This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perldelta: we suspect no known problems..?
[perl5.git] / pod / perldelta.pod
CommitLineData
44691e6f
AB
1=encoding utf8
2
3=head1 NAME
4
5ed58cbd 5perldelta - what is new for perl v5.18.0
e128ab2c 6
4eabcf70 7=head1 DESCRIPTION
6db9054f 8
5ed58cbd 9This document describes differences between the 5.16.0 release and the 5.18.0
e08634c5 10release.
6db9054f 11
d5f315e8 12If you are upgrading from an earlier release such as 5.14.0, first read
5ed58cbd 13L<perl5160delta>, which describes differences between 5.14.0 and 5.16.0.
3f01b192 14
5ed58cbd 15=head1 Core Enhancements
3f01b192 16
5ed58cbd 17=head2 New mechanism for experimental features
82d98f72 18
5ed58cbd 19Newly-added experimental features will now require this incantation:
82d98f72 20
5ed58cbd
RS
21 no warnings "experimental::feature_name";
22 use feature "feature_name"; # would warn without the prev line
23
24There is a new warnings category, called "experimental", containing
25warnings that the L<feature> pragma emits when enabling experimental
26features.
27
28Newly-added experimental features will also be given special warning IDs,
29which consist of "experimental::" followed by the name of the feature. (The
30plan is to extend this mechanism eventually to all warnings, to allow them
31to be enabled or disabled individually, and not just by category.)
32
33By saying
34
35 no warnings "experimental::feature_name";
36
37you are taking responsibility for any breakage that future changes to, or
38removal of, the feature may cause.
39
40Existing experimental features may begin emitting these warnings, too. Please
41consult L<perlexperiment> for information on which features are considered
42experimental.
43
44=head2 Hash overhaul
45
46Changes to the implementation of hashes in perl 5.18.0 will be one of the most
484d8ba1
DG
47visible changes to the behavior of existing code.
48
8e74b2ed 49By default, two distinct hash variables with identical keys and values may now
484d8ba1
DG
50provide their contents in a different order where it was previously identical.
51
52When encountering these changes, the key to cleaning up from them is to accept
53that B<hashes are unordered collections> and to act accordingly.
5ed58cbd
RS
54
55=head3 Hash randomization
56
57The seed used by Perl's hash function is now random. This means that the
58order which keys/values will be returned from functions like C<keys()>,
59C<values()>, and C<each()> will differ from run to run.
60
61This change was introduced to make Perl's hashes more robust to algorithmic
62complexity attacks, and also because we discovered that it exposes hash
63ordering dependency bugs and makes them easier to track down.
64
65Toolchain maintainers might want to invest in additional infrastructure to
66test for things like this. Running tests several times in a row and then
67comparing results will make it easier to spot hash order dependencies in
68code. Authors are strongly encouraged not to expose the key order of
69Perl's hashes to insecure audiences.
70
71Further, every hash has its own iteration order, which should make it much
72more difficult to determine what the current hash seed is.
73
0f2be12f
RS
74=head3 New hash functions
75
76Perl 5.18 includes support for multiple hash functions, and changed
77the default (to ONE_AT_A_TIME_HARD), you can choose a different
78algorithm by defining a symbol at compile time. For a current list,
79consult the F<INSTALL> document. Note that as of Perl 5.18 we can
80only recommend use of the default or SIPHASH. All the others are
81known to have security issues and are for research purposes only.
5ed58cbd 82
f105b7be 83=head3 PERL_HASH_SEED environment variable now takes a hex value
5ed58cbd 84
484d8ba1
DG
85C<PERL_HASH_SEED> no longer accepts an integer as a parameter;
86instead the value is expected to be a binary value encoded in a hex
87string, such as "0xf5867c55039dc724". This is to make the
88infrastructure support hash seeds of arbitrary lengths, which might
5ed58cbd
RS
89exceed that of an integer. (SipHash uses a 16 byte seed).
90
c40c48bb
RS
91=head3 PERL_PERTURB_KEYS environment variable added
92
f105b7be 93The C<PERL_PERTURB_KEYS> environment variable allows one to control the level of
c40c48bb
RS
94randomization applied to C<keys> and friends.
95
f105b7be 96When C<PERL_PERTURB_KEYS> is 0, perl will not randomize the key order at all. The
c40c48bb
RS
97chance that C<keys> changes due to an insert will be the same as in previous
98perls, basically only when the bucket size is changed.
99
f105b7be 100When C<PERL_PERTURB_KEYS> is 1, perl will randomize keys in a non-repeatable
c40c48bb
RS
101way. The chance that C<keys> changes due to an insert will be very high. This
102is the most secure and default mode.
103
f105b7be
KE
104When C<PERL_PERTURB_KEYS> is 2, perl will randomize keys in a repeatable way.
105Repeated runs of the same program should produce the same output every time.
c40c48bb 106
f105b7be
KE
107C<PERL_HASH_SEED> implies a non-default C<PERL_PERTURB_KEYS> setting. Setting
108C<PERL_HASH_SEED=0> (exactly one 0) implies C<PERL_PERTURB_KEYS=0> (hash key
109randomization disabled); settng C<PERL_HASH_SEED> to any other value implies
110C<PERL_PERTURB_KEYS=2> (deterministic and repeatable hash key randomization).
111Specifying C<PERL_PERTURB_KEYS> explicitly to a different level overrides this
c40c48bb
RS
112behavior.
113
5ed58cbd
RS
114=head3 Hash::Util::hash_seed() now returns a string
115
116Hash::Util::hash_seed() now returns a string instead of an integer. This
117is to make the infrastructure support hash seeds of arbitrary lengths
118which might exceed that of an integer. (SipHash uses a 16 byte seed).
119
120=head3 Output of PERL_HASH_SEED_DEBUG has been changed
121
122The environment variable PERL_HASH_SEED_DEBUG now makes perl show both the
f105b7be 123hash function perl was built with, I<and> the seed, in hex, in use for that
5ed58cbd
RS
124process. Code parsing this output, should it exist, must change to accommodate
125the new format. Example of the new format:
126
127 $ PERL_HASH_SEED_DEBUG=1 ./perl -e1
128 HASH_FUNCTION = MURMUR3 HASH_SEED = 0x1476bb9f
129
130=head2 Upgrade to Unicode 6.2
131
2e7bc647 132Perl now supports Unicode 6.2. A list of changes from Unicode
5ed58cbd
RS
1336.1 is at L<http://www.unicode.org/versions/Unicode6.2.0>.
134
135=head2 Character name aliases may now include non-Latin1-range characters
136
137It is possible to define your own names for characters for use in
138C<\N{...}>, C<charnames::vianame()>, etc. These names can now be
139comprised of characters from the whole Unicode range. This allows for
140names to be in your native language, and not just English. Certain
141restrictions apply to the characters that may be used (you can't define
142a name that has punctuation in it, for example). See L<charnames/CUSTOM
143ALIASES>.
144
145=head2 New DTrace probes
146
147The following new DTrace probes have been added:
14731ad1 148
337fb649 149=over 4
14731ad1 150
82d98f72 151=item *
14731ad1 152
5ed58cbd
RS
153C<op-entry>
154
155=item *
156
157C<loading-file>
158
159=item *
160
161C<loaded-file>
162
163=back
164
165=head2 C<${^LAST_FH}>
166
167This new variable provides access to the filehandle that was last read.
168This is the handle used by C<$.> and by C<tell> and C<eof> without
169arguments.
170
171=head2 Regular Expression Set Operations
172
173This is an B<experimental> feature to allow matching against the union,
174intersection, etc., of sets of code points, similar to
175L<Unicode::Regex::Set>. It can also be used to extend C</x> processing
176to [bracketed] character classes, and as a replacement of user-defined
177properties, allowing more complex expressions than they do. See
178L<perlrecharclass/Extended Bracketed Character Classes>.
179
180=head2 Lexical subroutines
181
182This new feature is still considered B<experimental>. To enable it:
183
184 use 5.018;
185 no warnings "experimental::lexical_subs";
186 use feature "lexical_subs";
187
188You can now declare subroutines with C<state sub foo>, C<my sub foo>, and
189C<our sub foo>. (C<state sub> requires that the "state" feature be
190enabled, unless you write it as C<CORE::state sub foo>.)
191
192C<state sub> creates a subroutine visible within the lexical scope in which
193it is declared. The subroutine is shared between calls to the outer sub.
194
195C<my sub> declares a lexical subroutine that is created each time the
196enclosing block is entered. C<state sub> is generally slightly faster than
197C<my sub>.
198
199C<our sub> declares a lexical alias to the package subroutine of the same
200name.
201
202For more information, see L<perlsub/Lexical Subroutines>.
203
204=head2 Computed Labels
205
206The loop controls C<next>, C<last> and C<redo>, and the special C<dump>
207operator, now allow arbitrary expressions to be used to compute labels at run
208time. Previously, any argument that was not a constant was treated as the
209empty string.
210
211=head2 More CORE:: subs
212
213Several more built-in functions have been added as subroutines to the
214CORE:: namespace - namely, those non-overridable keywords that can be
215implemented without custom parsers: C<defined>, C<delete>, C<exists>,
216C<glob>, C<pos>, C<protoytpe>, C<scalar>, C<split>, C<study>, and C<undef>.
217
218As some of these have prototypes, C<prototype('CORE::...')> has been
219changed to not make a distinction between overridable and non-overridable
220keywords. This is to make C<prototype('CORE::pos')> consistent with
221C<prototype(&CORE::pos)>.
222
223=head2 C<kill> with negative signal names
224
225C<kill> has always allowed a negative signal number, which kills the
226process group instead of a single process. It has also allowed signal
227names. But it did not behave consistently, because negative signal names
228were treated as 0. Now negative signals names like C<-INT> are supported
229and treated the same way as -2 [perl #112990].
230
231=head1 Security
232
233=head2 C<Storable> security warning in documentation
234
235The documentation for C<Storable> now includes a section which warns readers
236of the danger of accepting Storable documents from untrusted sources. The
237short version is that deserializing certain types of data can lead to loading
238modules and other code execution. This is documented behavior and wanted
239behavior, but this opens an attack vector for malicious entities.
240
241=head2 C<Locale::Maketext> allowed code injection via a malicious template
242
243If users could provide a translation string to Locale::Maketext, this could be
244used to invoke arbitrary Perl subroutines available in the current process.
245
246This has been fixed, but it is still possible to invoke any method provided by
247C<Locale::Maketext> itself or a subclass that you are using. One of these
248methods in turn will invoke the Perl core's C<sprintf> subroutine.
249
250In summary, allowing users to provide translation strings without auditing
251them is a bad idea.
252
253This vulnerability is documented in CVE-2012-6329.
254
255=head2 Avoid calling memset with a negative count
256
257Poorly written perl code that allows an attacker to specify the count to perl's
258C<x> string repeat operator can already cause a memory exhaustion
259denial-of-service attack. A flaw in versions of perl before 5.15.5 can escalate
260that into a heap buffer overrun; coupled with versions of glibc before 2.16, it
261possibly allows the execution of arbitrary code.
262
263The flaw addressed to this commit has been assigned identifier CVE-2012-5195
264and was researched by Tim Brown.
265
266=head1 Incompatible Changes
267
268=head2 See also: hash overhaul
269
270Some of the changes in the L<hash overhaul|/"Hash overhaul"> are not fully
271compatible with previous versions of perl. Please read that section.
272
273=head2 An unknown character name in C<\N{...}> is now a syntax error
274
275Previously, it warned, and the Unicode REPLACEMENT CHARACTER was
276substituted. Unicode now recommends that this situation be a syntax
277error. Also, the previous behavior led to some confusing warnings and
278behaviors, and since the REPLACEMENT CHARACTER has no use other than as
279a stand-in for some unknown character, any code that has this problem is
280buggy.
281
282=head2 Formerly deprecated characters in C<\N{}> character name aliases are now errors.
283
284Since v5.12.0, it has been deprecated to use certain characters in
285user-defined C<\N{...}> character names. These now cause a syntax
286error. For example, it is now an error to begin a name with a digit,
287such as in
288
289 my $undraftable = "\N{4F}"; # Syntax error!
290
291or to have commas anywhere in the name. See L<charnames/CUSTOM ALIASES>
292
293=head2 C<\N{BELL}> now refers to U+1F514 instead of U+0007
294
295Unicode 6.0 reused the name "BELL" for a different code point than it
296traditionally had meant. Since Perl v5.14, use of this name still
297referred to U+0007, but would raise a deprecation warning. Now, "BELL"
298refers to U+1F514, and the name for U+0007 is "ALERT". All the
299functions in L<charnames> have been correspondingly updated.
300
301=head2 New Restrictions in Multi-Character Case-Insensitive Matching in Regular Expression Bracketed Character Classes
302
303Unicode has now withdrawn their previous recommendation for regular
304expressions to automatically handle cases where a single character can
305match multiple characters case-insensitively, for example the letter
306LATIN SMALL LETTER SHARP S and the sequence C<ss>. This is because
307it turns out to be impracticable to do this correctly in all
308circumstances. Because Perl has tried to do this as best it can, it
309will continue to do so. (We are considering an option to turn it off.)
