Debian Codegraph @

upstream/debian-unstable/man1/perl5200delta.1

raw
1 .\" -*- mode: troff; coding: utf-8 -*-
2 .\" Automatically generated by Pod::Man 5.0102 (Pod::Simple 3.45)
3 .\"
4 .\" Standard preamble:
5 .\" ========================================================================
6 .de Sp \" Vertical space (when we can't use .PP)
7 .if t .sp .5v
8 .if n .sp
9 ..
10 .de Vb \" Begin verbatim text
11 .ft CW
12 .nf
13 .ne \\$1
14 ..
15 .de Ve \" End verbatim text
16 .ft R
17 .fi
18 ..
19 .\" \*(C` and \*(C' are quotes in nroff, nothing in troff, for use with C<>.
20 .ie n \{\
21 . ds C` ""
22 . ds C' ""
23 'br\}
24 .el\{\
25 . ds C`
26 . ds C'
27 'br\}
28 .\"
29 .\" Escape single quotes in literal strings from groff's Unicode transform.
30 .ie \n(.g .ds Aq \(aq
31 .el .ds Aq '
32 .\"
33 .\" If the F register is >0, we'll generate index entries on stderr for
34 .\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
35 .\" entries marked with X<> in POD. Of course, you'll have to process the
36 .\" output yourself in some meaningful fashion.
37 .\"
38 .\" Avoid warning from groff about undefined register 'F'.
39 .de IX
40 ..
41 .nr rF 0
42 .if \n(.g .if rF .nr rF 1
43 .if (\n(rF:(\n(.g==0)) \{\
44 . if \nF \{\
45 . de IX
46 . tm Index:\\$1\t\\n%\t"\\$2"
47 ..
48 . if !\nF==2 \{\
49 . nr % 0
50 . nr F 2
51 . \}
52 . \}
53 .\}
54 .rr rF
55 .\" ========================================================================
56 .\"
57 .IX Title "PERL5200DELTA 1"
58 .TH PERL5200DELTA 1 2025-11-16 "perl v5.40.1" "Perl Programmers Reference Guide"
59 .\" For nroff, turn off justification. Always turn off hyphenation; it makes
60 .\" way too many mistakes in technical documents.
61 .if n .ad l
62 .nh
63 .SH NAME
64 perl5200delta \- what is new for perl v5.20.0
65 .SH DESCRIPTION
66 .IX Header "DESCRIPTION"
67 This document describes differences between the 5.18.0 release and the
68 5.20.0 release.
69 .PP
70 If you are upgrading from an earlier release such as 5.16.0, first read
71 perl5180delta, which describes differences between 5.16.0 and 5.18.0.
72 .SH "Core Enhancements"
73 .IX Header "Core Enhancements"
74 .SS "Experimental Subroutine signatures"
75 .IX Subsection "Experimental Subroutine signatures"
76 Declarative syntax to unwrap argument list into lexical variables.
77 \&\f(CW\*(C`sub foo ($a,$b) {...}\*(C'\fR checks the number of arguments and puts the
78 arguments into lexical variables. Signatures are not equivalent to
79 the existing idiom of \f(CW\*(C`sub foo { my($a,$b) = @_; ... }\*(C'\fR. Signatures
80 are only available by enabling a non-default feature, and generate
81 warnings about being experimental. The syntactic clash with
82 prototypes is managed by disabling the short prototype syntax when
83 signatures are enabled.
84 .PP
85 See "Signatures" in perlsub for details.
86 .ie n .SS """sub""s now take a ""prototype"" attribute"
87 .el .SS "\f(CWsub\fPs now take a \f(CWprototype\fP attribute"
88 .IX Subsection "subs now take a prototype attribute"
89 When declaring or defining a \f(CW\*(C`sub\*(C'\fR, the prototype can now be specified inside
90 of a \f(CW\*(C`prototype\*(C'\fR attribute instead of in parens following the name.
91 .PP
92 For example, \f(CW\*(C`sub foo($$){}\*(C'\fR could be rewritten as
93 \&\f(CW\*(C`sub foo : prototype($$){}\*(C'\fR.
94 .SS "More consistent prototype parsing"
95 .IX Subsection "More consistent prototype parsing"
96 Multiple semicolons in subroutine prototypes have long been tolerated and
97 treated as a single semicolon. There was one case where this did not
98 happen. A subroutine whose prototype begins with "*" or ";*" can affect
99 whether a bareword is considered a method name or sub call. This now
100 applies also to ";;;*".
101 .PP
102 Whitespace has long been allowed inside subroutine prototypes, so
103 \&\f(CW\*(C`sub( $ $ )\*(C'\fR is equivalent to \f(CWsub($$)\fR, but until now it was stripped
104 when the subroutine was parsed. Hence, whitespace was \fInot\fR allowed in
105 prototypes set by \f(CW\*(C`Scalar::Util::set_prototype\*(C'\fR. Now it is permitted,
106 and the parser no longer strips whitespace. This means
107 \&\f(CW\*(C`prototype &mysub\*(C'\fR returns the original prototype, whitespace and all.
108 .ie n .SS """rand"" now uses a consistent random number generator"
109 .el .SS "\f(CWrand\fP now uses a consistent random number generator"
110 .IX Subsection "rand now uses a consistent random number generator"
111 Previously perl would use a platform specific random number generator, varying
112 between the libc \fBrand()\fR, \fBrandom()\fR or \fBdrand48()\fR.
113 .PP
114 This meant that the quality of perl's random numbers would vary from platform
115 to platform, from the 15 bits of \fBrand()\fR on Windows to 48\-bits on POSIX
116 platforms such as Linux with \fBdrand48()\fR.
117 .PP
118 Perl now uses its own internal \fBdrand48()\fR implementation on all platforms. This
119 does not make perl's \f(CW\*(C`rand\*(C'\fR cryptographically secure. [perl #115928]
120 .SS "New slice syntax"
121 .IX Subsection "New slice syntax"
122 The new \f(CW%hash{...}\fR and \f(CW%array[...]\fR syntax returns a list of key/value (or
123 index/value) pairs. See "Key/Value Hash Slices" in perldata.
124 .SS "Experimental Postfix Dereferencing"
125 .IX Subsection "Experimental Postfix Dereferencing"
126 When the \f(CW\*(C`postderef\*(C'\fR feature is in effect, the following syntactical
127 equivalencies are set up:
128 .PP
129 .Vb 5
130 \& $sref\->$*; # same as ${ $sref } # interpolates
131 \& $aref\->@*; # same as @{ $aref } # interpolates
132 \& $href\->%*; # same as %{ $href }
133 \& $cref\->&*; # same as &{ $cref }
134 \& $gref\->**; # same as *{ $gref }
135 \&
136 \& $aref\->$#*; # same as $#{ $aref }
137 \&
138 \& $gref\->*{ $slot }; # same as *{ $gref }{ $slot }
139 \&
140 \& $aref\->@[ ... ]; # same as @$aref[ ... ] # interpolates
141 \& $href\->@{ ... }; # same as @$href{ ... } # interpolates
142 \& $aref\->%[ ... ]; # same as %$aref[ ... ]
143 \& $href\->%{ ... }; # same as %$href{ ... }
144 .Ve
145 .PP
146 Those marked as interpolating only interpolate if the associated
147 \&\f(CW\*(C`postderef_qq\*(C'\fR feature is also enabled. This feature is \fBexperimental\fR and
148 will trigger \f(CW\*(C`experimental::postderef\*(C'\fR\-category warnings when used, unless
149 they are suppressed.
150 .PP
151 For more information, consult the Postfix Dereference Syntax section of
152 perlref.
153 .SS "Unicode 6.3 now supported"
154 .IX Subsection "Unicode 6.3 now supported"
155 Perl now supports and is shipped with Unicode 6.3 (though Perl may be
156 recompiled with any previous Unicode release as well). A detailed list of
157 Unicode 6.3 changes is at <http://www.unicode.org/versions/Unicode6.3.0/>.
158 .ie n .SS "New ""\ep{Unicode}"" regular expression pattern property"
159 .el .SS "New \f(CW\ep{Unicode}\fP regular expression pattern property"
160 .IX Subsection "New p{Unicode} regular expression pattern property"
161 This is a synonym for \f(CW\*(C`\ep{Any}\*(C'\fR and matches the set of Unicode-defined
162 code points 0 \- 0x10FFFF.
163 .SS "Better 64\-bit support"
164 .IX Subsection "Better 64-bit support"
165 On 64\-bit platforms, the internal array functions now use 64\-bit offsets,
166 allowing Perl arrays to hold more than 2**31 elements, if you have the memory
167 available.
168 .PP
169 The regular expression engine now supports strings longer than 2**31
170 characters. [perl #112790, #116907]
171 .PP
172 The functions PerlIO_get_bufsiz, PerlIO_get_cnt, PerlIO_set_cnt and
173 PerlIO_set_ptrcnt now have SSize_t, rather than int, return values and
174 parameters.
175 .ie n .SS """use\ locale"" now works on UTF\-8 locales"
176 .el .SS "\f(CWuse\ locale\fP now works on UTF\-8 locales"
177 .IX Subsection "use locale now works on UTF-8 locales"
178 Until this release, only single-byte locales, such as the ISO 8859
179 series were supported. Now, the increasingly common multi-byte UTF\-8
180 locales are also supported. A UTF\-8 locale is one in which the
181 character set is Unicode and the encoding is UTF\-8. The POSIX
182 \&\f(CW\*(C`LC_CTYPE\*(C'\fR category operations (case changing (like \f(CWlc()\fR, \f(CW"\eU"\fR),
183 and character classification (\f(CW\*(C`\ew\*(C'\fR, \f(CW\*(C`\eD\*(C'\fR, \f(CW\*(C`qr/[[:punct:]]/\*(C'\fR)) under
184 such a locale work just as if not under locale, but instead as if under
185 \&\f(CW\*(C`use\ feature\ \*(Aqunicode_strings\*(Aq\*(C'\fR, except taint rules are followed.
186 Sorting remains by code point order in this release. [perl #56820].
187 .ie n .SS """use\ locale"" now compiles on systems without locale ability"
188 .el .SS "\f(CWuse\ locale\fP now compiles on systems without locale ability"
189 .IX Subsection "use locale now compiles on systems without locale ability"
190 Previously doing this caused the program to not compile. Within its
191 scope the program behaves as if in the "C" locale. Thus programs
192 written for platforms that support locales can run on locale-less
193 platforms without change. Attempts to change the locale away from the
194 "C" locale will, of course, fail.
195 .SS "More locale initialization fallback options"
196 .IX Subsection "More locale initialization fallback options"
197 If there was an error with locales during Perl start-up, it immediately
198 gave up and tried to use the \f(CW"C"\fR locale. Now it first tries using
199 other locales given by the environment variables, as detailed in
200 "ENVIRONMENT" in perllocale. For example, if \f(CW\*(C`LC_ALL\*(C'\fR and \f(CW\*(C`LANG\*(C'\fR are
201 both set, and using the \f(CW\*(C`LC_ALL\*(C'\fR locale fails, Perl will now try the
202 \&\f(CW\*(C`LANG\*(C'\fR locale, and only if that fails, will it fall back to \f(CW"C"\fR. On
203 Windows machines, Perl will try, ahead of using \f(CW"C"\fR, the system
204 default locale if all the locales given by environment variables fail.
205 .ie n .SS """\-DL"" runtime option now added for tracing locale setting"
206 .el .SS "\f(CW\-DL\fP runtime option now added for tracing locale setting"
207 .IX Subsection "-DL runtime option now added for tracing locale setting"
208 This is designed for Perl core developers to aid in field debugging bugs
209 regarding locales.
210 .SS "\fB\-F\fP now implies \fB\-a\fP and \fB\-a\fP implies \fB\-n\fP"
211 .IX Subsection "-F now implies -a and -a implies -n"
212 Previously \fB\-F\fR without \fB\-a\fR was a no-op, and \fB\-a\fR without \fB\-n\fR or \fB\-p\fR
213 was a no-op, with this change, if you supply \fB\-F\fR then both \fB\-a\fR and \fB\-n\fR
214 are implied and if you supply \fB\-a\fR then \fB\-n\fR is implied.
215 .PP
216 You can still use \fB\-p\fR for its extra behaviour. [perl #116190]
217 .ie n .SS "$a and $b warnings exemption"
218 .el .SS "\f(CW$a\fP and \f(CW$b\fP warnings exemption"
219 .IX Subsection "$a and $b warnings exemption"
220 The special variables \f(CW$a\fR and \f(CW$b\fR, used in \f(CW\*(C`sort\*(C'\fR, are now exempt from "used
221 once" warnings, even where \f(CW\*(C`sort\*(C'\fR is not used. This makes it easier for
222 CPAN modules to provide functions using \f(CW$a\fR and \f(CW$b\fR for similar purposes.
223 [perl #120462]
224 .SH Security
225 .IX Header "Security"
226 .SS "Avoid possible read of \fBfree()\fPd memory during parsing"
227 .IX Subsection "Avoid possible read of free()d memory during parsing"
228 It was possible that \fBfree()\fRd memory could be read during parsing in the unusual
229 circumstance of the Perl program ending with a heredoc and the last line of the
230 file on disk having no terminating newline character. This has now been fixed.
231 .SH "Incompatible Changes"
232 .IX Header "Incompatible Changes"
233 .ie n .SS """do"" can no longer be used to call subroutines"
234 .el .SS "\f(CWdo\fP can no longer be used to call subroutines"
235 .IX Subsection "do can no longer be used to call subroutines"
236 The \f(CW\*(C`do SUBROUTINE(LIST)\*(C'\fR form has resulted in a deprecation warning
237 since Perl v5.0.0, and is now a syntax error.
238 .SS "Quote-like escape changes"
239 .IX Subsection "Quote-like escape changes"
240 The character after \f(CW\*(C`\ec\*(C'\fR in a double-quoted string ("..." or qq(...))
241 or regular expression must now be a printable character and may not be
242 \&\f(CW\*(C`{\*(C'\fR.
