This is a live mirror of the Perl 5 development currently hosted at
Known test failures update.
[perl5.git] / pod / perltodo.pod
1=head1 NAME
3perltodo - Perl TO-DO List
e50bb9a1 6
722d2a37 7This is a list of wishes for Perl. Send updates to
8I<>. If you want to work on any of these
9projects, be sure to check the perl5-porters archives for past ideas,
10flames, and propaganda. This will save you time and also prevent you
11from implementing something that Larry has already vetoed. One set
12of archives may be found at:
722d2a37 16=head1 To do during 5.6.x
e50bb9a1 17
722d2a37 18=head2 Support for I/O disciplines
e50bb9a1 19
20C<perlio> provides this, but the interface could be a lot more
e50bb9a1 22
722d2a37 23=head2 Eliminate need for "use utf8";
e50bb9a1 24
25While the C<utf8> pragma is autoloaded when necessary, it's still needed
26for things like Unicode characters in a source file. The UTF8 hint can
27always be set to true, but it needs to be set to false when F<>
28is being compiled. (To stop Perl trying to autoload the C<utf8>
e50bb9a1 30
f35392ae 31=head2 Create a char *sv_printify(sv, STRLEN *lenp, UV flags)
33For displaying PVs with control characters, embedded nulls, and Unicode.
34This would be useful for printing warnings, or data and regex dumping,
35not_a_number(), and so on.
37Requirements: should handle both byte and UTF8 strings. isPRINT()
38characters printed as-is, character less than 256 as \xHH, Unicode
39characters as \x{HHH}.
41Possible options, controlled by the flags:
42- whitespace (other than ' ' of isPRINTF()) printed as-is
43- use isPRINT_LC() instead of isPRINT()
44- print control characters like this: "\cA"
45- print control characters like this: "^A"
46- non-printables printed as '.' instead of \xHH
47- print the \OOO instead of \xHH
48- have a maximum length for the produced string (read it from *lenp)
49- append a "..." to the produced string if the maximum length is exceeded
722d2a37 51=head2 Autoload
e50bb9a1 52
53When the lexer sees, for instance, C<bytes::length>, it should
54automatically load the C<bytes> pragma.
e50bb9a1 55
722d2a37 56=head2 Make "\u{XXXX}" et al work
e50bb9a1 57
58Danger, Will Robinson! Discussing the semantics of C<"\x{F00}">,
59C<"\xF00"> and C<"\U{F00}"> on P5P I<will> lead to a long and boring
e50bb9a1 61
722d2a37 62=head2 Overloadable regex assertions
e50bb9a1 63
64This may or may not be possible with the current regular expression
65engine. The idea is that, for instance, C<\b> needs to be
66algorithmically computed if you're dealing with Thai text. Hence, the
67B<\b> assertion wants to be overloaded by a function.
e50bb9a1 68
722d2a37 69=head2 Unicode collation and normalization
e50bb9a1 70
722d2a37 71Simon Cozens promises to work on this.
e50bb9a1 72
73 Collation?
74 Normalization?
e50bb9a1 75
722d2a37 76=head2 Unicode case mappings
e50bb9a1 77
722d2a37 78 Case Mappings?
e50bb9a1 79
722d2a37 80=head2 Unicode regular expression character classes
e50bb9a1 81
722d2a37 82They have some tricks Perl doesn't yet implement.
e50bb9a1 83
722d2a37 84
e50bb9a1 85
722d2a37 86=head2 use Thread for iThreads
e50bb9a1 87
88Artur Bergman's C<iThreads> module is a start on this, but needs to
89be more mature.
e50bb9a1 90
91=head2 make perl_clone optionally clone ops
93So that pseudoforking, mod_perl, iThreads and nvi will work properly
94(but not as efficiently) until the regex engine is fixed to be threadsafe.
