{ "mode": "man", "parameter": "perl5220delta", "section": "1", "url": "https://www.chedong.com/phpMan.php/man/perl5220delta/1/json", "generated": "2026-06-13T21:25:31Z", "sections": { "NAME": { "content": "perl5220delta - what is new for perl v5.22.0\n", "subsections": [] }, "DESCRIPTION": { "content": "This document describes differences between the 5.20.0 release and the 5.22.0 release.\n\nIf you are upgrading from an earlier release such as 5.18.0, first read perl5200delta, which\ndescribes differences between 5.18.0 and 5.20.0.\n", "subsections": [ { "name": "Core Enhancements", "content": "" }, { "name": "New bitwise operators", "content": "A new experimental facility has been added that makes the four standard bitwise operators (\"&\n| ^ ~\") treat their operands consistently as numbers, and introduces four new dotted\noperators (\"&. |. ^. ~.\") that treat their operands consistently as strings. The same\napplies to the assignment variants (\"&= |= ^= &.= |.= ^.=\").\n\nTo use this, enable the \"bitwise\" feature and disable the \"experimental::bitwise\" warnings\ncategory. See \"Bitwise String Operators\" in perlop for details. [GH #14348]\n.\n" }, { "name": "New double-diamond operator", "content": "\"<<>>\" is like \"<>\" but uses three-argument \"open\" to open each file in @ARGV. This means\nthat each element of @ARGV will be treated as an actual file name, and \"|foo\" won't be\ntreated as a pipe open.\n" }, { "name": "New \"\\b\" boundaries in regular expressions", "content": "\"qr/\\b{gcb}/\"\n\n\"gcb\" stands for Grapheme Cluster Boundary. It is a Unicode property that finds the boundary\nbetween sequences of characters that look like a single character to a native speaker of a\nlanguage. Perl has long had the ability to deal with these through the \"\\X\" regular escape\nsequence. Now, there is an alternative way of handling these. See \"\\b{}, \\b, \\B{}, \\B\" in\nperlrebackslash for details.\n\n\"qr/\\b{wb}/\"\n\n\"wb\" stands for Word Boundary. It is a Unicode property that finds the boundary between\nwords. This is similar to the plain \"\\b\" (without braces) but is more suitable for natural\nlanguage processing. It knows, for example, that apostrophes can occur in the middle of\nwords. See \"\\b{}, \\b, \\B{}, \\B\" in perlrebackslash for details.\n\n\"qr/\\b{sb}/\"\n\n\"sb\" stands for Sentence Boundary. It is a Unicode property to aid in parsing natural\nlanguage sentences. See \"\\b{}, \\b, \\B{}, \\B\" in perlrebackslash for details.\n" }, { "name": "Non-Capturing Regular Expression Flag", "content": "Regular expressions now support a \"/n\" flag that disables capturing and filling in $1, $2,\netc inside of groups:\n\n\"hello\" =~ /(hi|hello)/n; # $1 is not set\n\nThis is equivalent to putting \"?:\" at the beginning of every capturing group.\n\nSee \"n\" in perlre for more information.\n" }, { "name": "\"use re 'strict'\"", "content": "This applies stricter syntax rules to regular expression patterns compiled within its scope.\nThis will hopefully alert you to typos and other unintentional behavior that backwards-\ncompatibility issues prevent us from reporting in normal regular expression compilations.\nBecause the behavior of this is subject to change in future Perl releases as we gain\nexperience, using this pragma will raise a warning of category \"experimental::restrict\".\nSee 'strict' in re.\n" }, { "name": "Unicode 7.0 (with correction) is now supported", "content": "For details on what is in this release, see .\nThe version of Unicode 7.0 that comes with Perl includes a correction dealing with glyph\nshaping in Arabic (see ).\n\n\"use  locale\" can restrict which locale categories are affected\nIt is now possible to pass a parameter to \"use locale\" to specify a subset of locale\ncategories to be locale-aware, with the remaining ones unaffected. See \"The \"use locale\"\npragma\" in perllocale for details.\n" }, { "name": "Perl now supports POSIX 2008 locale currency additions", "content": "On platforms that are able to handle POSIX.1-2008, the hash returned by \"POSIX::localeconv()\"\nincludes the international currency fields added by that version of the POSIX standard.\nThese are \"intncsprecedes\", \"intnsepbyspace\", \"intnsignposn\", \"intpcsprecedes\",\n\"intpsepbyspace\", and \"intpsignposn\".\n" }, { "name": "Better heuristics on older platforms for determining locale UTF-8ness", "content": "On platforms that implement neither the C99 standard nor the POSIX 2001 standard, determining\nif the current locale is UTF-8 or not depends on heuristics. These are improved in this\nrelease.\n" }, { "name": "Aliasing via reference", "content": "Variables and subroutines can now be aliased by assigning to a reference:\n\n\\$c = \\$d;\n\\&x = \\&y;\n\nAliasing can also be accomplished by using a backslash before a \"foreach\" iterator variable;\nthis is perhaps the most useful idiom this feature provides:\n\nforeach \\%hash (@arrayofhashrefs) { ... }\n\nThis feature is experimental and must be enabled via \"use feature 'refaliasing'\". It will\nwarn unless the \"experimental::refaliasing\" warnings category is disabled.\n\nSee \"Assigning to References\" in perlref\n" }, { "name": "\"prototype\" with no arguments", "content": "\"prototype()\" with no arguments now infers $. [GH #14376]\n.\n" }, { "name": "New \":const\" subroutine attribute", "content": "The \"const\" attribute can be applied to an anonymous subroutine. It causes the new sub to be\nexecuted immediately whenever one is created (i.e. when the \"sub\" expression is evaluated).\nIts value is captured and used to create a new constant subroutine that is returned. This\nfeature is experimental. See \"Constant Functions\" in perlsub.\n" }, { "name": "\"fileno\" now works on directory handles", "content": "When the relevant support is available in the operating system, the \"fileno\" builtin now\nworks on directory handles, yielding the underlying file descriptor in the same way as for\nfilehandles. On operating systems without such support, \"fileno\" on a directory handle\ncontinues to return the undefined value, as before, but also sets $! to indicate that the\noperation is not supported.\n\nCurrently, this uses either a \"ddfd\" member in the OS \"DIR\" structure, or a dirfd(3)\nfunction as specified by POSIX.1-2008.\n" }, { "name": "List form of pipe open implemented for Win32", "content": "The list form of pipe:\n\nopen my $fh, \"-|\", \"program\", @arguments;\n\nis now implemented on Win32. It has the same limitations as \"system LIST\" on Win32, since\nthe Win32 API doesn't accept program arguments as a list.\n" }, { "name": "Assignment to list repetition", "content": "\"(...) x ...\" can now be used within a list that is assigned to, as long as the left-hand\nside is a valid lvalue. This allows \"(undef,undef,$foo) = thatfunction()\" to be written as\n\"((undef)x2, $foo) = thatfunction()\".\n" }, { "name": "Infinity and NaN (not-a-number) handling improved", "content": "Floating point values are able to hold the special values infinity, negative infinity, and\nNaN (not-a-number). Now we more robustly recognize and propagate the value in computations,\nand on output normalize them to the strings \"Inf\", \"-Inf\", and \"NaN\".\n\nSee also the POSIX enhancements.\n" }, { "name": "Floating point parsing has been improved", "content": "Parsing and printing of floating point values has been improved.\n\nAs a completely new feature, hexadecimal floating point literals (like \"0x1.23p-4\") are now\nsupported, and they can be output with \"printf \"%a\"\". See \"Scalar value constructors\" in\nperldata for more details.\n" }, { "name": "Packing infinity or not-a-number into a character is now fatal", "content": "Before, when trying to pack infinity or not-a-number into a (signed) character, Perl would\nwarn, and assumed you tried to pack 0xFF; if you gave it as an argument to \"chr\", \"U+FFFD\"\nwas returned.\n\nBut now, all such actions (\"pack\", \"chr\", and \"print '%c'\") result in a fatal error.\n" }, { "name": "Experimental C Backtrace API", "content": "Perl now supports (via a C level API) retrieving the C level backtrace (similar to what\nsymbolic debuggers like gdb do).\n\nThe backtrace returns the stack trace of the C call frames, with the symbol names (function\nnames), the object names (like \"perl\"), and if it can, also the source code locations\n(file:line).\n\nThe supported platforms are Linux and OS X (some *BSD might work at least partly, but they\nhave not yet been tested).\n\nThe feature needs to be enabled with \"Configure -Dusecbacktrace\".\n\nSee \"C backtrace\" in perlhacktips for more information.\n" } ] }, "Security": { "content": "", "subsections": [ { "name": "Perl is now compiled with \"-fstack-protector-strong\" if available", "content": "Perl has been compiled with the anti-stack-smashing option \"-fstack-protector\" since 5.10.1.\nNow Perl uses the newer variant called \"-fstack-protector-strong\", if available.\n" }, { "name": "The Safe module could allow outside packages to be replaced", "content": "Critical bugfix: outside packages could be replaced. Safe has been patched to 2.38 to\naddress this.\n\nPerl is now always compiled with \"-DFORTIFYSOURCE=2\" if available\nThe 'code hardening' option called \"FORTIFYSOURCE\", available in gcc 4.*, is now always\nused for compiling Perl, if available.\n\nNote that this isn't necessarily a huge step since in many platforms the step had already\nbeen taken several years ago: many Linux distributions (like Fedora) have been using this\noption for Perl, and OS X has enforced the same for many years.\n" }, { "name": "Incompatible Changes", "content": "" }, { "name": "Subroutine signatures moved before attributes", "content": "The experimental sub signatures feature, as introduced in 5.20, parsed signatures after\nattributes. In this release, following feedback from users of the experimental feature, the\npositioning has been moved such that signatures occur after the subroutine name (if any) and\nbefore the attribute list (if any).\n" }, { "name": "\"&\" and \"\\&\" prototypes accepts only subs", "content": "The \"&\" prototype character now accepts only anonymous subs (\"sub {...}\"), things beginning\nwith \"\\&\", or an explicit \"undef\". Formerly it erroneously also allowed references to\narrays, hashes, and lists. [GH #2776] . [GH\n#14186] . [GH #14353]\n.\n\nIn addition, the \"\\&\" prototype was allowing subroutine calls, whereas now it only allows\nsubroutines: &foo is still permitted as an argument, while \"&foo()\" and \"foo()\" no longer\nare. [GH #10633] .\n" }, { "name": "\"use encoding\" is now lexical", "content": "The encoding pragma's effect is now limited to lexical scope. This pragma is deprecated, but\nin the meantime, it could adversely affect unrelated modules that are included in the same\nprogram; this change fixes that.\n" }, { "name": "List slices returning empty lists", "content": "List slices now return an empty list only if the original list was empty (or if there are no\nindices). Formerly, a list slice would return an empty list if all indices fell outside the\noriginal list; now it returns a list of \"undef\" values in that case. [GH #12335]\n.\n" }, { "name": "\"\\N{}\" with a sequence of multiple spaces is now a fatal error", "content": "E.g. \"\\N{TOO  MANY SPACES}\" or \"\\N{TRAILING SPACE }\". This has been deprecated since v5.18.\n\n\"use  UNIVERSAL  '...'\" is now a fatal error\nImporting functions from \"UNIVERSAL\" has been deprecated since v5.12, and is now a fatal\nerror. \"use UNIVERSAL\" without any arguments is still allowed.\n\nIn double-quotish \"\\cX\", X must now be a printable ASCII character\nIn prior releases, failure to do this raised a deprecation warning.\n\nSplitting the tokens \"(?\" and \"(*\" in regular expressions is now a fatal compilation error.\nThese had been deprecated since v5.18.\n" }, { "name": "\"qr/foo/x\" now ignores all Unicode pattern white space", "content": "The \"/x\" regular expression modifier allows the pattern to contain white space and comments\n(both of which are ignored) for improved readability. Until now, not all the white space\ncharacters that Unicode designates for this purpose were handled. The additional ones now\nrecognized are:\n\nU+0085 NEXT LINE\nU+200E LEFT-TO-RIGHT MARK\nU+200F RIGHT-TO-LEFT MARK\nU+2028 LINE SEPARATOR\nU+2029 PARAGRAPH SEPARATOR\n\nThe use of these characters with \"/x\" outside bracketed character classes and when not\npreceded by a backslash has raised a deprecation warning since v5.18. Now they will be\nignored.\n\nComment lines within \"(?[  ])\" are now ended only by a \"\\n\"\n\"(?[ ])\" is an experimental feature, introduced in v5.18. It operates as if \"/x\" is always\nenabled. But there was a difference: comment lines (following a \"#\" character) were\nterminated by anything matching \"\\R\" which includes all vertical whitespace, such as form\nfeeds. For consistency, this is now changed to match what terminates comment lines outside\n\"(?[ ])\", namely a \"\\n\" (even if escaped), which is the same as what terminates a heredoc\nstring and formats.\n" }, { "name": "\"(?[...])\" operators now follow standard Perl precedence", "content": "This experimental feature allows set operations in regular expression patterns. Prior to\nthis, the intersection operator had the same precedence as the other binary operators. Now\nit has higher precedence. This could lead to different outcomes than existing code expects\n(though the documentation has always noted that this change might happen, recommending fully\nparenthesizing the expressions). See \"Extended Bracketed Character Classes\" in\nperlrecharclass.\n" }, { "name": "Omitting \"%\" and \"@\" on hash and array names is no longer permitted", "content": "Really old Perl let you omit the \"@\" on array names and the \"%\" on hash names in some spots.\nThis has issued a deprecation warning since Perl 5.000, and is no longer permitted.\n" }, { "name": "\"$!\" text is now in English outside the scope of \"use locale\"", "content": "Previously, the text, unlike almost everything else, always came out based on the current\nunderlying locale of the program. (Also affected on some systems is \"$^E\".) For programs\nthat are unprepared to handle locale differences, this can cause garbage text to be\ndisplayed. It's better to display text that is translatable via some tool than garbage text\nwhich is much harder to figure out.