Commit | Line | Data |
7711098a |
1 | =head1 NAME |
2 | |
3 | perltodo - Perl TO-DO List |
4 | |
5 | =head1 DESCRIPTION |
e50bb9a1 |
6 | |
049aabcb |
7 | This is a list of wishes for Perl. The most up to date version of this file |
8 | is at http://perl5.git.perl.org/perl.git/blob_plain/HEAD:/pod/perltodo.pod |
9 | |
10 | The tasks we think are smaller or easier are listed first. Anyone is welcome |
11 | to work on any of these, but it's a good idea to first contact |
12 | I<perl5-porters@perl.org> to avoid duplication of effort, and to learn from |
13 | any previous attempts. By all means contact a pumpking privately first if you |
14 | prefer. |
e50bb9a1 |
15 | |
0bdfc961 |
16 | Whilst patches to make the list shorter are most welcome, ideas to add to |
17 | the list are also encouraged. Check the perl5-porters archives for past |
18 | ideas, and any discussion about them. One set of archives may be found at: |
e50bb9a1 |
19 | |
0bdfc961 |
20 | http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/ |
938c8732 |
21 | |
617eabfa |
22 | What can we offer you in return? Fame, fortune, and everlasting glory? Maybe |
23 | not, but if your patch is incorporated, then we'll add your name to the |
24 | F<AUTHORS> file, which ships in the official distribution. How many other |
25 | programming languages offer you 1 line of immortality? |
938c8732 |
26 | |
0bdfc961 |
27 | =head1 Tasks that only need Perl knowledge |
e50bb9a1 |
28 | |
de2b17d8 |
29 | =head2 Migrate t/ from custom TAP generation |
30 | |
31 | Many tests below F<t/> still generate TAP by "hand", rather than using library |
32 | functions. As explained in L<perlhack/Writing a test>, tests in F<t/> are |
33 | written in a particular way to test that more complex constructions actually |
34 | work before using them routinely. Hence they don't use C<Test::More>, but |
35 | instead there is an intentionally simpler library, F<t/test.pl>. However, |
36 | quite a few tests in F<t/> have not been refactored to use it. Refactoring |
37 | any of these tests, one at a time, is a useful thing TODO. |
38 | |
08063926 |
39 | =head2 Test that regen.pl was run |
40 | |
41 | There are various generated files shipped with the perl distribution, for |
42 | things like header files generate from data. The generation scripts are |
43 | written in perl, and all can be run by F<regen.pl>. However, because they're |
44 | written in perl, we can't run them before we've built perl. We can't run them |
45 | as part of the F<Makefile>, because changing files underneath F<make> confuses |
46 | it completely, and we don't want to run them automatically anyway, as they |
47 | change files shipped by the distribution, something we seek not do to. |
48 | |
49 | If someone changes the data, but forgets to re-run F<regen.pl> then the |
50 | generated files are out of sync. It would be good to have a test in |
51 | F<t/porting> that checks that the generated files are in sync, and fails |
52 | otherwise, to alert someone before they make a poor commit. I suspect that this |
53 | would require adapting the scripts run from F<regen.pl> to have dry-run |
54 | options, and invoking them with these, or by refactoring them into a library |
55 | that does the generation, which can be called by the scripts, and by the test. |
56 | |
0be987a2 |
57 | =head2 Automate perldelta generation |
58 | |
59 | The perldelta file accompanying each release summaries the major changes. |
60 | It's mostly manually generated currently, but some of that could be |
61 | automated with a bit of perl, specifically the generation of |
62 | |
63 | =over |
64 | |
65 | =item Modules and Pragmata |
66 | |
67 | =item New Documentation |
68 | |
69 | =item New Tests |
70 | |
71 | =back |
72 | |
73 | See F<Porting/how_to_write_a_perldelta.pod> for details. |
74 | |
5a176cbc |
75 | =head2 Remove duplication of test setup. |
76 | |
77 | Schwern notes, that there's duplication of code - lots and lots of tests have |
78 | some variation on the big block of C<$Is_Foo> checks. We can safely put this |
79 | into a file, change it to build an C<%Is> hash and require it. Maybe just put |
80 | it into F<test.pl>. Throw in the handy tainting subroutines. |
81 | |
87a942b1 |
82 | =head2 POD -E<gt> HTML conversion in the core still sucks |
e50bb9a1 |
83 | |
938c8732 |
84 | Which is crazy given just how simple POD purports to be, and how simple HTML |
adebf063 |
85 | can be. It's not actually I<as> simple as it sounds, particularly with the |
86 | flexibility POD allows for C<=item>, but it would be good to improve the |
87 | visual appeal of the HTML generated, and to avoid it having any validation |
88 | errors. See also L</make HTML install work>, as the layout of installation tree |
89 | is needed to improve the cross-linking. |
938c8732 |
90 | |
dc0fb092 |
91 | The addition of C<Pod::Simple> and its related modules may make this task |
92 | easier to complete. |
93 | |
0befdfba |
94 | =head2 Make ExtUtils::ParseXS use strict; |
95 | |
96 | F<lib/ExtUtils/ParseXS.pm> contains this line |
97 | |
98 | # use strict; # One of these days... |
99 | |
100 | Simply uncomment it, and fix all the resulting issues :-) |
101 | |
102 | The more practical approach, to break the task down into manageable chunks, is |
103 | to work your way though the code from bottom to top, or if necessary adding |
104 | extra C<{ ... }> blocks, and turning on strict within them. |
105 | |
0bdfc961 |
106 | =head2 Make Schwern poorer |
e50bb9a1 |
107 | |
613bd4f7 |
108 | We should have tests for everything. When all the core's modules are tested, |
0bdfc961 |
109 | Schwern has promised to donate to $500 to TPF. We may need volunteers to |
110 | hold him upside down and shake vigorously in order to actually extract the |
111 | cash. |
3958b146 |
112 | |
0bdfc961 |
113 | =head2 Improve the coverage of the core tests |
e50bb9a1 |
114 | |
02f21748 |
115 | Use Devel::Cover to ascertain the core modules's test coverage, then add |
116 | tests that are currently missing. |
30222c0f |
117 | |
0bdfc961 |
118 | =head2 test B |
e50bb9a1 |
119 | |
0bdfc961 |
120 | A full test suite for the B module would be nice. |
e50bb9a1 |
121 | |
0bdfc961 |
122 | =head2 A decent benchmark |
e50bb9a1 |
123 | |
617eabfa |
124 | C<perlbench> seems impervious to any recent changes made to the perl core. It |
0bdfc961 |
125 | would be useful to have a reasonable general benchmarking suite that roughly |
126 | represented what current perl programs do, and measurably reported whether |
127 | tweaks to the core improve, degrade or don't really affect performance, to |
128 | guide people attempting to optimise the guts of perl. Gisle would welcome |
129 | new tests for perlbench. |
6168cf99 |
130 | |
0bdfc961 |
131 | =head2 fix tainting bugs |
6168cf99 |
132 | |
0bdfc961 |
133 | Fix the bugs revealed by running the test suite with the C<-t> switch (via |
134 | C<make test.taintwarn>). |
e50bb9a1 |
135 | |
0bdfc961 |
136 | =head2 Dual life everything |
e50bb9a1 |
137 | |
0bdfc961 |
138 | As part of the "dists" plan, anything that doesn't belong in the smallest perl |
139 | distribution needs to be dual lifed. Anything else can be too. Figure out what |
140 | changes would be needed to package that module and its tests up for CPAN, and |
141 | do so. Test it with older perl releases, and fix the problems you find. |
e50bb9a1 |
142 | |
a393eb28 |
143 | To make a minimal perl distribution, it's useful to look at |
144 | F<t/lib/commonsense.t>. |
145 | |
dfb56e28 |
146 | =head2 Move dual-life pod/*.PL into ext |
c2aba5b8 |
147 | |
dfb56e28 |
148 | Nearly all the dual-life modules have been moved to F<ext>. However, we |
