3 perltodo - Perl TO-DO List
7 This is a list of wishes for Perl. The tasks we think are smaller or easier
8 are listed first. Anyone is welcome to work on any of these, but it's a good
9 idea to first contact I<perl5-porters@perl.org> to avoid duplication of
10 effort. By all means contact a pumpking privately first if you prefer.
12 Whilst patches to make the list shorter are most welcome, ideas to add to
13 the list are also encouraged. Check the perl5-porters archives for past
14 ideas, and any discussion about them. One set of archives may be found at:
16 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
22 =head1 Tasks that only need Perl knowledge
24 =head2 common test code for timed bail out
26 Write portable self destruct code for tests to stop them burning CPU in
27 infinite loops. This needs to avoid using alarm, as some of the tests are
28 testing alarm/sleep or timers.
30 =head2 POD -> HTML conversion in the core still sucks
32 Which is crazy given just how simple POD purports to be, and how simple HTML
33 can be. It's not actually I<as> simple as it sounds, particularly with the
34 flexibility POD allows for C<=item>, but it would be good to improve the
35 visual appeal of the HTML generated, and to avoid it having any validation
36 errors. See also L</make HTML install work>, as the layout of installation tree
37 is needed to improve the cross-linking.
39 =head2 Make Schwern poorer
41 We should have for everything. When all the core's modules are tested,
42 Schwern has promised to donate to $500 to TPF. We may need volunteers to
43 hold him upside down and shake vigorously in order to actually extract the
46 See F<t/lib/1_compile.t> for the 3 remaining modules that need tests.
48 =head2 Improve the coverage of the core tests
50 Use Devel::Cover to ascertain the core's test coverage, then add tests that
51 are currently missing.
55 A full test suite for the B module would be nice.
57 =head2 A decent benchmark
59 perlbench seems impervious to any recent changes made to the perl core. It
60 would be useful to have a reasonable general benchmarking suite that roughly
61 represented what current perl programs do, and measurably reported whether
62 tweaks to the core improve, degrade or don't really affect performance, to
63 guide people attempting to optimise the guts of perl. Gisle would welcome
64 new tests for perlbench.
66 =head2 fix tainting bugs
68 Fix the bugs revealed by running the test suite with the C<-t> switch (via
69 C<make test.taintwarn>).
71 =head2 Dual life everything
73 As part of the "dists" plan, anything that doesn't belong in the smallest perl
74 distribution needs to be dual lifed. Anything else can be too. Figure out what
75 changes would be needed to package that module and its tests up for CPAN, and
76 do so. Test it with older perl releases, and fix the problems you find.
78 =head2 Improving C<threads::shared>
80 Investigate whether C<threads::shared> could share aggregates properly with
81 only Perl level changes to shared.pm
83 =head2 POSIX memory footprint
85 Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
86 various times worked to cut it down. There is probably still fat to cut out -
87 for example POSIX passes Exporter some very memory hungry data structures.
95 =head1 Tasks that need a little sysadmin-type knowledge
97 Or if you prefer, tasks that you would learn from, and broaden your skills
100 =head2 make HTML install work
102 There is an C<installhtml> target in the Makefile. It's marked as
103 "experimental". It would be good to get this tested, make it work reliably, and
104 remove the "experimental" tag. This would include
110 Checking that cross linking between various parts of the documentation works.
111 In particular that links work between the modules (files with POD in F<lib/>)
112 and the core documentation (files in F<pod/>)
116 Work out how to split perlfunc into chunks, preferably one per function group,
117 preferably with general case code that could be used elsewhere. Challenges
118 here are correctly identifying the groups of functions that go together, and
119 making the right named external cross-links point to the right page. Things to
120 be aware of are C<-X>, groups such as C<getpwnam> to C<endservent>, two or
121 more C<=items> giving the different parameter lists, such as
123 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
125 =item substr EXPR,OFFSET,LENGTH
127 =item substr EXPR,OFFSET
129 and different parameter lists having different meanings. (eg C<select>)
133 =head2 compressed man pages
135 Be able to install them. This would probably need a configure test to see how
136 the system does compressed man pages (same directory/different directory?
