This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Detail the minimal rewards programme we have on offer.
[perl5.git] / pod / perltodo.pod
CommitLineData
7711098a
GS
1=head1 NAME
2
3perltodo - Perl TO-DO List
4
5=head1 DESCRIPTION
e50bb9a1 6
0bdfc961
NC
7This is a list of wishes for Perl. The tasks we think are smaller or easier
8are listed first. Anyone is welcome to work on any of these, but it's a good
9idea to first contact I<perl5-porters@perl.org> to avoid duplication of
10effort. By all means contact a pumpking privately first if you prefer.
e50bb9a1 11
0bdfc961
NC
12Whilst patches to make the list shorter are most welcome, ideas to add to
13the list are also encouraged. Check the perl5-porters archives for past
14ideas, 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
NC
18What can we offer you in return? Fame, fortune, and everlasting glory? Maybe
19not, but if your patch is incorporated, then we'll add your name to the
20F<AUTHORS> file, which ships in the official distribution. How many other
21programming languages offer you 1 line of immortality?
938c8732 22
938c8732 23
e50bb9a1 24
0bdfc961 25=head1 Tasks that only need Perl knowledge
e50bb9a1 26
0bdfc961 27=head2 common test code for timed bail out
e50bb9a1 28
0bdfc961
NC
29Write portable self destruct code for tests to stop them burning CPU in
30infinite loops. This needs to avoid using alarm, as some of the tests are
31testing alarm/sleep or timers.
e50bb9a1 32
0bdfc961 33=head2 POD -> HTML conversion in the core still sucks
e50bb9a1 34
938c8732 35Which is crazy given just how simple POD purports to be, and how simple HTML
adebf063
NC
36can be. It's not actually I<as> simple as it sounds, particularly with the
37flexibility POD allows for C<=item>, but it would be good to improve the
38visual appeal of the HTML generated, and to avoid it having any validation
39errors. See also L</make HTML install work>, as the layout of installation tree
40is needed to improve the cross-linking.
938c8732 41
0bdfc961 42=head2 Make Schwern poorer
e50bb9a1 43
0bdfc961
NC
44We should have for everything. When all the core's modules are tested,
45Schwern has promised to donate to $500 to TPF. We may need volunteers to
46hold him upside down and shake vigorously in order to actually extract the
47cash.
3958b146 48
0bdfc961 49See F<t/lib/1_compile.t> for the 3 remaining modules that need tests.
e50bb9a1 50
0bdfc961 51=head2 Improve the coverage of the core tests
e50bb9a1 52
0bdfc961
NC
53Use Devel::Cover to ascertain the core's test coverage, then add tests that
54are currently missing.
30222c0f 55
0bdfc961 56=head2 test B
e50bb9a1 57
0bdfc961 58A full test suite for the B module would be nice.
e50bb9a1 59
0bdfc961 60=head2 A decent benchmark
e50bb9a1 61
617eabfa 62C<perlbench> seems impervious to any recent changes made to the perl core. It
0bdfc961
NC
63would be useful to have a reasonable general benchmarking suite that roughly
64represented what current perl programs do, and measurably reported whether
65tweaks to the core improve, degrade or don't really affect performance, to
66guide people attempting to optimise the guts of perl. Gisle would welcome
67new tests for perlbench.
6168cf99 68
0bdfc961 69=head2 fix tainting bugs
6168cf99 70
0bdfc961
NC
71Fix the bugs revealed by running the test suite with the C<-t> switch (via
72C<make test.taintwarn>).
e50bb9a1 73
0bdfc961 74=head2 Dual life everything
e50bb9a1 75
0bdfc961
NC
76As part of the "dists" plan, anything that doesn't belong in the smallest perl
77distribution needs to be dual lifed. Anything else can be too. Figure out what
78changes would be needed to package that module and its tests up for CPAN, and
79do so. Test it with older perl releases, and fix the problems you find.
e50bb9a1 80
0bdfc961 81=head2 Improving C<threads::shared>
722d2a37 82
0bdfc961
NC
83Investigate whether C<threads::shared> could share aggregates properly with
84only Perl level changes to shared.pm
722d2a37 85
0bdfc961 86=head2 POSIX memory footprint
e50bb9a1 87
0bdfc961
NC
88Ilya observed that use POSIX; eats memory like there's no tomorrow, and at
89various times worked to cut it down. There is probably still fat to cut out -
90for example POSIX passes Exporter some very memory hungry data structures.
e50bb9a1 91
4f05ac62
NC
92=head2 Refactor C<xsubpp> to be a thin wrapper around C<ExtUtils::ParseXS>
93
94C<ExtUtils::ParseXS> encapsulates a version of the C<xsubpp> into a module.