\n" }, { "name": "\"$!\" text will be returned in UTF-8 when appropriate", "content": "The stringification of $! and $^E will have the UTF-8 flag set when the text is actually non-\nASCII UTF-8. This will enable programs that are set up to be locale-aware to properly output\nmessages in the user's native language. Code that needs to continue the 5.20 and earlier\nbehavior can do the stringification within the scopes of both \"use bytes\" and\n\"use locale \":messages\"\". Within these two scopes, no other Perl operations will be affected\nby locale; only $! and $^E stringification. The \"bytes\" pragma causes the UTF-8 flag to not\nbe set, just as in previous Perl releases. This resolves [GH #12035]\n.\n" }, { "name": "Support for \"?PATTERN?\" without explicit operator has been removed", "content": "The \"m?PATTERN?\" construct, which allows matching a regex only once, previously had an\nalternative form that was written directly with a question mark delimiter, omitting the\nexplicit \"m\" operator. This usage has produced a deprecation warning since 5.14.0. It is\nnow a syntax error, so that the question mark can be available for use in new operators.\n" }, { "name": "\"defined(@array)\" and \"defined(%hash)\" are now fatal errors", "content": "These have been deprecated since v5.6.1 and have raised deprecation warnings since v5.16.\n" }, { "name": "Using a hash or an array as a reference are now fatal errors", "content": "For example, \"%foo->{\"bar\"}\" now causes a fatal compilation error. These have been\ndeprecated since before v5.8, and have raised deprecation warnings since then.\n\nChanges to the \"*\" prototype\nThe \"*\" character in a subroutine's prototype used to allow barewords to take precedence over\nmost, but not all, subroutine names. It was never consistent and exhibited buggy behavior.\n\nNow it has been changed, so subroutines always take precedence over barewords, which brings\nit into conformity with similarly prototyped built-in functions:\n\nsub splat(*) { ... }\nsub foo { ... }\nsplat(foo); # now always splat(foo())\nsplat(bar); # still splat('bar') as before\nclose(foo); # close(foo())\nclose(bar); # close('bar')\n" } ] }, "Deprecations": { "content": "", "subsections": [ { "name": "Setting \"${^ENCODING}\" to anything but \"undef\"", "content": "This variable allows Perl scripts to be written in an encoding other than ASCII or UTF-8.\nHowever, it affects all modules globally, leading to wrong answers and segmentation faults.\nNew scripts should be written in UTF-8; old scripts should be converted to UTF-8, which is\neasily done with the piconv utility.\n" }, { "name": "Use of non-graphic characters in single-character variable names", "content": "The syntax for single-character variable names is more lenient than for longer variable\nnames, allowing the one-character name to be a punctuation character or even invisible (a\nnon-graphic). Perl v5.20 deprecated the ASCII-range controls as such a name. Now, all non-\ngraphic characters that formerly were allowed are deprecated. The practical effect of this\noccurs only when not under \"use utf8\", and affects just the C1 controls (code points 0x80\nthrough 0xFF), NO-BREAK SPACE, and SOFT HYPHEN.\n" }, { "name": "Inlining of \"sub () { $var }\" with observable side-effects", "content": "In many cases Perl makes \"sub () { $var }\" into an inlinable constant subroutine, capturing\nthe value of $var at the time the \"sub\" expression is evaluated. This can break the closure\nbehavior in those cases where $var is subsequently modified, since the subroutine won't\nreturn the changed value. (Note that this all only applies to anonymous subroutines with an\nempty prototype (\"sub ()\").)\n\nThis usage is now deprecated in those cases where the variable could be modified elsewhere.\nPerl detects those cases and emits a deprecation warning. Such code will likely change in\nthe future and stop producing a constant.\n\nIf your variable is only modified in the place where it is declared, then Perl will continue\nto make the sub inlinable with no warnings.\n\nsub makeconstant {\nmy $var = shift;\nreturn sub () { $var }; # fine\n}\n\nsub makeconstantdeprecated {\nmy $var;\n$var = shift;\nreturn sub () { $var }; # deprecated\n}\n\nsub makeconstantdeprecated2 {\nmy $var = shift;\nlogthatvalue($var); # could modify $var\nreturn sub () { $var }; # deprecated\n}\n\nIn the second example above, detecting that $var is assigned to only once is too hard to\ndetect. That it happens in a spot other than the \"my\" declaration is enough for Perl to find\nit suspicious.\n\nThis deprecation warning happens only for a simple variable for the body of the sub. (A\n\"BEGIN\" block or \"use\" statement inside the sub is ignored, because it does not become part\nof the sub's body.) For more complex cases, such as \"sub () { dosomething() if 0; $var }\"\nthe behavior has changed such that inlining does not happen if the variable is modifiable\nelsewhere. Such cases should be rare.\n" }, { "name": "Use of multiple \"/x\" regexp modifiers", "content": "It is now deprecated to say something like any of the following:\n\nqr/foo/xx;\n/(?xax:foo)/;\nuse re qw(/amxx);\n\nThat is, now \"x\" should only occur once in any string of contiguous regular expression\npattern modifiers. We do not believe there are any occurrences of this in all of CPAN. This\nis in preparation for a future Perl release having \"/xx\" permit white-space for readability\nin bracketed character classes (those enclosed in square brackets: \"[...]\").\n" }, { "name": "Using a NO-BREAK space in a character alias for \"\\N{...}\" is now deprecated", "content": "This non-graphic character is essentially indistinguishable from a regular space, and so\nshould not be allowed. See \"CUSTOM ALIASES\" in charnames.\n" }, { "name": "A literal \"{\" should now be escaped in a pattern", "content": "If you want a literal left curly bracket (also called a left brace) in a regular expression\npattern, you should now escape it by either preceding it with a backslash (\"\\{\") or enclosing\nit within square brackets \"[{]\", or by using \"\\Q\"; otherwise a deprecation warning will be\nraised. This was first announced as forthcoming in the v5.16 release; it will allow future\nextensions to the language to happen.\n" }, { "name": "Making all warnings fatal is discouraged", "content": "The documentation for fatal warnings notes that \"use warnings FATAL => 'all'\" is discouraged,\nand provides stronger language about the risks of fatal warnings in general.\n" }, { "name": "Performance Enhancements", "content": "• If a method or class name is known at compile time, a hash is precomputed to speed up\nrun-time method lookup. Also, compound method names like \"SUPER::new\" are parsed at\ncompile time, to save having to parse them at run time.\n\n• Array and hash lookups (especially nested ones) that use only constants or simple\nvariables as keys, are now considerably faster. See \"Internal Changes\" for more details.\n\n• \"(...)x1\", \"(\"constant\")x0\" and \"($scalar)x0\" are now optimised in list context. If the\nright-hand argument is a constant 1, the repetition operator disappears. If the right-\nhand argument is a constant 0, the whole expression is optimised to the empty list, so\nlong as the left-hand argument is a simple scalar or constant. (That is, \"(foo())x0\" is\nnot subject to this optimisation.)\n\n• \"substr\" assignment is now optimised into 4-argument \"substr\" at the end of a subroutine\n(or as the argument to \"return\"). Previously, this optimisation only happened in void\ncontext.\n\n• In \"\\L...\", \"\\Q...\", etc., the extra \"stringify\" op is now optimised away, making these\njust as fast as \"lcfirst\", \"quotemeta\", etc.\n\n• Assignment to an empty list is now sometimes faster. In particular, it never calls\n\"FETCH\" on tied arguments on the right-hand side, whereas it used to sometimes.\n\n• There is a performance improvement of up to 20% when \"length\" is applied to a non-\nmagical, non-tied string, and either \"use bytes\" is in scope or the string doesn't use\nUTF-8 internally.\n\n• On most perl builds with 64-bit integers, memory usage for non-magical, non-tied scalars\ncontaining only a floating point value has been reduced by between 8 and 32 bytes,\ndepending on OS.\n\n• In \"@array = split\", the assignment can be optimized away, so that \"split\" writes\ndirectly to the array. This optimisation was happening only for package arrays other\nthan @, and only sometimes. Now this optimisation happens almost all the time.\n\n• \"join\" is now subject to constant folding. So for example \"join \"-\", \"a\", \"b\"\" is\nconverted at compile-time to \"a-b\". Moreover, \"join\" with a scalar or constant for the\nseparator and a single-item list to join is simplified to a stringification, and the\nseparator doesn't even get evaluated.\n\n• \"qq(@array)\" is implemented using two ops: a stringify op and a join op. If the \"qq\"\ncontains nothing but a single array, the stringification is optimized away.\n\n• \"our $var\" and \"our($s,@a,%h)\" in void context are no longer evaluated at run time. Even\na whole sequence of \"our $foo;\" statements will simply be skipped over. The same applies\nto \"state\" variables.\n\n• Many internal functions have been refactored to improve performance and reduce their\nmemory footprints. [GH #13659] [GH #13856]\n [GH #13874]\n\n\n• \"-T\" and \"-B\" filetests will return sooner when an empty file is detected. [GH #13686]\n\n\n• Hash lookups where the key is a constant are faster.\n\n• Subroutines with an empty prototype and a body containing just \"undef\" are now eligible\nfor inlining. [GH #14077] \n\n• Subroutines in packages no longer need to be stored in typeglobs: declaring a subroutine\nwill now put a simple sub reference directly in the stash if possible, saving memory.\nThe typeglob still notionally exists, so accessing it will cause the stash entry to be\nupgraded to a typeglob (i.e. this is just an internal implementation detail). This\noptimization does not currently apply to XSUBs or exported subroutines, and method calls\nwill undo it, since they cache things in typeglobs. [GH #13392]\n\n\n• The functions \"utf8::nativetounicode()\" and \"utf8::unicodetonative()\" (see utf8) are\nnow optimized out on ASCII platforms. There is now not even a minimal performance hit in\nwriting code portable between ASCII and EBCDIC platforms.\n\n• Win32 Perl uses 8 KB less of per-process memory than before for every perl process,\nbecause some data is now memory mapped from disk and shared between processes from the\nsame perl binary.\n" }, { "name": "Modules and Pragmata", "content": "" }, { "name": "Updated Modules and Pragmata", "content": "Many of the libraries distributed with perl have been upgraded since v5.20.0. For a complete\nlist of changes, run:\n\ncorelist --diff 5.20.0 5.22.0\n\nYou can substitute your favorite version in place of 5.20.0, too.\n\nSome notable changes include:\n\n• Archive::Tar has been upgraded to version 2.04.\n\nTests can now be run in parallel.\n\n• attributes has been upgraded to version 0.27.\n\nThe usage of \"memEQs\" in the XS has been corrected. [GH #14072]\n\n\nAvoid reading beyond the end of a buffer. [perl #122629]\n\n• B has been upgraded to version 1.58.\n\nIt provides a new \"B::safename\" function, based on the existing \"B::GV->SAFENAME\", that\nconverts \"\\cOPEN\" to \"^OPEN\".\n\nNulled COPs are now of class \"B::COP\", rather than \"B::OP\".\n\n\"B::REGEXP\" objects now provide a \"qranoncv\" method for accessing the implicit CV\nassociated with \"qr//\" things containing code blocks, and a \"compflags\" method that\nreturns the pertinent flags originating from the \"qr//blahblah\" op.\n\n\"B::PMOP\" now provides a \"pmregexp\" method returning a \"B::REGEXP\" object. Two new\nclasses, \"B::PADNAME\" and \"B::PADNAMELIST\", have been introduced.\n\nA bug where, after an ithread creation or pseudofork, special/immortal SVs in the child\nithread/pseudoprocess did not have the correct class of \"B::SPECIAL\", has been fixed.\nThe \"id\" and \"outid\" PADLIST methods have been added.\n\n• B::Concise has been upgraded to version 0.996.\n\nNull ops that are part of the execution chain are now given sequence numbers.\n\nPrivate flags for nulled ops are now dumped with mnemonics as they would be for the non-\nnulled counterparts.\n\n• B::Deparse has been upgraded to version 1.35.\n\nIt now deparses \"+sub : attr { ... }\" correctly at the start of a statement. Without the\ninitial \"+\", \"sub\" would be a statement label.\n\n\"BEGIN\" blocks are now emitted in the right place most of the time, but the change\nunfortunately introduced a regression, in that \"BEGIN\" blocks occurring just before the\nend of the enclosing block may appear below it instead.\n\n\"B::Deparse\" no longer puts erroneous \"local\" here and there, such as for \"LIST =\ntr/a//d\". [perl #119815]\n\nAdjacent \"use\" statements are no longer accidentally nested if one contains a \"do\" block.\n[perl #115066]\n\nParenthesised arrays in lists passed to \"\\\" are now correctly deparsed with parentheses\n(e.g., \"\\(@a, (@b), @c)\" now retains the parentheses around @b), thus preserving the\nflattening behavior of referenced parenthesised arrays. Formerly, it only worked for one\narray: \"\\(@a)\".\n\n\"local our\" is now deparsed correctly, with the \"our\" included.\n\n\"for($foo; !$bar; $baz) {...}\" was deparsed without the \"!\" (or \"not\"). This has been\nfixed.\n\nCore keywords that conflict with lexical subroutines are now deparsed with the \"CORE::\"\nprefix.\n\n\"foreach state $x (...) {...}\" now deparses correctly with \"state\" and not \"my\".\n\n\"our @array = split(...)\" now deparses correctly with \"our\" in those cases where the\nassignment is optimized away.\n\nIt now deparses \"our(LIST)\" and typed lexical (\"my Dog $spot\") correctly.\n\nDeparse $# as that instead of as $#{}. [GH #14545]\n\n\nBEGIN blocks at the end of the enclosing scope are now deparsed in the right place.\n[perl #77452]\n\nBEGIN blocks were sometimes deparsed as ANON, but are now always called BEGIN.\n\nLexical subroutines are now fully deparsed. [perl #116553]\n\n\"Anything =~ y///r\" with \"/r\" no longer omits the left-hand operand.\n\nThe op trees that make up regexp code blocks are now deparsed for real. Formerly, the\noriginal string that made up the regular expression was used. That caused problems with\n\"qr/(?