137 same filename/different filename), as well as tweaking the F<installman> script
138 to compress as necessary.
140 =head2 Add a code coverage target to the Makefile
142 Make it easy for anyone to run Devel::Cover on the core's tests. The steps
143 to do this manually are roughly
149 do a normal C<Configure>, but include Devel::Cover as a module to install
150 (see F<INSTALL> for how to do this)
158 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
162 Process the resulting Devel::Cover database
166 This just give you the coverage of the F<.pm>s. To also get the C level
173 Additionally tell C<Configure> to use the appropriate C compiler flags for
180 (instead of C<make perl>)
184 After running the tests run C<gcov> to generate all the F<.gcov> files.
185 (Including down in the subdirectories of F<ext/>
189 (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
190 to get their stats into the cover_db directory.
194 Then process the Devel::Cover database
198 It would be good to add a single switch to C<Configure> to specify that you
199 wanted to perform perl level coverage, and another to specify C level
200 coverage, and have C<Configure> and the F<Makefile> do all the right things
203 =head2 Make Config.pm cope with differences between build and installed perl
205 Quite often vendors ship a perl binary compiled with their (pay-for)
206 compilers. People install a free compiler, such as gcc. To work out how to
207 build extensions, Perl interrogates C<%Config>, so in this situation
208 C<%Config> describes compilers that aren't there, and extension building
209 fails. This forces people into choosing between re-compiling perl themselves
210 using the compiler they have, or only using modules that the vendor ships.
212 It would be good to find a way teach C<Config.pm> about the installation setup,
213 possibly involving probing at install time or later, so that the C<%Config> in
214 a binary distribution better describes the installed machine, when the
215 installed machine differs from the build machine in some significant way.
217 =head2 Relocatable perl
219 The C level patches needed to create a relocatable perl binary are done, as
220 is the work on Config.pm. All that's left to do is the C<Configure> tweaking
221 to let people specify how they want to do the install.
223 =head2 make parallel builds work
225 Currently parallel builds (such as C<make -j3>) don't work reliably. We believe
226 that this is due to incomplete dependency specification in the F<Makefile>.
227 It would be good if someone were able to track down the causes of these
228 problems, so that parallel builds worked properly.
230 =head2 Refactor C<xsubpp> to be a thin wrapper around C<ExtUtils::ParseXS>
232 C<ExtUtils::ParseXS> encapsulates a version of the C<xsubpp> into a module.
233 In effect this is a code fork, and it's likely that C<xsubpp> has had some
234 bug fixes since the code from C<ExtUtils::ParseXS> was derived. It would be
235 good to merge the differences in, reduce down to 1 canonical implementation,
236 and convert C<xsubpp> to a very thin command line wrapper to
237 C<ExtUtils::ParseXS>.
239 In theory this needs no real C knowledge, as one way of approaching this task
240 is to ensure that C<ExtUtils::ParseXS> generates identical output to C<xsubpp>
241 for input XS files, which does not require understanding the contents of the
242 output C file. However, some C knowledge is likely to help with testing, and
243 locating/producing comprehensive test cases.
249 =head1 Tasks that need a little C knowledge
251 These tasks would need a little C knowledge, but don't need any specific
252 background or experience with XS, or how the Perl interpreter works
254 =head2 Make it clear from -v if this is the exact official release
256 Currently perl from p4/rsync ships with a patchlevel.h file that usually
257 defines one local patch, of the form "MAINT12345" or "RC1". The output of
258 perl -v doesn't report that a perl isn't an official release, and this
259 information can get lost in bugs reports. Because of this, the minor version
260 isn't bumped up until RC time, to minimise the possibility of versions of perl
261 escaping that believe themselves to be newer than they actually are.