310However, a new restriction is being added on such matches when they
311occur in [bracketed] character classes. People were specifying
312things such as C</[\0-\xff]/i>, and being surprised that it matches the
313two character sequence C<ss> (since LATIN SMALL LETTER SHARP S occurs in
314this range). This behavior is also inconsistent with using a
315property instead of a range: C<\p{Block=Latin1}> also includes LATIN
316SMALL LETTER SHARP S, but C</[\p{Block=Latin1}]/i> does not match C<ss>.
317The new rule is that for there to be a multi-character case-insensitive
318match within a bracketed character class, the character must be
319explicitly listed, and not as an end point of a range. This more
320closely obeys the Principle of Least Astonishment. See
321L<perlrecharclass/Bracketed Character Classes>. Note that a bug [perl
322#89774], now fixed as part of this change, prevented the previous
323behavior from working fully.
324
325=head2 Explicit rules for variable names and identifiers
326
327Due to an oversight, length-one variable names in 5.16 were completely
328unrestricted, and opened the door to several kinds of insanity. As of
3295.18, these now follow the rules of other identifiers, in addition
330to accepting characters that match the C<\p{POSIX_Punct}> property.
331
332There are no longer any differences in the parsing of identifiers
333specified as C<$...> or C<${...}>; previously, they were dealt with in
334different parts of the core, and so had slightly different behavior. For
335instance, C<${foo:bar}> was a legal variable name. Since they are now
336both parsed by the same code, that is no longer the case.
337
338=head2 C<\s> in regular expressions now matches a Vertical Tab
339
340No one could recall why C<\s> didn't match C<\cK>, the vertical tab.
341Now it does. Given the extreme rarity of that character, very little
342breakage is expected.
343
344=head2 C</(?{})/> and C</(??{})/> have been heavily reworked
345
346The implementation of this feature has been almost completely rewritten.
347Although its main intent is to fix bugs, some behaviors, especially
348related to the scope of lexical variables, will have changed. This is
349described more fully in the L</Selected Bug Fixes> section.
350
351=head2 Stricter parsing of substitution replacement
352
353It is no longer possible to abuse the way the parser parses C<s///e> like
354this:
355
356 %_=(_,"Just another ");
357 $_="Perl hacker,\n";
358 s//_}->{_/e;print
359
360=head2 C<given> now aliases the global C<$_>
361
362Instead of assigning to an implicit lexical C<$_>, C<given> now makes the
363global C<$_> an alias for its argument, just like C<foreach>. However, it
364still uses lexical C<$_> if there is lexical C<$_> in scope (again, just like
365C<foreach>) [perl #114020].
366
367=head2 Lexical C<$_> is now experimental
368
369Since it was introduced in Perl 5.10, it has caused much confusion with no
370obvious solution:
371
372=over
829397b9
TC
373
374=item *
375
5ed58cbd
RS
376Various modules (e.g., List::Util) expect callback routines to use the
377global C<$_>. C<use List::Util 'first'; my $_; first { $_ == 1 } @list>
378does not work as one would expect.
71e6aba6
RS
379
380=item *
381
5ed58cbd
RS
382A C<my $_> declaration earlier in the same file can cause confusing closure
383warnings.
71e6aba6
RS
384
385=item *
386
5ed58cbd
RS
387The "_" subroutine prototype character allows called subroutines to access
388your lexical C<$_>, so it is not really private after all.
2426c394 389
71e6aba6
RS
390=item *
391
5ed58cbd
RS
392Nevertheless, subroutines with a "(@)" prototype and methods cannot access
393the caller's lexical C<$_>, unless they are written in XS.
71e6aba6
RS
394
395=item *
396
5ed58cbd
RS
397But even XS routines cannot access a lexical C<$_> declared, not in the
398calling subroutine, but in an outer scope, iff that subroutine happened not
399to mention C<$_> or use any operators that default to C<$_>.
400
401=back
402
403It is our hope that lexical C<$_> can be rehabilitated, but this may
404cause changes in its behavior. Please use it with caution until it
405becomes stable.
406
407=head2 readline() with C<$/ = \N> now reads N characters, not N bytes
408
409Previously, when reading from a stream with I/O layers such as
410C<encoding>, the readline() function, otherwise known as the C<< <> >>
411operator, would read I<N> bytes from the top-most layer. [perl #79960]
412
413Now, I<N> characters are read instead.
414
415There is no change in behaviour when reading from streams with no
416extra layers, since bytes map exactly to characters.
417
418=head2 Overridden C<glob> is now passed one argument
419
420C<glob> overrides used to be passed a magical undocumented second argument
421that identified the caller. Nothing on CPAN was using this, and it got in
422the way of a bug fix, so it was removed. If you really need to identify
423the caller, see L<Devel::Callsite> on CPAN.
424
425=head2 Here-doc parsing
426
427The body of a here-document inside a quote-like operator now always begins
428on the line after the "<<foo" marker. Previously, it was documented to
429begin on the line following the containing quote-like operator, but that
430was only sometimes the case [perl #114040].
431
432=head2 Alphanumeric operators must now be separated from the closing
433delimiter of regular expressions
434
435You may no longer write something like:
436
437 m/a/and 1
438
439Instead you must write
440
441 m/a/ and 1
442
443with whitespace separating the operator from the closing delimiter of
444the regular expression. Not having whitespace has resulted in a
445deprecation warning since Perl v5.14.0.
446
447=head2 qw(...) can no longer be used as parentheses
448
449C<qw> lists used to fool the parser into thinking they were always
450surrounded by parentheses. This permitted some surprising constructions
451such as C<foreach $x qw(a b c) {...}>, which should really be written
452C<foreach $x (qw(a b c)) {...}>. These would sometimes get the lexer into
453the wrong state, so they didn't fully work, and the similar C<foreach qw(a
454b c) {...}> that one might expect to be permitted never worked at all.
455
456This side effect of C<qw> has now been abolished. It has been deprecated
457since Perl 5.13.11. It is now necessary to use real parentheses
458everywhere that the grammar calls for them.
459
460=head2 Interaction of lexical and default warnings
461
462Turning on any lexical warnings used first to disable all default warnings
463if lexical warnings were not already enabled:
464
465 $*; # deprecation warning
466 use warnings "void";
467 $#; # void warning; no deprecation warning
468
f105b7be 469Now, the C<debugging>, C<deprecated>, C<glob>, C<inplace> and C<malloc> warnings
5ed58cbd
RS
470categories are left on when turning on lexical warnings (unless they are
471turned off by C<no warnings>, of course).
472
473This may cause deprecation warnings to occur in code that used to be free
474of warnings.
475
476Those are the only categories consisting only of default warnings. Default
f105b7be
KE
477warnings in other categories are still disabled by C<< use warnings "category" >>,
478as we do not yet have the infrastructure for controlling
5ed58cbd
RS
479individual warnings.
480
481=head2 C<state sub> and C<our sub>
482
483Due to an accident of history, C<state sub> and C<our sub> were equivalent
484to a plain C<sub>, so one could even create an anonymous sub with
485C<our sub { ... }>. These are now disallowed outside of the "lexical_subs"
486feature. Under the "lexical_subs" feature they have new meanings described
487in L<perlsub/Lexical Subroutines>.
488
489=head2 Defined values stored in environment are forced to byte strings
490
491A value stored in an environment variable has always been stringified. In this
492release, it is converted to be only a byte string. First, it is forced to be a
493only a string. Then if the string is utf8 and the equivalent of
494C<utf8::downgrade()> works, that result is used; otherwise, the equivalent of
495C<utf8::encode()> is used, and a warning is issued about wide characters
496(L</Diagnostics>).
497
498=head2 C<require> dies for unreadable files
499
500When C<require> encounters an unreadable file, it now dies. It used to
501ignore the file and continue searching the directories in C<@INC>
502[perl #113422].
503
504=head2 C<gv_fetchmeth_*> and SUPER
505
506The various C<gv_fetchmeth_*> XS functions used to treat a package whose
f105b7be
KE
507named ended with C<::SUPER> specially. A method lookup on the C<Foo::SUPER>
508package would be treated as a C<SUPER> method lookup on the C<Foo> package. This
509is no longer the case. To do a C<SUPER> lookup, pass the C<Foo> stash and the
510C<GV_SUPER> flag.
5ed58cbd
RS
511
512=head2 C<split>'s first argument is more consistently interpreted
513
514After some changes earlier in 5.17, C<split>'s behavior has been
515simplified: if the PATTERN argument evaluates to a literal string
516containing one space, it is treated the way that a I<literal> string
517containing one space once was.
518
519=head1 Deprecations
520
521=head2 Deprecated modules
522
523The following modules will be removed from the core distribution in a
524future release, and should be installed from CPAN instead. Distributions
525on CPAN which require these should add them to their prerequisites.
f105b7be 526The core versions of these modules will issue C<"deprecated">-category
5ed58cbd
RS
527warnings.
528
529You can silence these deprecation warnings by installing the modules
530in question from CPAN.
531
532=over
533
534=item L<Archive::Extract>
535
536=item L<B::Lint>
537
538=item L<B::Lint::Debug>
539
540=item L<CPANPLUS> and all included C<CPANPLUS::*> modules
541
542=item L<Devel::InnerPackage>
543
544=item L<encoding>
545
546=item L<Log::Message>
547
548=item L<Log::Message::Config>
549
550=item L<Log::Message::Handlers>
551
552=item L<Log::Message::Item>
553
554=item L<Log::Message::Simple>
555
556=item L<Module::Pluggable>
557
558=item L<Module::Pluggable::Object>
559
560=item L<Object::Accessor>
561
562=item L<Pod::LaTeX>
563
564=item L<Term::UI>
565
566=item L<Term::UI::History>
567
568=back
569
570=head2 Deprecated Utilities
571
572The following utilities will be removed from the core distribution in a
573future release as their associated modules have been deprecated. They
574will remain available with the applicable CPAN distribution.
575
576=over
577
578=item L<cpanp>
579
580=item C<cpanp-run-perl>
581
582=item L<cpan2dist>
583
584These items are part of the C<CPANPLUS> distribution.
585
586=item L<pod2latex>
587
588This item is part of the C<Pod::LaTeX> distribution.
589
590=back
591
592=head2 PL_sv_objcount
593
594This interpreter-global variable used to track the total number of
595Perl objects in the interpreter. It is no longer maintained and will
596be removed altogether in Perl 5.20.
597
598=head2 Five additional characters should be escaped in patterns with C</x>
599
600When a regular expression pattern is compiled with C</x>, Perl treats 6
601characters as white space to ignore, such as SPACE and TAB. However,
602Unicode recommends 11 characters be treated thusly. We will conform
603with this in a future Perl version. In the meantime, use of any of the
604missing characters will raise a deprecation warning, unless turned off.
605The five characters are:
f105b7be
KE
606
607 U+0085 NEXT LINE,
608 U+200E LEFT-TO-RIGHT MARK,
609 U+200F RIGHT-TO-LEFT MARK,
610 U+2028 LINE SEPARATOR,
611
5ed58cbd 612and
f105b7be
KE
613
614 U+2029 PARAGRAPH SEPARATOR.
5ed58cbd
RS
615
616=head2 User-defined charnames with surprising whitespace
617
618A user-defined character name with trailing or multiple spaces in a row is
619likely a typo. This now generates a warning when defined, on the assumption
620that uses of it will be unlikely to include the excess whitespace.
621
622=head2 Various XS-callable functions are now deprecated
623
624All the functions used to classify characters will be removed from a
625future version of Perl, and should not be used. With participating C
626compilers (e.g., gcc), compiling any file that uses any of these will
627generate a warning. These were not intended for public use; there are
628equivalent, faster, macros for most of them.
2e7bc647 629See L<perlapi/Character classes>. The complete list is:
5ed58cbd
RS
630C<is_uni_alnum>, C<is_uni_alnumc>, C<is_uni_alnumc_lc>,
631C<is_uni_alnum_lc>, C<is_uni_alpha>, C<is_uni_alpha_lc>,
632C<is_uni_ascii>, C<is_uni_ascii_lc>, C<is_uni_blank>,
633C<is_uni_blank_lc>, C<is_uni_cntrl>, C<is_uni_cntrl_lc>,
634C<is_uni_digit>, C<is_uni_digit_lc>, C<is_uni_graph>,
635C<is_uni_graph_lc>, C<is_uni_idfirst>, C<is_uni_idfirst_lc>,
636C<is_uni_lower>, C<is_uni_lower_lc>, C<is_uni_print>,
637C<is_uni_print_lc>, C<is_uni_punct>, C<is_uni_punct_lc>,
638C<is_uni_space>, C<is_uni_space_lc>, C<is_uni_upper>,
639C<is_uni_upper_lc>, C<is_uni_xdigit>, C<is_uni_xdigit_lc>,
640C<is_utf8_alnum>, C<is_utf8_alnumc>, C<is_utf8_alpha>,
641C<is_utf8_ascii>, C<is_utf8_blank>, C<is_utf8_char>,
642C<is_utf8_cntrl>, C<is_utf8_digit>, C<is_utf8_graph>,
643C<is_utf8_idcont>, C<is_utf8_idfirst>, C<is_utf8_lower>,
644C<is_utf8_mark>, C<is_utf8_perl_space>, C<is_utf8_perl_word>,
645C<is_utf8_posix_digit>, C<is_utf8_print>, C<is_utf8_punct>,
646C<is_utf8_space>, C<is_utf8_upper>, C<is_utf8_xdigit>,
647C<is_utf8_xidcont>, C<is_utf8_xidfirst>.