243 .PP
244 A literal \f(CW\*(C`{\*(C'\fR after \f(CW\*(C`\eB\*(C'\fR or \f(CW\*(C`\eb\*(C'\fR is now fatal.
245 .PP
246 These were deprecated in perl v5.14.0.
247 .SS "Tainting happens under more circumstances; now conforms to documentation"
248 .IX Subsection "Tainting happens under more circumstances; now conforms to documentation"
249 This affects regular expression matching and changing the case of a
250 string (\f(CW\*(C`lc\*(C'\fR, \f(CW"\eU"\fR, \fIetc\fR.) within the scope of \f(CW\*(C`use locale\*(C'\fR.
251 The result is now tainted based on the operation, no matter what the
252 contents of the string were, as the documentation (perlsec,
253 "SECURITY" in perllocale) indicates it should. Previously, for the case
254 change operation, if the string contained no characters whose case
255 change could be affected by the locale, the result would not be tainted.
256 For example, the result of \f(CWuc()\fR on an empty string or one containing
257 only above\-Latin1 code points is now tainted, and wasn't before. This
258 leads to more consistent tainting results. Regular expression patterns
259 taint their non-binary results (like \f(CW$&\fR, \f(CW$2\fR) if and only if the
260 pattern contains elements whose matching depends on the current
261 (potentially tainted) locale. Like the case changing functions, the
262 actual contents of the string being matched now do not matter, whereas
263 formerly it did. For example, if the pattern contains a \f(CW\*(C`\ew\*(C'\fR, the
264 results will be tainted even if the match did not have to use that
265 portion of the pattern to succeed or fail, because what a \f(CW\*(C`\ew\*(C'\fR matches
266 depends on locale. However, for example, a \f(CW\*(C`.\*(C'\fR in a pattern will not
267 enable tainting, because the dot matches any single character, and what
268 the current locale is doesn't change in any way what matches and what
269 doesn't.
270 .ie n .SS """\ep{}"", ""\eP{}"" matching has changed for non-Unicode code points."
271 .el .SS "\f(CW\ep{}\fP, \f(CW\eP{}\fP matching has changed for non-Unicode code points."
272 .IX Subsection "p{}, P{} matching has changed for non-Unicode code points."
273 \&\f(CW\*(C`\ep{}\*(C'\fR and \f(CW\*(C`\eP{}\*(C'\fR are defined by Unicode only on Unicode-defined code
274 points (\f(CW\*(C`U+0000\*(C'\fR through \f(CW\*(C`U+10FFFF\*(C'\fR). Their behavior on matching
275 these legal Unicode code points is unchanged, but there are changes for
276 code points \f(CW0x110000\fR and above. Previously, Perl treated the result
277 of matching \f(CW\*(C`\ep{}\*(C'\fR and \f(CW\*(C`\eP{}\*(C'\fR against these as \f(CW\*(C`undef\*(C'\fR, which
278 translates into "false". For \f(CW\*(C`\eP{}\*(C'\fR, this was then complemented into
279 "true". A warning was supposed to be raised when this happened.
280 However, various optimizations could prevent the warning, and the
281 results were often counter-intuitive, with both a match and its seeming
282 complement being false. Now all non-Unicode code points are treated as
283 typical unassigned Unicode code points. This generally is more
284 Do-What-I-Mean. A warning is raised only if the results are arguably
285 different from a strict Unicode approach, and from what Perl used to do.
286 Code that needs to be strictly Unicode compliant can make this warning
287 fatal, and then Perl always raises the warning.
288 .PP
289 Details are in "Beyond Unicode code points" in perlunicode.
290 .ie n .SS """\ep{All}"" has been expanded to match all possible code points"
291 .el .SS "\f(CW\ep{All}\fP has been expanded to match all possible code points"
292 .IX Subsection "p{All} has been expanded to match all possible code points"
293 The Perl-defined regular expression pattern element \f(CW\*(C`\ep{All}\*(C'\fR, unused
294 on CPAN, used to match just the Unicode code points; now it matches all
295 possible code points; that is, it is equivalent to \f(CW\*(C`qr/./s\*(C'\fR. Thus
296 \&\f(CW\*(C`\ep{All}\*(C'\fR is no longer synonymous with \f(CW\*(C`\ep{Any}\*(C'\fR, which continues to
297 match just the Unicode code points, as Unicode says it should.
298 .SS "Data::Dumper's output may change"
299 .IX Subsection "Data::Dumper's output may change"
300 Depending on the data structures dumped and the settings set for
301 Data::Dumper, the dumped output may have changed from previous
302 versions.
303 .PP
304 If you have tests that depend on the exact output of Data::Dumper,
305 they may fail.
306 .PP
307 To avoid this problem in your code, test against the data structure
308 from evaluating the dumped structure, instead of the dump itself.
309 .ie n .SS "Locale decimal point character no longer leaks outside of ""use\ locale"" scope"
310 .el .SS "Locale decimal point character no longer leaks outside of \f(CWuse\ locale\fP scope"
311 .IX Subsection "Locale decimal point character no longer leaks outside of use locale scope"
312 This is actually a bug fix, but some code has come to rely on the bug
313 being present, so this change is listed here. The current locale that
314 the program is running under is not supposed to be visible to Perl code
315 except within the scope of a \f(CW\*(C`use\ locale\*(C'\fR. However, until now under
316 certain circumstances, the character used for a decimal point (often a
317 comma) leaked outside the scope. If your code is affected by this
318 change, simply add a \f(CW\*(C`use\ locale\*(C'\fR.
319 .SS "Assignments of Windows sockets error codes to $! now prefer \fIerrno.h\fP values over \fBWSAGetLastError()\fP values"
320 .IX Subsection "Assignments of Windows sockets error codes to $! now prefer errno.h values over WSAGetLastError() values"
321 In previous versions of Perl, Windows sockets error codes as returned by
322 \&\fBWSAGetLastError()\fR were assigned to $!, and some constants such as ECONNABORTED,
323 not in \fIerrno.h\fR in VC++ (or the various Windows ports of gcc) were defined to
324 corresponding WSAE* values to allow $! to be tested against the E* constants
325 exported by Errno and POSIX.
326 .PP
327 This worked well until VC++ 2010 and later, which introduced new E* constants
328 with values > 100 into \fIerrno.h\fR, including some being (re)defined by perl
329 to WSAE* values. That caused problems when linking XS code against other
330 libraries which used the original definitions of \fIerrno.h\fR constants.
331 .PP
332 To avoid this incompatibility, perl now maps WSAE* error codes to E* values
333 where possible, and assigns those values to $!. The E* constants exported by
334 Errno and POSIX are updated to match so that testing $! against them,
335 wherever previously possible, will continue to work as expected, and all E*
336 constants found in \fIerrno.h\fR are now exported from those modules with their
337 original \fIerrno.h\fR values.
338 .PP
339 In order to avoid breakage in existing Perl code which assigns WSAE* values to
340 $!, perl now intercepts the assignment and performs the same mapping to E*
341 values as it uses internally when assigning to $! itself.
342 .PP
343 However, one backwards-incompatibility remains: existing Perl code which
344 compares $! against the numeric values of the WSAE* error codes that were
345 previously assigned to $! will now be broken in those cases where a
346 corresponding E* value has been assigned instead. This is only an issue for
347 those E* values < 100, which were always exported from Errno and
348 POSIX with their original \fIerrno.h\fR values, and therefore could not be used
349 for WSAE* error code tests (e.g. WSAEINVAL is 10022, but the corresponding
350 EINVAL is 22). (E* values > 100, if present, were redefined to WSAE*
351 values anyway, so compatibility can be achieved by using the E* constants,
352 which will work both before and after this change, albeit using different
353 numeric values under the hood.)
354 .ie n .SS "Functions ""PerlIO_vsprintf"" and ""PerlIO_sprintf"" have been removed"
355 .el .SS "Functions \f(CWPerlIO_vsprintf\fP and \f(CWPerlIO_sprintf\fP have been removed"
356 .IX Subsection "Functions PerlIO_vsprintf and PerlIO_sprintf have been removed"
357 These two functions, undocumented, unused in CPAN, and problematic, have been
358 removed.
359 .SH Deprecations
360 .IX Header "Deprecations"
361 .ie n .SS "The ""/\eC/"" character class"
362 .el .SS "The \f(CW/\eC/\fP character class"
363 .IX Subsection "The /C/ character class"
364 The \f(CW\*(C`/\eC/\*(C'\fR regular expression character class is deprecated. From perl
365 5.22 onwards it will generate a warning, and from perl 5.24 onwards it
366 will be a regular expression compiler error. If you need to examine the
367 individual bytes that make up a UTF8\-encoded character, then use
368 \&\f(CWutf8::encode()\fR on the string (or a copy) first.
369 .SS "Literal control characters in variable names"
370 .IX Subsection "Literal control characters in variable names"
371 This deprecation affects things like $\ecT, where \ecT is a literal control (such
372 as a \f(CW\*(C`NAK\*(C'\fR or \f(CW\*(C`NEGATIVE ACKNOWLEDGE\*(C'\fR character) in
373 the source code. Surprisingly, it appears that originally this was intended as
374 the canonical way of accessing variables like $^T, with the caret form only
375 being added as an alternative.
376 .PP
377 The literal control form is being deprecated for two main reasons. It has what
378 are likely unfixable bugs, such as $\ecI not working as an alias for $^I, and
379 their usage not being portable to non-ASCII platforms: While $^T will work
380 everywhere, \ecT is whitespace in EBCDIC. [perl #119123]
381 .ie n .SS "References to non-integers and non-positive integers in $/"
382 .el .SS "References to non-integers and non-positive integers in \f(CW$/\fP"
383 .IX Subsection "References to non-integers and non-positive integers in $/"
384 Setting \f(CW$/\fR to a reference to zero or a reference to a negative integer is
385 now deprecated, and will behave \fBexactly\fR as though it was set to \f(CW\*(C`undef\*(C'\fR.
386 If you want slurp behavior set \f(CW$/\fR to \f(CW\*(C`undef\*(C'\fR explicitly.
387 .PP
388 Setting \f(CW$/\fR to a reference to a non integer is now forbidden and will
389 throw an error. Perl has never documented what would happen in this
390 context and while it used to behave the same as setting \f(CW$/\fR to
391 the address of the references in future it may behave differently, so we
392 have forbidden this usage.
393 .SS "Character matching routines in POSIX"
394 .IX Subsection "Character matching routines in POSIX"
395 Use of any of these functions in the \f(CW\*(C`POSIX\*(C'\fR module is now deprecated:
396 \&\f(CW\*(C`isalnum\*(C'\fR, \f(CW\*(C`isalpha\*(C'\fR, \f(CW\*(C`iscntrl\*(C'\fR, \f(CW\*(C`isdigit\*(C'\fR, \f(CW\*(C`isgraph\*(C'\fR, \f(CW\*(C`islower\*(C'\fR,
397 \&\f(CW\*(C`isprint\*(C'\fR, \f(CW\*(C`ispunct\*(C'\fR, \f(CW\*(C`isspace\*(C'\fR, \f(CW\*(C`isupper\*(C'\fR, and \f(CW\*(C`isxdigit\*(C'\fR. The
398 functions are buggy and don't work on UTF\-8 encoded strings. See their
399 entries in POSIX for more information.
400 .PP
401 A warning is raised on the first call to any of them from each place in
402 the code that they are called. (Hence a repeated statement in a loop
403 will raise just the one warning.)
404 .SS "Interpreter-based threads are now \fIdiscouraged\fP"
405 .IX Subsection "Interpreter-based threads are now discouraged"
406 The "interpreter-based threads" provided by Perl are not the fast, lightweight
407 system for multitasking that one might expect or hope for. Threads are
408 implemented in a way that make them easy to misuse. Few people know how to
409 use them correctly or will be able to provide help.
410 .PP
411 The use of interpreter-based threads in perl is officially
412 discouraged.
413 .SS "Module removals"
414 .IX Subsection "Module removals"
415 The following modules will be removed from the core distribution in a
416 future release, and will at that time need to be installed from CPAN.
417 Distributions on CPAN which require these modules will need to list them as
418 prerequisites.
419 .PP
420 The core versions of these modules will now issue \f(CW"deprecated"\fR\-category
421 warnings to alert you to this fact. To silence these deprecation warnings,
422 install the modules in question from CPAN.
423 .PP
424 Note that the planned removal of these modules from core does not reflect a
425 judgement about the quality of the code and should not be taken as a suggestion
426 that their use be halted. Their disinclusion from core primarily hinges on
427 their necessity to bootstrapping a fully functional, CPAN-capable Perl
428 installation, not on concerns over their design.
429 .IP "CGI and its associated CGI:: packages" 4
430 .IX Item "CGI and its associated CGI:: packages"
431 .PD 0
432 .IP inc::latest 4
433 .IX Item "inc::latest"
434 .IP Package::Constants 4
435 .IX Item "Package::Constants"
436 .IP "Module::Build and its associated Module::Build:: packages" 4
437 .IX Item "Module::Build and its associated Module::Build:: packages"
438 .PD
439 .SS "Utility removals"
440 .IX Subsection "Utility removals"
441 The following utilities will be removed from the core distribution in a
442 future release, and will at that time need to be installed from CPAN.
443 .IP find2perl 4
444 .IX Item "find2perl"
445 .PD 0
446 .IP s2p 4
447 .IX Item "s2p"
448 .IP a2p 4
449 .IX Item "a2p"
450 .PD
451 .SH "Performance Enhancements"
452 .IX Header "Performance Enhancements"
453 .IP \(bu 4
454 Perl has a new copy-on-write mechanism that avoids the need to copy the
455 internal string buffer when assigning from one scalar to another. This
456 makes copying large strings appear much faster. Modifying one of the two
457 (or more) strings after an assignment will force a copy internally. This
458 makes it unnecessary to pass strings by reference for efficiency.
