7 This is a list of wishes for Perl. The most up to date version of this file
8 is at L<http://perl5.git.perl.org/perl.git/blob_plain/HEAD:/Porting/todo.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
19 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/>
21 What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
22 not, but if your patch is incorporated, then we'll add your name to the
23 F<AUTHORS> file, which ships in the official distribution. How many other
24 programming languages offer you 1 line of immortality?
26 =head1 Tasks that need only a little Perl knowledge
28 =head2 Fix POD errors in Perl documentation
30 Perl documentation is furnished in POD (Plain Old Documentation); see
31 L<perlpod>. We also have a utility that checks for various errors in
32 this documentation: F<t/porting/podcheck.t>. Unfortunately many files
33 have errors in them, and there is a database of known problems, kept in
34 F<t/porting/known_pod_issues.dat>. The most prevalent errors are lines
35 too wide to fit in a standard terminal window, but there are more
36 serious problems as well; and there are items listed there that are not
37 in fact errors. The task would be to go through and clean up the
38 documentation. This would be a good way to learn more about Perl.
40 =head1 Tasks that only need Perl knowledge
42 =head2 Classify bug tickets by type
44 Known bugs in Perl are tracked by L<https://rt.perl.org/rt3> (which also
45 includes Perl 6). A summary can be found at
46 L<https://rt.perl.org/rt3/NoAuth/perl5/Overview.html>.
47 It shows bugs classified by "type". However, the type of many of the
48 bugs is "unknown". This greatly lowers the chances of them getting
49 fixed, as the number of open bugs is overwhelming -- too many to wade
50 through for someone to try to find the bugs in the parts of
51 Perl that s/he knows well enough to try to fix. This task involves
52 going through these bugs and classifying them into one or more types.
54 =head2 Ongoing: investigate new bug reports
56 When a bug report is filed, it would be very helpful to have someone do
57 a quick investigation to see if it is a real problem, and to reply to
58 the poster about it, asking for example code that reproduces the
59 problem. Such code should be added to the test suite as TODO tests, and
60 the ticket should be classified by type. To get started on this task,
61 look at the tickets that are marked as "New Issues" in
62 L<https://rt.perl.org/rt3/NoAuth/perl5/Overview.html>.
64 =head2 Migrate t/ from custom TAP generation
66 Many tests below F<t/> still generate TAP by "hand", rather than using library
67 functions. As explained in L<perlhack/TESTING>, tests in F<t/> are
68 written in a particular way to test that more complex constructions actually
69 work before using them routinely. Hence they don't use C<Test::More>, but
70 instead there is an intentionally simpler library, F<t/test.pl>. However,
71 quite a few tests in F<t/> have not been refactored to use it. Refactoring
72 any of these tests, one at a time, is a useful thing TODO.
74 The subdirectories F<base>, F<cmd> and F<comp>, that contain the most
75 basic tests, should be excluded from this task.
77 =head2 Automate perldelta generation
79 The perldelta file accompanying each release summaries the major changes.
80 It's mostly manually generated currently, but some of that could be
81 automated with a bit of perl, specifically the generation of
85 =item Modules and Pragmata
87 =item New Documentation
93 See F<Porting/how_to_write_a_perldelta.pod> for details.
95 =head2 Make Schwern poorer
97 We should have tests for everything. When all the core's modules are tested,
98 Schwern has promised to donate to $500 to TPF. We may need volunteers to
99 hold him upside down and shake vigorously in order to actually extract the
102 =head2 Write descriptions for all tests
104 Many individual tests in the test suite lack descriptions (or names, or labels
105 -- call them what you will). Many files completely lack descriptions, meaning
106 that the only output you get is the test numbers. If all tests had
107 descriptions, understanding what the tests are testing and why they sometimes
108 fail would both get a whole lot easier.
110 =head2 Improve the coverage of the core tests
112 Use Devel::Cover to ascertain the core modules' test coverage, then add
113 tests that are currently missing.
117 A full test suite for the B module would be nice.
119 =head2 A decent benchmark
121 C<perlbench> seems impervious to any recent changes made to the perl core. It
122 would be useful to have a reasonable general benchmarking suite that roughly
123 represented what current perl programs do, and measurably reported whether
124 tweaks to the core improve, degrade or don't really affect performance, to
125 guide people attempting to optimise the guts of perl. Gisle would welcome
126 new tests for perlbench. Steffen Schwingon would welcome help with
127 L<Benchmark::Perl::Formance>
129 =head2 fix tainting bugs
131 Fix the bugs revealed by running the test suite with the C<-t> switch (via
132 C<make test.taintwarn>).
134 =head2 Dual life everything
136 As part of the "dists" plan, anything that doesn't belong in the smallest perl
137 distribution needs to be dual lifed. Anything else can be too. Figure out what
138 changes would be needed to package that module and its tests up for CPAN, and
139 do so. Test it with older perl releases, and fix the problems you find.
141 To make a minimal perl distribution, it's useful to look at
142 F<t/lib/commonsense.t>.
144 =head2 POSIX memory footprint
146 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
147 various times worked to cut it down. There is probably still fat to cut out -
148 for example POSIX passes Exporter some very memory hungry data structures.