{< qr/.../flags\" followed by \"\"\" =~ QR\" is no longer without the flags.\n\nDeparsing \"BEGIN { undef &foo }\" with the -w switch enabled started to emit\n'uninitialized' warnings in Perl 5.14. This has been fixed.\n\nDeparsing calls to subs with a \"(;+)\" prototype resulted in an infinite loop. The \"(;$\")\n\"()\" and \"(;)\" prototypes were given the wrong precedence, causing \"foo($a<$b)\" to be\ndeparsed without the parentheses.\n\nDeparse now provides a defined state sub in inner subs.\n\n• B::Opprivate has been added.\n\nB::Opprivate provides detailed information about the flags used in the \"opprivate\"\nfield of perl opcodes.\n\n• bigint, bignum, bigrat have been upgraded to version 0.39.\n\nDocument in CAVEATS that using strings as numbers won't always invoke the big number\noverloading, and how to invoke it. [rt.perl.org #123064]\n\n• Carp has been upgraded to version 1.36.\n\n\"Carp::Heavy\" now ignores version mismatches with Carp if Carp is newer than 1.12, since\n\"Carp::Heavy\"'s guts were merged into Carp at that point. [GH #13708]\n\n\nCarp now handles non-ASCII platforms better.\n\nOff-by-one error fix for Perl < 5.14.\n\n• constant has been upgraded to version 1.33.\n\nIt now accepts fully-qualified constant names, allowing constants to be defined in\npackages other than the caller.\n\n• CPAN has been upgraded to version 2.11.\n\nAdd support for \"Cwd::getdcwd()\" and introduce workaround for a misbehavior seen on\nStrawberry Perl 5.20.1.\n\nFix \"chdir()\" after building dependencies bug.\n\nIntroduce experimental support for plugins/hooks.\n\nIntegrate the \"App::Cpan\" sources.\n\nDo not check recursion on optional dependencies.\n\nSanity check META.yml to contain a hash. [cpan #95271]\n\n\n• CPAN::Meta::Requirements has been upgraded to version 2.132.\n\nWorks around limitations in \"version::vpp\" detecting v-string magic and adds support for\nforthcoming ExtUtils::MakeMaker bootstrap version.pm for Perls older than 5.10.0.\n\n• Data::Dumper has been upgraded to version 2.158.\n\nFixes CVE-2014-4330 by adding a configuration variable/option to limit recursion when\ndumping deep data structures.\n\nChanges to resolve Coverity issues. XS dumps incorrectly stored the name of code\nreferences stored in a GLOB. [GH #13911] \n\n• DynaLoader has been upgraded to version 1.32.\n\nRemove \"dlnonlazy\" global if unused in Dynaloader. [perl #122926]\n\n• Encode has been upgraded to version 2.72.\n\n\"piconv\" now has better error handling when the encoding name is nonexistent, and a build\nbreakage when upgrading Encode in perl-5.8.2 and earlier has been fixed.\n\nBuilding in C++ mode on Windows now works.\n\n• Errno has been upgraded to version 1.23.\n\nAdd \"-P\" to the preprocessor command-line on GCC 5. GCC added extra line directives,\nbreaking parsing of error code definitions. [rt.perl.org #123784]\n\n• experimental has been upgraded to version 0.013.\n\nHardcodes features for Perls older than 5.15.7.\n\n• ExtUtils::CBuilder has been upgraded to version 0.280221.\n\nFixes a regression on Android. [GH #14064] \n\n• ExtUtils::Manifest has been upgraded to version 1.70.\n\nFixes a bug with \"maniread()\"'s handling of quoted filenames and improves \"manifind()\" to\nfollow symlinks. [GH #14003] \n\n• ExtUtils::ParseXS has been upgraded to version 3.28.\n\nOnly declare \"file\" unused if we actually define it. Improve generated \"RETVAL\" code\ngeneration to avoid repeated references to ST(0). [perl #123278] Broaden and document\nthe \"/OBJ$/\" to \"/REF$/\" typemap optimization for the \"DESTROY\" method. [perl #123418]\n\n• Fcntl has been upgraded to version 1.13.\n\nAdd support for the Linux pipe buffer size \"fcntl()\" commands.\n\n• File::Find has been upgraded to version 1.29.\n\n\"find()\" and \"finddepth()\" will now warn if passed inappropriate or misspelled options.\n\n• File::Glob has been upgraded to version 1.24.\n\nAvoid \"SvIV()\" expanding to call \"getsv()\" three times in a few places. [perl #123606]\n\n• HTTP::Tiny has been upgraded to version 0.054.\n\n\"keepalive\" is now fork-safe and thread-safe.\n\n• IO has been upgraded to version 1.35.\n\nThe XS implementation has been fixed for the sake of older Perls.\n\n• IO::Socket has been upgraded to version 1.38.\n\nDocument the limitations of the \"connected()\" method. [perl #123096]\n\n• IO::Socket::IP has been upgraded to version 0.37.\n\nA better fix for subclassing \"connect()\". [cpan #95983]\n [cpan #97050]\n\n\nImplements Timeout for \"connect()\". [cpan #92075]\n\n\n• The libnet collection of modules has been upgraded to version 3.05.\n\nSupport for IPv6 and SSL to \"Net::FTP\", \"Net::NNTP\", \"Net::POP3\" and \"Net::SMTP\".\nImprovements in \"Net::SMTP\" authentication.\n\n• Locale::Codes has been upgraded to version 3.34.\n\nFixed a bug in the scripts used to extract data from spreadsheets that prevented the SHP\ncurrency code from being found. [cpan #94229]\n\n\nNew codes have been added.\n\n• Math::BigInt has been upgraded to version 1.9997.\n\nSynchronize POD changes from the CPAN release. \"Math::BigFloat->blog(x)\" would sometimes\nreturn \"blog(2*x)\" when the accuracy was greater than 70 digits. The result of\n\"Math::BigFloat->bdiv()\" in list context now satisfies \"x = quotient * divisor +\nremainder\".\n\nCorrect handling of subclasses. [cpan #96254]\n [cpan #96329]\n\n\n• Module::Metadata has been upgraded to version 1.000026.\n\nSupport installations on older perls with an ExtUtils::MakeMaker earlier than 6.6303\n\n• overload has been upgraded to version 1.26.\n\nA redundant \"ref $sub\" check has been removed.\n\n• The PathTools module collection has been upgraded to version 3.56.\n\nA warning from the gcc compiler is now avoided when building the XS.\n\nDon't turn leading \"//\" into \"/\" on Cygwin. [perl #122635]\n\n• perl5db.pl has been upgraded to version 1.49.\n\nThe debugger would cause an assertion failure. [GH #14605]\n\n\n\"fork()\" in the debugger under \"tmux\" will now create a new window for the forked\nprocess. [GH #13602] \n\nThe debugger now saves the current working directory on startup and restores it when you\nrestart your program with \"R\" or \"rerun\". [GH #13691]\n\n\n• PerlIO::scalar has been upgraded to version 0.22.\n\nReading from a position well past the end of the scalar now correctly returns end of\nfile. [perl #123443]\n\nSeeking to a negative position still fails, but no longer leaves the file position set to\na negation location.\n\n\"eof()\" on a \"PerlIO::scalar\" handle now properly returns true when the file position is\npast the 2GB mark on 32-bit systems.\n\nAttempting to write at file positions impossible for the platform now fail early rather\nthan wrapping at 4GB.\n\n• Pod::Perldoc has been upgraded to version 3.25.\n\nFilehandles opened for reading or writing now have \":encoding(UTF-8)\" set. [cpan #98019]\n\n\n• POSIX has been upgraded to version 1.53.\n\nThe C99 math functions and constants (for example \"acosh\", \"isinf\", \"isnan\", \"round\",\n\"trunc\"; \"ME\", \"MSQRT2\", \"MPI\") have been added.\n\n\"POSIX::tmpnam()\" now produces a deprecation warning. [perl #122005]\n\n• Safe has been upgraded to version 2.39.\n\n\"reval\" was not propagating void context properly.\n\n• Scalar-List-Utils has been upgraded to version 1.41.\n\nA new module, Sub::Util, has been added, containing functions related to CODE refs,\nincluding \"subname\" (inspired by \"Sub::Identity\") and \"setsubname\" (copied and renamed\nfrom \"Sub::Name\"). The use of \"GetMagic\" in \"List::Util::reduce()\" has also been fixed.\n[cpan #63211] \n\n• SDBMFile has been upgraded to version 1.13.\n\nSimplified the build process. [perl #123413]\n\n• Time::Piece has been upgraded to version 1.29.\n\nWhen pretty printing negative \"Time::Seconds\", the \"minus\" is no longer lost.\n\n• Unicode::Collate has been upgraded to version 1.12.\n\nVersion 0.67's improved discontiguous contractions is invalidated by default and is\nsupported as a parameter \"longcontraction\".\n\n• Unicode::Normalize has been upgraded to version 1.18.\n\nThe XSUB implementation has been removed in favor of pure Perl.\n\n• Unicode::UCD has been upgraded to version 0.61.\n\nA new function propertyvalues() has been added to return a given property's possible\nvalues.\n\nA new function charprop() has been added to return the value of a given property for a\ngiven code point.\n\nA new function charpropsall() has been added to return the values of all Unicode\nproperties for a given code point.\n\nA bug has been fixed so that propaliases() returns the correct short and long names for\nthe Perl extensions where it was incorrect.\n\nA bug has been fixed so that propvaluealiases() returns \"undef\" instead of a wrong\nresult for properties that are Perl extensions.\n\nThis module now works on EBCDIC platforms.\n\n• utf8 has been upgraded to version 1.17\n\nA mismatch between the documentation and the code in \"utf8::downgrade()\" was fixed in\nfavor of the documentation. The optional second argument is now correctly treated as a\nperl boolean (true/false semantics) and not as an integer.\n\n• version has been upgraded to version 0.9909.\n\nNumerous changes. See the Changes file in the CPAN distribution for details.\n\n• Win32 has been upgraded to version 0.51.\n\n\"GetOSName()\" now supports Windows 8.1, and building in C++ mode now works.\n\n• Win32API::File has been upgraded to version 0.1202\n\nBuilding in C++ mode now works.\n\n• XSLoader has been upgraded to version 0.20.\n\nAllow XSLoader to load modules from a different namespace. [perl #122455]\n" }, { "name": "Removed Modules and Pragmata", "content": "The following modules (and associated modules) have been removed from the core perl\ndistribution:\n\n• CGI\n\n• Module::Build\n" } ] }, "Documentation": { "content": "", "subsections": [ { "name": "New Documentation", "content": "perlunicook\n\nThis document, by Tom Christiansen, provides examples of handling Unicode in Perl.\n" }, { "name": "Changes to Existing Documentation", "content": "perlaix\n\n• A note on long doubles has been added.\n\nperlapi\n\n• Note that \"SvSetSV\" doesn't do set magic.\n\n• \"svusepvnflags\" - fix documentation to mention the use of \"Newx\" instead of \"malloc\".\n\n[GH #13835] \n\n• Clarify where \"NUL\" may be embedded or is required to terminate a string.\n\n• Some documentation that was previously missing due to formatting errors is now included.\n\n• Entries are now organized into groups rather than by the file where they are found.\n\n• Alphabetical sorting of entries is now done consistently (automatically by the POD\ngenerator) to make entries easier to find when scanning.\n\nperldata\n\n• The syntax of single-character variable names has been brought up-to-date and more fully\nexplained.\n\n• Hexadecimal floating point numbers are described, as are infinity and NaN.\n\nperlebcdic\n\n• This document has been significantly updated in the light of recent improvements to\nEBCDIC support.\n\nperlfilter\n\n• Added a LIMITATIONS section.\n\nperlfunc\n\n• Mention that \"study()\" is currently a no-op.\n\n• Calling \"delete\" or \"exists\" on array values is now described as \"strongly discouraged\"\nrather than \"deprecated\".\n\n• Improve documentation of \"our\".\n\n• \"-l\" now notes that it will return false if symlinks aren't supported by the file system.\n[GH #13695] \n\n• Note that \"exec LIST\" and \"system LIST\" may fall back to the shell on Win32. Only the\nindirect-object syntax \"exec PROGRAM LIST\" and \"system PROGRAM LIST\" will reliably avoid\nusing the shell.\n\nThis has also been noted in perlport.\n\n[GH #13907] \n\nperlguts\n\n• The OOK example has been updated to account for COW changes and a change in the storage\nof the offset.\n\n• Details on C level symbols and libperl.t added.\n\n• Information on Unicode handling has been added\n\n• Information on EBCDIC handling has been added\n\nperlhack\n\n• A note has been added about running on platforms with non-ASCII character sets\n\n• A note has been added about performance testing\n\nperlhacktips\n\n• Documentation has been added illustrating the perils of assuming that there is no change\nto the contents of static memory pointed to by the return values of Perl's wrappers for C\nlibrary functions.\n\n• Replacements for \"tmpfile\", \"atoi\", \"strtol\", and \"strtoul\" are now recommended.\n\n• Updated documentation for the \"test.valgrind\" \"make\" target. [GH #13658]\n\n\n• Information is given about writing test files portably to non-ASCII platforms.\n\n• A note has been added about how to get a C language stack backtrace.\n\nperlhpux\n\n• Note that the message \"Redeclaration of \"sendpath\" with a different storage class\nspecifier\" is harmless.\n\nperllocale\n\n• Updated for the enhancements in v5.22, along with some clarifications.\n\nperlmodstyle\n\n• Instead of pointing to the module list, we are now pointing to PrePAN\n.\n\nperlop\n\n• Updated for the enhancements in v5.22, along with some clarifications.\n\nperlpodspec\n\n• The specification of the pod language is changing so that the default encoding of pods\nthat aren't in UTF-8 (unless otherwise indicated) is CP1252 instead of ISO 8859-1\n(Latin1).\n\nperlpolicy\n\n• We now have a code of conduct for the p5p mailing list, as documented in \"STANDARDS OF\nCONDUCT\" in perlpolicy.\n\n• The conditions for marking an experimental feature as non-experimental are now set out.\n\n• Clarification has been made as to what sorts of changes are permissible in maintenance\nreleases.\n\nperlport\n\n• Out-of-date VMS-specific information has been fixed and/or simplified.\n\n• Notes about EBCDIC have been added.\n\nperlre\n\n• The description of the \"/x\" modifier has been clarified to note that comments cannot be\ncontinued onto the next line by escaping them; and there is now a list of all the\ncharacters that are considered whitespace by this modifier.\n\n• The new \"/n\" modifier is described.\n\n• A note has been added on how to make bracketed character class ranges portable to non-\nASCII machines.\n\nperlrebackslash\n\n• Added documentation of \"\\b{sb}\", \"\\b{wb}\", \"\\b{gcb}\", and \"\\b{g}\".\n\nperlrecharclass\n\n• Clarifications have been added to \"Character Ranges\" in perlrecharclass to the effect\n\"[A-Z]\", \"[a-z]\", \"[0-9]\" and any subranges thereof in regular expression bracketed\ncharacter classes are guaranteed to match exactly what a naive English speaker would\nexpect them to match, even on platforms (such as EBCDIC) where perl has to do extra work\nto accomplish this.