149 | still need to move F<pod/*.PL> into their respective directories |
764e6bc7 |
150 | in F<ext/>. They're referenced by (at least) C<plextract> in F<Makefile.SH> |
151 | and C<utils> in F<win32/Makefile> and F<win32/makefile.ml>, and listed |
152 | explicitly in F<win32/pod.mak>, F<vms/descrip_mms.template> and F<utils.lst> |
153 | |
0bdfc961 |
154 | =head2 POSIX memory footprint |
e50bb9a1 |
155 | |
0bdfc961 |
156 | Ilya observed that use POSIX; eats memory like there's no tomorrow, and at |
157 | various times worked to cut it down. There is probably still fat to cut out - |
158 | for example POSIX passes Exporter some very memory hungry data structures. |
e50bb9a1 |
159 | |
eed36644 |
160 | =head2 embed.pl/makedef.pl |
161 | |
162 | There is a script F<embed.pl> that generates several header files to prefix |
163 | all of Perl's symbols in a consistent way, to provide some semblance of |
164 | namespace support in C<C>. Functions are declared in F<embed.fnc>, variables |
907b3e23 |
165 | in F<interpvar.h>. Quite a few of the functions and variables |
eed36644 |
166 | are conditionally declared there, using C<#ifdef>. However, F<embed.pl> |
167 | doesn't understand the C macros, so the rules about which symbols are present |
168 | when is duplicated in F<makedef.pl>. Writing things twice is bad, m'kay. |
169 | It would be good to teach C<embed.pl> to understand the conditional |
170 | compilation, and hence remove the duplication, and the mistakes it has caused. |
e50bb9a1 |
171 | |
801de10e |
172 | =head2 use strict; and AutoLoad |
173 | |
174 | Currently if you write |
175 | |
176 | package Whack; |
177 | use AutoLoader 'AUTOLOAD'; |
178 | use strict; |
179 | 1; |
180 | __END__ |
181 | sub bloop { |
182 | print join (' ', No, strict, here), "!\n"; |
183 | } |
184 | |
185 | then C<use strict;> isn't in force within the autoloaded subroutines. It would |
186 | be more consistent (and less surprising) to arrange for all lexical pragmas |
187 | in force at the __END__ block to be in force within each autoloaded subroutine. |
188 | |
773b3597 |
189 | There's a similar problem with SelfLoader. |
190 | |
91d0cbf6 |
191 | =head2 profile installman |
192 | |
193 | The F<installman> script is slow. All it is doing text processing, which we're |
194 | told is something Perl is good at. So it would be nice to know what it is doing |
195 | that is taking so much CPU, and where possible address it. |
196 | |
a9ed9b74 |
197 | =head2 enable lexical enabling/disabling of inidvidual warnings |
198 | |
199 | Currently, warnings can only be enabled or disabled by category. There |
200 | are times when it would be useful to quash a single warning, not a |
201 | whole category. |
91d0cbf6 |
202 | |
0bdfc961 |
203 | =head1 Tasks that need a little sysadmin-type knowledge |
e50bb9a1 |
204 | |
0bdfc961 |
205 | Or if you prefer, tasks that you would learn from, and broaden your skills |
206 | base... |
e50bb9a1 |
207 | |
cd793d32 |
208 | =head2 make HTML install work |
e50bb9a1 |
209 | |
adebf063 |
210 | There is an C<installhtml> target in the Makefile. It's marked as |
211 | "experimental". It would be good to get this tested, make it work reliably, and |
212 | remove the "experimental" tag. This would include |
213 | |
214 | =over 4 |
215 | |
216 | =item 1 |
217 | |
218 | Checking that cross linking between various parts of the documentation works. |
219 | In particular that links work between the modules (files with POD in F<lib/>) |
220 | and the core documentation (files in F<pod/>) |
221 | |
222 | =item 2 |
223 | |
617eabfa |
224 | Work out how to split C<perlfunc> into chunks, preferably one per function |
225 | group, preferably with general case code that could be used elsewhere. |
226 | Challenges here are correctly identifying the groups of functions that go |
227 | together, and making the right named external cross-links point to the right |
228 | page. Things to be aware of are C<-X>, groups such as C<getpwnam> to |
229 | C<endservent>, two or more C<=items> giving the different parameter lists, such |
230 | as |
adebf063 |
231 | |
232 | =item substr EXPR,OFFSET,LENGTH,REPLACEMENT |
adebf063 |
233 | =item substr EXPR,OFFSET,LENGTH |
adebf063 |
234 | =item substr EXPR,OFFSET |
235 | |
236 | and different parameter lists having different meanings. (eg C<select>) |
237 | |
238 | =back |
3a89a73c |
239 | |
0bdfc961 |
240 | =head2 compressed man pages |
241 | |
242 | Be able to install them. This would probably need a configure test to see how |
243 | the system does compressed man pages (same directory/different directory? |
244 | same filename/different filename), as well as tweaking the F<installman> script |
245 | to compress as necessary. |
246 | |
30222c0f |
247 | =head2 Add a code coverage target to the Makefile |
248 | |
249 | Make it easy for anyone to run Devel::Cover on the core's tests. The steps |
250 | to do this manually are roughly |
251 | |
252 | =over 4 |
253 | |
254 | =item * |
255 | |
256 | do a normal C<Configure>, but include Devel::Cover as a module to install |
257 | (see F<INSTALL> for how to do this) |
258 | |
259 | =item * |
260 | |
261 | make perl |
262 | |
263 | =item * |
264 | |
265 | cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness |
266 | |
267 | =item * |
268 | |
269 | Process the resulting Devel::Cover database |
270 | |
271 | =back |
272 | |
273 | This just give you the coverage of the F<.pm>s. To also get the C level |
274 | coverage you need to |
275 | |
276 | =over 4 |
277 | |
278 | =item * |
279 | |
280 | Additionally tell C<Configure> to use the appropriate C compiler flags for |
281 | C<gcov> |
282 | |
283 | =item * |
284 | |
285 | make perl.gcov |
286 | |
287 | (instead of C<make perl>) |
288 | |
289 | =item * |
290 | |
291 | After running the tests run C<gcov> to generate all the F<.gcov> files. |
292 | (Including down in the subdirectories of F<ext/> |
293 | |
294 | =item * |
295 | |
296 | (From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files |
297 | to get their stats into the cover_db directory. |
298 | |
299 | =item * |
300 | |
301 | Then process the Devel::Cover database |
302 | |
303 | =back |
304 | |
305 | It would be good to add a single switch to C<Configure> to specify that you |
306 | wanted to perform perl level coverage, and another to specify C level |
307 | coverage, and have C<Configure> and the F<Makefile> do all the right things |
308 | automatically. |
309 | |
02f21748 |
310 | =head2 Make Config.pm cope with differences between built and installed perl |
0bdfc961 |
311 | |
312 | Quite often vendors ship a perl binary compiled with their (pay-for) |
313 | compilers. People install a free compiler, such as gcc. To work out how to |
314 | build extensions, Perl interrogates C<%Config>, so in this situation |
315 | C<%Config> describes compilers that aren't there, and extension building |
316 | fails. This forces people into choosing between re-compiling perl themselves |
317 | using the compiler they have, or only using modules that the vendor ships. |
318 | |
319 | It would be good to find a way teach C<Config.pm> about the installation setup, |
320 | possibly involving probing at install time or later, so that the C<%Config> in |
321 | a binary distribution better describes the installed machine, when the |
322 | installed machine differs from the build machine in some significant way. |
323 | |
728f4ecd |
324 | =head2 linker specification files |
325 | |
326 | Some platforms mandate that you provide a list of a shared library's external |
327 | symbols to the linker, so the core already has the infrastructure in place to |