150 =head2 makedef.pl and conditional compilation
152 The script F<makedef.pl> that generates the list of exported symbols on
153 platforms which need this. Functions are declared in F<embed.fnc>, variables
154 in F<intrpvar.h>. Quite a few of the functions and variables are conditionally
155 declared there, using C<#ifdef>. However, F<makedef.pl> doesn't understand the
156 C macros, so the rules about which symbols are present when is duplicated in
157 the Perl code. Writing things twice is bad, m'kay. It would be good to teach
158 F<.pl> to understand the conditional compilation, and hence remove the
159 duplication, and the mistakes it has caused.
161 =head2 use strict; and AutoLoad
163 Currently if you write
166 use AutoLoader 'AUTOLOAD';
171 print join (' ', No, strict, here), "!\n";
174 then C<use strict;> isn't in force within the autoloaded subroutines. It would
175 be more consistent (and less surprising) to arrange for all lexical pragmas
176 in force at the __END__ block to be in force within each autoloaded subroutine.
178 There's a similar problem with SelfLoader.
180 =head2 profile installman
182 The F<installman> script is slow. All it is doing text processing, which we're
183 told is something Perl is good at. So it would be nice to know what it is doing
184 that is taking so much CPU, and where possible address it.
186 =head2 enable lexical enabling/disabling of individual warnings
188 Currently, warnings can only be enabled or disabled by category. There
189 are times when it would be useful to quash a single warning, not a
192 =head2 document diagnostics
194 Many diagnostic messages are not currently documented. The list is at the end
197 =head2 Write TODO tests for open bugs
199 Sometimes bugs get fixed as a side effect of something else, and
200 the bug remains open because no one realizes that it has been fixed.
201 Ideally, every open bug should have a TODO test in the core test suite.
203 =head1 Tasks that need a little sysadmin-type knowledge
205 Or if you prefer, tasks that you would learn from, and broaden your skills
208 =head2 make HTML install work
210 There is an C<install.html> target in the Makefile. It's marked as
211 "experimental". It would be good to get this tested, make it work reliably, and
212 remove the "experimental" tag. This would include
218 Checking that cross linking between various parts of the documentation works.
219 In particular that links work between the modules (files with POD in F<lib/>)
220 and the core documentation (files in F<pod/>)
224 Improving the code that split C<perlfunc> into chunks, preferably with
225 general case code added to L<Pod::Functions> that could be used elsewhere.
227 Challenges here are correctly identifying the groups of functions that go
228 together, and making the right named external cross-links point to the right
229 page. Currently this works reasonably well in the general case, and correctly
230 parses two or more C<=items> giving the different parameter lists for the
231 same function, such used by C<substr>. However it fails completely where
232 I<different> functions are listed as a sequence of C<=items> but share the
233 same description. All the functions from C<getpwnam> to C<endprotoent> have
234 individual stub pages, with only the page for C<endservent> holding the
235 description common to all. Likewise C<q>, C<qq> and C<qw> have stub pages,
236 instead of sharing the body of C<qx>.
238 Note also the current code isn't ideal with the two forms of C<select>, mushing
239 them both into one F<select.html> with the two descriptions run together.
240 Fixing this may well be a special case.
244 =head2 compressed man pages
246 Be able to install them. This would probably need a configure test to see how
247 the system does compressed man pages (same directory/different directory?
248 same filename/different filename), as well as tweaking the F<installman> script
249 to compress as necessary.
251 =head2 Add a code coverage target to the Makefile
253 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
254 to do this manually are roughly
260 do a normal C<Configure>, but include Devel::Cover as a module to install
261 (see L<INSTALL> for how to do this)
269 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
273 Process the resulting Devel::Cover database
277 This just give you the coverage of the F<.pm>s. To also get the C level
284 Additionally tell C<Configure> to use the appropriate C compiler flags for
291 (instead of C<make perl>)
295 After running the tests run C<gcov> to generate all the F<.gcov> files.
296 (Including down in the subdirectories of F<ext/>
300 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
301 to get their stats into the cover_db directory.
305 Then process the Devel::Cover database
309 It would be good to add a single switch to C<Configure> to specify that you
310 wanted to perform perl level coverage, and another to specify C level
311 coverage, and have C<Configure> and the F<Makefile> do all the right things
314 =head2 Make Config.pm cope with differences between built and installed perl
316 Quite often vendors ship a perl binary compiled with their (pay-for)
317 compilers. People install a free compiler, such as gcc. To work out how to
318 build extensions, Perl interrogates C<%Config>, so in this situation
319 C<%Config> describes compilers that aren't there, and extension building
320 fails. This forces people into choosing between re-compiling perl themselves
321 using the compiler they have, or only using modules that the vendor ships.
323 It would be good to find a way teach C<Config.pm> about the installation setup,
324 possibly involving probing at install time or later, so that the C<%Config> in
325 a binary distribution better describes the installed machine, when the
326 installed machine differs from the build machine in some significant way.
