3 This is a list of wishes for Perl. It is maintained by Nathan
4 Torkington for the Perl porters. Send updates to
5 I<perl5-porters@perl.org>. If you want to work on any of these
6 projects, be sure to check the perl5-porters archives for past ideas,
7 flames, and propaganda. This will save you time and also prevent you
8 from implementing something that Larry has already vetoed. One set
9 of archives may be found at:
11 http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/
16 =head2 Mailing list archives
18 Chaim suggests contacting egroup and asking them to archive the other
19 perl.org mailing lists. Probably not advocacy, but definitely
22 =head2 Bug tracking system
24 Richard Foley I<richard@perl.org> is writing one. We looked at
25 several, like gnats and the Debian system, but at the time we
26 investigated them, none met our needs. Since then, Jitterbug has
27 matured, and may be worth reinvestigation.
29 The system we've developed will eventually be recipient of perlbug
30 mail. New bugs are entered into a mysql database, and sent on to
31 perl5-porters with the subject line rewritten to include a "ticket
32 number" (unique ID for the new bug). If the incoming message already
33 had a ticket number in the subject line, then the message is logged
34 against that bug. There is a separate email interface (not forwarding
35 to p5p) that permits porters to claim, categorize, and close tickets.
37 The next desire is a web interface. It is hoped that code can be
38 reused between the mail and the web interfaces.
40 The current delay in implementation is caused by perl.org lockups.
41 One suspect is the mail handling system, possibly going into loops.
43 We're probably going to need a bugmaster, someone who will look at
44 every new "bug" and kill those that we already know about, those
45 that are not bugs at all, etc.
47 =head2 Regression Tests
49 The test suite for Perl serves two needs: ensuring features work, and
50 ensuring old bugs have not been reintroduced. Both need work.
52 Brent LaVelle (lavelle@metronet.com) has stepped forward to work on
53 performance tests and improving the size of the test suite.
59 Do the tests that come with Perl exercise every line (or every block,
60 or ...) of the Perl interpreter, and if not then how can we make them
65 No bug fixes should be made without a corresponding testsuite addition.
66 This needs a dedicated enforcer, as the current pumpking is either too
67 lazy or too stupid or both and lets enforcement wander all over the
72 Tests that fail need to be of a form that can be readily mailed
73 to perlbug and diagnosed with minimal back-and-forth's to determine
74 which test failed, due to what cause, etc.
78 We need regression/sanity tests for suidperl
80 =item The 25% slowdown from perl4 to perl5
82 This value may or may not be accurate, but it certainly is
83 eye-catching. For some things perl5 is faster than perl4, but often
84 the reliability and extensability have come at a cost of speed. The
85 benchmark suite that Gisle released earlier has been hailed as both a
86 fantastic solution and as a source of entirely meaningless figures.
87 Do we need to test "real applications"? Can you do so? Anyone have
88 machines to dedicate to the task? Identify the things that have grown
89 slower, and see if there's a way to make them faster.
95 Andy Dougherty maintain(ed|s) a list of "todo" items for the configure
96 that comes with Perl. See Porting/pumpkin.pod in the latest
101 Have "make install" give you the option to install HTML as well. This
102 would be part of Configure. Andy Wardley (certified Perl studmuffin)
103 will look into the current problems of HTML installation--is
104 'installhtml' preventing this from happening cleanly, or is pod2html
105 the problem? If the latter, Brad Appleton's pod work may fix the
112 Declare global variables (lexically or otherwise).
116 Verify complete 64 bit support so that the value of sysseek, or C<-s>, or
117 stat(), or tell can fit into a perl number without losing precision.
