Commit | Line | Data |
7711098a |
1 | =head1 NAME |
2 | |
3 | perltodo - Perl TO-DO List |
4 | |
5 | =head1 DESCRIPTION |
e50bb9a1 |
6 | |
722d2a37 |
7 | This is a list of wishes for Perl. Send updates to |
e50bb9a1 |
8 | I<perl5-porters@perl.org>. If you want to work on any of these |
9 | projects, be sure to check the perl5-porters archives for past ideas, |
10 | flames, and propaganda. This will save you time and also prevent you |
11 | from implementing something that Larry has already vetoed. One set |
12 | of archives may be found at: |
13 | |
14 | http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/ |
15 | |
722d2a37 |
16 | =head1 To do during 5.6.x |
e50bb9a1 |
17 | |
722d2a37 |
18 | =head2 Support for I/O disciplines |
e50bb9a1 |
19 | |
722d2a37 |
20 | C<perlio> provides this, but the interface could be a lot more |
21 | straightforward. |
e50bb9a1 |
22 | |
722d2a37 |
23 | =head2 Eliminate need for "use utf8"; |
e50bb9a1 |
24 | |
722d2a37 |
25 | While the C<utf8> pragma is autoloaded when necessary, it's still needed |
26 | for things like Unicode characters in a source file. The UTF8 hint can |
27 | always be set to true, but it needs to be set to false when F<utf8.pm> |
28 | is being compiled. (To stop Perl trying to autoload the C<utf8> |
29 | pragma...) |
e50bb9a1 |
30 | |
c6287c21 |
31 | =head2 Create a char *sv_pvprintify(sv, STRLEN *lenp, UV flags) |
0562c0e3 |
32 | |
33 | For displaying PVs with control characters, embedded nulls, and Unicode. |
34 | This would be useful for printing warnings, or data and regex dumping, |
35 | not_a_number(), and so on. |
36 | |
f35392ae |
37 | Requirements: should handle both byte and UTF8 strings. isPRINT() |
38 | characters printed as-is, character less than 256 as \xHH, Unicode |
0661e9a4 |
39 | characters as \x{HHH}. Don't assume ASCII-like, either, get somebody |
40 | on EBCDIC to test the output. |
f35392ae |
41 | |
42 | Possible options, controlled by the flags: |
0661e9a4 |
43 | - whitespace (other than ' ' of isPRINT()) printed as-is |
f35392ae |
44 | - use isPRINT_LC() instead of isPRINT() |
45 | - print control characters like this: "\cA" |
46 | - print control characters like this: "^A" |
0661e9a4 |
47 | - non-PRINTables printed as '.' instead of \xHH |
48 | - use \OOO instead of \xHH |
49 | - use the C/Perl-metacharacters like \n, \t |
f35392ae |
50 | - have a maximum length for the produced string (read it from *lenp) |
51 | - append a "..." to the produced string if the maximum length is exceeded |
0661e9a4 |
52 | - really fancy: print unicode characters as \N{...} |
f35392ae |
53 | |
722d2a37 |
54 | =head2 Autoload byte.pm |
e50bb9a1 |
55 | |
722d2a37 |
56 | When the lexer sees, for instance, C<bytes::length>, it should |
57 | automatically load the C<bytes> pragma. |
e50bb9a1 |
58 | |
722d2a37 |
59 | =head2 Make "\u{XXXX}" et al work |
e50bb9a1 |
60 | |
722d2a37 |
61 | Danger, Will Robinson! Discussing the semantics of C<"\x{F00}">, |
62 | C<"\xF00"> and C<"\U{F00}"> on P5P I<will> lead to a long and boring |
63 | flamewar. |
e50bb9a1 |
64 | |
722d2a37 |
65 | =head2 Overloadable regex assertions |
e50bb9a1 |
66 | |
722d2a37 |
67 | This may or may not be possible with the current regular expression |
68 | engine. The idea is that, for instance, C<\b> needs to be |
69 | algorithmically computed if you're dealing with Thai text. Hence, the |
70 | B<\b> assertion wants to be overloaded by a function. |
e50bb9a1 |
71 | |
722d2a37 |
72 | =head2 Unicode collation and normalization |
e50bb9a1 |
73 | |
722d2a37 |
74 | Simon Cozens promises to work on this. |
e50bb9a1 |
75 | |
722d2a37 |
76 | Collation? http://www.unicode.org/unicode/reports/tr10/ |
77 | Normalization? http://www.unicode.org/unicode/reports/tr15/ |
e50bb9a1 |
78 | |
722d2a37 |
79 | =head2 Unicode case mappings |
e50bb9a1 |
80 | |
722d2a37 |
81 | Case Mappings? http://www.unicode.org/unicode/reports/tr21/ |
e50bb9a1 |
82 | |
722d2a37 |
83 | =head2 Unicode regular expression character classes |
e50bb9a1 |
84 | |
c6287c21 |
85 | They have some tricks Perl doesn't yet implement like character |
86 | class subtraction. |
e50bb9a1 |
87 | |
722d2a37 |
88 | http://www.unicode.org/unicode/reports/tr18/ |
e50bb9a1 |
89 | |
c6287c21 |
90 | =head2 Unicode Scripts support |
91 | |
92 | Currently the C<\p{In...}> supports only the Blocks database, like |
93 | C<\p{BasicLatin}>, C<\p{InGreek}>, C<\p{InThai}>, but there's also the |
94 | Scripts database, which has members like C<Latin>, C<Greek>, |
95 | C<Armenian>, C<Han>. It is desireable that also the script names |