\n\n• The documentation of Bracketed Character Classes has been expanded to cover the\nimprovements in \"qr/[\\N{named sequence}]/\" (see under \"Selected Bug Fixes\").\n\nperlref\n\n• A new section has been added Assigning to References\n\nperlsec\n\n• Comments added on algorithmic complexity and tied hashes.\n\nperlsyn\n\n• An ambiguity in the documentation of the \"...\" statement has been corrected. [GH #14054]\n\n\n• The empty conditional in \"for\" and \"while\" is now documented in perlsyn.\n\nperlunicode\n\n• This has had extensive revisions to bring it up-to-date with current Unicode support and\nto make it more readable. Notable is that Unicode 7.0 changed what it should do with\nnon-characters. Perl retains the old way of handling for reasons of backward\ncompatibility. See \"Noncharacter code points\" in perlunicode.\n\nperluniintro\n\n• Advice for how to make sure your strings and regular expression patterns are interpreted\nas Unicode has been updated.\n\nperlvar\n\n• $] is no longer listed as being deprecated. Instead, discussion has been added on the\nadvantages and disadvantages of using it versus $^V. $OLDPERLVERSION was re-added to\nthe documentation as the long form of $].\n\n• \"${^ENCODING}\" is now marked as deprecated.\n\n• The entry for \"%^H\" has been clarified to indicate it can only handle simple values.\n\nperlvms\n\n• Out-of-date and/or incorrect material has been removed.\n\n• Updated documentation on environment and shell interaction in VMS.\n\nperlxs\n\n• Added a discussion of locale issues in XS code.\n" } ] }, "Diagnostics": { "content": "The following additions or changes have been made to diagnostic output, including warnings\nand fatal error messages. For the complete list of diagnostic messages, see perldiag.\n", "subsections": [ { "name": "New Diagnostics", "content": "New Errors\n\n• Bad symbol for scalar\n\n(P) An internal request asked to add a scalar entry to something that wasn't a symbol\ntable entry.\n\n• Can't use a hash as a reference\n\n(F) You tried to use a hash as a reference, as in \"%foo->{\"bar\"}\" or \"%$ref->{\"hello\"}\".\nVersions of perl <= 5.6.1 used to allow this syntax, but shouldn't have.\n\n• Can't use an array as a reference\n\n(F) You tried to use an array as a reference, as in \"@foo->[23]\" or \"@$ref->[99]\".\nVersions of perl <= 5.6.1 used to allow this syntax, but shouldn't have.\n\n• Can't use 'defined(@array)' (Maybe you should just omit the defined()?)\n\n(F) \"defined()\" is not useful on arrays because it checks for an undefined scalar value.\nIf you want to see if the array is empty, just use \"if (@array) { # not empty }\" for\nexample.\n\n• Can't use 'defined(%hash)' (Maybe you should just omit the defined()?)\n\n(F) \"defined()\" is not usually right on hashes.\n\nAlthough \"defined %hash\" is false on a plain not-yet-used hash, it becomes true in\nseveral non-obvious circumstances, including iterators, weak references, stash names,\neven remaining true after \"undef %hash\". These things make \"defined %hash\" fairly\nuseless in practice, so it now generates a fatal error.\n\nIf a check for non-empty is what you wanted then just put it in boolean context (see\n\"Scalar values\" in perldata):\n\nif (%hash) {\n# not empty\n}\n\nIf you had \"defined %Foo::Bar::QUUX\" to check whether such a package variable exists then\nthat's never really been reliable, and isn't a good way to enquire about the features of\na package, or whether it's loaded, etc.\n\n• Cannot chr %f\n\n(F) You passed an invalid number (like an infinity or not-a-number) to \"chr\".\n\n• Cannot compress %f in pack\n\n(F) You tried converting an infinity or not-a-number to an unsigned character, which\nmakes no sense.\n\n• Cannot pack %f with '%c'\n\n(F) You tried converting an infinity or not-a-number to a character, which makes no\nsense.\n\n• Cannot print %f with '%c'\n\n(F) You tried printing an infinity or not-a-number as a character (%c), which makes no\nsense. Maybe you meant '%s', or just stringifying it?\n\n• charnames alias definitions may not contain a sequence of multiple spaces\n\n(F) You defined a character name which had multiple space characters in a row. Change\nthem to single spaces. Usually these names are defined in the \":alias\" import argument\nto \"use charnames\", but they could be defined by a translator installed into\n$^H{charnames}. See \"CUSTOM ALIASES\" in charnames.\n\n• charnames alias definitions may not contain trailing white-space\n\n(F) You defined a character name which ended in a space character. Remove the trailing\nspace(s). Usually these names are defined in the \":alias\" import argument to \"use\ncharnames\", but they could be defined by a translator installed into $^H{charnames}. See\n\"CUSTOM ALIASES\" in charnames.\n\n• :const is not permitted on named subroutines\n\n(F) The \"const\" attribute causes an anonymous subroutine to be run and its value captured\nat the time that it is cloned. Named subroutines are not cloned like this, so the\nattribute does not make sense on them.\n\n• Hexadecimal float: internal error\n\n(F) Something went horribly bad in hexadecimal float handling.\n\n• Hexadecimal float: unsupported long double format\n\n(F) You have configured Perl to use long doubles but the internals of the long double\nformat are unknown, therefore the hexadecimal float output is impossible.\n\n• Illegal suidscript\n\n(F) The script run under suidperl was somehow illegal.\n\n• In '(?...)', the '(' and '?' must be adjacent in regex; marked by <-- HERE in m/%s/\n\n(F) The two-character sequence \"(?\" in this context in a regular expression pattern\nshould be an indivisible token, with nothing intervening between the \"(\" and the \"?\", but\nyou separated them.\n\n• In '(*VERB...)', the '(' and '*' must be adjacent in regex; marked by <-- HERE in m/%s/\n\n(F) The two-character sequence \"(*\" in this context in a regular expression pattern\nshould be an indivisible token, with nothing intervening between the \"(\" and the \"*\", but\nyou separated them.\n\n• Invalid quantifier in {,} in regex; marked by <-- HERE in m/%s/\n\n(F) The pattern looks like a {min,max} quantifier, but the min or max could not be parsed\nas a valid number: either it has leading zeroes, or it represents too big a number to\ncope with. The <-- HERE shows where in the regular expression the problem was\ndiscovered. See perlre.\n\n• '%s' is an unknown bound type in regex\n\n(F) You used \"\\b{...}\" or \"\\B{...}\" and the \"...\" is not known to Perl. The current\nvalid ones are given in \"\\b{}, \\b, \\B{}, \\B\" in perlrebackslash.\n\n• Missing or undefined argument to require\n\n(F) You tried to call \"require\" with no argument or with an undefined value as an\nargument. \"require\" expects either a package name or a file-specification as an\nargument. See \"require\" in perlfunc.\n\nFormerly, \"require\" with no argument or \"undef\" warned about a Null filename.\n\nNew Warnings\n\n• \\C is deprecated in regex\n\n(D deprecated) The \"/\\C/\" character class was deprecated in v5.20, and now emits a\nwarning. It is intended that it will become an error in v5.24. This character class\nmatches a single byte even if it appears within a multi-byte character, breaks\nencapsulation, and can corrupt UTF-8 strings.\n\n• \"%s\" is more clearly written simply as \"%s\" in regex; marked by <-- HERE in m/%s/\n\n(W regexp) (only under \"use re 'strict'\" or within \"(?[...])\")\n\nYou specified a character that has the given plainer way of writing it, and which is also\nportable to platforms running with different character sets.\n\n• Argument \"%s\" treated as 0 in increment (++)\n\n(W numeric) The indicated string was fed as an argument to the \"++\" operator which\nexpects either a number or a string matching \"/^[a-zA-Z]*[0-9]*\\z/\". See \"Auto-increment\nand Auto-decrement\" in perlop for details.\n\n• Both or neither range ends should be Unicode in regex; marked by <-- HERE in m/%s/\n\n(W regexp) (only under \"use re 'strict'\" or within \"(?[...])\")\n\nIn a bracketed character class in a regular expression pattern, you had a range which has\nexactly one end of it specified using \"\\N{}\", and the other end is specified using a non-\nportable mechanism. Perl treats the range as a Unicode range, that is, all the\ncharacters in it are considered to be the Unicode characters, and which may be different\ncode points on some platforms Perl runs on. For example, \"[\\N{U+06}-\\x08]\" is treated as\nif you had instead said \"[\\N{U+06}-\\N{U+08}]\", that is it matches the characters whose\ncode points in Unicode are 6, 7, and 8. But that \"\\x08\" might indicate that you meant\nsomething different, so the warning gets raised.\n\n• Can't do %s(\"%s\") on non-UTF-8 locale; resolved to \"%s\".\n\n(W locale) You are 1) running under \"\"use locale\"\"; 2) the current locale is not a UTF-8\none; 3) you tried to do the designated case-change operation on the specified Unicode\ncharacter; and 4) the result of this operation would mix Unicode and locale rules, which\nlikely conflict.\n\nThe warnings category \"locale\" is new.\n\n• :const is experimental\n\n(S experimental::constattr) The \"const\" attribute is experimental. If you want to use\nthe feature, disable the warning with \"no warnings 'experimental::constattr'\", but know\nthat in doing so you are taking the risk that your code may break in a future Perl\nversion.\n\n• gmtime(%f) failed\n\n(W overflow) You called \"gmtime\" with a number that it could not handle: too large, too\nsmall, or NaN. The returned value is \"undef\".\n\n• Hexadecimal float: exponent overflow\n\n(W overflow) The hexadecimal floating point has larger exponent than the floating point\nsupports.\n\n• Hexadecimal float: exponent underflow\n\n(W overflow) The hexadecimal floating point has smaller exponent than the floating point\nsupports.\n\n• Hexadecimal float: mantissa overflow\n\n(W overflow) The hexadecimal floating point literal had more bits in the mantissa (the\npart between the \"0x\" and the exponent, also known as the fraction or the significand)\nthan the floating point supports.\n\n• Hexadecimal float: precision loss\n\n(W overflow) The hexadecimal floating point had internally more digits than could be\noutput. This can be caused by unsupported long double formats, or by 64-bit integers not\nbeing available (needed to retrieve the digits under some configurations).\n\n• Locale '%s' may not work well.%s\n\n(W locale) You are using the named locale, which is a non-UTF-8 one, and which perl has\ndetermined is not fully compatible with what it can handle. The second %s gives a\nreason.\n\nThe warnings category \"locale\" is new.\n\n• localtime(%f) failed\n\n(W overflow) You called \"localtime\" with a number that it could not handle: too large,\ntoo small, or NaN. The returned value is \"undef\".\n\n• Negative repeat count does nothing\n\n(W numeric) You tried to execute the \"x\" repetition operator fewer than 0 times, which\ndoesn't make sense.\n\n• NO-BREAK SPACE in a charnames alias definition is deprecated\n\n(D deprecated) You defined a character name which contained a no-break space character.\nChange it to a regular space. Usually these names are defined in the \":alias\" import\nargument to \"use charnames\", but they could be defined by a translator installed into\n$^H{charnames}. See \"CUSTOM ALIASES\" in charnames.\n\n• Non-finite repeat count does nothing\n\n(W numeric) You tried to execute the \"x\" repetition operator \"Inf\" (or \"-Inf\") or NaN\ntimes, which doesn't make sense.\n\n• PerlIO layer ':win32' is experimental\n\n(S experimental::win32perlio) The \":win32\" PerlIO layer is experimental. If you want to\ntake the risk of using this layer, simply disable this warning:\n\nno warnings \"experimental::win32perlio\";\n\n• Ranges of ASCII printables should be some subset of \"0-9\", \"A-Z\", or \"a-z\" in regex;\nmarked by <-- HERE in m/%s/\n\n(W regexp) (only under \"use re 'strict'\" or within \"(?[...])\")\n\nStricter rules help to find typos and other errors. Perhaps you didn't even intend a\nrange here, if the \"-\" was meant to be some other character, or should have been escaped\n(like \"\\-\"). If you did intend a range, the one that was used is not portable between\nASCII and EBCDIC platforms, and doesn't have an obvious meaning to a casual reader.\n\n[3-7] # OK; Obvious and portable\n[d-g] # OK; Obvious and portable\n[A-Y] # OK; Obvious and portable\n[A-z] # WRONG; Not portable; not clear what is meant\n[a-Z] # WRONG; Not portable; not clear what is meant\n[%-.] # WRONG; Not portable; not clear what is meant\n[\\x41-Z] # WRONG; Not portable; not obvious to non-geek\n\n(You can force portability by specifying a Unicode range, which means that the endpoints\nare specified by \"\\N{...}\", but the meaning may still not be obvious.) The stricter\nrules require that ranges that start or stop with an ASCII character that is not a\ncontrol have all their endpoints be a literal character, and not some escape sequence\n(like \"\\x41\"), and the ranges must be all digits, or all uppercase letters, or all\nlowercase letters.\n\n• Ranges of digits should be from the same group in regex; marked by <-- HERE in m/%s/\n\n(W regexp) (only under \"use re 'strict'\" or within \"(?[...])\")\n\nStricter rules help to find typos and other errors. You included a range, and at least\none of the end points is a decimal digit. Under the stricter rules, when this happens,\nboth end points should be digits in the same group of 10 consecutive digits.\n\n• Redundant argument in %s\n\n(W redundant) You called a function with more arguments than were needed, as indicated by\ninformation within other arguments you supplied (e.g. a printf format). Currently only\nemitted when a printf-type format required fewer arguments than were supplied, but might\nbe used in the future for e.g. \"pack\" in perlfunc.\n\nThe warnings category \"redundant\" is new. See also [GH #13534]\n.\n\n• Replacement list is longer than search list\n\nThis is not a new diagnostic, but in earlier releases was accidentally not displayed if\nthe transliteration contained wide characters. This is now fixed, so that you may see\nthis diagnostic in places where you previously didn't (but should have).