1PERL5220DELTA(1) Perl Programmers Reference Guide PERL5220DELTA(1)
2
6 perl5220delta - what is new for perl v5.22.0
7
9 This document describes differences between the 5.20.0 release and the
10 5.22.0 release.
11 If you are upgrading from an earlier release such as 5.18.0, first read
13 perl5200delta, which describes differences between 5.18.0 and 5.20.0.
14
16 New bitwise operators
17 A new experimental facility has been added that makes the four standard
18 bitwise operators ("& | ^ ~") treat their operands consistently as
19 numbers, and introduces four new dotted operators ("&. |. ^. ~.") that
20 treat their operands consistently as strings. The same applies to the
21 assignment variants ("&= |= ^= &.= |.= ^.=").
22 To use this, enable the "bitwise" feature and disable the
24 "experimental::bitwise" warnings category. See "Bitwise String
25 Operators" in perlop for details. [GH #14348]
26 <https://github.com/Perl/perl5/issues/14348>.
27 New double-diamond operator
29 "<<>>" is like "<>" but uses three-argument "open" to open each file in
30 @ARGV. This means that each element of @ARGV will be treated as an
31 actual file name, and "|foo" won't be treated as a pipe open.
32 New "\b" boundaries in regular expressions
34 "qr/\b{gcb}/"
35 "gcb" stands for Grapheme Cluster Boundary. It is a Unicode property
37 that finds the boundary between sequences of characters that look like
38 a single character to a native speaker of a language. Perl has long
39 had the ability to deal with these through the "\X" regular escape
40 sequence. Now, there is an alternative way of handling these. See
41 "\b{}, \b, \B{}, \B" in perlrebackslash for details.
42 "qr/\b{wb}/"
44 "wb" stands for Word Boundary. It is a Unicode property that finds the
46 boundary between words. This is similar to the plain "\b" (without
47 braces) but is more suitable for natural language processing. It
48 knows, for example, that apostrophes can occur in the middle of words.
49 See "\b{}, \b, \B{}, \B" in perlrebackslash for details.
50 "qr/\b{sb}/"
52 "sb" stands for Sentence Boundary. It is a Unicode property to aid in
54 parsing natural language sentences. See "\b{}, \b, \B{}, \B" in
55 perlrebackslash for details.
56 Non-Capturing Regular Expression Flag
58 Regular expressions now support a "/n" flag that disables capturing and
59 filling in $1, $2, etc inside of groups:
60 "hello" =~ /(hi|hello)/n; # $1 is not set
62 This is equivalent to putting "?:" at the beginning of every capturing
64 group.
65 See "n" in perlre for more information.
67 "use re 'strict'"
69 This applies stricter syntax rules to regular expression patterns
70 compiled within its scope. This will hopefully alert you to typos and
71 other unintentional behavior that backwards-compatibility issues
72 prevent us from reporting in normal regular expression compilations.
73 Because the behavior of this is subject to change in future Perl
74 releases as we gain experience, using this pragma will raise a warning
75 of category "experimental::re_strict". See 'strict' in re.
76 Unicode 7.0 (with correction) is now supported
78 For details on what is in this release, see
79 <http://www.unicode.org/versions/Unicode7.0.0/>. The version of
80 Unicode 7.0 that comes with Perl includes a correction dealing with
81 glyph shaping in Arabic (see
82 <http://www.unicode.org/errata/#current_errata>).
83 "use locale" can restrict which locale categories are affected
85 It is now possible to pass a parameter to "use locale" to specify a
86 subset of locale categories to be locale-aware, with the remaining ones
87 unaffected. See "The "use locale" pragma" in perllocale for details.
88 Perl now supports POSIX 2008 locale currency additions
90 On platforms that are able to handle POSIX.1-2008, the hash returned by
91 "POSIX::localeconv()" includes the international currency fields added
92 by that version of the POSIX standard. These are "int_n_cs_precedes",
93 "int_n_sep_by_space", "int_n_sign_posn", "int_p_cs_precedes",
94 "int_p_sep_by_space", and "int_p_sign_posn".
95 Better heuristics on older platforms for determining locale UTF-8ness
97 On platforms that implement neither the C99 standard nor the POSIX 2001
98 standard, determining if the current locale is UTF-8 or not depends on
99 heuristics. These are improved in this release.
100 Aliasing via reference
102 Variables and subroutines can now be aliased by assigning to a
103 reference:
104 \$c = \$d;
106 \&x = \&y;
107 Aliasing can also be accomplished by using a backslash before a
109 "foreach" iterator variable; this is perhaps the most useful idiom this
110 feature provides:
111 foreach \%hash (@array_of_hash_refs) { ... }
113 This feature is experimental and must be enabled via
115 "use feature 'refaliasing'". It will warn unless the
116 "experimental::refaliasing" warnings category is disabled.
117 See "Assigning to References" in perlref
119 "prototype" with no arguments
121 "prototype()" with no arguments now infers $_. [GH #14376]
122 <https://github.com/Perl/perl5/issues/14376>.
123 New ":const" subroutine attribute
125 The "const" attribute can be applied to an anonymous subroutine. It
126 causes the new sub to be executed immediately whenever one is created
127 (i.e. when the "sub" expression is evaluated). Its value is captured
128 and used to create a new constant subroutine that is returned. This
129 feature is experimental. See "Constant Functions" in perlsub.
130 "fileno" now works on directory handles
132 When the relevant support is available in the operating system, the
133 "fileno" builtin now works on directory handles, yielding the
134 underlying file descriptor in the same way as for filehandles. On
135 operating systems without such support, "fileno" on a directory handle
136 continues to return the undefined value, as before, but also sets $! to
137 indicate that the operation is not supported.
138 Currently, this uses either a "dd_fd" member in the OS "DIR" structure,
140 or a dirfd(3) function as specified by POSIX.1-2008.
141 List form of pipe open implemented for Win32
143 The list form of pipe:
144 open my $fh, "-|", "program", @arguments;
146 is now implemented on Win32. It has the same limitations as "system
148 LIST" on Win32, since the Win32 API doesn't accept program arguments as
149 a list.
150 Assignment to list repetition
152 "(...) x ..." can now be used within a list that is assigned to, as
153 long as the left-hand side is a valid lvalue. This allows
154 "(undef,undef,$foo) = that_function()" to be written as
155 "((undef)x2, $foo) = that_function()".
156 Infinity and NaN (not-a-number) handling improved
158 Floating point values are able to hold the special values infinity,
159 negative infinity, and NaN (not-a-number). Now we more robustly
160 recognize and propagate the value in computations, and on output
161 normalize them to the strings "Inf", "-Inf", and "NaN".
162 See also the POSIX enhancements.
164 Floating point parsing has been improved
166 Parsing and printing of floating point values has been improved.
167 As a completely new feature, hexadecimal floating point literals (like
169 "0x1.23p-4") are now supported, and they can be output with
170 "printf "%a"". See "Scalar value constructors" in perldata for more
171 details.
172 Packing infinity or not-a-number into a character is now fatal
174 Before, when trying to pack infinity or not-a-number into a (signed)
175 character, Perl would warn, and assumed you tried to pack 0xFF; if you
176 gave it as an argument to "chr", "U+FFFD" was returned.
177 But now, all such actions ("pack", "chr", and "print '%c'") result in a
179 fatal error.
180 Experimental C Backtrace API
182 Perl now supports (via a C level API) retrieving the C level backtrace
183 (similar to what symbolic debuggers like gdb do).
184 The backtrace returns the stack trace of the C call frames, with the
186 symbol names (function names), the object names (like "perl"), and if
187 it can, also the source code locations (file:line).
188 The supported platforms are Linux and OS X (some *BSD might work at
190 least partly, but they have not yet been tested).
191 The feature needs to be enabled with "Configure -Dusecbacktrace".
193 See "C backtrace" in perlhacktips for more information.
195
197 Perl is now compiled with "-fstack-protector-strong" if available
198 Perl has been compiled with the anti-stack-smashing option
199 "-fstack-protector" since 5.10.1. Now Perl uses the newer variant
200 called "-fstack-protector-strong", if available.
201 The Safe module could allow outside packages to be replaced
203 Critical bugfix: outside packages could be replaced. Safe has been
204 patched to 2.38 to address this.
205 Perl is now always compiled with "-D_FORTIFY_SOURCE=2" if available
207 The 'code hardening' option called "_FORTIFY_SOURCE", available in gcc
208 4.*, is now always used for compiling Perl, if available.
209 Note that this isn't necessarily a huge step since in many platforms
211 the step had already been taken several years ago: many Linux
212 distributions (like Fedora) have been using this option for Perl, and
213 OS X has enforced the same for many years.
214
216 Subroutine signatures moved before attributes
217 The experimental sub signatures feature, as introduced in 5.20, parsed
218 signatures after attributes. In this release, following feedback from
219 users of the experimental feature, the positioning has been moved such
220 that signatures occur after the subroutine name (if any) and before the
221 attribute list (if any).
222 "&" and "\&" prototypes accepts only subs
224 The "&" prototype character now accepts only anonymous subs ("sub
225 {...}"), things beginning with "\&", or an explicit "undef". Formerly
226 it erroneously also allowed references to arrays, hashes, and lists.
227 [GH #2776] <https://github.com/Perl/perl5/issues/2776>. [GH #14186]
228 <https://github.com/Perl/perl5/issues/14186>. [GH #14353]
229 <https://github.com/Perl/perl5/issues/14353>.
230 In addition, the "\&" prototype was allowing subroutine calls, whereas
232 now it only allows subroutines: &foo is still permitted as an argument,
233 while "&foo()" and "foo()" no longer are. [GH #10633]
234 <https://github.com/Perl/perl5/issues/10633>.
235 "use encoding" is now lexical
237 The encoding pragma's effect is now limited to lexical scope. This
238 pragma is deprecated, but in the meantime, it could adversely affect
239 unrelated modules that are included in the same program; this change
240 fixes that.
241 List slices returning empty lists
243 List slices now return an empty list only if the original list was
244 empty (or if there are no indices). Formerly, a list slice would
245 return an empty list if all indices fell outside the original list; now
246 it returns a list of "undef" values in that case. [GH #12335]
247 <https://github.com/Perl/perl5/issues/12335>.
248 "\N{}" with a sequence of multiple spaces is now a fatal error
250 E.g. "\N{TOO MANY SPACES}" or "\N{TRAILING SPACE }". This has been
251 deprecated since v5.18.
252 "use UNIVERSAL '...'" is now a fatal error
254 Importing functions from "UNIVERSAL" has been deprecated since v5.12,
255 and is now a fatal error. "use UNIVERSAL" without any arguments is
256 still allowed.
257 In double-quotish "\cX", X must now be a printable ASCII character
259 In prior releases, failure to do this raised a deprecation warning.
260 Splitting the tokens "(?" and "(*" in regular expressions is now a fatal
262 compilation error.
263 These had been deprecated since v5.18.
264 "qr/foo/x" now ignores all Unicode pattern white space
266 The "/x" regular expression modifier allows the pattern to contain
267 white space and comments (both of which are ignored) for improved
268 readability. Until now, not all the white space characters that
269 Unicode designates for this purpose were handled. The additional ones
270 now recognized are:
271 U+0085 NEXT LINE
273 U+200E LEFT-TO-RIGHT MARK
274 U+200F RIGHT-TO-LEFT MARK
275 U+2028 LINE SEPARATOR
276 U+2029 PARAGRAPH SEPARATOR
277 The use of these characters with "/x" outside bracketed character
279 classes and when not preceded by a backslash has raised a deprecation
280 warning since v5.18. Now they will be ignored.
281 Comment lines within "(?[ ])" are now ended only by a "\n"
283 "(?[ ])" is an experimental feature, introduced in v5.18. It operates
284 as if "/x" is always enabled. But there was a difference: comment
285 lines (following a "#" character) were terminated by anything matching
286 "\R" which includes all vertical whitespace, such as form feeds. For
287 consistency, this is now changed to match what terminates comment lines
288 outside "(?[ ])", namely a "\n" (even if escaped), which is the same as
289 what terminates a heredoc string and formats.
290 "(?[...])" operators now follow standard Perl precedence
292 This experimental feature allows set operations in regular expression
293 patterns. Prior to this, the intersection operator had the same
294 precedence as the other binary operators. Now it has higher
295 precedence. This could lead to different outcomes than existing code
296 expects (though the documentation has always noted that this change
297 might happen, recommending fully parenthesizing the expressions). See
298 "Extended Bracketed Character Classes" in perlrecharclass.
299 Omitting "%" and "@" on hash and array names is no longer permitted
301 Really old Perl let you omit the "@" on array names and the "%" on hash
302 names in some spots. This has issued a deprecation warning since Perl
303 5.000, and is no longer permitted.
304 "$!" text is now in English outside the scope of "use locale"
306 Previously, the text, unlike almost everything else, always came out
307 based on the current underlying locale of the program. (Also affected
308 on some systems is "$^E".) For programs that are unprepared to handle
309 locale differences, this can cause garbage text to be displayed. It's
310 better to display text that is translatable via some tool than garbage
311 text which is much harder to figure out.
312 "$!" text will be returned in UTF-8 when appropriate
314 The stringification of $! and $^E will have the UTF-8 flag set when the
315 text is actually non-ASCII UTF-8. This will enable programs that are
316 set up to be locale-aware to properly output messages in the user's
317 native language. Code that needs to continue the 5.20 and earlier
318 behavior can do the stringification within the scopes of both
319 "use bytes" and "use locale ":messages"". Within these two scopes, no
320 other Perl operations will be affected by locale; only $! and $^E
321 stringification. The "bytes" pragma causes the UTF-8 flag to not be
322 set, just as in previous Perl releases. This resolves [GH #12035]
323 <https://github.com/Perl/perl5/issues/12035>.
324 Support for "?PATTERN?" without explicit operator has been removed
326 The "m?PATTERN?" construct, which allows matching a regex only once,
327 previously had an alternative form that was written directly with a
328 question mark delimiter, omitting the explicit "m" operator. This
329 usage has produced a deprecation warning since 5.14.0. It is now a
330 syntax error, so that the question mark can be available for use in new
331 operators.
332 "defined(@array)" and "defined(%hash)" are now fatal errors
334 These have been deprecated since v5.6.1 and have raised deprecation
335 warnings since v5.16.
336 Using a hash or an array as a reference are now fatal errors
338 For example, "%foo->{"bar"}" now causes a fatal compilation error.
339 These have been deprecated since before v5.8, and have raised
340 deprecation warnings since then.
341 Changes to the "*" prototype
343 The "*" character in a subroutine's prototype used to allow barewords
344 to take precedence over most, but not all, subroutine names. It was
345 never consistent and exhibited buggy behavior.
346 Now it has been changed, so subroutines always take precedence over
348 barewords, which brings it into conformity with similarly prototyped
349 built-in functions:
350 sub splat(*) { ... }
352 sub foo { ... }
353 splat(foo); # now always splat(foo())
354 splat(bar); # still splat('bar') as before
355 close(foo); # close(foo())
356 close(bar); # close('bar')
357
359 Setting "${^ENCODING}" to anything but "undef"
360 This variable allows Perl scripts to be written in an encoding other
361 than ASCII or UTF-8. However, it affects all modules globally, leading
362 to wrong answers and segmentation faults. New scripts should be
363 written in UTF-8; old scripts should be converted to UTF-8, which is
364 easily done with the piconv utility.