328 =head2 linker specification files
330 Some platforms mandate that you provide a list of a shared library's external
331 symbols to the linker, so the core already has the infrastructure in place to
332 do this for generating shared perl libraries. Florian Ragwitz has been working
333 to offer this for the GNU toolchain, to allow Unix users to test that the
334 export list is correct, and to build a perl that does not pollute the global
335 namespace with private symbols, and will fail in the same way as msvc or mingw
336 builds or when using PERL_DL_NONLAZY=1. See the branch smoke-me/rafl/ld_export
338 =head2 Cross-compile support
340 We get requests for "how to cross compile Perl". The vast majority of these
341 seem to be for a couple of scenarios:
347 Platforms that could build natively using F<./Configure> (I<e.g.> Linux or
348 NetBSD on MIPS or ARM) but people want to use a beefier machine (and on the
349 same OS) to build more easily.
353 Platforms that can't build natively, but no (significant) porting changes
354 are needed to our current source code. Prime example of this is Android.
358 There are several scripts and tools for cross-compiling perl for other
359 platforms. However, these are somewhat inconsistent and scattered across the
360 codebase, none are documented well, none are clearly flexible enough to
361 be confident that they can support any TARGET/HOST platform pair other than
362 that which they were developed on, and it's not clear how bitrotted they are.
364 For example, C<Configure> understands C<-Dusecrosscompile> option. This option
365 arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is
366 assumed then to be copied to TARGET machine and used as a replacement of
367 full C<perl> executable. This code is almost 10 years old. Meanwhile, the
368 F<Cross/> directory contains two different approaches for cross compiling to
369 ARM Linux targets, relying on hand curated F<config.sh> files, but that code
370 is getting on for 5 years old, and requires insider knowledge of perl's
371 build system to draft a F<config.sh> for a new platform.
373 Jess Robinson has submitted a grant to TPF to work on cleaning this up.
375 =head2 Split "linker" from "compiler"
377 Right now, Configure probes for two commands, and sets two variables:
381 =item * C<cc> (in F<cc.U>)
383 This variable holds the name of a command to execute a C compiler which
384 can resolve multiple global references that happen to have the same
385 name. Usual values are F<cc> and F<gcc>.
386 Fervent ANSI compilers may be called F<c89>. AIX has F<xlc>.
388 =item * C<ld> (in F<dlsrc.U>)
390 This variable indicates the program to be used to link
391 libraries for dynamic loading. On some systems, it is F<ld>.
392 On ELF systems, it should be C<$cc>. Mostly, we'll try to respect
393 the hint file setting.
397 There is an implicit historical assumption from around Perl5.000alpha
398 something, that C<$cc> is also the correct command for linking object files
399 together to make an executable. This may be true on Unix, but it's not true
400 on other platforms, and there are a maze of work arounds in other places (such
401 as F<Makefile.SH>) to cope with this.
403 Ideally, we should create a new variable to hold the name of the executable
404 linker program, probe for it in F<Configure>, and centralise all the special
405 case logic there or in hints files.
407 A small bikeshed issue remains - what to call it, given that C<$ld> is already
408 taken (arguably for the wrong thing now, but on SunOS 4.1 it is the command
409 for creating dynamically-loadable modules) and C<$link> could be confused with
410 the Unix command line executable of the same name, which does something
411 completely different. Andy Dougherty makes the counter argument "In parrot, I
412 tried to call the command used to link object files and libraries into an
413 executable F<link>, since that's what my vaguely-remembered DOS and VMS
414 experience suggested. I don't think any real confusion has ensued, so it's
415 probably a reasonable name for perl5 to use."
417 "Alas, I've always worried that introducing it would make things worse,
418 since now the module building utilities would have to look for
419 C<$Config{link}> and institute a fall-back plan if it weren't found."
420 Although I can see that as confusing, given that C<$Config{d_link}> is true
421 when (hard) links are available.
423 =head2 Configure Windows using PowerShell
425 Currently, Windows uses hard-coded config files based to build the
426 config.h for compiling Perl. Makefiles are also hard-coded and need to be
427 hand edited prior to building Perl. While this makes it easy to create a perl.exe
428 that works across multiple Windows versions, being able to accurately
429 configure a perl.exe for a specific Windows versions and VS C++ would be
430 a nice enhancement. With PowerShell available on Windows XP and up, this
431 may now be possible. Step 1 might be to investigate whether this is possible
432 and use this to clean up our current makefile situation. Step 2 would be to
433 see if there would be a way to use our existing metaconfig units to configure a
434 Windows Perl or whether we go in a separate direction and make it so. Of
435 course, we all know what step 3 is.
437 =head1 Tasks that need a little C knowledge
439 These tasks would need a little C knowledge, but don't need any specific
440 background or experience with XS, or how the Perl interpreter works
442 =head2 Weed out needless PERL_UNUSED_ARG
444 The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about
445 unused arguments. Often the arguments can't be removed, as there is an
446 external constraint that determines the prototype of the function, so this
447 approach is valid. However, there are some cases where C<PERL_UNUSED_ARG>
448 could be removed. Specifically
454 The prototypes of (nearly all) static functions can be changed
458 Unused arguments generated by short cut macros are wasteful - the short cut
459 macro used can be changed.
465 Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall.
466 On these systems, it might be the default compilation mode, and there
467 is currently no guarantee that passing no use64bitall option to the
468 Configure process will build a 32bit perl. Implementing -Duse32bit*
469 options would be nice for perl 5.19.1.