648
649In addition these three functions that have never worked properly are
650deprecated:
651C<to_uni_lower_lc>, C<to_uni_title_lc>, and C<to_uni_upper_lc>.
652
f105b7be 653=head2 Certain rare uses of backslashes within regexes are now deprecated
5ed58cbd
RS
654
655There are three pairs of characters that Perl recognizes as
656metacharacters in regular expression patterns: C<{}>, C<[]>, and C<()>.
657These can be used as well to delimit patterns, as in:
658
659 m{foo}
660 s(foo)(bar)
661
662Since they are metacharacters, they have special meaning to regular
663expression patterns, and it turns out that you can't turn off that
664special meaning by the normal means of preceding them with a backslash,
f105b7be 665if you use them, paired, within a pattern delimited by them. For
5ed58cbd
RS
666example, in
667
668 m{foo\{1,3\}}
669
670the backslashes do not change the behavior, and this matches
671S<C<"f o">> followed by one to three more occurrences of C<"o">.
672
673Usages like this, where they are interpreted as metacharacters, are
674exceedingly rare; we think there are none, for example, in all of CPAN.
675Hence, this deprecation should affect very little code. It does give
676notice, however, that any such code needs to change, which will in turn
677allow us to change the behavior in future Perl versions so that the
678backslashes do have an effect, and without fear that we are silently
679breaking any existing code.
680
d5f315e8
KW
681=head2 Splitting the tokens C<(?> and C<(*> in regular expressions
682
683A deprecation warning is now raised if the C<(> and C<?> are separated
684by white space or comments in C<(?...)> regular expression constructs.
685Similarly, if the C<(> and C<*> are separated in C<(*VERB...)>
686constructs.
687
e0a1dec5
LT
688=head2 Pre-PerlIO IO implementations
689
690Perl supports being built without PerlIO proper, using a stdio or sfio
691wrapper instead. A perl build like this will not support IO layers and
692thus Unicode IO, making it rather handicapped.
693
694PerlIO supports a C<stdio> layer if stdio use is desired, and similarly a
695sfio layer could be produced.
696
5ed58cbd
RS
697=head1 Future Deprecations
698
699=over
71e6aba6
RS
700
701=item *
702
4263dd11 703Platforms without support infrastructure
5ed58cbd
RS
704
705Both Windows CE and z/OS have been historically under-maintained, and are
706currently neither successfully building nor regularly being smoke tested.
707Efforts are underway to change this situation, but it should not be taken for
708granted that the platforms are safe and supported. If they do not become
709buildable and regularly smoked, support for them may be actively removed in
710future releases. If you have an interest in these platforms and you can lend
711your time, expertise, or hardware to help support these platforms, please let
712the perl development effort know by emailing C<perl5-porters@perl.org>.
713
714Some platforms that appear otherwise entirely dead are also on the short list
715for removal between now and 5.20.0:
716
717=over
718
719=item DG/UX
720
721=item NeXT
722
723=back
1993add8 724
ec985017
RS
725We also think it likely that current versions of Perl will no longer
726build AmigaOS, DJGPP, NetWare (natively), OS/2 and Plan 9. If you
727are using Perl on such a platform and have an interest in ensuring
728Perl's future on them, please contact us.
729
730We believe that Perl has long been unable to build on mixed endian
731architectures (such as PDP-11s), and intend to remove any remain
732support code. Similarly, code supporting the long umaintained GNU
733dld will be removed soon if no-one makes themselves known as an
734active user.
735
1993add8
RS
736=item *
737
5ed58cbd
RS
738Swapping of $< and $>
739
740For more information about this future deprecation, see L<the relevant RT
741ticket|https://rt.perl.org/rt3/Ticket/Display.html?id=96212>.
71e6aba6
RS
742
743=item *
744
5ed58cbd 745C<microperl>, long broken and of unclear present purpose, will be removed.
71e6aba6
RS
746
747=item *
748
5ed58cbd
RS
749Revamping C<< "\Q" >> semantics in double-quotish strings when combined with
750other escapes.
751
752There are several bugs and inconsistencies involving combinations
753of C<\Q> and escapes like C<\x>, C<\L>, etc., within a C<\Q...\E> pair.
754These need to be fixed, and doing so will necessarily change current
755behavior. The changes have not yet been settled.
71e6aba6 756
d5f315e8
KW
757=item *
758
759Use of C<$^>, where C<^> stands for any actual (non-printing) C0 control
760character will be disallowed in a future Perl version. Use C<${^}>
761instead (where again C<^> stands for a control character),
762or better, C<$^A> , where C<^> this time is a caret (CIRCUMFLEX ACCENT),
763and C<A> stands for any of the characters listed at the end of
764L<perlebcdic/OPERATOR DIFFERENCES>.
765
337fb649 766=back
2426c394 767
5ed58cbd
RS
768=head1 Performance Enhancements
769
770=over 4
2426c394 771
5ed58cbd 772=item *
2426c394 773
5ed58cbd
RS
774Lists of lexical variable declarations (C<my($x, $y)>) are now optimised
775down to a single op and are hence faster than before.
2426c394 776
5ed58cbd 777=item *
2426c394 778
5ed58cbd
RS
779A new C preprocessor define C<NO_TAINT_SUPPORT> was added that, if set,
780disables Perl's taint support altogether. Using the -T or -t command
781line flags will cause a fatal error. Beware that both core tests as
782well as many a CPAN distribution's tests will fail with this change. On
783the upside, it provides a small performance benefit due to reduced
784branching.
2426c394 785
5ed58cbd
RS
786B<Do not enable this unless you know exactly what you are getting yourself
787into.>
788
789=item *
790
791C<pack> with constant arguments is now constant folded in most cases
792[perl #113470].
793
794=item *
795
796Speed up in regular expression matching against Unicode properties. The
797largest gain is for C<\X>, the Unicode "extended grapheme cluster." The
798gain for it is about 35% - 40%. Bracketed character classes, e.g.,
799C<[0-9\x{100}]> containing code points above 255 are also now faster.
800
801=item *
802
803On platforms supporting it, several former macros are now implemented as static
804inline functions. This should speed things up slightly on non-GCC platforms.
805
806=item *
807
66f62cf6
RS
808The optimisation of hashes in boolean context has been extended to
809affect C<scalar(%hash)>, C<%hash ? ... : ...>, and C<sub { %hash || ... }>.
5ed58cbd
RS
810
811=item *
812
f105b7be 813Filetest operators manage the stack in a fractionally more efficient manner.
5ed58cbd
RS
814
815=item *
816
817Globs used in a numeric context are now numified directly in most cases,
f105b7be 818rather than being numified via stringification.
5ed58cbd
RS
819
820=item *
821
822The C<x> repetition operator is now folded to a single constant at compile
823time if called in scalar context with constant operands and no parentheses
824around the left operand.
825
826=back
827
828=head1 Modules and Pragmata
829
830=head2 New Modules and Pragmata
2426c394 831
337fb649 832=over 4
982110e0 833
82d98f72 834=item *
2426c394 835
5ed58cbd
RS
836L<Config::Perl::V> version 0.16 has been added as a dual-lifed module.
837It provides structured data retrieval of C<perl -V> output including
838information only known to the C<perl> binary and not available via L<Config>.
839
840=back
841
842=head2 Updated Modules and Pragmata
843
844This is only an overview of selected module updates. For a complete
845list of updates, run:
846
f105b7be 847 $ corelist --diff 5.16.0 5.18.0
5ed58cbd 848
f105b7be 849You can substitute your favorite version in place of 5.16.0, too.
5ed58cbd
RS
850
851=over 4
33392251
BF
852
853=item *
854
5ed58cbd
RS
855L<XXX> has been upgraded from version A.xx to B.yy.
856
857=back
858
859=head2 Removed Modules and Pragmata
860
861=over
33392251
BF
862
863=item *
864
5ed58cbd
RS
865L<Version::Requirements> has been removed from the core distribution. It is
866available under a different name: L<CPAN::Meta::Requirements>.
2426c394 867
337fb649 868=back
2426c394 869
5ed58cbd 870=head1 Documentation
19718730 871
5ed58cbd
RS
872=head2 Changes to Existing Documentation
873
874=head3 L<perlcheat>
82d98f72 875
5a6a30f4 876=over 4
b7c7d786 877
5ed58cbd
RS
878=item *
879
880L<perlcheat> has been reorganized, and a few new sections were added.
881
882=back
883
884=head3 L<perldata>
885
886=over 4
82d98f72 887
5ed58cbd 888=item *
d2d1e842 889
5ed58cbd
RS
890Now explicitly documents the behaviour of hash initializer lists that
891contain duplicate keys.
f355e93d 892
5a6a30f4 893=back
f355e93d 894
5ed58cbd 895=head3 L<perldiag>
19718730 896
19718730 897=over 4
e14ac59b 898
5ed58cbd
RS
899=item *
900
901The explanation of symbolic references being prevented by "strict refs"
902now doesn't assume that the reader knows what symbolic references are.
903
904=back
9f351b45 905
5ed58cbd 906=head3 L<perlfaq>
9f351b45 907
5ed58cbd 908=over 4
9f351b45 909
5ed58cbd 910=item *
7cf3104f 911
5ed58cbd 912L<perlfaq> has been synchronized with version 5.0150040 from CPAN.
12719193 913
6253ee75 914=back
216cf7fc 915
5ed58cbd 916=head3 L<perlfunc>
f5b73711 917
5ed58cbd
RS
918=over 4
919
920=item *
a75569c0 921
5ed58cbd 922The return value of C<pipe> is now documented.
a75569c0 923
5ed58cbd 924=item *
a75569c0 925
5ed58cbd
RS
926Clarified documentation of C<our>.
927
928=back
929
930=head3 L<perlop>
931
932=over 4
933
934=item *
935
936Loop control verbs (C<dump>, C<goto>, C<next>, C<last> and C<redo>) have always
937had the same precedence as assignment operators, but this was not documented
938until now.
939
940=back
941
942=head3 Diagnostics
943
944The following additions or changes have been made to diagnostic output,
945including warnings and fatal error messages. For the complete list of
946diagnostic messages, see L<perldiag>.
947
948XXX New or changed warnings emitted by the core's C<C> code go here. Also
949include any changes in L<perldiag> that reconcile it to the C<C> code.
950
951=head2 New Diagnostics
952
953XXX Newly added diagnostic messages go under here, separated into New Errors
954and New Warnings
955
956=head3 New Errors
957
958=over 4
959
960=item *
961
962L<Unterminated delimiter for here document|perldiag/"Unterminated delimiter for here document">
963
964This message now occurs when a here document label has an initial quotation
965mark but the final quotation mark is missing.
966
967This replaces a bogus and misleading error message about not finding the label
968itself [perl #114104].
969
970=item *
971
972L<panic: child pseudo-process was never scheduled|perldiag/"panic: child pseudo-process was never scheduled">
973
974This error is thrown when a child pseudo-process in the ithreads implementation
975on Windows was not scheduled within the time period allowed and therefore was
976not able to initialize properly [perl #88840].
977
978=item *
979
980L<Group name must start with a non-digit word character in regex; marked by <-- HERE in mE<sol>%sE<sol>|perldiag/"Group name must start with a non-digit word character in regex; marked by <-- HERE in m/%s/">
981
982This error has been added for C<(?&0)>, which is invalid. It used to
983produce an incomprehensible error message [perl #101666].
984
985=item *
986
987L<Can't use an undefined value as a subroutine reference|perldiag/"Can't use an undefined value as %s reference">
988
989Calling an undefined value as a subroutine now produces this error message.
990It used to, but was accidentally disabled, first in Perl 5.004 for
991non-magical variables, and then in Perl 5.14 for magical (e.g., tied)
992variables. It has now been restored. In the mean time, undef was treated
993as an empty string [perl #113576].
994
995=item *
996
997L<Experimental "%s" subs not enabled|perldiag/"Experimental "%s" subs not enabled">
998
999To use lexical subs, you must first enable them:
1000
1001 no warnings 'experimental::lexical_subs';
1002 use feature 'lexical_subs';
1003 my sub foo { ... }
1004
1005=back
1006
1007=head3 New Warnings
1008
1009=over 4
1010
1011=item *
1012
1013XXX: This needs more detail.
1014
1015Strings with code points over 0xFF may not be mapped into in-memory file
1016handles
1017
1018=item *
1019
1020L<'%s' resolved to '\o{%s}%d'|perldiag/"'%s' resolved to '\o{%s}%d'">
1021
1022=item *
1023
1024L<'Trailing white-space in a charnames alias definition is deprecated'|perldiag/"Trailing white-space in a charnames alias definition is deprecated">
1025
1026=item *
1027
1028L<'A sequence of multiple spaces in a charnames alias definition is deprecated'|perldiag/"A sequence of multiple spaces in a charnames alias definition is deprecated">
1029
1030=item *
1031
1032L<'Passing malformed UTF-8 to "%s" is deprecated'|perldiag/"Passing malformed UTF-8 to "%s" is deprecated">
1033
1034=item *
1035
1036L<Subroutine "&%s" is not available|perldiag/"Subroutine "&%s" is not available">
1037
1038(W closure) During compilation, an inner named subroutine or eval is
1039attempting to capture an outer lexical subroutine that is not currently
1040available. This can happen for one of two reasons. First, the lexical
1041subroutine may be declared in an outer anonymous subroutine that has not
1042yet been created. (Remember that named subs are created at compile time,
1043while anonymous subs are created at run-time.) For example,
1044
1045 sub { my sub a {...} sub f { \&a } }
1046
1047At the time that f is created, it can't capture the current the "a" sub,
1048since the anonymous subroutine hasn't been created yet. Conversely, the
1049following won't give a warning since the anonymous subroutine has by now
1050been created and is live:
1051
1052 sub { my sub a {...} eval 'sub f { \&a }' }->();
1053
1054The second situation is caused by an eval accessing a variable that has
1055gone out of scope, for example,
1056
1057 sub f {
1058 my sub a {...}
1059 sub { eval '\&a' }
1060 }
1061 f()->();
1062
1063Here, when the '\&a' in the eval is being compiled, f() is not currently
1064being executed, so its &a is not available for capture.