95In effect this is a code fork, and it's likely that C<xsubpp> has had some
96bug fixes since the code from C<ExtUtils::ParseXS> was derived. It would be
97good to merge the differences in, reduce down to 1 canonical implementation,
98and convert C<xsubpp> to a very thin command line wrapper to
99C<ExtUtils::ParseXS>.
100
101In theory this needs no real C knowledge, as one way of approaching this task
102is to ensure that C<ExtUtils::ParseXS> generates identical output to C<xsubpp>
103for input XS files, which does not require understanding the contents of the
104output C file. However, some C knowledge is likely to help with testing, and
105locating/producing comprehensive test cases.
e50bb9a1 106
e50bb9a1 107
e50bb9a1 108
e50bb9a1 109
adebf063 110
adebf063 111
0bdfc961 112=head1 Tasks that need a little sysadmin-type knowledge
e50bb9a1 113
0bdfc961
NC
114Or if you prefer, tasks that you would learn from, and broaden your skills
115base...
e50bb9a1 116
617eabfa
NC
117=head2 Relocatable perl
118
119The C level patches needed to create a relocatable perl binary are done, as
120is the work on F<Config.pm>. All that's left to do is the C<Configure> tweaking
121to let people specify how they want to do the install.
122
cd793d32 123=head2 make HTML install work
e50bb9a1 124
adebf063
NC
125There is an C<installhtml> target in the Makefile. It's marked as
126"experimental". It would be good to get this tested, make it work reliably, and
127remove the "experimental" tag. This would include
128
129=over 4
130
131=item 1
132
133Checking that cross linking between various parts of the documentation works.
134In particular that links work between the modules (files with POD in F<lib/>)
135and the core documentation (files in F<pod/>)
136
137=item 2
138
617eabfa
NC
139Work out how to split C<perlfunc> into chunks, preferably one per function
140group, preferably with general case code that could be used elsewhere.
141Challenges here are correctly identifying the groups of functions that go
142together, and making the right named external cross-links point to the right
143page. Things to be aware of are C<-X>, groups such as C<getpwnam> to
144C<endservent>, two or more C<=items> giving the different parameter lists, such
145as
adebf063
NC
146
147 =item substr EXPR,OFFSET,LENGTH,REPLACEMENT
148
149 =item substr EXPR,OFFSET,LENGTH
150
151 =item substr EXPR,OFFSET
152
153and different parameter lists having different meanings. (eg C<select>)
154
155=back
3a89a73c 156
0bdfc961
NC
157=head2 compressed man pages
158
159Be able to install them. This would probably need a configure test to see how
160the system does compressed man pages (same directory/different directory?
161same filename/different filename), as well as tweaking the F<installman> script
162to compress as necessary.
163
30222c0f
NC
164=head2 Add a code coverage target to the Makefile
165
166Make it easy for anyone to run Devel::Cover on the core's tests. The steps
167to do this manually are roughly
168
169=over 4
170
171=item *
172
173do a normal C<Configure>, but include Devel::Cover as a module to install
174(see F<INSTALL> for how to do this)
175
176=item *
177
178 make perl
179
180=item *
181
182 cd t; HARNESS_PERL_SWITCHES=-MDevel::Cover ./perl -I../lib harness
183
184=item *
185
186Process the resulting Devel::Cover database
187
188=back
189
190This just give you the coverage of the F<.pm>s. To also get the C level
191coverage you need to
192
193=over 4
194
195=item *
196
197Additionally tell C<Configure> to use the appropriate C compiler flags for
198C<gcov>
199
200=item *
201
202 make perl.gcov
203
204(instead of C<make perl>)
205
206=item *
207
208After running the tests run C<gcov> to generate all the F<.gcov> files.
209(Including down in the subdirectories of F<ext/>
210
211=item *
212
213(From the top level perl directory) run C<gcov2perl> on all the C<.gcov> files
214to get their stats into the cover_db directory.