365 Use of non-graphic characters in single-character variable names
367 The syntax for single-character variable names is more lenient than for
368 longer variable names, allowing the one-character name to be a
369 punctuation character or even invisible (a non-graphic). Perl v5.20
370 deprecated the ASCII-range controls as such a name. Now, all non-
371 graphic characters that formerly were allowed are deprecated. The
372 practical effect of this occurs only when not under "use utf8", and
373 affects just the C1 controls (code points 0x80 through 0xFF), NO-BREAK
374 SPACE, and SOFT HYPHEN.
375 Inlining of "sub () { $var }" with observable side-effects
377 In many cases Perl makes "sub () { $var }" into an inlinable constant
378 subroutine, capturing the value of $var at the time the "sub"
379 expression is evaluated. This can break the closure behavior in those
380 cases where $var is subsequently modified, since the subroutine won't
381 return the changed value. (Note that this all only applies to anonymous
382 subroutines with an empty prototype ("sub ()").)
383 This usage is now deprecated in those cases where the variable could be
385 modified elsewhere. Perl detects those cases and emits a deprecation
386 warning. Such code will likely change in the future and stop producing
387 a constant.
388 If your variable is only modified in the place where it is declared,
390 then Perl will continue to make the sub inlinable with no warnings.
391 sub make_constant {
393 my $var = shift;
394 return sub () { $var }; # fine
395 }
396 sub make_constant_deprecated {
398 my $var;
399 $var = shift;
400 return sub () { $var }; # deprecated
401 }
402 sub make_constant_deprecated2 {
404 my $var = shift;
405 log_that_value($var); # could modify $var
406 return sub () { $var }; # deprecated
407 }
408 In the second example above, detecting that $var is assigned to only
410 once is too hard to detect. That it happens in a spot other than the
411 "my" declaration is enough for Perl to find it suspicious.
412 This deprecation warning happens only for a simple variable for the
414 body of the sub. (A "BEGIN" block or "use" statement inside the sub is
415 ignored, because it does not become part of the sub's body.) For more
416 complex cases, such as "sub () { do_something() if 0; $var }" the
417 behavior has changed such that inlining does not happen if the variable
418 is modifiable elsewhere. Such cases should be rare.
419 Use of multiple "/x" regexp modifiers
421 It is now deprecated to say something like any of the following:
422 qr/foo/xx;
424 /(?xax:foo)/;
425 use re qw(/amxx);
426 That is, now "x" should only occur once in any string of contiguous
428 regular expression pattern modifiers. We do not believe there are any
429 occurrences of this in all of CPAN. This is in preparation for a
430 future Perl release having "/xx" permit white-space for readability in
431 bracketed character classes (those enclosed in square brackets:
432 "[...]").
433 Using a NO-BREAK space in a character alias for "\N{...}" is now deprecated
435 This non-graphic character is essentially indistinguishable from a
436 regular space, and so should not be allowed. See "CUSTOM ALIASES" in
437 charnames.
438 A literal "{" should now be escaped in a pattern
440 If you want a literal left curly bracket (also called a left brace) in
441 a regular expression pattern, you should now escape it by either
442 preceding it with a backslash ("\{") or enclosing it within square
443 brackets "[{]", or by using "\Q"; otherwise a deprecation warning will
444 be raised. This was first announced as forthcoming in the v5.16
445 release; it will allow future extensions to the language to happen.
446 Making all warnings fatal is discouraged
448 The documentation for fatal warnings notes that "use warnings FATAL =>
449 'all'" is discouraged, and provides stronger language about the risks
450 of fatal warnings in general.
451
453 • If a method or class name is known at compile time, a hash is
454 precomputed to speed up run-time method lookup. Also, compound
455 method names like "SUPER::new" are parsed at compile time, to save
456 having to parse them at run time.
457 • Array and hash lookups (especially nested ones) that use only
459 constants or simple variables as keys, are now considerably faster.
460 See "Internal Changes" for more details.
461 • "(...)x1", "("constant")x0" and "($scalar)x0" are now optimised in
463 list context. If the right-hand argument is a constant 1, the
464 repetition operator disappears. If the right-hand argument is a
465 constant 0, the whole expression is optimised to the empty list, so
466 long as the left-hand argument is a simple scalar or constant.
467 (That is, "(foo())x0" is not subject to this optimisation.)
468 • "substr" assignment is now optimised into 4-argument "substr" at
470 the end of a subroutine (or as the argument to "return").
471 Previously, this optimisation only happened in void context.
472 • In "\L...", "\Q...", etc., the extra "stringify" op is now
474 optimised away, making these just as fast as "lcfirst",
475 "quotemeta", etc.
476 • Assignment to an empty list is now sometimes faster. In
478 particular, it never calls "FETCH" on tied arguments on the right-
479 hand side, whereas it used to sometimes.
480 • There is a performance improvement of up to 20% when "length" is
482 applied to a non-magical, non-tied string, and either "use bytes"
483 is in scope or the string doesn't use UTF-8 internally.
484 • On most perl builds with 64-bit integers, memory usage for non-
486 magical, non-tied scalars containing only a floating point value
487 has been reduced by between 8 and 32 bytes, depending on OS.
488 • In "@array = split", the assignment can be optimized away, so that
490 "split" writes directly to the array. This optimisation was
491 happening only for package arrays other than @_, and only
492 sometimes. Now this optimisation happens almost all the time.
493 • "join" is now subject to constant folding. So for example
495 "join "-", "a", "b"" is converted at compile-time to "a-b".
496 Moreover, "join" with a scalar or constant for the separator and a
497 single-item list to join is simplified to a stringification, and
498 the separator doesn't even get evaluated.
499 • "qq(@array)" is implemented using two ops: a stringify op and a
501 join op. If the "qq" contains nothing but a single array, the
502 stringification is optimized away.
503 • "our $var" and "our($s,@a,%h)" in void context are no longer
505 evaluated at run time. Even a whole sequence of "our $foo;"
506 statements will simply be skipped over. The same applies to
507 "state" variables.
508 • Many internal functions have been refactored to improve performance
510 and reduce their memory footprints. [GH #13659]
511 <https://github.com/Perl/perl5/issues/13659> [GH #13856]
512 <https://github.com/Perl/perl5/issues/13856> [GH #13874]
513 <https://github.com/Perl/perl5/issues/13874>
514 • "-T" and "-B" filetests will return sooner when an empty file is
516 detected. [GH #13686] <https://github.com/Perl/perl5/issues/13686>
517 • Hash lookups where the key is a constant are faster.
519 • Subroutines with an empty prototype and a body containing just
521 "undef" are now eligible for inlining. [GH #14077]
522 <https://github.com/Perl/perl5/issues/14077>
523 • Subroutines in packages no longer need to be stored in typeglobs:
525 declaring a subroutine will now put a simple sub reference directly
526 in the stash if possible, saving memory. The typeglob still
527 notionally exists, so accessing it will cause the stash entry to be
528 upgraded to a typeglob (i.e. this is just an internal
529 implementation detail). This optimization does not currently apply
530 to XSUBs or exported subroutines, and method calls will undo it,
531 since they cache things in typeglobs. [GH #13392]
532 <https://github.com/Perl/perl5/issues/13392>
533 • The functions "utf8::native_to_unicode()" and
535 "utf8::unicode_to_native()" (see utf8) are now optimized out on
536 ASCII platforms. There is now not even a minimal performance hit
537 in writing code portable between ASCII and EBCDIC platforms.
538 • Win32 Perl uses 8 KB less of per-process memory than before for
540 every perl process, because some data is now memory mapped from
541 disk and shared between processes from the same perl binary.
542
544 Updated Modules and Pragmata
545 Many of the libraries distributed with perl have been upgraded since
546 v5.20.0. For a complete list of changes, run:
547 corelist --diff 5.20.0 5.22.0
549 You can substitute your favorite version in place of 5.20.0, too.
551 Some notable changes include:
553 • Archive::Tar has been upgraded to version 2.04.
555 Tests can now be run in parallel.
557 • attributes has been upgraded to version 0.27.
559 The usage of "memEQs" in the XS has been corrected. [GH #14072]
561 <https://github.com/Perl/perl5/issues/14072>
562 Avoid reading beyond the end of a buffer. [perl #122629]
564 • B has been upgraded to version 1.58.
566 It provides a new "B::safename" function, based on the existing
568 "B::GV->SAFENAME", that converts "\cOPEN" to "^OPEN".
569 Nulled COPs are now of class "B::COP", rather than "B::OP".
571 "B::REGEXP" objects now provide a "qr_anoncv" method for accessing
573 the implicit CV associated with "qr//" things containing code
574 blocks, and a "compflags" method that returns the pertinent flags
575 originating from the "qr//blahblah" op.
576 "B::PMOP" now provides a "pmregexp" method returning a "B::REGEXP"
578 object. Two new classes, "B::PADNAME" and "B::PADNAMELIST", have
579 been introduced.
580 A bug where, after an ithread creation or pseudofork,
582 special/immortal SVs in the child ithread/pseudoprocess did not
583 have the correct class of "B::SPECIAL", has been fixed. The "id"
584 and "outid" PADLIST methods have been added.
585 • B::Concise has been upgraded to version 0.996.
587 Null ops that are part of the execution chain are now given
589 sequence numbers.
590 Private flags for nulled ops are now dumped with mnemonics as they
592 would be for the non-nulled counterparts.
593 • B::Deparse has been upgraded to version 1.35.
595 It now deparses "+sub : attr { ... }" correctly at the start of a
597 statement. Without the initial "+", "sub" would be a statement
598 label.
599 "BEGIN" blocks are now emitted in the right place most of the time,
601 but the change unfortunately introduced a regression, in that
602 "BEGIN" blocks occurring just before the end of the enclosing block
603 may appear below it instead.
604 "B::Deparse" no longer puts erroneous "local" here and there, such
606 as for "LIST = tr/a//d". [perl #119815]
607 Adjacent "use" statements are no longer accidentally nested if one
609 contains a "do" block. [perl #115066]
610 Parenthesised arrays in lists passed to "\" are now correctly
612 deparsed with parentheses (e.g., "\(@a, (@b), @c)" now retains the
613 parentheses around @b), thus preserving the flattening behavior of
614 referenced parenthesised arrays. Formerly, it only worked for one
615 array: "\(@a)".
616 "local our" is now deparsed correctly, with the "our" included.
618 "for($foo; !$bar; $baz) {...}" was deparsed without the "!" (or
620 "not"). This has been fixed.
621 Core keywords that conflict with lexical subroutines are now
623 deparsed with the "CORE::" prefix.
624 "foreach state $x (...) {...}" now deparses correctly with "state"
626 and not "my".
627 "our @array = split(...)" now deparses correctly with "our" in
629 those cases where the assignment is optimized away.
630 It now deparses "our(LIST)" and typed lexical ("my Dog $spot")
632 correctly.
633 Deparse $#_ as that instead of as $#{_}. [GH #14545]
635 <https://github.com/Perl/perl5/issues/14545>
636 BEGIN blocks at the end of the enclosing scope are now deparsed in
638 the right place. [perl #77452]
639 BEGIN blocks were sometimes deparsed as __ANON__, but are now
641 always called BEGIN.
642 Lexical subroutines are now fully deparsed. [perl #116553]
644 "Anything =~ y///r" with "/r" no longer omits the left-hand
646 operand.
647 The op trees that make up regexp code blocks are now deparsed for
649 real. Formerly, the original string that made up the regular
650 expression was used. That caused problems with
651 "qr/(?{<<heredoc})/" and multiline code blocks, which were deparsed
652 incorrectly. [perl #123217] [perl #115256]
653 $; at the end of a statement no longer loses its semicolon. [perl
655 #123357]
656 Some cases of subroutine declarations stored in the stash in
658 shorthand form were being omitted.
659 Non-ASCII characters are now consistently escaped in strings,
661 instead of some of the time. (There are still outstanding problems
662 with regular expressions and identifiers that have not been fixed.)
663 When prototype sub calls are deparsed with "&" (e.g., under the -P
665 option), "scalar" is now added where appropriate, to force the
666 scalar context implied by the prototype.
667 "require(foo())", "do(foo())", "goto(foo())" and similar constructs
669 with loop controls are now deparsed correctly. The outer
670 parentheses are not optional.
671 Whitespace is no longer escaped in regular expressions, because it
673 was getting erroneously escaped within "(?x:...)" sections.
674 "sub foo { foo() }" is now deparsed with those mandatory
676 parentheses.
677 "/@array/" is now deparsed as a regular expression, and not just
679 @array.
680 "/@{-}/", "/@{+}/" and $#{1} are now deparsed with the braces,
682 which are mandatory in these cases.
683 In deparsing feature bundles, "B::Deparse" was emitting "no
685 feature;" first instead of "no feature ':all';". This has been
686 fixed.
687 "chdir FH" is now deparsed without quotation marks.
689 "\my @a" is now deparsed without parentheses. (Parenthese would
691 flatten the array.)
692 "system" and "exec" followed by a block are now deparsed correctly.
694 Formerly there was an erroneous "do" before the block.
695 "use constant QR => qr/.../flags" followed by """ =~ QR" is no
697 longer without the flags.
698 Deparsing "BEGIN { undef &foo }" with the -w switch enabled started
700 to emit 'uninitialized' warnings in Perl 5.14. This has been
701 fixed.
702 Deparsing calls to subs with a "(;+)" prototype resulted in an
704 infinite loop. The "(;$") "(_)" and "(;_)" prototypes were given
705 the wrong precedence, causing "foo($a<$b)" to be deparsed without
706 the parentheses.
707 Deparse now provides a defined state sub in inner subs.
709 • B::Op_private has been added.
711 B::Op_private provides detailed information about the flags used in
713 the "op_private" field of perl opcodes.
714 • bigint, bignum, bigrat have been upgraded to version 0.39.
716 Document in CAVEATS that using strings as numbers won't always
718 invoke the big number overloading, and how to invoke it.
719 [rt.perl.org #123064]
720 • Carp has been upgraded to version 1.36.
722 "Carp::Heavy" now ignores version mismatches with Carp if Carp is
724 newer than 1.12, since "Carp::Heavy"'s guts were merged into Carp
725 at that point. [GH #13708]
726 <https://github.com/Perl/perl5/issues/13708>
727 Carp now handles non-ASCII platforms better.
729 Off-by-one error fix for Perl < 5.14.
731 • constant has been upgraded to version 1.33.
733 It now accepts fully-qualified constant names, allowing constants
735 to be defined in packages other than the caller.
736 • CPAN has been upgraded to version 2.11.
738 Add support for "Cwd::getdcwd()" and introduce workaround for a
740 misbehavior seen on Strawberry Perl 5.20.1.
741 Fix "chdir()" after building dependencies bug.
743 Introduce experimental support for plugins/hooks.
745 Integrate the "App::Cpan" sources.
747 Do not check recursion on optional dependencies.
749 Sanity check META.yml to contain a hash. [cpan #95271]
751 <https://rt.cpan.org/Ticket/Display.html?id=95271>
752 • CPAN::Meta::Requirements has been upgraded to version 2.132.
754 Works around limitations in "version::vpp" detecting v-string magic
756 and adds support for forthcoming ExtUtils::MakeMaker bootstrap
757 version.pm for Perls older than 5.10.0.
758 • Data::Dumper has been upgraded to version 2.158.
760 Fixes CVE-2014-4330 by adding a configuration variable/option to
762 limit recursion when dumping deep data structures.