\n\n• Use of \\b{} for non-UTF-8 locale is wrong. Assuming a UTF-8 locale\n\n(W locale) You are matching a regular expression using locale rules, and a Unicode\nboundary is being matched, but the locale is not a Unicode one. This doesn't make sense.\nPerl will continue, assuming a Unicode (UTF-8) locale, but the results could well be\nwrong except if the locale happens to be ISO-8859-1 (Latin1) where this message is\nspurious and can be ignored.\n\nThe warnings category \"locale\" is new.\n\n• Using /u for '%s' instead of /%s in regex; marked by <-- HERE in m/%s/\n\n(W regexp) You used a Unicode boundary (\"\\b{...}\" or \"\\B{...}\") in a portion of a regular\nexpression where the character set modifiers \"/a\" or \"/aa\" are in effect. These two\nmodifiers indicate an ASCII interpretation, and this doesn't make sense for a Unicode\ndefinition. The generated regular expression will compile so that the boundary uses all\nof Unicode. No other portion of the regular expression is affected.\n\n• The bitwise feature is experimental\n\n(S experimental::bitwise) This warning is emitted if you use bitwise operators (\"& | ^ ~\n&. |. ^. ~.\") with the \"bitwise\" feature enabled. Simply suppress the warning if you\nwant to use the feature, but know that in doing so you are taking the risk of using an\nexperimental feature which may change or be removed in a future Perl version:\n\nno warnings \"experimental::bitwise\";\nuse feature \"bitwise\";\n$x |.= $y;\n\n• Unescaped left brace in regex is deprecated, passed through in regex; marked by <-- HERE\nin m/%s/\n\n(D deprecated, regexp) You used a literal \"{\" character in a regular expression pattern.\nYou should change to use \"\\{\" instead, because a future version of Perl (tentatively\nv5.26) will consider this to be a syntax error. If the pattern delimiters are also\nbraces, any matching right brace (\"}\") should also be escaped to avoid confusing the\nparser, for example,\n\nqr{abc\\{def\\}ghi}\n\n• Use of literal non-graphic characters in variable names is deprecated\n\n(D deprecated) Using literal non-graphic (including control) characters in the source to\nrefer to the ^FOO variables, like $^X and \"${^GLOBALPHASE}\" is now deprecated.\n\n• Useless use of attribute \"const\"\n\n(W misc) The \"const\" attribute has no effect except on anonymous closure prototypes. You\napplied it to a subroutine via attributes.pm. This is only useful inside an attribute\nhandler for an anonymous subroutine.\n\n• Useless use of /d modifier in transliteration operator\n\nThis is not a new diagnostic, but in earlier releases was accidentally not displayed if\nthe transliteration contained wide characters. This is now fixed, so that you may see\nthis diagnostic in places where you previously didn't (but should have).\n\n• \"use re 'strict'\" is experimental\n\n(S experimental::restrict) The things that are different when a regular expression\npattern is compiled under 'strict' are subject to change in future Perl releases in\nincompatible ways; there are also proposals to change how to enable strict checking\ninstead of using this subpragma. This means that a pattern that compiles today may not\nin a future Perl release. This warning is to alert you to that risk.\n\n• Warning: unable to close filehandle properly: %s\n\nWarning: unable to close filehandle %s properly: %s\n\n(S io) Previously, perl silently ignored any errors when doing an implicit close of a\nfilehandle, i.e. where the reference count of the filehandle reached zero and the user's\ncode hadn't already called \"close()\"; e.g.\n\n{\nopen my $fh, '>', $file or die \"open: '$file': $!\\n\";\nprint $fh, $data or die;\n} # implicit close here\n\nIn a situation such as disk full, due to buffering, the error may only be detected during\nthe final close, so not checking the result of the close is dangerous.\n\nSo perl now warns in such situations.\n\n• Wide character (U+%X) in %s\n\n(W locale) While in a single-byte locale (i.e., a non-UTF-8 one), a multi-byte character\nwas encountered. Perl considers this character to be the specified Unicode code point.\nCombining non-UTF-8 locales and Unicode is dangerous. Almost certainly some characters\nwill have two different representations. For example, in the ISO 8859-7 (Greek) locale,\nthe code point 0xC3 represents a Capital Gamma. But so also does 0x393. This will make\nstring comparisons unreliable.\n\nYou likely need to figure out how this multi-byte character got mixed up with your\nsingle-byte locale (or perhaps you thought you had a UTF-8 locale, but Perl disagrees).\n\nThe warnings category \"locale\" is new.\n" }, { "name": "Changes to Existing Diagnostics", "content": "• <> should be quotes\n\nThis warning has been changed to <> at require-statement should be quotes to make the\nissue more identifiable.\n\n• Argument \"%s\" isn't numeric%s\n\nThe perldiag entry for this warning has added this clarifying note:\n\nNote that for the Inf and NaN (infinity and not-a-number) the\ndefinition of \"numeric\" is somewhat unusual: the strings themselves\n(like \"Inf\") are considered numeric, and anything following them is\nconsidered non-numeric.\n\n• Global symbol \"%s\" requires explicit package name\n\nThis message has had '(did you forget to declare \"my %s\"?)' appended to it, to make it\nmore helpful to new Perl programmers. [GH #13732]\n\n\n• '\"my\" variable &foo::bar can't be in a package' has been reworded to say 'subroutine'\ninstead of 'variable'.\n\n• \\N{} in character class restricted to one character in regex; marked by <-- HERE in m/%s/\n\nThis message has had character class changed to inverted character class or as a range\nend-point is to reflect improvements in \"qr/[\\N{named sequence}]/\" (see under \"Selected\nBug Fixes\").\n\n• panic: frexp\n\nThis message has had ': %f' appended to it, to show what the offending floating point\nnumber is.\n\n• Possible precedence problem on bitwise %c operator reworded as Possible precedence\nproblem on bitwise %s operator.\n\n• Unsuccessful %s on filename containing newline\n\nThis warning is now only produced when the newline is at the end of the filename.\n\n• \"Variable %s will not stay shared\" has been changed to say \"Subroutine\" when it is\nactually a lexical sub that will not stay shared.\n\n• Variable length lookbehind not implemented in regex m/%s/\n\nThe perldiag entry for this warning has had information about Unicode behavior added.\n" }, { "name": "Diagnostic Removals", "content": "• \"Ambiguous use of -foo resolved as -&foo()\"\n\nThere is actually no ambiguity here, and this impedes the use of negated constants; e.g.,\n\"-Inf\".\n\n• \"Constant is not a FOO reference\"\n\nCompile-time checking of constant dereferencing (e.g., \"myconstant->()\") has been\nremoved, since it was not taking overloading into account. [GH #9891]\n [GH #14044]\n\n" }, { "name": "Utility Changes", "content": "find2perl, s2p and a2p removal\n• The x2p/ directory has been removed from the Perl core.\n\nThis removes find2perl, s2p and a2p. They have all been released to CPAN as separate\ndistributions (\"App::find2perl\", \"App::s2p\", \"App::a2p\").\n" }, { "name": "h2ph", "content": "• h2ph now handles hexadecimal constants in the compiler's predefined macro definitions, as\nvisible in $Config{cppsymbols}. [GH #14491]\n.\n" }, { "name": "encguess", "content": "• No longer depends on non-core modules.\n" }, { "name": "Configuration and Compilation", "content": "• Configure now checks for \"lrintl()\", \"lroundl()\", \"llrintl()\", and \"llroundl()\".\n\n• Configure with \"-Dmksymlinks\" should now be faster. [GH #13890]\n.\n\n• The \"pthreads\" and \"cl\" libraries will be linked by default if present. This allows XS\nmodules that require threading to work on non-threaded perls. Note that you must still\npass \"-Dusethreads\" if you want a threaded perl.\n\n• To get more precision and range for floating point numbers one can now use the GCC\nquadmath library which implements the quadruple precision floating point numbers on x86\nand IA-64 platforms. See INSTALL for details.\n\n• MurmurHash64A and MurmurHash64B can now be configured as the internal hash function.\n\n• \"make test.valgrind\" now supports parallel testing.\n\nFor example:\n\nTESTJOBS=9 make test.valgrind\n\nSee \"valgrind\" in perlhacktips for more information.\n\n[GH #13658] \n\n• The MAD (Misc Attribute Decoration) build option has been removed\n\nThis was an unmaintained attempt at preserving the Perl parse tree more faithfully so\nthat automatic conversion of Perl 5 to Perl 6 would have been easier.\n\nThis build-time configuration option had been unmaintained for years, and had probably\nseriously diverged on both Perl 5 and Perl 6 sides.\n\n• A new compilation flag, \"-DPERLOPPARENT\" is available. For details, see the discussion\nbelow at \"Internal Changes\".\n\n• Pathtools no longer tries to load XS on miniperl. This speeds up building perl slightly.\n" } ] }, "Testing": { "content": "• t/porting/recontext.t has been added to test that utf8 and its dependencies only use the\nsubset of the \"$1..$n\" capture vars that \"Perlsaverecontext()\" is hard-coded to\nlocalize, because that function has no efficient way of determining at runtime what vars\nto localize.\n\n• Tests for performance issues have been added in the file t/perf/taint.t.\n\n• Some regular expression tests are written in such a way that they will run very slowly if\ncertain optimizations break. These tests have been moved into new files, t/re/speed.t and\nt/re/speedthr.t, and are run with a \"watchdog()\".\n\n• \"test.pl\" now allows \"plan skipall => $reason\", to make it more compatible with\n\"Test::More\".\n\n• A new test script, op/infnan.t, has been added to test if infinity and NaN are working\ncorrectly. See \"Infinity and NaN (not-a-number) handling improved\".\n", "subsections": [ { "name": "Platform Support", "content": "" }, { "name": "Regained Platforms", "content": "IRIX and Tru64 platforms are working again.\nSome \"make test\" failures remain: [GH #14557]\n and [GH #14727]\n for IRIX; [GH #14629]\n, [cpan #99605]\n, and [cpan #104836]\n for Tru64.\n\nz/OS running EBCDIC Code Page 1047\nCore perl now works on this EBCDIC platform. Earlier perls also worked, but, even though\nsupport wasn't officially withdrawn, recent perls would not compile and run well. Perl\n5.20 would work, but had many bugs which have now been fixed. Many CPAN modules that\nship with Perl still fail tests, including \"Pod::Simple\". However the version of\n\"Pod::Simple\" currently on CPAN should work; it was fixed too late to include in Perl\n5.22. Work is under way to fix many of the still-broken CPAN modules, which likely will\nbe installed on CPAN when completed, so that you may not have to wait until Perl 5.24 to\nget a working version.\n" }, { "name": "Discontinued Platforms", "content": "NeXTSTEP/OPENSTEP\nNeXTSTEP was a proprietary operating system bundled with NeXT's workstations in the early\nto mid 90s; OPENSTEP was an API specification that provided a NeXTSTEP-like environment\non a non-NeXTSTEP system. Both are now long dead, so support for building Perl on them\nhas been removed.\n" }, { "name": "Platform-Specific Notes", "content": "EBCDIC\nSpecial handling is required of the perl interpreter on EBCDIC platforms to get\n\"qr/[i-j]/\" to match only \"i\" and \"j\", since there are 7 characters between the code\npoints for \"i\" and \"j\". This special handling had only been invoked when both ends of\nthe range are literals. Now it is also invoked if any of the \"\\N{...}\" forms for\nspecifying a character by name or Unicode code point is used instead of a literal. See\n\"Character Ranges\" in perlrecharclass.\n\nHP-UX\nThe archname now distinguishes use64bitint from use64bitall.\n\nAndroid\nBuild support has been improved for cross-compiling in general and for Android in\nparticular.\n\nVMS\n• When spawning a subprocess without waiting, the return value is now the correct PID.\n\n• Fix a prototype so linking doesn't fail under the VMS C++ compiler.\n\n• \"finite\", \"finitel\", and \"isfinite\" detection has been added to \"configure.com\",\nenvironment handling has had some minor changes, and a fix for legacy feature\nchecking status.\n\nWin32\n• miniperl.exe is now built with \"-fno-strict-aliasing\", allowing 64-bit builds to\ncomplete on GCC 4.8. [GH #14556] \n\n• \"nmake minitest\" now works on Win32. Due to dependency issues you need to build\n\"nmake test-prep\" first, and a small number of the tests fail. [GH #14318]\n\n\n• Perl can now be built in C++ mode on Windows by setting the makefile macro\n\"USECPLUSPLUS\" to the value \"define\".\n\n• The list form of piped open has been implemented for Win32. Note: unlike \"system\nLIST\" this does not fall back to the shell. [GH #13574]\n\n\n• New \"DebugSymbols\" and \"DebugFull\" configuration options added to Windows makefiles.\n\n• Previously, compiling XS modules (including CPAN ones) using Visual C++ for Win64\nresulted in around a dozen warnings per file from hvfunc.h. These warnings have\nbeen silenced.\n\n• Support for building without PerlIO has been removed from the Windows makefiles.\nNon-PerlIO builds were all but deprecated in Perl 5.18.0 and are already not\nsupported by Configure on POSIX systems.\n\n• Between 2 and 6 milliseconds and seven I/O calls have been saved per attempt to open\na perl module for each path in @INC.\n\n• Intel C builds are now always built with C99 mode on.\n\n• %I64d is now being used instead of %lld for MinGW.\n\n• In the experimental \":win32\" layer, a crash in \"open\" was fixed. Also opening\n/dev/null (which works under Win32 Perl's default \":unix\" layer) was implemented for\n\":win32\". [GH #13968] \n\n• A new makefile option, \"USELONGDOUBLE\", has been added to the Windows dmake\nmakefile for gcc builds only. Set this to \"define\" if you want perl to use long\ndoubles to give more accuracy and range for floating point numbers.\n\nOpenBSD\nOn OpenBSD, Perl will now default to using the system \"malloc\" due to the security\nfeatures it provides. Perl's own malloc wrapper has been in use since v5.14 due to\nperformance reasons, but the OpenBSD project believes the tradeoff is worth it and would\nprefer that users who need the speed specifically ask for it.\n\n[GH #13888] .\n\nSolaris\n• We now look for the Sun Studio compiler in both /opt/solstudio* and\n/opt/solarisstudio*.