1065
1066=item *
1067
1068L<"%s" subroutine &%s masks earlier declaration in same %s|perldiag/"%s" subroutine &%s masks earlier declaration in same %s>
1069
1070(W misc) A "my" or "state" subroutine has been redeclared in the
1071current scope or statement, effectively eliminating all access to
1072the previous instance. This is almost always a typographical error.
1073Note that the earlier subroutine will still exist until the end of
1074the scope or until all closure references to it are destroyed.
1075
1076=item *
1077
1078L<The %s feature is experimental|perldiag/"The %s feature is experimental">
1079
1080(S experimental) This warning is emitted if you enable an experimental
1081feature via C<use feature>. Simply suppress the warning if you want
1082to use the feature, but know that in doing so you are taking the risk
1083of using an experimental feature which may change or be removed in a
1084future Perl version:
1085
1086 no warnings "experimental::lexical_subs";
1087 use feature "lexical_subs";
1088
1089=item *
1090
1091L<sleep(%u) too large|perldiag/"sleep(%u) too large">
1092
1093(W overflow) You called C<sleep> with a number that was larger than it can
1094reliably handle and C<sleep> probably slept for less time than requested.
1095
1096=item *
1097
1098L<Wide character in setenv|perldiag/"Wide character in %s">
1099
1100Attempts to put wide characters into environment variables via C<%ENV> now
1101provoke this warning.
1102
1103=item *
1104
1105"L<Invalid negative number (%s) in chr|perldiag/"Invalid negative number (%s) in chr">"
1106
1107C<chr()> now warns when passed a negative value [perl #83048].
1108
1109=item *
1110
1111"L<Integer overflow in srand|perldiag/"Integer overflow in srand">"
1112
1113C<srand()> now warns when passed a value that doesn't fit in a C<UV> (since the
1114value will be truncated rather than overflowing) [perl #40605].
1115
1116=item *
1117
1118"L<-i used with no filenames on the command line, reading from STDIN|perldiag/"-i used with no filenames on the command line, reading from STDIN">"
1119
1120Running perl with the C<-i> flag now warns if no input files are provided on
1121the command line [perl #113410].
1122
1123=back
1124
1125=head2 Changes to Existing Diagnostics
1126
1127=over 4
1128
1129=item *
1130
1131L<$* is no longer supported|perldiag/"$* is no longer supported">
1132
1133The warning that use of C<$*> and C<$#> is no longer supported is now
1134generated for every location that references them. Previously it would fail
1135to be generated if another variable using the same typeglob was seen first
1136(e.g. C<@*> before C<$*>), and would not be generated for the second and
1137subsequent uses. (It's hard to fix the failure to generate warnings at all
1138without also generating them every time, and warning every time is
1139consistent with the warnings that C<$[> used to generate.)
1140
1141=item *
1142
1143The warnings for C<\b{> and C<\B{> were added. They are a deprecation
1144warning which should be turned off by that category. One should not
1145have to turn off regular regexp warnings as well to get rid of these.
1146
1147=item *
1148
1149L<Constant(%s): Call to &{$^H{%s}} did not return a defined value|perldiag/Constant(%s): Call to &{$^H{%s}} did not return a defined value>
1150
1151Constant overloading that returns C<undef> results in this error message.
1152For numeric constants, it used to say "Constant(undef)". "undef" has been
1153replaced with the number itself.
1154
1155=item *
1156
1157The error produced when a module cannot be loaded now includes a hint that
1158the module may need to be installed: "Can't locate hopping.pm in @INC (you
1159may need to install the hopping module) (@INC contains: ...)"
1160
1161=item *
1162
1163L<vector argument not supported with alpha versions|perldiag/vector argument not supported with alpha versions>
1164
1165This warning was not suppressable, even with C<no warnings>. Now it is
1166suppressible, and has been moved from the "internal" category to the
1167"printf" category.
1168
1169=item *
1170
1171C<< Can't do {n,m} with n > m in regex; marked by <-- HERE in m/%s/ >>
1172
1173This fatal error has been turned into a warning that reads:
1174
1175L<< Quantifier {n,m} with n > m can't match in regex | perldiag/Quantifier {n,m} with n > m can't match in regex >>
1176
1177(W regexp) Minima should be less than or equal to maxima. If you really want
1178your regexp to match something 0 times, just put {0}.
1179
1180=item *
1181
1182The "Runaway prototype" warning that occurs in bizarre cases has been
1183removed as being unhelpful and inconsistent.
1184
1185=item *
1186
1187The "Not a format reference" error has been removed, as the only case in
1188which it could be triggered was a bug.
1189
1190=item *
1191
1192The "Unable to create sub named %s" error has been removed for the same
1193reason.
1194
1195=item *
1196
1197The 'Can't use "my %s" in sort comparison' error has been downgraded to a
1198warning, '"my %s" used in sort comparison' (with 'state' instead of 'my'
1199for state variables). In addition, the heuristics for guessing whether
1200lexical $a or $b has been misused have been improved to generate fewer
1201false positives. Lexical $a and $b are no longer disallowed if they are
1202outside the sort block. Also, a named unary or list operator inside the
1203sort block no longer causes the $a or $b to be ignored [perl #86136].
1204
1205=back
1206
1207=head1 Utility Changes
1208
1209=head3 L<h2xs>
1210
1211=over 4
1212
1213=item *
1214
1215F<h2xs> no longer produces invalid code for empty defines. [perl #20636]
1216
1217=back
1218
1219=head1 Configuration and Compilation
1220
1221=over 4
1222
1223=item *
1224
1225Added C<useversionedarchname> option to Configure
1226
1227When set, it includes 'api_versionstring' in 'archname'. E.g.
1228x86_64-linux-5.13.6-thread-multi. It is unset by default.
1229
1230This feature was requested by Tim Bunce, who observed that
f105b7be 1231C<INSTALL_BASE> creates a library structure that does not
5ed58cbd
RS
1232differentiate by perl version. Instead, it places architecture
1233specific files in "$install_base/lib/perl5/$archname". This makes
f105b7be 1234it difficult to use a common C<INSTALL_BASE> library path with
5ed58cbd
RS
1235multiple versions of perl.
1236
f105b7be 1237By setting C<-Duseversionedarchname>, the $archname will be
c2959982 1238distinct for architecture I<and> API version, allowing mixed use of
f105b7be 1239C<INSTALL_BASE>.
5ed58cbd
RS
1240
1241=item *
1242
ff772877
RS
1243Add a C<PERL_NO_INLINE_FUNCTIONS> option
1244
f105b7be 1245If C<PERL_NO_INLINE_FUNCTIONS> is defined, don't include "inline.h"
ff772877
RS
1246
1247This permits test code to include the perl headers for definitions without
1248creating a link dependency on the perl library (which may not exist yet).
1249
1250=item *
1251
5ed58cbd
RS
1252Configure will honour the external C<MAILDOMAIN> environment variable, if set.
1253
1254=item *
1255
1256C<installman> no longer ignores the silent option
1257
1258=item *
1259
1260Both C<META.yml> and C<META.json> files are now included in the distribution.
1261
1262=item *
1263
1264F<Configure> will now correctly detect C<isblank()> when compiling with a C++
1265compiler.
1266
1267=item *
1268
1269The pager detection in F<Configure> has been improved to allow responses which
1270specify options after the program name, e.g. B</usr/bin/less -R>, if the user
1271accepts the default value. This helps B<perldoc> when handling ANSI escapes
1272[perl #72156].
1273
1274=back
1275
1276=head1 Testing
1277
1278=over 4
1279
1280=item *
1281
1282The test suite now has a section for tests that require very large amounts
1283of memory. These tests won't run by default; they can be enabled by
1284setting the C<PERL_TEST_MEMORY> environment variable to the number of
1285gibibytes of memory that may be safely used.
1286
1287=back
1288
1289=head1 Platform Support
1290
1291=head2 Discontinued Platforms
1292
1293=over 4
1294
1295=item BeOS
1296
1297BeOS was an operating system for personal computers developed by Be Inc,
1298initially for their BeBox hardware. The OS Haiku was written as an open
1299source replacement for/continuation of BeOS, and its perl port is current and
1300actively maintained.
1301
1302=item UTS Global
1303
1304Support code relating to UTS global has been removed. UTS was a mainframe
1305version of System V created by Amdahl, subsequently sold to UTS Global. The
1306port has not been touched since before Perl 5.8.0, and UTS Global is now
1307defunct.
1308
1309=item VM/ESA
1310
1311Support for VM/ESA has been removed. The port was tested on 2.3.0, which
1312IBM ended service on in March 2002. 2.4.0 ended service in June 2003, and
1313was superseded by Z/VM. The current version of Z/VM is V6.2.0, and scheduled
1314for end of service on 2015/04/30.
1315
1316=item MPE/IX
1317
1318Support for MPE/IX has been removed.
1319
1320=item EPOC
1321
1322Support code relating to EPOC has been removed. EPOC was a family of
1323operating systems developed by Psion for mobile devices. It was the
1324predecessor of Symbian. The port was last updated in April 2002.
1325
1326=item Rhapsody
1327
1328Support for Rhapsody has been removed.
1329
1330=back
1331
1332=head2 Platform-Specific Notes
1333
1334=head3 AIX
1335
1336Configure now always adds C<-qlanglvl=extc99> to the CC flags on AIX when
1337using xlC. This will make it easier to compile a number of XS-based modules
1338that assume C99 [perl #113778].
1339
1340=head3 clang++
1341
1342There is now a workaround for a compiler bug that prevented compiling
1343with clang++ since Perl 5.15.7 [perl #112786].
1344
1345=head3 C++
1346
1347When compiling the Perl core as C++ (which is only semi-supported), the
1348mathom functions are now compiled as C<extern "C">, to ensure proper
1349binary compatibility. (However, binary compatibility isn't generally
1350guaranteed anyway in the situations where this would matter.)
1351
1352=head3 Darwin
1353
1354Stop hardcoding an alignment on 8 byte boundaries to fix builds using
1355-Dusemorebits.
1356
1357=head3 Haiku
1358
1359Perl should now work out of the box on Haiku R1 Alpha 4.
1360
1361=head3 MidnightBSD
1362
1363C<libc_r> was removed from recent versions of MidnightBSD and older versions
1364work better with C<pthread>. Threading is now enabled using C<pthread> which
1365corrects build errors with threading enabled on 0.4-CURRENT.
1366
1367=head3 Solaris
1368
1369In Configure, avoid running sed commands with flags not supported on Solaris.
1370
1371=head3 VMS
1372
1373=over
1374
1375=item *
1376
1377Where possible, the case of filenames and command-line arguments is now
1378preserved by enabling the CRTL features C<DECC$EFS_CASE_PRESERVE> and
1379C<DECC$ARGV_PARSE_STYLE> at start-up time. The latter only takes effect
1380when extended parse is enabled in the process from which Perl is run.
1381
1382=item *
1383
1384The character set for Extended Filename Syntax (EFS) is now enabled by default
1385on VMS. Among other things, this provides better handling of dots in directory
05f5908f 1386names, multiple dots in filenames, and spaces in filenames. To obtain the old
5ed58cbd
RS
1387behavior, set the logical name C<DECC$EFS_CHARSET> to C<DISABLE>.
1388
1389=item *
1390
05f5908f 1391Fixed linking on builds configured with C<-Dusemymalloc=y>.
5ed58cbd
RS
1392
1393=item *
1394
05f5908f
CB
1395Experimental support for building Perl with the HP C++ compiler is available
1396by configuring with C<-Dusecxx>.
5ed58cbd
RS
1397
1398=item *
1399
1400All C header files from the top-level directory of the distribution are now
1401installed on VMS, providing consistency with a long-standing practice on other
1402platforms. Previously only a subset were installed, which broke non-core
1403extension builds for extensions that depended on the missing include files.
1404
1405=item *
1406
1407Quotes are now removed from the command verb (but not the parameters) for
1408commands spawned via C<system>, backticks, or a piped C<open>. Previously,
1409quotes on the verb were passed through to DCL, which would fail to recognize
1410the command. Also, if the verb is actually a path to an image or command
1411procedure on an ODS-5 volume, quoting it now allows the path to contain spaces.
1412
1413=item *
1414
1415The B<a2p> build has been fixed for the HP C++ compiler on OpenVMS.
1416
1417=back
1418
1419=head3 Win32
1420
1421=over
1422
1423=item *
1424
1425Perl can now be built using Microsoft's Visual C++ 2012 compiler by specifying
1426CCTYPE=MSVC110 (or MSVC110FREE if you are using the free Express edition for
1427Windows Desktop) in F<win32/Makefile>.