459 .Sp
460 This feature was already available in 5.18.0, but wasn't enabled by
461 default. It is the default now, and so you no longer need build perl with
462 the \fIConfigure\fR argument:
463 .Sp
464 .Vb 1
465 \& \-Accflags=\-DPERL_NEW_COPY_ON_WRITE
466 .Ve
467 .Sp
468 It can be disabled (for now) in a perl build with:
469 .Sp
470 .Vb 1
471 \& \-Accflags=\-DPERL_NO_COW
472 .Ve
473 .Sp
474 On some operating systems Perl can be compiled in such a way that any
475 attempt to modify string buffers shared by multiple SVs will crash. This
476 way XS authors can test that their modules handle copy-on-write scalars
477 correctly. See "Copy on Write" in perlguts for detail.
478 .IP \(bu 4
479 Perl has an optimizer for regular expression patterns. It analyzes the pattern
480 to find things such as the minimum length a string has to be to match, etc. It
481 now better handles code points that are above the Latin1 range.
482 .IP \(bu 4
483 Executing a regex that contains the \f(CW\*(C`^\*(C'\fR anchor (or its variant under the
484 \&\f(CW\*(C`/m\*(C'\fR flag) has been made much faster in several situations.
485 .IP \(bu 4
486 Precomputed hash values are now used in more places during method lookup.
487 .IP \(bu 4
488 Constant hash key lookups (\f(CW$hash{key}\fR as opposed to \f(CW$hash{$key}\fR) have
489 long had the internal hash value computed at compile time, to speed up
490 lookup. This optimisation has only now been applied to hash slices as
491 well.
492 .IP \(bu 4
493 Combined \f(CW\*(C`and\*(C'\fR and \f(CW\*(C`or\*(C'\fR operators in void context, like those
494 generated for \f(CW\*(C`unless ($a && $b)\*(C'\fR and \f(CW\*(C`if ($a || b)\*(C'\fR now
495 short circuit directly to the end of the statement. [perl #120128]
496 .IP \(bu 4
497 In certain situations, when \f(CW\*(C`return\*(C'\fR is the last statement in a subroutine's
498 main scope, it will be optimized out. This means code like:
499 .Sp
500 .Vb 1
501 \& sub baz { return $cat; }
502 .Ve
503 .Sp
504 will now behave like:
505 .Sp
506 .Vb 1
507 \& sub baz { $cat; }
508 .Ve
509 .Sp
510 which is notably faster.
511 .Sp
512 [perl #120765]
513 .IP \(bu 4
514 Code like:
515 .Sp
516 .Vb 2
517 \& my $x; # or @x, %x
518 \& my $y;
519 .Ve
520 .Sp
521 is now optimized to:
522 .Sp
523 .Vb 1
524 \& my ($x, $y);
525 .Ve
526 .Sp
527 In combination with the padrange optimization introduced in
528 v5.18.0, this means longer uninitialized my
529 variable statements are also optimized, so:
530 .Sp
531 .Vb 1
532 \& my $x; my @y; my %z;
533 .Ve
534 .Sp
535 becomes:
536 .Sp
537 .Vb 1
538 \& my ($x, @y, %z);
539 .Ve
540 .Sp
541 [perl #121077]
542 .IP \(bu 4
543 The creation of certain sorts of lists, including array and hash slices, is now
544 faster.
545 .IP \(bu 4
546 The optimisation for arrays indexed with a small constant integer is now
547 applied for integers in the range \-128..127, rather than 0..255. This should
548 speed up Perl code using expressions like \f(CW$x[\-1]\fR, at the expense of
549 (presumably much rarer) code using expressions like \f(CW$x[200]\fR.
550 .IP \(bu 4
551 The first iteration over a large hash (using \f(CW\*(C`keys\*(C'\fR or \f(CW\*(C`each\*(C'\fR) is now
552 faster. This is achieved by preallocating the hash's internal iterator
553 state, rather than lazily creating it when the hash is first iterated. (For
554 small hashes, the iterator is still created only when first needed. The
555 assumption is that small hashes are more likely to be used as objects, and
556 therefore never allocated. For large hashes, that's less likely to be true,
557 and the cost of allocating the iterator is swamped by the cost of allocating
558 space for the hash itself.)
559 .IP \(bu 4
560 When doing a global regex match on a string that came from the \f(CW\*(C`readline\*(C'\fR
561 or \f(CW\*(C`<>\*(C'\fR operator, the data is no longer copied unnecessarily.
562 [perl #121259]
563 .IP \(bu 4
564 Dereferencing (as in \f(CW\*(C`$obj\->[0]\*(C'\fR or \f(CW\*(C`$obj\->{k}\*(C'\fR) is now faster
565 when \f(CW$obj\fR is an instance of a class that has overloaded methods, but
566 doesn't overload any of the dereferencing methods \f(CW\*(C`@{}\*(C'\fR, \f(CW\*(C`%{}\*(C'\fR, and so on.
567 .IP \(bu 4
568 Perl's optimiser no longer skips optimising code that follows certain
569 \&\f(CW\*(C`eval {}\*(C'\fR expressions (including those with an apparent infinite loop).
570 .IP \(bu 4
571 The implementation now does a better job of avoiding meaningless work at
572 runtime. Internal effect-free "null" operations (created as a side-effect of
573 parsing Perl programs) are normally deleted during compilation. That
574 deletion is now applied in some situations that weren't previously handled.
575 .IP \(bu 4
576 Perl now does less disk I/O when dealing with Unicode properties that cover
577 up to three ranges of consecutive code points.
578 .SH "Modules and Pragmata"
579 .IX Header "Modules and Pragmata"
580 .SS "New Modules and Pragmata"
581 .IX Subsection "New Modules and Pragmata"
582 .IP \(bu 4
583 experimental 0.007 has been added to the Perl core.
584 .IP \(bu 4
585 IO::Socket::IP 0.29 has been added to the Perl core.
586 .SS "Updated Modules and Pragmata"
587 .IX Subsection "Updated Modules and Pragmata"
588 .IP \(bu 4
589 Archive::Tar has been upgraded from version 1.90 to 1.96.
590 .IP \(bu 4
591 arybase has been upgraded from version 0.06 to 0.07.
592 .IP \(bu 4
593 Attribute::Handlers has been upgraded from version 0.94 to 0.96.
594 .IP \(bu 4
595 attributes has been upgraded from version 0.21 to 0.22.
596 .IP \(bu 4
597 autodie has been upgraded from version 2.13 to 2.23.
598 .IP \(bu 4
599 AutoLoader has been upgraded from version 5.73 to 5.74.
600 .IP \(bu 4
601 autouse has been upgraded from version 1.07 to 1.08.
602 .IP \(bu 4
603 B has been upgraded from version 1.42 to 1.48.
604 .IP \(bu 4
605 B::Concise has been upgraded from version 0.95 to 0.992.
606 .IP \(bu 4
607 B::Debug has been upgraded from version 1.18 to 1.19.
608 .IP \(bu 4
609 B::Deparse has been upgraded from version 1.20 to 1.26.
610 .IP \(bu 4
611 base has been upgraded from version 2.18 to 2.22.
612 .IP \(bu 4
613 Benchmark has been upgraded from version 1.15 to 1.18.
614 .IP \(bu 4
615 bignum has been upgraded from version 0.33 to 0.37.
616 .IP \(bu 4
617 Carp has been upgraded from version 1.29 to 1.3301.
618 .IP \(bu 4
619 CGI has been upgraded from version 3.63 to 3.65.
620 NOTE: CGI is deprecated and may be removed from a future version of Perl.
621 .IP \(bu 4
622 charnames has been upgraded from version 1.36 to 1.40.
623 .IP \(bu 4
624 Class::Struct has been upgraded from version 0.64 to 0.65.
625 .IP \(bu 4
626 Compress::Raw::Bzip2 has been upgraded from version 2.060 to 2.064.
627 .IP \(bu 4
628 Compress::Raw::Zlib has been upgraded from version 2.060 to 2.065.
629 .IP \(bu 4
630 Config::Perl::V has been upgraded from version 0.17 to 0.20.
631 .IP \(bu 4
632 constant has been upgraded from version 1.27 to 1.31.
633 .IP \(bu 4
634 CPAN has been upgraded from version 2.00 to 2.05.
635 .IP \(bu 4
636 CPAN::Meta has been upgraded from version 2.120921 to 2.140640.
637 .IP \(bu 4
638 CPAN::Meta::Requirements has been upgraded from version 2.122 to 2.125.
639 .IP \(bu 4
640 CPAN::Meta::YAML has been upgraded from version 0.008 to 0.012.
641 .IP \(bu 4
642 Data::Dumper has been upgraded from version 2.145 to 2.151.
643 .IP \(bu 4
644 DB has been upgraded from version 1.04 to 1.07.
645 .IP \(bu 4
646 DB_File has been upgraded from version 1.827 to 1.831.
647 .IP \(bu 4
648 DBM_Filter has been upgraded from version 0.05 to 0.06.
649 .IP \(bu 4
650 deprecate has been upgraded from version 0.02 to 0.03.
651 .IP \(bu 4
652 Devel::Peek has been upgraded from version 1.11 to 1.16.
653 .IP \(bu 4
654 Devel::PPPort has been upgraded from version 3.20 to 3.21.
655 .IP \(bu 4
656 diagnostics has been upgraded from version 1.31 to 1.34.
657 .IP \(bu 4
658 Digest::MD5 has been upgraded from version 2.52 to 2.53.
659 .IP \(bu 4
660 Digest::SHA has been upgraded from version 5.84 to 5.88.
661 .IP \(bu 4
662 DynaLoader has been upgraded from version 1.18 to 1.25.
663 .IP \(bu 4
664 Encode has been upgraded from version 2.49 to 2.60.
665 .IP \(bu 4
666 encoding has been upgraded from version 2.6_01 to 2.12.
667 .IP \(bu 4
668 English has been upgraded from version 1.06 to 1.09.
669 .Sp
670 \&\f(CW$OLD_PERL_VERSION\fR was added as an alias of \f(CW$]\fR.
671 .IP \(bu 4
672 Errno has been upgraded from version 1.18 to 1.20_03.
673 .IP \(bu 4
674 Exporter has been upgraded from version 5.68 to 5.70.
675 .IP \(bu 4
676 ExtUtils::CBuilder has been upgraded from version 0.280210 to 0.280216.
677 .IP \(bu 4
678 ExtUtils::Command has been upgraded from version 1.17 to 1.18.
679 .IP \(bu 4
680 ExtUtils::Embed has been upgraded from version 1.30 to 1.32.
681 .IP \(bu 4
682 ExtUtils::Install has been upgraded from version 1.59 to 1.67.
683 .IP \(bu 4
684 ExtUtils::MakeMaker has been upgraded from version 6.66 to 6.98.
685 .IP \(bu 4
686 ExtUtils::Miniperl has been upgraded from version to 1.01.
687 .IP \(bu 4
688 ExtUtils::ParseXS has been upgraded from version 3.18 to 3.24.
689 .IP \(bu 4
690 ExtUtils::Typemaps has been upgraded from version 3.19 to 3.24.
691 .IP \(bu 4
692 ExtUtils::XSSymSet has been upgraded from version 1.2 to 1.3.
693 .IP \(bu 4
694 feature has been upgraded from version 1.32 to 1.36.
695 .IP \(bu 4
696 fields has been upgraded from version 2.16 to 2.17.
697 .IP \(bu 4
698 File::Basename has been upgraded from version 2.84 to 2.85.
699 .IP \(bu 4
700 File::Copy has been upgraded from version 2.26 to 2.29.
701 .IP \(bu 4
702 File::DosGlob has been upgraded from version 1.10 to 1.12.
703 .IP \(bu 4
704 File::Fetch has been upgraded from version 0.38 to 0.48.
705 .IP \(bu 4
706 File::Find has been upgraded from version 1.23 to 1.27.
707 .IP \(bu 4
708 File::Glob has been upgraded from version 1.20 to 1.23.
709 .IP \(bu 4
710 File::Spec has been upgraded from version 3.40 to 3.47.
711 .IP \(bu 4
712 File::Temp has been upgraded from version 0.23 to 0.2304.
713 .IP \(bu 4
714 FileCache has been upgraded from version 1.08 to 1.09.
715 .IP \(bu 4
716 Filter::Simple has been upgraded from version 0.89 to 0.91.
717 .IP \(bu 4
718 Filter::Util::Call has been upgraded from version 1.45 to 1.49.
719 .IP \(bu 4
720 Getopt::Long has been upgraded from version 2.39 to 2.42.
721 .IP \(bu 4
722 Getopt::Std has been upgraded from version 1.07 to 1.10.
723 .IP \(bu 4
724 Hash::Util::FieldHash has been upgraded from version 1.10 to 1.15.
725 .IP \(bu 4
726 HTTP::Tiny has been upgraded from version 0.025 to 0.043.
727 .IP \(bu 4
728 I18N::Langinfo has been upgraded from version 0.10 to 0.11.
729 .IP \(bu 4
730 I18N::LangTags has been upgraded from version 0.39 to 0.40.
731 .IP \(bu 4
732 if has been upgraded from version 0.0602 to 0.0603.
733 .IP \(bu 4
734 inc::latest has been upgraded from version 0.4003 to 0.4205.
735 NOTE: inc::latest is deprecated and may be removed from a future version of Perl.
736 .IP \(bu 4
737 integer has been upgraded from version 1.00 to 1.01.
738 .IP \(bu 4
739 IO has been upgraded from version 1.28 to 1.31.
740 .IP \(bu 4
741 IO::Compress::Gzip and friends have been upgraded from version 2.060 to
742 2.064.
743 .IP \(bu 4
744 IPC::Cmd has been upgraded from version 0.80 to 0.92.
745 .IP \(bu 4
746 IPC::Open3 has been upgraded from version 1.13 to 1.16.
747 .IP \(bu 4
748 IPC::SysV has been upgraded from version 2.03 to 2.04.
749 .IP \(bu 4
750 JSON::PP has been upgraded from version 2.27202 to 2.27203.
751 .IP \(bu 4
752 List::Util has been upgraded from version 1.27 to 1.38.
753 .IP \(bu 4
754 locale has been upgraded from version 1.02 to 1.03.