118 Work with the perl-64bit mailing list on perl.org.
124 =item Named prototypes
126 Add proper named prototypes that actually work usefully.
128 =item Indirect objects
130 Fix prototype bug that forgets indirect objects.
134 Prototypes for method calls.
138 Return context prototype declarations.
142 lexically-scoped subs, e.g. my sub
146 =head2 Built-in globbing
148 Currently the C<E<lt>*.cE<gt>> syntax calls the c shell. This causes
149 problems on sites without csh, systems where fork() is expensive, and
150 setuid environments. Decide between Glob::BSD and File::KGlob, move
151 it into the core, and make Perl use it for globbing. Ben Holzman and
152 Tye McQueen have claimed the pumpkin for this.
154 =head1 Perl Internals
158 C<magic_setisa> should be made to update %FIELDS [???]
160 =head2 Garbage Collection
162 There was talk of a mark-and-sweep garbage collector at TPC2, but the
163 (to users) unpredictable nature of its behaviour put some off.
164 Sarathy, I believe, did the work. Here's what he has to say:
166 Yeah, I hope to implement it someday too. The points that were
167 raised in TPC2 were all to do with calling DESTROY() methods, but
168 I think we can accomodate that by extending bless() to stash
169 extra information for objects so we track their lifetime accurately
170 for those that want their DESTROY() to be predictable (this will be
171 a speed hit, naturally, and will therefore be optional, naturally. :)
173 [N.B. Don't even ask me about this now! When I have the time to
174 write a cogent summary, I'll post it.]
176 =head2 Reliable signals
178 Sarathy and Dan Sugalski are working on this. Chip posted a patch
179 earlier, but it was not accepted into 5.005. The issue is tricky,
180 because it has the potential to greatly slow down the core.
182 There are at least three things to consider:
186 =item Alternate runops() for signal despatch
188 Sarathy and Dan are discussed this on perl5-porters.
190 =item Figure out how to die() in delayed sighandler
192 =item Add tests for Thread::Signal
194 =item Automatic tests against CPAN
196 Is there some way to automatically build all/most of CPAN with
197 the new Perl and check that the modules there pass all the tests?
201 =head2 Interpolated regex performance bugs
205 foreach $pat (@patterns) {
211 The qr// syntax added in 5.005 has solved this problem, but
212 it needs more thorough documentation.
214 =head2 Memory leaks from failed eval/regcomp
216 The only known memory leaks in Perl are in failed code or regexp
217 compilation. Fix this. Hugo Van Der Sanden will attempt this but
218 won't have tuits until January 1999.
220 =head2 Make XS easier to use
222 There was interest in SWIG from porters, but nothing has happened
225 =head2 Make embedded Perl easier to use
227 This is probably difficult for the same reasons that "XS For Dummies"
230 =head2 Namespace cleanup
232 CPP-space: restrict CPP symbols exported from headers
233 header-space: move into CORE/perl/
234 API-space: begin list of things that constitute public api
235 env-space: Configure should use PERL_CONFIG instead of CONFIG etc.
239 Complete work on safe recursive interpreters C<Perl-E<gt>new()>.
240 Sarathy says that a reference implementation exists.
244 Chris Nandor and Matthias Neeracher are working on better integrating
245 MacPerl into the Perl distribution.
249 There's a lot of documentation that comes with Perl. The quantity of
250 documentation makes it difficult for users to know which section of
251 which manpage to read in order to solve their problem. Tom
252 Christiansen has done much of the documentation work in the past.
254 =head2 A clear division into tutorial and reference
256 Some manpages (e.g., perltoot and perlreftut) clearly set out to
257 educate the reader about a subject. Other manpages (e.g., perlsub)
258 are references for which there is no tutorial, or are references with
259 a slight tutorial bent. If things are either tutorial or reference,
260 then the reader knows which manpage to read to learn about a subject,
261 and which manpage to read to learn all about an aspect of that
262 subject. Part of the solution to this is:
264 =head2 Remove the artificial distinction between operators and functions
266 History shows us that users, and often porters, aren't clear on the
267 operator-function distinction. The present split in reference
268 material between perlfunc and perlop hinders user navigation. Given
269 that perlfunc is by far the larger of the two, move operator reference
272 =head2 More tutorials
274 More documents of a tutorial nature could help. Here are some
279 =item Regular expressions
281 Robin Berjon (r.berjon@ltconsulting.net) has volunteered.
285 Mark-Jason Dominus (mjd@plover.com) has an outline for perliotut.
289 This is badly needed. There has been some discussion on the
290 subject on perl5-porters.