96 | could be used for the C<\p{In...}> construct. Note: needs to be |
97 | researched whether this is possible, that is, are there conflicts |
98 | between the Blocks and the Scripts, is the Blocks Greek the same as |
99 | the Scripts Greek? |
100 | |
722d2a37 |
101 | =head2 use Thread for iThreads |
e50bb9a1 |
102 | |
722d2a37 |
103 | Artur Bergman's C<iThreads> module is a start on this, but needs to |
104 | be more mature. |
e50bb9a1 |
105 | |
dd0afe54 |
106 | =head2 make perl_clone optionally clone ops |
107 | |
108 | So that pseudoforking, mod_perl, iThreads and nvi will work properly |
109 | (but not as efficiently) until the regex engine is fixed to be threadsafe. |
110 | |
722d2a37 |
111 | =head2 Work out exit/die semantics for threads |
e50bb9a1 |
112 | |
722d2a37 |
113 | =head2 Typed lexicals for compiler |
e50bb9a1 |
114 | |
722d2a37 |
115 | =head2 Compiler workarounds for Win32 |
e50bb9a1 |
116 | |
722d2a37 |
117 | =head2 AUTOLOADing in the compiler |
e50bb9a1 |
118 | |
722d2a37 |
119 | =head2 Fixing comppadlist when compiling |
e50bb9a1 |
120 | |
722d2a37 |
121 | =head2 Cleaning up exported namespace |
e50bb9a1 |
122 | |
722d2a37 |
123 | =head2 Complete signal handling |
e50bb9a1 |
124 | |
722d2a37 |
125 | Add C<PERL_ASYNC_CHECK> to opcodes which loop; replace C<sigsetjmp> with |
126 | C<sigjmp>; check C<wait> for signal safety. |
e50bb9a1 |
127 | |
722d2a37 |
128 | =head2 Out-of-source builds |
e50bb9a1 |
129 | |
722d2a37 |
130 | This was done for 5.6.0, but needs reworking for 5.7.x |
e50bb9a1 |
131 | |
722d2a37 |
132 | =head2 POSIX realtime support |
e50bb9a1 |
133 | |
722d2a37 |
134 | POSIX 1003.1 1996 Edition support--realtime stuff: POSIX semaphores, |
135 | message queues, shared memory, realtime clocks, timers, signals (the |
136 | metaconfig units mostly already exist for these) |
e50bb9a1 |
137 | |
722d2a37 |
138 | =head2 UNIX98 support |
e50bb9a1 |
139 | |
722d2a37 |
140 | Reader-writer locks, realtime/asynchronous IO |
e50bb9a1 |
141 | |
722d2a37 |
142 | =head2 IPv6 Support |
e50bb9a1 |
143 | |
722d2a37 |
144 | There are non-core modules, such as C<Net::IPv6>, but these will need |
145 | integrating when IPv6 actually starts to really happen. See RFC 2292 |
146 | and RFC 2553. |
e50bb9a1 |
147 | |
722d2a37 |
148 | =head2 Long double conversion |
e50bb9a1 |
149 | |
722d2a37 |
150 | Floating point formatting is still causing some weird test failures. |
e50bb9a1 |
151 | |
722d2a37 |
152 | =head2 Locales |
e50bb9a1 |
153 | |
722d2a37 |
154 | Locales and Unicode interact with each other in unpleasant ways. |
155 | One possible solution would be to adopt/support ICU: |
e50bb9a1 |
156 | |
722d2a37 |
157 | http://oss.software.ibm.com/developerworks/opensource/icu/project/ |
e50bb9a1 |
158 | |
722d2a37 |
159 | =head2 Thread-safe regexes |
e50bb9a1 |
160 | |
722d2a37 |
161 | The regular expression engine is currently non-threadsafe. |
e50bb9a1 |
162 | |
722d2a37 |
163 | =head2 Arithmetic on non-Arabic numerals |
e50bb9a1 |
164 | |
722d2a37 |
165 | C<[1234567890]> aren't the only numerals any more. |
e50bb9a1 |
166 | |
722d2a37 |
167 | =head2 POSIX Unicode character classes |
e50bb9a1 |
168 | |
722d2a37 |
169 | ([=a=] for equivalance classes, [.ch.] for collation.) |
170 | These are dependent on Unicode normalization and collation. |
e50bb9a1 |
171 | |
722d2a37 |
172 | =head2 Factoring out common suffices/prefices in regexps (trie optimization) |
c47ff5f1 |
173 | |
722d2a37 |
174 | Currently, the user has to optimize C<foo|far> and C<foo|goo> into |
175 | C<f(?:oo|ar)> and C<[fg]oo> by hand; this could be done automatically. |
e50bb9a1 |
176 | |
722d2a37 |
177 | =head2 Security audit shipped utilities |
e50bb9a1 |
178 | |
722d2a37 |
179 | All the code we ship with Perl needs to be sensible about temporary file |
180 | handling, locking, input validation, and so on. |
e50bb9a1 |
181 | |
722d2a37 |
182 | =head2 Custom opcodes |
e50bb9a1 |
183 | |
722d2a37 |
184 | Have a way to introduce user-defined opcodes without the subroutine call |
185 | overhead of an XSUB; the user should be able to create PP code. Simon |
186 | Cozens has some ideas on this. |
e50bb9a1 |
187 | |
722d2a37 |
188 | =head2 spawnvp() on Win32 |
e50bb9a1 |
189 | |
722d2a37 |
190 | Win32 has problems spawning processes, particularly when the arguments |
191 | to the child process contain spaces, quotes or tab characters. |
e50bb9a1 |
192 | |
722d2a37 |
193 | =head2 DLL Versioning |
e50bb9a1 |
194 | |
722d2a37 |