471 =head2 Profile Perl - am I hot or not?
473 The Perl source code is stable enough that it makes sense to profile it,
474 identify and optimise the hotspots. It would be good to measure the
475 performance of the Perl interpreter using free tools such as cachegrind,
476 gprof, and dtrace, and work to reduce the bottlenecks they reveal.
478 As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops,
479 the ops that are most commonly used. The idea is that by grouping them, their
480 object code will be adjacent in the executable, so they have a greater chance
481 of already being in the CPU cache (or swapped in) due to being near another op
484 Except that it's not clear if these really are the most commonly used ops. So
485 as part of exercising your skills with coverage and profiling tools you might
486 want to determine what ops I<really> are the most commonly used. And in turn
487 suggest evictions and promotions to achieve a better F<pp_hot.c>.
489 One piece of Perl code that might make a good testbed is F<installman>.
491 =head2 Improve win32/wince.c
493 Currently, numerous functions look virtually, if not completely,
494 identical in both F<win32/wince.c> and F<win32/win32.c> files, which can't
497 =head2 Use secure CRT functions when building with VC8 on Win32
499 Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis
500 that they were "unsafe" and introduced differently named secure versions of
501 them as replacements, e.g. instead of writing
503 FILE* f = fopen(__FILE__, "r");
508 errno_t err = fopen_s(&f, __FILE__, "r");
510 Currently, the warnings about these deprecations have been disabled by adding
511 -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that
512 warning suppressant and actually make use of the new secure CRT functions.
514 There is also a similar issue with POSIX CRT function names like fileno having
515 been deprecated in favour of ISO C++ conformant names like _fileno. These
516 warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It
517 might be nice to do as Microsoft suggest here too, although, unlike the secure
518 functions issue, there is presumably little or no benefit in this case.
520 =head2 Fix POSIX::access() and chdir() on Win32
522 These functions currently take no account of DACLs and therefore do not behave
523 correctly in situations where access is restricted by DACLs (as opposed to the
524 read-only attribute).
526 Furthermore, POSIX::access() behaves differently for directories having the
527 read-only attribute set depending on what CRT library is being used. For
528 example, the _access() function in the VC6 and VC7 CRTs (wrongly) claim that
529 such directories are not writable, whereas in fact all directories are writable
530 unless access is denied by DACLs. (In the case of directories, the read-only
531 attribute actually only means that the directory cannot be deleted.) This CRT
532 bug is fixed in the VC8 and VC9 CRTs (but, of course, the directory may still
533 not actually be writable if access is indeed denied by DACLs).
535 For the chdir() issue, see ActiveState bug #74552:
536 L<http://bugs.activestate.com/show_bug.cgi?id=74552>
538 Therefore, DACLs should be checked both for consistency across CRTs and for
541 (Note that perl's -w operator should not be modified to check DACLs. It has
542 been written so that it reflects the state of the read-only attribute, even
543 for directories (whatever CRT is being used), for symmetry with chmod().)
545 =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf()
547 Maybe create a utility that checks after each libperl.a creation that
548 none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets())
549 ever creep back to libperl.a.
551 nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/;
552 print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/'
554 Note, of course, that this will only tell whether B<your> platform
555 is using those naughty interfaces.
557 =head2 -D_FORTIFY_SOURCE=2
559 Recent glibcs support C<-D_FORTIFY_SOURCE=2> which gives
560 protection against various kinds of buffer overflow problems.
561 It should probably be used for compiling Perl whenever available,
562 Configure and/or hints files should be adjusted to probe for the
563 availability of these feature and enable it as appropriate.
565 =head2 Arenas for GPs? For MAGIC?
567 C<struct gp> and C<struct magic> are both currently allocated by C<malloc>.
568 It might be a speed or memory saving to change to using arenas. Or it might
569 not. It would need some suitable benchmarking first. In particular, C<GP>s
570 can probably be changed with minimal compatibility impact (probably nothing
571 outside of the core, or even outside of F<gv.c> allocates them), but they
572 probably aren't allocated/deallocated often enough for a speed saving. Whereas
573 C<MAGIC> is allocated/deallocated more often, but in turn, is also something
574 more externally visible, so changing the rules here may bite external code.
578 Several SV body structs are now the same size, notably PVMG and PVGV, PVAV and
579 PVHV, and PVCV and PVFM. It should be possible to allocate and return same
580 sized bodies from the same actual arena, rather than maintaining one arena for
581 each. This could save 4-6K per thread, of memory no longer tied up in the
582 not-yet-allocated part of an arena.
585 =head1 Tasks that need a knowledge of XS
587 These tasks would need C knowledge, and roughly the level of knowledge of
588 the perl API that comes from writing modules that use XS to interface to
591 =head2 Write an XS cookbook
593 Create pod/perlxscookbook.pod with short, task-focused 'recipes' in XS that
594 demonstrate common tasks and good practices. (Some of these might be
595 extracted from perlguts.) The target audience should be XS novices, who need
596 more examples than perlguts but something less overwhelming than perlapi.