328 | do this for generating shared perl libraries. My understanding is that the |
329 | GNU toolchain can accept an optional linker specification file, and restrict |
330 | visibility just to symbols declared in that file. It would be good to extend |
331 | F<makedef.pl> to support this format, and to provide a means within |
332 | C<Configure> to enable it. This would allow Unix users to test that the |
333 | export list is correct, and to build a perl that does not pollute the global |
334 | namespace with private symbols. |
335 | |
a229ae3b |
336 | =head2 Cross-compile support |
337 | |
338 | Currently C<Configure> understands C<-Dusecrosscompile> option. This option |
339 | arranges for building C<miniperl> for TARGET machine, so this C<miniperl> is |
340 | assumed then to be copied to TARGET machine and used as a replacement of full |
341 | C<perl> executable. |
342 | |
d1307786 |
343 | This could be done little differently. Namely C<miniperl> should be built for |
a229ae3b |
344 | HOST and then full C<perl> with extensions should be compiled for TARGET. |
d1307786 |
345 | This, however, might require extra trickery for %Config: we have one config |
87a942b1 |
346 | first for HOST and then another for TARGET. Tools like MakeMaker will be |
347 | mightily confused. Having around two different types of executables and |
348 | libraries (HOST and TARGET) makes life interesting for Makefiles and |
349 | shell (and Perl) scripts. There is $Config{run}, normally empty, which |
350 | can be used as an execution wrapper. Also note that in some |
351 | cross-compilation/execution environments the HOST and the TARGET do |
352 | not see the same filesystem(s), the $Config{run} may need to do some |
353 | file/directory copying back and forth. |
0bdfc961 |
354 | |
8537f021 |
355 | =head2 roffitall |
356 | |
357 | Make F<pod/roffitall> be updated by F<pod/buildtoc>. |
358 | |
98fca0e8 |
359 | =head2 Split "linker" from "compiler" |
360 | |
361 | Right now, Configure probes for two commands, and sets two variables: |
362 | |
363 | =over 4 |
364 | |
b91dd380 |
365 | =item * C<cc> (in F<cc.U>) |
98fca0e8 |
366 | |
367 | This variable holds the name of a command to execute a C compiler which |
368 | can resolve multiple global references that happen to have the same |
369 | name. Usual values are F<cc> and F<gcc>. |
370 | Fervent ANSI compilers may be called F<c89>. AIX has F<xlc>. |
371 | |
b91dd380 |
372 | =item * C<ld> (in F<dlsrc.U>) |
98fca0e8 |
373 | |
374 | This variable indicates the program to be used to link |
375 | libraries for dynamic loading. On some systems, it is F<ld>. |
376 | On ELF systems, it should be C<$cc>. Mostly, we'll try to respect |
377 | the hint file setting. |
378 | |
379 | =back |
380 | |
8d159ec1 |
381 | There is an implicit historical assumption from around Perl5.000alpha |
382 | something, that C<$cc> is also the correct command for linking object files |
383 | together to make an executable. This may be true on Unix, but it's not true |
384 | on other platforms, and there are a maze of work arounds in other places (such |
385 | as F<Makefile.SH>) to cope with this. |
98fca0e8 |
386 | |
387 | Ideally, we should create a new variable to hold the name of the executable |
388 | linker program, probe for it in F<Configure>, and centralise all the special |
389 | case logic there or in hints files. |
390 | |
391 | A small bikeshed issue remains - what to call it, given that C<$ld> is already |
8d159ec1 |
392 | taken (arguably for the wrong thing now, but on SunOS 4.1 it is the command |
393 | for creating dynamically-loadable modules) and C<$link> could be confused with |
394 | the Unix command line executable of the same name, which does something |
395 | completely different. Andy Dougherty makes the counter argument "In parrot, I |
396 | tried to call the command used to link object files and libraries into an |
397 | executable F<link>, since that's what my vaguely-remembered DOS and VMS |
398 | experience suggested. I don't think any real confusion has ensued, so it's |
399 | probably a reasonable name for perl5 to use." |
98fca0e8 |
400 | |
401 | "Alas, I've always worried that introducing it would make things worse, |
402 | since now the module building utilities would have to look for |
403 | C<$Config{link}> and institute a fall-back plan if it weren't found." |
8d159ec1 |
404 | Although I can see that as confusing, given that C<$Config{d_link}> is true |
405 | when (hard) links are available. |
98fca0e8 |
406 | |
75585ce3 |
407 | =head2 Configure Windows using PowerShell |
408 | |
409 | Currently, Windows uses hard-coded config files based to build the |
410 | config.h for compiling Perl. Makefiles are also hard-coded and need to be |
411 | hand edited prior to building Perl. While this makes it easy to create a perl.exe |
412 | that works across multiple Windows versions, being able to accurately |
413 | configure a perl.exe for a specific Windows versions and VS C++ would be |
414 | a nice enhancement. With PowerShell available on Windows XP and up, this |
415 | may now be possible. Step 1 might be to investigate whether this is possible |
416 | and use this to clean up our current makefile situation. Step 2 would be to |
417 | see if there would be a way to use our existing metaconfig units to configure a |
418 | Windows Perl or whether we go in a separate direction and make it so. Of |
419 | course, we all know what step 3 is. |
420 | |
ab45a0fa |
421 | =head2 decouple -g and -DDEBUGGING |
422 | |
423 | Currently F<Configure> automatically adds C<-DDEBUGGING> to the C compiler |
424 | flags if it spots C<-g> in the optimiser flags. The pre-processor directive |
eeab323f |
425 | C<DEBUGGING> enables F<perl>'s command line C<-D> options, but in the process |
ab45a0fa |
426 | makes F<perl> slower. It would be good to disentangle this logic, so that |
427 | C-level debugging with C<-g> and Perl level debugging with C<-D> can easily |
428 | be enabled independently. |
429 | |
0bdfc961 |
430 | =head1 Tasks that need a little C knowledge |
431 | |
432 | These tasks would need a little C knowledge, but don't need any specific |
433 | background or experience with XS, or how the Perl interpreter works |
434 | |
3d826b29 |
435 | =head2 Weed out needless PERL_UNUSED_ARG |
436 | |
437 | The C code uses the macro C<PERL_UNUSED_ARG> to stop compilers warning about |
438 | unused arguments. Often the arguments can't be removed, as there is an |
439 | external constraint that determines the prototype of the function, so this |
440 | approach is valid. However, there are some cases where C<PERL_UNUSED_ARG> |
441 | could be removed. Specifically |
442 | |
443 | =over 4 |
444 | |
445 | =item * |
446 | |
447 | The prototypes of (nearly all) static functions can be changed |
448 | |
449 | =item * |
450 | |
451 | Unused arguments generated by short cut macros are wasteful - the short cut |
452 | macro used can be changed. |
453 | |
454 | =back |
455 | |
fbf638cb |
456 | =head2 Modernize the order of directories in @INC |
457 | |
458 | The way @INC is laid out by default, one cannot upgrade core (dual-life) |
459 | modules without overwriting files. This causes problems for binary |
3d14fd97 |
460 | package builders. One possible proposal is laid out in this |
461 | message: |
462 | L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2002-04/msg02380.html>. |
fbf638cb |
463 | |
bcbaa2d5 |
464 | =head2 -Duse32bit* |
465 | |
466 | Natively 64-bit systems need neither -Duse64bitint nor -Duse64bitall. |
467 | On these systems, it might be the default compilation mode, and there |
