Commit | Line | Data |
7711098a |
1 | =head1 NAME |
2 | |
3 | perltodo - Perl TO-DO List |
4 | |
5 | =head1 DESCRIPTION |
e50bb9a1 |
6 | |
0bdfc961 |
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. |
e50bb9a1 |
11 | |
0bdfc961 |
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: |
e50bb9a1 |
15 | |
0bdfc961 |
16 | http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/ |
938c8732 |
17 | |
617eabfa |
18 | What can we offer you in return? Fame, fortune, and everlasting glory? Maybe |
19 | not, but if your patch is incorporated, then we'll add your name to the |
20 | F<AUTHORS> file, which ships in the official distribution. How many other |
21 | programming languages offer you 1 line of immortality? |
938c8732 |
22 | |
0bdfc961 |
23 | =head1 Tasks that only need Perl knowledge |
e50bb9a1 |
24 | |
5a176cbc |
25 | =head2 Remove duplication of test setup. |
26 | |
27 | Schwern notes, that there's duplication of code - lots and lots of tests have |
28 | some variation on the big block of C<$Is_Foo> checks. We can safely put this |
29 | into a file, change it to build an C<%Is> hash and require it. Maybe just put |
30 | it into F<test.pl>. Throw in the handy tainting subroutines. |
31 | |
412f19a0 |
32 | =head2 merge common code in installperl and installman |
33 | |
34 | There are some common subroutines and a common C<BEGIN> block in F<installperl> |
35 | and F<installman>. These should probably be merged. It would also be good to |
36 | check for duplication in all the utility scripts supplied in the source |
37 | tarball. It might be good to move them all to a subdirectory, but this would |
38 | require careful checking to find all places that call them, and change those |
39 | correctly. |
40 | |
0bdfc961 |
41 | =head2 common test code for timed bail out |
e50bb9a1 |
42 | |
0bdfc961 |
43 | Write portable self destruct code for tests to stop them burning CPU in |
44 | infinite loops. This needs to avoid using alarm, as some of the tests are |
45 | testing alarm/sleep or timers. |
e50bb9a1 |
46 | |
87a942b1 |
47 | =head2 POD -E<gt> HTML conversion in the core still sucks |
e50bb9a1 |
48 | |
938c8732 |
49 | Which is crazy given just how simple POD purports to be, and how simple HTML |
adebf063 |
50 | can be. It's not actually I<as> simple as it sounds, particularly with the |
51 | flexibility POD allows for C<=item>, but it would be good to improve the |
52 | visual appeal of the HTML generated, and to avoid it having any validation |
53 | errors. See also L</make HTML install work>, as the layout of installation tree |
54 | is needed to improve the cross-linking. |
938c8732 |
55 | |
dc0fb092 |
56 | The addition of C<Pod::Simple> and its related modules may make this task |
57 | easier to complete. |
58 | |
8537f021 |
59 | =head2 merge checkpods and podchecker |
60 | |
61 | F<pod/checkpods.PL> (and C<make check> in the F<pod/> subdirectory) |
62 | implements a very basic check for pod files, but the errors it discovers |
63 | aren't found by podchecker. Add this check to podchecker, get rid of |
64 | checkpods and have C<make check> use podchecker. |
65 | |
b032e2ff |
66 | =head2 perlmodlib.PL rewrite |
67 | |
68 | Currently perlmodlib.PL needs to be run from a source directory where perl |
69 | has been built, or some modules won't be found, and others will be |
70 | skipped. Make it run from a clean perl source tree (so it's reproducible). |
71 | |
aa237293 |
72 | =head2 Parallel testing |
73 | |
b2e2905c |
74 | (This probably impacts much more than the core: also the Test::Harness |
02f21748 |
75 | and TAP::* modules on CPAN.) |
76 | |
aa237293 |
77 | The core regression test suite is getting ever more comprehensive, which has |
78 | the side effect that it takes longer to run. This isn't so good. Investigate |
79 | whether it would be feasible to give the harness script the B<option> of |
80 | running sets of tests in parallel. This would be useful for tests in |
81 | F<t/op/*.t> and F<t/uni/*.t> and maybe some sets of tests in F<lib/>. |
82 | |
83 | Questions to answer |
84 | |
85 | =over 4 |
86 | |
87 | =item 1 |
88 | |
89 | How does screen layout work when you're running more than one test? |
90 | |
91 | =item 2 |
92 | |
93 | How does the caller of test specify how many tests to run in parallel? |
94 | |
95 | =item 3 |
96 | |
97 | How do setup/teardown tests identify themselves? |
98 | |
99 | =back |
100 | |
101 | Pugs already does parallel testing - can their approach be re-used? |
102 | |
0bdfc961 |
103 | =head2 Make Schwern poorer |
e50bb9a1 |
104 | |
613bd4f7 |
105 | We should have tests for everything. When all the core's modules are tested, |
0bdfc961 |
106 | Schwern has promised to donate to $500 to TPF. We may need volunteers to |
107 | hold him upside down and shake vigorously in order to actually extract the |
108 | cash. |
3958b146 |
109 | |
0bdfc961 |
110 | =head2 Improve the coverage of the core tests |
e50bb9a1 |
111 | |
02f21748 |
112 | Use Devel::Cover to ascertain the core modules's test coverage, then add |
113 | tests that are currently missing. |
30222c0f |
114 | |
0bdfc961 |
115 | =head2 test B |
e50bb9a1 |
116 | |
0bdfc961 |
117 | A full test suite for the B module would be nice. |
e50bb9a1 |
118 | |
636e63cb |
119 | =head2 Deparse inlined constants |
120 | |
121 | Code such as this |
122 | |
123 | use constant PI => 4; |
124 | warn PI |
125 | |
126 | will currently deparse as |
127 | |
128 | use constant ('PI', 4); |
129 | warn 4; |
130 | |
