173 One bit would be to write the missing code in sv.c:Perl_dirp_dup.
1175 Fix Perl_sv_dup, et al so that threads can return objects.
1179 Sarathy and Arthur have a proposal for an improved Copy On Write which
1180 specifically will be able to COW new ithreads. If this can be implemented
1181 it would be a good thing.
1183 =head2 (?{...}) closures in regexps
1185 Fix (or rewrite) the implementation of the C</(?{...})/> closures.
1187 =head2 A re-entrant regexp engine
1189 This will allow the use of a regex from inside (?{ }), (??{ }) and
1190 (?(?{ })|) constructs.
1192 =head2 Add class set operations to regexp engine
1194 Apparently these are quite useful. Anyway, Jeffery Friedl wants them.
1196 demerphq has this on his todo list, but right at the bottom.
1199 =head1 Tasks for microperl
1202 [ Each and every one of these may be obsolete, but they were listed
1203 in the old Todo.micro file]
1206 =head2 make creating uconfig.sh automatic
1208 =head2 make creating Makefile.micro automatic
1210 =head2 do away with fork/exec/wait?
1212 (system, popen should be enough?)
1214 =head2 some of the uconfig.sh really needs to be probed (using cc) in buildtime:
1216 (uConfigure? :-) native datatype widths and endianness come to mind