468 | is currently no guarantee that passing no use64bitall option to the |
469 | Configure process will build a 32bit perl. Implementing -Duse32bit* |
470 | options would be nice for perl 5.12. |
471 | |
fee0a0f7 |
472 | =head2 Profile Perl - am I hot or not? |
62403a3c |
473 | |
fee0a0f7 |
474 | The Perl source code is stable enough that it makes sense to profile it, |
475 | identify and optimise the hotspots. It would be good to measure the |
476 | performance of the Perl interpreter using free tools such as cachegrind, |
477 | gprof, and dtrace, and work to reduce the bottlenecks they reveal. |
478 | |
479 | As part of this, the idea of F<pp_hot.c> is that it contains the I<hot> ops, |
480 | the ops that are most commonly used. The idea is that by grouping them, their |
481 | object code will be adjacent in the executable, so they have a greater chance |
482 | of already being in the CPU cache (or swapped in) due to being near another op |
483 | already in use. |
62403a3c |
484 | |
485 | Except that it's not clear if these really are the most commonly used ops. So |
fee0a0f7 |
486 | as part of exercising your skills with coverage and profiling tools you might |
487 | want to determine what ops I<really> are the most commonly used. And in turn |
488 | suggest evictions and promotions to achieve a better F<pp_hot.c>. |
62403a3c |
489 | |
91d0cbf6 |
490 | One piece of Perl code that might make a good testbed is F<installman>. |
491 | |
98fed0ad |
492 | =head2 Allocate OPs from arenas |
493 | |
494 | Currently all new OP structures are individually malloc()ed and free()d. |
495 | All C<malloc> implementations have space overheads, and are now as fast as |
496 | custom allocates so it would both use less memory and less CPU to allocate |
497 | the various OP structures from arenas. The SV arena code can probably be |
498 | re-used for this. |
499 | |
539f2c54 |
500 | Note that Configuring perl with C<-Accflags=-DPL_OP_SLAB_ALLOC> will use |
501 | Perl_Slab_alloc() to pack optrees into a contiguous block, which is |
502 | probably superior to the use of OP arenas, esp. from a cache locality |
503 | standpoint. See L<Profile Perl - am I hot or not?>. |
504 | |
a229ae3b |
505 | =head2 Improve win32/wince.c |
0bdfc961 |
506 | |
a229ae3b |
507 | Currently, numerous functions look virtually, if not completely, |
02f21748 |
508 | identical in both C<win32/wince.c> and C<win32/win32.c> files, which can't |
6d71adcd |
509 | be good. |
510 | |
c5b31784 |
511 | =head2 Use secure CRT functions when building with VC8 on Win32 |
512 | |
513 | Visual C++ 2005 (VC++ 8.x) deprecated a number of CRT functions on the basis |
514 | that they were "unsafe" and introduced differently named secure versions of |
515 | them as replacements, e.g. instead of writing |
516 | |
517 | FILE* f = fopen(__FILE__, "r"); |
518 | |
519 | one should now write |
520 | |
521 | FILE* f; |
522 | errno_t err = fopen_s(&f, __FILE__, "r"); |
523 | |
524 | Currently, the warnings about these deprecations have been disabled by adding |
525 | -D_CRT_SECURE_NO_DEPRECATE to the CFLAGS. It would be nice to remove that |
526 | warning suppressant and actually make use of the new secure CRT functions. |
527 | |
528 | There is also a similar issue with POSIX CRT function names like fileno having |
529 | been deprecated in favour of ISO C++ conformant names like _fileno. These |
26a6faa8 |
530 | warnings are also currently suppressed by adding -D_CRT_NONSTDC_NO_DEPRECATE. It |
c5b31784 |
531 | might be nice to do as Microsoft suggest here too, although, unlike the secure |
532 | functions issue, there is presumably little or no benefit in this case. |
533 | |
038ae9a4 |
534 | =head2 Fix POSIX::access() and chdir() on Win32 |
535 | |
536 | These functions currently take no account of DACLs and therefore do not behave |
537 | correctly in situations where access is restricted by DACLs (as opposed to the |
538 | read-only attribute). |
539 | |
540 | Furthermore, POSIX::access() behaves differently for directories having the |
541 | read-only attribute set depending on what CRT library is being used. For |
542 | example, the _access() function in the VC6 and VC7 CRTs (wrongly) claim that |
543 | such directories are not writable, whereas in fact all directories are writable |
544 | unless access is denied by DACLs. (In the case of directories, the read-only |
545 | attribute actually only means that the directory cannot be deleted.) This CRT |
546 | bug is fixed in the VC8 and VC9 CRTs (but, of course, the directory may still |
547 | not actually be writable if access is indeed denied by DACLs). |
548 | |
549 | For the chdir() issue, see ActiveState bug #74552: |
550 | http://bugs.activestate.com/show_bug.cgi?id=74552 |
551 | |
552 | Therefore, DACLs should be checked both for consistency across CRTs and for |
553 | the correct answer. |
554 | |
555 | (Note that perl's -w operator should not be modified to check DACLs. It has |
556 | been written so that it reflects the state of the read-only attribute, even |
557 | for directories (whatever CRT is being used), for symmetry with chmod().) |
558 | |
16815324 |
559 | =head2 strcat(), strcpy(), strncat(), strncpy(), sprintf(), vsprintf() |
560 | |
561 | Maybe create a utility that checks after each libperl.a creation that |
562 | none of the above (nor sprintf(), vsprintf(), or *SHUDDER* gets()) |
563 | ever creep back to libperl.a. |
564 | |
565 | nm libperl.a | ./miniperl -alne '$o = $F[0] if /:$/; print "$o $F[1]" if $F[0] eq "U" && $F[1] =~ /^(?:strn?c(?:at|py)|v?sprintf|gets)$/' |
566 | |
567 | Note, of course, that this will only tell whether B<your> platform |
568 | is using those naughty interfaces. |
569 | |
de96509d |
570 | =head2 -D_FORTIFY_SOURCE=2, -fstack-protector |
571 | |
572 | Recent glibcs support C<-D_FORTIFY_SOURCE=2> and recent gcc |
573 | (4.1 onwards?) supports C<-fstack-protector>, both of which give |
574 | protection against various kinds of buffer overflow problems. |
575 | These should probably be used for compiling Perl whenever available, |
576 | Configure and/or hints files should be adjusted to probe for the |
577 | availability of these features and enable them as appropriate. |
16815324 |
578 | |
8964cfe0 |
579 | =head2 Arenas for GPs? For MAGIC? |
580 | |
581 | C<struct gp> and C<struct magic> are both currently allocated by C<malloc>. |
582 | It might be a speed or memory saving to change to using arenas. Or it might |
583 | not. It would need some suitable benchmarking first. In particular, C<GP>s |
584 | can probably be changed with minimal compatibility impact (probably nothing |
585 | outside of the core, or even outside of F<gv.c> allocates them), but they |
586 | probably aren't allocated/deallocated often enough for a speed saving. Whereas |
587 | C<MAGIC> is allocated/deallocated more often, but in turn, is also something |
588 | more externally visible, so changing the rules here may bite external code. |
589 | |
3880c8ec |
590 | =head2 Shared arenas |
591 | |
592 | Several SV body structs are now the same size, notably PVMG and PVGV, PVAV and |
593 | PVHV, and PVCV and PVFM. It should be possible to allocate and return same |
594 | sized bodies from the same actual arena, rather than maintaining one arena for |
595 | each. This could save 4-6K per thread, of memory no longer tied up in the |
596 | not-yet-allocated part of an arena. |
597 | |
8964cfe0 |
598 | |
6d71adcd |
599 | =head1 Tasks that need a knowledge of XS |
600 | |
601 | These tasks would need C knowledge, and roughly the level of knowledge of |
602 | the perl API that comes from writing modules that use XS to interface to |