763 Changes to resolve Coverity issues. XS dumps incorrectly stored
765 the name of code references stored in a GLOB. [GH #13911]
766 <https://github.com/Perl/perl5/issues/13911>
767 • DynaLoader has been upgraded to version 1.32.
769 Remove "dl_nonlazy" global if unused in Dynaloader. [perl #122926]
771 • Encode has been upgraded to version 2.72.
773 "piconv" now has better error handling when the encoding name is
775 nonexistent, and a build breakage when upgrading Encode in
776 perl-5.8.2 and earlier has been fixed.
777 Building in C++ mode on Windows now works.
779 • Errno has been upgraded to version 1.23.
781 Add "-P" to the preprocessor command-line on GCC 5. GCC added
783 extra line directives, breaking parsing of error code definitions.
784 [rt.perl.org #123784]
785 • experimental has been upgraded to version 0.013.
787 Hardcodes features for Perls older than 5.15.7.
789 • ExtUtils::CBuilder has been upgraded to version 0.280221.
791 Fixes a regression on Android. [GH #14064]
793 <https://github.com/Perl/perl5/issues/14064>
794 • ExtUtils::Manifest has been upgraded to version 1.70.
796 Fixes a bug with "maniread()"'s handling of quoted filenames and
798 improves "manifind()" to follow symlinks. [GH #14003]
799 <https://github.com/Perl/perl5/issues/14003>
800 • ExtUtils::ParseXS has been upgraded to version 3.28.
802 Only declare "file" unused if we actually define it. Improve
804 generated "RETVAL" code generation to avoid repeated references to
805 ST(0). [perl #123278] Broaden and document the "/OBJ$/" to
806 "/REF$/" typemap optimization for the "DESTROY" method. [perl
807 #123418]
808 • Fcntl has been upgraded to version 1.13.
810 Add support for the Linux pipe buffer size "fcntl()" commands.
812 • File::Find has been upgraded to version 1.29.
814 "find()" and "finddepth()" will now warn if passed inappropriate or
816 misspelled options.
817 • File::Glob has been upgraded to version 1.24.
819 Avoid "SvIV()" expanding to call "get_sv()" three times in a few
821 places. [perl #123606]
822 • HTTP::Tiny has been upgraded to version 0.054.
824 "keep_alive" is now fork-safe and thread-safe.
826 • IO has been upgraded to version 1.35.
828 The XS implementation has been fixed for the sake of older Perls.
830 • IO::Socket has been upgraded to version 1.38.
832 Document the limitations of the "connected()" method. [perl
834 #123096]
835 • IO::Socket::IP has been upgraded to version 0.37.
837 A better fix for subclassing "connect()". [cpan #95983]
839 <https://rt.cpan.org/Ticket/Display.html?id=95983> [cpan #97050]
840 <https://rt.cpan.org/Ticket/Display.html?id=97050>
841 Implements Timeout for "connect()". [cpan #92075]
843 <https://rt.cpan.org/Ticket/Display.html?id=92075>
844 • The libnet collection of modules has been upgraded to version 3.05.
846 Support for IPv6 and SSL to "Net::FTP", "Net::NNTP", "Net::POP3"
848 and "Net::SMTP". Improvements in "Net::SMTP" authentication.
849 • Locale::Codes has been upgraded to version 3.34.
851 Fixed a bug in the scripts used to extract data from spreadsheets
853 that prevented the SHP currency code from being found. [cpan
854 #94229] <https://rt.cpan.org/Ticket/Display.html?id=94229>
855 New codes have been added.
857 • Math::BigInt has been upgraded to version 1.9997.
859 Synchronize POD changes from the CPAN release.
861 "Math::BigFloat->blog(x)" would sometimes return "blog(2*x)" when
862 the accuracy was greater than 70 digits. The result of
863 "Math::BigFloat->bdiv()" in list context now satisfies "x =
864 quotient * divisor + remainder".
865 Correct handling of subclasses. [cpan #96254]
867 <https://rt.cpan.org/Ticket/Display.html?id=96254> [cpan #96329]
868 <https://rt.cpan.org/Ticket/Display.html?id=96329>
869 • Module::Metadata has been upgraded to version 1.000026.
871 Support installations on older perls with an ExtUtils::MakeMaker
873 earlier than 6.63_03
874 • overload has been upgraded to version 1.26.
876 A redundant "ref $sub" check has been removed.
878 • The PathTools module collection has been upgraded to version 3.56.
880 A warning from the gcc compiler is now avoided when building the
882 XS.
883 Don't turn leading "//" into "/" on Cygwin. [perl #122635]
885 • perl5db.pl has been upgraded to version 1.49.
887 The debugger would cause an assertion failure. [GH #14605]
889 <https://github.com/Perl/perl5/issues/14605>
890 "fork()" in the debugger under "tmux" will now create a new window
892 for the forked process. [GH #13602]
893 <https://github.com/Perl/perl5/issues/13602>
894 The debugger now saves the current working directory on startup and
896 restores it when you restart your program with "R" or "rerun". [GH
897 #13691] <https://github.com/Perl/perl5/issues/13691>
898 • PerlIO::scalar has been upgraded to version 0.22.
900 Reading from a position well past the end of the scalar now
902 correctly returns end of file. [perl #123443]
903 Seeking to a negative position still fails, but no longer leaves
905 the file position set to a negation location.
906 "eof()" on a "PerlIO::scalar" handle now properly returns true when
908 the file position is past the 2GB mark on 32-bit systems.
909 Attempting to write at file positions impossible for the platform
911 now fail early rather than wrapping at 4GB.
912 • Pod::Perldoc has been upgraded to version 3.25.
914 Filehandles opened for reading or writing now have
916 ":encoding(UTF-8)" set. [cpan #98019]
917 <https://rt.cpan.org/Ticket/Display.html?id=98019>
918 • POSIX has been upgraded to version 1.53.
920 The C99 math functions and constants (for example "acosh", "isinf",
922 "isnan", "round", "trunc"; "M_E", "M_SQRT2", "M_PI") have been
923 added.
924 "POSIX::tmpnam()" now produces a deprecation warning. [perl
926 #122005]
927 • Safe has been upgraded to version 2.39.
929 "reval" was not propagating void context properly.
931 • Scalar-List-Utils has been upgraded to version 1.41.
933 A new module, Sub::Util, has been added, containing functions
935 related to CODE refs, including "subname" (inspired by
936 "Sub::Identity") and "set_subname" (copied and renamed from
937 "Sub::Name"). The use of "GetMagic" in "List::Util::reduce()" has
938 also been fixed. [cpan #63211]
939 <https://rt.cpan.org/Ticket/Display.html?id=63211>
940 • SDBM_File has been upgraded to version 1.13.
942 Simplified the build process. [perl #123413]
944 • Time::Piece has been upgraded to version 1.29.
946 When pretty printing negative "Time::Seconds", the "minus" is no
948 longer lost.
949 • Unicode::Collate has been upgraded to version 1.12.
951 Version 0.67's improved discontiguous contractions is invalidated
953 by default and is supported as a parameter "long_contraction".
954 • Unicode::Normalize has been upgraded to version 1.18.
956 The XSUB implementation has been removed in favor of pure Perl.
958 • Unicode::UCD has been upgraded to version 0.61.
960 A new function property_values() has been added to return a given
962 property's possible values.
963 A new function charprop() has been added to return the value of a
965 given property for a given code point.
966 A new function charprops_all() has been added to return the values
968 of all Unicode properties for a given code point.
969 A bug has been fixed so that propaliases() returns the correct
971 short and long names for the Perl extensions where it was
972 incorrect.
973 A bug has been fixed so that prop_value_aliases() returns "undef"
975 instead of a wrong result for properties that are Perl extensions.
976 This module now works on EBCDIC platforms.
978 • utf8 has been upgraded to version 1.17
980 A mismatch between the documentation and the code in
982 "utf8::downgrade()" was fixed in favor of the documentation. The
983 optional second argument is now correctly treated as a perl boolean
984 (true/false semantics) and not as an integer.
985 • version has been upgraded to version 0.9909.
987 Numerous changes. See the Changes file in the CPAN distribution
989 for details.
990 • Win32 has been upgraded to version 0.51.
992 "GetOSName()" now supports Windows 8.1, and building in C++ mode
994 now works.
995 • Win32API::File has been upgraded to version 0.1202
997 Building in C++ mode now works.
999 • XSLoader has been upgraded to version 0.20.
1001 Allow XSLoader to load modules from a different namespace. [perl
1003 #122455]
1004 Removed Modules and Pragmata
1006 The following modules (and associated modules) have been removed from
1007 the core perl distribution:
1008 • CGI
1010 • Module::Build
1012
1014 New Documentation
1015 perlunicook
1016 This document, by Tom Christiansen, provides examples of handling
1018 Unicode in Perl.
1019 Changes to Existing Documentation
1021 perlaix
1022 • A note on long doubles has been added.
1024 perlapi
1026 • Note that "SvSetSV" doesn't do set magic.
1028 • "sv_usepvn_flags" - fix documentation to mention the use of "Newx"
1030 instead of "malloc".
1031 [GH #13835] <https://github.com/Perl/perl5/issues/13835>
1033 • Clarify where "NUL" may be embedded or is required to terminate a
1035 string.
1036 • Some documentation that was previously missing due to formatting
1038 errors is now included.
1039 • Entries are now organized into groups rather than by the file where
1041 they are found.
1042 • Alphabetical sorting of entries is now done consistently
1044 (automatically by the POD generator) to make entries easier to find
1045 when scanning.
1046 perldata
1048 • The syntax of single-character variable names has been brought up-
1050 to-date and more fully explained.
1051 • Hexadecimal floating point numbers are described, as are infinity
1053 and NaN.
1054 perlebcdic
1056 • This document has been significantly updated in the light of recent
1058 improvements to EBCDIC support.
1059 perlfilter
1061 • Added a LIMITATIONS section.
1063 perlfunc
1065 • Mention that "study()" is currently a no-op.
1067 • Calling "delete" or "exists" on array values is now described as
1069 "strongly discouraged" rather than "deprecated".
1070 • Improve documentation of "our".
1072 • "-l" now notes that it will return false if symlinks aren't
1074 supported by the file system. [GH #13695]
1075 <https://github.com/Perl/perl5/issues/13695>
1076 • Note that "exec LIST" and "system LIST" may fall back to the shell
1078 on Win32. Only the indirect-object syntax "exec PROGRAM LIST" and
1079 "system PROGRAM LIST" will reliably avoid using the shell.
1080 This has also been noted in perlport.
1082 [GH #13907] <https://github.com/Perl/perl5/issues/13907>
1084 perlguts
1086 • The OOK example has been updated to account for COW changes and a
1088 change in the storage of the offset.
1089 • Details on C level symbols and libperl.t added.
1091 • Information on Unicode handling has been added
1093 • Information on EBCDIC handling has been added
1095 perlhack
1097 • A note has been added about running on platforms with non-ASCII
1099 character sets
1100 • A note has been added about performance testing
1102 perlhacktips
1104 • Documentation has been added illustrating the perils of assuming
1106 that there is no change to the contents of static memory pointed to
1107 by the return values of Perl's wrappers for C library functions.
1108 • Replacements for "tmpfile", "atoi", "strtol", and "strtoul" are now
1110 recommended.
1111 • Updated documentation for the "test.valgrind" "make" target. [GH
1113 #13658] <https://github.com/Perl/perl5/issues/13658>
1114 • Information is given about writing test files portably to non-ASCII
1116 platforms.
1117 • A note has been added about how to get a C language stack
1119 backtrace.
1120 perlhpux
1122 • Note that the message "Redeclaration of "sendpath" with a different
1124 storage class specifier" is harmless.
1125 perllocale
1127 • Updated for the enhancements in v5.22, along with some
1129 clarifications.
1130 perlmodstyle
1132 • Instead of pointing to the module list, we are now pointing to
1134 PrePAN <http://prepan.org/>.
1135 perlop
1137 • Updated for the enhancements in v5.22, along with some
1139 clarifications.
1140 perlpodspec
1142 • The specification of the pod language is changing so that the
1144 default encoding of pods that aren't in UTF-8 (unless otherwise
1145 indicated) is CP1252 instead of ISO 8859-1 (Latin1).
1146 perlpolicy
1148 • We now have a code of conduct for the p5p mailing list, as
1150 documented in "STANDARDS OF CONDUCT" in perlpolicy.
1151 • The conditions for marking an experimental feature as non-
1153 experimental are now set out.
1154 • Clarification has been made as to what sorts of changes are
1156 permissible in maintenance releases.
1157 perlport
1159 • Out-of-date VMS-specific information has been fixed and/or
1161 simplified.
1162 • Notes about EBCDIC have been added.
1164 perlre
1166 • The description of the "/x" modifier has been clarified to note
1168 that comments cannot be continued onto the next line by escaping
1169 them; and there is now a list of all the characters that are
1170 considered whitespace by this modifier.
1171 • The new "/n" modifier is described.
1173 • A note has been added on how to make bracketed character class
1175 ranges portable to non-ASCII machines.
1176 perlrebackslash
1178 • Added documentation of "\b{sb}", "\b{wb}", "\b{gcb}", and "\b{g}".
1180 perlrecharclass
1182 • Clarifications have been added to "Character Ranges" in
1184 perlrecharclass to the effect "[A-Z]", "[a-z]", "[0-9]" and any
1185 subranges thereof in regular expression bracketed character classes
1186 are guaranteed to match exactly what a naive English speaker would
1187 expect them to match, even on platforms (such as EBCDIC) where perl
1188 has to do extra work to accomplish this.
1189 • The documentation of Bracketed Character Classes has been expanded
1191 to cover the improvements in "qr/[\N{named sequence}]/" (see under
1192 "Selected Bug Fixes").
1193 perlref
1195 • A new section has been added Assigning to References
1197 perlsec
1199 • Comments added on algorithmic complexity and tied hashes.
1201 perlsyn
1203 • An ambiguity in the documentation of the "..." statement has been
1205 corrected. [GH #14054]
1206 <https://github.com/Perl/perl5/issues/14054>
1207 • The empty conditional in "for" and "while" is now documented in
1209 perlsyn.
1210 perlunicode
1212 • This has had extensive revisions to bring it up-to-date with
1214 current Unicode support and to make it more readable. Notable is
1215 that Unicode 7.0 changed what it should do with non-characters.
1216 Perl retains the old way of handling for reasons of backward
1217 compatibility. See "Noncharacter code points" in perlunicode.
1218 perluniintro
1220 • Advice for how to make sure your strings and regular expression
1222 patterns are interpreted as Unicode has been updated.
1223 perlvar
1225 • $] is no longer listed as being deprecated. Instead, discussion
1227 has been added on the advantages and disadvantages of using it
1228 versus $^V. $OLD_PERL_VERSION was re-added to the documentation as
1229 the long form of $].
1230 • "${^ENCODING}" is now marked as deprecated.
1232 • The entry for "%^H" has been clarified to indicate it can only
1234 handle simple values.
1235 perlvms
1237 • Out-of-date and/or incorrect material has been removed.
1239 • Updated documentation on environment and shell interaction in VMS.
1241 perlxs
1243 • Added a discussion of locale issues in XS code.
1245
1247 The following additions or changes have been made to diagnostic output,
1248 including warnings and fatal error messages. For the complete list of
1249 diagnostic messages, see perldiag.
1250 New Diagnostics
1252 New Errors
1253 • Bad symbol for scalar
1255 (P) An internal request asked to add a scalar entry to something
1257 that wasn't a symbol table entry.