\n\n• Builds on Solaris 10 with \"-Dusedtrace\" would fail early since make didn't follow\nimplied dependencies to build \"perldtrace.h\". Added an explicit dependency to\n\"depend\". [GH #13334] \n\n• C99 options have been cleaned up; hints look for \"solstudio\" as well as \"SUNWspro\";\nand support for native \"setenv\" has been added.\n" }, { "name": "Internal Changes", "content": "• Experimental support has been added to allow ops in the optree to locate their parent, if\nany. This is enabled by the non-default build option \"-DPERLOPPARENT\". It is envisaged\nthat this will eventually become enabled by default, so XS code which directly accesses\nthe \"opsibling\" field of ops should be updated to be future-proofed.\n\nOn \"PERLOPPARENT\" builds, the \"opsibling\" field has been renamed \"opsibparent\" and a\nnew flag, \"opmoresib\", added. On the last op in a sibling chain, \"opmoresib\" is false\nand \"opsibparent\" points to the parent (if any) rather than being \"NULL\".\n\nTo make existing code work transparently whether using \"PERLOPPARENT\" or not, a number\nof new macros and functions have been added that should be used, rather than directly\nmanipulating \"opsibling\".\n\nFor the case of just reading \"opsibling\" to determine the next sibling, two new macros\nhave been added. A simple scan through a sibling chain like this:\n\nfor (; kid->opsibling; kid = kid->opsibling) { ... }\n\nshould now be written as:\n\nfor (; OpHASSIBLING(kid); kid = OpSIBLING(kid)) { ... }\n\nFor altering optrees, a general-purpose function \"opsiblingsplice()\" has been added,\nwhich allows for manipulation of a chain of sibling ops. By analogy with the Perl\nfunction \"splice()\", it allows you to cut out zero or more ops from a sibling chain and\nreplace them with zero or more new ops. It transparently handles all the updating of\nsibling, parent, oplast pointers etc.\n\nIf you need to manipulate ops at a lower level, then three new macros, \"OpMORESIBset\",\n\"OpLASTSIBset\" and \"OpMAYBESIBset\" are intended to be a low-level portable way to set\n\"opsibling\" / \"opsibparent\" while also updating \"opmoresib\". The first sets the\nsibling pointer to a new sibling, the second makes the op the last sibling, and the third\nconditionally does the first or second action. Note that unlike \"opsiblingsplice()\"\nthese macros won't maintain consistency in the parent at the same time (e.g. by updating\n\"opfirst\" and \"oplast\" where appropriate).\n\nA C-level \"Perlopparent()\" function and a Perl-level \"B::OP::parent()\" method have been\nadded. The C function only exists under \"PERLOPPARENT\" builds (using it is build-time\nerror on vanilla perls). \"B::OP::parent()\" exists always, but on a vanilla build it\nalways returns \"NULL\". Under \"PERLOPPARENT\", they return the parent of the current op,\nif any. The variable $B::OP::doesparent allows you to determine whether \"B\" supports\nretrieving an op's parent.\n\n\"PERLOPPARENT\" was introduced in 5.21.2, but the interface was changed considerably in\n5.21.11. If you updated your code before the 5.21.11 changes, it may require further\nrevision. The main changes after 5.21.2 were:\n\n• The \"OPSIBLING\" and \"OPHASSIBLING\" macros have been renamed \"OpSIBLING\" and\n\"OpHASSIBLING\" for consistency with other op-manipulating macros.\n\n• The \"oplastsib\" field has been renamed \"opmoresib\", and its meaning inverted.\n\n• The macro \"OpSIBLINGset\" has been removed, and has been superseded by\n\"OpMORESIBset\" et al.\n\n• The \"opsiblingsplice()\" function now accepts a null \"parent\" argument where the\nsplicing doesn't affect the first or last ops in the sibling chain\n\n• Macros have been created to allow XS code to better manipulate the POSIX locale category\n\"LCNUMERIC\". See \"Locale-related functions and macros\" in perlapi.\n\n• The previous \"atoi\" et al replacement function, \"grokatou\", has now been superseded by\n\"grokatoUV\". See perlclib for details.\n\n• A new function, \"Perlsvgetbackrefs()\", has been added which allows you retrieve the\nweak references, if any, which point at an SV.\n\n• The \"screaminstr()\" function has been removed. Although marked as public API, it was\nundocumented and had no usage in CPAN modules. Calling it has been fatal since 5.17.0.\n\n• The \"newDEFSVOP()\", \"blockstart()\", \"blockend()\" and \"intromy()\" functions have been\nadded to the API.\n\n• The internal \"convert\" function in op.c has been renamed \"opconvertlist\" and added to\nthe API.\n\n• The \"svmagic()\" function no longer forbids \"ext\" magic on read-only values. After all,\nperl can't know whether the custom magic will modify the SV or not. [GH #14202]\n.\n\n• Accessing \"CvPADLIST\" in perlapi on an XSUB is now forbidden.\n\nThe \"CvPADLIST\" field has been reused for a different internal purpose for XSUBs. So in\nparticular, you can no longer rely on it being NULL as a test of whether a CV is an XSUB.\nUse \"CvISXSUB()\" instead.\n\n• SVs of type \"SVtNV\" are now sometimes bodiless when the build configuration and platform\nallow it: specifically, when \"sizeof(NV) <= sizeof(IV)\". \"Bodiless\" means that the NV\nvalue is stored directly in the head of an SV, without requiring a separate body to be\nallocated. This trick has already been used for IVs since 5.9.2 (though in the case of\nIVs, it is always used, regardless of platform and build configuration).\n\n• The $DB::single, $DB::signal and $DB::trace variables now have set- and get-magic that\nstores their values as IVs, and those IVs are used when testing their values in\n\"ppdbstate()\". This prevents perl from recursing infinitely if an overloaded object is\nassigned to any of those variables. [GH #14013]\n.\n\n• \"Perltmpsgrow()\", which is marked as public API but is undocumented, has been removed\nfrom the public API. This change does not affect XS code that uses the \"EXTENDMORTAL\"\nmacro to pre-extend the mortal stack.\n\n• Perl's internals no longer sets or uses the \"SVsPADMY\" flag. \"SvPADMY()\" now returns a\ntrue value for anything not marked \"PADTMP\" and \"SVsPADMY\" is now defined as 0.\n\n• The macros \"SETsv\" and \"SETsvUN\" have been removed. They were no longer used in the core\nsince commit 6f1401dc2a five years ago, and have not been found present on CPAN.\n\n• The \"SvFAKE\" bit (unused on HVs) got informally reserved by David Mitchell for future\nwork on vtables.\n\n• The \"svcatpvnflags()\" function accepts \"SVCATBYTES\" and \"SVCATUTF8\" flags, which\nspecify whether the appended string is bytes or UTF-8, respectively. (These flags have in\nfact been present since 5.16.0, but were formerly not regarded as part of the API.)\n\n• A new opcode class, \"METHOP\", has been introduced. It holds information used at runtime\nto improve the performance of class/object method calls.\n\n\"OPMETHOD\" and \"OPMETHODNAMED\" have changed from being \"UNOP/SVOP\" to being \"METHOP\".\n\n• \"cvname()\" is a new API function that can be passed a CV or GV. It returns an SV\ncontaining the name of the subroutine, for use in diagnostics.\n\n[GH #12767] [GH #13392]\n\n\n• \"cvsetcallcheckerflags()\" is a new API function that works like\n\"cvsetcallchecker()\", except that it allows the caller to specify whether the call\nchecker requires a full GV for reporting the subroutine's name, or whether it could be\npassed a CV instead. Whatever value is passed will be acceptable to \"cvname()\".\n\"cvsetcallchecker()\" guarantees there will be a GV, but it may have to create one on\nthe fly, which is inefficient. [GH #12767] \n\n• \"CvGV\" (which is not part of the API) is now a more complex macro, which may call a\nfunction and reify a GV. For those cases where it has been used as a boolean, \"CvHASGV\"\nhas been added, which will return true for CVs that notionally have GVs, but without\nreifying the GV. \"CvGV\" also returns a GV now for lexical subs. [GH #13392]\n\n\n• The \"synclocale\" in perlapi function has been added to the public API. Changing the\nprogram's locale should be avoided by XS code. Nevertheless, certain non-Perl libraries\ncalled from XS need to do so, such as \"Gtk\". When this happens, Perl needs to be told\nthat the locale has changed. Use this function to do so, before returning to Perl.\n\n• The defines and labels for the flags in the \"opprivate\" field of OPs are now auto-\ngenerated from data in regen/opprivate. The noticeable effect of this is that some of\nthe flag output of \"Concise\" might differ slightly, and the flag output of \"perl -Dx\" may\ndiffer considerably (they both use the same set of labels now). Also, debugging builds\nnow have a new assertion in \"opfree()\" to ensure that the op doesn't have any\nunrecognized flags set in \"opprivate\".\n\n• The deprecated variable \"PLsvobjcount\" has been removed.\n\n• Perl now tries to keep the locale category \"LCNUMERIC\" set to \"C\" except around\noperations that need it to be set to the program's underlying locale. This protects the\nmany XS modules that cannot cope with the decimal radix character not being a dot. Prior\nto this release, Perl initialized this category to \"C\", but a call to\n\"POSIX::setlocale()\" would change it. Now such a call will change the underlying locale\nof the \"LCNUMERIC\" category for the program, but the locale exposed to XS code will\nremain \"C\". There are new macros to manipulate the LCNUMERIC locale, including\n\"STORELCNUMERICSETTONEEDED\" and \"STORELCNUMERICFORCETOUNDERLYING\". See\n\"Locale-related functions and macros\" in perlapi.\n\n• A new macro \"isUTF8CHAR\" has been written which efficiently determines if the string\ngiven by its parameters begins with a well-formed UTF-8 encoded character.\n\n• The following private API functions had their context parameter removed:\n\"Perlcastulong\", \"Perlcasti32\", \"Perlcastiv\", \"Perlcastuv\",\n\"Perlcvconstsv\", \"Perlmgfind\", \"Perlmgfindext\", \"Perlmgmagical\",\n\"Perlminimktime\", \"Perlmydirfd\", \"Perlsvbackoff\", \"Perlutf8hop\".\n\nNote that the prefix-less versions of those functions that are part of the public API,\nsuch as \"casti32()\", remain unaffected.\n\n• The \"PADNAME\" and \"PADNAMELIST\" types are now separate types, and no longer simply\naliases for SV and AV. [GH #14250] .\n\n• Pad names are now always UTF-8. The \"PadnameUTF8\" macro always returns true.\nPreviously, this was effectively the case already, but any support for two different\ninternal representations of pad names has now been removed.\n\n• A new op class, \"UNOPAUX\", has been added. This is a subclass of \"UNOP\" with an \"opaux\"\nfield added, which points to an array of unions of UV, SV* etc. It is intended for where\nan op needs to store more data than a simple \"opsv\" or whatever. Currently the only op\nof this type is \"OPMULTIDEREF\" (see next item).\n\n• A new op has been added, \"OPMULTIDEREF\", which performs one or more nested array and\nhash lookups where the key is a constant or simple variable. For example the expression\n$a[0]{$k}[$i], which previously involved ten \"rv2Xv\", \"Xelem\", \"gvsv\" and \"const\" ops is\nnow performed by a single \"multideref\" op. It can also handle \"local\", \"exists\" and\n\"delete\". A non-simple index expression, such as \"[$i+1]\" is still done using\n\"aelem\"/\"helem\", and single-level array lookup with a small constant index is still done\nusing \"aelemfast\".\n" }, { "name": "Selected Bug Fixes", "content": "• \"close\" now sets $!\n\nWhen an I/O error occurs, the fact that there has been an error is recorded in the\nhandle. \"close\" returns false for such a handle. Previously, the value of $! would be\nuntouched by \"close\", so the common convention of writing \"close $fh or die $!\" did not\nwork reliably. Now the handle records the value of $!, too, and \"close\" restores it.\n\n• \"no re\" now can turn off everything that \"use re\" enables\n\nPreviously, running \"no re\" would turn off only a few things. Now it can turn off all the\nenabled things. For example, the only way to stop debugging, once enabled, was to exit\nthe enclosing block; that is now fixed.\n\n• \"pack(\"D\", $x)\" and \"pack(\"F\", $x)\" now zero the padding on x86 long double builds.\nUnder some build options on GCC 4.8 and later, they used to either overwrite the zero-\ninitialized padding, or bypass the initialized buffer entirely. This caused op/pack.t to\nfail. [GH #14554] \n\n• Extending an array cloned from a parent thread could result in \"Modification of a read-\nonly value attempted\" errors when attempting to modify the new elements. [GH #14605]\n\n\n• An assertion failure and subsequent crash with \"*x=\" has been fixed. [GH #14493]\n\n\n• A possible crashing/looping bug related to compiling lexical subs has been fixed. [GH\n#14596] \n\n• UTF-8 now works correctly in function names, in unquoted HERE-document terminators, and\nin variable names used as array indexes. [GH #14601]\n\n\n• Repeated global pattern matches in scalar context on large tainted strings were\nexponentially slow depending on the current match position in the string. [GH #14238]\n\n\n• Various crashes due to the parser getting confused by syntax errors have been fixed. [GH\n#14496] [GH #14497]\n [GH #14548]\n [GH #14564]\n\n\n• \"split\" in the scope of lexical $ has been fixed not to fail assertions. [GH #14483]\n\n\n• \"my $x : attr\" syntax inside various list operators no longer fails assertions. [GH\n#14500] \n\n• An \"@\" sign in quotes followed by a non-ASCII digit (which is not a valid identifier)\nwould cause the parser to crash, instead of simply trying the \"@\" as literal. This has\nbeen fixed. [GH #14553] \n\n• \"*bar::=*foo::=*globwithhash\" has been crashing since Perl 5.14, but no longer does.\n[GH #14512] \n\n• \"foreach\" in scalar context was not pushing an item on to the stack, resulting in bugs.\n(\"print 4, scalar do { foreach(@x){} } + 1\" would print 5.) It has been fixed to return\n\"undef\". [GH #14569] \n\n• Several cases of data used to store environment variable contents in core C code being\npotentially overwritten before being used have been fixed. [GH #14476]\n\n\n• Some patterns starting with \"/.*..../\" matched against long strings have been slow since\nv5.8, and some of the form \"/.*..../i\" have been slow since v5.18. They are now all fast\nagain. [GH #14475] .\n\n• The original visible value of $/ is now preserved when it is set to an invalid value.