603 | C. |
604 | |
e851c105 |
605 | =head2 Write an XS cookbook |
606 | |
607 | Create pod/perlxscookbook.pod with short, task-focused 'recipes' in XS that |
608 | demonstrate common tasks and good practices. (Some of these might be |
609 | extracted from perlguts.) The target audience should be XS novices, who need |
610 | more examples than perlguts but something less overwhelming than perlapi. |
611 | Recipes should provide "one pretty good way to do it" instead of TIMTOWTDI. |
612 | |
5b7d14ff |
613 | Rather than focusing on interfacing Perl to C libraries, such a cookbook |
614 | should probably focus on how to optimize Perl routines by re-writing them |
615 | in XS. This will likely be more motivating to those who mostly work in |
616 | Perl but are looking to take the next step into XS. |
617 | |
618 | Deconstructing and explaining some simpler XS modules could be one way to |
619 | bootstrap a cookbook. (List::Util? Class::XSAccessor? Tree::Ternary_XS?) |
620 | Another option could be deconstructing the implementation of some simpler |
621 | functions in op.c. |
622 | |
318bf708 |
623 | =head2 Remove the use of SVs as temporaries in dump.c |
624 | |
625 | F<dump.c> contains debugging routines to dump out the contains of perl data |
626 | structures, such as C<SV>s, C<AV>s and C<HV>s. Currently, the dumping code |
627 | B<uses> C<SV>s for its temporary buffers, which was a logical initial |
628 | implementation choice, as they provide ready made memory handling. |
629 | |
630 | However, they also lead to a lot of confusion when it happens that what you're |
631 | trying to debug is seen by the code in F<dump.c>, correctly or incorrectly, as |
632 | a temporary scalar it can use for a temporary buffer. It's also not possible |
633 | to dump scalars before the interpreter is properly set up, such as during |
634 | ithreads cloning. It would be good to progressively replace the use of scalars |
635 | as string accumulation buffers with something much simpler, directly allocated |
636 | by C<malloc>. The F<dump.c> code is (or should be) only producing 7 bit |
637 | US-ASCII, so output character sets are not an issue. |
638 | |
639 | Producing and proving an internal simple buffer allocation would make it easier |
640 | to re-write the internals of the PerlIO subsystem to avoid using C<SV>s for |
641 | B<its> buffers, use of which can cause problems similar to those of F<dump.c>, |
642 | at similar times. |
643 | |
5d96f598 |
644 | =head2 safely supporting POSIX SA_SIGINFO |
645 | |
646 | Some years ago Jarkko supplied patches to provide support for the POSIX |
647 | SA_SIGINFO feature in Perl, passing the extra data to the Perl signal handler. |
648 | |
649 | Unfortunately, it only works with "unsafe" signals, because under safe |
650 | signals, by the time Perl gets to run the signal handler, the extra |
651 | information has been lost. Moreover, it's not easy to store it somewhere, |
652 | as you can't call mutexs, or do anything else fancy, from inside a signal |
653 | handler. |
654 | |
655 | So it strikes me that we could provide safe SA_SIGINFO support |
656 | |
657 | =over 4 |
658 | |
659 | =item 1 |
660 | |
661 | Provide global variables for two file descriptors |
662 | |
663 | =item 2 |
664 | |
665 | When the first request is made via C<sigaction> for C<SA_SIGINFO>, create a |
666 | pipe, store the reader in one, the writer in the other |
667 | |
668 | =item 3 |
669 | |
670 | In the "safe" signal handler (C<Perl_csighandler()>/C<S_raise_signal()>), if |
671 | the C<siginfo_t> pointer non-C<NULL>, and the writer file handle is open, |
672 | |
673 | =over 8 |
674 | |
675 | =item 1 |
676 | |
677 | serialise signal number, C<struct siginfo_t> (or at least the parts we care |
678 | about) into a small auto char buff |
679 | |
680 | =item 2 |
681 | |
682 | C<write()> that (non-blocking) to the writer fd |
683 | |
684 | =over 12 |
685 | |
686 | =item 1 |
687 | |
688 | if it writes 100%, flag the signal in a counter of "signals on the pipe" akin |
689 | to the current per-signal-number counts |
690 | |
691 | =item 2 |
692 | |
693 | if it writes 0%, assume the pipe is full. Flag the data as lost? |
694 | |
695 | =item 3 |
696 | |
697 | if it writes partially, croak a panic, as your OS is broken. |
698 | |
699 | =back |
700 | |
701 | =back |
702 | |
703 | =item 4 |
704 | |
705 | in the regular C<PERL_ASYNC_CHECK()> processing, if there are "signals on |
706 | the pipe", read the data out, deserialise, build the Perl structures on |
707 | the stack (code in C<Perl_sighandler()>, the "unsafe" handler), and call as |
708 | usual. |
709 | |
710 | =back |
711 | |
712 | I think that this gets us decent C<SA_SIGINFO> support, without the current risk |
713 | of running Perl code inside the signal handler context. (With all the dangers |
714 | of things like C<malloc> corruption that that currently offers us) |
715 | |
716 | For more information see the thread starting with this message: |
717 | http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-03/msg00305.html |
718 | |
6d71adcd |
719 | =head2 autovivification |
720 | |
721 | Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict; |
722 | |
723 | This task is incremental - even a little bit of work on it will help. |
724 | |
725 | =head2 Unicode in Filenames |
726 | |
727 | chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open, |
728 | opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen, |
729 | system, truncate, unlink, utime, -X. All these could potentially accept |
730 | Unicode filenames either as input or output (and in the case of system |
731 | and qx Unicode in general, as input or output to/from the shell). |
732 | Whether a filesystem - an operating system pair understands Unicode in |
733 | filenames varies. |
734 | |
735 | Known combinations that have some level of understanding include |
736 | Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac |
737 | OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to |
738 | create Unicode filenames, what forms of Unicode are accepted and used |
739 | (UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used, |
740 | and so on, varies. Finding the right level of interfacing to Perl |
741 | requires some thought. Remember that an OS does not implicate a |
742 | filesystem. |
743 | |
744 | (The Windows -C command flag "wide API support" has been at least |
745 | temporarily retired in 5.8.1, and the -C has been repurposed, see |
746 | L<perlrun>.) |
747 | |
87a942b1 |
748 | Most probably the right way to do this would be this: |
749 | L</"Virtualize operating system access">. |
750 | |
6d71adcd |
751 | =head2 Unicode in %ENV |
752 | |
753 | Currently the %ENV entries are always byte strings. |
87a942b1 |
754 | See L</"Virtualize operating system access">. |
6d71adcd |
755 | |
1f2e7916 |
756 | =head2 Unicode and glob() |
757 | |
758 | Currently glob patterns and filenames returned from File::Glob::glob() |
87a942b1 |
759 | are always byte strings. See L</"Virtualize operating system access">. |
1f2e7916 |
760 | |
dbb0c492 |
761 | =head2 Unicode and lc/uc operators |
762 | |
763 | Some built-in operators (C<lc>, C<uc>, etc.) behave differently, based on |
764 | what the internal encoding of their argument is. That should not be the |
765 | case. Maybe add a pragma to switch behaviour. |
766 | |
6d71adcd |
767 | =head2 use less 'memory' |
768 | |
769 | Investigate trade offs to switch out perl's choices on memory usage. |
770 | Particularly perl should be able to give memory back. |
771 | |
772 | This task is incremental - even a little bit of work on it will help. |
773 | |