1258 • Can't use a hash as a reference
1260 (F) You tried to use a hash as a reference, as in "%foo->{"bar"}"
1262 or "%$ref->{"hello"}". Versions of perl <= 5.6.1 used to allow
1263 this syntax, but shouldn't have.
1264 • Can't use an array as a reference
1266 (F) You tried to use an array as a reference, as in "@foo->[23]" or
1268 "@$ref->[99]". Versions of perl <= 5.6.1 used to allow this
1269 syntax, but shouldn't have.
1270 • Can't use 'defined(@array)' (Maybe you should just omit the
1272 defined()?)
1273 (F) "defined()" is not useful on arrays because it checks for an
1275 undefined scalar value. If you want to see if the array is empty,
1276 just use "if (@array) { # not empty }" for example.
1277 • Can't use 'defined(%hash)' (Maybe you should just omit the
1279 defined()?)
1280 (F) "defined()" is not usually right on hashes.
1282 Although "defined %hash" is false on a plain not-yet-used hash, it
1284 becomes true in several non-obvious circumstances, including
1285 iterators, weak references, stash names, even remaining true after
1286 "undef %hash". These things make "defined %hash" fairly useless in
1287 practice, so it now generates a fatal error.
1288 If a check for non-empty is what you wanted then just put it in
1290 boolean context (see "Scalar values" in perldata):
1291 if (%hash) {
1293 # not empty
1294 }
1295 If you had "defined %Foo::Bar::QUUX" to check whether such a
1297 package variable exists then that's never really been reliable, and
1298 isn't a good way to enquire about the features of a package, or
1299 whether it's loaded, etc.
1300 • Cannot chr %f
1302 (F) You passed an invalid number (like an infinity or not-a-number)
1304 to "chr".
1305 • Cannot compress %f in pack
1307 (F) You tried converting an infinity or not-a-number to an unsigned
1309 character, which makes no sense.
1310 • Cannot pack %f with '%c'
1312 (F) You tried converting an infinity or not-a-number to a
1314 character, which makes no sense.
1315 • Cannot print %f with '%c'
1317 (F) You tried printing an infinity or not-a-number as a character
1319 (%c), which makes no sense. Maybe you meant '%s', or just
1320 stringifying it?
1321 • charnames alias definitions may not contain a sequence of multiple
1323 spaces
1324 (F) You defined a character name which had multiple space
1326 characters in a row. Change them to single spaces. Usually these
1327 names are defined in the ":alias" import argument to "use
1328 charnames", but they could be defined by a translator installed
1329 into $^H{charnames}. See "CUSTOM ALIASES" in charnames.
1330 • charnames alias definitions may not contain trailing white-space
1332 (F) You defined a character name which ended in a space character.
1334 Remove the trailing space(s). Usually these names are defined in
1335 the ":alias" import argument to "use charnames", but they could be
1336 defined by a translator installed into $^H{charnames}. See "CUSTOM
1337 ALIASES" in charnames.
1338 • :const is not permitted on named subroutines
1340 (F) The "const" attribute causes an anonymous subroutine to be run
1342 and its value captured at the time that it is cloned. Named
1343 subroutines are not cloned like this, so the attribute does not
1344 make sense on them.
1345 • Hexadecimal float: internal error
1347 (F) Something went horribly bad in hexadecimal float handling.
1349 • Hexadecimal float: unsupported long double format
1351 (F) You have configured Perl to use long doubles but the internals
1353 of the long double format are unknown, therefore the hexadecimal
1354 float output is impossible.
1355 • Illegal suidscript
1357 (F) The script run under suidperl was somehow illegal.
1359 • In '(?...)', the '(' and '?' must be adjacent in regex; marked by
1361 <-- HERE in m/%s/
1362 (F) The two-character sequence "(?" in this context in a regular
1364 expression pattern should be an indivisible token, with nothing
1365 intervening between the "(" and the "?", but you separated them.
1366 • In '(*VERB...)', the '(' and '*' must be adjacent in regex; marked
1368 by <-- HERE in m/%s/
1369 (F) The two-character sequence "(*" in this context in a regular
1371 expression pattern should be an indivisible token, with nothing
1372 intervening between the "(" and the "*", but you separated them.
1373 • Invalid quantifier in {,} in regex; marked by <-- HERE in m/%s/
1375 (F) The pattern looks like a {min,max} quantifier, but the min or
1377 max could not be parsed as a valid number: either it has leading
1378 zeroes, or it represents too big a number to cope with. The
1379 <-- HERE shows where in the regular expression the problem was
1380 discovered. See perlre.
1381 • '%s' is an unknown bound type in regex
1383 (F) You used "\b{...}" or "\B{...}" and the "..." is not known to
1385 Perl. The current valid ones are given in "\b{}, \b, \B{}, \B" in
1386 perlrebackslash.
1387 • Missing or undefined argument to require
1389 (F) You tried to call "require" with no argument or with an
1391 undefined value as an argument. "require" expects either a package
1392 name or a file-specification as an argument. See "require" in
1393 perlfunc.
1394 Formerly, "require" with no argument or "undef" warned about a Null
1396 filename.
1397 New Warnings
1399 • \C is deprecated in regex
1401 (D deprecated) The "/\C/" character class was deprecated in v5.20,
1403 and now emits a warning. It is intended that it will become an
1404 error in v5.24. This character class matches a single byte even if
1405 it appears within a multi-byte character, breaks encapsulation, and
1406 can corrupt UTF-8 strings.
1407 • "%s" is more clearly written simply as "%s" in regex; marked by <--
1409 HERE in m/%s/
1410 (W regexp) (only under "use re 'strict'" or within "(?[...])")
1412 You specified a character that has the given plainer way of writing
1414 it, and which is also portable to platforms running with different
1415 character sets.
1416 • Argument "%s" treated as 0 in increment (++)
1418 (W numeric) The indicated string was fed as an argument to the "++"
1420 operator which expects either a number or a string matching
1421 "/^[a-zA-Z]*[0-9]*\z/". See "Auto-increment and Auto-decrement" in
1422 perlop for details.
1423 • Both or neither range ends should be Unicode in regex; marked by
1425 <-- HERE in m/%s/
1426 (W regexp) (only under "use re 'strict'" or within "(?[...])")
1428 In a bracketed character class in a regular expression pattern, you
1430 had a range which has exactly one end of it specified using "\N{}",
1431 and the other end is specified using a non-portable mechanism.
1432 Perl treats the range as a Unicode range, that is, all the
1433 characters in it are considered to be the Unicode characters, and
1434 which may be different code points on some platforms Perl runs on.
1435 For example, "[\N{U+06}-\x08]" is treated as if you had instead
1436 said "[\N{U+06}-\N{U+08}]", that is it matches the characters whose
1437 code points in Unicode are 6, 7, and 8. But that "\x08" might
1438 indicate that you meant something different, so the warning gets
1439 raised.
1440 • Can't do %s("%s") on non-UTF-8 locale; resolved to "%s".
1442 (W locale) You are 1) running under ""use locale""; 2) the current
1444 locale is not a UTF-8 one; 3) you tried to do the designated case-
1445 change operation on the specified Unicode character; and 4) the
1446 result of this operation would mix Unicode and locale rules, which
1447 likely conflict.
1448 The warnings category "locale" is new.
1450 • :const is experimental
1452 (S experimental::const_attr) The "const" attribute is experimental.
1454 If you want to use the feature, disable the warning with "no
1455 warnings 'experimental::const_attr'", but know that in doing so you
1456 are taking the risk that your code may break in a future Perl
1457 version.
1458 • gmtime(%f) failed
1460 (W overflow) You called "gmtime" with a number that it could not
1462 handle: too large, too small, or NaN. The returned value is
1463 "undef".
1464 • Hexadecimal float: exponent overflow
1466 (W overflow) The hexadecimal floating point has larger exponent
1468 than the floating point supports.
1469 • Hexadecimal float: exponent underflow
1471 (W overflow) The hexadecimal floating point has smaller exponent
1473 than the floating point supports.
1474 • Hexadecimal float: mantissa overflow
1476 (W overflow) The hexadecimal floating point literal had more bits
1478 in the mantissa (the part between the "0x" and the exponent, also
1479 known as the fraction or the significand) than the floating point
1480 supports.
1481 • Hexadecimal float: precision loss
1483 (W overflow) The hexadecimal floating point had internally more
1485 digits than could be output. This can be caused by unsupported
1486 long double formats, or by 64-bit integers not being available
1487 (needed to retrieve the digits under some configurations).
1488 • Locale '%s' may not work well.%s
1490 (W locale) You are using the named locale, which is a non-UTF-8
1492 one, and which perl has determined is not fully compatible with
1493 what it can handle. The second %s gives a reason.
1494 The warnings category "locale" is new.
1496 • localtime(%f) failed
1498 (W overflow) You called "localtime" with a number that it could not
1500 handle: too large, too small, or NaN. The returned value is
1501 "undef".
1502 • Negative repeat count does nothing
1504 (W numeric) You tried to execute the "x" repetition operator fewer
1506 than 0 times, which doesn't make sense.
1507 • NO-BREAK SPACE in a charnames alias definition is deprecated
1509 (D deprecated) You defined a character name which contained a no-
1511 break space character. Change it to a regular space. Usually
1512 these names are defined in the ":alias" import argument to "use
1513 charnames", but they could be defined by a translator installed
1514 into $^H{charnames}. See "CUSTOM ALIASES" in charnames.
1515 • Non-finite repeat count does nothing
1517 (W numeric) You tried to execute the "x" repetition operator "Inf"
1519 (or "-Inf") or NaN times, which doesn't make sense.
1520 • PerlIO layer ':win32' is experimental
1522 (S experimental::win32_perlio) The ":win32" PerlIO layer is
1524 experimental. If you want to take the risk of using this layer,
1525 simply disable this warning:
1526 no warnings "experimental::win32_perlio";
1528 • Ranges of ASCII printables should be some subset of "0-9", "A-Z",
1530 or "a-z" in regex; marked by <-- HERE in m/%s/
1531 (W regexp) (only under "use re 'strict'" or within "(?[...])")
1533 Stricter rules help to find typos and other errors. Perhaps you
1535 didn't even intend a range here, if the "-" was meant to be some
1536 other character, or should have been escaped (like "\-"). If you
1537 did intend a range, the one that was used is not portable between
1538 ASCII and EBCDIC platforms, and doesn't have an obvious meaning to
1539 a casual reader.
1540 [3-7] # OK; Obvious and portable
1542 [d-g] # OK; Obvious and portable
1543 [A-Y] # OK; Obvious and portable
1544 [A-z] # WRONG; Not portable; not clear what is meant
1545 [a-Z] # WRONG; Not portable; not clear what is meant
1546 [%-.] # WRONG; Not portable; not clear what is meant
1547 [\x41-Z] # WRONG; Not portable; not obvious to non-geek
1548 (You can force portability by specifying a Unicode range, which
1550 means that the endpoints are specified by "\N{...}", but the
1551 meaning may still not be obvious.) The stricter rules require that
1552 ranges that start or stop with an ASCII character that is not a
1553 control have all their endpoints be a literal character, and not
1554 some escape sequence (like "\x41"), and the ranges must be all
1555 digits, or all uppercase letters, or all lowercase letters.
1556 • Ranges of digits should be from the same group in regex; marked by
1558 <-- HERE in m/%s/
1559 (W regexp) (only under "use re 'strict'" or within "(?[...])")
1561 Stricter rules help to find typos and other errors. You included a
1563 range, and at least one of the end points is a decimal digit.
1564 Under the stricter rules, when this happens, both end points should
1565 be digits in the same group of 10 consecutive digits.
1566 • Redundant argument in %s
1568 (W redundant) You called a function with more arguments than were
1570 needed, as indicated by information within other arguments you
1571 supplied (e.g. a printf format). Currently only emitted when a
1572 printf-type format required fewer arguments than were supplied, but
1573 might be used in the future for e.g. "pack" in perlfunc.
1574 The warnings category "redundant" is new. See also [GH #13534]
1576 <https://github.com/Perl/perl5/issues/13534>.
1577 • Replacement list is longer than search list
1579 This is not a new diagnostic, but in earlier releases was
1581 accidentally not displayed if the transliteration contained wide
1582 characters. This is now fixed, so that you may see this diagnostic
1583 in places where you previously didn't (but should have).
1584 • Use of \b{} for non-UTF-8 locale is wrong. Assuming a UTF-8 locale
1586 (W locale) You are matching a regular expression using locale
1588 rules, and a Unicode boundary is being matched, but the locale is
1589 not a Unicode one. This doesn't make sense. Perl will continue,
1590 assuming a Unicode (UTF-8) locale, but the results could well be
1591 wrong except if the locale happens to be ISO-8859-1 (Latin1) where
1592 this message is spurious and can be ignored.
1593 The warnings category "locale" is new.
1595 • Using /u for '%s' instead of /%s in regex; marked by <-- HERE in
1597 m/%s/
1598 (W regexp) You used a Unicode boundary ("\b{...}" or "\B{...}") in
1600 a portion of a regular expression where the character set modifiers
1601 "/a" or "/aa" are in effect. These two modifiers indicate an ASCII
1602 interpretation, and this doesn't make sense for a Unicode
1603 definition. The generated regular expression will compile so that
1604 the boundary uses all of Unicode. No other portion of the regular
1605 expression is affected.
1606 • The bitwise feature is experimental
1608 (S experimental::bitwise) This warning is emitted if you use
1610 bitwise operators ("& | ^ ~ &. |. ^. ~.") with the "bitwise"
1611 feature enabled. Simply suppress the warning if you want to use
1612 the feature, but know that in doing so you are taking the risk of
1613 using an experimental feature which may change or be removed in a
1614 future Perl version:
1615 no warnings "experimental::bitwise";
1617 use feature "bitwise";
1618 $x |.= $y;
1619 • Unescaped left brace in regex is deprecated, passed through in
1621 regex; marked by <-- HERE in m/%s/
1622 (D deprecated, regexp) You used a literal "{" character in a
1624 regular expression pattern. You should change to use "\{" instead,
1625 because a future version of Perl (tentatively v5.26) will consider
1626 this to be a syntax error. If the pattern delimiters are also
1627 braces, any matching right brace ("}") should also be escaped to
1628 avoid confusing the parser, for example,
1629 qr{abc\{def\}ghi}
1631 • Use of literal non-graphic characters in variable names is
1633 deprecated
1634 (D deprecated) Using literal non-graphic (including control)
1636 characters in the source to refer to the ^FOO variables, like $^X
1637 and "${^GLOBAL_PHASE}" is now deprecated.
1638 • Useless use of attribute "const"
1640 (W misc) The "const" attribute has no effect except on anonymous
1642 closure prototypes. You applied it to a subroutine via
1643 attributes.pm. This is only useful inside an attribute handler for
1644 an anonymous subroutine.
1645 • Useless use of /d modifier in transliteration operator
1647 This is not a new diagnostic, but in earlier releases was
1649 accidentally not displayed if the transliteration contained wide
1650 characters. This is now fixed, so that you may see this diagnostic
1651 in places where you previously didn't (but should have).
1652 • "use re 'strict'" is experimental
1654 (S experimental::re_strict) The things that are different when a
1656 regular expression pattern is compiled under 'strict' are subject
1657 to change in future Perl releases in incompatible ways; there are
1658 also proposals to change how to enable strict checking instead of
1659 using this subpragma. This means that a pattern that compiles
1660 today may not in a future Perl release. This warning is to alert
1661 you to that risk.