294 Ronald Kimball (rjk@linguist.dartmouth.edu) has volunteered.
296 =head2 Include a search tool
298 perldoc should be able to 'grep' fulltext indices of installed POD
299 files. This would let people say:
301 perldoc -find printing numbers with commas
303 and get back the perlfaq entry on 'commify'.
305 This solution, however, requires documentation to contain the keywords
306 the user is searching for. Even when the users know what they're
307 looking for, often they can't spell it.
309 =head2 Include a locate tool
311 perldoc should be able to help people find the manpages on a
312 particular high-level subject:
316 would tell them manpages, web pages, and books with material on web
317 programming. Similarly C<perldoc -find databases>, C<perldoc -find
318 references> and so on.
320 We need something in the vicinity of:
322 % perl -help random stuff
323 No documentation for perl function `random stuff' found
324 The following entry in perlfunc.pod matches /random/a:
329 Returns a random fractional number greater than or equal to C<0> and less
330 than the value of EXPR. (EXPR should be positive.) If EXPR is
331 omitted, the value C<1> is used. Automatically calls C<srand()> unless
332 C<srand()> has already been called. See also C<srand()>.
334 (Note: If your rand function consistently returns numbers that are too
335 large or too small, then your version of Perl was probably compiled
336 with the wrong number of RANDBITS.)
337 The following pod pages seem to have /stuff/a:
338 perlfunc.pod (7 hits)
339 perlfaq7.pod (6 hits)
342 perlfaq8.pod (2 hits)
344 perl5004delta.pod (1 hit)
345 perl5005delta.pod (1 hit)
347 perldelta.pod (1 hit)
354 Proceed to open perlfunc.pod? [y] n
355 Do you want to speak perl interactively? [y] n
356 Should I dial 911? [y] n
357 Do you need psychiatric help? [y] y
358 <PELIZA> Hi, what bothers you today?
359 A Python programmer in the next cubby is driving me nuts!
360 <PELIZA> Hmm, thats fixable. Just [rest censored]
362 =head2 Separate function manpages by default
364 Perl should install 'manpages' for every function/operator into the
365 3pl or 3p manual section. By default. The splitman program in the
366 Perl source distribution does the work of turning big perlfunc into
369 =head2 Users can't find the manpages
371 Make C<perldoc> tell users what they need to add to their .login or
372 .cshrc to set their MANPATH correctly.
374 =head2 Install ALL Documentation
376 Make the standard documentation kit include the VMS, OS/2, Win32,
377 Threads, etc information. installperl and pod/Makefile should know
378 enough to copy README.foo to perlfoo.pod before building everything,
381 =head2 Outstanding issues to be documented
383 Tom has a list of 5.005_5* features or changes that require
386 Create one document that coherently explains the delta between the
387 last camel release and the current release. perldelta was supposed
388 to be that, but no longer. The things in perldelta never seemed to
389 get placed in the right places in the real manpages, either. This
392 =head2 Adapt www.linuxhq.com for Perl
394 This should help glorify documentation and get more people involved in
397 =head2 Replace man with a perl program
399 Can we reimplement man in Perl? Tom has a start. I believe some of
400 the Linux systems distribute a manalike. Alternatively, build on
401 perldoc to remove the unfeatures like "is slow" and "has no apropos".