597 Recipes should provide "one pretty good way to do it" instead of TIMTOWTDI.
599 Rather than focusing on interfacing Perl to C libraries, such a cookbook
600 should probably focus on how to optimize Perl routines by re-writing them
601 in XS. This will likely be more motivating to those who mostly work in
602 Perl but are looking to take the next step into XS.
604 Deconstructing and explaining some simpler XS modules could be one way to
605 bootstrap a cookbook. (List::Util? Class::XSAccessor? Tree::Ternary_XS?)
606 Another option could be deconstructing the implementation of some simpler
609 =head2 Document how XSUBs can use C<cv_set_call_checker> to inline themselves as OPs
611 For a simple XSUB, often the subroutine dispatch takes more time than the
612 XSUB itself. v5.14.0 now allows XSUBs to register a function which will be
613 called when the parser is finished building an C<entersub> op which calls
616 Registration is done with C<Perl_cv_set_call_checker>, is documented at the
617 API level in L<perlapi>, and L<perl5140delta/Custom per-subroutine check hooks>
618 notes that it can be used to inline a subroutine, by replacing it with a
619 custom op. However there is no further detail of the code needed to do this.
620 It would be useful to add one or more annotated examples of how to create
623 This should provide a measurable speed up to simple XSUBs inside
624 tight loops. Initially one would have to write the OP alternative
625 implementation by hand, but it's likely that this should be reasonably
626 straightforward for the type of XSUB that would benefit the most. Longer
627 term, once the run-time implementation is proven, it should be possible to
628 progressively update ExtUtils::ParseXS to generate OP implementations for
631 =head2 Remove the use of SVs as temporaries in dump.c
633 F<dump.c> contains debugging routines to dump out the contains of perl data
634 structures, such as C<SV>s, C<AV>s and C<HV>s. Currently, the dumping code
635 B<uses> C<SV>s for its temporary buffers, which was a logical initial
636 implementation choice, as they provide ready made memory handling.
638 However, they also lead to a lot of confusion when it happens that what you're
639 trying to debug is seen by the code in F<dump.c>, correctly or incorrectly, as
640 a temporary scalar it can use for a temporary buffer. It's also not possible
641 to dump scalars before the interpreter is properly set up, such as during
642 ithreads cloning. It would be good to progressively replace the use of scalars
643 as string accumulation buffers with something much simpler, directly allocated
644 by C<malloc>. The F<dump.c> code is (or should be) only producing 7 bit
645 US-ASCII, so output character sets are not an issue.
647 Producing and proving an internal simple buffer allocation would make it easier
648 to re-write the internals of the PerlIO subsystem to avoid using C<SV>s for
649 B<its> buffers, use of which can cause problems similar to those of F<dump.c>,
652 =head2 safely supporting POSIX SA_SIGINFO
654 Some years ago Jarkko supplied patches to provide support for the POSIX
655 SA_SIGINFO feature in Perl, passing the extra data to the Perl signal handler.
657 Unfortunately, it only works with "unsafe" signals, because under safe
658 signals, by the time Perl gets to run the signal handler, the extra
659 information has been lost. Moreover, it's not easy to store it somewhere,
660 as you can't call mutexs, or do anything else fancy, from inside a signal
663 So it strikes me that we could provide safe SA_SIGINFO support
669 Provide global variables for two file descriptors
673 When the first request is made via C<sigaction> for C<SA_SIGINFO>, create a
674 pipe, store the reader in one, the writer in the other
678 In the "safe" signal handler (C<Perl_csighandler()>/C<S_raise_signal()>), if
679 the C<siginfo_t> pointer non-C<NULL>, and the writer file handle is open,
685 serialise signal number, C<struct siginfo_t> (or at least the parts we care
686 about) into a small auto char buff
690 C<write()> that (non-blocking) to the writer fd
696 if it writes 100%, flag the signal in a counter of "signals on the pipe" akin
697 to the current per-signal-number counts
701 if it writes 0%, assume the pipe is full. Flag the data as lost?
705 if it writes partially, croak a panic, as your OS is broken.
713 in the regular C<PERL_ASYNC_CHECK()> processing, if there are "signals on
714 the pipe", read the data out, deserialise, build the Perl structures on
715 the stack (code in C<Perl_sighandler()>, the "unsafe" handler), and call as
720 I think that this gets us decent C<SA_SIGINFO> support, without the current risk
721 of running Perl code inside the signal handler context. (With all the dangers
722 of things like C<malloc> corruption that that currently offers us)
724 For more information see the thread starting with this message:
725 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-03/msg00305.html>
727 =head2 autovivification
729 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
731 This task is incremental - even a little bit of work on it will help.
733 =head2 Unicode in Filenames
735 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
736 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
737 system, truncate, unlink, utime, -X. All these could potentially accept
738 Unicode filenames either as input or output (and in the case of system
739 and qx Unicode in general, as input or output to/from the shell).