1662 • Warning: unable to close filehandle properly: %s
1664 Warning: unable to close filehandle %s properly: %s
1666 (S io) Previously, perl silently ignored any errors when doing an
1668 implicit close of a filehandle, i.e. where the reference count of
1669 the filehandle reached zero and the user's code hadn't already
1670 called "close()"; e.g.
1671 {
1673 open my $fh, '>', $file or die "open: '$file': $!\n";
1674 print $fh, $data or die;
1675 } # implicit close here
1676 In a situation such as disk full, due to buffering, the error may
1678 only be detected during the final close, so not checking the result
1679 of the close is dangerous.
1680 So perl now warns in such situations.
1682 • Wide character (U+%X) in %s
1684 (W locale) While in a single-byte locale (i.e., a non-UTF-8 one), a
1686 multi-byte character was encountered. Perl considers this
1687 character to be the specified Unicode code point. Combining
1688 non-UTF-8 locales and Unicode is dangerous. Almost certainly some
1689 characters will have two different representations. For example,
1690 in the ISO 8859-7 (Greek) locale, the code point 0xC3 represents a
1691 Capital Gamma. But so also does 0x393. This will make string
1692 comparisons unreliable.
1693 You likely need to figure out how this multi-byte character got
1695 mixed up with your single-byte locale (or perhaps you thought you
1696 had a UTF-8 locale, but Perl disagrees).
1697 The warnings category "locale" is new.
1699 Changes to Existing Diagnostics
1701 • <> should be quotes
1702 This warning has been changed to <> at require-statement should be
1704 quotes to make the issue more identifiable.
1705 • Argument "%s" isn't numeric%s
1707 The perldiag entry for this warning has added this clarifying note:
1709 Note that for the Inf and NaN (infinity and not-a-number) the
1711 definition of "numeric" is somewhat unusual: the strings themselves
1712 (like "Inf") are considered numeric, and anything following them is
1713 considered non-numeric.
1714 • Global symbol "%s" requires explicit package name
1716 This message has had '(did you forget to declare "my %s"?)'
1718 appended to it, to make it more helpful to new Perl programmers.
1719 [GH #13732] <https://github.com/Perl/perl5/issues/13732>
1720 • '"my" variable &foo::bar can't be in a package' has been reworded
1722 to say 'subroutine' instead of 'variable'.
1723 • \N{} in character class restricted to one character in regex;
1725 marked by <-- HERE in m/%s/
1726 This message has had character class changed to inverted character
1728 class or as a range end-point is to reflect improvements in
1729 "qr/[\N{named sequence}]/" (see under "Selected Bug Fixes").
1730 • panic: frexp
1732 This message has had ': %f' appended to it, to show what the
1734 offending floating point number is.
1735 • Possible precedence problem on bitwise %c operator reworded as
1737 Possible precedence problem on bitwise %s operator.
1738 • Unsuccessful %s on filename containing newline
1740 This warning is now only produced when the newline is at the end of
1742 the filename.
1743 • "Variable %s will not stay shared" has been changed to say
1745 "Subroutine" when it is actually a lexical sub that will not stay
1746 shared.
1747 • Variable length lookbehind not implemented in regex m/%s/
1749 The perldiag entry for this warning has had information about
1751 Unicode behavior added.
1752 Diagnostic Removals
1754 • "Ambiguous use of -foo resolved as -&foo()"
1755 There is actually no ambiguity here, and this impedes the use of
1757 negated constants; e.g., "-Inf".
1758 • "Constant is not a FOO reference"
1760 Compile-time checking of constant dereferencing (e.g.,
1762 "my_constant->()") has been removed, since it was not taking
1763 overloading into account. [GH #9891]
1764 <https://github.com/Perl/perl5/issues/9891> [GH #14044]
1765 <https://github.com/Perl/perl5/issues/14044>
1766
1768 find2perl, s2p and a2p removal
1769 • The x2p/ directory has been removed from the Perl core.
1770 This removes find2perl, s2p and a2p. They have all been released to
1772 CPAN as separate distributions ("App::find2perl", "App::s2p",
1773 "App::a2p").
1774 h2ph
1776 • h2ph now handles hexadecimal constants in the compiler's predefined
1777 macro definitions, as visible in $Config{cppsymbols}. [GH #14491]
1778 <https://github.com/Perl/perl5/issues/14491>.
1779 encguess
1781 • No longer depends on non-core modules.
1782
1784 • Configure now checks for "lrintl()", "lroundl()", "llrintl()", and
1785 "llroundl()".
1786 • Configure with "-Dmksymlinks" should now be faster. [GH #13890]
1788 <https://github.com/Perl/perl5/issues/13890>.
1789 • The "pthreads" and "cl" libraries will be linked by default if
1791 present. This allows XS modules that require threading to work on
1792 non-threaded perls. Note that you must still pass "-Dusethreads" if
1793 you want a threaded perl.
1794 • To get more precision and range for floating point numbers one can
1796 now use the GCC quadmath library which implements the quadruple
1797 precision floating point numbers on x86 and IA-64 platforms. See
1798 INSTALL for details.
1799 • MurmurHash64A and MurmurHash64B can now be configured as the
1801 internal hash function.
1802 • "make test.valgrind" now supports parallel testing.
1804 For example:
1806 TEST_JOBS=9 make test.valgrind
1808 See "valgrind" in perlhacktips for more information.
1810 [GH #13658] <https://github.com/Perl/perl5/issues/13658>
1812 • The MAD (Misc Attribute Decoration) build option has been removed
1814 This was an unmaintained attempt at preserving the Perl parse tree
1816 more faithfully so that automatic conversion of Perl 5 to Perl 6
1817 would have been easier.
1818 This build-time configuration option had been unmaintained for
1820 years, and had probably seriously diverged on both Perl 5 and Perl
1821 6 sides.
1822 • A new compilation flag, "-DPERL_OP_PARENT" is available. For
1824 details, see the discussion below at "Internal Changes".
1825 • Pathtools no longer tries to load XS on miniperl. This speeds up
1827 building perl slightly.
1828
1830 • t/porting/re_context.t has been added to test that utf8 and its
1831 dependencies only use the subset of the "$1..$n" capture vars that
1832 "Perl_save_re_context()" is hard-coded to localize, because that
1833 function has no efficient way of determining at runtime what vars
1834 to localize.
1835 • Tests for performance issues have been added in the file
1837 t/perf/taint.t.
1838 • Some regular expression tests are written in such a way that they
1840 will run very slowly if certain optimizations break. These tests
1841 have been moved into new files, t/re/speed.t and t/re/speed_thr.t,
1842 and are run with a "watchdog()".
1843 • "test.pl" now allows "plan skip_all => $reason", to make it more
1845 compatible with "Test::More".
1846 • A new test script, op/infnan.t, has been added to test if infinity
1848 and NaN are working correctly. See "Infinity and NaN (not-a-
1849 number) handling improved".
1850
1852 Regained Platforms
1853 IRIX and Tru64 platforms are working again.
1854 Some "make test" failures remain: [GH #14557]
1855 <https://github.com/Perl/perl5/issues/14557> and [GH #14727]
1856 <https://github.com/Perl/perl5/issues/14727> for IRIX; [GH #14629]
1857 <https://github.com/Perl/perl5/issues/14629>, [cpan #99605]
1858 <https://rt.cpan.org/Public/Bug/Display.html?id=99605>, and [cpan
1859 #104836] <https://rt.cpan.org/Ticket/Display.html?id=104836> for
1860 Tru64.
1861 z/OS running EBCDIC Code Page 1047
1863 Core perl now works on this EBCDIC platform. Earlier perls also
1864 worked, but, even though support wasn't officially withdrawn,
1865 recent perls would not compile and run well. Perl 5.20 would work,
1866 but had many bugs which have now been fixed. Many CPAN modules
1867 that ship with Perl still fail tests, including "Pod::Simple".
1868 However the version of "Pod::Simple" currently on CPAN should work;
1869 it was fixed too late to include in Perl 5.22. Work is under way
1870 to fix many of the still-broken CPAN modules, which likely will be
1871 installed on CPAN when completed, so that you may not have to wait
1872 until Perl 5.24 to get a working version.
1873 Discontinued Platforms
1875 NeXTSTEP/OPENSTEP
1876 NeXTSTEP was a proprietary operating system bundled with NeXT's
1877 workstations in the early to mid 90s; OPENSTEP was an API
1878 specification that provided a NeXTSTEP-like environment on a non-
1879 NeXTSTEP system. Both are now long dead, so support for building
1880 Perl on them has been removed.
1881 Platform-Specific Notes
1883 EBCDIC
1884 Special handling is required of the perl interpreter on EBCDIC
1885 platforms to get "qr/[i-j]/" to match only "i" and "j", since there
1886 are 7 characters between the code points for "i" and "j". This
1887 special handling had only been invoked when both ends of the range
1888 are literals. Now it is also invoked if any of the "\N{...}" forms
1889 for specifying a character by name or Unicode code point is used
1890 instead of a literal. See "Character Ranges" in perlrecharclass.
1891 HP-UX
1893 The archname now distinguishes use64bitint from use64bitall.
1894 Android
1896 Build support has been improved for cross-compiling in general and
1897 for Android in particular.
1898 VMS
1900 • When spawning a subprocess without waiting, the return value is
1901 now the correct PID.
1902 • Fix a prototype so linking doesn't fail under the VMS C++
1904 compiler.
1905 • "finite", "finitel", and "isfinite" detection has been added to
1907 "configure.com", environment handling has had some minor
1908 changes, and a fix for legacy feature checking status.
1909 Win32
1911 • miniperl.exe is now built with "-fno-strict-aliasing", allowing
1912 64-bit builds to complete on GCC 4.8. [GH #14556]
1913 <https://github.com/Perl/perl5/issues/14556>
1914 • "nmake minitest" now works on Win32. Due to dependency issues
1916 you need to build "nmake test-prep" first, and a small number
1917 of the tests fail. [GH #14318]
1918 <https://github.com/Perl/perl5/issues/14318>
1919 • Perl can now be built in C++ mode on Windows by setting the
1921 makefile macro "USE_CPLUSPLUS" to the value "define".
1922 • The list form of piped open has been implemented for Win32.
1924 Note: unlike "system LIST" this does not fall back to the
1925 shell. [GH #13574]
1926 <https://github.com/Perl/perl5/issues/13574>
1927 • New "DebugSymbols" and "DebugFull" configuration options added
1929 to Windows makefiles.
1930 • Previously, compiling XS modules (including CPAN ones) using
1932 Visual C++ for Win64 resulted in around a dozen warnings per
1933 file from hv_func.h. These warnings have been silenced.
1934 • Support for building without PerlIO has been removed from the
1936 Windows makefiles. Non-PerlIO builds were all but deprecated
1937 in Perl 5.18.0 and are already not supported by Configure on
1938 POSIX systems.
1939 • Between 2 and 6 milliseconds and seven I/O calls have been
1941 saved per attempt to open a perl module for each path in @INC.
1942 • Intel C builds are now always built with C99 mode on.
1944 • %I64d is now being used instead of %lld for MinGW.
1946 • In the experimental ":win32" layer, a crash in "open" was
1948 fixed. Also opening /dev/null (which works under Win32 Perl's
1949 default ":unix" layer) was implemented for ":win32". [GH
1950 #13968] <https://github.com/Perl/perl5/issues/13968>
1951 • A new makefile option, "USE_LONG_DOUBLE", has been added to the
1953 Windows dmake makefile for gcc builds only. Set this to
1954 "define" if you want perl to use long doubles to give more
1955 accuracy and range for floating point numbers.
1956 OpenBSD
1958 On OpenBSD, Perl will now default to using the system "malloc" due
1959 to the security features it provides. Perl's own malloc wrapper has
1960 been in use since v5.14 due to performance reasons, but the OpenBSD
1961 project believes the tradeoff is worth it and would prefer that
1962 users who need the speed specifically ask for it.
1963 [GH #13888] <https://github.com/Perl/perl5/issues/13888>.
1965 Solaris
1967 • We now look for the Sun Studio compiler in both /opt/solstudio*
1968 and /opt/solarisstudio*.
1969 • Builds on Solaris 10 with "-Dusedtrace" would fail early since
1971 make didn't follow implied dependencies to build
1972 "perldtrace.h". Added an explicit dependency to "depend". [GH
1973 #13334] <https://github.com/Perl/perl5/issues/13334>
1974 • C99 options have been cleaned up; hints look for "solstudio" as
1976 well as "SUNWspro"; and support for native "setenv" has been
1977 added.
1978
1980 • Experimental support has been added to allow ops in the optree to
1981 locate their parent, if any. This is enabled by the non-default
1982 build option "-DPERL_OP_PARENT". It is envisaged that this will
1983 eventually become enabled by default, so XS code which directly
1984 accesses the "op_sibling" field of ops should be updated to be
1985 future-proofed.
1986 On "PERL_OP_PARENT" builds, the "op_sibling" field has been renamed
1988 "op_sibparent" and a new flag, "op_moresib", added. On the last op
1989 in a sibling chain, "op_moresib" is false and "op_sibparent" points
1990 to the parent (if any) rather than being "NULL".
1991 To make existing code work transparently whether using
1993 "PERL_OP_PARENT" or not, a number of new macros and functions have
1994 been added that should be used, rather than directly manipulating
1995 "op_sibling".
1996 For the case of just reading "op_sibling" to determine the next
1998 sibling, two new macros have been added. A simple scan through a
1999 sibling chain like this:
2000 for (; kid->op_sibling; kid = kid->op_sibling) { ... }
2002 should now be written as:
2004 for (; OpHAS_SIBLING(kid); kid = OpSIBLING(kid)) { ... }
2006 For altering optrees, a general-purpose function
2008 "op_sibling_splice()" has been added, which allows for manipulation
2009 of a chain of sibling ops. By analogy with the Perl function
2010 "splice()", it allows you to cut out zero or more ops from a
2011 sibling chain and replace them with zero or more new ops. It
2012 transparently handles all the updating of sibling, parent, op_last
2013 pointers etc.
2014 If you need to manipulate ops at a lower level, then three new
2016 macros, "OpMORESIB_set", "OpLASTSIB_set" and "OpMAYBESIB_set" are
2017 intended to be a low-level portable way to set "op_sibling" /
2018 "op_sibparent" while also updating "op_moresib". The first sets
2019 the sibling pointer to a new sibling, the second makes the op the
2020 last sibling, and the third conditionally does the first or second
2021 action. Note that unlike "op_sibling_splice()" these macros won't
2022 maintain consistency in the parent at the same time (e.g. by
2023 updating "op_first" and "op_last" where appropriate).
2024 A C-level "Perl_op_parent()" function and a Perl-level
2026 "B::OP::parent()" method have been added. The C function only
2027 exists under "PERL_OP_PARENT" builds (using it is build-time error
2028 on vanilla perls). "B::OP::parent()" exists always, but on a
2029 vanilla build it always returns "NULL". Under "PERL_OP_PARENT",
2030 they return the parent of the current op, if any. The variable
2031 $B::OP::does_parent allows you to determine whether "B" supports
2032 retrieving an op's parent.
2033 "PERL_OP_PARENT" was introduced in 5.21.2, but the interface was
2035 changed considerably in 5.21.11. If you updated your code before
2036 the 5.21.11 changes, it may require further revision. The main
2037 changes after 5.21.2 were:
2038 • The "OP_SIBLING" and "OP_HAS_SIBLING" macros have been renamed
2040 "OpSIBLING" and "OpHAS_SIBLING" for consistency with other op-
2041 manipulating macros.