774 | =head2 Re-implement C<:unique> in a way that is actually thread-safe |
775 | |
776 | The old implementation made bad assumptions on several levels. A good 90% |
777 | solution might be just to make C<:unique> work to share the string buffer |
778 | of SvPVs. That way large constant strings can be shared between ithreads, |
779 | such as the configuration information in F<Config>. |
780 | |
781 | =head2 Make tainting consistent |
782 | |
783 | Tainting would be easier to use if it didn't take documented shortcuts and |
784 | allow taint to "leak" everywhere within an expression. |
785 | |
786 | =head2 readpipe(LIST) |
787 | |
788 | system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid |
789 | running a shell. readpipe() (the function behind qx//) could be similarly |
790 | extended. |
791 | |
6d71adcd |
792 | =head2 Audit the code for destruction ordering assumptions |
793 | |
794 | Change 25773 notes |
795 | |
796 | /* Need to check SvMAGICAL, as during global destruction it may be that |
797 | AvARYLEN(av) has been freed before av, and hence the SvANY() pointer |
798 | is now part of the linked list of SV heads, rather than pointing to |
799 | the original body. */ |
800 | /* FIXME - audit the code for other bugs like this one. */ |
801 | |
802 | adding the C<SvMAGICAL> check to |
803 | |
804 | if (AvARYLEN(av) && SvMAGICAL(AvARYLEN(av))) { |
805 | MAGIC *mg = mg_find (AvARYLEN(av), PERL_MAGIC_arylen); |
806 | |
807 | Go through the core and look for similar assumptions that SVs have particular |
808 | types, as all bets are off during global destruction. |
809 | |
749904bf |
810 | =head2 Extend PerlIO and PerlIO::Scalar |
811 | |
812 | PerlIO::Scalar doesn't know how to truncate(). Implementing this |
813 | would require extending the PerlIO vtable. |
814 | |
815 | Similarly the PerlIO vtable doesn't know about formats (write()), or |
816 | about stat(), or chmod()/chown(), utime(), or flock(). |
817 | |
818 | (For PerlIO::Scalar it's hard to see what e.g. mode bits or ownership |
819 | would mean.) |
820 | |
821 | PerlIO doesn't do directories or symlinks, either: mkdir(), rmdir(), |
822 | opendir(), closedir(), seekdir(), rewinddir(), glob(); symlink(), |
823 | readlink(). |
824 | |
94da6c29 |
825 | See also L</"Virtualize operating system access">. |
826 | |
3236f110 |
827 | =head2 -C on the #! line |
828 | |
829 | It should be possible to make -C work correctly if found on the #! line, |
830 | given that all perl command line options are strict ASCII, and -C changes |
831 | only the interpretation of non-ASCII characters, and not for the script file |
832 | handle. To make it work needs some investigation of the ordering of function |
833 | calls during startup, and (by implication) a bit of tweaking of that order. |
834 | |
d6c1e11f |
835 | =head2 Organize error messages |
836 | |
837 | Perl's diagnostics (error messages, see L<perldiag>) could use |
a8d0aeb9 |
838 | reorganizing and formalizing so that each error message has its |
d6c1e11f |
839 | stable-for-all-eternity unique id, categorized by severity, type, and |
840 | subsystem. (The error messages would be listed in a datafile outside |
c4bd451b |
841 | of the Perl source code, and the source code would only refer to the |
842 | messages by the id.) This clean-up and regularizing should apply |
d6c1e11f |
843 | for all croak() messages. |
844 | |
845 | This would enable all sorts of things: easier translation/localization |
846 | of the messages (though please do keep in mind the caveats of |
847 | L<Locale::Maketext> about too straightforward approaches to |
848 | translation), filtering by severity, and instead of grepping for a |
849 | particular error message one could look for a stable error id. (Of |
850 | course, changing the error messages by default would break all the |
851 | existing software depending on some particular error message...) |
852 | |
853 | This kind of functionality is known as I<message catalogs>. Look for |
854 | inspiration for example in the catgets() system, possibly even use it |
855 | if available-- but B<only> if available, all platforms will B<not> |
de96509d |
856 | have catgets(). |
d6c1e11f |
857 | |
858 | For the really pure at heart, consider extending this item to cover |
859 | also the warning messages (see L<perllexwarn>, C<warnings.pl>). |
3236f110 |
860 | |
0bdfc961 |
861 | =head1 Tasks that need a knowledge of the interpreter |
3298bd4d |
862 | |
0bdfc961 |
863 | These tasks would need C knowledge, and knowledge of how the interpreter works, |
864 | or a willingness to learn. |
3298bd4d |
865 | |
de6375e3 |
866 | =head2 truncate() prototype |
867 | |
868 | The prototype of truncate() is currently C<$$>. It should probably |
869 | be C<*$> instead. (This is changed in F<opcode.pl>) |
870 | |
2d0587d8 |
871 | =head2 decapsulation of smart match argument |
872 | |
873 | Currently C<$foo ~~ $object> will die with the message "Smart matching a |
874 | non-overloaded object breaks encapsulation". It would be nice to allow |
875 | to bypass this by using explictly the syntax C<$foo ~~ %$object> or |
876 | C<$foo ~~ @$object>. |
877 | |
565590b5 |
878 | =head2 error reporting of [$a ; $b] |
879 | |
880 | Using C<;> inside brackets is a syntax error, and we don't propose to change |
881 | that by giving it any meaning. However, it's not reported very helpfully: |
882 | |
883 | $ perl -e '$a = [$b; $c];' |
884 | syntax error at -e line 1, near "$b;" |
885 | syntax error at -e line 1, near "$c]" |
886 | Execution of -e aborted due to compilation errors. |
887 | |
888 | It should be possible to hook into the tokeniser or the lexer, so that when a |
889 | C<;> is parsed where it is not legal as a statement terminator (ie inside |
890 | C<{}> used as a hashref, C<[]> or C<()>) it issues an error something like |
891 | I<';' isn't legal inside an expression - if you need multiple statements use a |
892 | do {...} block>. See the thread starting at |
893 | http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2008-09/msg00573.html |
894 | |
718140ec |
895 | =head2 lexicals used only once |
896 | |
897 | This warns: |
898 | |
899 | $ perl -we '$pie = 42' |
900 | Name "main::pie" used only once: possible typo at -e line 1. |
901 | |
902 | This does not: |
903 | |
904 | $ perl -we 'my $pie = 42' |
905 | |
906 | Logically all lexicals used only once should warn, if the user asks for |
d6f4ea2e |
907 | warnings. An unworked RT ticket (#5087) has been open for almost seven |
908 | years for this discrepancy. |
718140ec |
909 | |
a3d15f9a |
910 | =head2 UTF-8 revamp |
911 | |
912 | The handling of Unicode is unclean in many places. For example, the regexp |
913 | engine matches in Unicode semantics whenever the string or the pattern is |
914 | flagged as UTF-8, but that should not be dependent on an internal storage |
915 | detail of the string. Likewise, case folding behaviour is dependent on the |
916 | UTF8 internal flag being on or off. |
917 | |
918 | =head2 Properly Unicode safe tokeniser and pads. |
919 | |
920 | The tokeniser isn't actually very UTF-8 clean. C<use utf8;> is a hack - |
921 | variable names are stored in stashes as raw bytes, without the utf-8 flag |
922 | set. The pad API only takes a C<char *> pointer, so that's all bytes too. The |
923 | tokeniser ignores the UTF-8-ness of C<PL_rsfp>, or any SVs returned from |
924 | source filters. All this could be fixed. |
925 | |
636e63cb |
926 | =head2 state variable initialization in list context |
927 | |
928 | Currently this is illegal: |
929 | |
930 | state ($a, $b) = foo(); |
931 | |
a2874905 |
932 | In Perl 6, C<state ($a) = foo();> and C<(state $a) = foo();> have different |