740 Whether a filesystem - an operating system pair understands Unicode in
743 Known combinations that have some level of understanding include
744 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
745 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
746 create Unicode filenames, what forms of Unicode are accepted and used
747 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
748 and so on, varies. Finding the right level of interfacing to Perl
749 requires some thought. Remember that an OS does not implicate a
752 (The Windows -C command flag "wide API support" has been at least
753 temporarily retired in 5.8.1, and the -C has been repurposed, see
756 Most probably the right way to do this would be this:
757 L</"Virtualize operating system access">.
759 =head2 Unicode in %ENV
761 Currently the %ENV entries are always byte strings.
762 See L</"Virtualize operating system access">.
764 (See RT ticket #113536 for information on Win32's handling of %ENV,
765 which was fixed to work with native ANSI codepage characters in the
766 environment, but still doesn't work with other characters outside of
767 that codepage present in the environment.)
769 =head2 Unicode and glob()
771 Currently glob patterns and filenames returned from File::Glob::glob()
772 are always byte strings. See L</"Virtualize operating system access">.
774 =head2 use less 'memory'
776 Investigate trade offs to switch out perl's choices on memory usage.
777 Particularly perl should be able to give memory back.
779 This task is incremental - even a little bit of work on it will help.
781 =head2 Re-implement C<:unique> in a way that is actually thread-safe
783 The old implementation made bad assumptions on several levels. A good 90%
784 solution might be just to make C<:unique> work to share the string buffer
785 of SvPVs. That way large constant strings can be shared between ithreads,
786 such as the configuration information in F<Config>.
788 =head2 Make tainting consistent
790 Tainting would be easier to use if it didn't take documented shortcuts and
791 allow taint to "leak" everywhere within an expression.
793 =head2 readpipe(LIST)
795 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
796 running a shell. readpipe() (the function behind qx//) could be similarly
799 =head2 Audit the code for destruction ordering assumptions
803 /* Need to check SvMAGICAL, as during global destruction it may be that
804 AvARYLEN(av) has been freed before av, and hence the SvANY() pointer
805 is now part of the linked list of SV heads, rather than pointing to
806 the original body. */
807 /* FIXME - audit the code for other bugs like this one. */
809 adding the C<SvMAGICAL> check to
811 if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) {
812 MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen);
814 Go through the core and look for similar assumptions that SVs have particular
815 types, as all bets are off during global destruction.
817 =head2 Extend PerlIO and PerlIO::Scalar
819 PerlIO::Scalar doesn't know how to truncate(). Implementing this
820 would require extending the PerlIO vtable.
822 Similarly the PerlIO vtable doesn't know about formats (write()), or
823 about stat(), or chmod()/chown(), utime(), or flock().
825 (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership
828 PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(),
829 opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(),
832 See also L</"Virtualize operating system access">.
834 =head2 Organize error messages
836 Perl's diagnostics (error messages, see L<perldiag>) could use
837 reorganizing and formalizing so that each error message has its
838 stable-for-all-eternity unique id, categorized by severity, type, and
839 subsystem. (The error messages would be listed in a datafile outside
840 of the Perl source code, and the source code would only refer to the
841 messages by the id.) This clean-up and regularizing should apply
842 for all croak() messages.
844 This would enable all sorts of things: easier translation/localization
845 of the messages (though please do keep in mind the caveats of
846 L<Locale::Maketext> about too straightforward approaches to
847 translation), filtering by severity, and instead of grepping for a
848 particular error message one could look for a stable error id. (Of
849 course, changing the error messages by default would break all the
850 existing software depending on some particular error message...)
852 This kind of functionality is known as I<message catalogs>. Look for
853 inspiration for example in the catgets() system, possibly even use it
854 if available-- but B<only> if available, all platforms will B<not>
857 For the really pure at heart, consider extending this item to cover
858 also the warning messages (see L<perllexwarn>, C<warnings.pl>).
860 =head1 Tasks that need a knowledge of the interpreter
862 These tasks would need C knowledge, and knowledge of how the interpreter works,
863 or a willingness to learn.
865 =head2 forbid labels with keyword names
867 Currently C<goto keyword> "computes" the label value:
869 $ perl -e 'goto print'
870 Can't find label 1 at -e line 1.
872 It is controversial if the right way to avoid the confusion is to forbid
873 labels with keyword names, or if it would be better to always treat
874 bareword expressions after a "goto" as a label and never as a keyword.
876 =head2 truncate() prototype
878 The prototype of truncate() is currently C<$$>. It should probably
879 be C<*$> instead. (This is changed in F<opcode.pl>)
881 =head2 error reporting of [$a ; $b]
883 Using C<;> inside brackets is a syntax error, and we don't propose to change
884 that by giving it any meaning. However, it's not reported very helpfully:
886 $ perl -e '$a = [$b; $c];'
887 syntax error at -e line 1, near "$b;"
888 syntax error at -e line 1, near "$c]"
889 Execution of -e aborted due to compilation errors.
891 It should be possible to hook into the tokeniser or the lexer, so that when a
892 C<;> is parsed where it is not legal as a statement terminator (ie inside
893 C<{}> used as a hashref, C<[]> or C<()>) it issues an error something like
894 I<';' isn't legal inside an expression - if you need multiple statements use a
895 do {...} block>. See the thread starting at
896 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-09/msg00573.html>
898 =head2 strict as warnings
900 See L<http://markmail.org/message/vbrupaslr3bybmvk>, where Josua ben Jore
901 writes: I've been of the opinion that everything strict.pm does ought to be
902 able to considered just warnings that have been promoted to 'FATAL'.