195 | Windows needs a way to know what version of a XS or C<libperl> DLL it's |
196 | loading. |
e50bb9a1 |
197 | |
722d2a37 |
198 | =head2 Introduce @( and @) |
e50bb9a1 |
199 | |
722d2a37 |
200 | C<$(> may return "foo bar baz". Unfortunately, since groups can |
201 | theoretically have spaces in their names, this could be one, two or |
202 | three groups. |
e50bb9a1 |
203 | |
722d2a37 |
204 | =head2 Floating point handling |
e50bb9a1 |
205 | |
722d2a37 |
206 | C<NaN> and C<inf> support is particularly troublesome. |
207 | (fp_classify(), fp_class(), fp_class_d(), class(), isinf(), |
208 | isfinite(), finite(), isnormal(), unordered(), <ieeefp.h>, |
209 | <fp_class.h> (there are metaconfig units for all these) (I think), |
210 | fp_setmask(), fp_getmask(), fp_setround(), fp_getround() |
211 | (no metaconfig units yet for these). Don't forget finitel(), fp_classl(), |
212 | fp_class_l(), (yes, both do, unfortunately, exist), and unorderedl().) |
e50bb9a1 |
213 | |
722d2a37 |
214 | As of Perl 5.6.1 is a Perl macro, Perl_isnan(). |
e50bb9a1 |
215 | |
722d2a37 |
216 | =head2 IV/UV preservation |
e50bb9a1 |
217 | |
722d2a37 |
218 | Nicholas Clark has done a lot of work on this, but work is continuing. |
219 | C<+>, C<-> and C<*> work, but guards need to be in place for C<%>, C</>, |
220 | C<&>, C<oct>, C<hex> and C<pack>. |
e50bb9a1 |
221 | |
722d2a37 |
222 | =head2 Replace pod2html with something using Pod::Parser |
83df6a1d |
223 | |
722d2a37 |
224 | The CPAN module C<Malik::Pod::Html> may be a more suitable basis for a |
225 | C<pod2html> convertor; the current one duplicates the functionality |
226 | abstracted in C<Pod::Parser>, which makes updating the POD language |
227 | difficult. |
e50bb9a1 |
228 | |
722d2a37 |
229 | =head2 Automate module testing on CPAN |
e50bb9a1 |
230 | |
722d2a37 |
231 | When a new Perl is being beta tested, porters have to manually grab |
232 | their favourite CPAN modules and test them - this should be done |
233 | automatically. |
e50bb9a1 |
234 | |
722d2a37 |
235 | =head2 sendmsg and recvmsg |
83df6a1d |
236 | |
722d2a37 |
237 | We have all the other BSD socket functions but these. There are |
238 | metaconfig units for these functions which can be added. To avoid these |
239 | being new opcodes, a solution similar to the way C<sockatmark> was added |
240 | would be preferable. (Autoload the C<IO::whatever> module.) |
e50bb9a1 |
241 | |
722d2a37 |
242 | =head2 Rewrite perlre documentation |
e50bb9a1 |
243 | |
722d2a37 |
244 | The new-style patterns need full documentation, and the whole document |
245 | needs to be a lot clearer. |
e50bb9a1 |
246 | |
722d2a37 |
247 | =head2 Convert example code to IO::Handle filehandles |
e50bb9a1 |
248 | |
722d2a37 |
249 | =head2 Document Win32 choices |
e50bb9a1 |
250 | |
722d2a37 |
251 | =head2 Check new modules |
e50bb9a1 |
252 | |
722d2a37 |
253 | =head2 Make roffitall find pods and libs itself |
e50bb9a1 |
254 | |
722d2a37 |
255 | Simon Cozens has done some work on this but it needs a rethink. |
e50bb9a1 |
256 | |
722d2a37 |
257 | =head1 To do at some point |
e50bb9a1 |
258 | |
722d2a37 |
259 | These are ideas that have been regularly tossed around, that most |
260 | people believe should be done maybe during 5.8.x |
e50bb9a1 |
261 | |
722d2a37 |
262 | =head2 Remove regular expression recursion |
e50bb9a1 |
263 | |
722d2a37 |
264 | Because the regular expression engine is recursive, badly designed |
265 | expressions can lead to lots of recursion filling up the stack. Ilya |
266 | claims that it is easy to convert the engine to being iterative, but |
267 | this has still not yet been done. There may be a regular expression |
268 | engine hit squad meeting at TPC5. |
e50bb9a1 |
269 | |
722d2a37 |
270 | =head2 Memory leaks after failed eval |
e50bb9a1 |
271 | |
722d2a37 |
272 | Perl will leak memory if you C<eval "hlagh hlagh hlagh hlagh">. This is |
273 | partially because it attempts to build up an op tree for that code and |
274 | doesn't properly free it. The same goes for non-syntactically-correct |
275 | regular expressions. Hugo looked into this, but decided it needed a |
276 | mark-and-sweep GC implementation. |
e50bb9a1 |
277 | |
722d2a37 |
278 | Alan notes that: The basic idea was to extend the parser token stack |
279 | (C<YYSTYPE>) to include a type field so we knew what sort of thing each |
280 | element of the stack was. The F<<perly.c> code would then have to be |
281 | postprocessed to record the type of each entry on the stack as it was |
282 | created, and the parser patched so that it could unroll the stack |