755 .IP \(bu 4
756 Locale::Codes has been upgraded from version 3.25 to 3.30.
757 .IP \(bu 4
758 Locale::Maketext has been upgraded from version 1.23 to 1.25.
759 .IP \(bu 4
760 Math::BigInt has been upgraded from version 1.9991 to 1.9993.
761 .IP \(bu 4
762 Math::BigInt::FastCalc has been upgraded from version 0.30 to 0.31.
763 .IP \(bu 4
764 Math::BigRat has been upgraded from version 0.2604 to 0.2606.
765 .IP \(bu 4
766 MIME::Base64 has been upgraded from version 3.13 to 3.14.
767 .IP \(bu 4
768 Module::Build has been upgraded from version 0.4003 to 0.4205.
769 NOTE: Module::Build is deprecated and may be removed from a future version of Perl.
770 .IP \(bu 4
771 Module::CoreList has been upgraded from version 2.89 to 3.10.
772 .IP \(bu 4
773 Module::Load has been upgraded from version 0.24 to 0.32.
774 .IP \(bu 4
775 Module::Load::Conditional has been upgraded from version 0.54 to 0.62.
776 .IP \(bu 4
777 Module::Metadata has been upgraded from version 1.000011 to 1.000019.
778 .IP \(bu 4
779 mro has been upgraded from version 1.11 to 1.16.
780 .IP \(bu 4
781 Net::Ping has been upgraded from version 2.41 to 2.43.
782 .IP \(bu 4
783 Opcode has been upgraded from version 1.25 to 1.27.
784 .IP \(bu 4
785 Package::Constants has been upgraded from version 0.02 to 0.04.
786 NOTE: Package::Constants is deprecated and may be removed from a future version of Perl.
787 .IP \(bu 4
788 Params::Check has been upgraded from version 0.36 to 0.38.
789 .IP \(bu 4
790 parent has been upgraded from version 0.225 to 0.228.
791 .IP \(bu 4
792 Parse::CPAN::Meta has been upgraded from version 1.4404 to 1.4414.
793 .IP \(bu 4
794 Perl::OSType has been upgraded from version 1.003 to 1.007.
795 .IP \(bu 4
796 perlfaq has been upgraded from version 5.0150042 to 5.0150044.
797 .IP \(bu 4
798 PerlIO has been upgraded from version 1.07 to 1.09.
799 .IP \(bu 4
800 PerlIO::encoding has been upgraded from version 0.16 to 0.18.
801 .IP \(bu 4
802 PerlIO::scalar has been upgraded from version 0.16 to 0.18.
803 .IP \(bu 4
804 PerlIO::via has been upgraded from version 0.12 to 0.14.
805 .IP \(bu 4
806 Pod::Escapes has been upgraded from version 1.04 to 1.06.
807 .IP \(bu 4
808 Pod::Functions has been upgraded from version 1.06 to 1.08.
809 .IP \(bu 4
810 Pod::Html has been upgraded from version 1.18 to 1.21.
811 .IP \(bu 4
812 Pod::Parser has been upgraded from version 1.60 to 1.62.
813 .IP \(bu 4
814 Pod::Perldoc has been upgraded from version 3.19 to 3.23.
815 .IP \(bu 4
816 Pod::Usage has been upgraded from version 1.61 to 1.63.
817 .IP \(bu 4
818 POSIX has been upgraded from version 1.32 to 1.38_03.
819 .IP \(bu 4
820 re has been upgraded from version 0.23 to 0.26.
821 .IP \(bu 4
822 Safe has been upgraded from version 2.35 to 2.37.
823 .IP \(bu 4
824 Scalar::Util has been upgraded from version 1.27 to 1.38.
825 .IP \(bu 4
826 SDBM_File has been upgraded from version 1.09 to 1.11.
827 .IP \(bu 4
828 Socket has been upgraded from version 2.009 to 2.013.
829 .IP \(bu 4
830 Storable has been upgraded from version 2.41 to 2.49.
831 .IP \(bu 4
832 strict has been upgraded from version 1.07 to 1.08.
833 .IP \(bu 4
834 subs has been upgraded from version 1.01 to 1.02.
835 .IP \(bu 4
836 Sys::Hostname has been upgraded from version 1.17 to 1.18.
837 .IP \(bu 4
838 Sys::Syslog has been upgraded from version 0.32 to 0.33.
839 .IP \(bu 4
840 Term::Cap has been upgraded from version 1.13 to 1.15.
841 .IP \(bu 4
842 Term::ReadLine has been upgraded from version 1.12 to 1.14.
843 .IP \(bu 4
844 Test::Harness has been upgraded from version 3.26 to 3.30.
845 .IP \(bu 4
846 Test::Simple has been upgraded from version 0.98 to 1.001002.
847 .IP \(bu 4
848 Text::ParseWords has been upgraded from version 3.28 to 3.29.
849 .IP \(bu 4
850 Text::Tabs has been upgraded from version 2012.0818 to 2013.0523.
851 .IP \(bu 4
852 Text::Wrap has been upgraded from version 2012.0818 to 2013.0523.
853 .IP \(bu 4
854 Thread has been upgraded from version 3.02 to 3.04.
855 .IP \(bu 4
856 Thread::Queue has been upgraded from version 3.02 to 3.05.
857 .IP \(bu 4
858 threads has been upgraded from version 1.86 to 1.93.
859 .IP \(bu 4
860 threads::shared has been upgraded from version 1.43 to 1.46.
861 .IP \(bu 4
862 Tie::Array has been upgraded from version 1.05 to 1.06.
863 .IP \(bu 4
864 Tie::File has been upgraded from version 0.99 to 1.00.
865 .IP \(bu 4
866 Tie::Hash has been upgraded from version 1.04 to 1.05.
867 .IP \(bu 4
868 Tie::Scalar has been upgraded from version 1.02 to 1.03.
869 .IP \(bu 4
870 Tie::StdHandle has been upgraded from version 4.3 to 4.4.
871 .IP \(bu 4
872 Time::HiRes has been upgraded from version 1.9725 to 1.9726.
873 .IP \(bu 4
874 Time::Piece has been upgraded from version 1.20_01 to 1.27.
875 .IP \(bu 4
876 Unicode::Collate has been upgraded from version 0.97 to 1.04.
877 .IP \(bu 4
878 Unicode::Normalize has been upgraded from version 1.16 to 1.17.
879 .IP \(bu 4
880 Unicode::UCD has been upgraded from version 0.51 to 0.57.
881 .IP \(bu 4
882 utf8 has been upgraded from version 1.10 to 1.13.
883 .IP \(bu 4
884 version has been upgraded from version 0.9902 to 0.9908.
885 .IP \(bu 4
886 vmsish has been upgraded from version 1.03 to 1.04.
887 .IP \(bu 4
888 warnings has been upgraded from version 1.18 to 1.23.
889 .IP \(bu 4
890 Win32 has been upgraded from version 0.47 to 0.49.
891 .IP \(bu 4
892 XS::Typemap has been upgraded from version 0.10 to 0.13.
893 .IP \(bu 4
894 XSLoader has been upgraded from version 0.16 to 0.17.
895 .SH Documentation
896 .IX Header "Documentation"
897 .SS "New Documentation"
898 .IX Subsection "New Documentation"
899 \fIperlrepository\fR
900 .IX Subsection "perlrepository"
901 .PP
902 This document was removed (actually, renamed perlgit and given a major
903 overhaul) in Perl v5.14, causing Perl documentation websites to show the now
904 out of date version in Perl v5.12 as the latest version. It has now been
905 restored in stub form, directing readers to current information.
906 .SS "Changes to Existing Documentation"
907 .IX Subsection "Changes to Existing Documentation"
908 \fIperldata\fR
909 .IX Subsection "perldata"
910 .IP \(bu 4
911 New sections have been added to document the new index/value array slice and
912 key/value hash slice syntax.
913 .PP
914 \fIperldebguts\fR
915 .IX Subsection "perldebguts"
916 .IP \(bu 4
917 The \f(CW\*(C`DB::goto\*(C'\fR and \f(CW\*(C`DB::lsub\*(C'\fR debugger subroutines are now documented. [perl
918 #77680]
919 .PP
920 \fIperlexperiment\fR
921 .IX Subsection "perlexperiment"
922 .IP \(bu 4
923 \&\f(CW\*(C`\es\*(C'\fR matching \f(CW\*(C`\ecK\*(C'\fR is marked experimental.
924 .IP \(bu 4
925 ithreads were accepted in v5.8.0 (but are discouraged as of v5.20.0).
926 .IP \(bu 4
927 Long doubles are not considered experimental.
928 .IP \(bu 4
929 Code in regular expressions, regular expression backtracking verbs,
930 and lvalue subroutines are no longer listed as experimental. (This
931 also affects perlre and perlsub.)
932 .PP
933 \fIperlfunc\fR
934 .IX Subsection "perlfunc"
935 .IP \(bu 4
936 \&\f(CW\*(C`chop\*(C'\fR and \f(CW\*(C`chomp\*(C'\fR now note that they can reset the hash iterator.
937 .IP \(bu 4
938 \&\f(CW\*(C`exec\*(C'\fR's handling of arguments is now more clearly documented.
939 .IP \(bu 4
940 \&\f(CW\*(C`eval EXPR\*(C'\fR now has caveats about expanding floating point numbers in some
941 locales.
942 .IP \(bu 4
943 \&\f(CW\*(C`goto EXPR\*(C'\fR is now documented to handle an expression that evaluates to a
944 code reference as if it was \f(CW\*(C`goto &$coderef\*(C'\fR. This behavior is at least ten
945 years old.
946 .IP \(bu 4
947 Since Perl v5.10, it has been possible for subroutines in \f(CW@INC\fR to return
948 a reference to a scalar holding initial source code to prepend to the file.
949 This is now documented.
950 .IP \(bu 4
951 The documentation of \f(CW\*(C`ref\*(C'\fR has been updated to recommend the use of
952 \&\f(CW\*(C`blessed\*(C'\fR, \f(CW\*(C`isa\*(C'\fR and \f(CW\*(C`reftype\*(C'\fR when dealing with references to blessed
953 objects.
954 .PP
955 \fIperlguts\fR
956 .IX Subsection "perlguts"
957 .IP \(bu 4
958 Numerous minor changes have been made to reflect changes made to the perl
959 internals in this release.
960 .IP \(bu 4
961 New sections on Read-Only Values and
962 Copy on Write have been added.
963 .PP
964 \fIperlhack\fR
965 .IX Subsection "perlhack"
966 .IP \(bu 4
967 The Super Quick Patch Guide section has
968 been updated.
969 .PP
970 \fIperlhacktips\fR
971 .IX Subsection "perlhacktips"
972 .IP \(bu 4
973 The documentation has been updated to include some more examples of \f(CW\*(C`gdb\*(C'\fR
974 usage.
975 .PP
976 \fIperllexwarn\fR
977 .IX Subsection "perllexwarn"
978 .IP \(bu 4
979 The perllexwarn documentation used to describe the hierarchy of warning
980 categories understood by the warnings pragma. That description has now
981 been moved to the warnings documentation itself, leaving perllexwarn
982 as a stub that points to it. This change consolidates all documentation for
983 lexical warnings in a single place.
984 .PP
985 \fIperllocale\fR
986 .IX Subsection "perllocale"
987 .IP \(bu 4
988 The documentation now mentions \fR\f(BIfc()\fR\fI\fR and \f(CW\*(C`\eF\*(C'\fR, and includes many
989 clarifications and corrections in general.
990 .PP
991 \fIperlop\fR
992 .IX Subsection "perlop"
993 .IP \(bu 4
994 The language design of Perl has always called for monomorphic operators.
995 This is now mentioned explicitly.
996 .PP
997 \fIperlopentut\fR
998 .IX Subsection "perlopentut"
999 .IP \(bu 4
1000 The \f(CW\*(C`open\*(C'\fR tutorial has been completely rewritten by Tom Christiansen, and now
1001 focuses on covering only the basics, rather than providing a comprehensive
1002 reference to all things openable. This rewrite came as the result of a
1003 vigorous discussion on perl5\-porters kicked off by a set of improvements
1004 written by Alexander Hartmaier to the existing perlopentut. A "more than
1005 you ever wanted to know about \f(CW\*(C`open\*(C'\fR" document may follow in subsequent
1006 versions of perl.
1007 .PP
1008 \fIperlre\fR
1009 .IX Subsection "perlre"
1010 .IP \(bu 4
1011 The fact that the regexp engine makes no effort to call (?{}) and (??{})
1012 constructs any specified number of times (although it will basically DWIM
1013 in case of a successful match) has been documented.
1014 .IP \(bu 4
1015 The \f(CW\*(C`/r\*(C'\fR modifier (for non-destructive substitution) is now documented. [perl
1016 #119151]
1017 .IP \(bu 4
1018 The documentation for \f(CW\*(C`/x\*(C'\fR and \f(CW\*(C`(?# comment)\*(C'\fR has been expanded and clarified.
1019 .PP
1020 \fIperlreguts\fR
1021 .IX Subsection "perlreguts"
1022 .IP \(bu 4
1023 The documentation has been updated in the light of recent changes to
1024 \&\fIregcomp.c\fR.