722d2a37 96=head2 Work out exit/die semantics for threads
e50bb9a1 97
722d2a37 98=head2 Typed lexicals for compiler
e50bb9a1 99
722d2a37 100=head2 Compiler workarounds for Win32
e50bb9a1 101
722d2a37 102=head2 AUTOLOADing in the compiler
e50bb9a1 103
722d2a37 104=head2 Fixing comppadlist when compiling
e50bb9a1 105
722d2a37 106=head2 Cleaning up exported namespace
e50bb9a1 107
722d2a37 108=head2 Complete signal handling
e50bb9a1 109
110Add C<PERL_ASYNC_CHECK> to opcodes which loop; replace C<sigsetjmp> with
111C<sigjmp>; check C<wait> for signal safety.
e50bb9a1 112
722d2a37 113=head2 Out-of-source builds
e50bb9a1 114
722d2a37 115This was done for 5.6.0, but needs reworking for 5.7.x
e50bb9a1 116
722d2a37 117=head2 POSIX realtime support
e50bb9a1 118
119POSIX 1003.1 1996 Edition support--realtime stuff: POSIX semaphores,
120message queues, shared memory, realtime clocks, timers, signals (the
121metaconfig units mostly already exist for these)
e50bb9a1 122
722d2a37 123=head2 UNIX98 support
e50bb9a1 124
722d2a37 125Reader-writer locks, realtime/asynchronous IO
e50bb9a1 126
722d2a37 127=head2 IPv6 Support
e50bb9a1 128
129There are non-core modules, such as C<Net::IPv6>, but these will need
130integrating when IPv6 actually starts to really happen. See RFC 2292
131and RFC 2553.
e50bb9a1 132
722d2a37 133=head2 Long double conversion
e50bb9a1 134
722d2a37 135Floating point formatting is still causing some weird test failures.
e50bb9a1 136
722d2a37 137=head2 Locales
e50bb9a1 138
139Locales and Unicode interact with each other in unpleasant ways.
140One possible solution would be to adopt/support ICU:
e50bb9a1 141
722d2a37 142
e50bb9a1 143
722d2a37 144=head2 Thread-safe regexes
e50bb9a1 145
722d2a37 146The regular expression engine is currently non-threadsafe.
e50bb9a1 147
722d2a37 148=head2 Arithmetic on non-Arabic numerals
e50bb9a1 149
722d2a37 150C<[1234567890]> aren't the only numerals any more.
e50bb9a1 151
722d2a37 152=head2 POSIX Unicode character classes
e50bb9a1 153
154([=a=] for equivalance classes, [.ch.] for collation.)
155These are dependent on Unicode normalization and collation.
e50bb9a1 156
722d2a37 157=head2 Factoring out common suffices/prefices in regexps (trie optimization)
c47ff5f1 158
159Currently, the user has to optimize C<foo|far> and C<foo|goo> into
160C<f(?:oo|ar)> and C<[fg]oo> by hand; this could be done automatically.
e50bb9a1 161
722d2a37 162=head2 Security audit shipped utilities
e50bb9a1 163
164All the code we ship with Perl needs to be sensible about temporary file
165handling, locking, input validation, and so on.
e50bb9a1 166
722d2a37 167=head2 Custom opcodes
e50bb9a1 168
169Have a way to introduce user-defined opcodes without the subroutine call
170overhead of an XSUB; the user should be able to create PP code. Simon
171Cozens has some ideas on this.
e50bb9a1 172
722d2a37 173=head2 spawnvp() on Win32
e50bb9a1 174
175Win32 has problems spawning processes, particularly when the arguments
176to the child process contain spaces, quotes or tab characters.
e50bb9a1 177
722d2a37 178=head2 DLL Versioning
e50bb9a1 179
180Windows needs a way to know what version of a XS or C<libperl> DLL it's
e50bb9a1 182
722d2a37 183=head2 Introduce @( and @)
e50bb9a1 184
185C<$(> may return "foo bar baz". Unfortunately, since groups can
186theoretically have spaces in their names, this could be one, two or
187three groups.
e50bb9a1 188
722d2a37 189=head2 Floating point handling
e50bb9a1 190
191C<NaN> and C<inf> support is particularly troublesome.