904 =head2 lexicals used only once
908 $ perl -we '$pie = 42'
909 Name "main::pie" used only once: possible typo at -e line 1.
913 $ perl -we 'my $pie = 42'
915 Logically all lexicals used only once should warn, if the user asks for
916 warnings. An unworked RT ticket (#5087) has been open for almost seven
917 years for this discrepancy.
921 The handling of Unicode is unclean in many places. In the regex engine
922 there are especially many problems. The swash data structure could be
923 replaced my something better. Inversion lists and maps are likely
924 candidates. The whole Unicode database could be placed in-core for a
925 huge speed-up. Only minimal work was done on the optimizer when utf8
926 was added, with the result that the synthetic start class often will
927 fail to narrow down the possible choices when given non-Latin1 input.
928 Karl Williamson has been working on this - talk to him.
930 =head2 state variable initialization in list context
932 Currently this is illegal:
934 state ($a, $b) = foo();
936 In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different
937 semantics, which is tricky to implement in Perl 5 as currently they produce
938 the same opcode trees. The Perl 6 design is firm, so it would be good to
939 implement the necessary code in Perl 5. There are comments in
940 C<Perl_newASSIGNOP()> that show the code paths taken by various assignment
941 constructions involving state variables.
943 =head2 A does() built-in
945 Like ref(), only useful. It would call the C<DOES> method on objects; it
946 would also tell whether something can be dereferenced as an
947 array/hash/etc., or used as a regexp, etc.
948 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html>
950 =head2 Tied filehandles and write() don't mix
952 There is no method on tied filehandles to allow them to be called back by
955 =head2 Propagate compilation hints to the debugger
957 Currently a debugger started with -dE on the command-line doesn't see the
958 features enabled by -E. More generally hints (C<$^H> and C<%^H>) aren't
959 propagated to the debugger. Probably it would be a good thing to propagate
960 hints from the innermost non-C<DB::> scope: this would make code eval'ed
961 in the debugger see the features (and strictures, etc.) currently in
964 =head2 Attach/detach debugger from running program
966 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
967 program if you pass the process ID. It would be good to do this with the Perl
968 debugger on a running Perl program, although I'm not sure how it would be
969 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
971 =head2 LVALUE functions for lists
973 The old perltodo notes that lvalue functions don't work for list or hash
974 slices. This would be good to fix.
976 =head2 regexp optimizer optional
978 The regexp optimizer is not optional. It should be configurable to be optional
979 and to allow its performance to be measured and its bugs to be easily
982 =head2 C</w> regex modifier
984 That flag would enable to match whole words, and also to interpolate
985 arrays as alternations. With it, C</P/w> would be roughly equivalent to:
987 do { local $"='|'; /\b(?:P)\b/ }
990 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html>
993 =head2 optional optimizer
995 Make the peephole optimizer optional. Currently it performs two tasks as
996 it walks the optree - genuine peephole optimisations, and necessary fixups of
997 ops. It would be good to find an efficient way to switch out the
998 optimisations whilst keeping the fixups.
1000 =head2 You WANT *how* many
1002 Currently contexts are void, scalar and list. split has a special mechanism in
1003 place to pass in the number of return values wanted. It would be useful to
1004 have a general mechanism for this, backwards compatible and little speed hit.
1005 This would allow proposals such as short circuiting sort to be implemented
1006 as a module on CPAN.
1008 =head2 lexical aliases
1010 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>).
1014 Self-ties are currently illegal because they caused too many segfaults. Maybe
1015 the causes of these could be tracked down and self-ties on all types
1018 =head2 Optimize away @_
1020 The old perltodo notes "Look at the "reification" code in C<av.c>".
1022 =head2 Virtualize operating system access
1024 Implement a set of "vtables" that virtualizes operating system access
1025 (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very
1026 least these interfaces should take SVs as "name" arguments instead of
1027 bare char pointers; probably the most flexible and extensible way
1028 would be for the Perl-facing interfaces to accept HVs. The system
1029 needs to be per-operating-system and per-file-system
1030 hookable/filterable, preferably both from XS and Perl level
1031 (L<perlport/"Files and Filesystems"> is good reading at this point,
1032 in fact, all of L<perlport> is.)
1034 This has actually already been implemented (but only for Win32),
1035 take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32
1036 variants go through a set of "vtables" for operating system access,
1037 non-Win32 systems currently go straight for the POSIX/Unix-style
1038 system/library call. Similar system as for Win32 should be
1039 implemented for all platforms. The existing Win32 implementation
1040 probably does not need to survive alongside this proposed new
1041 implementation, the approaches could be merged.
1043 What would this give us? One often-asked-for feature this would
1044 enable is using Unicode for filenames, and other "names" like %ENV,
1045 usernames, hostnames, and so forth.
1046 (See L<perlunicode/"When Unicode Does Not Happen">.)
1048 But this kind of virtualization would also allow for things like
1049 virtual filesystems, virtual networks, and "sandboxes" (though as long
1050 as dynamic loading of random object code is allowed, not very safe
1051 sandboxes since external code of course know not of Perl's vtables).
1052 An example of a smaller "sandbox" is that this feature can be used to
1053 implement per-thread working directories: Win32 already does this.