263 It would be useful to find an elegant way to have the "this is an interim
264 maintenance release" or "this is a release candidate" in the terse -v output,
265 and have it so that it's easy for the pumpking to remove this just as the
266 release tarball is rolled up. This way the version pulled out of rsync would
267 always say "I'm a development release" and it would be safe to bump the
268 reported minor version as soon as a release ships, which would aid perl
271 This task is really about thinking of an elegant way to arrange the C source
272 such that it's trivial for the Pumpking to flag "this is an official release"
273 when making a tarball, yet leave the default source saying "I'm not the
276 =head2 Ordering of "global" variables.
278 F<thrdvar.h> and F<intrpvarh> define the "global" variables that need to be
279 per-thread under ithreads, where the variables are actually elements in a
280 structure. As C dictates, the variables must be laid out in order of
281 declaration. There is a comment
282 C</* Important ones in the first cache line (if alignment is done right) */>
283 which implies that at some point in the past the ordering was carefully chosen
284 (at least in part). However, it's clear that the ordering is less than perfect,
285 as currently there are things such as 7 C<bool>s in a row, then something
286 typically requiring 4 byte alignment, and then an odd C<bool> later on.
287 (C<bool>s are typically defined as C<char>s). So it would be good for someone
288 to review the ordering of the variables, to see how much alignment padding can
291 =head2 bincompat functions
293 There are lots of functions which are retained for binary compatibility.
294 Clean these up. Move them to mathom.c, and don't compile for blead?
296 =head2 am I hot or not?
298 The idea of F<pp_hot.c> is that it contains the I<hot> ops, the ops that are
299 most commonly used. The idea is that by grouping them, their object code will
300 be adjacent in the executable, so they have a greater chance of already being
301 in the CPU cache (or swapped in) due to being near another op already in use.
303 Except that it's not clear if these really are the most commonly used ops. So
304 anyone feeling like exercising their skill with coverage and profiling tools
305 might want to determine what ops I<really> are the most commonly used. And in
306 turn suggest evictions and promotions to achieve a better F<pp_hot.c>.
308 =head2 emulate the per-thread memory pool on Unix
310 For Windows, ithreads allocates memory for each thread from a separate pool,
311 which it discards at thread exit. It also checks that memory is free()d to
312 the correct pool. Neither check is done on Unix, so code developed there won't
313 be subject to such strictures, so can harbour bugs that only show up when the
314 code reaches Windows.
316 It would be good to be able to optionally emulate the Window pool system on
317 Unix, to let developers who only have access to Unix, or want to use
318 Unix-specific debugging tools, check for these problems. To do this would
319 involve figuring out how the C<PerlMem_*> macros wrap C<malloc()> access, and
320 providing a layer that records/checks the identity of the thread making the
321 call, and recording all the memory allocated by each thread via this API so
322 that it can be summarily free()d at thread exit. One implementation idea
323 would be to increase the size of allocation, and store the C<my_perl> pointer
324 (to identify the thread) at the start, along with pointers to make a linked
325 list of blocks for this thread. To avoid alignment problems it would be
326 necessary to do something like
328 union memory_header_padded {
329 struct memory_header {
330 void *thread_id; /* For my_perl */
331 void *next; /* Pointer to next block for this thread */
333 long double padding; /* whatever type has maximal alignment constraint */
337 although C<long double> might not be the only type to add to the padding
340 =head2 reduce duplication in sv_setsv_flags
342 C<Perl_sv_setsv_flags> has a comment
343 C</* There's a lot of redundancy below but we're going for speed here */>
345 Whilst this was true 10 years ago, the growing disparity between RAM and CPU
346 speeds mean that the trade offs have changed. In addition, the duplicate code
347 adds to the maintenance burden. It would be good to see how much of the
348 redundancy can be pruned, particular in the less common paths. (Profiling
349 tools at the ready...). For example, why does the test for
350 "Can't redefine active sort subroutine" need to occur in two places?
355 =head1 Tasks that need a knowledge of XS
357 These tasks would need C knowledge, and roughly the level of knowledge of
358 the perl API that comes from writing modules that use XS to interface to
363 Clean this up. Check everything in core works
365 =head2 shrink C<GV>s, C<CV>s
367 By removing unused elements and careful re-ordering, the structures for C<AV>s
368 and C<HV>s have recently been shrunk considerably. It's probable that the same
369 approach would find savings in C<GV>s and C<CV>s, if not all the other
370 larger-than-C<PVMG> types.