283 | properly on error. |
e50bb9a1 |
284 | |
722d2a37 |
285 | This is possible to do, but would be pretty messy to implement, as it |
286 | would rely on even more sed hackery in F<perly.fixer>. |
e50bb9a1 |
287 | |
722d2a37 |
288 | =head2 pack "(stuff)*" |
e50bb9a1 |
289 | |
722d2a37 |
290 | That's to say, C<pack "(sI)40"> would be the same as C<pack "sI"x40> |
e50bb9a1 |
291 | |
722d2a37 |
292 | =head2 bitfields in pack |
e50bb9a1 |
293 | |
722d2a37 |
294 | =head2 Cross compilation |
e50bb9a1 |
295 | |
722d2a37 |
296 | Make Perl buildable with a cross-compiler. This will play havoc with |
297 | Configure, which needs to how how the target system will respond to |
298 | its tests; maybe C<microperl> will be a good starting point here. |
299 | (Indeed, Bart Schuller reports that he compiled up C<microperl> for |
300 | the Agenda PDA and it works fine.) A really big spanner in the works |
301 | is the bootstrapping build process of Perl: if the filesystem the |
302 | target systems sees is not the same what the build host sees, various |
303 | input, output, and (Perl) library files need to be copied back and forth. |
e50bb9a1 |
304 | |
722d2a37 |
305 | =head2 Perl preprocessor / macros |
e50bb9a1 |
306 | |
722d2a37 |
307 | Source filters help with this, but do not get us all the way. For |
308 | instance, it should be possible to implement the C<??> operator somehow; |
309 | source filters don't (quite) cut it. |
e50bb9a1 |
310 | |
722d2a37 |
311 | =head2 Perl lexer in Perl |
a45bd81d |
312 | |
722d2a37 |
313 | Damian Conway is planning to work on this, but it hasn't happened yet. |
e50bb9a1 |
314 | |
722d2a37 |
315 | =head2 Using POSIX calls internally |
e50bb9a1 |
316 | |
722d2a37 |
317 | When faced with a BSD vs. SySV -style interface to some library or |
318 | system function, perl's roots show in that it typically prefers the BSD |
319 | interface (but falls back to the SysV one). One example is getpgrp(). |
320 | Other examples include C<memcpy> vs. C<bcopy>. There are others, mostly in |
321 | F<<pp_sys.c>. |
e50bb9a1 |
322 | |
722d2a37 |
323 | Mostly, this item is a suggestion for which way to start a journey into |
324 | an C<#ifdef> forest. It is not primarily a suggestion to eliminate any of |
325 | the C<#ifdef> forests. |
e50bb9a1 |
326 | |
722d2a37 |
327 | POSIX calls are perhaps more likely to be portable to unexpected |
328 | architectures. They are also perhaps more likely to be actively |
329 | maintained by a current vendor. They are also perhaps more likely to be |
330 | available in thread-safe versions, if appropriate. |
e50bb9a1 |
331 | |
722d2a37 |
332 | =head2 -i rename file when changed |
e50bb9a1 |
333 | |
722d2a37 |
334 | It's only necessary to rename a file when inplace editing when the file |
335 | has changed. Detecting a change is perhaps the difficult bit. |
e50bb9a1 |
336 | |
722d2a37 |
337 | =head2 All ARGV input should act like E<lt>E<gt> |
e50bb9a1 |
338 | |
722d2a37 |
339 | =head2 Support for rerunning debugger |
e50bb9a1 |
340 | |
722d2a37 |
341 | There should be a way of restarting the debugger on demand. |
e50bb9a1 |
342 | |
c6287c21 |
343 | =head2 Test Suite for the Debugger |
344 | |
345 | The debugger is a complex piece of software and fixing something |
346 | here may inadvertently break something else over there. To tame |
347 | this chaotic behaviour, a test suite is necessary. |
348 | |
722d2a37 |
349 | =head2 my sub foo { } |
c47ff5f1 |
350 | |
722d2a37 |
351 | The basic principle is sound, but there are problems with the semantics |
352 | of self-referential and mutually referential lexical subs: how to |
353 | declare the subs? |
c47ff5f1 |
354 | |
722d2a37 |
355 | =head2 One-pass global destruction |
c47ff5f1 |
356 | |
722d2a37 |
357 | Sweeping away all the allocated memory in one go is a laudable goal, but |
358 | it's difficult and in most cases, it's easier to let the memory get |
359 | freed by exiting. |
e50bb9a1 |
360 | |
722d2a37 |
361 | =head2 Rewrite regexp parser |
e50bb9a1 |
362 | |
722d2a37 |
363 | There has been talk recently of rewriting the regular expression parser |
364 | to produce an optree instead of a chain of opcodes; it's unclear whether |
365 | or not this would be a win. |
e50bb9a1 |
366 | |
722d2a37 |
367 | =head2 Cache recently used regexps |
e50bb9a1 |
368 | |
722d2a37 |
369 | This is to speed up |
e50bb9a1 |
370 | |
722d2a37 |
371 | for my $re (@regexps) { |
372 | $matched++ if /$re/ |
373 | } |
e50bb9a1 |
374 | |
722d2a37 |
375 | C<qr//> already gives us a way of saving compiled regexps, but it should |