192(fp_classify(), fp_class(), fp_class_d(), class(), isinf(),
193isfinite(), finite(), isnormal(), unordered(), <ieeefp.h>,
194<fp_class.h> (there are metaconfig units for all these) (I think),
195fp_setmask(), fp_getmask(), fp_setround(), fp_getround()
196(no metaconfig units yet for these). Don't forget finitel(), fp_classl(),
197fp_class_l(), (yes, both do, unfortunately, exist), and unorderedl().)
e50bb9a1 198
722d2a37 199As of Perl 5.6.1 is a Perl macro, Perl_isnan().
e50bb9a1 200
722d2a37 201=head2 IV/UV preservation
e50bb9a1 202
203Nicholas Clark has done a lot of work on this, but work is continuing.
204C<+>, C<-> and C<*> work, but guards need to be in place for C<%>, C</>,
205C<&>, C<oct>, C<hex> and C<pack>.
e50bb9a1 206
722d2a37 207=head2 Replace pod2html with something using Pod::Parser
83df6a1d 208
209The CPAN module C<Malik::Pod::Html> may be a more suitable basis for a
210C<pod2html> convertor; the current one duplicates the functionality
211abstracted in C<Pod::Parser>, which makes updating the POD language
e50bb9a1 213
722d2a37 214=head2 Automate module testing on CPAN
e50bb9a1 215
216When a new Perl is being beta tested, porters have to manually grab
217their favourite CPAN modules and test them - this should be done
e50bb9a1 219
722d2a37 220=head2 sendmsg and recvmsg
83df6a1d 221
222We have all the other BSD socket functions but these. There are
223metaconfig units for these functions which can be added. To avoid these
224being new opcodes, a solution similar to the way C<sockatmark> was added
225would be preferable. (Autoload the C<IO::whatever> module.)
e50bb9a1 226
722d2a37 227=head2 Rewrite perlre documentation
e50bb9a1 228
229The new-style patterns need full documentation, and the whole document
230needs to be a lot clearer.
e50bb9a1 231
722d2a37 232=head2 Convert example code to IO::Handle filehandles
e50bb9a1 233
722d2a37 234=head2 Document Win32 choices
e50bb9a1 235
722d2a37 236=head2 Check new modules
e50bb9a1 237
722d2a37 238=head2 Make roffitall find pods and libs itself
e50bb9a1 239
722d2a37 240Simon Cozens has done some work on this but it needs a rethink.
e50bb9a1 241
722d2a37 242=head1 To do at some point
e50bb9a1 243
244These are ideas that have been regularly tossed around, that most
245people believe should be done maybe during 5.8.x
e50bb9a1 246
722d2a37 247=head2 Remove regular expression recursion
e50bb9a1 248
249Because the regular expression engine is recursive, badly designed
250expressions can lead to lots of recursion filling up the stack. Ilya
251claims that it is easy to convert the engine to being iterative, but
252this has still not yet been done. There may be a regular expression
253engine hit squad meeting at TPC5.
e50bb9a1 254
722d2a37 255=head2 Memory leaks after failed eval
e50bb9a1 256
257Perl will leak memory if you C<eval "hlagh hlagh hlagh hlagh">. This is
258partially because it attempts to build up an op tree for that code and
259doesn't properly free it. The same goes for non-syntactically-correct
260regular expressions. Hugo looked into this, but decided it needed a
261mark-and-sweep GC implementation.
e50bb9a1 262
263Alan notes that: The basic idea was to extend the parser token stack
264(C<YYSTYPE>) to include a type field so we knew what sort of thing each
265element of the stack was. The F<<perly.c> code would then have to be
266postprocessed to record the type of each entry on the stack as it was
267created, and the parser patched so that it could unroll the stack
268properly on error.
e50bb9a1 269
270This is possible to do, but would be pretty messy to implement, as it
271would rely on even more sed hackery in F<perly.fixer>.
e50bb9a1 272
722d2a37 273=head2 pack "(stuff)*"
e50bb9a1 274
722d2a37 275That's to say, C<pack "(sI)40"> would be the same as C<pack "sI"x40>
e50bb9a1 276
722d2a37 277=head2 bitfields in pack
e50bb9a1 278
722d2a37 279=head2 Cross compilation
e50bb9a1 280
281Make Perl buildable with a cross-compiler. This will play havoc with
282Configure, which needs to how how the target system will respond to
283its tests; maybe C<microperl> will be a good starting point here.