2042 • The "op_lastsib" field has been renamed "op_moresib", and its
2044 meaning inverted.
2045 • The macro "OpSIBLING_set" has been removed, and has been
2047 superseded by "OpMORESIB_set" et al.
2048 • The "op_sibling_splice()" function now accepts a null "parent"
2050 argument where the splicing doesn't affect the first or last
2051 ops in the sibling chain
2052 • Macros have been created to allow XS code to better manipulate the
2054 POSIX locale category "LC_NUMERIC". See "Locale-related functions
2055 and macros" in perlapi.
2056 • The previous "atoi" et al replacement function, "grok_atou", has
2058 now been superseded by "grok_atoUV". See perlclib for details.
2059 • A new function, "Perl_sv_get_backrefs()", has been added which
2061 allows you retrieve the weak references, if any, which point at an
2062 SV.
2063 • The "screaminstr()" function has been removed. Although marked as
2065 public API, it was undocumented and had no usage in CPAN modules.
2066 Calling it has been fatal since 5.17.0.
2067 • The "newDEFSVOP()", "block_start()", "block_end()" and "intro_my()"
2069 functions have been added to the API.
2070 • The internal "convert" function in op.c has been renamed
2072 "op_convert_list" and added to the API.
2073 • The "sv_magic()" function no longer forbids "ext" magic on read-
2075 only values. After all, perl can't know whether the custom magic
2076 will modify the SV or not. [GH #14202]
2077 <https://github.com/Perl/perl5/issues/14202>.
2078 • Accessing "CvPADLIST" in perlapi on an XSUB is now forbidden.
2080 The "CvPADLIST" field has been reused for a different internal
2082 purpose for XSUBs. So in particular, you can no longer rely on it
2083 being NULL as a test of whether a CV is an XSUB. Use "CvISXSUB()"
2084 instead.
2085 • SVs of type "SVt_NV" are now sometimes bodiless when the build
2087 configuration and platform allow it: specifically, when "sizeof(NV)
2088 <= sizeof(IV)". "Bodiless" means that the NV value is stored
2089 directly in the head of an SV, without requiring a separate body to
2090 be allocated. This trick has already been used for IVs since 5.9.2
2091 (though in the case of IVs, it is always used, regardless of
2092 platform and build configuration).
2093 • The $DB::single, $DB::signal and $DB::trace variables now have set-
2095 and get-magic that stores their values as IVs, and those IVs are
2096 used when testing their values in "pp_dbstate()". This prevents
2097 perl from recursing infinitely if an overloaded object is assigned
2098 to any of those variables. [GH #14013]
2099 <https://github.com/Perl/perl5/issues/14013>.
2100 • "Perl_tmps_grow()", which is marked as public API but is
2102 undocumented, has been removed from the public API. This change
2103 does not affect XS code that uses the "EXTEND_MORTAL" macro to pre-
2104 extend the mortal stack.
2105 • Perl's internals no longer sets or uses the "SVs_PADMY" flag.
2107 "SvPADMY()" now returns a true value for anything not marked
2108 "PADTMP" and "SVs_PADMY" is now defined as 0.
2109 • The macros "SETsv" and "SETsvUN" have been removed. They were no
2111 longer used in the core since commit 6f1401dc2a five years ago, and
2112 have not been found present on CPAN.
2113 • The "SvFAKE" bit (unused on HVs) got informally reserved by David
2115 Mitchell for future work on vtables.
2116 • The "sv_catpvn_flags()" function accepts "SV_CATBYTES" and
2118 "SV_CATUTF8" flags, which specify whether the appended string is
2119 bytes or UTF-8, respectively. (These flags have in fact been
2120 present since 5.16.0, but were formerly not regarded as part of the
2121 API.)
2122 • A new opcode class, "METHOP", has been introduced. It holds
2124 information used at runtime to improve the performance of
2125 class/object method calls.
2126 "OP_METHOD" and "OP_METHOD_NAMED" have changed from being
2128 "UNOP/SVOP" to being "METHOP".
2129 • "cv_name()" is a new API function that can be passed a CV or GV.
2131 It returns an SV containing the name of the subroutine, for use in
2132 diagnostics.
2133 [GH #12767] <https://github.com/Perl/perl5/issues/12767> [GH
2135 #13392] <https://github.com/Perl/perl5/issues/13392>
2136 • "cv_set_call_checker_flags()" is a new API function that works like
2138 "cv_set_call_checker()", except that it allows the caller to
2139 specify whether the call checker requires a full GV for reporting
2140 the subroutine's name, or whether it could be passed a CV instead.
2141 Whatever value is passed will be acceptable to "cv_name()".
2142 "cv_set_call_checker()" guarantees there will be a GV, but it may
2143 have to create one on the fly, which is inefficient. [GH #12767]
2144 <https://github.com/Perl/perl5/issues/12767>
2145 • "CvGV" (which is not part of the API) is now a more complex macro,
2147 which may call a function and reify a GV. For those cases where it
2148 has been used as a boolean, "CvHASGV" has been added, which will
2149 return true for CVs that notionally have GVs, but without reifying
2150 the GV. "CvGV" also returns a GV now for lexical subs. [GH
2151 #13392] <https://github.com/Perl/perl5/issues/13392>
2152 • The "sync_locale" in perlapi function has been added to the public
2154 API. Changing the program's locale should be avoided by XS code.
2155 Nevertheless, certain non-Perl libraries called from XS need to do
2156 so, such as "Gtk". When this happens, Perl needs to be told that
2157 the locale has changed. Use this function to do so, before
2158 returning to Perl.
2159 • The defines and labels for the flags in the "op_private" field of
2161 OPs are now auto-generated from data in regen/op_private. The
2162 noticeable effect of this is that some of the flag output of
2163 "Concise" might differ slightly, and the flag output of "perl -Dx"
2164 may differ considerably (they both use the same set of labels now).
2165 Also, debugging builds now have a new assertion in "op_free()" to
2166 ensure that the op doesn't have any unrecognized flags set in
2167 "op_private".
2168 • The deprecated variable "PL_sv_objcount" has been removed.
2170 • Perl now tries to keep the locale category "LC_NUMERIC" set to "C"
2172 except around operations that need it to be set to the program's
2173 underlying locale. This protects the many XS modules that cannot
2174 cope with the decimal radix character not being a dot. Prior to
2175 this release, Perl initialized this category to "C", but a call to
2176 "POSIX::setlocale()" would change it. Now such a call will change
2177 the underlying locale of the "LC_NUMERIC" category for the program,
2178 but the locale exposed to XS code will remain "C". There are new
2179 macros to manipulate the LC_NUMERIC locale, including
2180 "STORE_LC_NUMERIC_SET_TO_NEEDED" and
2181 "STORE_LC_NUMERIC_FORCE_TO_UNDERLYING". See "Locale-related
2182 functions and macros" in perlapi.
2183 • A new macro "isUTF8_CHAR" has been written which efficiently
2185 determines if the string given by its parameters begins with a
2186 well-formed UTF-8 encoded character.
2187 • The following private API functions had their context parameter
2189 removed: "Perl_cast_ulong", "Perl_cast_i32", "Perl_cast_iv",
2190 "Perl_cast_uv", "Perl_cv_const_sv", "Perl_mg_find",
2191 "Perl_mg_findext", "Perl_mg_magical", "Perl_mini_mktime",
2192 "Perl_my_dirfd", "Perl_sv_backoff", "Perl_utf8_hop".
2193 Note that the prefix-less versions of those functions that are part
2195 of the public API, such as "cast_i32()", remain unaffected.
2196 • The "PADNAME" and "PADNAMELIST" types are now separate types, and
2198 no longer simply aliases for SV and AV. [GH #14250]
2199 <https://github.com/Perl/perl5/issues/14250>.
2200 • Pad names are now always UTF-8. The "PadnameUTF8" macro always
2202 returns true. Previously, this was effectively the case already,
2203 but any support for two different internal representations of pad
2204 names has now been removed.
2205 • A new op class, "UNOP_AUX", has been added. This is a subclass of
2207 "UNOP" with an "op_aux" field added, which points to an array of
2208 unions of UV, SV* etc. It is intended for where an op needs to
2209 store more data than a simple "op_sv" or whatever. Currently the
2210 only op of this type is "OP_MULTIDEREF" (see next item).
2211 • A new op has been added, "OP_MULTIDEREF", which performs one or
2213 more nested array and hash lookups where the key is a constant or
2214 simple variable. For example the expression $a[0]{$k}[$i], which
2215 previously involved ten "rv2Xv", "Xelem", "gvsv" and "const" ops is
2216 now performed by a single "multideref" op. It can also handle
2217 "local", "exists" and "delete". A non-simple index expression, such
2218 as "[$i+1]" is still done using "aelem"/"helem", and single-level
2219 array lookup with a small constant index is still done using
2220 "aelemfast".
2221
2223 • "close" now sets $!
2224 When an I/O error occurs, the fact that there has been an error is
2226 recorded in the handle. "close" returns false for such a handle.
2227 Previously, the value of $! would be untouched by "close", so the
2228 common convention of writing "close $fh or die $!" did not work
2229 reliably. Now the handle records the value of $!, too, and "close"
2230 restores it.
2231 • "no re" now can turn off everything that "use re" enables
2233 Previously, running "no re" would turn off only a few things. Now
2235 it can turn off all the enabled things. For example, the only way
2236 to stop debugging, once enabled, was to exit the enclosing block;
2237 that is now fixed.
2238 • "pack("D", $x)" and "pack("F", $x)" now zero the padding on x86
2240 long double builds. Under some build options on GCC 4.8 and later,
2241 they used to either overwrite the zero-initialized padding, or
2242 bypass the initialized buffer entirely. This caused op/pack.t to
2243 fail. [GH #14554] <https://github.com/Perl/perl5/issues/14554>
2244 • Extending an array cloned from a parent thread could result in
2246 "Modification of a read-only value attempted" errors when
2247 attempting to modify the new elements. [GH #14605]
2248 <https://github.com/Perl/perl5/issues/14605>
2249 • An assertion failure and subsequent crash with "*x=<y>" has been
2251 fixed. [GH #14493] <https://github.com/Perl/perl5/issues/14493>
2252 • A possible crashing/looping bug related to compiling lexical subs
2254 has been fixed. [GH #14596]
2255 <https://github.com/Perl/perl5/issues/14596>
2256 • UTF-8 now works correctly in function names, in unquoted HERE-
2258 document terminators, and in variable names used as array indexes.
2259 [GH #14601] <https://github.com/Perl/perl5/issues/14601>
2260 • Repeated global pattern matches in scalar context on large tainted
2262 strings were exponentially slow depending on the current match
2263 position in the string. [GH #14238]
2264 <https://github.com/Perl/perl5/issues/14238>
2265 • Various crashes due to the parser getting confused by syntax errors
2267 have been fixed. [GH #14496]
2268 <https://github.com/Perl/perl5/issues/14496> [GH #14497]
2269 <https://github.com/Perl/perl5/issues/14497> [GH #14548]
2270 <https://github.com/Perl/perl5/issues/14548> [GH #14564]
2271 <https://github.com/Perl/perl5/issues/14564>
2272 • "split" in the scope of lexical $_ has been fixed not to fail
2274 assertions. [GH #14483]
2275 <https://github.com/Perl/perl5/issues/14483>
2276 • "my $x : attr" syntax inside various list operators no longer fails
2278 assertions. [GH #14500]
2279 <https://github.com/Perl/perl5/issues/14500>
2280 • An "@" sign in quotes followed by a non-ASCII digit (which is not a
2282 valid identifier) would cause the parser to crash, instead of
2283 simply trying the "@" as literal. This has been fixed. [GH
2284 #14553] <https://github.com/Perl/perl5/issues/14553>
2285 • "*bar::=*foo::=*glob_with_hash" has been crashing since Perl 5.14,
2287 but no longer does. [GH #14512]
2288 <https://github.com/Perl/perl5/issues/14512>
2289 • "foreach" in scalar context was not pushing an item on to the
2291 stack, resulting in bugs.
2292 ("print 4, scalar do { foreach(@x){} } + 1" would print 5.) It has
2293 been fixed to return "undef". [GH #14569]
2294 <https://github.com/Perl/perl5/issues/14569>
2295 • Several cases of data used to store environment variable contents
2297 in core C code being potentially overwritten before being used have
2298 been fixed. [GH #14476]
2299 <https://github.com/Perl/perl5/issues/14476>
2300 • Some patterns starting with "/.*..../" matched against long strings
2302 have been slow since v5.8, and some of the form "/.*..../i" have
2303 been slow since v5.18. They are now all fast again. [GH #14475]
2304 <https://github.com/Perl/perl5/issues/14475>.
2305 • The original visible value of $/ is now preserved when it is set to
2307 an invalid value. Previously if you set $/ to a reference to an
2308 array, for example, perl would produce a runtime error and not set
2309 "PL_rs", but Perl code that checked $/ would see the array
2310 reference. [GH #14245]
2311 <https://github.com/Perl/perl5/issues/14245>.
2312 • In a regular expression pattern, a POSIX class, like "[:ascii:]",
2314 must be inside a bracketed character class, like "qr/[[:ascii:]]/".
2315 A warning is issued when something looking like a POSIX class is
2316 not inside a bracketed class. That warning wasn't getting
2317 generated when the POSIX class was negated: "[:^ascii:]". This is
2318 now fixed.
2319 • Perl 5.14.0 introduced a bug whereby "eval { LABEL: }" would crash.
2321 This has been fixed. [GH #14438]
2322 <https://github.com/Perl/perl5/issues/14438>.
2323 • Various crashes due to the parser getting confused by syntax errors
2325 have been fixed. [GH #14421]
2326 <https://github.com/Perl/perl5/issues/14421>. [GH #14472]
2327 <https://github.com/Perl/perl5/issues/14472>. [GH #14480]
2328 <https://github.com/Perl/perl5/issues/14480>. [GH #14447]
2329 <https://github.com/Perl/perl5/issues/14447>.
2330 • Code like "/$a[/" used to read the next line of input and treat it
2332 as though it came immediately after the opening bracket. Some
2333 invalid code consequently would parse and run, but some code caused
2334 crashes, so this is now disallowed. [GH #14462]
2335 <https://github.com/Perl/perl5/issues/14462>.
2336 • Fix argument underflow for "pack". [GH #14525]
2338 <https://github.com/Perl/perl5/issues/14525>.
2339 • Fix handling of non-strict "\x{}". Now "\x{}" is equivalent to
2341 "\x{0}" instead of faulting.
2342 • "stat -t" is now no longer treated as stackable, just like "-t
2344 stat". [GH #14499] <https://github.com/Perl/perl5/issues/14499>.
2345 • The following no longer causes a SEGV: "qr{x+(y(?0))*}".
2347 • Fixed infinite loop in parsing backrefs in regexp patterns.
2349 • Several minor bug fixes in behavior of Infinity and NaN, including
2351 warnings when stringifying Infinity-like or NaN-like strings. For
2352 example, "NaNcy" doesn't numify to NaN anymore.
2353 • A bug in regular expression patterns that could lead to segfaults
2355 and other crashes has been fixed. This occurred only in patterns
2356 compiled with "/i" while taking into account the current POSIX
2357 locale (which usually means they have to be compiled within the
2358 scope of "use locale"), and there must be a string of at least 128
2359 consecutive bytes to match. [GH #14389]
2360 <https://github.com/Perl/perl5/issues/14389>.