376 | be done automatically. |
e50bb9a1 |
377 | |
722d2a37 |
378 | =head2 Re-entrant functions |
e50bb9a1 |
379 | |
722d2a37 |
380 | Add configure probes for C<_r> forms of system calls and fit them to the |
381 | core. Unfortunately, calling conventions for these functions and not |
382 | standardised. |
04c70446 |
383 | |
722d2a37 |
384 | =head2 Cross-compilation support |
04c70446 |
385 | |
722d2a37 |
386 | Bart Schuller reports that using C<microperl> and a cross-compiler, he |
387 | got Perl working on the Agenda PDA. However, one cannot build a full |
388 | Perl because Configure needs to get the results for the target platform, |
389 | for the host. |
e50bb9a1 |
390 | |
722d2a37 |
391 | =head2 Bit-shifting bitvectors |
e50bb9a1 |
392 | |
722d2a37 |
393 | Given: |
e50bb9a1 |
394 | |
722d2a37 |
395 | vec($v, 1000, 1) = 1; |
e50bb9a1 |
396 | |
722d2a37 |
397 | One should be able to do |
e50bb9a1 |
398 | |
722d2a37 |
399 | $v <<= 1; |
e50bb9a1 |
400 | |
722d2a37 |
401 | and have the 999'th bit set. |
e50bb9a1 |
402 | |
722d2a37 |
403 | Currently if you try with shift bitvectors you shift the NV/UV, instead |
404 | of the bits in the PV. Not very logical. |
e50bb9a1 |
405 | |
722d2a37 |
406 | =head2 debugger pragma |
e50bb9a1 |
407 | |
722d2a37 |
408 | The debugger is implemented in Perl in F<perl5db.pl>; turning it into a |
409 | pragma should be easy, but making it work lexically might be more |
410 | difficult. Fiddling with C<$^P> would be necessary. |
e50bb9a1 |
411 | |
722d2a37 |
412 | =head2 use less pragma |
e50bb9a1 |
413 | |
722d2a37 |
414 | Identify areas where speed/memory tradeoffs can be made and have a hint |
415 | to switch between them. |
e50bb9a1 |
416 | |
722d2a37 |
417 | =head2 switch structures |
e50bb9a1 |
418 | |
722d2a37 |
419 | Although we have C<Switch.pm> in core, Larry points to the dormant |
420 | C<nswitch> and C<cswitch> ops in F<pp.c>; using these opcodes would be |
421 | much faster. |
e50bb9a1 |
422 | |
722d2a37 |
423 | =head2 Cache eval tree |
e50bb9a1 |
424 | |
722d2a37 |
425 | =head2 rcatmaybe |
e50bb9a1 |
426 | |
722d2a37 |
427 | =head2 Shrink opcode tables |
e50bb9a1 |
428 | |
722d2a37 |
429 | =head2 Optimize away @_ |
e50bb9a1 |
430 | |
722d2a37 |
431 | Look at the "reification" code in C<av.c> |
e50bb9a1 |
432 | |
722d2a37 |
433 | =head2 Prototypes versus indirect objects |
e50bb9a1 |
434 | |
722d2a37 |
435 | Currently, indirect object syntax bypasses prototype checks. |
e50bb9a1 |
436 | |
722d2a37 |
437 | =head2 Install HMTL |
e50bb9a1 |
438 | |
722d2a37 |
439 | HTML versions of the documentation need to be installed by default; a |
440 | call to C<installhtml> from C<installperl> may be all that's necessary. |
e50bb9a1 |
441 | |
722d2a37 |
442 | =head2 Prototype method calls |
e50bb9a1 |
443 | |
722d2a37 |
444 | =head2 Return context prototype declarations |
e50bb9a1 |
445 | |
722d2a37 |
446 | =head2 magic_setisa |
e50bb9a1 |
447 | |
722d2a37 |
448 | =head2 Garbage collection |
e50bb9a1 |
449 | |
722d2a37 |
450 | There have been persistent mumblings about putting a mark-and-sweep |
451 | garbage detector into Perl; Alan Burlison has some ideas about this. |
e50bb9a1 |
452 | |
722d2a37 |
453 | =head2 IO tutorial |
e50bb9a1 |
454 | |
722d2a37 |
455 | Mark-Jason Dominus has the beginnings of one of these. |
e50bb9a1 |
456 | |
722d2a37 |
457 | =head2 pack/unpack tutorial |
e50bb9a1 |
458 | |
722d2a37 |
459 | Simon Cozens has the beginnings of one of these. |
e50bb9a1 |
460 | |
722d2a37 |
461 | =head2 Rewrite perldoc |
e50bb9a1 |
462 | |
722d2a37 |
463 | There are a few suggestions for what to do with C<perldoc>: maybe a |
464 | full-text search, an index function, locating pages on a particular |
465 | high-level subject, and so on. |
e50bb9a1 |
466 | |
3958b146 |
467 | =head2 Install .3p manpages |
e50bb9a1 |
468 | |
3958b146 |
469 | This is a bone of contention; we can create C<.3p> manpages for each |
722d2a37 |
470 | built-in function, but should we install them by default? Tcl does this, |
471 | and it clutters up C<apropos>. |
e50bb9a1 |
472 | |
722d2a37 |
473 | =head2 Unicode tutorial |
e50bb9a1 |
474 | |
722d2a37 |
475 | Simon Cozens promises to do this before he gets old. |
e50bb9a1 |
476 | |
722d2a37 |
477 | =head2 Update POSIX.pm for 1003.1-2 |
3958b146 |
478 | |
722d2a37 |
479 | =head2 Retargetable installation |
e50bb9a1 |
480 | |
722d2a37 |
481 | Allow C<@INC> to be changed after Perl is built. |
e50bb9a1 |
482 | |
722d2a37 |
483 | =head2 POSIX emulation on non-POSIX systems |