215
216=item *
217
218Then process the Devel::Cover database
219
220=back
221
222It would be good to add a single switch to C<Configure> to specify that you
223wanted to perform perl level coverage, and another to specify C level
224coverage, and have C<Configure> and the F<Makefile> do all the right things
225automatically.
226
0bdfc961
NC
227=head2 Make Config.pm cope with differences between build and installed perl
228
229Quite often vendors ship a perl binary compiled with their (pay-for)
230compilers. People install a free compiler, such as gcc. To work out how to
231build extensions, Perl interrogates C<%Config>, so in this situation
232C<%Config> describes compilers that aren't there, and extension building
233fails. This forces people into choosing between re-compiling perl themselves
234using the compiler they have, or only using modules that the vendor ships.
235
236It would be good to find a way teach C<Config.pm> about the installation setup,
237possibly involving probing at install time or later, so that the C<%Config> in
238a binary distribution better describes the installed machine, when the
239installed machine differs from the build machine in some significant way.
240
46925299 241=head2 make parallel builds work
0bdfc961 242
46925299
NC
243Currently parallel builds (such as C<make -j3>) don't work reliably. We believe
244that this is due to incomplete dependency specification in the F<Makefile>.
245It would be good if someone were able to track down the causes of these
246problems, so that parallel builds worked properly.
0bdfc961 247
8523e164 248
0bdfc961
NC
249
250
251=head1 Tasks that need a little C knowledge
252
253These tasks would need a little C knowledge, but don't need any specific
254background or experience with XS, or how the Perl interpreter works
255
256=head2 Make it clear from -v if this is the exact official release
89007cb3 257
617eabfa
NC
258Currently perl from C<p4>/C<rsync> ships with a F<patchlevel.h> file that
259usually defines one local patch, of the form "MAINT12345" or "RC1". The output
260of perl -v doesn't report that a perl isn't an official release, and this
89007cb3 261information can get lost in bugs reports. Because of this, the minor version
fa11829f 262isn't bumped up until RC time, to minimise the possibility of versions of perl
89007cb3
NC
263escaping that believe themselves to be newer than they actually are.
264
265It would be useful to find an elegant way to have the "this is an interim
266maintenance release" or "this is a release candidate" in the terse -v output,
267and have it so that it's easy for the pumpking to remove this just as the
268release tarball is rolled up. This way the version pulled out of rsync would
269always say "I'm a development release" and it would be safe to bump the
270reported minor version as soon as a release ships, which would aid perl
271developers.
272
0bdfc961
NC
273This task is really about thinking of an elegant way to arrange the C source
274such that it's trivial for the Pumpking to flag "this is an official release"
275when making a tarball, yet leave the default source saying "I'm not the
276official release".
277
0f788cd2
NC
278=head2 Ordering of "global" variables.
279
280F<thrdvar.h> and F<intrpvarh> define the "global" variables that need to be
281per-thread under ithreads, where the variables are actually elements in a
282structure. As C dictates, the variables must be laid out in order of
283declaration. There is a comment
284C</* Important ones in the first cache line (if alignment is done right) */>
285which implies that at some point in the past the ordering was carefully chosen
286(at least in part). However, it's clear that the ordering is less than perfect,
287as currently there are things such as 7 C<bool>s in a row, then something
288typically requiring 4 byte alignment, and then an odd C<bool> later on.
289(C<bool>s are typically defined as C<char>s). So it would be good for someone
290to review the ordering of the variables, to see how much alignment padding can
291be removed.
292
0bdfc961
NC
293=head2 bincompat functions
294
295There are lots of functions which are retained for binary compatibility.
296Clean these up. Move them to mathom.c, and don't compile for blead?
297
62403a3c
NC
298=head2 am I hot or not?
299
300The idea of F<pp_hot.c> is that it contains the I<hot> ops, the ops that are
301most commonly used. The idea is that by grouping them, their object code will
302be adjacent in the executable, so they have a greater chance of already being
303in the CPU cache (or swapped in) due to being near another op already in use.
304
305Except that it's not clear if these really are the most commonly used ops. So
306anyone feeling like exercising their skill with coverage and profiling tools
307might want to determine what ops I<really> are the most commonly used. And in
308turn suggest evictions and promotions to achieve a better F<pp_hot.c>.