131 | because the tokenizer inlines the value of the constant subroutine C<PI>. |
132 | This allows various compile time optimisations, such as constant folding |
133 | and dead code elimination. Where these haven't happened (such as the example |
134 | above) it ought be possible to make B::Deparse work out the name of the |
135 | original constant, because just enough information survives in the symbol |
136 | table to do this. Specifically, the same scalar is used for the constant in |
137 | the optree as is used for the constant subroutine, so by iterating over all |
138 | symbol tables and generating a mapping of SV address to constant name, it |
139 | would be possible to provide B::Deparse with this functionality. |
140 | |
0bdfc961 |
141 | =head2 A decent benchmark |
e50bb9a1 |
142 | |
617eabfa |
143 | C<perlbench> seems impervious to any recent changes made to the perl core. It |
0bdfc961 |
144 | would be useful to have a reasonable general benchmarking suite that roughly |
145 | represented what current perl programs do, and measurably reported whether |
146 | tweaks to the core improve, degrade or don't really affect performance, to |
147 | guide people attempting to optimise the guts of perl. Gisle would welcome |
148 | new tests for perlbench. |
6168cf99 |
149 | |
0bdfc961 |
150 | =head2 fix tainting bugs |
6168cf99 |
151 | |
0bdfc961 |
152 | Fix the bugs revealed by running the test suite with the C<-t> switch (via |
153 | C<make test.taintwarn>). |
e50bb9a1 |
154 | |
0bdfc961 |
155 | =head2 Dual life everything |
e50bb9a1 |
156 | |
0bdfc961 |
157 | As part of the "dists" plan, anything that doesn't belong in the smallest perl |
158 | distribution needs to be dual lifed. Anything else can be too. Figure out what |
159 | changes would be needed to package that module and its tests up for CPAN, and |
160 | do so. Test it with older perl releases, and fix the problems you find. |
e50bb9a1 |
161 | |
a393eb28 |
162 | To make a minimal perl distribution, it's useful to look at |
163 | F<t/lib/commonsense.t>. |
164 | |
0bdfc961 |
165 | =head2 Improving C<threads::shared> |
722d2a37 |
166 | |
0bdfc961 |
167 | Investigate whether C<threads::shared> could share aggregates properly with |
168 | only Perl level changes to shared.pm |
722d2a37 |
169 | |
0bdfc961 |
170 | =head2 POSIX memory footprint |
e50bb9a1 |
171 | |
0bdfc961 |
172 | Ilya observed that use POSIX; eats memory like there's no tomorrow, and at |
173 | various times worked to cut it down. There is probably still fat to cut out - |
174 | for example POSIX passes Exporter some very memory hungry data structures. |
e50bb9a1 |
175 | |
eed36644 |
176 | =head2 embed.pl/makedef.pl |
177 | |
178 | There is a script F<embed.pl> that generates several header files to prefix |
179 | all of Perl's symbols in a consistent way, to provide some semblance of |
180 | namespace support in C<C>. Functions are declared in F<embed.fnc>, variables |
907b3e23 |
181 | in F<interpvar.h>. Quite a few of the functions and variables |
eed36644 |
182 | are conditionally declared there, using C<#ifdef>. However, F<embed.pl> |
183 | doesn't understand the C macros, so the rules about which symbols are present |
184 | when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay. |
185 | It would be good to teach C<embed.pl> to understand the conditional |
186 | compilation, and hence remove the duplication, and the mistakes it has caused. |
e50bb9a1 |
187 | |
801de10e |
188 | =head2 use strict; and AutoLoad |
189 | |
190 | Currently if you write |
191 | |
192 | package Whack; |
193 | use AutoLoader 'AUTOLOAD'; |
194 | use strict; |
195 | 1; |
196 | __END__ |
197 | sub bloop { |
198 | print join (' ', No, strict, here), "!\n"; |
199 | } |
200 | |
201 | then C<use strict;> isn't in force within the autoloaded subroutines. It would |
202 | be more consistent (and less surprising) to arrange for all lexical pragmas |
203 | in force at the __END__ block to be in force within each autoloaded subroutine. |
204 | |
773b3597 |
205 | There's a similar problem with SelfLoader. |
206 | |
0bdfc961 |
207 | =head1 Tasks that need a little sysadmin-type knowledge |
e50bb9a1 |
208 | |
0bdfc961 |
209 | Or if you prefer, tasks that you would learn from, and broaden your skills |
210 | base... |
e50bb9a1 |
211 | |
cd793d32 |
212 | =head2 make HTML install work |
e50bb9a1 |
213 | |
adebf063 |
214 | There is an C<installhtml> target in the Makefile. It's marked as |
215 | "experimental". It would be good to get this tested, make it work reliably, and |
216 | remove the "experimental" tag. This would include |
217 | |
218 | =over 4 |
219 | |
220 | =item 1 |
221 | |
222 | Checking that cross linking between various parts of the documentation works. |
223 | In particular that links work between the modules (files with POD in F<lib/>) |
224 | and the core documentation (files in F<pod/>) |
225 | |
226 | =item 2 |
227 | |
617eabfa |
228 | Work out how to split C<perlfunc> into chunks, preferably one per function |
229 | group, preferably with general case code that could be used elsewhere. |
230 | Challenges here are correctly identifying the groups of functions that go |
231 | together, and making the right named external cross-links point to the right |
232 | page. Things to be aware of are C<-X>, groups such as C<getpwnam> to |
233 | C<endservent>, two or more C<=items> giving the different parameter lists, such |
234 | as |
adebf063 |
235 | |
236 | =item substr EXPR,OFFSET,LENGTH,REPLACEMENT |