284(Indeed, Bart Schuller reports that he compiled up C<microperl> for
285the Agenda PDA and it works fine.) A really big spanner in the works
286is the bootstrapping build process of Perl: if the filesystem the
287target systems sees is not the same what the build host sees, various
288input, output, and (Perl) library files need to be copied back and forth.
e50bb9a1 289
722d2a37 290=head2 Perl preprocessor / macros
e50bb9a1 291
292Source filters help with this, but do not get us all the way. For
293instance, it should be possible to implement the C<??> operator somehow;
294source filters don't (quite) cut it.
e50bb9a1 295
722d2a37 296=head2 Perl lexer in Perl
a45bd81d 297
722d2a37 298Damian Conway is planning to work on this, but it hasn't happened yet.
e50bb9a1 299
722d2a37 300=head2 Using POSIX calls internally
e50bb9a1 301
302When faced with a BSD vs. SySV -style interface to some library or
303system function, perl's roots show in that it typically prefers the BSD
304interface (but falls back to the SysV one). One example is getpgrp().
305Other examples include C<memcpy> vs. C<bcopy>. There are others, mostly in
e50bb9a1 307
308Mostly, this item is a suggestion for which way to start a journey into
309an C<#ifdef> forest. It is not primarily a suggestion to eliminate any of
310the C<#ifdef> forests.
e50bb9a1 311
312POSIX calls are perhaps more likely to be portable to unexpected
313architectures. They are also perhaps more likely to be actively
314maintained by a current vendor. They are also perhaps more likely to be
315available in thread-safe versions, if appropriate.
e50bb9a1 316
722d2a37 317=head2 -i rename file when changed
e50bb9a1 318
319It's only necessary to rename a file when inplace editing when the file
320has changed. Detecting a change is perhaps the difficult bit.
e50bb9a1 321
722d2a37 322=head2 All ARGV input should act like E<lt>E<gt>
e50bb9a1 323
722d2a37 324=head2 Support for rerunning debugger
e50bb9a1 325
722d2a37 326There should be a way of restarting the debugger on demand.
e50bb9a1 327
722d2a37 328=head2 my sub foo { }
c47ff5f1 329
330The basic principle is sound, but there are problems with the semantics
331of self-referential and mutually referential lexical subs: how to
332declare the subs?
c47ff5f1 333
722d2a37 334=head2 One-pass global destruction
c47ff5f1 335
336Sweeping away all the allocated memory in one go is a laudable goal, but
337it's difficult and in most cases, it's easier to let the memory get
338freed by exiting.
e50bb9a1 339
722d2a37 340=head2 Rewrite regexp parser
e50bb9a1 341
342There has been talk recently of rewriting the regular expression parser
343to produce an optree instead of a chain of opcodes; it's unclear whether
344or not this would be a win.
e50bb9a1 345
722d2a37 346=head2 Cache recently used regexps
e50bb9a1 347
722d2a37 348This is to speed up
e50bb9a1 349
350 for my $re (@regexps) {
351 $matched++ if /$re/
352 }
e50bb9a1 353
354C<qr//> already gives us a way of saving compiled regexps, but it should
355be done automatically.
e50bb9a1 356
722d2a37 357=head2 Re-entrant functions
e50bb9a1 358
359Add configure probes for C<_r> forms of system calls and fit them to the
360core. Unfortunately, calling conventions for these functions and not
04c70446 362
722d2a37 363=head2 Cross-compilation support
04c70446 364
365Bart Schuller reports that using C<microperl> and a cross-compiler, he
366got Perl working on the Agenda PDA. However, one cannot build a full
367Perl because Configure needs to get the results for the target platform,
368for the host.
e50bb9a1 369
722d2a37 370=head2 Bit-shifting bitvectors
e50bb9a1 371
722d2a37 372Given:
e50bb9a1 373
722d2a37 374 vec($v, 1000, 1) = 1;
e50bb9a1 375
722d2a37 376One should be able to do
e50bb9a1 377
722d2a37 378 $v <<= 1;
e50bb9a1 379
722d2a37 380and have the 999'th bit set.