309
c99e3826
NC
310=head2 emulate the per-thread memory pool on Unix
311
312For Windows, ithreads allocates memory for each thread from a separate pool,
313which it discards at thread exit. It also checks that memory is free()d to
314the correct pool. Neither check is done on Unix, so code developed there won't
315be subject to such strictures, so can harbour bugs that only show up when the
316code reaches Windows.
317
318It would be good to be able to optionally emulate the Window pool system on
319Unix, to let developers who only have access to Unix, or want to use
320Unix-specific debugging tools, check for these problems. To do this would
321involve figuring out how the C<PerlMem_*> macros wrap C<malloc()> access, and
322providing a layer that records/checks the identity of the thread making the
323call, and recording all the memory allocated by each thread via this API so
324that it can be summarily free()d at thread exit. One implementation idea
325would be to increase the size of allocation, and store the C<my_perl> pointer
326(to identify the thread) at the start, along with pointers to make a linked
327list of blocks for this thread. To avoid alignment problems it would be
328necessary to do something like
329
330 union memory_header_padded {
331 struct memory_header {
332 void *thread_id; /* For my_perl */
333 void *next; /* Pointer to next block for this thread */
334 } data;
335 long double padding; /* whatever type has maximal alignment constraint */
336 };
337
338
339although C<long double> might not be the only type to add to the padding
340union.
62403a3c 341
077e3186
NC
342=head2 reduce duplication in sv_setsv_flags
343
344C<Perl_sv_setsv_flags> has a comment
345C</* There's a lot of redundancy below but we're going for speed here */>
346
347Whilst this was true 10 years ago, the growing disparity between RAM and CPU
348speeds mean that the trade offs have changed. In addition, the duplicate code
349adds to the maintenance burden. It would be good to see how much of the
350redundancy can be pruned, particular in the less common paths. (Profiling
351tools at the ready...). For example, why does the test for
352"Can't redefine active sort subroutine" need to occur in two places?
353
0bdfc961
NC
354
355
356
0bdfc961 357=head1 Tasks that need a knowledge of XS
e50bb9a1 358
0bdfc961
NC
359These tasks would need C knowledge, and roughly the level of knowledge of
360the perl API that comes from writing modules that use XS to interface to
361C.
362
363=head2 IPv6
364
365Clean this up. Check everything in core works
366
4a750395
NC
367=head2 shrink C<GV>s, C<CV>s
368
369By removing unused elements and careful re-ordering, the structures for C<AV>s
370and C<HV>s have recently been shrunk considerably. It's probable that the same
371approach would find savings in C<GV>s and C<CV>s, if not all the other
372larger-than-C<PVMG> types.
373
e593da1a
NC
374=head2 merge Perl_sv_2[inpu]v
375
376There's a lot of code shared between C<Perl_sv_2iv_flags>,
377C<Perl_sv_2uv_flags>, C<Perl_sv_2nv>, and C<Perl_sv_2pv_flags>. It would be
378interesting to see if some of it can be merged into common shared static
379functions. In particular, C<Perl_sv_2uv_flags> started out as a cut&paste
380from C<Perl_sv_2iv_flags> around 5.005_50 time, and it may be possible to
381replace both with a single function that returns a value or union which is
382split out by the macros in F<sv.h>
383
0bdfc961
NC
384=head2 UTF8 caching code
385
386The string position/offset cache is not optional. It should be.
387
388=head2 Implicit Latin 1 => Unicode translation
389
390Conversions from byte strings to UTF-8 currently map high bit characters
391to Unicode without translation (or, depending on how you look at it, by
392implicitly assuming that the byte strings are in Latin-1). As perl assumes
393the C locale by default, upgrading a string to UTF-8 may change the
394meaning of its contents regarding character classes, case mapping, etc.
395This should probably emit a warning (at least).
396
397This task is incremental - even a little bit of work on it will help.
e50bb9a1 398
cd793d32 399=head2 autovivification
e50bb9a1 400
cd793d32 401Make all autovivification consistent w.r.t LVALUE/RVALUE and strict/no strict;
e50bb9a1 402
0bdfc961 403This task is incremental - even a little bit of work on it will help.
e50bb9a1 404
0bdfc961 405=head2 Unicode in Filenames
e50bb9a1 406
0bdfc961
NC
407chdir, chmod, chown, chroot, exec, glob, link, lstat, mkdir, open,
408opendir, qx, readdir, readlink, rename, rmdir, stat, symlink, sysopen,
409system, truncate, unlink, utime, -X. All these could potentially accept
410Unicode filenames either as input or output (and in the case of system
411and qx Unicode in general, as input or output to/from the shell).