adebf063 |
237 | =item substr EXPR,OFFSET,LENGTH |
adebf063 |
238 | =item substr EXPR,OFFSET |
239 | |
240 | and different parameter lists having different meanings. (eg C<select>) |
241 | |
242 | =back |
3a89a73c |
243 | |
0bdfc961 |
244 | =head2 compressed man pages |
245 | |
246 | Be able to install them. This would probably need a configure test to see how |
247 | the system does compressed man pages (same directory/different directory? |
248 | same filename/different filename), as well as tweaking the F<installman> script |
249 | to compress as necessary. |
250 | |
30222c0f |
251 | =head2 Add a code coverage target to the Makefile |
252 | |
253 | Make it easy for anyone to run Devel::Cover on the core's tests. The steps |
254 | to do this manually are roughly |
255 | |
256 | =over 4 |
257 | |
258 | =item * |
259 | |
260 | do a normal C<Configure>, but include Devel::Cover as a module to install |
261 | (see F<INSTALL> for how to do this) |
262 | |
263 | =item * |
264 | |
265 | make perl |
266 | |
267 | =item * |
268 | |
269 | cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness |
270 | |
271 | =item * |
272 | |
273 | Process the resulting Devel::Cover database |
274 | |
275 | =back |
276 | |
277 | This just give you the coverage of the F<.pm>s. To also get the C level |
278 | coverage you need to |
279 | |
280 | =over 4 |
281 | |
282 | =item * |
283 | |
284 | Additionally tell C<Configure> to use the appropriate C compiler flags for |
285 | C<gcov> |
286 | |
287 | =item * |
288 | |
289 | make perl.gcov |
290 | |
291 | (instead of C<make perl>) |
292 | |
293 | =item * |
294 | |
295 | After running the tests run C<gcov> to generate all the F<.gcov> files. |
296 | (Including down in the subdirectories of F<ext/> |
297 | |
298 | =item * |
299 | |
300 | (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files |
301 | to get their stats into the cover_db directory. |
302 | |
303 | =item * |
304 | |
305 | Then process the Devel::Cover database |
306 | |
307 | =back |
308 | |
309 | It would be good to add a single switch to C<Configure> to specify that you |
310 | wanted to perform perl level coverage, and another to specify C level |
311 | coverage, and have C<Configure> and the F<Makefile> do all the right things |
312 | automatically. |
313 | |
02f21748 |
314 | =head2 Make Config.pm cope with differences between built and installed perl |
0bdfc961 |
315 | |
316 | Quite often vendors ship a perl binary compiled with their (pay-for) |
317 | compilers. People install a free compiler, such as gcc. To work out how to |
318 | build extensions, Perl interrogates C<%Config>, so in this situation |
319 | C<%Config> describes compilers that aren't there, and extension building |
320 | fails. This forces people into choosing between re-compiling perl themselves |
321 | using the compiler they have, or only using modules that the vendor ships. |
322 | |
323 | It would be good to find a way teach C<Config.pm> about the installation setup, |
324 | possibly involving probing at install time or later, so that the C<%Config> in |
325 | a binary distribution better describes the installed machine, when the |
326 | installed machine differs from the build machine in some significant way. |
327 | |
728f4ecd |
328 | =head2 linker specification files |
329 | |
330 | Some platforms mandate that you provide a list of a shared library's external |
331 | symbols to the linker, so the core already has the infrastructure in place to |
332 | do this for generating shared perl libraries. My understanding is that the |
333 | GNU toolchain can accept an optional linker specification file, and restrict |
334 | visibility just to symbols declared in that file. It would be good to extend |
335 | F<makedef.pl> to support this format, and to provide a means within |
336 | C<Configure> to enable it. This would allow Unix users to test that the |
337 | export list is correct, and to build a perl that does not pollute the global |
338 | namespace with private symbols. |
339 | |
a229ae3b |
340 | =head2 Cross-compile support |
341 | |
342 | Currently C<Configure> understands C<-Dusecrosscompile> option. This option |
343 | arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is |
344 | assumed then to be copied to TARGET machine and used as a replacement of full |
345 | C<perl> executable. |
346 | |
d1307786 |
347 | This could be done little differently. Namely C<miniperl> should be built for |
a229ae3b |
348 | HOST and then full C<perl> with extensions should be compiled for TARGET. |
d1307786 |
349 | This, however, might require extra trickery for %Config: we have one config |
87a942b1 |
350 | first for HOST and then another for TARGET. Tools like MakeMaker will be |
351 | mightily confused. Having around two different types of executables and |
352 | libraries (HOST and TARGET) makes life interesting for Makefiles and |
353 | shell (and Perl) scripts. There is $Config{run}, normally empty, which |
354 | can be used as an execution wrapper. Also note that in some |
355 | cross-compilation/execution environments the HOST and the TARGET do |
356 | not see the same filesystem(s), the $Config{run} may need to do some |
357 | file/directory copying back and forth. |
0bdfc961 |
358 | |
8537f021 |
359 | =head2 roffitall |
360 | |
361 | Make F<pod/roffitall> be updated by F<pod/buildtoc>. |
362 | |
0bdfc961 |
363 | =head1 Tasks that need a little C knowledge |
364 | |
365 | These tasks would need a little C knowledge, but don't need any specific |