e50bb9a1 |
484 | |
722d2a37 |
485 | Make C<POSIX.pm> behave as POSIXly as possible everywhere, meaning we |
486 | have to implement POSIX equivalents for some functions if necessary. |
e50bb9a1 |
487 | |
722d2a37 |
488 | =head2 Rename Win32 headers |
e50bb9a1 |
489 | |
722d2a37 |
490 | =head2 Finish off lvalue functions |
491 | |
492 | They don't work in the debugger, and they don't work for list or hash |
493 | slices. |
e50bb9a1 |
494 | |
722d2a37 |
495 | =head2 Update sprintf documentation |
e50bb9a1 |
496 | |
722d2a37 |
497 | Hugo van der Sanden plans to look at this. |
e50bb9a1 |
498 | |
722d2a37 |
499 | =head2 Use fchown/fchmod internally |
e50bb9a1 |
500 | |
722d2a37 |
501 | This has been done in places, but needs a thorough code review. |
502 | Also fchdir is available in some platforms. |
e50bb9a1 |
503 | |
722d2a37 |
504 | =head1 Vague ideas |
e50bb9a1 |
505 | |
722d2a37 |
506 | Ideas which have been discussed, and which may or may not happen. |
e50bb9a1 |
507 | |
722d2a37 |
508 | =head2 ref() in list context |
e50bb9a1 |
509 | |
722d2a37 |
510 | It's unclear what this should do or how to do it without breaking old |
511 | code. |
e50bb9a1 |
512 | |
722d2a37 |
513 | =head2 Make tr/// return histogram |
e50bb9a1 |
514 | |
722d2a37 |
515 | There is a patch for this, but it may require Unicodification. |
e50bb9a1 |
516 | |
722d2a37 |
517 | =head2 Compile to real threaded code |
3958b146 |
518 | |
722d2a37 |
519 | =head2 Structured types |
3958b146 |
520 | |
722d2a37 |
521 | =head2 Modifiable $1 et al. |
e50bb9a1 |
522 | |
722d2a37 |
523 | ($x = "elephant") =~ /e(ph)/; |
524 | $1 = "g"; # $x = "elegant" |
e50bb9a1 |
525 | |
722d2a37 |
526 | What happens if there are multiple (nested?) brackets? What if the |
527 | string changes between the match and the assignment? |
e50bb9a1 |
528 | |
722d2a37 |
529 | =head2 Procedural interfaces for IO::*, etc. |
e50bb9a1 |
530 | |
722d2a37 |
531 | Some core modules have been accused of being overly-OO. Adding |
532 | procedural interfaces could demystify them. |
e50bb9a1 |
533 | |
722d2a37 |
534 | =head2 RPC modules |
e50bb9a1 |
535 | |
722d2a37 |
536 | =head2 Attach/detach debugger from running program |
e50bb9a1 |
537 | |
722d2a37 |
538 | With C<gdb>, you can attach the debugger to a running program if you |
539 | pass the process ID. It would be good to do this with the Perl debugger |
540 | on a running Perl program, although I'm not sure how it would be done. |
e50bb9a1 |
541 | |
722d2a37 |
542 | =head2 Alternative RE syntax module |
e50bb9a1 |
543 | |
722d2a37 |
544 | use Regex::Newbie; |
545 | $re = Regex::Newbie->new |
546 | ->start |
547 | ->match("foo") |
548 | ->repeat(Regex::Newbie->class("char"),3) |
549 | ->end; |
550 | /$re/; |
e50bb9a1 |
551 | |
722d2a37 |
552 | =head2 GUI::Native |
e50bb9a1 |
553 | |
722d2a37 |
554 | A non-core module that would use "native" GUI to create graphical |
555 | applications. |
e50bb9a1 |
556 | |
722d2a37 |
557 | =head2 foreach(reverse ...) |
e50bb9a1 |
558 | |
722d2a37 |
559 | Currently |
e50bb9a1 |
560 | |
722d2a37 |
561 | foreach (reverse @_) { ... } |
e50bb9a1 |
562 | |
722d2a37 |
563 | puts C<@_> on the stack, reverses it putting the reversed version on the |
564 | stack, then iterates forwards. Instead, it could be special-cased to put |
565 | C<@_> on the stack then iterate backwards. |
e50bb9a1 |
566 | |
722d2a37 |
567 | =head2 Constant function cache |
e50bb9a1 |
568 | |
722d2a37 |
569 | =head2 Approximate regular expression matching |
e50bb9a1 |
570 | |
722d2a37 |
571 | =head1 Ongoing |
e50bb9a1 |
572 | |
722d2a37 |
573 | These items B<always> need doing: |
e50bb9a1 |
574 | |
722d2a37 |
575 | =head2 Update guts documentation |
e50bb9a1 |
576 | |
722d2a37 |
577 | Simon Cozens tries to do this when possible, and contributions to the |
578 | C<perlapi> documentation is welcome. |
e50bb9a1 |
579 | |
722d2a37 |
580 | =head2 Add more tests |
e50bb9a1 |
581 | |
722d2a37 |
582 | Michael Schwern will donate $500 to Yet Another Society when all core |
583 | modules have tests. |
e50bb9a1 |
584 | |
722d2a37 |
585 | =head2 Update auxiliary tools |
e50bb9a1 |
586 | |
722d2a37 |
587 | The code we ship with Perl should look like good Perl 5. |
e50bb9a1 |
588 | |
722d2a37 |
589 | =head1 Recently done things |
e50bb9a1 |
590 | |
722d2a37 |
591 | These are things which have been on the todo lists in previous releases |
592 | but have recently been completed. |
e50bb9a1 |
593 | |
722d2a37 |
594 | =head2 Safe signal handling |
e50bb9a1 |
595 | |
722d2a37 |