1025 .PP
1026 \fIperlsub\fR
1027 .IX Subsection "perlsub"
1028 .IP \(bu 4
1029 The need to predeclare recursive functions with prototypes in order for the
1030 prototype to be honoured in the recursive call is now documented. [perl #2726]
1031 .IP \(bu 4
1032 A list of subroutine names used by the perl implementation is now included.
1033 [perl #77680]
1034 .PP
1035 \fIperltrap\fR
1036 .IX Subsection "perltrap"
1037 .IP \(bu 4
1038 There is now a JavaScript section.
1039 .PP
1040 \fIperlunicode\fR
1041 .IX Subsection "perlunicode"
1042 .IP \(bu 4
1043 The documentation has been updated to reflect \f(CW\*(C`Bidi_Class\*(C'\fR changes in
1044 Unicode 6.3.
1045 .PP
1046 \fIperlvar\fR
1047 .IX Subsection "perlvar"
1048 .IP \(bu 4
1049 A new section explaining the performance issues of $`, $& and $', including
1050 workarounds and changes in different versions of Perl, has been added.
1051 .IP \(bu 4
1052 Three English variable names which have long been documented but do not
1053 actually exist have been removed from the documentation. These were
1054 \&\f(CW$OLD_PERL_VERSION\fR, \f(CW$OFMT\fR, and \f(CW$ARRAY_BASE\fR.
1055 .Sp
1056 (Actually, \f(CW\*(C`OLD_PERL_VERSION\*(C'\fR \fIdoes\fR exist, starting with this revision, but
1057 remained undocumented until perl 5.22.0.)
1058 .PP
1059 \fIperlxs\fR
1060 .IX Subsection "perlxs"
1061 .IP \(bu 4
1062 Several problems in the \f(CW\*(C`MY_CXT\*(C'\fR example have been fixed.
1063 .SH Diagnostics
1064 .IX Header "Diagnostics"
1065 The following additions or changes have been made to diagnostic output,
1066 including warnings and fatal error messages. For the complete list of
1067 diagnostic messages, see perldiag.
1068 .SS "New Diagnostics"
1069 .IX Subsection "New Diagnostics"
1070 \fINew Errors\fR
1071 .IX Subsection "New Errors"
1072 .IP \(bu 4
1073 delete argument is index/value array slice, use array slice
1074 .Sp
1075 (F) You used index/value array slice syntax (\f(CW%array[...]\fR) as the argument to
1076 \&\f(CW\*(C`delete\*(C'\fR. You probably meant \f(CW@array[...]\fR with an @ symbol instead.
1077 .IP \(bu 4
1078 delete argument is key/value hash slice, use hash slice
1079 .Sp
1080 (F) You used key/value hash slice syntax (\f(CW%hash{...}\fR) as the argument to
1081 \&\f(CW\*(C`delete\*(C'\fR. You probably meant \f(CW@hash{...}\fR with an @ symbol instead.
1082 .IP \(bu 4
1083 Magical list constants are not supported
1084 .Sp
1085 (F) You assigned a magical array to a stash element, and then tried to use the
1086 subroutine from the same slot. You are asking Perl to do something it cannot
1087 do, details subject to change between Perl versions.
1088 .IP \(bu 4
1089 Added Setting $/ to a \f(CW%s\fR reference is forbidden
1090 .PP
1091 \fINew Warnings\fR
1092 .IX Subsection "New Warnings"
1093 .IP \(bu 4
1094 \&\f(CW%s\fR on reference is experimental:
1095 .Sp
1096 The "auto-deref" feature is experimental.
1097 .Sp
1098 Starting in v5.14.0, it was possible to use push, pop, keys, and other
1099 built-in functions not only on aggregate types, but on references to
1100 them. The feature was not deployed to its original intended
1101 specification, and now may become redundant to postfix dereferencing.
1102 It has always been categorized as an experimental feature, and in
1103 v5.20.0 is carries a warning as such.
1104 .Sp
1105 Warnings will now be issued at compile time when these operations are
1106 detected.
1107 .Sp
1108 .Vb 1
1109 \& no if $] >= 5.01908, warnings => "experimental::autoderef";
1110 .Ve
1111 .Sp
1112 Consider, though, replacing the use of these features, as they may
1113 change behavior again before becoming stable.
1114 .IP \(bu 4
1115 A sequence of multiple spaces in a charnames alias definition is deprecated
1116 .Sp
1117 Trailing white-space in a charnames alias definition is deprecated
1118 .Sp
1119 These two deprecation warnings involving \f(CW\*(C`\eN{...}\*(C'\fR were incorrectly
1120 implemented. They did not warn by default (now they do) and could not be
1121 made fatal via \f(CW\*(C`use warnings FATAL => \*(Aqdeprecated\*(Aq\*(C'\fR (now they can).
1122 .IP \(bu 4
1123 Attribute prototype(%s) discards earlier prototype attribute in same sub
1124 .Sp
1125 (W misc) A sub was declared as \f(CW\*(C`sub foo : prototype(A) : prototype(B) {}\*(C'\fR, for
1126 example. Since each sub can only have one prototype, the earlier
1127 declaration(s) are discarded while the last one is applied.
1128 .IP \(bu 4
1129 Invalid \e0 character in \f(CW%s\fR for \f(CW%s:\fR \f(CW%s\fR\e0%s
1130 .Sp
1131 (W syscalls) Embedded \e0 characters in pathnames or other system call arguments
1132 produce a warning as of 5.20. The parts after the \e0 were formerly ignored by
1133 system calls.
1134 .IP \(bu 4
1135 Matched non-Unicode code point 0x%X against Unicode property; may not be portable.
1136 .Sp
1137 This replaces the message "Code point 0x%X is not Unicode, all \ep{} matches
1138 fail; all \eP{} matches succeed".
1139 .IP \(bu 4
1140 Missing ']' in prototype for \f(CW%s\fR : \f(CW%s\fR
1141 .Sp
1142 (W illegalproto) A grouping was started with \f(CW\*(C`[\*(C'\fR but never closed with \f(CW\*(C`]\*(C'\fR.
1143 .IP \(bu 4
1144 Possible precedence issue with control flow operator
1145 .Sp
1146 (W syntax) There is a possible problem with the mixing of a control flow
1147 operator (e.g. \f(CW\*(C`return\*(C'\fR) and a low-precedence operator like \f(CW\*(C`or\*(C'\fR. Consider:
1148 .Sp
1149 .Vb 1
1150 \& sub { return $a or $b; }
1151 .Ve
1152 .Sp
1153 This is parsed as:
1154 .Sp
1155 .Vb 1
1156 \& sub { (return $a) or $b; }
1157 .Ve
1158 .Sp
1159 Which is effectively just:
1160 .Sp
1161 .Vb 1
1162 \& sub { return $a; }
1163 .Ve
1164 .Sp
1165 Either use parentheses or the high-precedence variant of the operator.
1166 .Sp
1167 Note this may be also triggered for constructs like:
1168 .Sp
1169 .Vb 1
1170 \& sub { 1 if die; }
1171 .Ve
1172 .IP \(bu 4
1173 Postfix dereference is experimental
1174 .Sp
1175 (S experimental::postderef) This warning is emitted if you use the experimental
1176 postfix dereference syntax. Simply suppress the warning if you want to use the
1177 feature, but know that in doing so you are taking the risk of using an
1178 experimental feature which may change or be removed in a future Perl version:
1179 .Sp
1180 .Vb 6
1181 \& no warnings "experimental::postderef";
1182 \& use feature "postderef", "postderef_qq";
1183 \& $ref\->$*;
1184 \& $aref\->@*;
1185 \& $aref\->@[@indices];
1186 \& ... etc ...
1187 .Ve
1188 .IP \(bu 4
1189 Prototype '%s' overridden by attribute 'prototype(%s)' in \f(CW%s\fR
1190 .Sp
1191 (W prototype) A prototype was declared in both the parentheses after the sub
1192 name and via the prototype attribute. The prototype in parentheses is useless,
1193 since it will be replaced by the prototype from the attribute before it's ever
1194 used.
1195 .IP \(bu 4
1196 Scalar value @%s[%s] better written as $%s[%s]
1197 .Sp
1198 (W syntax) In scalar context, you've used an array index/value slice (indicated
1199 by %) to select a single element of an array. Generally it's better to ask for
1200 a scalar value (indicated by $). The difference is that \f(CW$foo[&bar]\fR always
1201 behaves like a scalar, both in the value it returns and when evaluating its
1202 argument, while \f(CW%foo[&bar]\fR provides a list context to its subscript, which
1203 can do weird things if you're expecting only one subscript. When called in
1204 list context, it also returns the index (what \f(CW&bar\fR returns) in addition to
1205 the value.
1206 .IP \(bu 4
1207 Scalar value @%s{%s} better written as $%s{%s}
1208 .Sp
1209 (W syntax) In scalar context, you've used a hash key/value slice (indicated by
1210 %) to select a single element of a hash. Generally it's better to ask for a
1211 scalar value (indicated by $). The difference is that \f(CW$foo{&bar}\fR always
1212 behaves like a scalar, both in the value it returns and when evaluating its
1213 argument, while \f(CW@foo{&bar}\fR and provides a list context to its subscript,
1214 which can do weird things if you're expecting only one subscript. When called
1215 in list context, it also returns the key in addition to the value.
1216 .IP \(bu 4
1217 Setting $/ to a reference to \f(CW%s\fR as a form of slurp is deprecated, treating as undef
1218 .IP \(bu 4
1219 Unexpected exit \f(CW%u\fR
1220 .Sp
1221 (S) \fBexit()\fR was called or the script otherwise finished gracefully when
1222 \&\f(CW\*(C`PERL_EXIT_WARN\*(C'\fR was set in \f(CW\*(C`PL_exit_flags\*(C'\fR.
1223 .IP \(bu 4
1224 Unexpected exit failure \f(CW%d\fR
1225 .Sp
1226 (S) An uncaught \fBdie()\fR was called when \f(CW\*(C`PERL_EXIT_WARN\*(C'\fR was set in
1227 \&\f(CW\*(C`PL_exit_flags\*(C'\fR.
1228 .IP \(bu 4
1229 Use of literal control characters in variable names is deprecated
1230 .Sp
1231 (D deprecated) Using literal control characters in the source to refer to the
1232 ^FOO variables, like $^X and ${^GLOBAL_PHASE} is now deprecated. This only
1233 affects code like $\ecT, where \ecT is a control (like a \f(CW\*(C`SOH\*(C'\fR) in the
1234 source code: ${"\ecT"} and $^T remain valid.
1235 .IP \(bu 4
1236 Useless use of greediness modifier
1237 .Sp
1238 This fixes [Perl #42957].
1239 .SS "Changes to Existing Diagnostics"
1240 .IX Subsection "Changes to Existing Diagnostics"
1241 .IP \(bu 4
1242 Warnings and errors from the regexp engine are now UTF\-8 clean.
1243 .IP \(bu 4
1244 The "Unknown switch condition" error message has some slight changes. This
1245 error triggers when there is an unknown condition in a \f(CW\*(C`(?(foo))\*(C'\fR conditional.
1246 The error message used to read:
1247 .Sp
1248 .Vb 1
1249 \& Unknown switch condition (?(%s in regex;
1250 .Ve
1251 .Sp
1252 But what \f(CW%s\fR could be was mostly up to luck. For \f(CW\*(C`(?(foobar))\*(C'\fR, you might have
1253 seen "fo" or "f". For Unicode characters, you would generally get a corrupted
1254 string. The message has been changed to read:
1255 .Sp
1256 .Vb 1
1257 \& Unknown switch condition (?(...)) in regex;
1258 .Ve
1259 .Sp
1260 Additionally, the \f(CW\*(Aq<\-\- HERE\*(Aq\fR marker in the error will now point to the
1261 correct spot in the regex.
1262 .IP \(bu 4
1263 The "%s "\ex%X" does not map to Unicode" warning is now correctly listed as a
1264 severe warning rather than as a fatal error.
1265 .IP \(bu 4
1266 Under rare circumstances, one could get a "Can't coerce readonly REF to
1267 string" instead of the customary "Modification of a read-only value". This
1268 alternate error message has been removed.
1269 .IP \(bu 4
1270 "Ambiguous use of * resolved as operator *": This and similar warnings
1271 about "%" and "&" used to occur in some circumstances where there was no
1272 operator of the type cited, so the warning was completely wrong. This has
1273 been fixed [perl #117535, #76910].
1274 .IP \(bu 4
1275 Warnings about malformed subroutine prototypes are now more consistent in
1276 how the prototypes are rendered. Some of these warnings would truncate
1277 prototypes containing nulls. In other cases one warning would suppress
1278 another. The warning about illegal characters in prototypes no longer says
1279 "after '_'" if the bad character came before the underscore.
1280 .IP \(bu 4
1281 Perl folding rules are not up-to-date for 0x%X; please use the perlbug
1282 utility to report; in regex; marked by <\-\- HERE in
1283 m/%s/
1284 .Sp
1285 This message is now only in the regexp category, and not in the deprecated
1286 category. It is still a default (i.e., severe) warning [perl #89648].
1287 .IP \(bu 4
1288 %%s[%s] in scalar context better written as $%s[%s]
1289 .Sp
1290 This warning now occurs for any \f(CW%array[$index]\fR or \f(CW%hash{key}\fR known to
1291 be in scalar context at compile time. Previously it was worded "Scalar
1292 value %%s[%s] better written as $%s[%s]".
1293 .IP \(bu 4
1294 Switch condition not recognized in regex; marked by <\-\- HERE in m/%s/:
1295 .Sp
1296 The description for this diagnostic has been extended to cover all cases where the warning may occur.
1297 Issues with the positioning of the arrow indicator have also been resolved.
1298 .IP \(bu 4
1299 The error messages for \f(CWmy($a?$b$c)\fR and \f(CWmy(do{})\fR now mention "conditional
1300 expression" and "do block", respectively, instead of reading 'Can't declare
1301 null operation in "my"'.
1302 .IP \(bu 4
1303 When \f(CW\*(C`use re "debug"\*(C'\fR executes a regex containing a backreference, the
1304 debugging output now shows what string is being matched.
1305 .IP \(bu 4
1306 The now fatal error message \f(CW\*(C`Character following "\ec" must be ASCII\*(C'\fR has been
1307 reworded as \f(CW\*(C`Character following "\ec" must be printable ASCII\*(C'\fR to emphasize
1308 that in \f(CW\*(C`\ec\fR\f(CIX\fR\f(CW\*(C'\fR, \fIX\fR must be a \fIprintable (non-control)\fR ASCII character.