403 =head2 Unicode tutorial
405 We could use more work on helping people understand Perl's new
406 Unicode support that Larry has created.
410 =head2 Update the POSIX extension to conform with the POSIX 1003.1 Edition 2
412 The current state of the POSIX extension is as of Edition 1, 1991,
413 whereas the Edition 2 came out in 1996. ISO/IEC 9945:1-1996(E),
414 ANSI/IEEE Std 1003.1, 1996 Edition. ISBN 1-55937-573-6. The updates
415 were legion: threads, IPC, and real time extensions.
417 =head2 Module versions
419 Automate the checking of versions in the standard distribution so
420 it's easy for a pumpking to check whether CPAN has a newer version
421 that we should be including?
425 Which modules should be added to the standard distribution? This ties
426 in with the SDK discussed on the perl-sdk list at perl.org.
430 Make the profiler (Devel::DProf) part of the standard release, and
439 Implement array using vec(). Nathan Torkington has working code to
444 Implement array using substr()
448 Implement array using a file
452 Defines shift et al in terms of splice method
456 =head2 Procedural options
458 Support procedural interfaces for the common cases of Perl's
459 gratuitously OOO modules. Tom objects to "use IO::File" reading many
460 thousands of lines of code.
464 Write a module for transparent, portable remote procedure calls. (Not
465 core). This touches on the CORBA and ILU work.
467 =head2 y2k localtime/gmtime
469 Write a module, Y2k::Catch, which overloads localtime and gmtime's
470 returned year value and catches "bad" attempts to use it.
472 =head2 Export File::Find variables
474 Make File::Find export C<$name> etc manually, at least if asked to.
478 Finish a proper Ioctl module.
480 =head2 Debugger attach/detach
482 Permit a user to debug an already-running program.
484 =head2 Regular Expression debugger
486 Create a visual profiler/debugger tool that stepped you through the
487 execution of a regular expression point by point. Ilya has a module
488 to color-code and display regular expression parses and executions.
489 There's something at http://tkworld.org/ that might be a good start,
490 it's a Tk/Tcl RE wizard, that builds regexen of many flavours.
492 =head2 Alternative RE Syntax
494 Make an alternative regular expression syntax that is accessed through
495 a module. For instance,
498 $re = start_of_line()
499 ->literal("1998/10/08")
500 ->optional( whitespace() )
502 ->remember( many( or( "-", digit() ) ) );
505 print "time is $1\n";
508 Newbies to regular expressions typically only use a subset of the full
509 language. Perhaps you wouldn't have to implement the full feature set.
511 =head2 Bundled modules
513 Nicholas Clark (nick@flirble.org) had a patch for storing modules in
514 zipped format. This needs exploring and concluding.
518 Adopt IO::Tty, make it as portable as Don Libes' "expect" (can we link
519 against expect code?), and perfect a Perl version of expect. IO::Tty
520 and expect could then be distributed as part of the core distribution,
521 replacing Comm.pl and other hacks.
525 A simple-to-use interface to native graphical abilities would
526 be welcomed. Oh, Perl's access Tk is nice enough, and reasonably
527 portable, but it's not particularly as fast as one would like.
528 Simple access to the mouse's cut buffer or mouse-presses shouldn't
529 required loading a few terabytes of Tk code.
531 =head2 Update semibroken auxiliary tools; h2ph, a2p, etc.
533 Kurt Starsinic is working on h2ph. mjd has fixed bugs in a2p in the
534 past. a2p apparently doesn't work on nawk and gawk extensions.
535 Graham Barr has an Include module that does h2ph work at runtime.
537 =head2 POD Converters
539 Brad's PodParser code needs to become part of the core, and the Pod::*
540 and pod2* programs rewritten to use this standard parser. Currently
541 the converters take different options, some behave in different
542 fashions, and some are more picky than others in terms of the POD
547 A short-term fix: pod2html generates absolute HTML links. Make it
548 generate relative links.
552 Something like lint for Pod would be good. Something that catches
553 common errors as well as gross ones. Brad Appleton is putting
554 together something as part of his PodParser work.
560 Design a webperl environment that's as tightly integrated and as
561 easy-to-use as Perl's current command-line environment.