596 | A new signal model went into 5.7.1 without much fanfare. Operations and |
597 | C<malloc>s are no longer interrupted by signals, which are handled |
598 | between opcodes. This means that C<PERL_ASYNC_CHECK> now actually does |
599 | something. However, there are still a few things that need to be done. |
e50bb9a1 |
600 | |
722d2a37 |
601 | =head2 Tie Modules |
e50bb9a1 |
602 | |
722d2a37 |
603 | Modules which implement arrays in terms of strings, substrings or files |
604 | can be found on the CPAN. |
e50bb9a1 |
605 | |
722d2a37 |
606 | =head2 gettimeofday |
e50bb9a1 |
607 | |
722d2a37 |
608 | C<Time::Hires> has been integrated into the core. |
e50bb9a1 |
609 | |
722d2a37 |
610 | =head2 setitimer and getimiter |
e50bb9a1 |
611 | |
722d2a37 |
612 | Adding C<Time::Hires> got us this too. |
e50bb9a1 |
613 | |
722d2a37 |
614 | =head2 Testing __DIE__ hook |
615 | |
616 | Tests have been added. |
617 | |
618 | =head2 CPP equivalent in Perl |
e50bb9a1 |
619 | |
722d2a37 |
620 | A C Yardley will probably have done this by the time you can read this. |
621 | This allows for a generalization of the C constant detection used in |
622 | building C<Errno.pm>. |
e50bb9a1 |
623 | |
722d2a37 |
624 | =head2 Explicit switch statements |
e50bb9a1 |
625 | |
722d2a37 |
626 | C<Switch.pm> has been integrated into the core to give you all manner of |
627 | C<switch...case> semantics. |
e50bb9a1 |
628 | |
722d2a37 |
629 | =head2 autocroak |
e50bb9a1 |
630 | |
722d2a37 |
631 | This is C<Fatal.pm>. |
e50bb9a1 |
632 | |
722d2a37 |
633 | =head2 UTF/EBCDIC |
e50bb9a1 |
634 | |
722d2a37 |
635 | Nick Ing-Simmons has made UTF-EBCDIC (UTR13) work with Perl. |
e50bb9a1 |
636 | |
722d2a37 |
637 | EBCDIC? http://www.unicode.org/unicode/reports/tr16/ |
e50bb9a1 |
638 | |
722d2a37 |
639 | =head2 UTF Regexes |
e50bb9a1 |
640 | |
722d2a37 |
641 | Although there are probably some small bugs to be rooted out, Jarkko |
642 | Hietaniemi has made regular expressions polymorphic between bytes and |
643 | characters. |
e50bb9a1 |
644 | |
722d2a37 |
645 | =head2 perlcc to produce executable |
e50bb9a1 |
646 | |
722d2a37 |
647 | C<perlcc> was recently rewritten, and can now produce standalone |
648 | executables. |
e50bb9a1 |
649 | |
722d2a37 |
650 | =head2 END blocks saved in compiled output |
e50bb9a1 |
651 | |
722d2a37 |
652 | =head2 Secure temporary file module |
e50bb9a1 |
653 | |
722d2a37 |
654 | Tim Jenness' C<File::Temp> is now in core. |
e50bb9a1 |
655 | |
722d2a37 |
656 | =head2 Integrate Time::HiRes |
e50bb9a1 |
657 | |
722d2a37 |
658 | This module is now part of core. |
e50bb9a1 |
659 | |
722d2a37 |
660 | =head2 Turn Cwd into XS |
e50bb9a1 |
661 | |
722d2a37 |
662 | Benjamin Sugars has done this. |
e50bb9a1 |
663 | |
722d2a37 |
664 | =head2 Mmap for input |
e50bb9a1 |
665 | |
722d2a37 |
666 | Nick Ing-Simmons' C<perlio> supports an C<mmap> IO method. |
e50bb9a1 |
667 | |
722d2a37 |
668 | =head2 Byte to/from UTF8 and UTF8 to/from local conversion |
e50bb9a1 |
669 | |
722d2a37 |
670 | C<Encode> provides this. |
e50bb9a1 |
671 | |
722d2a37 |
672 | =head2 Add sockatmark support |
e50bb9a1 |
673 | |
722d2a37 |
674 | Added in 5.7.1 |
e50bb9a1 |
675 | |
722d2a37 |
676 | =head2 Mailing list archives |
677 | |
678 | http://lists.perl.org/, http://archive.develooper.com/ |
679 | |
680 | =head2 Bug tracking |
681 | |
682 | Richard Foley has written the bug tracking system at http://bugs.perl.org/ |
e50bb9a1 |
683 | |
722d2a37 |
684 | =head2 Integrate MacPerl |
e50bb9a1 |
685 | |
722d2a37 |
686 | Chris Nandor and Matthias Neeracher have integrated the MacPerl changes |
687 | into 5.6.0. |
e50bb9a1 |
688 | |
722d2a37 |
689 | =head2 Web "nerve center" for Perl |
e50bb9a1 |
690 | |
722d2a37 |
691 | http://use.perl.org/ is what you're looking for. |
e50bb9a1 |
692 | |
722d2a37 |
693 | =head2 Regular expression tutorial |
e50bb9a1 |
694 | |
722d2a37 |
695 | C<perlretut>, provided by Mark Kvale. |
e50bb9a1 |
696 | |
722d2a37 |
697 | =head2 Debugging Tutorial |
e50bb9a1 |
698 | |
722d2a37 |
699 | C<perldebtut>, written by Richard Foley. |
e50bb9a1 |
700 | |
722d2a37 |
701 | =head2 Integrate new modules |
e50bb9a1 |
702 | |
722d2a37 |
703 | Jarkko has been integrating madly into 5.7.x |
e50bb9a1 |
704 | |
722d2a37 |
705 | =head2 Integrate profiler |
e50bb9a1 |
706 | |
722d2a37 |
707 | C<Devel::DProf> is now a core module. |
e50bb9a1 |
708 | |
722d2a37 |
709 | =head2 Y2K error detection |
e50bb9a1 |
710 | |
722d2a37 |
711 | There's a configure option to detect unsafe concatenation with "19", and |
712 | a CPAN module. (C<D'oh::Year>) |