1309 .SH "Utility Changes"
1310 .IX Header "Utility Changes"
1311 \fIa2p\fR
1312 .IX Subsection "a2p"
1313 .IP \(bu 4
1314 A possible crash from an off-by-one error when trying to access before the
1315 beginning of a buffer has been fixed. [perl #120244]
1316 .PP
1317 \fIbisect.pl\fR
1318 .IX Subsection "bisect.pl"
1319 .PP
1320 The git bisection tool \fIPorting/bisect.pl\fR has had many enhancements.
1321 .PP
1322 It is provided as part of the source distribution but not installed because
1323 it is not self-contained as it relies on being run from within a git
1324 checkout. Note also that it makes no attempt to fix tests, correct runtime
1325 bugs or make something useful to install \- its purpose is to make minimal
1326 changes to get any historical revision of interest to build and run as close
1327 as possible to "as-was", and thereby make \f(CW\*(C`git bisect\*(C'\fR easy to use.
1328 .IP \(bu 4
1329 Can optionally run the test case with a timeout.
1330 .IP \(bu 4
1331 Can now run in-place in a clean git checkout.
1332 .IP \(bu 4
1333 Can run the test case under \f(CW\*(C`valgrind\*(C'\fR.
1334 .IP \(bu 4
1335 Can apply user supplied patches and fixes to the source checkout before
1336 building.
1337 .IP \(bu 4
1338 Now has fixups to enable building several more historical ranges of bleadperl,
1339 which can be useful for pinpointing the origins of bugs or behaviour changes.
1340 .PP
1341 \fIfind2perl\fR
1342 .IX Subsection "find2perl"
1343 .IP \(bu 4
1344 find2perl now handles \f(CW\*(C`?\*(C'\fR wildcards correctly. [perl #113054]
1345 .PP
1346 \fIperlbug\fR
1347 .IX Subsection "perlbug"
1348 .IP \(bu 4
1349 \&\fIperlbug\fR now has a \f(CW\*(C`\-p\*(C'\fR option for attaching patches with a bug report.
1350 .IP \(bu 4
1351 perlbug has been modified to supply the report template with CRLF line
1352 endings on Windows.
1353 [GH #13612] <https://github.com/Perl/perl5/issues/13612>
1354 .IP \(bu 4
1355 perlbug now makes as few assumptions as possible about the encoding of the
1356 report. This will likely change in the future to assume UTF\-8 by default but
1357 allow a user override.
1358 .SH "Configuration and Compilation"
1359 .IX Header "Configuration and Compilation"
1360 .IP \(bu 4
1361 The \fIMakefile.PL\fR for SDBM_File now generates a better \fIMakefile\fR, which
1362 avoids a race condition during parallel makes, which could cause the build to
1363 fail. This is the last known parallel make problem (on *nix platforms), and
1364 therefore we believe that a parallel make should now always be error free.
1365 .IP \(bu 4
1366 \&\fIinstallperl\fR and \fIinstallman\fR's option handling has been refactored to use
1367 Getopt::Long. Both are used by the \fIMakefile\fR \f(CW\*(C`install\*(C'\fR targets, and
1368 are not installed, so these changes are only likely to affect custom
1369 installation scripts.
1370 .RS 4
1371 .IP \(bu 4
1372 Single letter options now also have long names.
1373 .IP \(bu 4
1374 Invalid options are now rejected.
1375 .IP \(bu 4
1376 Command line arguments that are not options are now rejected.
1377 .IP \(bu 4
1378 Each now has a \f(CW\*(C`\-\-help\*(C'\fR option to display the usage message.
1379 .RE
1380 .RS 4
1381 .Sp
1382 The behaviour for all valid documented invocations is unchanged.
1383 .RE
1384 .IP \(bu 4
1385 Where possible, the build now avoids recursive invocations of \fImake\fR when
1386 building pure-Perl extensions, without removing any parallelism from the
1387 build. Currently around 80 extensions can be processed directly by the
1388 \&\fImake_ext.pl\fR tool, meaning that 80 invocations of \fImake\fR and 160
1389 invocations of \fIminiperl\fR are no longer made.
1390 .IP \(bu 4
1391 The build system now works correctly when compiling under GCC or Clang with
1392 link-time optimization enabled (the \f(CW\*(C`\-flto\*(C'\fR option). [perl #113022]
1393 .IP \(bu 4
1394 Distinct library basenames with \f(CW\*(C`d_libname_unique\*(C'\fR.
1395 .Sp
1396 When compiling perl with this option, the library files for XS modules are
1397 named something "unique" \-\- for example, Hash/Util/Util.so becomes
1398 Hash/Util/PL_Hash_\|_Util.so. This behavior is similar to what currently
1399 happens on VMS, and serves as groundwork for the Android port.
1400 .IP \(bu 4
1401 \&\f(CW\*(C`sysroot\*(C'\fR option to indicate the logical root directory under gcc and clang.
1402 .Sp
1403 When building with this option set, both Configure and the compilers search
1404 for all headers and libraries under this new sysroot, instead of /.
1405 .Sp
1406 This is a huge time saver if cross-compiling, but can also help
1407 on native builds if your toolchain's files have non-standard locations.
1408 .IP \(bu 4
1409 The cross-compilation model has been renovated.
1410 There's several new options, and some backwards-incompatible changes:
1411 .Sp
1412 We now build binaries for miniperl and generate_uudmap to be used on the host,
1413 rather than running every miniperl call on the target; this means that, short
1414 of 'make test', we no longer need access to the target system once Configure is
1415 done. You can provide already-built binaries through the \f(CW\*(C`hostperl\*(C'\fR and
1416 \&\f(CW\*(C`hostgenerate\*(C'\fR options to Configure.
1417 .Sp
1418 Additionally, if targeting an EBCDIC platform from an ASCII host,
1419 or viceversa, you'll need to run Configure with \f(CW\*(C`\-Uhostgenerate\*(C'\fR, to
1420 indicate that generate_uudmap should be run on the target.
1421 .Sp
1422 Finally, there's also a way of having Configure end early, right after
1423 building the host binaries, by cross-compiling without specifying a
1424 \&\f(CW\*(C`targethost\*(C'\fR.
1425 .Sp
1426 The incompatible changes include no longer using xconfig.h, xlib, or
1427 Cross.pm, so canned config files and Makefiles will have to be updated.
1428 .IP \(bu 4
1429 Related to the above, there is now a way of specifying the location of sh
1430 (or equivalent) on the target system: \f(CW\*(C`targetsh\*(C'\fR.
1431 .Sp
1432 For example, Android has its sh in /system/bin/sh, so if cross-compiling
1433 from a more normal Unixy system with sh in /bin/sh, "targetsh" would end
1434 up as /system/bin/sh, and "sh" as /bin/sh.
1435 .IP \(bu 4
1436 By default, \fBgcc\fR 4.9 does some optimizations that break perl. The \fB\-fwrapv\fR
1437 option disables those optimizations (and probably others), so for \fBgcc\fR 4.3
1438 and later (since the there might be similar problems lurking on older versions
1439 too, but \fB\-fwrapv\fR was broken before 4.3, and the optimizations probably won't
1440 go away), \fIConfigure\fR now adds \fB\-fwrapv\fR unless the user requests
1441 \&\fB\-fno\-wrapv\fR, which disables \fB\-fwrapv\fR, or \fB\-fsanitize=undefined\fR, which
1442 turns the overflows \fB\-fwrapv\fR ignores into runtime errors.
1443 [GH #13690] <https://github.com/Perl/perl5/issues/13690>
1444 .SH Testing
1445 .IX Header "Testing"
1446 .IP \(bu 4
1447 The \f(CW\*(C`test.valgrind\*(C'\fR make target now allows tests to be run in parallel.
1448 This target allows Perl's test suite to be run under Valgrind, which detects
1449 certain sorts of C programming errors, though at significant cost in running
1450 time. On suitable hardware, allowing parallel execution claws back a lot of
1451 that additional cost. [perl #121431]
1452 .IP \(bu 4
1453 Various tests in \fIt/porting/\fR are no longer skipped when the perl
1454 \&\fI.git\fR directory is outside the perl tree and pointed to by
1455 \&\f(CW$GIT_DIR\fR. [perl #120505]
1456 .IP \(bu 4
1457 The test suite no longer fails when the user's interactive shell maintains a
1458 \&\f(CW$PWD\fR environment variable, but the \fI/bin/sh\fR used for running tests
1459 doesn't.
1460 .SH "Platform Support"
1461 .IX Header "Platform Support"
1462 .SS "New Platforms"
1463 .IX Subsection "New Platforms"
1464 .IP Android 4
1465 .IX Item "Android"
1466 Perl can now be built for Android, either natively or through
1467 cross-compilation, for all three currently available architectures (ARM,
1468 MIPS, and x86), on a wide range of versions.
1469 .IP Bitrig 4
1470 .IX Item "Bitrig"
1471 Compile support has been added for Bitrig, a fork of OpenBSD.
1472 .IP FreeMiNT 4
1473 .IX Item "FreeMiNT"
1474 Support has been added for FreeMiNT, a free open-source OS for the Atari ST
1475 system and its successors, based on the original MiNT that was officially
1476 adopted by Atari.
1477 .IP Synology 4
1478 .IX Item "Synology"
1479 Synology ships its NAS boxes with a lean Linux distribution (DSM) on relative
1480 cheap CPU's (like the Marvell Kirkwood mv6282 \- ARMv5tel or Freescale QorIQ
1481 P1022 ppc \- e500v2) not meant for workstations or development. These boxes
1482 should build now. The basic problems are the non-standard location for tools.
1483 .SS "Discontinued Platforms"
1484 .IX Subsection "Discontinued Platforms"
1485 .ie n .IP """sfio""" 4
1486 .el .IP \f(CWsfio\fR 4
1487 .IX Item "sfio"
1488 Code related to supporting the \f(CW\*(C`sfio\*(C'\fR I/O system has been removed.
1489 .Sp
1490 Perl 5.004 added support to use the native API of \f(CW\*(C`sfio\*(C'\fR, AT&T's Safe/Fast
1491 I/O library. This code still built with v5.8.0, albeit with many regression
1492 tests failing, but was inadvertently broken before the v5.8.1 release,
1493 meaning that it has not worked on any version of Perl released since then.
1494 In over a decade we have received no bug reports about this, hence it is clear
1495 that no-one is using this functionality on any version of Perl that is still
1496 supported to any degree.
1497 .IP "AT&T 3b1" 4
1498 .IX Item "AT&T 3b1"
1499 Configure support for the 3b1, also known as the AT&T Unix PC (and the similar
1500 AT&T 7300), has been removed.
1501 .IP DG/UX 4
1502 .IX Item "DG/UX"
1503 DG/UX was a Unix sold by Data General. The last release was in April 2001.
1504 It only runs on Data General's own hardware.
1505 .IP EBCDIC 4
1506 .IX Item "EBCDIC"
1507 In the absence of a regular source of smoke reports, code intended to support
1508 native EBCDIC platforms will be removed from perl before 5.22.0.
1509 .SS "Platform-Specific Notes"
1510 .IX Subsection "Platform-Specific Notes"
1511 .IP Cygwin 4
1512 .IX Item "Cygwin"
1513 .RS 4
1514 .PD 0
1515 .IP \(bu 4
1516 .PD
1517 \&\fBrecv()\fR on a connected handle would populate the returned sender
1518 address with whatever happened to be in the working buffer. \fBrecv()\fR
1519 now uses a workaround similar to the Win32 \fBrecv()\fR wrapper and returns
1520 an empty string when \fBrecvfrom\fR\|(2) doesn't modify the supplied address
1521 length. [perl #118843]
1522 .IP \(bu 4
1523 Fixed a build error in cygwin.c on Cygwin 1.7.28.
1524 .Sp
1525 Tests now handle the errors that occur when \f(CW\*(C`cygserver\*(C'\fR isn't
1526 running.
1527 .RE
1528 .RS 4
1529 .RE
1530 .IP GNU/Hurd 4
1531 .IX Item "GNU/Hurd"
1532 The BSD compatibility library \f(CW\*(C`libbsd\*(C'\fR is no longer required for builds.
1533 .IP Linux 4
1534 .IX Item "Linux"
1535 The hints file now looks for \f(CW\*(C`libgdbm_compat\*(C'\fR only if \f(CW\*(C`libgdbm\*(C'\fR itself is
1536 also wanted. The former is never useful without the latter, and in some
1537 circumstances, including it could actually prevent building.
1538 .IP "Mac OS" 4
1539 .IX Item "Mac OS"
1540 The build system now honors an \f(CW\*(C`ld\*(C'\fR setting supplied by the user running
1541 \&\fIConfigure\fR.
1542 .IP MidnightBSD 4
1543 .IX Item "MidnightBSD"
1544 \&\f(CW\*(C`objformat\*(C'\fR was removed from version 0.4\-RELEASE of MidnightBSD and had been
1545 deprecated on earlier versions. This caused the build environment to be
1546 erroneously configured for \f(CW\*(C`a.out\*(C'\fR rather than \f(CW\*(C`elf\*(C'\fR. This has been now
1547 been corrected.
1548 .IP "Mixed-endian platforms" 4
1549 .IX Item "Mixed-endian platforms"
1550 The code supporting \f(CW\*(C`pack\*(C'\fR and \f(CW\*(C`unpack\*(C'\fR operations on mixed endian
1551 platforms has been removed. We believe that Perl has long been unable to
1552 build on mixed endian architectures (such as PDP\-11s), so we don't think
1553 that this change will affect any platforms which were able to build v5.18.0.
1554 .IP VMS 4
1555 .IX Item "VMS"
1556 .RS 4
1557 .PD 0
1558 .IP \(bu 4
1559 .PD
1560 The \f(CW\*(C`PERL_ENV_TABLES\*(C'\fR feature to control the population of \f(CW%ENV\fR at perl
1561 start-up was broken in Perl 5.16.0 but has now been fixed.