366 | background or experience with XS, or how the Perl interpreter works |
367 | |
3d826b29 |
368 | =head2 Weed out needless PERL_UNUSED_ARG |
369 | |
370 | The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about |
371 | unused arguments. Often the arguments can't be removed, as there is an |
372 | external constraint that determines the prototype of the function, so this |
373 | approach is valid. However, there are some cases where C<PERL_UNUSED_ARG> |
374 | could be removed. Specifically |
375 | |
376 | =over 4 |
377 | |
378 | =item * |
379 | |
380 | The prototypes of (nearly all) static functions can be changed |
381 | |
382 | =item * |
383 | |
384 | Unused arguments generated by short cut macros are wasteful - the short cut |
385 | macro used can be changed. |
386 | |
387 | =back |
388 | |
fbf638cb |
389 | =head2 Modernize the order of directories in @INC |
390 | |
391 | The way @INC is laid out by default, one cannot upgrade core (dual-life) |
392 | modules without overwriting files. This causes problems for binary |
3d14fd97 |
393 | package builders. One possible proposal is laid out in this |
394 | message: |
395 | L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>. |
fbf638cb |
396 | |
bcbaa2d5 |
397 | =head2 -Duse32bit* |
398 | |
399 | Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall. |
400 | On these systems, it might be the default compilation mode, and there |
401 | is currently no guarantee that passing no use64bitall option to the |
402 | Configure process will build a 32bit perl. Implementing -Duse32bit* |
403 | options would be nice for perl 5.12. |
404 | |
0bdfc961 |
405 | =head2 Make it clear from -v if this is the exact official release |
89007cb3 |
406 | |
617eabfa |
407 | Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that |
408 | usually defines one local patch, of the form "MAINT12345" or "RC1". The output |
409 | of perl -v doesn't report that a perl isn't an official release, and this |
89007cb3 |
410 | information can get lost in bugs reports. Because of this, the minor version |
fa11829f |
411 | isn't bumped up until RC time, to minimise the possibility of versions of perl |
89007cb3 |
412 | escaping that believe themselves to be newer than they actually are. |
413 | |
414 | It would be useful to find an elegant way to have the "this is an interim |
415 | maintenance release" or "this is a release candidate" in the terse -v output, |
416 | and have it so that it's easy for the pumpking to remove this just as the |
417 | release tarball is rolled up. This way the version pulled out of rsync would |
418 | always say "I'm a development release" and it would be safe to bump the |
419 | reported minor version as soon as a release ships, which would aid perl |
420 | developers. |
421 | |
0bdfc961 |
422 | This task is really about thinking of an elegant way to arrange the C source |
423 | such that it's trivial for the Pumpking to flag "this is an official release" |
424 | when making a tarball, yet leave the default source saying "I'm not the |
425 | official release". |
426 | |
fee0a0f7 |
427 | =head2 Profile Perl - am I hot or not? |
62403a3c |
428 | |
fee0a0f7 |
429 | The Perl source code is stable enough that it makes sense to profile it, |
430 | identify and optimise the hotspots. It would be good to measure the |
431 | performance of the Perl interpreter using free tools such as cachegrind, |
432 | gprof, and dtrace, and work to reduce the bottlenecks they reveal. |
433 | |
434 | As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops, |
435 | the ops that are most commonly used. The idea is that by grouping them, their |
436 | object code will be adjacent in the executable, so they have a greater chance |
437 | of already being in the CPU cache (or swapped in) due to being near another op |
438 | already in use. |
62403a3c |
439 | |
440 | Except that it's not clear if these really are the most commonly used ops. So |
fee0a0f7 |
441 | as part of exercising your skills with coverage and profiling tools you might |
442 | want to determine what ops I<really> are the most commonly used. And in turn |
443 | suggest evictions and promotions to achieve a better F<pp_hot.c>. |
62403a3c |
444 | |
98fed0ad |
445 | =head2 Allocate OPs from arenas |
446 | |
447 | Currently all new OP structures are individually malloc()ed and free()d. |
448 | All C<malloc> implementations have space overheads, and are now as fast as |
449 | custom allocates so it would both use less memory and less CPU to allocate |
450 | the various OP structures from arenas. The SV arena code can probably be |
451 | re-used for this. |
452 | |
539f2c54 |
453 | Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use |
454 | Perl_Slab_alloc() to pack optrees into a contiguous block, which is |
455 | probably superior to the use of OP arenas, esp. from a cache locality |
456 | standpoint. See L<Profile Perl - am I hot or not?>. |
457 | |
a229ae3b |
458 | =head2 Improve win32/wince.c |
0bdfc961 |
459 | |
a229ae3b |
460 | Currently, numerous functions look virtually, if not completely, |
02f21748 |
461 | identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't |
6d71adcd |
462 | be good. |
463 | |
c5b31784 |
464 | =head2 Use secure CRT functions when building with VC8 on Win32 |
465 | |
466 | Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis |
467 | that they were "unsafe" and introduced differently named secure versions of |
468 | them as replacements, e.g. instead of writing |