565 More work on a safe and secure execution environment for mobile
566 agents would be neat; the Safe.pm module is a start, but there's a
567 still a lot to be done in that area. Adopt Penguin?
569 =head2 POSIX on non-POSIX
571 Standard programming constructs for non-POSIX systems would help a
572 lot of programmers stuck on primitive, legacy systems. For example,
573 Microsoft still hasn't made a usable POSIX interface on their clunky
574 systems, which means that standard operations such as alarm() and
575 fork(), both critical for sophisticated client-server programming,
576 must both be kludged around.
578 I'm unsure whether Tom means to emulate alarm( )and fork(), or merely
579 to provide a document like perlport.pod to say which features are
580 portable and which are not.
582 =head2 Portable installations
584 Figure out a portable semi-gelled installation, that is, one without
585 full paths. Larry has said that he's thinking about this. Ilya
586 pointed out that perllib_mangle() is good for this.
590 =head2 Get PERL_OBJECT building under gcc
592 B<Part done>, according to Sarathy. It builds under egcs on win32,
593 but doesn't run for occult reasons. If anyone knows the right
594 breed of chicken to sacrifice, please speak up.
596 =head2 Rename new headers to be consistent with the rest
598 =head2 Sort out the spawnvp() mess
600 =head2 Work out DLL versioning
602 =head2 Get PERL_OBJECT building on non-win32
606 =head1 Would be nice to have
610 =item C<pack "(stuff)*">
612 =item Contiguous bitfields in pack/unpack
616 =item Bundled perl preprocessor
618 =item Use posix calls internally where possible
622 =item -i rename file only when successfully changed
624 =item All ARGV input should act like <>
626 =item report HANDLE [formats].
628 =item support in perlmain to rerun debugger
630 =item lvalue functions
632 Tuomas Lukka, on behalf of the PDL project, greatly desires this and
633 Ilya has a patch for it (probably against an older version of Perl).
634 Tuomas points out that what PDL really wants is lvalue I<methods>,
639 =head1 Possible pragmas
643 (use less memory, CPU)
647 =head2 constant function cache
649 =head2 foreach(reverse...)
651 =head2 Cache eval tree
653 Unless lexical outer scope used (mark in &compiling?).
657 =head2 Shrink opcode tables
659 Via multiple implementations selected in peep.
661 =head2 Cache hash value
663 Not a win, according to Guido.
665 =head2 Optimize away @_ where possible
667 =head2 Optimize sort by { $a <=> $b }
669 Greg Bacon added several more sort optimizations. These have
670 made it into 5.005_55, thanks to Hans Mulder.
672 =head2 Rewrite regexp parser for better integrated optimization
674 The regexp parser was rewritten for 5.005. Ilya's the regexp guru.
676 =head1 Vague possibilities
680 =item ref function in list context
682 This seems impossible to do without substantially breaking code.
684 =item make tr/// return histogram in list context?
686 =item Loop control on do{} et al
688 =item Explicit switch statements
690 Nobody has yet managed to come up with a switch syntax that would
691 allow for mixed hash, constant, regexp checks. Submit implementation
694 =item compile to real threaded code
696 =item structured types
698 =item Modifiable $1 et al
700 The intent is for this to be a means of editing the matched portions of
705 =head1 To Do Or Not To Do
707 These are things that have been discussed in the past and roundly
708 criticized for being of questionable value.
710 =head2 Making my() work on "package" variables
712 Being able to say my($Foo::Bar), something that sounds ludicrous and
713 the 5.006 pumpking has mocked.
715 =head2 "or" testing defined not truth
717 We tell people that C<||> can be used to give a default value to a
720 $children = shift || 5; # default is 5 children
722 which is almost (but not):
725 $children = 5 unless $children;
727 but if the first argument was given and is "0", then it will be
728 considered false by C<||> and C<5> used instead. Really we want
729 an C<||>-like operator that behaves like:
732 $children = 5 unless defined $children;
734 Namely, a C<||> that tests defined-ness rather than truth. One was
735 discussed, and a patch submitted, but the objections were many. While
736 there were objections, many still feel the need. At least it was
737 decided that C<??> is the best name for the operator.