372 =head2 merge Perl_sv_2[inpu]v
374 There's a lot of code shared between C<Perl_sv_2iv_flags>,
375 C<Perl_sv_2uv_flags>, C<Perl_sv_2nv>, and C<Perl_sv_2pv_flags>. It would be
376 interesting to see if some of it can be merged into common shared static
377 functions. In particular, C<Perl_sv_2uv_flags> started out as a cut&paste
378 from C<Perl_sv_2iv_flags> around 5.005_50 time, and it may be possible to
379 replace both with a single function that returns a value or union which is
380 split out by the macros in F<sv.h>
382 =head2 UTF8 caching code
384 The string position/offset cache is not optional. It should be.
386 =head2 Implicit Latin 1 => Unicode translation
388 Conversions from byte strings to UTF-8 currently map high bit characters
389 to Unicode without translation (or, depending on how you look at it, by
390 implicitly assuming that the byte strings are in Latin-1). As perl assumes
391 the C locale by default, upgrading a string to UTF-8 may change the
392 meaning of its contents regarding character classes, case mapping, etc.
393 This should probably emit a warning (at least).
395 This task is incremental - even a little bit of work on it will help.
397 =head2 autovivification
399 Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
401 This task is incremental - even a little bit of work on it will help.
403 =head2 Unicode in Filenames
405 chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
406 opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
407 system, truncate, unlink, utime, -X. All these could potentially accept
408 Unicode filenames either as input or output (and in the case of system
409 and qx Unicode in general, as input or output to/from the shell).
410 Whether a filesystem - an operating system pair understands Unicode in
413 Known combinations that have some level of understanding include
414 Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
415 OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
416 create Unicode filenames, what forms of Unicode are accepted and used
417 (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
418 and so on, varies. Finding the right level of interfacing to Perl
419 requires some thought. Remember that an OS does not implicate a
422 (The Windows -C command flag "wide API support" has been at least
423 temporarily retired in 5.8.1, and the -C has been repurposed, see
426 =head2 Unicode in %ENV
428 Currently the %ENV entries are always byte strings.
430 =head2 use less 'memory'
432 Investigate trade offs to switch out perl's choices on memory usage.
433 Particularly perl should be able to give memory back.
435 This task is incremental - even a little bit of work on it will help.
437 =head2 Re-implement C<:unique> in a way that is actually thread-safe
439 The old implementation made bad assumptions on several levels. A good 90%
440 solution might be just to make C<:unique> work to share the string buffer
441 of SvPVs. That way large constant strings can be shared between ithreads,
442 such as the configuration information in F<Config>.
444 =head2 Make tainting consistent
446 Tainting would be easier to use if it didn't take documented shortcuts and
447 allow taint to "leak" everywhere within an expression.
449 =head2 readpipe(LIST)
451 system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
452 running a shell. readpipe() (the function behind qx//) could be similarly
459 =head1 Tasks that need a knowledge of the interpreter
461 These tasks would need C knowledge, and knowledge of how the interpreter works,
462 or a willingness to learn.
464 =head2 lexical pragmas
466 Reimplement the mechanism of lexical pragmas to be more extensible. Fix
467 current pragmas that don't work well (or at all) with lexical scopes or in
468 run-time eval(STRING) (C<sort>, C<re>, C<encoding> for example). MJD has a
469 preliminary patch that implements this.
471 =head2 Attach/detach debugger from running program
473 The old perltodo notes "With C<gdb>, you can attach the debugger to a running
474 program if you pass the process ID. It would be good to do this with the Perl
475 debugger on a running Perl program, although I'm not sure how it would be
476 done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
478 =head2 inlining autoloaded constants
480 Currently the optimiser can inline constants when expressed as subroutines
481 with prototype ($) that return a constant. Likewise, many packages wrapping
482 C libraries export lots of constants as subroutines which are AUTOLOADed on
483 demand. However, these have no prototypes, so can't be seen as constants by
484 the optimiser. Some way of cheaply (low syntax, low memory overhead) to the
485 perl compiler that a name is a constant would be great, so that it knows to
486 call the AUTOLOAD routine at compile time, and then inline the constant.