1428
1429=item *
1430
f105b7be 1431The option to build without C<USE_SOCKETS_AS_HANDLES> has been removed.
5ed58cbd
RS
1432
1433=item *
1434
1435Fixed a problem where perl could crash while cleaning up threads (including the
1436main thread) in threaded debugging builds on Win32 and possibly other platforms
1437[perl #114496].
1438
1439=item *
1440
1441A rare race condition that would lead to L<sleep|perlfunc/sleep> taking more
1442time than requested, and possibly even hanging, has been fixed [perl #33096].
1443
1444=item *
1445
1446C<link> on Win32 now attempts to set C<$!> to more appropriate values
1447based on the Win32 API error code. [perl #112272]
1448
1449Perl no longer mangles the environment block, e.g. when launching a new
1450sub-process, when the environment contains non-ASCII characters. Known
1451problems still remain, however, when the environment contains characters
1452outside of the current ANSI codepage (e.g. see the item about Unicode in
1453C<%ENV> in L<http://perl5.git.perl.org/perl.git/blob/HEAD:/Porting/todo.pod>).
1454[perl #113536]
1455
1456=item *
1457
1458Building perl with some Windows compilers used to fail due to a problem
1459with miniperl's C<glob> operator (which uses the C<perlglob> program)
1460deleting the PATH environment variable [perl #113798].
1461
1462=item *
1463
f105b7be 1464A new makefile option, C<USE_64_BIT_INT>, has been added to the Windows
5ed58cbd
RS
1465makefiles. Set this to "define" when building a 32-bit perl if you want
1466it to use 64-bit integers.
1467
1468Machine code size reductions, already made to the DLLs of XS modules in
1469Perl 5.17.2, have now been extended to the perl DLL itself.
1470
1471Building with VC++ 6.0 was inadvertently broken in Perl 5.17.2 but has
1472now been fixed again.
1473
1474=back
1475
1476=head3 WinCE
1477
1478Building on WinCE is now possible once again, although more work is required
1479to fully restore a clean build.
1480
1481=head1 Internal Changes
1482
1483=over
1484
1485=item *
1486
4263dd11 1487Synonyms for the misleadingly named C<av_len()> have been created:
5ed58cbd
RS
1488C<av_top_index()> and C<av_tindex>. All three of these return the
1489number of the highest index in the array, not the number of elements it
1490contains.
1491
1492=item *
1493
1494SvUPGRADE() is no longer an expression. Originally this macro (and its
1495underlying function, sv_upgrade()) were documented as boolean, although
1496in reality they always croaked on error and never returned false. In 2005
1497the documentation was updated to specify a void return value, but
1498SvUPGRADE() was left always returning 1 for backwards compatibility. This
1499has now been removed, and SvUPGRADE() is now a statement with no return
1500value.
1501
1502So this is now a syntax error:
1503
1504 if (!SvUPGRADE(sv)) { croak(...); }
1505
1506If you have code like that, simply replace it with
1507
1508 SvUPGRADE(sv);
1509
1510or to to avoid compiler warnings with older perls, possibly
1511
1512 (void)SvUPGRADE(sv);
1513
1514=item *
1515
1516Perl has a new copy-on-write mechanism that allows any SvPOK scalar to be
1517upgraded to a copy-on-write scalar. A reference count on the string buffer
1518is stored in the string buffer itself.
1519
1520This breaks a few XS modules by allowing copy-on-write scalars to go
1521through code paths that never encountered them before.
1522
1523This behaviour can still be disabled by running F<Configure> with
1524B<-Accflags=-DPERL_NO_COW>. This option will probably be removed in Perl
15255.20.
1526
1527=item *
1528
1529Copy-on-write no longer uses the SvFAKE and SvREADONLY flags. Hence,
1530SvREADONLY indicates a true read-only SV.
1531
1532Use the SvIsCOW macro (as before) to identify a copy-on-write scalar.
1533
1534=item *
1535
f105b7be 1536C<PL_glob_index> is gone.
5ed58cbd
RS
1537
1538=item *
1539
1540The private Perl_croak_no_modify has had its context parameter removed. It is
1541now has a void prototype. Users of the public API croak_no_modify remain
1542unaffected.
1543
1544=item *
1545
1546Copy-on-write (shared hash key) scalars are no longer marked read-only.
1547C<SvREADONLY> returns false on such an SV, but C<SvIsCOW> still returns
1548true.
1549
1550=item *
1551
1552A new op type, C<OP_PADRANGE> has been introduced. The perl peephole
1553optimiser will, where possible, substitute a single padrange op for a
1554pushmark followed by one or more pad ops, and possibly also skipping list
1555and nextstate ops. In addition, the op can carry out the tasks associated
f105b7be 1556with the RHS of a C<< my(...) = @_ >> assignment, so those ops may be optimised
5ed58cbd
RS
1557away too.
1558
1559=item *
1560
1561Case-insensitive matching inside a [bracketed] character class with a
1562multi-character fold no longer excludes one of the possibilities in the
1563circumstances that it used to. [perl #89774].
1564
1565=item *
1566
1567C<PL_formfeed> has been removed.
1568
1569=item *
1570
1571The regular expression engine no longer reads one byte past the end of the
1572target string. While for all internally well-formed scalars this should
1573never have been a problem, this change facilitates clever tricks with
1574string buffers in CPAN modules. [perl #73542]
1575
1576=item *
1577
1578Inside a BEGIN block, C<PL_compcv> now points to the currently-compiling
1579subroutine, rather than the BEGIN block itself.
1580
1581=item *
1582
1583C<mg_length> has been deprecated.
1584
1585=item *
1586
1587C<sv_len> now always returns a byte count and C<sv_len_utf8> a character
1588count. Previously, C<sv_len> and C<sv_len_utf8> were both buggy and would
1589sometimes returns bytes and sometimes characters. C<sv_len_utf8> no longer
be12dd22 1590assumes that its argument is in UTF-8. Neither of these creates UTF-8 caches
5ed58cbd
RS
1591for tied or overloaded values or for non-PVs any more.
1592
1593=item *
1594
1595C<sv_mortalcopy> now copies string buffers of shared hash key scalars when
1596called from XS modules [perl #79824].
1597
1598=item *
1599
1600C<RXf_SPLIT> and C<RXf_SKIPWHITE> are no longer used. They are now
1601#defined as 0.
1602
1603=item *
1604
1605The new C<RXf_MODIFIES_VARS> flag can be set by custom regular expression
1606engines to indicate that the execution of the regular expression may cause
1607variables to be modified. This lets C<s///> know to skip certain
1608optimisations. Perl's own regular expression engine sets this flag for the
1609special backtracking verbs that set $REGMARK and $REGERROR.
1610
1611=item *
1612
1613The APIs for accessing lexical pads have changed considerably.
1614
1615C<PADLIST>s are now longer C<AV>s, but their own type instead.
1616C<PADLIST>s now contain a C<PAD> and a C<PADNAMELIST> of C<PADNAME>s,
1617rather than C<AV>s for the pad and the list of pad names. C<PAD>s,
1618C<PADNAMELIST>s, and C<PADNAME>s are to be accessed as such through the
1619newly added pad API instead of the plain C<AV> and C<SV> APIs. See
1620L<perlapi> for details.
1621
1622=item *
1623
1624In the regex API, the numbered capture callbacks are passed an index
1625indicating what match variable is being accessed. There are special
1626index values for the C<$`, $&, $&> variables. Previously the same three
1627values were used to retrieve C<${^PREMATCH}, ${^MATCH}, ${^POSTMATCH}>
1628too, but these have now been assigned three separate values. See
1629L<perlreapi/Numbered capture callbacks>.
1630
1631=item *
1632
1633C<PL_sawampersand> was previously a boolean indicating that any of
1634C<$`, $&, $&> had been seen; it now contains three one-bit flags
1635indicating the presence of each of the variables individually.
1636
1637=item *
1638
1639The C<CV *> typemap entry now supports C<&{}> overloading and typeglobs,
1640just like C<&{...}> [perl #96872].
1641
1642=item *
1643
1644The C<SVf_AMAGIC> flag to indicate overloading is now on the stash, not the
1645object. It is now set automatically whenever a method or @ISA changes, so
1646its meaning has changed, too. It now means "potentially overloaded". When
1647the overload table is calculated, the flag is automatically turned off if
1648there is no overloading, so there should be no noticeable slowdown.
1649
1650The staleness of the overload tables is now checked when overload methods
1651are invoked, rather than during C<bless>.
1652
1653"A" magic is gone. The changes to the handling of the C<SVf_AMAGIC> flag
1654eliminate the need for it.
1655
1656C<PL_amagic_generation> has been removed as no longer necessary. For XS
1657modules, it is now a macro alias to C<PL_na>.
1658
1659The fallback overload setting is now stored in a stash entry separate from
1660overloadedness itself.
1661
1662=item *
1663
1664The character-processing code has been cleaned up in places. The changes
1665should be operationally invisible.
1666
1667=item *
1668
1669The C<study> function was made a no-op in 5.16. It was simply disabled via
1670a C<return> statement; the code was left in place. Now the code supporting
1671what C<study> used to do has been removed.
1672
1673=item *
1674
1675Under threaded perls, there is no longer a separate PV allocated for every
1676COP to store its package name (C<< cop->stashpv >>). Instead, there is an
1677offset (C<< cop->stashoff >>) into the new C<PL_stashpad> array, which
1678holds stash pointers.
1679
1680=item *
1681
1682In the pluggable regex API, the C<regexp_engine> struct has acquired a new
1683field C<op_comp>, which is currently just for perl's internal use, and
f105b7be 1684should be initialized to NULL by other regex plugin modules.
5ed58cbd
RS
1685
1686=item *
1687
7779650e 1688A new function C<alloccopstash> has been added to the API, but is considered
5ed58cbd
RS
1689experimental. See L<perlapi>.
1690
1691=item *
1692
1693Perl used to implement get magic in a way that would sometimes hide bugs in
4263dd11 1694code that could call mg_get() too many times on magical values. This hiding of
5ed58cbd
RS
1695errors no longer occurs, so long-standing bugs may become visible now. If
1696you see magic-related errors in XS code, check to make sure it, together
1697with the Perl API functions it uses, calls mg_get() only once on SvGMAGICAL()
1698values.
1699
1700=item *
1701
1702OP allocation for CVs now uses a slab allocator. This simplifies
1703memory management for OPs allocated to a CV, so cleaning up after a
1704compilation error is simpler and safer [perl #111462][perl #112312].
1705
1706=item *
1707
f105b7be 1708C<PERL_DEBUG_READONLY_OPS> has been rewritten to work with the new slab
5ed58cbd
RS
1709allocator, allowing it to catch more violations than before.
1710
1711=item *
1712
f105b7be
KE
1713The old slab allocator for ops, which was only enabled for C<PERL_IMPLICIT_SYS>
1714and C<PERL_DEBUG_READONLY_OPS>, has been retired.
5ed58cbd
RS
1715
1716=back
1717
1718=head1 Selected Bug Fixes
1719
1720=over 4
1721
1722=item *
1723
1724Here-doc terminators no longer require a terminating newline character when
1725they occur at the end of a file. This was already the case at the end of a
1726string eval [perl #65838].
1727
1728=item *
1729
f105b7be 1730C<-DPERL_GLOBAL_STRUCT> builds now free the global struct B<after>
5ed58cbd
RS
1731they've finished using it.
1732
1733=item *
1734
1735A trailing '/' on a path in @INC will no longer have an additional '/'
1736appended.
1737
1738=item *
1739
1740The C<:crlf> layer now works when unread data doesn't fit into its own
1741buffer. [perl #112244].
1742
1743=item *
1744
1745C<ungetc()> now handles UTF-8 encoded data. [perl #116322].
1746
1747=item *
1748
1749A bug in the core typemap caused any C types that map to the T_BOOL core
1750typemap entry to not be set, updated, or modified when the T_BOOL variable was
1751used in an OUTPUT: section with an exception for RETVAL. T_BOOL in an INPUT:
1752section was not affected. Using a T_BOOL return type for an XSUB (RETVAL)
1753was not affected. A side effect of fixing this bug is, if a T_BOOL is specified
1754in the OUTPUT: section (which previous did nothing to the SV), and a read only
1755SV (literal) is passed to the XSUB, croaks like "Modification of a read-only
1756value attempted" will happen. [perl #115796]
1757
1758=item *
1759
1760On many platforms, providing a directory name as the script name caused perl
1761to do nothing and report success. It should now universally report an error
1762and exit nonzero. [perl #61362]
1763
1764=item *
1765
1766C<sort {undef} ...> under fatal warnings no longer crashes. It had
1767begun crashing in Perl 5.16.
1768
1769=item *
1770
1771Stashes blessed into each other
1772(C<bless \%Foo::, 'Bar'; bless \%Bar::, 'Foo'>) no longer result in double
1773frees. This bug started happening in Perl 5.16.
1774
1775=item *
1776
1777Numerous memory leaks have been fixed, mostly involving fatal warnings and
1778syntax errors.
1779
1780=item *
1781
1782Some failed regular expression matches such as C<'f' =~ /../g> were not
1783resetting C<pos>. Also, "match-once" patterns (C<m?...?g>) failed to reset
1784it, too, when invoked a second time [perl #23180].