1562 .IP \(bu 4
1563 Skip access checks on remotes in \fBopendir()\fR. [perl #121002]
1564 .IP \(bu 4
1565 A check for glob metacharacters in a path returned by the
1566 \&\f(CWglob()\fR operator has been replaced with a check for VMS
1567 wildcard characters. This saves a significant number of unnecessary
1568 \&\f(CWlstat()\fR calls such that some simple glob operations become
1569 60\-80% faster.
1570 .RE
1571 .RS 4
1572 .RE
1573 .IP Win32 4
1574 .IX Item "Win32"
1575 .RS 4
1576 .PD 0
1577 .IP \(bu 4
1578 .PD
1579 \&\f(CW\*(C`rename\*(C'\fR and \f(CW\*(C`link\*(C'\fR on Win32 now set $! to ENOSPC and EDQUOT when
1580 appropriate. [perl #119857]
1581 .IP \(bu 4
1582 The BUILD_STATIC and ALL_STATIC makefile options for linking some or (nearly)
1583 all extensions statically (into perl520.dll, and into a separate
1584 perl\-static.exe too) were broken for MinGW builds. This has now been fixed.
1585 .Sp
1586 The ALL_STATIC option has also been improved to include the Encode and Win32
1587 extensions (for both VC++ and MinGW builds).
1588 .IP \(bu 4
1589 Support for building with Visual C++ 2013 has been added. There are currently
1590 two possible test failures (see "Testing Perl on Windows" in perlwin32) which
1591 will hopefully be resolved soon.
1592 .IP \(bu 4
1593 Experimental support for building with Intel C++ Compiler has been added. The
1594 nmake makefile (win32/Makefile) and the dmake makefile (win32/makefile.mk) can
1595 be used. A "nmake test" will not pass at this time due to \fIcpan/CGI/t/url.t\fR.
1596 .IP \(bu 4
1597 Killing a process tree with "kill" in perlfunc and a negative signal, was broken
1598 starting in 5.18.0. In this bug, \f(CW\*(C`kill\*(C'\fR always returned 0 for a negative
1599 signal even for valid PIDs, and no processes were terminated. This has been
1600 fixed [perl #121230].
1601 .IP \(bu 4
1602 The time taken to build perl on Windows has been reduced quite significantly
1603 (time savings in the region of 30\-40% are typically seen) by reducing the
1604 number of, usually failing, I/O calls for each \f(CWrequire()\fR
1605 (for \fBminiperl.exe\fR only).
1606 [GH #13566] <https://github.com/Perl/perl5/issues/13566>
1607 .IP \(bu 4
1608 About 15 minutes of idle sleeping was removed from running \f(CW\*(C`make test\*(C'\fR due to
1609 a bug in which the timeout monitor used for tests could not be cancelled once
1610 the test completes, and the full timeout period elapsed before running the next
1611 test file.
1612 [GH #13647] <https://github.com/Perl/perl5/issues/13647>
1613 .IP \(bu 4
1614 On a perl built without pseudo-fork (pseudo-fork builds were not affected by
1615 this bug), killing a process tree with \f(CWkill()\fR and a negative
1616 signal resulted in \f(CWkill()\fR inverting the returned value. For example, if
1617 \&\f(CWkill()\fR killed 1 process tree PID then it returned 0 instead of 1, and if
1618 \&\f(CWkill()\fR was passed 2 invalid PIDs then it returned 2 instead of 0. This has
1619 probably been the case since the process tree kill feature was implemented on
1620 Win32. It has now been corrected to follow the documented behaviour.
1621 [GH #13595] <https://github.com/Perl/perl5/issues/13595>
1622 .IP \(bu 4
1623 When building a 64\-bit perl, an uninitialized memory read in \fBminiperl.exe\fR,
1624 used during the build process, could lead to a 4GB \fBwperl.exe\fR being created.
1625 This has now been fixed. (Note that \fBperl.exe\fR itself was unaffected, but
1626 obviously \fBwperl.exe\fR would have been completely broken.)
1627 [GH #13677] <https://github.com/Perl/perl5/issues/13677>
1628 .IP \(bu 4
1629 Perl can now be built with \fBgcc\fR version 4.8.1 from <http://www.mingw.org>.
1630 This was previously broken due to an incorrect definition of \fBDllMain()\fR in one
1631 of perl's source files. Earlier \fBgcc\fR versions were also affected when using
1632 version 4 of the w32api package. Versions of \fBgcc\fR available from
1633 <http://mingw\-w64.sourceforge.net/> were not affected.
1634 [GH #13733] <https://github.com/Perl/perl5/issues/13733>
1635 .IP \(bu 4
1636 The test harness now has no failures when perl is built on a FAT drive with the
1637 Windows OS on an NTFS drive.
1638 [GH #6348] <https://github.com/Perl/perl5/issues/6348>
1639 .IP \(bu 4
1640 When cloning the context stack in \fBfork()\fR emulation, \fBPerl_cx_dup()\fR
1641 would crash accessing parameter information for context stack entries
1642 that included no parameters, as with \f(CW\*(C`&foo;\*(C'\fR.
1643 [GH #13763] <https://github.com/Perl/perl5/issues/13763>
1644 .IP \(bu 4
1645 Introduced by
1646 [GH #12161] <https://github.com/Perl/perl5/issues/12161>, a memory
1647 leak on every call to \f(CW\*(C`system\*(C'\fR and backticks (\f(CW \`\` \fR), on most Win32 Perls
1648 starting from 5.18.0 has been fixed. The memory leak only occurred if you
1649 enabled pseudo-fork in your build of Win32 Perl, and were running that build on
1650 Server 2003 R2 or newer OS. The leak does not appear on WinXP SP3.
1651 [GH #13741] <https://github.com/Perl/perl5/issues/13741>
1652 .RE
1653 .RS 4
1654 .RE
1655 .IP WinCE 4
1656 .IX Item "WinCE"
1657 .RS 4
1658 .PD 0
1659 .IP \(bu 4
1660 .PD
1661 The building of XS modules has largely been restored. Several still cannot
1662 (yet) be built but it is now possible to build Perl on WinCE with only a couple
1663 of further patches (to Socket and ExtUtils::MakeMaker), hopefully to be
1664 incorporated soon.
1665 .IP \(bu 4
1666 Perl can now be built in one shot with no user intervention on WinCE by running
1667 \&\f(CW\*(C`nmake \-f Makefile.ce all\*(C'\fR.
1668 .Sp
1669 Support for building with EVC (Embedded Visual C++) 4 has been restored. Perl
1670 can also be built using Smart Devices for Visual C++ 2005 or 2008.
1671 .RE
1672 .RS 4
1673 .RE
1674 .SH "Internal Changes"
1675 .IX Header "Internal Changes"
1676 .IP \(bu 4
1677 The internal representation has changed for the match variables \f(CW$1\fR, \f(CW$2\fR etc.,
1678 $`, $&, $', ${^PREMATCH}, ${^MATCH} and ${^POSTMATCH}. It uses slightly less
1679 memory, avoids string comparisons and numeric conversions during lookup, and
1680 uses 23 fewer lines of C. This change should not affect any external code.
1681 .IP \(bu 4
1682 Arrays now use NULL internally to represent unused slots, instead of
1683 &PL_sv_undef. &PL_sv_undef is no longer treated as a special value, so
1684 av_store(av, 0, &PL_sv_undef) will cause element 0 of that array to hold a
1685 read-only undefined scalar. \f(CW\*(C`$array[0] = anything\*(C'\fR will croak and
1686 \&\f(CW\*(C`\e$array[0]\*(C'\fR will compare equal to \f(CW\*(C`\eundef\*(C'\fR.
1687 .IP \(bu 4
1688 The SV returned by \fBHeSVKEY_force()\fR now correctly reflects the UTF8ness of the
1689 underlying hash key when that key is not stored as a SV. [perl #79074]
1690 .IP \(bu 4
1691 Certain rarely used functions and macros available to XS code are now
1692 deprecated. These are:
1693 \&\f(CW\*(C`utf8_to_uvuni_buf\*(C'\fR (use \f(CW\*(C`utf8_to_uvchr_buf\*(C'\fR instead),
1694 \&\f(CW\*(C`valid_utf8_to_uvuni\*(C'\fR (use \f(CW\*(C`utf8_to_uvchr_buf\*(C'\fR instead),
1695 \&\f(CW\*(C`NATIVE_TO_NEED\*(C'\fR (this did not work properly anyway),
1696 and \f(CW\*(C`ASCII_TO_NEED\*(C'\fR (this did not work properly anyway).
1697 .Sp
1698 Starting in this release, almost never does application code need to
1699 distinguish between the platform's character set and Latin1, on which the
1700 lowest 256 characters of Unicode are based. New code should not use
1701 \&\f(CW\*(C`utf8n_to_uvuni\*(C'\fR (use \f(CW\*(C`utf8_to_uvchr_buf\*(C'\fR instead),
1702 nor
1703 \&\f(CW\*(C`uvuni_to_utf8\*(C'\fR (use \f(CW\*(C`uvchr_to_utf8\*(C'\fR instead),
1704 .IP \(bu 4
1705 The Makefile shortcut targets for many rarely (or never) used testing and
1706 profiling targets have been removed, or merged into the only other Makefile
1707 target that uses them. Specifically, these targets are gone, along with
1708 documentation that referenced them or explained how to use them:
1709 .Sp
1710 .Vb 10
1711 \& check.third check.utf16 check.utf8 coretest minitest.prep
1712 \& minitest.utf16 perl.config.dashg perl.config.dashpg
1713 \& perl.config.gcov perl.gcov perl.gprof perl.gprof.config
1714 \& perl.pixie perl.pixie.atom perl.pixie.config perl.pixie.irix
1715 \& perl.third perl.third.config perl.valgrind.config purecovperl
1716 \& pureperl quantperl test.deparse test.taintwarn test.third
1717 \& test.torture test.utf16 test.utf8 test_notty.deparse
1718 \& test_notty.third test_notty.valgrind test_prep.third
1719 \& test_prep.valgrind torturetest ucheck ucheck.third ucheck.utf16
1720 \& ucheck.valgrind utest utest.third utest.utf16 utest.valgrind
1721 .Ve
1722 .Sp
1723 It's still possible to run the relevant commands by "hand" \- no underlying
1724 functionality has been removed.
1725 .IP \(bu 4
1726 It is now possible to keep Perl from initializing locale handling.
1727 For the most part, Perl doesn't pay attention to locale. (See
1728 perllocale.) Nonetheless, until now, on startup, it has always
1729 initialized locale handling to the system default, just in case the
1730 program being executed ends up using locales. (This is one of the first
1731 things a locale-aware program should do, long before Perl knows if it
1732 will actually be needed or not.) This works well except when Perl is
1733 embedded in another application which wants a locale that isn't the
1734 system default. Now, if the environment variable
1735 \&\f(CW\*(C`PERL_SKIP_LOCALE_INIT\*(C'\fR is set at the time Perl is started, this
1736 initialization step is skipped. Prior to this, on Windows platforms,
1737 the only workaround for this deficiency was to use a hacked-up copy of
1738 internal Perl code. Applications that need to use older Perls can
1739 discover if the embedded Perl they are using needs the workaround by
1740 testing that the C preprocessor symbol \f(CW\*(C`HAS_SKIP_LOCALE_INIT\*(C'\fR is not
1741 defined. [RT #38193]
1742 .IP \(bu 4
1743 \&\f(CW\*(C`BmRARE\*(C'\fR and \f(CW\*(C`BmPREVIOUS\*(C'\fR have been removed. They were not used anywhere
1744 and are not part of the API. For XS modules, they are now #defined as 0.
1745 .IP \(bu 4
1746 \&\f(CW\*(C`sv_force_normal\*(C'\fR, which usually croaks on read-only values, used to allow
1747 read-only values to be modified at compile time. This has been changed to
1748 croak on read-only values regardless. This change uncovered several core
1749 bugs.
1750 .IP \(bu 4
1751 Perl's new copy-on-write mechanism (which is now enabled by default),
1752 allows any \f(CW\*(C`SvPOK\*(C'\fR scalar to be automatically upgraded to a copy-on-write
1753 scalar when copied. A reference count on the string buffer is stored in
1754 the string buffer itself.
1755 .Sp
1756 For example:
1757 .Sp
1758 .Vb 10
1759 \& $ perl \-MDevel::Peek \-e\*(Aq$a="abc"; $b = $a; Dump $a; Dump $b\*(Aq
1760 \& SV = PV(0x260cd80) at 0x2620ad8
1761 \& REFCNT = 1
1762 \& FLAGS = (POK,IsCOW,pPOK)
1763 \& PV = 0x2619bc0 "abc"\e0
1764 \& CUR = 3
1765 \& LEN = 16
1766 \& COW_REFCNT = 1
1767 \& SV = PV(0x260ce30) at 0x2620b20
1768 \& REFCNT = 1
1769 \& FLAGS = (POK,IsCOW,pPOK)
1770 \& PV = 0x2619bc0 "abc"\e0
1771 \& CUR = 3
1772 \& LEN = 16
1773 \& COW_REFCNT = 1
1774 .Ve
1775 .Sp
1776 Note that both scalars share the same PV buffer and have a COW_REFCNT
1777 greater than zero.
1778 .Sp
1779 This means that XS code which wishes to modify the \f(CWSvPVX()\fR buffer of an
1780 SV should call \f(CWSvPV_force()\fR or similar first, to ensure a valid (and
1781 unshared) buffer, and to call \f(CWSvSETMAGIC()\fR afterwards. This in fact has
1782 always been the case (for example hash keys were already copy-on-write);
1783 this change just spreads the COW behaviour to a wider variety of SVs.
1784 .Sp
1785 One important difference is that before 5.18.0, shared hash-key scalars
1786 used to have the \f(CW\*(C`SvREADONLY\*(C'\fR flag set; this is no longer the case.