412Whether a filesystem - an operating system pair understands Unicode in
413filenames varies.
e50bb9a1 414
0bdfc961
NC
415Known combinations that have some level of understanding include
416Microsoft NTFS, Apple HFS+ (In Mac OS 9 and X) and Apple UFS (in Mac
417OS X), NFS v4 is rumored to be Unicode, and of course Plan 9. How to
418create Unicode filenames, what forms of Unicode are accepted and used
419(UCS-2, UTF-16, UTF-8), what (if any) is the normalization form used,
420and so on, varies. Finding the right level of interfacing to Perl
421requires some thought. Remember that an OS does not implicate a
422filesystem.
e50bb9a1 423
0bdfc961
NC
424(The Windows -C command flag "wide API support" has been at least
425temporarily retired in 5.8.1, and the -C has been repurposed, see
426L<perlrun>.)
969e704b 427
0bdfc961 428=head2 Unicode in %ENV
969e704b 429
0bdfc961 430Currently the %ENV entries are always byte strings.
e50bb9a1 431
0bdfc961 432=head2 use less 'memory'
e50bb9a1 433
0bdfc961
NC
434Investigate trade offs to switch out perl's choices on memory usage.
435Particularly perl should be able to give memory back.
e50bb9a1 436
0bdfc961 437This task is incremental - even a little bit of work on it will help.
0abe3f7c 438
0bdfc961 439=head2 Re-implement C<:unique> in a way that is actually thread-safe
0abe3f7c 440
0bdfc961
NC
441The old implementation made bad assumptions on several levels. A good 90%
442solution might be just to make C<:unique> work to share the string buffer
443of SvPVs. That way large constant strings can be shared between ithreads,
444such as the configuration information in F<Config>.
0abe3f7c 445
0bdfc961 446=head2 Make tainting consistent
0abe3f7c 447
0bdfc961
NC
448Tainting would be easier to use if it didn't take documented shortcuts and
449allow taint to "leak" everywhere within an expression.
0abe3f7c 450
0bdfc961 451=head2 readpipe(LIST)
0abe3f7c 452
0bdfc961
NC
453system() accepts a LIST syntax (and a PROGRAM LIST syntax) to avoid
454running a shell. readpipe() (the function behind qx//) could be similarly
455extended.
0abe3f7c 456
e50bb9a1 457
e50bb9a1 458
e50bb9a1 459
f86a8bc5 460
0bdfc961 461=head1 Tasks that need a knowledge of the interpreter
3298bd4d 462
0bdfc961
NC
463These tasks would need C knowledge, and knowledge of how the interpreter works,
464or a willingness to learn.
3298bd4d 465
0bdfc961
NC
466=head2 lexical pragmas
467
468Reimplement the mechanism of lexical pragmas to be more extensible. Fix
469current pragmas that don't work well (or at all) with lexical scopes or in
470run-time eval(STRING) (C<sort>, C<re>, C<encoding> for example). MJD has a
471preliminary patch that implements this.
0562c0e3 472
d10fc472 473=head2 Attach/detach debugger from running program
1626a787 474
cd793d32
NC
475The old perltodo notes "With C<gdb>, you can attach the debugger to a running
476program if you pass the process ID. It would be good to do this with the Perl
0bdfc961
NC
477debugger on a running Perl program, although I'm not sure how it would be
478done." ssh and screen do this with named pipes in /tmp. Maybe we can too.
1626a787 479
0bdfc961 480=head2 inlining autoloaded constants
d10fc472 481
0bdfc961
NC
482Currently the optimiser can inline constants when expressed as subroutines
483with prototype ($) that return a constant. Likewise, many packages wrapping
484C libraries export lots of constants as subroutines which are AUTOLOADed on
485demand. However, these have no prototypes, so can't be seen as constants by
486the optimiser. Some way of cheaply (low syntax, low memory overhead) to the
487perl compiler that a name is a constant would be great, so that it knows to
488call the AUTOLOAD routine at compile time, and then inline the constant.