e50bb9a1 381
382Currently if you try with shift bitvectors you shift the NV/UV, instead
383of the bits in the PV. Not very logical.
e50bb9a1 384
722d2a37 385=head2 debugger pragma
e50bb9a1 386
387The debugger is implemented in Perl in F<>; turning it into a
388pragma should be easy, but making it work lexically might be more
389difficult. Fiddling with C<$^P> would be necessary.
e50bb9a1 390
722d2a37 391=head2 use less pragma
e50bb9a1 392
393Identify areas where speed/memory tradeoffs can be made and have a hint
394to switch between them.
e50bb9a1 395
722d2a37 396=head2 switch structures
e50bb9a1 397
398Although we have C<> in core, Larry points to the dormant
399C<nswitch> and C<cswitch> ops in F<pp.c>; using these opcodes would be
400much faster.
e50bb9a1 401
722d2a37 402=head2 Cache eval tree
e50bb9a1 403
722d2a37 404=head2 rcatmaybe
e50bb9a1 405
722d2a37 406=head2 Shrink opcode tables
e50bb9a1 407
722d2a37 408=head2 Optimize away @_
e50bb9a1 409
722d2a37 410Look at the "reification" code in C<av.c>
e50bb9a1 411
722d2a37 412=head2 Prototypes versus indirect objects
e50bb9a1 413
722d2a37 414Currently, indirect object syntax bypasses prototype checks.
e50bb9a1 415
722d2a37 416=head2 Install HMTL
e50bb9a1 417
418HTML versions of the documentation need to be installed by default; a
419call to C<installhtml> from C<installperl> may be all that's necessary.
e50bb9a1 420
722d2a37 421=head2 Prototype method calls
e50bb9a1 422
722d2a37 423=head2 Return context prototype declarations
e50bb9a1 424
722d2a37 425=head2 magic_setisa
e50bb9a1 426
722d2a37 427=head2 Garbage collection
e50bb9a1 428
429There have been persistent mumblings about putting a mark-and-sweep
430garbage detector into Perl; Alan Burlison has some ideas about this.
e50bb9a1 431
722d2a37 432=head2 IO tutorial
e50bb9a1 433
722d2a37 434Mark-Jason Dominus has the beginnings of one of these.
e50bb9a1 435
722d2a37 436=head2 pack/unpack tutorial
e50bb9a1 437
722d2a37 438Simon Cozens has the beginnings of one of these.
e50bb9a1 439
722d2a37 440=head2 Rewrite perldoc
e50bb9a1 441
442There are a few suggestions for what to do with C<perldoc>: maybe a
443full-text search, an index function, locating pages on a particular
444high-level subject, and so on.
e50bb9a1 445
3958b146 446=head2 Install .3p manpages
e50bb9a1 447
3958b146 448This is a bone of contention; we can create C<.3p> manpages for each
449built-in function, but should we install them by default? Tcl does this,
450and it clutters up C<apropos>.
e50bb9a1 451
722d2a37 452=head2 Unicode tutorial
e50bb9a1 453
722d2a37 454Simon Cozens promises to do this before he gets old.
e50bb9a1 455
722d2a37 456=head2 Update for 1003.1-2
3958b146 457
722d2a37 458=head2 Retargetable installation
e50bb9a1 459
722d2a37 460Allow C<@INC> to be changed after Perl is built.
e50bb9a1 461
722d2a37 462=head2 POSIX emulation on non-POSIX systems
e50bb9a1 463
464Make C<> behave as POSIXly as possible everywhere, meaning we
465have to implement POSIX equivalents for some functions if necessary.
e50bb9a1 466
722d2a37 467=head2 Rename Win32 headers
e50bb9a1 468
469=head2 Finish off lvalue functions
471They don't work in the debugger, and they don't work for list or hash
e50bb9a1 473
722d2a37 474=head2 Update sprintf documentation
e50bb9a1 475
722d2a37 476Hugo van der Sanden plans to look at this.
e50bb9a1 477
722d2a37 478=head2 Use fchown/fchmod internally
e50bb9a1 479
480This has been done in places, but needs a thorough code review.