2361 • "s///g" now works on very long strings (where there are more than 2
2363 billion iterations) instead of dying with 'Substitution loop'. [GH
2364 #11742] <https://github.com/Perl/perl5/issues/11742>. [GH #14190]
2365 <https://github.com/Perl/perl5/issues/14190>.
2366 • "gmtime" no longer crashes with not-a-number values. [GH #14365]
2368 <https://github.com/Perl/perl5/issues/14365>.
2369 • "\()" (a reference to an empty list), and "y///" with lexical $_ in
2371 scope, could both do a bad write past the end of the stack. They
2372 have both been fixed to extend the stack first.
2373 • "prototype()" with no arguments used to read the previous item on
2375 the stack, so "print "foo", prototype()" would print foo's
2376 prototype. It has been fixed to infer $_ instead. [GH #14376]
2377 <https://github.com/Perl/perl5/issues/14376>.
2378 • Some cases of lexical state subs declared inside predeclared subs
2380 could crash, for example when evalling a string including the name
2381 of an outer variable, but no longer do.
2382 • Some cases of nested lexical state subs inside anonymous subs could
2384 cause 'Bizarre copy' errors or possibly even crashes.
2385 • When trying to emit warnings, perl's default debugger (perl5db.pl)
2387 was sometimes giving 'Undefined subroutine &DB::db_warn called'
2388 instead. This bug, which started to occur in Perl 5.18, has been
2389 fixed. [GH #14400] <https://github.com/Perl/perl5/issues/14400>.
2390 • Certain syntax errors in substitutions, such as "s/${<>{})//",
2392 would crash, and had done so since Perl 5.10. (In some cases the
2393 crash did not start happening till 5.16.) The crash has, of
2394 course, been fixed. [GH #14391]
2395 <https://github.com/Perl/perl5/issues/14391>.
2396 • Fix a couple of string grow size calculation overflows; in
2398 particular, a repeat expression like "33 x ~3" could cause a large
2399 buffer overflow since the new output buffer size was not correctly
2400 handled by "SvGROW()". An expression like this now properly
2401 produces a memory wrap panic. [GH #14401]
2402 <https://github.com/Perl/perl5/issues/14401>.
2403 • "formline("@...", "a");" would crash. The "FF_CHECKNL" case in
2405 "pp_formline()" didn't set the pointer used to mark the chop
2406 position, which led to the "FF_MORE" case crashing with a
2407 segmentation fault. This has been fixed. [GH #14388]
2408 <https://github.com/Perl/perl5/issues/14388>.
2409 • A possible buffer overrun and crash when parsing a literal pattern
2411 during regular expression compilation has been fixed. [GH #14416]
2412 <https://github.com/Perl/perl5/issues/14416>.
2413 • "fchmod()" and "futimes()" now set $! when they fail due to being
2415 passed a closed file handle. [GH #14073]
2416 <https://github.com/Perl/perl5/issues/14073>.
2417 • "op_free()" and "scalarvoid()" no longer crash due to a stack
2419 overflow when freeing a deeply recursive op tree. [GH #11866]
2420 <https://github.com/Perl/perl5/issues/11866>.
2421 • In Perl 5.20.0, $^N accidentally had the internal UTF-8 flag turned
2423 off if accessed from a code block within a regular expression,
2424 effectively UTF-8-encoding the value. This has been fixed. [GH
2425 #14211] <https://github.com/Perl/perl5/issues/14211>.
2426 • A failed "semctl" call no longer overwrites existing items on the
2428 stack, which means that "(semctl(-1,0,0,0))[0]" no longer gives an
2429 "uninitialized" warning.
2430 • "else{foo()}" with no space before "foo" is now better at assigning
2432 the right line number to that statement. [GH #14070]
2433 <https://github.com/Perl/perl5/issues/14070>.
2434 • Sometimes the assignment in "@array = split" gets optimised so that
2436 "split" itself writes directly to the array. This caused a bug,
2437 preventing this assignment from being used in lvalue context. So
2438 "(@a=split//,"foo")=bar()" was an error. (This bug probably goes
2439 back to Perl 3, when the optimisation was added.) It has now been
2440 fixed. [GH #14183] <https://github.com/Perl/perl5/issues/14183>.
2441 • When an argument list fails the checks specified by a subroutine
2443 signature (which is still an experimental feature), the resulting
2444 error messages now give the file and line number of the caller, not
2445 of the called subroutine. [GH #13643]
2446 <https://github.com/Perl/perl5/issues/13643>.
2447 • The flip-flop operators (".." and "..." in scalar context) used to
2449 maintain a separate state for each recursion level (the number of
2450 times the enclosing sub was called recursively), contrary to the
2451 documentation. Now each closure has one internal state for each
2452 flip-flop. [GH #14110]
2453 <https://github.com/Perl/perl5/issues/14110>.
2454 • The flip-flop operator (".." in scalar context) would return the
2456 same scalar each time, unless the containing subroutine was called
2457 recursively. Now it always returns a new scalar. [GH #14110]
2458 <https://github.com/Perl/perl5/issues/14110>.
2459 • "use", "no", statement labels, special blocks ("BEGIN") and pod are
2461 now permitted as the first thing in a "map" or "grep" block, the
2462 block after "print" or "say" (or other functions) returning a
2463 handle, and within "${...}", "@{...}", etc. [GH #14088]
2464 <https://github.com/Perl/perl5/issues/14088>.
2465 • The repetition operator "x" now propagates lvalue context to its
2467 left-hand argument when used in contexts like "foreach". That
2468 allows "for(($#that_array)x2) { ... }" to work as expected if the
2469 loop modifies $_.
2470 • "(...) x ..." in scalar context used to corrupt the stack if one
2472 operand was an object with "x" overloading, causing erratic
2473 behavior. [GH #13811]
2474 <https://github.com/Perl/perl5/issues/13811>.
2475 • Assignment to a lexical scalar is often optimised away; for example
2477 in "my $x; $x = $y + $z", the assign operator is optimised away and
2478 the add operator writes its result directly to $x. Various bugs
2479 related to this optimisation have been fixed. Certain operators on
2480 the right-hand side would sometimes fail to assign the value at all
2481 or assign the wrong value, or would call STORE twice or not at all
2482 on tied variables. The operators affected were "$foo++", "$foo--",
2483 and "-$foo" under "use integer", "chomp", "chr" and "setpgrp".
2484 • List assignments were sometimes buggy if the same scalar ended up
2486 on both sides of the assignment due to use of "tied", "values" or
2487 "each". The result would be the wrong value getting assigned.
2488 • "setpgrp($nonzero)" (with one argument) was accidentally changed in
2490 5.16 to mean setpgrp(0). This has been fixed.
2491 • "__SUB__" could return the wrong value or even corrupt memory under
2493 the debugger (the "-d" switch) and in subs containing "eval
2494 $string".
2495 • When "sub () { $var }" becomes inlinable, it now returns a
2497 different scalar each time, just as a non-inlinable sub would,
2498 though Perl still optimises the copy away in cases where it would
2499 make no observable difference.
2500 • "my sub f () { $var }" and "sub () : attr { $var }" are no longer
2502 eligible for inlining. The former would crash; the latter would
2503 just throw the attributes away. An exception is made for the
2504 little-known ":method" attribute, which does nothing much.
2505 • Inlining of subs with an empty prototype is now more consistent
2507 than before. Previously, a sub with multiple statements, of which
2508 all but the last were optimised away, would be inlinable only if it
2509 were an anonymous sub containing a string "eval" or "state"
2510 declaration or closing over an outer lexical variable (or any
2511 anonymous sub under the debugger). Now any sub that gets folded to
2512 a single constant after statements have been optimised away is
2513 eligible for inlining. This applies to things like "sub () {
2514 jabber() if DEBUG; 42 }".
2515 Some subroutines with an explicit "return" were being made
2517 inlinable, contrary to the documentation, Now "return" always
2518 prevents inlining.
2519 • On some systems, such as VMS, "crypt" can return a non-ASCII
2521 string. If a scalar assigned to had contained a UTF-8 string
2522 previously, then "crypt" would not turn off the UTF-8 flag, thus
2523 corrupting the return value. This would happen with
2524 "$lexical = crypt ...".
2525 • "crypt" no longer calls "FETCH" twice on a tied first argument.
2527 • An unterminated here-doc on the last line of a quote-like operator
2529 ("qq[${ <<END }]", "/(?{ <<END })/") no longer causes a double
2530 free. It started doing so in 5.18.
2531 • "index()" and "rindex()" no longer crash when used on strings over
2533 2GB in size. [GH #13700]
2534 <https://github.com/Perl/perl5/issues/13700>.
2535 • A small, previously intentional, memory leak in
2537 "PERL_SYS_INIT"/"PERL_SYS_INIT3" on Win32 builds was fixed. This
2538 might affect embedders who repeatedly create and destroy perl
2539 engines within the same process.
2540 • "POSIX::localeconv()" now returns the data for the program's
2542 underlying locale even when called from outside the scope of
2543 "use locale".
2544 • "POSIX::localeconv()" now works properly on platforms which don't
2546 have "LC_NUMERIC" and/or "LC_MONETARY", or for which Perl has been
2547 compiled to disregard either or both of these locale categories.
2548 In such circumstances, there are now no entries for the
2549 corresponding values in the hash returned by "localeconv()".
2550 • "POSIX::localeconv()" now marks appropriately the values it returns
2552 as UTF-8 or not. Previously they were always returned as bytes,
2553 even if they were supposed to be encoded as UTF-8.
2554 • On Microsoft Windows, within the scope of "use locale", the
2556 following POSIX character classes gave results for many locales
2557 that did not conform to the POSIX standard: "[[:alnum:]]",
2558 "[[:alpha:]]", "[[:blank:]]", "[[:digit:]]", "[[:graph:]]",
2559 "[[:lower:]]", "[[:print:]]", "[[:punct:]]", "[[:upper:]]",
2560 "[[:word:]]", and "[[:xdigit:]]". This was because the underlying
2561 Microsoft implementation does not follow the standard. Perl now
2562 takes special precautions to correct for this.
2563 • Many issues have been detected by Coverity
2565 <http://www.coverity.com/> and fixed.
2566 • "system()" and friends should now work properly on more Android
2568 builds.
2569 Due to an oversight, the value specified through "-Dtargetsh" to
2571 Configure would end up being ignored by some of the build process.
2572 This caused perls cross-compiled for Android to end up with
2573 defective versions of "system()", "exec()" and backticks: the
2574 commands would end up looking for "/bin/sh" instead of
2575 "/system/bin/sh", and so would fail for the vast majority of
2576 devices, leaving $! as "ENOENT".
2577 • "qr(...\(...\)...)", "qr[...\[...\]...]", and "qr{...\{...\}...}"
2579 now work. Previously it was impossible to escape these three left-
2580 characters with a backslash within a regular expression pattern
2581 where otherwise they would be considered metacharacters, and the
2582 pattern opening delimiter was the character, and the closing
2583 delimiter was its mirror character.
2584 • "s///e" on tainted UTF-8 strings corrupted "pos()". This bug,
2586 introduced in 5.20, is now fixed. [GH #13948]
2587 <https://github.com/Perl/perl5/issues/13948>.
2588 • A non-word boundary in a regular expression ("\B") did not always
2590 match the end of the string; in particular "q{} =~ /\B/" did not
2591 match. This bug, introduced in perl 5.14, is now fixed. [GH
2592 #13917] <https://github.com/Perl/perl5/issues/13917>.
2593 • "" P" =~ /(?=.*P)P/" should match, but did not. This is now fixed.
2595 [GH #13954] <https://github.com/Perl/perl5/issues/13954>.
2596 • Failing to compile "use Foo" in an "eval" could leave a spurious
2598 "BEGIN" subroutine definition, which would produce a "Subroutine
2599 BEGIN redefined" warning on the next use of "use", or other "BEGIN"
2600 block. [GH #13926] <https://github.com/Perl/perl5/issues/13926>.
2601 • "method { BLOCK } ARGS" syntax now correctly parses the arguments
2603 if they begin with an opening brace. [GH #9085]
2604 <https://github.com/Perl/perl5/issues/9085>.
2605 • External libraries and Perl may have different ideas of what the
2607 locale is. This is problematic when parsing version strings if the
2608 locale's numeric separator has been changed. Version parsing has
2609 been patched to ensure it handles the locales correctly. [GH
2610 #13863] <https://github.com/Perl/perl5/issues/13863>.
2611 • A bug has been fixed where zero-length assertions and code blocks
2613 inside of a regex could cause "pos" to see an incorrect value. [GH
2614 #14016] <https://github.com/Perl/perl5/issues/14016>.
2615 • Dereferencing of constants now works correctly for typeglob
2617 constants. Previously the glob was stringified and its name looked
2618 up. Now the glob itself is used. [GH #9891]
2619 <https://github.com/Perl/perl5/issues/9891>
2620 • When parsing a sigil ("$" "@" "%" "&)" followed by braces, the
2622 parser no longer tries to guess whether it is a block or a hash
2623 constructor (causing a syntax error when it guesses the latter),
2624 since it can only be a block.
2625 • "undef $reference" now frees the referent immediately, instead of
2627 hanging on to it until the next statement. [GH #14032]
2628 <https://github.com/Perl/perl5/issues/14032>
2629 • Various cases where the name of a sub is used (autoload,
2631 overloading, error messages) used to crash for lexical subs, but
2632 have been fixed.
2633 • Bareword lookup now tries to avoid vivifying packages if it turns
2635 out the bareword is not going to be a subroutine name.
2636 • Compilation of anonymous constants (e.g., "sub () { 3 }") no longer
2638 deletes any subroutine named "__ANON__" in the current package.
2639 Not only was "*__ANON__{CODE}" cleared, but there was a memory
2640 leak, too. This bug goes back to Perl 5.8.0.
2641 • Stub declarations like "sub f;" and "sub f ();" no longer wipe out
2643 constants of the same name declared by "use constant". This bug
2644 was introduced in Perl 5.10.0.
2645 • "qr/[\N{named sequence}]/" now works properly in many instances.
2647 Some names known to "\N{...}" refer to a sequence of multiple
2649 characters, instead of the usual single character. Bracketed
2650 character classes generally only match single characters, but now
2651 special handling has been added so that they can match named
2652 sequences, but not if the class is inverted or the sequence is
2653 specified as the beginning or end of a range. In these cases, the
2654 only behavior change from before is a slight rewording of the fatal
2655 error message given when this class is part of a "?[...])"
2656 construct. When the "[...]" stands alone, the same non-fatal
2657 warning as before is raised, and only the first character in the
2658 sequence is used, again just as before.
2659 • Tainted constants evaluated at compile time no longer cause
2661 unrelated statements to become tainted. [GH #14059]
2662 <https://github.com/Perl/perl5/issues/14059>
2663 • "open $$fh, ...", which vivifies a handle with a name like
2665 "main::_GEN_0", was not giving the handle the right reference
2666 count, so a double free could happen.
2667 • When deciding that a bareword was a method name, the parser would
2669 get confused if an "our" sub with the same name existed, and look
2670 up the method in the package of the "our" sub, instead of the
2671 package of the invocant.