1785
1786=item *
1787
1788Accessing C<$&> after a pattern match now works if it had not been seen
1789before the match. I.e., this applies to C<${'&'}> (under C<no strict>) and
1790C<eval '$&'>. The same applies to C<$'> and C<$`> [perl #4289].
1791
1792=item *
1793
1794Several bugs involving C<local *ISA> and C<local *Foo::> causing stale
1795MRO caches have been fixed.
1796
1797=item *
1798
1799Defining a subroutine when its typeglob has been aliased no longer results
1800in stale method caches. This bug was introduced in Perl 5.10.
1801
1802=item *
1803
1804Localising a typeglob containing a subroutine when the typeglob's package
1805has been deleted from its parent stash no longer produces an error. This
1806bug was introduced in Perl 5.14.
1807
1808=item *
1809
1810Under some circumstances, C<local *method=...> would fail to reset method
1811caches upon scope exit.
1812
1813=item *
1814
1815C</[.foo.]/> is no longer an error, but produces a warning (as before) and
1816is treated as C</[.fo]/> [perl #115818].
1817
1818=item *
1819
1820C<goto $tied_var> now calls FETCH before deciding what type of goto
1821(subroutine or label) this is.
1822
1823=item *
1824
1825Renaming packages through glob assignment
1826(C<*Foo:: = *Bar::; *Bar:: = *Baz::>) in combination with C<m?...?> and
1827C<reset> no longer makes threaded builds crash.
1828
1829=item *
1830
1831A number of bugs related to assigning a list to hash have been fixed. Many of
1832these involve lists with repeated keys like C<(1, 1, 1, 1)>.
1833
1834=over 4
1835
1836=item *
1837
1838The expression C<scalar(%h = (1, 1, 1, 1))> now returns C<4>, not C<2>.
1839
1840=item *
1841
1842The return value of C<%h = (1, 1, 1)> in list context was wrong. Previously
1843this would return C<(1, undef, 1)>, now it returns C<(1, undef)>.
1844
1845=item *
1846
1847Perl now issues the same warning on C<($s, %h) = (1, {})> as it does for
1848C<(%h) = ({})>, "Reference found where even-sized list expected".
1849
1850=item *
1851
1852A number of additional edge cases in list assignment to hashes were
1853corrected. For more details see commit 23b7025ebc.
1854
1855=back
1856
1857=item *
1858
1859Attributes applied to lexical variables no longer leak memory.
1860[perl #114764]
1861
1862=item *
1863
1864C<dump>, C<goto>, C<last>, C<next>, C<redo> or C<require> followed by a
1865bareword (or version) and then an infix operator is no longer a syntax
1866error. It used to be for those infix operators (like C<+>) that have a
1867different meaning where a term is expected. [perl #105924]
1868
1869=item *
1870
1871C<require a::b . 1> and C<require a::b + 1> no longer produce erroneous
1872ambiguity warnings. [perl #107002]
1873
1874=item *
1875
1876Class method calls are now allowed on any string, and not just strings
1877beginning with an alphanumeric character. [perl #105922]
1878
1879=item *
1880
1881An empty pattern created with C<qr//> used in C<m///> no longer triggers
1882the "empty pattern reuses last pattern" behaviour. [perl #96230]
1883
1884=item *
1885
1886Tying a hash during iteration no longer results in a memory leak.
1887
1888=item *
1889
1890Freeing a tied hash during iteration no longer results in a memory leak.
1891
1892=item *
1893
1894List assignment to a tied array or hash that dies on STORE no longer
1895results in a memory leak.
1896
1897=item *
1898
1899If the hint hash (C<%^H>) is tied, compile-time scope entry (which copies
1900the hint hash) no longer leaks memory if FETCH dies. [perl #107000]
1901
1902=item *
1903
1904Constant folding no longer inappropriately triggers the special
1905C<split " "> behaviour. [perl #94490]
1906
1907=item *
1908
1909C<defined scalar(@array)>, C<defined do { &foo }>, and similar constructs
1910now treat the argument to C<defined> as a simple scalar. [perl #97466]
1911
1912=item *
1913
1914Running a custom debugging that defines no C<*DB::DB> glob or provides a
1915subroutine stub for C<&DB::DB> no longer results in a crash, but an error
1916instead. [perl #114990]
1917
1918=item *
1919
1920C<reset ""> now matches its documentation. C<reset> only resets C<m?...?>
1921patterns when called with no argument. An empty string for an argument now
1922does nothing. (It used to be treated as no argument.) [perl #97958]
1923
1924=item *
1925
1926C<printf> with an argument returning an empty list no longer reads past the
1927end of the stack, resulting in erratic behaviour. [perl #77094]
1928
1929=item *
1930
1931C<--subname> no longer produces erroneous ambiguity warnings.
1932[perl #77240]
1933
1934=item *
1935
1936C<v10> is now allowed as a label or package name. This was inadvertently
1937broken when v-strings were added in Perl 5.6. [perl #56880]
1938
1939=item *
1940
1941C<length>, C<pos>, C<substr> and C<sprintf> could be confused by ties,
1942overloading, references and typeglobs if the stringification of such
be12dd22 1943changed the internal representation to or from UTF-8. [perl #114410]
5ed58cbd
RS
1944
1945=item *
1946
1947utf8::encode now calls FETCH and STORE on tied variables. utf8::decode now
1948calls STORE (it was already calling FETCH).
1949
1950=item *
1951
1952C<$tied =~ s/$non_utf8/$utf8/> no longer loops infinitely if the tied
1953variable returns a Latin-1 string, shared hash key scalar, or reference or
2ae351f8 1954typeglob that stringifies as ASCII or Latin-1. This was a regression from
5ed58cbd
RS
19555.12.x.
1956
1957=item *
1958
1959C<s///> without /e is now better at detecting when it needs to forego
1960certain optimisations, fixing some buggy cases:
1961
1962=over
1963
1964=item *
1965
1966Match variables in certain constructs (C<&&>, C<||>, C<..> and others) in
1967the replacement part; e.g., C<s/(.)/$l{$a||$1}/g>. [perl #26986]
1968
1969=item *
1970
1971Aliases to match variables in the replacement.
1972
1973=item *
1974
1975C<$REGERROR> or C<$REGMARK> in the replacement. [perl #49190]
1976
1977=item *
1978
1979An empty pattern (C<s//$foo/>) that causes the last-successful pattern to
1980be used, when that pattern contains code blocks that modify the variables
1981in the replacement.
1982
1983=back
1984
1985=item *
1986
1987The taintedness of the replacement string no longer affects the taintedness
1988of the return value of C<s///e>.
1989
1990=item *
1991
1992The C<$|> autoflush variable is created on-the-fly when needed. If this
1993happened (e.g., if it was mentioned in a module or eval) when the
1994currently-selected filehandle was a typeglob with an empty IO slot, it used
1995to crash. [perl #115206]
1996
1997=item *
1998
1999Line numbers at the end of a string eval are no longer off by one.
2000[perl #114658]
2001
2002=item *
2003
2004@INC filters (subroutines returned by subroutines in @INC) that set $_ to a
2005copy-on-write scalar no longer cause the parser to modify that string
2006buffer in place.
2007
2008=item *
2009
2010C<length($object)> no longer returns the undefined value if the object has
2011string overloading that returns undef. [perl #115260]
2012
2013=item *
2014
2015The use of C<PL_stashcache>, the stash name lookup cache for method calls, has
2016been restored,
2017
2018Commit da6b625f78f5f133 in August 2011 inadvertently broke the code that looks
2019up values in C<PL_stashcache>. As it's a only cache, quite correctly everything
2020carried on working without it.
2021
2022=item *
2023
2024The error "Can't localize through a reference" had disappeared in 5.16.0
2025when C<local %$ref> appeared on the last line of an lvalue subroutine.
2026This error disappeared for C<\local %$ref> in perl 5.8.1. It has now
2027been restored.
2028
2029=item *
2030
2031The parsing of here-docs has been improved significantly, fixing several
2032parsing bugs and crashes and one memory leak, and correcting wrong
2033subsequent line numbers under certain conditions.
2034
2035=item *
2036
2037Inside an eval, the error message for an unterminated here-doc no longer
2038has a newline in the middle of it [perl #70836].
2039
2040=item *
2041
2042A substitution inside a substitution pattern (C<s/${s|||}//>) no longer
2043confuses the parser.
2044
2045=item *
2046
2047It may be an odd place to allow comments, but C<s//"" # hello/e> has
2048always worked, I<unless> there happens to be a null character before the
2049first #. Now it works even in the presence of nulls.
2050
2051=item *
2052
2053An invalid range in C<tr///> or C<y///> no longer results in a memory leak.
2054
2055=item *
2056
2057String eval no longer treats a semicolon-delimited quote-like operator at
2058the very end (C<eval 'q;;'>) as a syntax error.
2059
2060=item *
2061
2062C<< warn {$_ => 1} + 1 >> is no longer a syntax error. The parser used to
2063get confused with certain list operators followed by an anonymous hash and
2064then an infix operator that shares its form with a unary operator.
2065
2066=item *
2067
2068C<(caller $n)[6]> (which gives the text of the eval) used to return the
2069actual parser buffer. Modifying it could result in crashes. Now it always
2070returns a copy. The string returned no longer has "\n;" tacked on to the
2071end. The returned text also includes here-doc bodies, which used to be
2072omitted.
2073
2074=item *
2075
be12dd22
RS
2076Reset the UTF-8 position cache when accessing magical variables to avoid the
2077string buffer and the UTF-8 position cache getting out of sync
5ed58cbd
RS
2078[perl #114410].
2079
2080=item *
2081
be12dd22
RS
2082Various cases of get magic being called twice for magical UTF-8
2083strings have been fixed.
5ed58cbd
RS
2084
2085=item *
2086
2087This code (when not in the presence of C<$&> etc)
2088
2089 $_ = 'x' x 1_000_000;
2090 1 while /(.)/;
2091
2092used to skip the buffer copy for performance reasons, but suffered from C<$1>
2093etc changing if the original string changed. That's now been fixed.
2094
2095=item *
2096
2097Perl doesn't use PerlIO anymore to report out of memory messages, as PerlIO
2098might attempt to allocate more memory.
2099
2100=item *
2101
2102In a regular expression, if something is quantified with C<{n,m}> where
2103C<S<n E<gt> m>>, it can't possibly match. Previously this was a fatal
2104error, but now is merely a warning (and that something won't match).
2105[perl #82954].
2106
2107=item *
2108
2109It used to be possible for formats defined in subroutines that have
2110subsequently been undefined and redefined to close over variables in the
2111wrong pad (the newly-defined enclosing sub), resulting in crashes or
2112"Bizarre copy" errors.
2113
2114=item *
2115
2116Redefinition of XSUBs at run time could produce warnings with the wrong
2117line number.
2118
2119=item *
2120
2121The %vd sprintf format does not support version objects for alpha versions.
2122It used to output the format itself (%vd) when passed an alpha version, and
2123also emit an "Invalid conversion in printf" warning. It no longer does,
2124but produces the empty string in the output. It also no longer leaks
2125memory in this case.
2126
2127=item *
2128
2129C<< $obj->SUPER::method >> calls in the main package could fail if the
2130SUPER package had already been accessed by other means.
2131
2132=item *
2133
f105b7be 2134Stash aliasing (C<< *foo:: = *bar:: >>) no longer causes SUPER calls to ignore
5ed58cbd
RS
2135changes to methods or @ISA or use the wrong package.
2136
2137=item *
2138
2139Method calls on packages whose names end in ::SUPER are no longer treated
2140as SUPER method calls, resulting in failure to find the method.
2141Furthermore, defining subroutines in such packages no longer causes them to
2142be found by SUPER method calls on the containing package [perl #114924].
2143
2144=item *
2145
2146C<\w> now matches the code points U+200C (ZERO WIDTH NON-JOINER) and U+200D
2147(ZERO WIDTH JOINER). C<\W> no longer matches these. This change is because
2148Unicode corrected their definition of what C<\w> should match.
2149
2150=item *
2151
2152C<dump LABEL> no longer leaks its label.
2153
2154=item *
2155
2156Constant folding no longer changes the behaviour of functions like C<stat()>
2157and C<truncate()> that can take either filenames or handles.
2158C<stat 1 ? foo : bar> nows treats its argument as a file name (since it is an
2159arbitrary expression), rather than the handle "foo".
2160
2161=item *
2162
2163C<truncate FOO, $len> no longer falls back to treating "FOO" as a file name if
2164the filehandle has been deleted. This was broken in Perl 5.16.0.
2165
2166=item *
2167
2168Subroutine redefinitions after sub-to-glob and glob-to-glob assignments no
2169longer cause double frees or panic messages.
2170
2171=item *
2172
2173C<s///> now turns vstrings into plain strings when performing a substitution,
2174even if the resulting string is the same (C<s/a/a/>).
2175
2176=item *
2177
2178Prototype mismatch warnings no longer erroneously treat constant subs as having
2179no prototype when they actually have "".
2180
2181=item *
2182
2183Constant subroutines and forward declarations no longer prevent prototype
2184mismatch warnings from omitting the sub name.
2185
2186=item *
2187
2188C<undef> on a subroutine now clears call checkers.
2189
2190=item *
2191
2192The C<ref> operator started leaking memory on blessed objects in Perl 5.16.0.
2193This has been fixed [perl #114340].