a8d0aeb9 |
933 | semantics, which is tricky to implement in Perl 5 as currently they produce |
a2874905 |
934 | the same opcode trees. The Perl 6 design is firm, so it would be good to |
a8d0aeb9 |
935 | implement the necessary code in Perl 5. There are comments in |
a2874905 |
936 | C<Perl_newASSIGNOP()> that show the code paths taken by various assignment |
937 | constructions involving state variables. |
636e63cb |
938 | |
4fedb12c |
939 | =head2 Implement $value ~~ 0 .. $range |
940 | |
941 | It would be nice to extend the syntax of the C<~~> operator to also |
942 | understand numeric (and maybe alphanumeric) ranges. |
a393eb28 |
943 | |
944 | =head2 A does() built-in |
945 | |
946 | Like ref(), only useful. It would call the C<DOES> method on objects; it |
947 | would also tell whether something can be dereferenced as an |
948 | array/hash/etc., or used as a regexp, etc. |
949 | L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-03/msg00481.html> |
950 | |
951 | =head2 Tied filehandles and write() don't mix |
952 | |
953 | There is no method on tied filehandles to allow them to be called back by |
954 | formats. |
4fedb12c |
955 | |
53967bb9 |
956 | =head2 Propagate compilation hints to the debugger |
957 | |
958 | Currently a debugger started with -dE on the command-line doesn't see the |
959 | features enabled by -E. More generally hints (C<$^H> and C<%^H>) aren't |
960 | propagated to the debugger. Probably it would be a good thing to propagate |
961 | hints from the innermost non-C<DB::> scope: this would make code eval'ed |
962 | in the debugger see the features (and strictures, etc.) currently in |
963 | scope. |
964 | |
d10fc472 |
965 | =head2 Attach/detach debugger from running program |
1626a787 |
966 | |
cd793d32 |
967 | The old perltodo notes "With C<gdb>, you can attach the debugger to a running |
968 | program if you pass the process ID. It would be good to do this with the Perl |
0bdfc961 |
969 | debugger on a running Perl program, although I'm not sure how it would be |
970 | done." ssh and screen do this with named pipes in /tmp. Maybe we can too. |
1626a787 |
971 | |
0bdfc961 |
972 | =head2 LVALUE functions for lists |
973 | |
974 | The old perltodo notes that lvalue functions don't work for list or hash |
975 | slices. This would be good to fix. |
976 | |
0bdfc961 |
977 | =head2 regexp optimiser optional |
978 | |
979 | The regexp optimiser is not optional. It should configurable to be, to allow |
980 | its performance to be measured, and its bugs to be easily demonstrated. |
981 | |
02f21748 |
982 | =head2 delete &function |
983 | |
984 | Allow to delete functions. One can already undef them, but they're still |
985 | in the stash. |
986 | |
ef36c6a7 |
987 | =head2 C</w> regex modifier |
988 | |
989 | That flag would enable to match whole words, and also to interpolate |
990 | arrays as alternations. With it, C</P/w> would be roughly equivalent to: |
991 | |
992 | do { local $"='|'; /\b(?:P)\b/ } |
993 | |
994 | See L<http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-01/msg00400.html> |
995 | for the discussion. |
996 | |
0bdfc961 |
997 | =head2 optional optimizer |
998 | |
999 | Make the peephole optimizer optional. Currently it performs two tasks as |
1000 | it walks the optree - genuine peephole optimisations, and necessary fixups of |
1001 | ops. It would be good to find an efficient way to switch out the |
1002 | optimisations whilst keeping the fixups. |
1003 | |
1004 | =head2 You WANT *how* many |
1005 | |
1006 | Currently contexts are void, scalar and list. split has a special mechanism in |
1007 | place to pass in the number of return values wanted. It would be useful to |
1008 | have a general mechanism for this, backwards compatible and little speed hit. |
1009 | This would allow proposals such as short circuiting sort to be implemented |
1010 | as a module on CPAN. |
1011 | |
1012 | =head2 lexical aliases |
1013 | |
1014 | Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>. |
1015 | |
1016 | =head2 entersub XS vs Perl |
1017 | |
1018 | At the moment pp_entersub is huge, and has code to deal with entering both |
1019 | perl and XS subroutines. Subroutine implementations rarely change between |
1020 | perl and XS at run time, so investigate using 2 ops to enter subs (one for |
1021 | XS, one for perl) and swap between if a sub is redefined. |
2810d901 |
1022 | |
de535794 |
1023 | =head2 Self-ties |
2810d901 |
1024 | |
de535794 |
1025 | Self-ties are currently illegal because they caused too many segfaults. Maybe |
a8d0aeb9 |
1026 | the causes of these could be tracked down and self-ties on all types |
de535794 |
1027 | reinstated. |
0bdfc961 |
1028 | |
1029 | =head2 Optimize away @_ |
1030 | |
1031 | The old perltodo notes "Look at the "reification" code in C<av.c>". |
1032 | |
87a942b1 |
1033 | =head2 Virtualize operating system access |
1034 | |
1035 | Implement a set of "vtables" that virtualizes operating system access |
1036 | (open(), mkdir(), unlink(), readdir(), getenv(), etc.) At the very |
1037 | least these interfaces should take SVs as "name" arguments instead of |
1038 | bare char pointers; probably the most flexible and extensible way |
e1a3d5d1 |
1039 | would be for the Perl-facing interfaces to accept HVs. The system |
1040 | needs to be per-operating-system and per-file-system |
1041 | hookable/filterable, preferably both from XS and Perl level |
87a942b1 |
1042 | (L<perlport/"Files and Filesystems"> is good reading at this point, |
1043 | in fact, all of L<perlport> is.) |
1044 | |
e1a3d5d1 |
1045 | This has actually already been implemented (but only for Win32), |
1046 | take a look at F<iperlsys.h> and F<win32/perlhost.h>. While all Win32 |
1047 | variants go through a set of "vtables" for operating system access, |
1048 | non-Win32 systems currently go straight for the POSIX/UNIX-style |
1049 | system/library call. Similar system as for Win32 should be |
1050 | implemented for all platforms. The existing Win32 implementation |
1051 | probably does not need to survive alongside this proposed new |
1052 | implementation, the approaches could be merged. |
87a942b1 |
1053 | |
1054 | What would this give us? One often-asked-for feature this would |
94da6c29 |
1055 | enable is using Unicode for filenames, and other "names" like %ENV, |
1056 | usernames, hostnames, and so forth. |
1057 | (See L<perlunicode/"When Unicode Does Not Happen">.) |
1058 | |
1059 | But this kind of virtualization would also allow for things like |
1060 | virtual filesystems, virtual networks, and "sandboxes" (though as long |
1061 | as dynamic loading of random object code is allowed, not very safe |
1062 | sandboxes since external code of course know not of Perl's vtables). |
1063 | An example of a smaller "sandbox" is that this feature can be used to |
1064 | implement per-thread working directories: Win32 already does this. |
1065 | |
1066 | See also L</"Extend PerlIO and PerlIO::Scalar">. |
87a942b1 |
1067 | |
ac6197af |
1068 | =head2 Investigate PADTMP hash pessimisation |
1069 | |
9a2f2e6b |
1070 | The peephole optimiser converts constants used for hash key lookups to shared |
057163d7 |
1071 | hash key scalars. Under ithreads, something is undoing this work. |
ac6197af |
1072 | See http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2007-09/msg00793.html |
1073 | |
057163d7 |
1074 | =head2 Store the current pad in the OP slab allocator |
1075 | |
1076 | =for clarification |
1077 | I hope that I got that "current pad" part correct |
1078 | |
1079 | Currently we leak ops in various cases of parse failure. I suggested that we |