469 | |
470 | FILE* f = fopen(__FILE__, "r"); |
471 | |
472 | one should now write |
473 | |
474 | FILE* f; |
475 | errno_t err = fopen_s(&f, __FILE__, "r"); |
476 | |
477 | Currently, the warnings about these deprecations have been disabled by adding |
478 | -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that |
479 | warning suppressant and actually make use of the new secure CRT functions. |
480 | |
481 | There is also a similar issue with POSIX CRT function names like fileno having |
482 | been deprecated in favour of ISO C++ conformant names like _fileno. These |
26a6faa8 |
483 | warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It |
c5b31784 |
484 | might be nice to do as Microsoft suggest here too, although, unlike the secure |
485 | functions issue, there is presumably little or no benefit in this case. |
486 | |
16815324 |
487 | =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf() |
488 | |
489 | Maybe create a utility that checks after each libperl.a creation that |
490 | none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets()) |
491 | ever creep back to libperl.a. |
492 | |
493 | nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/' |
494 | |
495 | Note, of course, that this will only tell whether B<your> platform |
496 | is using those naughty interfaces. |
497 | |
de96509d |
498 | =head2 -D_FORTIFY_SOURCE=2, -fstack-protector |
499 | |
500 | Recent glibcs support C<-D_FORTIFY_SOURCE=2> and recent gcc |
501 | (4.1 onwards?) supports C<-fstack-protector>, both of which give |
502 | protection against various kinds of buffer overflow problems. |
503 | These should probably be used for compiling Perl whenever available, |
504 | Configure and/or hints files should be adjusted to probe for the |
505 | availability of these features and enable them as appropriate. |
16815324 |
506 | |
8964cfe0 |
507 | =head2 Arenas for GPs? For MAGIC? |
508 | |
509 | C<struct gp> and C<struct magic> are both currently allocated by C<malloc>. |
510 | It might be a speed or memory saving to change to using arenas. Or it might |
511 | not. It would need some suitable benchmarking first. In particular, C<GP>s |
512 | can probably be changed with minimal compatibility impact (probably nothing |
513 | outside of the core, or even outside of F<gv.c> allocates them), but they |
514 | probably aren't allocated/deallocated often enough for a speed saving. Whereas |
515 | C<MAGIC> is allocated/deallocated more often, but in turn, is also something |
516 | more externally visible, so changing the rules here may bite external code. |
517 | |
518 | |
6d71adcd |
519 | =head1 Tasks that need a knowledge of XS |
520 | |
521 | These tasks would need C knowledge, and roughly the level of knowledge of |
522 | the perl API that comes from writing modules that use XS to interface to |
523 | C. |
524 | |
6d71adcd |
525 | =head2 autovivification |
526 | |
527 | Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict; |
528 | |
529 | This task is incremental - even a little bit of work on it will help. |
530 | |
531 | =head2 Unicode in Filenames |
532 | |
533 | chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open, |
534 | opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen, |
535 | system, truncate, unlink, utime, -X. All these could potentially accept |
536 | Unicode filenames either as input or output (and in the case of system |
537 | and qx Unicode in general, as input or output to/from the shell). |
538 | Whether a filesystem - an operating system pair understands Unicode in |
539 | filenames varies. |
540 | |
541 | Known combinations that have some level of understanding include |
542 | Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac |
543 | OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to |
544 | create Unicode filenames, what forms of Unicode are accepted and used |
545 | (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used, |
546 | and so on, varies. Finding the right level of interfacing to Perl |
547 | requires some thought. Remember that an OS does not implicate a |
548 | filesystem. |
549 | |
550 | (The Windows -C command flag "wide API support" has been at least |
551 | temporarily retired in 5.8.1, and the -C has been repurposed, see |
552 | L<perlrun>.) |
553 | |
87a942b1 |
554 | Most probably the right way to do this would be this: |
555 | L</"Virtualize operating system access">. |
556 | |
6d71adcd |
557 | =head2 Unicode in %ENV |
558 | |
559 | Currently the %ENV entries are always byte strings. |
87a942b1 |
560 | See L</"Virtualize operating system access">. |
6d71adcd |
561 | |
1f2e7916 |
562 | =head2 Unicode and glob() |
563 | |
564 | Currently glob patterns and filenames returned from File::Glob::glob() |
87a942b1 |
565 | are always byte strings. See L</"Virtualize operating system access">. |
1f2e7916 |
566 | |
dbb0c492 |
567 | =head2 Unicode and lc/uc operators |
568 | |
569 | Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on |
570 | what the internal encoding of their argument is. That should not be the |
571 | case. Maybe add a pragma to switch behaviour. |
572 | |
6d71adcd |
573 | =head2 use less 'memory' |
574 | |
575 | Investigate trade offs to switch out perl's choices on memory usage. |
576 | Particularly perl should be able to give memory back. |
577 | |
578 | This task is incremental - even a little bit of work on it will help. |
579 | |
580 | =head2 Re-implement C<:unique> in a way that is actually thread-safe |