2194
2195=item *
2196
2197C<use> no longer tries to parse its arguments as a statement, making
2198C<use constant { () };> a syntax error [perl #114222].
2199
2200=item *
2201
2202On debugging builds, "uninitialized" warnings inside formats no longer cause
2203assertion failures.
2204
2205=item *
2206
2207On debugging builds, subroutines nested inside formats no longer cause
2208assertion failures [perl #78550].
2209
2210=item *
2211
2212Formats and C<use> statements are now permitted inside formats.
2213
2214=item *
2215
2216C<print $x> and C<sub { print $x }-E<gt>()> now always produce the same output.
2217It was possible for the latter to refuse to close over $x if the variable was
2218not active; e.g., if it was defined outside a currently-running named
2219subroutine.
2220
2221=item *
2222
2223Similarly, C<print $x> and C<print eval '$x'> now produce the same output.
2224This also allows "my $x if 0" variables to be seen in the debugger [perl
2225#114018].
2226
2227=item *
2228
2229Formats called recursively no longer stomp on their own lexical variables, but
2230each recursive call has its own set of lexicals.
2231
2232=item *
2233
2234Attempting to free an active format or the handle associated with it no longer
2235results in a crash.
2236
2237=item *
2238
2239Format parsing no longer gets confused by braces, semicolons and low-precedence
2240operators. It used to be possible to use braces as format delimiters (instead
2241of C<=> and C<.>), but only sometimes. Semicolons and low-precedence operators
2242in format argument lines no longer confuse the parser into ignoring the line's
2243return value. In format argument lines, braces can now be used for anonymous
2244hashes, instead of being treated always as C<do> blocks.
2245
2246=item *
2247
2248Formats can now be nested inside code blocks in regular expressions and other
2249quoted constructs (C</(?{...})/> and C<qq/${...}/>) [perl #114040].
2250
2251=item *
2252
2253Formats are no longer created after compilation errors.
2254
2255=item *
2256
2257Under debugging builds, the B<-DA> command line option started crashing in Perl
22585.16.0. It has been fixed [perl #114368].
2259
2260=item *
2261
2262A potential deadlock scenario involving the premature termination of a pseudo-
2263forked child in a Windows build with ithreads enabled has been fixed. This
2264resolves the common problem of the F<t/op/fork.t> test hanging on Windows [perl
2265#88840].
2266
2267=item *
2268
5ed58cbd
RS
2269The code which generates errors from C<require()> could potentially read one or
2270two bytes before the start of the filename for filenames less than three bytes
2271long and ending C</\.p?\z/>. This has now been fixed. Note that it could
2272never have happened with module names given to C<use()> or C<require()> anyway.
2273
2274=item *
2275
2276The handling of pathnames of modules given to C<require()> has been made
2277thread-safe on VMS.
2278
2279=item *
2280
d85cd26b
RS
2281Non-blocking sockets have been fixed on VMS.
2282
2283=item *
2284
5ed58cbd
RS
2285A bug in the compilation of a C</(?{})/> expression which affected the TryCatch
2286test suite has been fixed [perl #114242].
2287
2288=item *
2289
2290Pod can now be nested in code inside a quoted construct outside of a string
2291eval. This used to work only within string evals [perl #114040].
2292
2293=item *
2294
2295C<goto ''> now looks for an empty label, producing the "goto must have
2296label" error message, instead of exiting the program [perl #111794].
2297
2298=item *
2299
2300C<goto "\0"> now dies with "Can't find label" instead of "goto must have
2301label".
2302
2303=item *
2304
2305The C function C<hv_store> used to result in crashes when used on C<%^H>
2306[perl #111000].
2307
2308=item *
2309
2310A call checker attached to a closure prototype via C<cv_set_call_checker>
2311is now copied to closures cloned from it. So C<cv_set_call_checker> now
2312works inside an attribute handler for a closure.
2313
2314=item *
2315
2316Writing to C<$^N> used to have no effect. Now it croaks with "Modification
2317of a read-only value" by default, but that can be overridden by a custom
2318regular expression engine, as with C<$1> [perl #112184].
2319
2320=item *
2321
2322C<undef> on a control character glob (C<undef *^H>) no longer emits an
2323erroneous warning about ambiguity [perl #112456].
2324
2325=item *
2326
2327For efficiency's sake, many operators and built-in functions return the
2328same scalar each time. Lvalue subroutines and subroutines in the CORE::
2329namespace were allowing this implementation detail to leak through.
2330C<print &CORE::uc("a"), &CORE::uc("b")> used to print "BB". The same thing
2331would happen with an lvalue subroutine returning the return value of C<uc>.
2332Now the value is copied in such cases.
2333
2334=item *
2335
2336C<method {}> syntax with an empty block or a block returning an empty list
2337used to crash or use some random value left on the stack as its invocant.
2338Now it produces an error.
2339
2340=item *
2341
2342C<vec> now works with extremely large offsets (E<gt>2 GB) [perl #111730].
2343
2344=item *
2345
2346Changes to overload settings now take effect immediately, as do changes to
2347inheritance that affect overloading. They used to take effect only after
2348C<bless>.
2349
2350Objects that were created before a class had any overloading used to remain
2351non-overloaded even if the class gained overloading through C<use overload>
2352or @ISA changes, and even after C<bless>. This has been fixed
2353[perl #112708].
2354
2355=item *
2356
2357Classes with overloading can now inherit fallback values.
2358
2359=item *
2360
2361Overloading was not respecting a fallback value of 0 if there were
2362overloaded objects on both sides of an assignment operator like C<+=>
2363[perl #111856].
2364
2365=item *
2366
2367C<pos> now croaks with hash and array arguments, instead of producing
2368erroneous warnings.
2369
2370=item *
2371
2372C<while(each %h)> now implies C<while(defined($_ = each %h))>, like
2373C<readline> and C<readdir>.
2374
2375=item *
2376
2377Subs in the CORE:: namespace no longer crash after C<undef *_> when called
2378with no argument list (C<&CORE::time> with no parentheses).
2379
2380=item *
2381
2382C<unpack> no longer produces the "'/' must follow a numeric type in unpack"
2383error when it is the data that are at fault [perl #60204].
2384
2385=item *
2386
2387C<join> and C<"@array"> now call FETCH only once on a tied C<$">
2388[perl #8931].
2389
2390=item *
2391
2392Some subroutine calls generated by compiling core ops affected by a
2393C<CORE::GLOBAL> override had op checking performed twice. The checking
2394is always idempotent for pure Perl code, but the double checking can
2395matter when custom call checkers are involved.
2396
2397=item *
2398
2399A race condition used to exist around fork that could cause a signal sent to
2400the parent to be handled by both parent and child. Signals are now blocked
2401briefly around fork to prevent this from happening [perl #82580].
2402
2403=item *
2404
2405The implementation of code blocks in regular expressions, such as C<(?{})>
2406and C<(??{})>, has been heavily reworked to eliminate a whole slew of bugs.
2407The main user-visible changes are:
2408
2409=over 4
2410
2411=item *
2412
2413Code blocks within patterns are now parsed in the same pass as the
2414surrounding code; in particular it is no longer necessary to have balanced
2415braces: this now works:
2416
2417 /(?{ $x='{' })/
2418
2419This means that this error message is no longer generated:
2420
2421 Sequence (?{...}) not terminated or not {}-balanced in regex
2422
2423but a new error may be seen:
2424
2425 Sequence (?{...}) not terminated with ')'
2426
2427In addition, literal code blocks within run-time patterns are only
2428compiled once, at perl compile-time:
2429
2430 for my $p (...) {
2431 # this 'FOO' block of code is compiled once,
2432 # at the same time as the surrounding 'for' loop
2433 /$p{(?{FOO;})/;
2434 }
2435
2436=item *
2437
2438Lexical variables are now sane as regards scope, recursion and closure
2439behavior. In particular, C</A(?{B})C/> behaves (from a closure viewpoint)
2440exactly like C</A/ && do { B } && /C/>, while C<qr/A(?{B})C/> is like
2441C<sub {/A/ && do { B } && /C/}>. So this code now works how you might
2442expect, creating three regexes that match 0, 1, and 2:
2443
2444 for my $i (0..2) {
2445 push @r, qr/^(??{$i})$/;
2446 }
2447 "1" =~ $r[1]; # matches
2448
2449=item *
2450
2451The C<use re 'eval'> pragma is now only required for code blocks defined
2452at runtime; in particular in the following, the text of the C<$r> pattern is
2453still interpolated into the new pattern and recompiled, but the individual
2454compiled code-blocks within C<$r> are reused rather than being recompiled,
2455and C<use re 'eval'> isn't needed any more:
2456
2457 my $r = qr/abc(?{....})def/;
2458 /xyz$r/;
2459
2460=item *
2461
2462Flow control operators no longer crash. Each code block runs in a new
2463dynamic scope, so C<next> etc. will not see
2464any enclosing loops. C<return> returns a value
2465from the code block, not from any enclosing subroutine.
2466
2467=item *
2468
2469Perl normally caches the compilation of run-time patterns, and doesn't
2470recompile if the pattern hasn't changed, but this is now disabled if
2471required for the correct behavior of closures. For example:
2472
2473 my $code = '(??{$x})';
2474 for my $x (1..3) {
2475 # recompile to see fresh value of $x each time
2476 $x =~ /$code/;
2477 }
2478
2479=item *
2480
2481The C</msix> and C<(?msix)> etc. flags are now propagated into the return
2482value from C<(??{})>; this now works:
2483
2484 "AB" =~ /a(??{'b'})/i;
2485
2486=item *
2487
2488Warnings and errors will appear to come from the surrounding code (or for
2489run-time code blocks, from an eval) rather than from an C<re_eval>:
2490
2491 use re 'eval'; $c = '(?{ warn "foo" })'; /$c/;
2492 /(?{ warn "foo" })/;
2493
2494formerly gave:
2495
2496 foo at (re_eval 1) line 1.
2497 foo at (re_eval 2) line 1.
2498
2499and now gives:
2500
2501 foo at (eval 1) line 1.
2502 foo at /some/prog line 2.
2503
2504=back
2505
2506=item *
2507
2e7bc647
KW
2508Perl now can be recompiled to use any Unicode version. In v5.16, it
2509worked on Unicodes 6.0 and 6.1, but there were various bugs if earlier
2510releases were used; the older the release the more problems.
5ed58cbd
RS
2511
2512=item *
2513
2514C<vec> no longer produces "uninitialized" warnings in lvalue context
2515[perl #9423].
2516
2517=item *
2518
2519An optimization involving fixed strings in regular expressions could cause
2520a severe performance penalty in edge cases. This has been fixed
2521[perl #76546].
2522
2523=item *
2524
2525In certain cases, including empty subpatterns within a regular expression (such
2526as C<(?:)> or C<(?:|)>) could disable some optimizations. This has been fixed.
2527
2528=item *
2529
2530The "Can't find an opnumber" message that C<prototype> produces when passed
2531a string like "CORE::nonexistent_keyword" now passes UTF-8 and embedded
2532NULs through unchanged [perl #97478].
2533
2534=item *
2535
2536C<prototype> now treats magical variables like C<$1> the same way as
2537non-magical variables when checking for the CORE:: prefix, instead of
2538treating them as subroutine names.
2539
2540=item *
2541
2542Under threaded perls, a runtime code block in a regular expression could
2543corrupt the package name stored in the op tree, resulting in bad reads
2544in C<caller>, and possibly crashes [perl #113060].
2545
2546=item *
2547
2548Referencing a closure prototype (C<\&{$_[1]}> in an attribute handler for a
2549closure) no longer results in a copy of the subroutine (or assertion
2550failures on debugging builds).
2551
2552=item *
2553
2554C<eval '__PACKAGE__'> now returns the right answer on threaded builds if
2555the current package has been assigned over (as in
2556C<*ThisPackage:: = *ThatPackage::>) [perl #78742].
2557
2558=item *
2559
2560If a package is deleted by code that it calls, it is possible for C<caller>
2561to see a stack frame belonging to that deleted package. C<caller> could
2562crash if the stash's memory address was reused for a scalar and a
2563substitution was performed on the same scalar [perl #113486].
2564
2565=item *
2566
2567C<UNIVERSAL::can> no longer treats its first argument differently
2568depending on whether it is a string or number internally.
2569
2570=item *
2571
2572C<open> with C<< <& >> for the mode checks to see whether the third argument is
2573a number, in determining whether to treat it as a file descriptor or a handle
2574name. Magical variables like C<$1> were always failing the numeric check and
2575being treated as handle names.
2576
2577=item *
2578
2579C<warn>'s handling of magical variables (C<$1>, ties) has undergone several
2580fixes. C<FETCH> is only called once now on a tied argument or a tied C<$@>
2581[perl #97480]. Tied variables returning objects that stringify as "" are
2582no longer ignored. A tied C<$@> that happened to return a reference the
2583I<previous> time it was used is no longer ignored.
2584
2585=item *
2586
2587C<warn ""> now treats C<$@> with a number in it the same way, regardless of
2588whether it happened via C<$@=3> or C<$@="3">. It used to ignore the
2589former. Now it appends "\t...caught", as it has always done with
2590C<$@="3">.
2591
2592=item *
2593
2594Numeric operators on magical variables (e.g., S<C<$1 + 1>>) used to use
2595floating point operations even where integer operations were more appropriate,
2596resulting in loss of accuracy on 64-bit platforms [perl #109542].