739 =head2 "dynamic" lexicals
746 Localizing, as Tim Bunce points out, is a separate concept from
747 whether the variable is global or lexical. Chip Salzenberg had
748 an implementation once, but Larry thought it had potential to
751 =head2 "class"-based, rather than package-based "lexicals"
753 This is like what the Alias module provides, but the variables would
754 be lexicals reserved by perl at compile-time, which really are indices
755 pointing into the pseudo-hash object visible inside every method so
762 Which of the standard modules are thread-safe? Which CPAN modules?
763 How easy is it to fix those non-safe modules?
767 Threading is still experimental. Every reproducible bug identifies
768 something else for us to fix. Find and submit more of these problems.
774 Consistent semantics for exit/die in threads.
776 =head2 External threads
778 Better support for externally created threads.
784 Spot-check globals like statcache and global GVs for thread-safety.
785 "B<Part done>", says Sarathy.
787 =head2 Per-thread GVs
789 According to Sarathy, this would make @_ be the same in threaded
790 and non-threaded, as well as helping solve problems like filehandles
791 (the same filehandle currently cannot be used in two threads).
797 The compiler's back-end code-generators for creating bytecode or
798 compilable C code could use optimization work.
802 Figure out how and where byteperl will be built for the various
805 =head2 Precompiled modules
807 Save byte-compiled modules on disk.
811 Auto-produce executable.
813 =head2 Typed lexicals
815 Typed lexicals should affect B::CC::load_pad.
819 Workarounds to help Win32 dynamic loading.
823 END blocks need saving in compiled output.
831 Fix comppadlist (names in comppad_name can have fake SvCUR
832 from where newASSIGNOP steals the field).
834 =head2 Cached compilation
836 Can we install modules as bytecode?
838 =head1 Recently Finished Tasks
840 =head2 Figure a way out of $^(capital letter)
842 Figure out a clean way to extend $^(capital letter) beyond
843 the 26 alphabets. (${^WORD} maybe?)
845 Mark-Jason Dominus sent a patch which went into 5.005_56.
849 Make filenames in the distribution and in the standard module set
850 be 8.3 friendly where feasible. Good luck changing the standard
851 modules, though. B<Done>.
853 =head2 Proper tied array support
855 This was B<done> in 5.005 by Nick Ing-Simmons.
859 Perl should be more generous in accepting foreign line terminations.
860 Mostly B<done> in 5.005.
862 =head2 Namespace cleanup
864 symbol-space: "pl_" prefix for all global vars
865 "Perl_" prefix for all functions
867 CPP-space: stop malloc()/free() pollution unless asked
871 Given a piece of Perl code, say what it does. B::Deparse is doing
876 Rename and alter ISA.pm. B<Done>. It is now base.pm.
878 =head2 Automate maintenance of most PERL_OBJECT code
880 B<Done>, says Sarathy.
884 Added in 5.004_70. B<Done>
890 =head2 reference to compiled regexp
892 B<done> This is the qr// support in 5.005.
894 =head2 eval qw() at compile time
896 qw() is presently compiled as a call to split. This means the split
897 happens at runtime. Change this so qw() is compiled as a real list
898 assignment. This also avoids surprises like:
900 $a = () = qw(What will $a hold?);
902 B<Done>. Tom Hughes submitted a patch that went into 5.005_55.
906 B<Done>. This is the Fatal.pm module, so any builtin that that does
907 not return success automatically die()s. If you're feeling brave, tie
908 this in with the unified exceptions scheme.
910 =head2 Status variable
912 $^C to track compiler/checker status. B<Done> in 5.005_54.