581 | |
582 | The old implementation made bad assumptions on several levels. A good 90% |
583 | solution might be just to make C<:unique> work to share the string buffer |
584 | of SvPVs. That way large constant strings can be shared between ithreads, |
585 | such as the configuration information in F<Config>. |
586 | |
587 | =head2 Make tainting consistent |
588 | |
589 | Tainting would be easier to use if it didn't take documented shortcuts and |
590 | allow taint to "leak" everywhere within an expression. |
591 | |
592 | =head2 readpipe(LIST) |
593 | |
594 | system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid |
595 | running a shell. readpipe() (the function behind qx//) could be similarly |
596 | extended. |
597 | |
6d71adcd |
598 | =head2 Audit the code for destruction ordering assumptions |
599 | |
600 | Change 25773 notes |
601 | |
602 | /* Need to check SvMAGICAL, as during global destruction it may be that |
603 | AvARYLEN(av) has been freed before av, and hence the SvANY() pointer |
604 | is now part of the linked list of SV heads, rather than pointing to |
605 | the original body. */ |
606 | /* FIXME - audit the code for other bugs like this one. */ |
607 | |
608 | adding the C<SvMAGICAL> check to |
609 | |
610 | if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) { |
611 | MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen); |
612 | |
613 | Go through the core and look for similar assumptions that SVs have particular |
614 | types, as all bets are off during global destruction. |
615 | |
749904bf |
616 | =head2 Extend PerlIO and PerlIO::Scalar |
617 | |
618 | PerlIO::Scalar doesn't know how to truncate(). Implementing this |
619 | would require extending the PerlIO vtable. |
620 | |
621 | Similarly the PerlIO vtable doesn't know about formats (write()), or |
622 | about stat(), or chmod()/chown(), utime(), or flock(). |
623 | |
624 | (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership |
625 | would mean.) |
626 | |
627 | PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(), |
628 | opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(), |
629 | readlink(). |
630 | |
94da6c29 |
631 | See also L</"Virtualize operating system access">. |
632 | |
3236f110 |
633 | =head2 -C on the #! line |
634 | |
635 | It should be possible to make -C work correctly if found on the #! line, |
636 | given that all perl command line options are strict ASCII, and -C changes |
637 | only the interpretation of non-ASCII characters, and not for the script file |
638 | handle. To make it work needs some investigation of the ordering of function |
639 | calls during startup, and (by implication) a bit of tweaking of that order. |
640 | |
81622873 |
641 | =head2 Propagate const outwards from Perl_moreswitches() |
642 | |
643 | Change 32057 changed the parameter and return value of C<Perl_moreswitches()> |
644 | from <char *> to <const char *>. It should now be possible to propagate |
645 | const-correctness outwards to C<S_parse_body()>, C<Perl_moreswitches()> |
646 | and C<Perl_yylex()>. |
647 | |
16815324 |
648 | =head2 Duplicate logic in S_method_common() and Perl_gv_fetchmethod_autoload() |
649 | |
650 | A comment in C<S_method_common> notes |
651 | |
652 | /* This code tries to figure out just what went wrong with |
653 | gv_fetchmethod. It therefore needs to duplicate a lot of |
654 | the internals of that function. We can't move it inside |
655 | Perl_gv_fetchmethod_autoload(), however, since that would |
656 | cause UNIVERSAL->can("NoSuchPackage::foo") to croak, and we |
657 | don't want that. |
658 | */ |
659 | |
660 | If C<Perl_gv_fetchmethod_autoload> gets rewritten to take (more) flag bits, |
661 | then it ought to be possible to move the logic from C<S_method_common> to |
662 | the "right" place. When making this change it would probably be good to also |
663 | pass in at least the method name length, if not also pre-computed hash values |
664 | when known. (I'm contemplating a plan to pre-compute hash values for common |
665 | fixed strings such as C<ISA> and pass them in to functions.) |
666 | |
d6c1e11f |
667 | =head2 Organize error messages |
668 | |
669 | Perl's diagnostics (error messages, see L<perldiag>) could use |
a8d0aeb9 |
670 | reorganizing and formalizing so that each error message has its |
d6c1e11f |
671 | stable-for-all-eternity unique id, categorized by severity, type, and |
672 | subsystem. (The error messages would be listed in a datafile outside |
c4bd451b |
673 | of the Perl source code, and the source code would only refer to the |
674 | messages by the id.) This clean-up and regularizing should apply |
d6c1e11f |
675 | for all croak() messages. |
676 | |
677 | This would enable all sorts of things: easier translation/localization |
678 | of the messages (though please do keep in mind the caveats of |
679 | L<Locale::Maketext> about too straightforward approaches to |
680 | translation), filtering by severity, and instead of grepping for a |
681 | particular error message one could look for a stable error id. (Of |
682 | course, changing the error messages by default would break all the |
683 | existing software depending on some particular error message...) |
684 | |
685 | This kind of functionality is known as I<message catalogs>. Look for |
686 | inspiration for example in the catgets() system, possibly even use it |
687 | if available-- but B<only> if available, all platforms will B<not> |
de96509d |
688 | have catgets(). |