1055 See also L</"Extend PerlIO and PerlIO::Scalar">.
1057 =head2 repack the optree
1059 Repacking the optree after execution order is determined could allow
1060 removal of NULL ops, and optimal ordering of OPs with respect to cache-line
1061 filling. I think that
1062 the best way to do this is to make it an optional step just before the
1063 completed optree is attached to anything else, and to use the slab allocator
1064 unchanged--but allocate a single slab the right size, avoiding partial
1065 slabs--, so that freeing ops is identical whether or not this step runs.
1066 Note that the slab allocator allocates ops downwards in memory, so one would
1067 have to actually "allocate" the ops in reverse-execution order to get them
1068 contiguous in memory in execution order.
1071 L<http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html>
1073 Note that running this copy, and then freeing all the old location ops would
1074 cause their slabs to be freed, which would eliminate possible memory wastage if
1075 the previous suggestion is implemented, and we swap slabs more frequently.
1077 =head2 eliminate incorrect line numbers in warnings
1085 } elsif ($undef == 0) {
1088 used to produce this output:
1090 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1091 Use of uninitialized value in numeric eq (==) at wrong.pl line 4.
1093 where the line of the second warning was misreported - it should be line 5.
1094 Rafael fixed this - the problem arose because there was no nextstate OP
1095 between the execution of the C<if> and the C<elsif>, hence C<PL_curcop> still
1096 reports that the currently executing line is line 4. The solution was to inject
1097 a nextstate OPs for each C<elsif>, although it turned out that the nextstate
1098 OP needed to be a nulled OP, rather than a live nextstate OP, else other line
1099 numbers became misreported. (Jenga!)
1101 The problem is more general than C<elsif> (although the C<elsif> case is the
1102 most common and the most confusing). Ideally this code
1112 would produce this output
1114 Use of uninitialized value $undef in addition (+) at wrong.pl line 4.
1115 Use of uninitialized value $undef in addition (+) at wrong.pl line 7.
1117 (rather than lines 4 and 5), but this would seem to require every OP to carry
1118 (at least) line number information.
1120 What might work is to have an optional line number in memory just before the
1121 BASEOP structure, with a flag bit in the op to say whether it's present.
1122 Initially during compile every OP would carry its line number. Then add a late
1123 pass to the optimizer (potentially combined with L</repack the optree>) which
1124 looks at the two ops on every edge of the graph of the execution path. If
1125 the line number changes, flags the destination OP with this information.
1126 Once all paths are traced, replace every op with the flag with a
1127 nextstate-light op (that just updates C<PL_curcop>), which in turn then passes
1128 control on to the true op. All ops would then be replaced by variants that
1129 do not store the line number. (Which, logically, why it would work best in
1130 conjunction with L</repack the optree>, as that is already copying/reallocating
1133 (Although I should note that we're not certain that doing this for the general
1136 =head2 optimize tail-calls
1138 Tail-calls present an opportunity for broadly applicable optimization;
1139 anywhere that C<< return foo(...) >> is called, the outer return can
1140 be replaced by a goto, and foo will return directly to the outer
1141 caller, saving (conservatively) 25% of perl's call&return cost, which
1142 is relatively higher than in C. The scheme language is known to do
1143 this heavily. B::Concise provides good insight into where this
1144 optimization is possible, ie anywhere entersub,leavesub op-sequence
1147 perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)'
1149 Bottom line on this is probably a new pp_tailcall function which
1150 combines the code in pp_entersub, pp_leavesub. This should probably
1151 be done 1st in XS, and using B::Generate to patch the new OP into the
1154 =head2 Add C<0odddd>
1156 It has been proposed that octal constants be specifiable through the syntax
1157 C<0oddddd>, parallel to the existing construct to specify hex constants
1160 =head2 Revisit the regex super-linear cache code
1162 Perl executes regexes using the traditional backtracking algorithm, which
1163 makes it possible to implement a variety of powerful pattern-matching
1164 features (like embedded code blocks), at the cost of taking exponential time
1165 to run on some pathological patterns. The exponential-time problem is
1166 mitigated by the I<super-linear cache>, which detects when we're processing
1167 such a pathological pattern, and does some additional bookkeeping to avoid
1168 much of the work. However, that code has bit-rotted a little; some patterns
1169 don't make as much use of it as they should. The proposal is to analyse
1170 where the current cache code has problems, and extend it to cover those cases.
1173 L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2013-01/msg00339.html>
1177 Tasks that will get your name mentioned in the description of the "Highlights
1180 =head2 make ithreads more robust
1182 Generally make ithreads more robust.
1184 This task is incremental - even a little bit of work on it will help, and
1185 will be greatly appreciated.
1187 One bit would be to determine how to clone directory handles on systems
1188 without a C<fchdir> function (in sv.c:Perl_dirp_dup).
1190 Fix Perl_sv_dup, et al so that threads can return objects.
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]
1205 =head2 do away with fork/exec/wait?
1207 (system, popen should be enough?)
1209 =head2 some of the uconfig.sh really needs to be probed (using cc) in buildtime:
1211 (uConfigure? :-) native datatype widths and endianness come to mind