1080 | could solve this by always using the op slab allocator, and walking it to |
1081 | free ops. Dave comments that as some ops are already freed during optree |
1082 | creation one would have to mark which ops are freed, and not double free them |
1083 | when walking the slab. He notes that one problem with this is that for some ops |
1084 | you have to know which pad was current at the time of allocation, which does |
1085 | change. I suggested storing a pointer to the current pad in the memory allocated |
1086 | for the slab, and swapping to a new slab each time the pad changes. Dave thinks |
1087 | that this would work. |
1088 | |
52960e22 |
1089 | =head2 repack the optree |
1090 | |
1091 | Repacking the optree after execution order is determined could allow |
057163d7 |
1092 | removal of NULL ops, and optimal ordering of OPs with respect to cache-line |
1093 | filling. The slab allocator could be reused for this purpose. I think that |
1094 | the best way to do this is to make it an optional step just before the |
1095 | completed optree is attached to anything else, and to use the slab allocator |
1096 | unchanged, so that freeing ops is identical whether or not this step runs. |
1097 | Note that the slab allocator allocates ops downwards in memory, so one would |
1098 | have to actually "allocate" the ops in reverse-execution order to get them |
1099 | contiguous in memory in execution order. |
1100 | |
1101 | See http://www.nntp.perl.org/group/perl.perl5.porters/2007/12/msg131975.html |
1102 | |
1103 | Note that running this copy, and then freeing all the old location ops would |
1104 | cause their slabs to be freed, which would eliminate possible memory wastage if |
1105 | the previous suggestion is implemented, and we swap slabs more frequently. |
52960e22 |
1106 | |
12e06b6f |
1107 | =head2 eliminate incorrect line numbers in warnings |
1108 | |
1109 | This code |
1110 | |
1111 | use warnings; |
1112 | my $undef; |
1113 | |
1114 | if ($undef == 3) { |
1115 | } elsif ($undef == 0) { |
1116 | } |
1117 | |
18a16cc5 |
1118 | used to produce this output: |
12e06b6f |
1119 | |
1120 | Use of uninitialized value in numeric eq (==) at wrong.pl line 4. |
1121 | Use of uninitialized value in numeric eq (==) at wrong.pl line 4. |
1122 | |
18a16cc5 |
1123 | where the line of the second warning was misreported - it should be line 5. |
1124 | Rafael fixed this - the problem arose because there was no nextstate OP |
1125 | between the execution of the C<if> and the C<elsif>, hence C<PL_curcop> still |
1126 | reports that the currently executing line is line 4. The solution was to inject |
1127 | a nextstate OPs for each C<elsif>, although it turned out that the nextstate |
1128 | OP needed to be a nulled OP, rather than a live nextstate OP, else other line |
1129 | numbers became misreported. (Jenga!) |
12e06b6f |
1130 | |
1131 | The problem is more general than C<elsif> (although the C<elsif> case is the |
1132 | most common and the most confusing). Ideally this code |
1133 | |
1134 | use warnings; |
1135 | my $undef; |
1136 | |
1137 | my $a = $undef + 1; |
1138 | my $b |
1139 | = $undef |
1140 | + 1; |
1141 | |
1142 | would produce this output |
1143 | |
1144 | Use of uninitialized value $undef in addition (+) at wrong.pl line 4. |
1145 | Use of uninitialized value $undef in addition (+) at wrong.pl line 7. |
1146 | |
1147 | (rather than lines 4 and 5), but this would seem to require every OP to carry |
1148 | (at least) line number information. |
1149 | |
1150 | What might work is to have an optional line number in memory just before the |
1151 | BASEOP structure, with a flag bit in the op to say whether it's present. |
1152 | Initially during compile every OP would carry its line number. Then add a late |
1153 | pass to the optimiser (potentially combined with L</repack the optree>) which |
1154 | looks at the two ops on every edge of the graph of the execution path. If |
1155 | the line number changes, flags the destination OP with this information. |
1156 | Once all paths are traced, replace every op with the flag with a |
1157 | nextstate-light op (that just updates C<PL_curcop>), which in turn then passes |
1158 | control on to the true op. All ops would then be replaced by variants that |
1159 | do not store the line number. (Which, logically, why it would work best in |
1160 | conjunction with L</repack the optree>, as that is already copying/reallocating |
1161 | all the OPs) |
1162 | |
18a16cc5 |
1163 | (Although I should note that we're not certain that doing this for the general |
1164 | case is worth it) |
1165 | |
52960e22 |
1166 | =head2 optimize tail-calls |
1167 | |
1168 | Tail-calls present an opportunity for broadly applicable optimization; |
1169 | anywhere that C<< return foo(...) >> is called, the outer return can |
1170 | be replaced by a goto, and foo will return directly to the outer |
1171 | caller, saving (conservatively) 25% of perl's call&return cost, which |
1172 | is relatively higher than in C. The scheme language is known to do |
1173 | this heavily. B::Concise provides good insight into where this |
1174 | optimization is possible, ie anywhere entersub,leavesub op-sequence |
1175 | occurs. |
1176 | |
1177 | perl -MO=Concise,-exec,a,b,-main -e 'sub a{ 1 }; sub b {a()}; b(2)' |
1178 | |
1179 | Bottom line on this is probably a new pp_tailcall function which |
1180 | combines the code in pp_entersub, pp_leavesub. This should probably |
1181 | be done 1st in XS, and using B::Generate to patch the new OP into the |
1182 | optrees. |
1183 | |
0bdfc961 |
1184 | =head1 Big projects |
1185 | |
1186 | Tasks that will get your name mentioned in the description of the "Highlights |
87a942b1 |
1187 | of 5.12" |
0bdfc961 |
1188 | |
1189 | =head2 make ithreads more robust |
1190 | |
4e577f8b |
1191 | Generally make ithreads more robust. See also L</iCOW> |
0bdfc961 |
1192 | |
1193 | This task is incremental - even a little bit of work on it will help, and |
1194 | will be greatly appreciated. |
1195 | |
6c047da7 |
1196 | One bit would be to write the missing code in sv.c:Perl_dirp_dup. |
1197 | |
59c7f7d5 |
1198 | Fix Perl_sv_dup, et al so that threads can return objects. |
1199 | |
0bdfc961 |
1200 | =head2 iCOW |
1201 | |
1202 | Sarathy and Arthur have a proposal for an improved Copy On Write which |
1203 | specifically will be able to COW new ithreads. If this can be implemented |
1204 | it would be a good thing. |
1205 | |
1206 | =head2 (?{...}) closures in regexps |
1207 | |
1208 | Fix (or rewrite) the implementation of the C</(?{...})/> closures. |
1209 | |
1210 | =head2 A re-entrant regexp engine |
1211 | |
1212 | This will allow the use of a regex from inside (?{ }), (??{ }) and |
1213 | (?(?{ })|) constructs. |
6bda09f9 |
1214 | |
6bda09f9 |
1215 | =head2 Add class set operations to regexp engine |
1216 | |
1217 | Apparently these are quite useful. Anyway, Jeffery Friedl wants them. |
1218 | |
1219 | demerphq has this on his todo list, but right at the bottom. |
44a7a252 |
1220 | |
1221 | |
1222 | =head1 Tasks for microperl |
1223 | |
1224 | |
1225 | [ Each and every one of these may be obsolete, but they were listed |
1226 | in the old Todo.micro file] |
1227 | |
1228 | |
1229 | =head2 make creating uconfig.sh automatic |
1230 | |
1231 | =head2 make creating Makefile.micro automatic |
1232 | |
1233 | =head2 do away with fork/exec/wait? |
1234 | |
1235 | (system, popen should be enough?) |
1236 | |
1237 | =head2 some of the uconfig.sh really needs to be probed (using cc) in buildtime: |
1238 | |
1239 | (uConfigure? :-) native datatype widths and endianness come to mind |
1240 | |