80b46460 489
0bdfc961 490=head2 Constant folding
80b46460 491
0bdfc961
NC
492The peephole optimiser should trap errors during constant folding, and give
493up on the folding, rather than bailing out at compile time. It is quite
494possible that the unfoldable constant is in unreachable code, eg something
495akin to C<$a = 0/0 if 0;>
496
497=head2 LVALUE functions for lists
498
499The old perltodo notes that lvalue functions don't work for list or hash
500slices. This would be good to fix.
501
502=head2 LVALUE functions in the debugger
503
504The old perltodo notes that lvalue functions don't work in the debugger. This
505would be good to fix.
506
507=head2 _ prototype character
508
509Study the possibility of adding a new prototype character, C<_>, meaning
510"this argument defaults to $_".
511
512=head2 @INC source filter to Filter::Simple
513
514The second return value from a sub in @INC can be a source filter. This isn't
515documented. It should be changed to use Filter::Simple, tested and documented.
516
517=head2 regexp optimiser optional
518
519The regexp optimiser is not optional. It should configurable to be, to allow
520its performance to be measured, and its bugs to be easily demonstrated.
521
522=head2 UNITCHECK
523
524Introduce a new special block, UNITCHECK, which is run at the end of a
525compilation unit (module, file, eval(STRING) block). This will correspond to
526the Perl 6 CHECK. Perl 5's CHECK cannot be changed or removed because the
527O.pm/B.pm backend framework depends on it.
528
529=head2 optional optimizer
530
531Make the peephole optimizer optional. Currently it performs two tasks as
532it walks the optree - genuine peephole optimisations, and necessary fixups of
533ops. It would be good to find an efficient way to switch out the
534optimisations whilst keeping the fixups.
535
536=head2 You WANT *how* many
537
538Currently contexts are void, scalar and list. split has a special mechanism in
539place to pass in the number of return values wanted. It would be useful to
540have a general mechanism for this, backwards compatible and little speed hit.
541This would allow proposals such as short circuiting sort to be implemented
542as a module on CPAN.
543
544=head2 lexical aliases
545
546Allow lexical aliases (maybe via the syntax C<my \$alias = \$foo>.
547
548=head2 entersub XS vs Perl
549
550At the moment pp_entersub is huge, and has code to deal with entering both
551perl and XS subroutines. Subroutine implementations rarely change between
552perl and XS at run time, so investigate using 2 ops to enter subs (one for
553XS, one for perl) and swap between if a sub is redefined.
2810d901
NC
554
555=head2 Self ties
556
557self ties are currently illegal because they caused too many segfaults. Maybe
558the causes of these could be tracked down and self-ties on all types re-
559instated.
0bdfc961
NC
560
561=head2 Optimize away @_
562
563The old perltodo notes "Look at the "reification" code in C<av.c>".
564
565=head2 switch ops
566
567The old perltodo notes "Although we have C<Switch.pm> in core, Larry points to
568the dormant C<nswitch> and C<cswitch> ops in F<pp.c>; using these opcodes would
569be much faster."
570
571=head2 What hooks would assertions need?
572
573Assertions are in the core, and work. However, assertions needed to be added
574as a core patch, rather than an XS module in ext, or a CPAN module, because
575the core has no hooks in the necessary places. It would be useful to
576investigate what hooks would need to be added to make it possible to provide
577the full assertion support from a CPAN module, so that we aren't constraining
578the imagination of future CPAN authors.
579
580
581
582
583
584
585
586=head1 Big projects
587
588Tasks that will get your name mentioned in the description of the "Highlights
589of 5.10"
590
591=head2 make ithreads more robust
592
593Generally make ithreads more robust. See also L<iCOW>
594
595This task is incremental - even a little bit of work on it will help, and
596will be greatly appreciated.
597
598=head2 iCOW
599
600Sarathy and Arthur have a proposal for an improved Copy On Write which
601specifically will be able to COW new ithreads. If this can be implemented
602it would be a good thing.
603
604=head2 (?{...}) closures in regexps
605
606Fix (or rewrite) the implementation of the C</(?{...})/> closures.
607
608=head2 A re-entrant regexp engine
609
610This will allow the use of a regex from inside (?{ }), (??{ }) and
611(?(?{ })|) constructs.