\nPreviously if you set $/ to a reference to an array, for example, perl would produce a\nruntime error and not set \"PLrs\", but Perl code that checked $/ would see the array\nreference. [GH #14245] .\n\n• In a regular expression pattern, a POSIX class, like \"[:ascii:]\", must be inside a\nbracketed character class, like \"qr/[[:ascii:]]/\". A warning is issued when something\nlooking like a POSIX class is not inside a bracketed class. That warning wasn't getting\ngenerated when the POSIX class was negated: \"[:^ascii:]\". This is now fixed.\n\n• Perl 5.14.0 introduced a bug whereby \"eval { LABEL: }\" would crash. This has been fixed.\n[GH #14438] .\n\n• Various crashes due to the parser getting confused by syntax errors have been fixed. [GH\n#14421] . [GH #14472]\n. [GH #14480]\n. [GH #14447]\n.\n\n• Code like \"/$a[/\" used to read the next line of input and treat it as though it came\nimmediately after the opening bracket. Some invalid code consequently would parse and\nrun, but some code caused crashes, so this is now disallowed. [GH #14462]\n.\n\n• Fix argument underflow for \"pack\". [GH #14525]\n.\n\n• Fix handling of non-strict \"\\x{}\". Now \"\\x{}\" is equivalent to \"\\x{0}\" instead of\nfaulting.\n\n• \"stat -t\" is now no longer treated as stackable, just like \"-t stat\". [GH #14499]\n.\n\n• The following no longer causes a SEGV: \"qr{x+(y(?0))*}\".\n\n• Fixed infinite loop in parsing backrefs in regexp patterns.\n\n• Several minor bug fixes in behavior of Infinity and NaN, including warnings when\nstringifying Infinity-like or NaN-like strings. For example, \"NaNcy\" doesn't numify to\nNaN anymore.\n\n• A bug in regular expression patterns that could lead to segfaults and other crashes has\nbeen fixed. This occurred only in patterns compiled with \"/i\" while taking into account\nthe current POSIX locale (which usually means they have to be compiled within the scope\nof \"use locale\"), and there must be a string of at least 128 consecutive bytes to match.\n[GH #14389] .\n\n• \"s///g\" now works on very long strings (where there are more than 2 billion iterations)\ninstead of dying with 'Substitution loop'. [GH #11742]\n. [GH #14190]\n.\n\n• \"gmtime\" no longer crashes with not-a-number values. [GH #14365]\n.\n\n• \"\\()\" (a reference to an empty list), and \"y///\" with lexical $ in scope, could both do\na bad write past the end of the stack. They have both been fixed to extend the stack\nfirst.\n\n• \"prototype()\" with no arguments used to read the previous item on the stack, so\n\"print \"foo\", prototype()\" would print foo's prototype. It has been fixed to infer $\ninstead. [GH #14376] .\n\n• Some cases of lexical state subs declared inside predeclared subs could crash, for\nexample when evalling a string including the name of an outer variable, but no longer do.\n\n• Some cases of nested lexical state subs inside anonymous subs could cause 'Bizarre copy'\nerrors or possibly even crashes.\n\n• When trying to emit warnings, perl's default debugger (perl5db.pl) was sometimes giving\n'Undefined subroutine &DB::dbwarn called' instead. This bug, which started to occur in\nPerl 5.18, has been fixed. [GH #14400] .\n\n• Certain syntax errors in substitutions, such as \"s/${<>{})//\", would crash, and had done\nso since Perl 5.10. (In some cases the crash did not start happening till 5.16.) The\ncrash has, of course, been fixed. [GH #14391]\n.\n\n• Fix a couple of string grow size calculation overflows; in particular, a repeat\nexpression like \"33 x ~3\" could cause a large buffer overflow since the new output buffer\nsize was not correctly handled by \"SvGROW()\". An expression like this now properly\nproduces a memory wrap panic. [GH #14401] .\n\n• \"formline(\"@...\", \"a\");\" would crash. The \"FFCHECKNL\" case in \"ppformline()\" didn't\nset the pointer used to mark the chop position, which led to the \"FFMORE\" case crashing\nwith a segmentation fault. This has been fixed. [GH #14388]\n.\n\n• A possible buffer overrun and crash when parsing a literal pattern during regular\nexpression compilation has been fixed. [GH #14416]\n.\n\n• \"fchmod()\" and \"futimes()\" now set $! when they fail due to being passed a closed file\nhandle. [GH #14073] .\n\n• \"opfree()\" and \"scalarvoid()\" no longer crash due to a stack overflow when freeing a\ndeeply recursive op tree. [GH #11866] .\n\n• In Perl 5.20.0, $^N accidentally had the internal UTF-8 flag turned off if accessed from\na code block within a regular expression, effectively UTF-8-encoding the value. This has\nbeen fixed. [GH #14211] .\n\n• A failed \"semctl\" call no longer overwrites existing items on the stack, which means that\n\"(semctl(-1,0,0,0))[0]\" no longer gives an \"uninitialized\" warning.\n\n• \"else{foo()}\" with no space before \"foo\" is now better at assigning the right line number\nto that statement. [GH #14070] .\n\n• Sometimes the assignment in \"@array = split\" gets optimised so that \"split\" itself writes\ndirectly to the array. This caused a bug, preventing this assignment from being used in\nlvalue context. So \"(@a=split//,\"foo\")=bar()\" was an error. (This bug probably goes\nback to Perl 3, when the optimisation was added.) It has now been fixed. [GH #14183]\n.\n\n• When an argument list fails the checks specified by a subroutine signature (which is\nstill an experimental feature), the resulting error messages now give the file and line\nnumber of the caller, not of the called subroutine. [GH #13643]\n.\n\n• The flip-flop operators (\"..\" and \"...\" in scalar context) used to maintain a separate\nstate for each recursion level (the number of times the enclosing sub was called\nrecursively), contrary to the documentation. Now each closure has one internal state for\neach flip-flop. [GH #14110] .\n\n• The flip-flop operator (\"..\" in scalar context) would return the same scalar each time,\nunless the containing subroutine was called recursively. Now it always returns a new\nscalar. [GH #14110] .\n\n• \"use\", \"no\", statement labels, special blocks (\"BEGIN\") and pod are now permitted as the\nfirst thing in a \"map\" or \"grep\" block, the block after \"print\" or \"say\" (or other\nfunctions) returning a handle, and within \"${...}\", \"@{...}\", etc. [GH #14088]\n.\n\n• The repetition operator \"x\" now propagates lvalue context to its left-hand argument when\nused in contexts like \"foreach\". That allows \"for(($#thatarray)x2) { ... }\" to work as\nexpected if the loop modifies $.\n\n• \"(...) x ...\" in scalar context used to corrupt the stack if one operand was an object\nwith \"x\" overloading, causing erratic behavior. [GH #13811]\n.\n\n• Assignment to a lexical scalar is often optimised away; for example in \"my $x; $x = $y +\n$z\", the assign operator is optimised away and the add operator writes its result\ndirectly to $x. Various bugs related to this optimisation have been fixed. Certain\noperators on the right-hand side would sometimes fail to assign the value at all or\nassign the wrong value, or would call STORE twice or not at all on tied variables. The\noperators affected were \"$foo++\", \"$foo--\", and \"-$foo\" under \"use integer\", \"chomp\",\n\"chr\" and \"setpgrp\".\n\n• List assignments were sometimes buggy if the same scalar ended up on both sides of the\nassignment due to use of \"tied\", \"values\" or \"each\". The result would be the wrong value\ngetting assigned.\n\n• \"setpgrp($nonzero)\" (with one argument) was accidentally changed in 5.16 to mean\nsetpgrp(0). This has been fixed.\n\n• \"SUB\" could return the wrong value or even corrupt memory under the debugger (the\n\"-d\" switch) and in subs containing \"eval $string\".\n\n• When \"sub () { $var }\" becomes inlinable, it now returns a different scalar each time,\njust as a non-inlinable sub would, though Perl still optimises the copy away in cases\nwhere it would make no observable difference.\n\n• \"my sub f () { $var }\" and \"sub () : attr { $var }\" are no longer eligible for inlining.\nThe former would crash; the latter would just throw the attributes away. An exception is\nmade for the little-known \":method\" attribute, which does nothing much.\n\n• Inlining of subs with an empty prototype is now more consistent than before. Previously,\na sub with multiple statements, of which all but the last were optimised away, would be\ninlinable only if it were an anonymous sub containing a string \"eval\" or \"state\"\ndeclaration or closing over an outer lexical variable (or any anonymous sub under the\ndebugger). Now any sub that gets folded to a single constant after statements have been\noptimised away is eligible for inlining. This applies to things like \"sub () { jabber()\nif DEBUG; 42 }\".\n\nSome subroutines with an explicit \"return\" were being made inlinable, contrary to the\ndocumentation, Now \"return\" always prevents inlining.\n\n• On some systems, such as VMS, \"crypt\" can return a non-ASCII string. If a scalar\nassigned to had contained a UTF-8 string previously, then \"crypt\" would not turn off the\nUTF-8 flag, thus corrupting the return value. This would happen with\n\"$lexical = crypt ...\".\n\n• \"crypt\" no longer calls \"FETCH\" twice on a tied first argument.\n\n• An unterminated here-doc on the last line of a quote-like operator (\"qq[${ <.\n\n• A small, previously intentional, memory leak in \"PERLSYSINIT\"/\"PERLSYSINIT3\" on Win32\nbuilds was fixed. This might affect embedders who repeatedly create and destroy perl\nengines within the same process.\n\n• \"POSIX::localeconv()\" now returns the data for the program's underlying locale even when\ncalled from outside the scope of \"use locale\".\n\n• \"POSIX::localeconv()\" now works properly on platforms which don't have \"LCNUMERIC\"\nand/or \"LCMONETARY\", or for which Perl has been compiled to disregard either or both of\nthese locale categories. In such circumstances, there are now no entries for the\ncorresponding values in the hash returned by \"localeconv()\".\n\n• \"POSIX::localeconv()\" now marks appropriately the values it returns as UTF-8 or not.\nPreviously they were always returned as bytes, even if they were supposed to be encoded\nas UTF-8.\n\n• On Microsoft Windows, within the scope of \"use locale\", the following POSIX character\nclasses gave results for many locales that did not conform to the POSIX standard:\n\"[[:alnum:]]\", \"[[:alpha:]]\", \"[[:blank:]]\", \"[[:digit:]]\", \"[[:graph:]]\", \"[[:lower:]]\",\n\"[[:print:]]\", \"[[:punct:]]\", \"[[:upper:]]\", \"[[:word:]]\", and \"[[:xdigit:]]\". This was\nbecause the underlying Microsoft implementation does not follow the standard. Perl now\ntakes special precautions to correct for this.\n\n• Many issues have been detected by Coverity and fixed.\n\n• \"system()\" and friends should now work properly on more Android builds.\n\nDue to an oversight, the value specified through \"-Dtargetsh\" to Configure would end up\nbeing ignored by some of the build process. This caused perls cross-compiled for Android\nto end up with defective versions of \"system()\", \"exec()\" and backticks: the commands\nwould end up looking for \"/bin/sh\" instead of \"/system/bin/sh\", and so would fail for the\nvast majority of devices, leaving $! as \"ENOENT\".\n\n• \"qr(...\\(...\\)...)\", \"qr[...\\[...\\]...]\", and \"qr{...\\{...\\}...}\" now work. Previously\nit was impossible to escape these three left-characters with a backslash within a regular\nexpression pattern where otherwise they would be considered metacharacters, and the\npattern opening delimiter was the character, and the closing delimiter was its mirror\ncharacter.\n\n• \"s///e\" on tainted UTF-8 strings corrupted \"pos()\". This bug, introduced in 5.20, is now\nfixed. [GH #13948] .\n\n• A non-word boundary in a regular expression (\"\\B\") did not always match the end of the\nstring; in particular \"q{} =~ /\\B/\" did not match. This bug, introduced in perl 5.14, is\nnow fixed. [GH #13917] .\n\n• \"\" P\" =~ /(?=.*P)P/\" should match, but did not. This is now fixed. [GH #13954]\n.\n\n• Failing to compile \"use Foo\" in an \"eval\" could leave a spurious \"BEGIN\" subroutine\ndefinition, which would produce a \"Subroutine BEGIN redefined\" warning on the next use of\n\"use\", or other \"BEGIN\" block. [GH #13926] .\n\n• \"method { BLOCK } ARGS\" syntax now correctly parses the arguments if they begin with an\nopening brace. [GH #9085] .\n\n• External libraries and Perl may have different ideas of what the locale is. This is\nproblematic when parsing version strings if the locale's numeric separator has been\nchanged. Version parsing has been patched to ensure it handles the locales correctly.\n[GH #13863] .\n\n• A bug has been fixed where zero-length assertions and code blocks inside of a regex could\ncause \"pos\" to see an incorrect value. [GH #14016]\n.\n\n• Dereferencing of constants now works correctly for typeglob constants. Previously the\nglob was stringified and its name looked up. Now the glob itself is used. [GH #9891]\n\n\n• When parsing a sigil (\"$\" \"@\" \"%\" \"&)\" followed by braces, the parser no longer tries to\nguess whether it is a block or a hash constructor (causing a syntax error when it guesses\nthe latter), since it can only be a block.\n\n• \"undef $reference\" now frees the referent immediately, instead of hanging on to it until\nthe next statement. [GH #14032] \n\n• Various cases where the name of a sub is used (autoload, overloading, error messages)\nused to crash for lexical subs, but have been fixed.\n\n• Bareword lookup now tries to avoid vivifying packages if it turns out the bareword is not\ngoing to be a subroutine name.\n\n• Compilation of anonymous constants (e.g., \"sub () { 3 }\") no longer deletes any\nsubroutine named \"ANON\" in the current package. Not only was \"*ANON{CODE}\"\ncleared, but there was a memory leak, too. This bug goes back to Perl 5.8.0.\n\n• Stub declarations like \"sub f;\" and \"sub f ();\" no longer wipe out constants of the same\nname declared by \"use constant\". This bug was introduced in Perl 5.10.0.\n\n• \"qr/[\\N{named sequence}]/\" now works properly in many instances.\n\nSome names known to \"\\N{...}\" refer to a sequence of multiple characters, instead of the\nusual single character. Bracketed character classes generally only match single\ncharacters, but now special handling has been added so that they can match named\nsequences, but not if the class is inverted or the sequence is specified as the beginning\nor end of a range. In these cases, the only behavior change from before is a slight\nrewording of the fatal error message given when this class is part of a \"?