2672 • The parser no longer gets confused by "\U=" within a double-quoted
2674 string. It used to produce a syntax error, but now compiles it
2675 correctly. [GH #10882]
2676 <https://github.com/Perl/perl5/issues/10882>
2677 • It has always been the intention for the "-B" and "-T" file test
2679 operators to treat UTF-8 encoded files as text. (perlfunc has been
2680 updated to say this.) Previously, it was possible for some files
2681 to be considered UTF-8 that actually weren't valid UTF-8. This is
2682 now fixed. The operators now work on EBCDIC platforms as well.
2683 • Under some conditions warning messages raised during regular
2685 expression pattern compilation were being output more than once.
2686 This has now been fixed.
2687 • Perl 5.20.0 introduced a regression in which a UTF-8 encoded
2689 regular expression pattern that contains a single ASCII lowercase
2690 letter did not match its uppercase counterpart. That has been fixed
2691 in both 5.20.1 and 5.22.0. [GH #14051]
2692 <https://github.com/Perl/perl5/issues/14051>
2693 • Constant folding could incorrectly suppress warnings if lexical
2695 warnings ("use warnings" or "no warnings") were not in effect and
2696 $^W were false at compile time and true at run time.
2697 • Loading Unicode tables during a regular expression match could
2699 cause assertion failures under debugging builds if the previous
2700 match used the very same regular expression. [GH #14081]
2701 <https://github.com/Perl/perl5/issues/14081>
2702 • Thread cloning used to work incorrectly for lexical subs, possibly
2704 causing crashes or double frees on exit.
2705 • Since Perl 5.14.0, deleting $SomePackage::{__ANON__} and then
2707 undefining an anonymous subroutine could corrupt things internally,
2708 resulting in Devel::Peek crashing or B.pm giving nonsensical data.
2709 This has been fixed.
2710 • "(caller $n)[3]" now reports names of lexical subs, instead of
2712 treating them as "(unknown)".
2713 • "sort subname LIST" now supports using a lexical sub as the
2715 comparison routine.
2716 • Aliasing (e.g., via "*x = *y") could confuse list assignments that
2718 mention the two names for the same variable on either side, causing
2719 wrong values to be assigned. [GH #5788]
2720 <https://github.com/Perl/perl5/issues/5788>
2721 • Long here-doc terminators could cause a bad read on short lines of
2723 input. This has been fixed. It is doubtful that any crash could
2724 have occurred. This bug goes back to when here-docs were
2725 introduced in Perl 3.000 twenty-five years ago.
2726 • An optimization in "split" to treat "split /^/" like "split /^/m"
2728 had the unfortunate side-effect of also treating "split /\A/" like
2729 "split /^/m", which it should not. This has been fixed. (Note,
2730 however, that "split /^x/" does not behave like "split /^x/m",
2731 which is also considered to be a bug and will be fixed in a future
2732 version.) [GH #14086] <https://github.com/Perl/perl5/issues/14086>
2733 • The little-known "my Class $var" syntax (see fields and attributes)
2735 could get confused in the scope of "use utf8" if "Class" were a
2736 constant whose value contained Latin-1 characters.
2737 • Locking and unlocking values via Hash::Util or
2739 "Internals::SvREADONLY" no longer has any effect on values that
2740 were read-only to begin with. Previously, unlocking such values
2741 could result in crashes, hangs or other erratic behavior.
2742 • Some unterminated "(?(...)...)" constructs in regular expressions
2744 would either crash or give erroneous error messages. "/(?(1)/" is
2745 one such example.
2746 • "pack "w", $tied" no longer calls FETCH twice.
2748 • List assignments like "($x, $z) = (1, $y)" now work correctly if $x
2750 and $y have been aliased by "foreach".
2751 • Some patterns including code blocks with syntax errors, such as
2753 "/ (?{(^{})/", would hang or fail assertions on debugging builds.
2754 Now they produce errors.
2755 • An assertion failure when parsing "sort" with debugging enabled has
2757 been fixed. [GH #14087]
2758 <https://github.com/Perl/perl5/issues/14087>.
2759 • "*a = *b; @a = split //, $b[1]" could do a bad read and produce
2761 junk results.
2762 • In "() = @array = split", the "() =" at the beginning no longer
2764 confuses the optimizer into assuming a limit of 1.
2765 • Fatal warnings no longer prevent the output of syntax errors. [GH
2767 #14155] <https://github.com/Perl/perl5/issues/14155>.
2768 • Fixed a NaN double-to-long-double conversion error on VMS. For
2770 quiet NaNs (and only on Itanium, not Alpha) negative infinity
2771 instead of NaN was produced.
2772 • Fixed the issue that caused "make distclean" to incorrectly leave
2774 some files behind. [GH #14108]
2775 <https://github.com/Perl/perl5/issues/14108>.
2776 • AIX now sets the length in "getsockopt" correctly. [GH #13484]
2778 <https://github.com/Perl/perl5/issues/13484>. [cpan #91183]
2779 <https://rt.cpan.org/Ticket/Display.html?id=91183>. [cpan #85570]
2780 <https://rt.cpan.org/Ticket/Display.html?id=85570>.
2781 • The optimization phase of a regexp compilation could run "forever"
2783 and exhaust all memory under certain circumstances; now fixed. [GH
2784 #13984] <https://github.com/Perl/perl5/issues/13984>.
2785 • The test script t/op/crypt.t now uses the SHA-256 algorithm if the
2787 default one is disabled, rather than giving failures. [GH #13715]
2788 <https://github.com/Perl/perl5/issues/13715>.
2789 • Fixed an off-by-one error when setting the size of a shared array.
2791 [GH #14151] <https://github.com/Perl/perl5/issues/14151>.
2792 • Fixed a bug that could cause perl to enter an infinite loop during
2794 compilation. In particular, a while(1) within a sublist, e.g.
2795 sub foo { () = ($a, my $b, ($c, do { while(1) {} })) }
2797 The bug was introduced in 5.20.0 [GH #14165]
2799 <https://github.com/Perl/perl5/issues/14165>.
2800 • On Win32, if a variable was "local"-ized in a pseudo-process that
2802 later forked, restoring the original value in the child pseudo-
2803 process caused memory corruption and a crash in the child pseudo-
2804 process (and therefore the OS process). [GH #8641]
2805 <https://github.com/Perl/perl5/issues/8641>.
2806 • Calling "write" on a format with a "^**" field could produce a
2808 panic in "sv_chop()" if there were insufficient arguments or if the
2809 variable used to fill the field was empty. [GH #14255]
2810 <https://github.com/Perl/perl5/issues/14255>.
2811 • Non-ASCII lexical sub names now appear without trailing junk when
2813 they appear in error messages.
2814 • The "\@" subroutine prototype no longer flattens parenthesized
2816 arrays (taking a reference to each element), but takes a reference
2817 to the array itself. [GH #9111]
2818 <https://github.com/Perl/perl5/issues/9111>.
2819 • A block containing nothing except a C-style "for" loop could
2821 corrupt the stack, causing lists outside the block to lose elements
2822 or have elements overwritten. This could happen with "map {
2823 for(...){...} } ..." and with lists containing "do { for(...){...}
2824 }". [GH #14269] <https://github.com/Perl/perl5/issues/14269>.
2825 • "scalar()" now propagates lvalue context, so that
2827 "for(scalar($#foo)) { ... }" can modify $#foo through $_.
2828 • "qr/@array(?{block})/" no longer dies with "Bizarre copy of ARRAY".
2830 [GH #14292] <https://github.com/Perl/perl5/issues/14292>.
2831 • "eval '$variable'" in nested named subroutines would sometimes look
2833 up a global variable even with a lexical variable in scope.
2834 • In perl 5.20.0, "sort CORE::fake" where 'fake' is anything other
2836 than a keyword, started chopping off the last 6 characters and
2837 treating the result as a sort sub name. The previous behavior of
2838 treating "CORE::fake" as a sort sub name has been restored. [GH
2839 #14323] <https://github.com/Perl/perl5/issues/14323>.
2840 • Outside of "use utf8", a single-character Latin-1 lexical variable
2842 is disallowed. The error message for it, "Can't use global
2843 $foo...", was giving garbage instead of the variable name.
2844 • "readline" on a nonexistent handle was causing "${^LAST_FH}" to
2846 produce a reference to an undefined scalar (or fail an assertion).
2847 Now "${^LAST_FH}" ends up undefined.
2848 • "(...) x ..." in void context now applies scalar context to the
2850 left-hand argument, instead of the context the current sub was
2851 called in. [GH #14174]
2852 <https://github.com/Perl/perl5/issues/14174>.
2853
2855 • "pack"-ing a NaN on a perl compiled with Visual C 6 does not behave
2856 properly, leading to a test failure in t/op/infnan.t. [GH #14705]
2857 <https://github.com/Perl/perl5/issues/14705>
2858 • A goal is for Perl to be able to be recompiled to work reasonably
2860 well on any Unicode version. In Perl 5.22, though, the earliest
2861 such version is Unicode 5.1 (current is 7.0).
2862 • EBCDIC platforms
2864 • The "cmp" (and hence "sort") operators do not necessarily give
2866 the correct results when both operands are UTF-EBCDIC encoded
2867 strings and there is a mixture of ASCII and/or control
2868 characters, along with other characters.
2869 • Ranges containing "\N{...}" in the "tr///" (and "y///")
2871 transliteration operators are treated differently than the
2872 equivalent ranges in regular expression patterns. They should,
2873 but don't, cause the values in the ranges to all be treated as
2874 Unicode code points, and not native ones. ("Version 8 Regular
2875 Expressions" in perlre gives details as to how it should work.)
2876 • Encode and encoding are mostly broken.
2878 • Many CPAN modules that are shipped with core show failing
2880 tests.
2881 • "pack"/"unpack" with "U0" format may not work properly.
2883 • The following modules are known to have test failures with this
2885 version of Perl. In many cases, patches have been submitted, so
2886 there will hopefully be new releases soon:
2887 • B::Generate version 1.50
2889 • B::Utils version 0.25
2891 • Coro version 6.42
2893 • Dancer version 1.3130
2895 • Data::Alias version 1.18
2897 • Data::Dump::Streamer version 2.38
2899 • Data::Util version 0.63
2901 • Devel::Spy version 0.07
2903 • invoker version 0.34
2905 • Lexical::Var version 0.009
2907 • LWP::ConsoleLogger version 0.000018
2909 • Mason version 2.22
2911 • NgxQueue version 0.02
2913 • Padre version 1.00
2915 • Parse::Keyword 0.08
2917
2919 Brian McCauley died on May 8, 2015. He was a frequent poster to
2920 Usenet, Perl Monks, and other Perl forums, and made several CPAN
2921 contributions under the nick NOBULL, including to the Perl FAQ. He
2922 attended almost every YAPC::Europe, and indeed, helped organise
2923 YAPC::Europe 2006 and the QA Hackathon 2009. His wit and his delight
2924 in intricate systems were particularly apparent in his love of board
2925 games; many Perl mongers will have fond memories of playing Fluxx and
2926 other games with Brian. He will be missed.
2927
2929 Perl 5.22.0 represents approximately 12 months of development since
2930 Perl 5.20.0 and contains approximately 590,000 lines of changes across
2931 2,400 files from 94 authors.
2932 Excluding auto-generated files, documentation and release tools, there
2934 were approximately 370,000 lines of changes to 1,500 .pm, .t, .c and .h
2935 files.
2936 Perl continues to flourish into its third decade thanks to a vibrant
2938 community of users and developers. The following people are known to
2939 have contributed the improvements that became Perl 5.22.0:
2940 Aaron Crane, Abhijit Menon-Sen, Abigail, Alberto Simo~es, Alex Solovey,
2942 Alex Vandiver, Alexandr Ciornii, Alexandre (Midnite) Jousset, Andreas
2943 Koenig, Andreas Voegele, Andrew Fresh, Andy Dougherty, Anthony Heading,
2944 Aristotle Pagaltzis, brian d foy, Brian Fraser, Chad Granum, Chris
2945 'BinGOs' Williams, Craig A. Berry, Dagfinn Ilmari Mannsaaker, Daniel
2946 Dragan, Darin McBride, Dave Rolsky, David Golden, David Mitchell, David
2947 Wheeler, Dmitri Tikhonov, Doug Bell, E. Choroba, Ed J, Eric Herman,
2948 Father Chrysostomos, George Greer, Glenn D. Golden, Graham Knop,
2949 H.Merijn Brand, Herbert Breunung, Hugo van der Sanden, James E Keenan,
2950 James McCoy, James Raspass, Jan Dubois, Jarkko Hietaniemi, Jasmine
2951 Ngan, Jerry D. Hedden, Jim Cromie, John Goodyear, kafka, Karen
2952 Etheridge, Karl Williamson, Kent Fredric, kmx, Lajos Veres, Leon
2953 Timmermans, Lukas Mai, Mathieu Arnold, Matthew Horsfall, Max Maischein,
2954 Michael Bunk, Nicholas Clark, Niels Thykier, Niko Tyni, Norman Koch,
2955 Olivier Mengue, Peter John Acklam, Peter Martini, Petr PisaX, Philippe
2956 Bruhat (BooK), Pierre Bogossian, Rafael Garcia-Suarez, Randy Stauner,
2957 Reini Urban, Ricardo Signes, Rob Hoelz, Rostislav Skudnov, Sawyer X,
2958 Shirakata Kentaro, Shlomi Fish, Sisyphus, Slaven Rezic, Smylers,
2959 Steffen Mueller, Steve Hay, Sullivan Beck, syber, Tadeusz SoXnierz,
2960 Thomas Sibley, Todd Rinaldo, Tony Cook, Vincent Pit, Vladimir Marek,
2961 Yaroslav Kuzmin, Yves Orton, AEvar Arnfjoerd` Bjarmason.
2962 The list above is almost certainly incomplete as it is automatically
2964 generated from version control history. In particular, it does not
2965 include the names of the (very much appreciated) contributors who
2966 reported issues to the Perl bug tracker.
2967 Many of the changes included in this version originated in the CPAN
2969 modules included in Perl's core. We're grateful to the entire CPAN
2970 community for helping Perl to flourish.
2971 For a more complete list of all of Perl's historical contributors,
2973 please see the AUTHORS file in the Perl source distribution.
2974
2976 If you find what you think is a bug, you might check the articles
2977 recently posted to the comp.lang.perl.misc newsgroup and the perl bug
2978 database at <https://rt.perl.org/>. There may also be information at
2979 <http://www.perl.org/>, the Perl Home Page.
2980 If you believe you have an unreported bug, please run the perlbug
2982 program included with your release. Be sure to trim your bug down to a
2983 tiny but sufficient test case. Your bug report, along with the output
2984 of "perl -V", will be sent off to perlbug@perl.org to be analysed by
2985 the Perl porting team.
2986 If the bug you are reporting has security implications, which make it
2988 inappropriate to send to a publicly archived mailing list, then please
2989 send it to perl5-security-report@perl.org. This points to a closed
2990 subscription unarchived mailing list, which includes all the core
2991 committers, who will be able to help assess the impact of issues,
2992 figure out a resolution, and help co-ordinate the release of patches to
2993 mitigate or fix the problem across all platforms on which Perl is
2994 supported. Please only use this address for security issues in the
2995 Perl core, not for modules independently distributed on CPAN.
2996
2998 The Changes file for an explanation of how to view exhaustive details
2999 on what changed.
3000 The INSTALL file for how to build Perl.
3002 The README file for general stuff.
3004 The Artistic and Copying files for copyright information.
3006perl v5.36.0 2022-08-30 PERL5220DELTA(1)