e50bb9a1 |
713 | |
722d2a37 |
714 | =head2 Regular expression debugger |
e50bb9a1 |
715 | |
722d2a37 |
716 | While not part of core, Mark-Jason Dominus has written C<Rx> and has |
717 | also come up with a generalised strategy for regular expression |
718 | debugging. |
e50bb9a1 |
719 | |
722d2a37 |
720 | =head2 POD checker |
e50bb9a1 |
721 | |
722d2a37 |
722 | That's, uh, F<podchecker> |
e50bb9a1 |
723 | |
722d2a37 |
724 | =head2 "Dynamic" lexicals |
e50bb9a1 |
725 | |
722d2a37 |
726 | =head2 Cache precompiled modules |
e50bb9a1 |
727 | |
722d2a37 |
728 | =head1 Deprecated Wishes |
e50bb9a1 |
729 | |
722d2a37 |
730 | These are items which used to be in the todo file, but have been |
731 | deprecated for some reason. |
e50bb9a1 |
732 | |
722d2a37 |
733 | =head2 Loop control on do{} |
e50bb9a1 |
734 | |
722d2a37 |
735 | This would break old code; use C<do{{ }}> instead. |
e50bb9a1 |
736 | |
722d2a37 |
737 | =head2 Lexically scoped typeglobs |
e50bb9a1 |
738 | |
722d2a37 |
739 | Not needed now we have lexical IO handles. |
e50bb9a1 |
740 | |
722d2a37 |
741 | =head2 format BOTTOM |
3958b146 |
742 | |
722d2a37 |
743 | =head2 report HANDLE |
e50bb9a1 |
744 | |
722d2a37 |
745 | Damian Conway's text formatting modules seem to be the Way To Go. |
e50bb9a1 |
746 | |
722d2a37 |
747 | =head2 Generalised want()/caller()) |
3958b146 |
748 | |
722d2a37 |
749 | =head2 Named prototypes |
e50bb9a1 |
750 | |
722d2a37 |
751 | These both seem to be delayed until Perl 6. |
e50bb9a1 |
752 | |
722d2a37 |
753 | =head2 Built-in globbing |
e50bb9a1 |
754 | |
722d2a37 |
755 | The C<File::Glob> module has been used to replace the C<glob> function. |
e50bb9a1 |
756 | |
722d2a37 |
757 | =head2 Regression tests for suidperl |
e50bb9a1 |
758 | |
722d2a37 |
759 | C<suidperl> is deprecated in favour of common sense. |
e50bb9a1 |
760 | |
722d2a37 |
761 | =head2 Cached hash values |
e50bb9a1 |
762 | |
722d2a37 |
763 | We have shared hash keys, which perform the same job. |
e50bb9a1 |
764 | |
722d2a37 |
765 | =head2 Add compression modules |
e50bb9a1 |
766 | |
722d2a37 |
767 | The compression modules are a little heavy; meanwhile, Nick Clark is |
768 | working on experimental pragmata to do transparent decompression on |
769 | input. |
e50bb9a1 |
770 | |
722d2a37 |
771 | =head2 Reorganise documentation into tutorials/references |
e50bb9a1 |
772 | |
722d2a37 |
773 | Could not get consensus on P5P about this. |
e50bb9a1 |
774 | |
722d2a37 |
775 | =head2 Remove distinction between functions and operators |
776 | |
777 | Caution: highly flammable. |
778 | |
779 | =head2 Make XS easier to use |
e50bb9a1 |
780 | |
722d2a37 |
781 | Use C<Inline> instead, or SWIG. |
e50bb9a1 |
782 | |
722d2a37 |
783 | =head2 Make embedding easier to use |
e50bb9a1 |
784 | |
722d2a37 |
785 | Use C<Inline::CPR>. |
e50bb9a1 |
786 | |
722d2a37 |
787 | =head2 man for perl |
04c70446 |
788 | |
722d2a37 |
789 | See the Perl Power Tools. (http://language.perl.com/ppt/) |
04c70446 |
790 | |
722d2a37 |
791 | =head2 my $Package::variable |
04c70446 |
792 | |
722d2a37 |
793 | Use C<our> instead. |
04c70446 |
794 | |
722d2a37 |
795 | =head2 "or" tests defined, not truth |
04c70446 |
796 | |
722d2a37 |
797 | Suggesting this on P5P B<will> cause a boring and interminable flamewar. |
04c70446 |
798 | |
722d2a37 |
799 | =head2 "class"-based lexicals |
04c70446 |
800 | |
cbb3fa72 |
801 | Use flyweight objects, secure hashes or, dare I say it, pseudo-hashes instead. |
04c70446 |
802 | |
722d2a37 |
803 | =head2 byteperl |
04c70446 |
804 | |
722d2a37 |
805 | C<ByteLoader> covers this. |
04c70446 |
806 | |
722d2a37 |
807 | =head2 Lazy evaluation / tail recursion removal |
04c70446 |
808 | |
722d2a37 |
809 | C<List::Util> in core gives some of these; tail recursion removal is |
810 | done manually, with C<goto &whoami;>. (However, MJD has found that |
811 | C<goto &whoami> introduces a performance penalty, so maybe there should |
812 | be a way to do this after all: C<sub foo {START: ... goto START;> is |
813 | better.) |
0562c0e3 |
814 | |
815 | =head2 Make "use utf8" the default |
816 | |
817 | There is a patch available for this, search p5p archives for |
818 | the Subject "[EXPERIMENTAL PATCH] make unicode (utf8) default" |
819 | but this would be unacceptable because of backward compatibility: |
820 | scripts could not contain B<any legacy eight-bit data>. Also would |
821 | introduce a measurable slowdown of at least few percentages since all |
822 | regular expression operations would be done in full UTF-8. |
823 | |