[...])\"\nconstruct. When the \"[...]\" stands alone, the same non-fatal warning as before is\nraised, and only the first character in the sequence is used, again just as before.\n\n• Tainted constants evaluated at compile time no longer cause unrelated statements to\nbecome tainted. [GH #14059] \n\n• \"open $$fh, ...\", which vivifies a handle with a name like \"main::GEN0\", was not giving\nthe handle the right reference count, so a double free could happen.\n\n• When deciding that a bareword was a method name, the parser would get confused if an\n\"our\" sub with the same name existed, and look up the method in the package of the \"our\"\nsub, instead of the package of the invocant.\n\n• The parser no longer gets confused by \"\\U=\" within a double-quoted string. It used to\nproduce a syntax error, but now compiles it correctly. [GH #10882]\n\n\n• It has always been the intention for the \"-B\" and \"-T\" file test operators to treat UTF-8\nencoded files as text. (perlfunc has been updated to say this.) Previously, it was\npossible for some files to be considered UTF-8 that actually weren't valid UTF-8. This\nis now fixed. The operators now work on EBCDIC platforms as well.\n\n• Under some conditions warning messages raised during regular expression pattern\ncompilation were being output more than once. This has now been fixed.\n\n• Perl 5.20.0 introduced a regression in which a UTF-8 encoded regular expression pattern\nthat contains a single ASCII lowercase letter did not match its uppercase counterpart.\nThat has been fixed in both 5.20.1 and 5.22.0. [GH #14051]\n\n\n• Constant folding could incorrectly suppress warnings if lexical warnings (\"use warnings\"\nor \"no warnings\") were not in effect and $^W were false at compile time and true at run\ntime.\n\n• Loading Unicode tables during a regular expression match could cause assertion failures\nunder debugging builds if the previous match used the very same regular expression. [GH\n#14081] \n\n• Thread cloning used to work incorrectly for lexical subs, possibly causing crashes or\ndouble frees on exit.\n\n• Since Perl 5.14.0, deleting $SomePackage::{ANON} and then undefining an anonymous\nsubroutine could corrupt things internally, resulting in Devel::Peek crashing or B.pm\ngiving nonsensical data. This has been fixed.\n\n• \"(caller $n)[3]\" now reports names of lexical subs, instead of treating them as\n\"(unknown)\".\n\n• \"sort subname LIST\" now supports using a lexical sub as the comparison routine.\n\n• Aliasing (e.g., via \"*x = *y\") could confuse list assignments that mention the two names\nfor the same variable on either side, causing wrong values to be assigned. [GH #5788]\n\n\n• Long here-doc terminators could cause a bad read on short lines of input. This has been\nfixed. It is doubtful that any crash could have occurred. This bug goes back to when\nhere-docs were introduced in Perl 3.000 twenty-five years ago.\n\n• An optimization in \"split\" to treat \"split /^/\" like \"split /^/m\" had the unfortunate\nside-effect of also treating \"split /\\A/\" like \"split /^/m\", which it should not. This\nhas been fixed. (Note, however, that \"split /^x/\" does not behave like \"split /^x/m\",\nwhich is also considered to be a bug and will be fixed in a future version.) [GH #14086]\n\n\n• The little-known \"my Class $var\" syntax (see fields and attributes) could get confused in\nthe scope of \"use utf8\" if \"Class\" were a constant whose value contained Latin-1\ncharacters.\n\n• Locking and unlocking values via Hash::Util or \"Internals::SvREADONLY\" no longer has any\neffect on values that were read-only to begin with. Previously, unlocking such values\ncould result in crashes, hangs or other erratic behavior.\n\n• Some unterminated \"(?(...)...)\" constructs in regular expressions would either crash or\ngive erroneous error messages. \"/(?(1)/\" is one such example.\n\n• \"pack \"w\", $tied\" no longer calls FETCH twice.\n\n• List assignments like \"($x, $z) = (1, $y)\" now work correctly if $x and $y have been\naliased by \"foreach\".\n\n• Some patterns including code blocks with syntax errors, such as \"/ (?{(^{})/\", would hang\nor fail assertions on debugging builds. Now they produce errors.\n\n• An assertion failure when parsing \"sort\" with debugging enabled has been fixed. [GH\n#14087] .\n\n• \"*a = *b; @a = split //, $b[1]\" could do a bad read and produce junk results.\n\n• In \"() = @array = split\", the \"() =\" at the beginning no longer confuses the optimizer\ninto assuming a limit of 1.\n\n• Fatal warnings no longer prevent the output of syntax errors. [GH #14155]\n.\n\n• Fixed a NaN double-to-long-double conversion error on VMS. For quiet NaNs (and only on\nItanium, not Alpha) negative infinity instead of NaN was produced.\n\n• Fixed the issue that caused \"make distclean\" to incorrectly leave some files behind. [GH\n#14108] .\n\n• AIX now sets the length in \"getsockopt\" correctly. [GH #13484]\n. [cpan #91183]\n. [cpan #85570]\n.\n\n• The optimization phase of a regexp compilation could run \"forever\" and exhaust all memory\nunder certain circumstances; now fixed. [GH #13984]\n.\n\n• The test script t/op/crypt.t now uses the SHA-256 algorithm if the default one is\ndisabled, rather than giving failures. [GH #13715]\n.\n\n• Fixed an off-by-one error when setting the size of a shared array. [GH #14151]\n.\n\n• Fixed a bug that could cause perl to enter an infinite loop during compilation. In\nparticular, a while(1) within a sublist, e.g.\n\nsub foo { () = ($a, my $b, ($c, do { while(1) {} })) }\n\nThe bug was introduced in 5.20.0 [GH #14165]\n.\n\n• On Win32, if a variable was \"local\"-ized in a pseudo-process that later forked, restoring\nthe original value in the child pseudo-process caused memory corruption and a crash in\nthe child pseudo-process (and therefore the OS process). [GH #8641]\n.\n\n• Calling \"write\" on a format with a \"^\" field could produce a panic in \"svchop()\" if\nthere were insufficient arguments or if the variable used to fill the field was empty.\n[GH #14255] .\n\n• Non-ASCII lexical sub names now appear without trailing junk when they appear in error\nmessages.\n\n• The \"\\@\" subroutine prototype no longer flattens parenthesized arrays (taking a reference\nto each element), but takes a reference to the array itself. [GH #9111]\n.\n\n• A block containing nothing except a C-style \"for\" loop could corrupt the stack, causing\nlists outside the block to lose elements or have elements overwritten. This could happen\nwith \"map { for(...){...} } ...\" and with lists containing \"do { for(...){...} }\". [GH\n#14269] .\n\n• \"scalar()\" now propagates lvalue context, so that \"for(scalar($#foo)) { ... }\" can modify\n$#foo through $.\n\n• \"qr/@array(?{block})/\" no longer dies with \"Bizarre copy of ARRAY\". [GH #14292]\n.\n\n• \"eval '$variable'\" in nested named subroutines would sometimes look up a global variable\neven with a lexical variable in scope.\n\n• In perl 5.20.0, \"sort CORE::fake\" where 'fake' is anything other than a keyword, started\nchopping off the last 6 characters and treating the result as a sort sub name. The\nprevious behavior of treating \"CORE::fake\" as a sort sub name has been restored. [GH\n#14323] .\n\n• Outside of \"use utf8\", a single-character Latin-1 lexical variable is disallowed. The\nerror message for it, \"Can't use global $foo...\", was giving garbage instead of the\nvariable name.\n\n• \"readline\" on a nonexistent handle was causing \"${^LASTFH}\" to produce a reference to an\nundefined scalar (or fail an assertion). Now \"${^LASTFH}\" ends up undefined.\n\n• \"(...) x ...\" in void context now applies scalar context to the left-hand argument,\ninstead of the context the current sub was called in. [GH #14174]\n.\n" }, { "name": "Known Problems", "content": "• \"pack\"-ing a NaN on a perl compiled with Visual C 6 does not behave properly, leading to\na test failure in t/op/infnan.t. [GH #14705]\n\n\n• A goal is for Perl to be able to be recompiled to work reasonably well on any Unicode\nversion. In Perl 5.22, though, the earliest such version is Unicode 5.1 (current is\n7.0).\n\n• EBCDIC platforms\n\n• The \"cmp\" (and hence \"sort\") operators do not necessarily give the correct results\nwhen both operands are UTF-EBCDIC encoded strings and there is a mixture of ASCII\nand/or control characters, along with other characters.\n\n• Ranges containing \"\\N{...}\" in the \"tr///\" (and \"y///\") transliteration operators are\ntreated differently than the equivalent ranges in regular expression patterns. They\nshould, but don't, cause the values in the ranges to all be treated as Unicode code\npoints, and not native ones. (\"Version 8 Regular Expressions\" in perlre gives\ndetails as to how it should work.)\n\n• Encode and encoding are mostly broken.\n\n• Many CPAN modules that are shipped with core show failing tests.\n\n• \"pack\"/\"unpack\" with \"U0\" format may not work properly.\n\n• The following modules are known to have test failures with this version of Perl. In many\ncases, patches have been submitted, so there will hopefully be new releases soon:\n\n• B::Generate version 1.50\n\n• B::Utils version 0.25\n\n• Coro version 6.42\n\n• Dancer version 1.3130\n\n• Data::Alias version 1.18\n\n• Data::Dump::Streamer version 2.38\n\n• Data::Util version 0.63\n\n• Devel::Spy version 0.07\n\n• invoker version 0.34\n\n• Lexical::Var version 0.009\n\n• LWP::ConsoleLogger version 0.000018\n\n• Mason version 2.22\n\n• NgxQueue version 0.02\n\n• Padre version 1.00\n\n• Parse::Keyword 0.08\n" } ] }, "Obituary": { "content": "Brian McCauley died on May 8, 2015. He was a frequent poster to Usenet, Perl Monks, and\nother Perl forums, and made several CPAN contributions under the nick NOBULL, including to\nthe Perl FAQ. He attended almost every YAPC::Europe, and indeed, helped organise\nYAPC::Europe 2006 and the QA Hackathon 2009. His wit and his delight in intricate systems\nwere particularly apparent in his love of board games; many Perl mongers will have fond\nmemories of playing Fluxx and other games with Brian. He will be missed.\n", "subsections": [] }, "Acknowledgements": { "content": "Perl 5.22.0 represents approximately 12 months of development since Perl 5.20.0 and contains\napproximately 590,000 lines of changes across 2,400 files from 94 authors.\n\nExcluding auto-generated files, documentation and release tools, there were approximately\n370,000 lines of changes to 1,500 .pm, .t, .c and .h files.\n\nPerl continues to flourish into its third decade thanks to a vibrant community of users and\ndevelopers. The following people are known to have contributed the improvements that became\nPerl 5.22.0:\n\nAaron Crane, Abhijit Menon-Sen, Abigail, Alberto Simões, Alex Solovey, Alex Vandiver,\nAlexandr Ciornii, Alexandre (Midnite) Jousset, Andreas König, Andreas Voegele, Andrew Fresh,\nAndy Dougherty, Anthony Heading, Aristotle Pagaltzis, brian d foy, Brian Fraser, Chad Granum,\nChris 'BinGOs' Williams, Craig A. Berry, Dagfinn Ilmari Mannsåker, Daniel Dragan, Darin\nMcBride, Dave Rolsky, David Golden, David Mitchell, David Wheeler, Dmitri Tikhonov, Doug\nBell, E. Choroba, Ed J, Eric Herman, Father Chrysostomos, George Greer, Glenn D. Golden,\nGraham Knop, H.Merijn Brand, Herbert Breunung, Hugo van der Sanden, James E Keenan, James\nMcCoy, James Raspass, Jan Dubois, Jarkko Hietaniemi, Jasmine Ngan, Jerry D. Hedden, Jim\nCromie, John Goodyear, kafka, Karen Etheridge, Karl Williamson, Kent Fredric, kmx, Lajos\nVeres, Leon Timmermans, Lukas Mai, Mathieu Arnold, Matthew Horsfall, Max Maischein, Michael\nBunk, Nicholas Clark, Niels Thykier, Niko Tyni, Norman Koch, Olivier Mengué, Peter John\nAcklam, Peter Martini, Petr Písař, Philippe Bruhat (BooK), Pierre Bogossian, Rafael Garcia-\nSuarez, Randy Stauner, Reini Urban, Ricardo Signes, Rob Hoelz, Rostislav Skudnov, Sawyer X,\nShirakata Kentaro, Shlomi Fish, Sisyphus, Slaven Rezic, Smylers, Steffen Müller, Steve Hay,\nSullivan Beck, syber, Tadeusz Sośnierz, Thomas Sibley, Todd Rinaldo, Tony Cook, Vincent Pit,\nVladimir Marek, Yaroslav Kuzmin, Yves Orton, Ævar Arnfjörð Bjarmason.\n\nThe list above is almost certainly incomplete as it is automatically generated from version\ncontrol history. In particular, it does not include the names of the (very much appreciated)\ncontributors who reported issues to the Perl bug tracker.\n\nMany of the changes included in this version originated in the CPAN modules included in\nPerl's core. We're grateful to the entire CPAN community for helping Perl to flourish.\n\nFor a more complete list of all of Perl's historical contributors, please see the AUTHORS\nfile in the Perl source distribution.\n", "subsections": [ { "name": "Reporting Bugs", "content": "If you find what you think is a bug, you might check the articles recently posted to the\ncomp.lang.perl.misc newsgroup and the perl bug database at . There may\nalso be information at , the Perl Home Page.\n\nIf you believe you have an unreported bug, please run the perlbug program included with your\nrelease. Be sure to trim your bug down to a tiny but sufficient test case. Your bug report,\nalong with the output of \"perl -V\", will be sent off to perlbug@perl.org to be analysed by\nthe Perl porting team.\n\nIf the bug you are reporting has security implications, which make it inappropriate to send\nto a publicly archived mailing list, then please send it to perl5-security-report@perl.org.\nThis points to a closed subscription unarchived mailing list, which includes all the core\ncommitters, who will be able to help assess the impact of issues, figure out a resolution,\nand help co-ordinate the release of patches to mitigate or fix the problem across all\nplatforms on which Perl is supported. Please only use this address for security issues in\nthe Perl core, not for modules independently distributed on CPAN.\n" } ] }, "SEE ALSO": { "content": "The Changes file for an explanation of how to view exhaustive details on what changed.\n\nThe INSTALL file for how to build Perl.\n\nThe README file for general stuff.\n\nThe Artistic and Copying files for copyright information.\n\n\n\nperl v5.34.0 2025-07-25 PERL5220DELTA(1)", "subsections": [] } }, "summary": "perl5220delta - what is new for perl v5.22.0", "flags": [], "examples": [], "see_also": [] }