481Also fchdir is available in some platforms.
e50bb9a1 482
722d2a37 483=head1 Vague ideas
e50bb9a1 484
722d2a37 485Ideas which have been discussed, and which may or may not happen.
e50bb9a1 486
722d2a37 487=head2 ref() in list context
e50bb9a1 488
489It's unclear what this should do or how to do it without breaking old
e50bb9a1 491
722d2a37 492=head2 Make tr/// return histogram
e50bb9a1 493
722d2a37 494There is a patch for this, but it may require Unicodification.
e50bb9a1 495
722d2a37 496=head2 Compile to real threaded code
3958b146 497
722d2a37 498=head2 Structured types
3958b146 499
722d2a37 500=head2 Modifiable $1 et al.
e50bb9a1 501
502 ($x = "elephant") =~ /e(ph)/;
503 $1 = "g"; # $x = "elegant"
e50bb9a1 504
505What happens if there are multiple (nested?) brackets? What if the
506string changes between the match and the assignment?
e50bb9a1 507
722d2a37 508=head2 Procedural interfaces for IO::*, etc.
e50bb9a1 509
510Some core modules have been accused of being overly-OO. Adding
511procedural interfaces could demystify them.
e50bb9a1 512
722d2a37 513=head2 RPC modules
e50bb9a1 514
722d2a37 515=head2 Attach/detach debugger from running program
e50bb9a1 516
517With C<gdb>, you can attach the debugger to a running program if you
518pass the process ID. It would be good to do this with the Perl debugger
519on a running Perl program, although I'm not sure how it would be done.
e50bb9a1 520
722d2a37 521=head2 Alternative RE syntax module
e50bb9a1 522
523 use Regex::Newbie;
524 $re = Regex::Newbie->new
525 ->start
526 ->match("foo")
527 ->repeat(Regex::Newbie->class("char"),3)
528 ->end;
529 /$re/;
e50bb9a1 530
722d2a37 531=head2 GUI::Native
e50bb9a1 532
533A non-core module that would use "native" GUI to create graphical
e50bb9a1 535
722d2a37 536=head2 foreach(reverse ...)
e50bb9a1 537
722d2a37 538Currently
e50bb9a1 539
722d2a37 540 foreach (reverse @_) { ... }
e50bb9a1 541
542puts C<@_> on the stack, reverses it putting the reversed version on the
543stack, then iterates forwards. Instead, it could be special-cased to put
544C<@_> on the stack then iterate backwards.
e50bb9a1 545
722d2a37 546=head2 Constant function cache
e50bb9a1 547
722d2a37 548=head2 Approximate regular expression matching
e50bb9a1 549
722d2a37 550=head1 Ongoing
e50bb9a1 551
722d2a37 552These items B<always> need doing:
e50bb9a1 553
722d2a37 554=head2 Update guts documentation
e50bb9a1 555
556Simon Cozens tries to do this when possible, and contributions to the
557C<perlapi> documentation is welcome.
e50bb9a1 558
722d2a37 559=head2 Add more tests
e50bb9a1 560
561Michael Schwern will donate $500 to Yet Another Society when all core
562modules have tests.
e50bb9a1 563
722d2a37 564=head2 Update auxiliary tools
e50bb9a1 565
722d2a37 566The code we ship with Perl should look like good Perl 5.
e50bb9a1 567
722d2a37 568=head1 Recently done things
e50bb9a1 569
570These are things which have been on the todo lists in previous releases
571but have recently been completed.
e50bb9a1 572
722d2a37 573=head2 Safe signal handling
e50bb9a1 574
575A new signal model went into 5.7.1 without much fanfare. Operations and
576C<malloc>s are no longer interrupted by signals, which are handled
577between opcodes. This means that C<PERL_ASYNC_CHECK> now actually does
578something. However, there are still a few things that need to be done.
e50bb9a1 579
722d2a37 580=head2 Tie Modules
e50bb9a1 581
582Modules which implement arrays in terms of strings, substrings or files
583can be found on the CPAN.
e50bb9a1 584
722d2a37 585=head2 gettimeofday
e50bb9a1 586
722d2a37 587C<Time::Hires> has been integrated into the core.
e50bb9a1 588
722d2a37 589=head2 setitimer and getimiter
e50bb9a1 590
722d2a37 591Adding C<Time::Hires> got us this too.
e50bb9a1 592
593=head2 Testing __DIE__ hook
595Tests have been added.