488 =head2 Constant folding
490 The peephole optimiser should trap errors during constant folding, and give
491 up on the folding, rather than bailing out at compile time. It is quite
492 possible that the unfoldable constant is in unreachable code, eg something
493 akin to C<$a = 0/0 if 0;>
495 =head2 LVALUE functions for lists
497 The old perltodo notes that lvalue functions don't work for list or hash
498 slices. This would be good to fix.
500 =head2 LVALUE functions in the debugger
502 The old perltodo notes that lvalue functions don't work in the debugger. This
503 would be good to fix.
505 =head2 _ prototype character
507 Study the possibility of adding a new prototype character, C<_>, meaning
508 "this argument defaults to $_".
510 =head2 @INC source filter to Filter::Simple
512 The second return value from a sub in @INC can be a source filter. This isn't
513 documented. It should be changed to use Filter::Simple, tested and documented.
515 =head2 regexp optimiser optional
517 The regexp optimiser is not optional. It should configurable to be, to allow
518 its performance to be measured, and its bugs to be easily demonstrated.
522 Introduce a new special block, UNITCHECK, which is run at the end of a
523 compilation unit (module, file, eval(STRING) block). This will correspond to
524 the Perl 6 CHECK. Perl 5's CHECK cannot be changed or removed because the
525 O.pm/B.pm backend framework depends on it.
527 =head2 optional optimizer
529 Make the peephole optimizer optional. Currently it performs two tasks as
530 it walks the optree - genuine peephole optimisations, and necessary fixups of
531 ops. It would be good to find an efficient way to switch out the
532 optimisations whilst keeping the fixups.
534 =head2 You WANT *how* many
536 Currently contexts are void, scalar and list. split has a special mechanism in
537 place to pass in the number of return values wanted. It would be useful to
538 have a general mechanism for this, backwards compatible and little speed hit.
539 This would allow proposals such as short circuiting sort to be implemented
542 =head2 lexical aliases
544 Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
546 =head2 entersub XS vs Perl
548 At the moment pp_entersub is huge, and has code to deal with entering both
549 perl and XS subroutines. Subroutine implementations rarely change between
550 perl and XS at run time, so investigate using 2 ops to enter subs (one for
551 XS, one for perl) and swap between if a sub is redefined.
555 self ties are currently illegal because they caused too many segfaults. Maybe
556 the causes of these could be tracked down and self-ties on all types re-
559 =head2 Optimize away @_
561 The old perltodo notes "Look at the "reification" code in C<av.c>".
565 The old perltodo notes "Although we have C<Switch.pm> in core, Larry points to
566 the dormant C<nswitch> and C<cswitch> ops in F<pp.c>; using these opcodes would
569 =head2 What hooks would assertions need?
571 Assertions are in the core, and work. However, assertions needed to be added
572 as a core patch, rather than an XS module in ext, or a CPAN module, because
573 the core has no hooks in the necessary places. It would be useful to
574 investigate what hooks would need to be added to make it possible to provide
575 the full assertion support from a CPAN module, so that we aren't constraining
576 the imagination of future CPAN authors.
586 Tasks that will get your name mentioned in the description of the "Highlights
589 =head2 make ithreads more robust
591 Generally make ithreads more robust. See also L<iCOW>
593 This task is incremental - even a little bit of work on it will help, and
594 will be greatly appreciated.
598 Sarathy and Arthur have a proposal for an improved Copy On Write which
599 specifically will be able to COW new ithreads. If this can be implemented
600 it would be a good thing.
602 =head2 (?{...}) closures in regexps
604 Fix (or rewrite) the implementation of the C</(?{...})/> closures.
606 =head2 A re-entrant regexp engine
608 This will allow the use of a regex from inside (?{ }), (??{ }) and
609 (?(?{ })|) constructs.