d6c1e11f |
689 | |
690 | For the really pure at heart, consider extending this item to cover |
691 | also the warning messages (see L<perllexwarn>, C<warnings.pl>). |
3236f110 |
692 | |
0bdfc961 |
693 | =head1 Tasks that need a knowledge of the interpreter |
3298bd4d |
694 | |
0bdfc961 |
695 | These tasks would need C knowledge, and knowledge of how the interpreter works, |
696 | or a willingness to learn. |
3298bd4d |
697 | |
718140ec |
698 | =head2 lexicals used only once |
699 | |
700 | This warns: |
701 | |
702 | $ perl -we '$pie = 42' |
703 | Name "main::pie" used only once: possible typo at -e line 1. |
704 | |
705 | This does not: |
706 | |
707 | $ perl -we 'my $pie = 42' |
708 | |
709 | Logically all lexicals used only once should warn, if the user asks for |
d6f4ea2e |
710 | warnings. An unworked RT ticket (#5087) has been open for almost seven |
711 | years for this discrepancy. |
718140ec |
712 | |
a3d15f9a |
713 | =head2 UTF-8 revamp |
714 | |
715 | The handling of Unicode is unclean in many places. For example, the regexp |
716 | engine matches in Unicode semantics whenever the string or the pattern is |
717 | flagged as UTF-8, but that should not be dependent on an internal storage |
718 | detail of the string. Likewise, case folding behaviour is dependent on the |
719 | UTF8 internal flag being on or off. |
720 | |
721 | =head2 Properly Unicode safe tokeniser and pads. |
722 | |
723 | The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack - |
724 | variable names are stored in stashes as raw bytes, without the utf-8 flag |
725 | set. The pad API only takes a C<char *> pointer, so that's all bytes too. The |
726 | tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from |
727 | source filters. All this could be fixed. |
728 | |
636e63cb |
729 | =head2 state variable initialization in list context |
730 | |
731 | Currently this is illegal: |
732 | |
733 | state ($a, $b) = foo(); |
734 | |
a2874905 |
735 | In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different |
a8d0aeb9 |
736 | semantics, which is tricky to implement in Perl 5 as currently they produce |
a2874905 |
737 | the same opcode trees. The Perl 6 design is firm, so it would be good to |
a8d0aeb9 |
738 | implement the necessary code in Perl 5. There are comments in |
a2874905 |
739 | C<Perl_newASSIGNOP()> that show the code paths taken by various assignment |
740 | constructions involving state variables. |
636e63cb |
741 | |
4fedb12c |
742 | =head2 Implement $value ~~ 0 .. $range |
743 | |
744 | It would be nice to extend the syntax of the C<~~> operator to also |
745 | understand numeric (and maybe alphanumeric) ranges. |
a393eb28 |
746 | |
747 | =head2 A does() built-in |
748 | |
749 | Like ref(), only useful. It would call the C<DOES> method on objects; it |
750 | would also tell whether something can be dereferenced as an |
751 | array/hash/etc., or used as a regexp, etc. |
752 | L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html> |
753 | |
754 | =head2 Tied filehandles and write() don't mix |
755 | |
756 | There is no method on tied filehandles to allow them to be called back by |
757 | formats. |
4fedb12c |
758 | |
d10fc472 |
759 | =head2 Attach/detach debugger from running program |
1626a787 |
760 | |
cd793d32 |
761 | The old perltodo notes "With C<gdb>, you can attach the debugger to a running |
762 | program if you pass the process ID. It would be good to do this with the Perl |
0bdfc961 |
763 | debugger on a running Perl program, although I'm not sure how it would be |
764 | done." ssh and screen do this with named pipes in /tmp. Maybe we can too. |
1626a787 |
765 | |
a8cb5b9e |
766 | =head2 Optimize away empty destructors |
767 | |
768 | Defining an empty DESTROY method might be useful (notably in |
769 | AUTOLOAD-enabled classes), but it's still a bit expensive to call. That |
770 | could probably be optimized. |
771 | |
0bdfc961 |
772 | =head2 LVALUE functions for lists |
773 | |
774 | The old perltodo notes that lvalue functions don't work for list or hash |
775 | slices. This would be good to fix. |
776 | |
777 | =head2 LVALUE functions in the debugger |
778 | |
779 | The old perltodo notes that lvalue functions don't work in the debugger. This |
780 | would be good to fix. |
781 | |
0bdfc961 |
782 | =head2 regexp optimiser optional |
783 | |
784 | The regexp optimiser is not optional. It should configurable to be, to allow |
785 | its performance to be measured, and its bugs to be easily demonstrated. |
786 | |
02f21748 |
787 | =head2 delete &function |
788 | |
789 | Allow to delete functions. One can already undef them, but they're still |
790 | in the stash. |
791 | |
ef36c6a7 |
792 | =head2 C</w> regex modifier |
793 | |
794 | That flag would enable to match whole words, and also to interpolate |
795 | arrays as alternations. With it, C</P/w> would be roughly equivalent to: |
796 | |
797 | do { local $"='|'; /\b(?:P)\b/ } |
798 | |
799 | See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html> |
800 | for the discussion. |
801 | |
0bdfc961 |
802 | =head2 optional optimizer |
803 | |
804 | Make the peephole optimizer optional. Currently it performs two tasks as |
805 | it walks the optree - genuine peephole optimisations, and necessary fixups of |
806 | ops. It would be good to find an efficient way to switch out the |
807 | optimisations whilst keeping the fixups. |
808 | |
809 | =head2 You WANT *how* many |
810 | |
811 | Currently contexts are void, scalar and list. split has a special mechanism in |