597=head2 CPP equivalent in Perl
e50bb9a1 598
599A C Yardley will probably have done this by the time you can read this.
600This allows for a generalization of the C constant detection used in
601building C<>.
e50bb9a1 602
722d2a37 603=head2 Explicit switch statements
e50bb9a1 604
605C<> has been integrated into the core to give you all manner of
606C<> semantics.
e50bb9a1 607
722d2a37 608=head2 autocroak
e50bb9a1 609
722d2a37 610This is C<>.
e50bb9a1 611
722d2a37 612=head2 UTF/EBCDIC
e50bb9a1 613
722d2a37 614Nick Ing-Simmons has made UTF-EBCDIC (UTR13) work with Perl.
e50bb9a1 615
722d2a37 616 EBCDIC?
e50bb9a1 617
722d2a37 618=head2 UTF Regexes
e50bb9a1 619
620Although there are probably some small bugs to be rooted out, Jarkko
621Hietaniemi has made regular expressions polymorphic between bytes and
e50bb9a1 623
722d2a37 624=head2 perlcc to produce executable
e50bb9a1 625
626C<perlcc> was recently rewritten, and can now produce standalone
e50bb9a1 628
722d2a37 629=head2 END blocks saved in compiled output
e50bb9a1 630
722d2a37 631=head2 Secure temporary file module
e50bb9a1 632
722d2a37 633Tim Jenness' C<File::Temp> is now in core.
e50bb9a1 634
722d2a37 635=head2 Integrate Time::HiRes
e50bb9a1 636
722d2a37 637This module is now part of core.
e50bb9a1 638
722d2a37 639=head2 Turn Cwd into XS
e50bb9a1 640
722d2a37 641Benjamin Sugars has done this.
e50bb9a1 642
722d2a37 643=head2 Mmap for input
e50bb9a1 644
722d2a37 645Nick Ing-Simmons' C<perlio> supports an C<mmap> IO method.
e50bb9a1 646
722d2a37 647=head2 Byte to/from UTF8 and UTF8 to/from local conversion
e50bb9a1 648
722d2a37 649C<Encode> provides this.
e50bb9a1 650
722d2a37 651=head2 Add sockatmark support
e50bb9a1 652
722d2a37 653Added in 5.7.1
e50bb9a1 654
655=head2 Mailing list archives
659=head2 Bug tracking
661Richard Foley has written the bug tracking system at
e50bb9a1 662
722d2a37 663=head2 Integrate MacPerl
e50bb9a1 664
665Chris Nandor and Matthias Neeracher have integrated the MacPerl changes
666into 5.6.0.
e50bb9a1 667
722d2a37 668=head2 Web "nerve center" for Perl
e50bb9a1 669
722d2a37 670 is what you're looking for.
e50bb9a1 671
722d2a37 672=head2 Regular expression tutorial
e50bb9a1 673
722d2a37 674C<perlretut>, provided by Mark Kvale.
e50bb9a1 675
722d2a37 676=head2 Debugging Tutorial
e50bb9a1 677
722d2a37 678C<perldebtut>, written by Richard Foley.
e50bb9a1 679
722d2a37 680=head2 Integrate new modules
e50bb9a1 681
722d2a37 682Jarkko has been integrating madly into 5.7.x
e50bb9a1 683
722d2a37 684=head2 Integrate profiler
e50bb9a1 685
722d2a37 686C<Devel::DProf> is now a core module.