2597
2598=item *
2599
2600Unary negation no longer treats a string as a number if the string happened
2601to be used as a number at some point. So, if C<$x> contains the string "dogs",
2602C<-$x> returns "-dogs" even if C<$y=0+$x> has happened at some point.
2603
2604=item *
2605
2606In Perl 5.14, C<-'-10'> was fixed to return "10", not "+10". But magical
2607variables (C<$1>, ties) were not fixed till now [perl #57706].
2608
2609=item *
2610
2611Unary negation now treats strings consistently, regardless of the internal
2612C<UTF8> flag.
2613
2614=item *
2615
2616A regression introduced in Perl v5.16.0 involving
2617C<tr/I<SEARCHLIST>/I<REPLACEMENTLIST>/> has been fixed. Only the first
2618instance is supposed to be meaningful if a character appears more than
2619once in C<I<SEARCHLIST>>. Under some circumstances, the final instance
2620was overriding all earlier ones. [perl #113584]
2621
2622=item *
2623
2624Regular expressions like C<qr/\87/> previously silently inserted a NUL
2625character, thus matching as if it had been written C<qr/\00087/>. Now it
2626matches as if it had been written as C<qr/87/>, with a message that the
2627sequence C<"\8"> is unrecognized.
2628
2629=item *
2630
2631C<__SUB__> now works in special blocks (C<BEGIN>, C<END>, etc.).
2632
2633=item *
2634
2635Thread creation on Windows could theoretically result in a crash if done
2636inside a C<BEGIN> block. It still does not work properly, but it no longer
2637crashes [perl #111610].
2638
2639=item *
2640
2641C<\&{''}> (with the empty string) now autovivifies a stub like any other
2642sub name, and no longer produces the "Unable to create sub" error
2643[perl #94476].
2644
2645=item *
2646
2647A regression introduced in v5.14.0 has been fixed, in which some calls
2648to the C<re> module would clobber C<$_> [perl #113750].
2649
2650=item *
2651
2652C<do FILE> now always either sets or clears C<$@>, even when the file can't be
2653read. This ensures that testing C<$@> first (as recommended by the
2654documentation) always returns the correct result.
2655
2656=item *
2657
2658The array iterator used for the C<each @array> construct is now correctly
2659reset when C<@array> is cleared (RT #75596). This happens for example when the
2660array is globally assigned to, as in C<@array = (...)>, but not when its
2661B<values> are assigned to. In terms of the XS API, it means that C<av_clear()>
2662will now reset the iterator.
2663
2664This mirrors the behaviour of the hash iterator when the hash is cleared.
2665
2666=item *
2667
2668C<< $class->can >>, C<< $class->isa >>, and C<< $class->DOES >> now return
2669correct results, regardless of whether that package referred to by C<$class>
2670exists [perl #47113].
2671
2672=item *
2673
2674Arriving signals no longer clear C<$@> [perl #45173].
2675
2676=item *
2677
2678Allow C<my ()> declarations with an empty variable list [perl #113554].
2679
2680=item *
2681
2682During parsing, subs declared after errors no longer leave stubs
2683[perl #113712].
2684
2685=item *
2686
2687Closures containing no string evals no longer hang on to their containing
2688subroutines, allowing variables closed over by outer subroutines to be
2689freed when the outer sub is freed, even if the inner sub still exists
2690[perl #89544].
2691
2692=item *
2693
2694Duplication of in-memory filehandles by opening with a "<&=" or ">&=" mode
2695stopped working properly in 5.16.0. It was causing the new handle to
2696reference a different scalar variable. This has been fixed [perl #113764].
2697
2698=item *
2699
2700C<qr//> expressions no longer crash with custom regular expression engines
2701that do not set C<offs> at regular expression compilation time
2702[perl #112962].
2703
2704=item *
2705
2706C<delete local> no longer crashes with certain magical arrays and hashes
2707[perl #112966].
2708
2709=item *
2710
2711C<local> on elements of certain magical arrays and hashes used not to
2712arrange to have the element deleted on scope exit, even if the element did
2713not exist before C<local>.
2714
2715=item *
2716
2717C<scalar(write)> no longer returns multiple items [perl #73690].
2718
2719=item *
2720
2721String to floating point conversions no longer misparse certain strings under
2722C<use locale> [perl #109318].
2723
2724=item *
2725
2726C<@INC> filters that die no longer leak memory [perl #92252].
2727
2728=item *
2729
2730The implementations of overloaded operations are now called in the correct
2731context. This allows, among other things, being able to properly override
2732C<< <> >> [perl #47119].
2733
2734=item *
2735
2736Specifying only the C<fallback> key when calling C<use overload> now behaves
2737properly [perl #113010].
2738
2739=item *
2740
2741C<< sub foo { my $a = 0; while ($a) { ... } } >> and
2742C<< sub foo { while (0) { ... } } >> now return the same thing [perl #73618].
2743
2744=item *
2745
2746String negation now behaves the same under C<use integer;> as it does
2747without [perl #113012].
2748
2749=item *
2750
2751C<chr> now returns the Unicode replacement character (U+FFFD) for -1,
2752regardless of the internal representation. -1 used to wrap if the argument
2753was tied or a string internally.
2754
2755=item *
2756
2757Using a C<format> after its enclosing sub was freed could crash as of
2758perl 5.12.0, if the format referenced lexical variables from the outer sub.
2759
2760=item *
2761
2762Using a C<format> after its enclosing sub was undefined could crash as of
2763perl 5.10.0, if the format referenced lexical variables from the outer sub.
2764
2765=item *
2766
2767Using a C<format> defined inside a closure, which format references
2768lexical variables from outside, never really worked unless the C<write>
2769call was directly inside the closure. In 5.10.0 it even started crashing.
2770Now the copy of that closure nearest the top of the call stack is used to
2771find those variables.
2772
2773=item *
2774
2775Formats that close over variables in special blocks no longer crash if a
2776stub exists with the same name as the special block before the special
2777block is compiled.
2778
2779=item *
2780
2781The parser no longer gets confused, treating C<eval foo ()> as a syntax
2782error if preceded by C<print;> [perl #16249].
2783
2784=item *
2785
2786The return value of C<syscall> is no longer truncated on 64-bit platforms
2787[perl #113980].
2788
2789=item *
2790
2791Constant folding no longer causes C<print 1 ? FOO : BAR> to print to the
2792FOO handle [perl #78064].
2793
2794=item *
2795
2796C<do subname> now calls the named subroutine and uses the file name it
2797returns, instead of opening a file named "subname".
2798
2799=item *
2800
2801Subroutines looked up by rv2cv check hooks (registered by XS modules) are
2802now taken into consideration when determining whether C<foo bar> should be
2803the sub call C<foo(bar)> or the method call C<< "bar"->foo >>.
2804
2805=item *
2806
2807C<CORE::foo::bar> is no longer treated specially, allowing global overrides
2808to be called directly via C<CORE::GLOBAL::uc(...)> [perl #113016].
2809
2810=item *
2811
2812Calling an undefined sub whose typeglob has been undefined now produces the
2813customary "Undefined subroutine called" error, instead of "Not a CODE
2814reference".
2815
2816=item *
2817
2818Two bugs involving @ISA have been fixed. C<*ISA = *glob_without_array> and
2819C<undef *ISA; @{*ISA}> would prevent future modifications to @ISA from
2820updating the internal caches used to look up methods. The
2821*glob_without_array case was a regression from Perl 5.12.
2822
2823=item *
2824
2825Regular expression optimisations sometimes caused C<$> with C</m> to
2826produce failed or incorrect matches [perl #114068].
2827
2828=item *
2829
2830C<__SUB__> now works in a C<sort> block when the enclosing subroutine is
2831predeclared with C<sub foo;> syntax [perl #113710].
2832
2833=item *
2834
2835Unicode properties only apply to Unicode code points, which leads to
2836some subtleties when regular expressions are matched against
2837above-Unicode code points. There is a warning generated to draw your
2838attention to this. However, this warning was being generated
2839inappropriately in some cases, such as when a program was being parsed.
2840Non-Unicode matches such as C<\w> and C<[:word;]> should not generate the
2841warning, as their definitions don't limit them to apply to only Unicode
2842code points. Now the message is only generated when matching against
2843C<\p{}> and C<\P{}>. There remains a bug, [perl #114148], for the very
2844few properties in Unicode that match just a single code point. The
2845warning is not generated if they are matched against an above-Unicode
2846code point.
2847
2848=item *
2849
2850Uninitialized warnings mentioning hash elements would only mention the
2851element name if it was not in the first bucket of the hash, due to an
2852off-by-one error.
2853
2854=item *
2855
2856A regular expression optimizer bug could cause multiline "^" to behave
2857incorrectly in the presence of line breaks, such that
2858C<"/\n\n" =~ m#\A(?:^/$)#im> would not match [perl #115242].
2859
2860=item *
2861
2862Failed C<fork> in list context no longer corrupts the stack.
2863C<@a = (1, 2, fork, 3)> used to gobble up the 2 and assign C<(1, undef, 3)>
2864if the C<fork> call failed.
2865
2866=item *
2867
2868Numerous memory leaks have been fixed, mostly involving tied variables that
2869die, regular expression character classes and code blocks, and syntax
2870errors.
2871
2872=item *
2873
2874Assigning a regular expression (C<${qr//}>) to a variable that happens to
2875hold a floating point number no longer causes assertion failures on
2876debugging builds.
2877
2878=item *
2879
2880Assigning a regular expression to a scalar containing a number no longer
f105b7be 2881causes subsequent numification to produce random numbers.
5ed58cbd
RS
2882
2883=item *
2884
2885Assigning a regular expression to a magic variable no longer wipes away the
2ae351f8 2886magic. This was a regression from 5.10.
5ed58cbd
RS
2887
2888=item *
2889
2890Assigning a regular expression to a blessed scalar no longer results in
2ae351f8 2891crashes. This was also a regression from 5.10.
5ed58cbd
RS
2892
2893=item *
2894
2895Regular expression can now be assigned to tied hash and array elements with
2896flattening into strings.
2897
2898=item *
2899
f105b7be 2900Numifying a regular expression no longer results in an uninitialized
5ed58cbd
RS
2901warning.
2902
2903=item *
2904
2905Negative array indices no longer cause EXISTS methods of tied variables to
2ae351f8 2906be ignored. This was a regression from 5.12.
5ed58cbd
RS
2907
2908=item *
2909
2910Negative array indices no longer result in crashes on arrays tied to
2911non-objects.
2912
2913=item *
2914
be12dd22
RS
2915C<$byte_overload .= $utf8> no longer results in doubly-encoded UTF-8 if the
2916left-hand scalar happened to have produced a UTF-8 string the last time
5ed58cbd
RS
2917overloading was invoked.
2918
2919=item *
2920
2921C<goto &sub> now uses the current value of @_, instead of using the array
2922the subroutine was originally called with. This means
2923C<local @_ = (...); goto &sub> now works [perl #43077].
2924
2925=item *
2926
2927If a debugger is invoked recursively, it no longer stomps on its own
2928lexical variables. Formerly under recursion all calls would share the same
2929set of lexical variables [perl #115742].
2930
2931=item *
2932
2933C<*_{ARRAY}> returned from a subroutine no longer spontaneously
2934becomes empty.
2935
2936=back
2937
2938=head1 Known Problems
2939
2940=over 4
2941
2942=item *
2943
7f50b25b 2944There are no known regressions. Please report any bugs you find!
5ed58cbd
RS
2945
2946=back
2947
2948=head1 Acknowledgements
a75569c0 2949
5ed58cbd 2950XXX Generate this with:
a75569c0 2951
5ed58cbd 2952 perl Porting/acknowledgements.pl v5.18.0..HEAD
f5b73711 2953
44691e6f
AB
2954=head1 Reporting Bugs
2955
e08634c5
SH
2956If you find what you think is a bug, you might check the articles recently
2957posted to the comp.lang.perl.misc newsgroup and the perl bug database at
2958http://rt.perl.org/perlbug/ . There may also be information at
2959http://www.perl.org/ , the Perl Home Page.
44691e6f 2960
e08634c5
SH
2961If you believe you have an unreported bug, please run the L<perlbug> program
2962included with your release. Be sure to trim your bug down to a tiny but
2963sufficient test case. Your bug report, along with the output of C<perl -V>,
2964will be sent off to perlbug@perl.org to be analysed by the Perl porting team.
44691e6f
AB
2965
2966If the bug you are reporting has security implications, which make it
e08634c5
SH
2967inappropriate to send to a publicly archived mailing list, then please send it
2968to perl5-security-report@perl.org. This points to a closed subscription
2969unarchived mailing list, which includes all the core committers, who will be
2970able to help assess the impact of issues, figure out a resolution, and help
f9001595 2971co-ordinate the release of patches to mitigate or fix the problem across all
e08634c5
SH
2972platforms on which Perl is supported. Please only use this address for
2973security issues in the Perl core, not for modules independently distributed on
2974CPAN.
44691e6f
AB
2975
2976=head1 SEE ALSO
2977
e08634c5
SH
2978The F<Changes> file for an explanation of how to view exhaustive details on
2979what changed.
44691e6f
AB
2980
2981The F<INSTALL> file for how to build Perl.
2982
2983The F<README> file for general stuff.
2984
2985The F<Artistic> and F<Copying> files for copyright information.
2986
2987=cut