812 | place to pass in the number of return values wanted. It would be useful to |
813 | have a general mechanism for this, backwards compatible and little speed hit. |
814 | This would allow proposals such as short circuiting sort to be implemented |
815 | as a module on CPAN. |
816 | |
817 | =head2 lexical aliases |
818 | |
819 | Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>. |
820 | |
821 | =head2 entersub XS vs Perl |
822 | |
823 | At the moment pp_entersub is huge, and has code to deal with entering both |
824 | perl and XS subroutines. Subroutine implementations rarely change between |
825 | perl and XS at run time, so investigate using 2 ops to enter subs (one for |
826 | XS, one for perl) and swap between if a sub is redefined. |
2810d901 |
827 | |
de535794 |
828 | =head2 Self-ties |
2810d901 |
829 | |
de535794 |
830 | Self-ties are currently illegal because they caused too many segfaults. Maybe |
a8d0aeb9 |
831 | the causes of these could be tracked down and self-ties on all types |
de535794 |
832 | reinstated. |
0bdfc961 |
833 | |
834 | =head2 Optimize away @_ |
835 | |
836 | The old perltodo notes "Look at the "reification" code in C<av.c>". |
837 | |
f092b1f4 |
838 | =head2 The yada yada yada operators |
839 | |
840 | Perl 6's Synopsis 3 says: |
841 | |
842 | I<The ... operator is the "yada, yada, yada" list operator, which is used as |
843 | the body in function prototypes. It complains bitterly (by calling fail) |
844 | if it is ever executed. Variant ??? calls warn, and !!! calls die.> |
845 | |
846 | Those would be nice to add to Perl 5. That could be done without new ops. |
847 | |
87a942b1 |
848 | =head2 Virtualize operating system access |
849 | |
850 | Implement a set of "vtables" that virtualizes operating system access |
851 | (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very |
852 | least these interfaces should take SVs as "name" arguments instead of |
853 | bare char pointers; probably the most flexible and extensible way |
e1a3d5d1 |
854 | would be for the Perl-facing interfaces to accept HVs. The system |
855 | needs to be per-operating-system and per-file-system |
856 | hookable/filterable, preferably both from XS and Perl level |
87a942b1 |
857 | (L<perlport/"Files and Filesystems"> is good reading at this point, |
858 | in fact, all of L<perlport> is.) |
859 | |
e1a3d5d1 |
860 | This has actually already been implemented (but only for Win32), |
861 | take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32 |
862 | variants go through a set of "vtables" for operating system access, |
863 | non-Win32 systems currently go straight for the POSIX/UNIX-style |
864 | system/library call. Similar system as for Win32 should be |
865 | implemented for all platforms. The existing Win32 implementation |
866 | probably does not need to survive alongside this proposed new |
867 | implementation, the approaches could be merged. |
87a942b1 |
868 | |
869 | What would this give us? One often-asked-for feature this would |
94da6c29 |
870 | enable is using Unicode for filenames, and other "names" like %ENV, |
871 | usernames, hostnames, and so forth. |
872 | (See L<perlunicode/"When Unicode Does Not Happen">.) |
873 | |
874 | But this kind of virtualization would also allow for things like |
875 | virtual filesystems, virtual networks, and "sandboxes" (though as long |
876 | as dynamic loading of random object code is allowed, not very safe |
877 | sandboxes since external code of course know not of Perl's vtables). |
878 | An example of a smaller "sandbox" is that this feature can be used to |
879 | implement per-thread working directories: Win32 already does this. |
880 | |
881 | See also L</"Extend PerlIO and PerlIO::Scalar">. |
87a942b1 |
882 | |
ac6197af |
883 | =head2 Investigate PADTMP hash pessimisation |
884 | |
885 | The peephole optimier converts constants used for hash key lookups to shared |
886 | hash key scalars. Under ithreads, something is undoing this work. See |
887 | See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html |
888 | |
0bdfc961 |
889 | =head1 Big projects |
890 | |
891 | Tasks that will get your name mentioned in the description of the "Highlights |
87a942b1 |
892 | of 5.12" |
0bdfc961 |
893 | |
894 | =head2 make ithreads more robust |
895 | |
4e577f8b |
896 | Generally make ithreads more robust. See also L</iCOW> |
0bdfc961 |
897 | |
898 | This task is incremental - even a little bit of work on it will help, and |
899 | will be greatly appreciated. |
900 | |
6c047da7 |
901 | One bit would be to write the missing code in sv.c:Perl_dirp_dup. |
902 | |
59c7f7d5 |
903 | Fix Perl_sv_dup, et al so that threads can return objects. |
904 | |
0bdfc961 |
905 | =head2 iCOW |
906 | |
907 | Sarathy and Arthur have a proposal for an improved Copy On Write which |
908 | specifically will be able to COW new ithreads. If this can be implemented |
909 | it would be a good thing. |
910 | |
911 | =head2 (?{...}) closures in regexps |
912 | |
913 | Fix (or rewrite) the implementation of the C</(?{...})/> closures. |
914 | |
915 | =head2 A re-entrant regexp engine |
916 | |
917 | This will allow the use of a regex from inside (?{ }), (??{ }) and |
918 | (?(?{ })|) constructs. |
6bda09f9 |
919 | |
6bda09f9 |
920 | =head2 Add class set operations to regexp engine |
921 | |
922 | Apparently these are quite useful. Anyway, Jeffery Friedl wants them. |
923 | |
924 | demerphq has this on his todo list, but right at the bottom. |