e50bb9a1 687
722d2a37 688=head2 Y2K error detection
e50bb9a1 689
690There's a configure option to detect unsafe concatenation with "19", and
691a CPAN module. (C<D'oh::Year>)
e50bb9a1 692
722d2a37 693=head2 Regular expression debugger
e50bb9a1 694
695While not part of core, Mark-Jason Dominus has written C<Rx> and has
696also come up with a generalised strategy for regular expression
e50bb9a1 698
722d2a37 699=head2 POD checker
e50bb9a1 700
722d2a37 701That's, uh, F<podchecker>
e50bb9a1 702
722d2a37 703=head2 "Dynamic" lexicals
e50bb9a1 704
722d2a37 705=head2 Cache precompiled modules
e50bb9a1 706
722d2a37 707=head1 Deprecated Wishes
e50bb9a1 708
709These are items which used to be in the todo file, but have been
710deprecated for some reason.
e50bb9a1 711
722d2a37 712=head2 Loop control on do{}
e50bb9a1 713
722d2a37 714This would break old code; use C<do{{ }}> instead.
e50bb9a1 715
722d2a37 716=head2 Lexically scoped typeglobs
e50bb9a1 717
722d2a37 718Not needed now we have lexical IO handles.
e50bb9a1 719
722d2a37 720=head2 format BOTTOM
3958b146 721
722d2a37 722=head2 report HANDLE
e50bb9a1 723
722d2a37 724Damian Conway's text formatting modules seem to be the Way To Go.
e50bb9a1 725
722d2a37 726=head2 Generalised want()/caller())
3958b146 727
722d2a37 728=head2 Named prototypes
e50bb9a1 729
722d2a37 730These both seem to be delayed until Perl 6.
e50bb9a1 731
722d2a37 732=head2 Built-in globbing
e50bb9a1 733
722d2a37 734The C<File::Glob> module has been used to replace the C<glob> function.
e50bb9a1 735
722d2a37 736=head2 Regression tests for suidperl
e50bb9a1 737
722d2a37 738C<suidperl> is deprecated in favour of common sense.
e50bb9a1 739
722d2a37 740=head2 Cached hash values
e50bb9a1 741
722d2a37 742We have shared hash keys, which perform the same job.
e50bb9a1 743
722d2a37 744=head2 Add compression modules
e50bb9a1 745
746The compression modules are a little heavy; meanwhile, Nick Clark is
747working on experimental pragmata to do transparent decompression on
e50bb9a1 749
722d2a37 750=head2 Reorganise documentation into tutorials/references
e50bb9a1 751
722d2a37 752Could not get consensus on P5P about this.
e50bb9a1 753
754=head2 Remove distinction between functions and operators
756Caution: highly flammable.
758=head2 Make XS easier to use
e50bb9a1 759
722d2a37 760Use C<Inline> instead, or SWIG.
e50bb9a1 761
722d2a37 762=head2 Make embedding easier to use
e50bb9a1 763
722d2a37 764Use C<Inline::CPR>.
e50bb9a1 765
722d2a37 766=head2 man for perl
04c70446 767
722d2a37 768See the Perl Power Tools. (
04c70446 769
722d2a37 770=head2 my $Package::variable
04c70446 771
722d2a37 772Use C<our> instead.
04c70446 773
722d2a37 774=head2 "or" tests defined, not truth
04c70446 775
722d2a37 776Suggesting this on P5P B<will> cause a boring and interminable flamewar.
04c70446 777
722d2a37 778=head2 "class"-based lexicals
04c70446 779
cbb3fa72 780Use flyweight objects, secure hashes or, dare I say it, pseudo-hashes instead.
04c70446 781
722d2a37 782=head2 byteperl
04c70446 783
722d2a37 784C<ByteLoader> covers this.
04c70446 785
722d2a37 786=head2 Lazy evaluation / tail recursion removal
04c70446 787
788C<List::Util> in core gives some of these; tail recursion removal is
789done manually, with C<goto &whoami;>. (However, MJD has found that
790C<goto &whoami> introduces a performance penalty, so maybe there should
791be a way to do this after all: C<sub foo {START: ... goto START;> is
794=head2 Make "use utf8" the default
796There is a patch available for this, search p5p archives for
797the Subject "[EXPERIMENTAL PATCH] make unicode (utf8) default"
798but this would be unacceptable because of backward compatibility:
799scripts could not contain B<any legacy eight-bit data>. Also would
800introduce a measurable slowdown of at least few percentages since all
801regular expression operations would be done in full UTF-8.