1787 .Sp
1788 This new behaviour can still be disabled by running \fIConfigure\fR with
1789 \&\fB\-Accflags=\-DPERL_NO_COW\fR. This option will probably be removed in Perl
1790 5.22.
1791 .IP \(bu 4
1792 \&\f(CW\*(C`PL_sawampersand\*(C'\fR is now a constant. The switch this variable provided
1793 (to enable/disable the pre-match copy depending on whether \f(CW$&\fR had been
1794 seen) has been removed and replaced with copy-on-write, eliminating a few
1795 bugs.
1796 .Sp
1797 The previous behaviour can still be enabled by running \fIConfigure\fR with
1798 \&\fB\-Accflags=\-DPERL_SAWAMPERSAND\fR.
1799 .IP \(bu 4
1800 The functions \f(CW\*(C`my_swap\*(C'\fR, \f(CW\*(C`my_htonl\*(C'\fR and \f(CW\*(C`my_ntohl\*(C'\fR have been removed.
1801 It is unclear why these functions were ever marked as \fIA\fR, part of the
1802 API. XS code can't call them directly, as it can't rely on them being
1803 compiled. Unsurprisingly, no code on CPAN references them.
1804 .IP \(bu 4
1805 The signature of the \f(CWPerl_re_intuit_start()\fR regex function has changed;
1806 the function pointer \f(CW\*(C`intuit\*(C'\fR in the regex engine plugin structure
1807 has also changed accordingly. A new parameter, \f(CW\*(C`strbeg\*(C'\fR has been added;
1808 this has the same meaning as the same-named parameter in
1809 \&\f(CW\*(C`Perl_regexec_flags\*(C'\fR. Previously intuit would try to guess the start of
1810 the string from the passed SV (if any), and would sometimes get it wrong
1811 (e.g. with an overloaded SV).
1812 .IP \(bu 4
1813 The signature of the \f(CWPerl_regexec_flags()\fR regex function has
1814 changed; the function pointer \f(CW\*(C`exec\*(C'\fR in the regex engine plugin
1815 structure has also changed to match. The \f(CW\*(C`minend\*(C'\fR parameter now has
1816 type \f(CW\*(C`SSize_t\*(C'\fR to better support 64\-bit systems.
1817 .IP \(bu 4
1818 XS code may use various macros to change the case of a character or code
1819 point (for example \f(CWtoLOWER_utf8()\fR). Only a couple of these were
1820 documented until now;
1821 and now they should be used in preference to calling the underlying
1822 functions. See "Character case changing" in perlapi.
1823 .IP \(bu 4
1824 The code dealt rather inconsistently with uids and gids. Some
1825 places assumed that they could be safely stored in UVs, others
1826 in IVs, others in ints. Four new macros are introduced:
1827 \&\fBSvUID()\fR, \fBsv_setuid()\fR, \fBSvGID()\fR, and \fBsv_setgid()\fR
1828 .IP \(bu 4
1829 \&\f(CW\*(C`sv_pos_b2u_flags\*(C'\fR has been added to the API. It is similar to \f(CW\*(C`sv_pos_b2u\*(C'\fR,
1830 but supports long strings on 64\-bit platforms.
1831 .IP \(bu 4
1832 \&\f(CW\*(C`PL_exit_flags\*(C'\fR can now be used by perl embedders or other XS code to have
1833 perl \f(CW\*(C`warn\*(C'\fR or \f(CW\*(C`abort\*(C'\fR on an attempted exit. [perl #52000]
1834 .IP \(bu 4
1835 Compiling with \f(CW\*(C`\-Accflags=\-PERL_BOOL_AS_CHAR\*(C'\fR now allows C99 and C++
1836 compilers to emulate the aliasing of \f(CW\*(C`bool\*(C'\fR to \f(CW\*(C`char\*(C'\fR that perl does for
1837 C89 compilers. [perl #120314]
1838 .IP \(bu 4
1839 The \f(CW\*(C`sv\*(C'\fR argument in "sv_2pv_flags" in perlapi, "sv_2iv_flags" in perlapi,
1840 "sv_2uv_flags" in perlapi, and "sv_2nv_flags" in perlapi and their older wrappers
1841 sv_2pv, sv_2iv, sv_2uv, sv_2nv, is now non-NULL. Passing NULL now will crash.
1842 When the non-NULL marker was introduced en masse in 5.9.3 the functions
1843 were marked non-NULL, but since the creation of the SV API in 5.0 alpha 2, if
1844 NULL was passed, the functions returned 0 or false-type values. The code that
1845 supports \f(CW\*(C`sv\*(C'\fR argument being non-NULL dates to 5.0 alpha 2 directly, and
1846 indirectly to Perl 1.0 (pre 5.0 api). The lack of documentation that the
1847 functions accepted a NULL \f(CW\*(C`sv\*(C'\fR was corrected in 5.11.0 and between 5.11.0
1848 and 5.19.5 the functions were marked NULLOK. As an optimization the NULLOK code
1849 has now been removed, and the functions became non-NULL marked again, because
1850 core getter-type macros never pass NULL to these functions and would crash
1851 before ever passing NULL.
1852 .Sp
1853 The only way a NULL \f(CW\*(C`sv\*(C'\fR can be passed to sv_2*v* functions is if XS code
1854 directly calls sv_2*v*. This is unlikely as XS code uses Sv*V* macros to get
1855 the underlying value out of the SV. One possible situation which leads to
1856 a NULL \f(CW\*(C`sv\*(C'\fR being passed to sv_2*v* functions, is if XS code defines its own
1857 getter type Sv*V* macros, which check for NULL \fBbefore\fR dereferencing and
1858 checking the SV's flags through public API Sv*OK* macros or directly using
1859 private API \f(CW\*(C`SvFLAGS\*(C'\fR, and if \f(CW\*(C`sv\*(C'\fR is NULL, then calling the sv_2*v functions
1860 with a NULL literal or passing the \f(CW\*(C`sv\*(C'\fR containing a NULL value.
1861 .IP \(bu 4
1862 newATTRSUB is now a macro
1863 .Sp
1864 The public API newATTRSUB was previously a macro to the private
1865 function Perl_newATTRSUB. Function Perl_newATTRSUB has been removed. newATTRSUB
1866 is now macro to a different internal function.
1867 .IP \(bu 4
1868 Changes in warnings raised by \f(CWutf8n_to_uvchr()\fR
1869 .Sp
1870 This bottom level function decodes the first character of a UTF\-8 string
1871 into a code point. It is accessible to \f(CW\*(C`XS\*(C'\fR level code, but it's
1872 discouraged from using it directly. There are higher level functions
1873 that call this that should be used instead, such as
1874 "utf8_to_uvchr_buf" in perlapi. For completeness though, this documents
1875 some changes to it. Now, tests for malformations are done before any
1876 tests for other potential issues. One of those issues involves code
1877 points so large that they have never appeared in any official standard
1878 (the current standard has scaled back the highest acceptable code point
1879 from earlier versions). It is possible (though not done in CPAN) to
1880 warn and/or forbid these code points, while accepting smaller code
1881 points that are still above the legal Unicode maximum. The warning
1882 message for this now includes the code point if representable on the
1883 machine. Previously it always displayed raw bytes, which is what it
1884 still does for non-representable code points.
1885 .IP \(bu 4
1886 Regexp engine changes that affect the pluggable regex engine interface
1887 .Sp
1888 Many flags that used to be exposed via regexp.h and used to populate the
1889 extflags member of struct regexp have been removed. These fields were
1890 technically private to Perl's own regexp engine and should not have been
1891 exposed there in the first place.
1892 .Sp
1893 The affected flags are:
1894 .Sp
1895 .Vb 8
1896 \& RXf_NOSCAN
1897 \& RXf_CANY_SEEN
1898 \& RXf_GPOS_SEEN
1899 \& RXf_GPOS_FLOAT
1900 \& RXf_ANCH_BOL
1901 \& RXf_ANCH_MBOL
1902 \& RXf_ANCH_SBOL
1903 \& RXf_ANCH_GPOS
1904 .Ve
1905 .Sp
1906 As well as the follow flag masks:
1907 .Sp
1908 .Vb 2
1909 \& RXf_ANCH_SINGLE
1910 \& RXf_ANCH
1911 .Ve
1912 .Sp
1913 All have been renamed to PREGf_ equivalents and moved to regcomp.h.
1914 .Sp
1915 The behavior previously achieved by setting one or more of the RXf_ANCH_
1916 flags (via the RXf_ANCH mask) have now been replaced by a *single* flag bit
1917 in extflags:
1918 .Sp
1919 .Vb 1
1920 \& RXf_IS_ANCHORED
1921 .Ve
1922 .Sp
1923 pluggable regex engines which previously used to set these flags should
1924 now set this flag ALONE.
1925 .IP \(bu 4
1926 The Perl core now consistently uses \f(CWav_tindex()\fR ("the top index of an
1927 array") as a more clearly-named synonym for \f(CWav_len()\fR.
1928 .IP \(bu 4
1929 The obscure interpreter variable \f(CW\*(C`PL_timesbuf\*(C'\fR is expected to be removed
1930 early in the 5.21.x development series, so that Perl 5.22.0 will not provide
1931 it to XS authors. While the variable still exists in 5.20.0, we hope that
1932 this advance warning of the deprecation will help anyone who is using that
1933 variable.
1934 .SH "Selected Bug Fixes"
1935 .IX Header "Selected Bug Fixes"
1936 .SS "Regular Expressions"
1937 .IX Subsection "Regular Expressions"
1938 .IP \(bu 4
1939 Fixed a small number of regexp constructions that could either fail to
1940 match or crash perl when the string being matched against was
1941 allocated above the 2GB line on 32\-bit systems. [RT #118175]
1942 .IP \(bu 4
1943 Various memory leaks involving the parsing of the \f(CW\*(C`(?[...])\*(C'\fR regular
1944 expression construct have been fixed.
1945 .IP \(bu 4
1946 \&\f(CW\*(C`(?[...])\*(C'\fR now allows interpolation of precompiled patterns consisting of
1947 \&\f(CW\*(C`(?[...])\*(C'\fR with bracketed character classes inside (\f(CW\*(C`$pat =
1948 qr/(?[\ [a]\ ])/; /(?[\ $pat\ ])/\*(C'\fR). Formerly, the brackets would
1949 confuse the regular expression parser.
1950 .IP \(bu 4
1951 The "Quantifier unexpected on zero-length expression" warning message could
1952 appear twice starting in Perl v5.10 for a regular expression also
1953 containing alternations (e.g., "a|b") triggering the trie optimisation.
1954 .IP \(bu 4
1955 Perl v5.18 inadvertently introduced a bug whereby interpolating mixed up\-
1956 and down-graded UTF\-8 strings in a regex could result in malformed UTF\-8
1957 in the pattern: specifically if a downgraded character in the range
1958 \&\f(CW\*(C`\ex80..\exff\*(C'\fR followed a UTF\-8 string, e.g.
1959 .Sp
1960 .Vb 3
1961 \& utf8::upgrade( my $u = "\ex{e5}");
1962 \& utf8::downgrade(my $d = "\ex{e5}");
1963 \& /$u$d/
1964 .Ve
1965 .Sp
1966 [RT #118297]
1967 .IP \(bu 4
1968 In regular expressions containing multiple code blocks, the values of
1969 \&\f(CW$1\fR, \f(CW$2\fR, etc., set by nested regular expression calls would leak from
1970 one block to the next. Now these variables always refer to the outer
1971 regular expression at the start of an embedded block [perl #117917].
1972 .IP \(bu 4
1973 \&\f(CW\*(C`/$qr/p\*(C'\fR was broken in Perl 5.18.0; the \f(CW\*(C`/p\*(C'\fR flag was ignored. This has been
1974 fixed. [perl #118213]
1975 .IP \(bu 4
1976 Starting in Perl 5.18.0, a construct like \f(CW\*(C`/[#](?{})/x\*(C'\fR would have its \f(CW\*(C`#\*(C'\fR
1977 incorrectly interpreted as a comment. The code block would be skipped,
1978 unparsed. This has been corrected.
1979 .IP \(bu 4
1980 Starting in Perl 5.001, a regular expression like \f(CW\*(C`/[#$a]/x\*(C'\fR or \f(CW\*(C`/[#]$a/x\*(C'\fR
1981 would have its \f(CW\*(C`#\*(C'\fR incorrectly interpreted as a comment, so the variable would
1982 not interpolate. This has been corrected. [perl #45667]
1983 .IP \(bu 4
1984 Perl 5.18.0 inadvertently made dereferenced regular expressions
1985 (\f(CW\*(C`${\ qr//\ }\*(C'\fR) false as booleans. This has been fixed.
1986 .IP \(bu 4
1987 The use of \f(CW\*(C`\eG\*(C'\fR in regular expressions, where it's not at the start of the
1988 pattern, is now slightly less buggy (although it is still somewhat
1989 problematic).
1990 .IP \(bu 4
1991 Where a regular expression included code blocks (\f(CW\*(C`/(?{...})/\*(C'\fR), and where the
1992 use of constant overloading triggered a re-compilation of the code block, the
1993 second compilation didn't see its outer lexical scope. This was a regression
1994 in Perl 5.18.0.
1995 .IP \(bu 4
1996 The string position set by \f(CW\*(C`pos\*(C'\fR could shift if the string changed
1997 representation internally to or from utf8. This could happen, e.g., with
1998 references to objects with string overloading.
1999 .IP \(bu 4
2000 Taking references to the return values of two \f(CW\*(C`pos\*(C'\fR calls with the same
2001 argument, and then assigning a reference to one and \f(CW\*(C`undef\*(C'\fR to the other,
2002 could result in assertion failures or memory leaks.
2003 .IP \(bu 4
2004 Elements of @\- and @+ now update correctly when they refer to non-existent
2005 captures. Previously, a referenced element (\f(CW\*(C`$ref = \e$\-[1]\*(C'\fR) could refer to
2006 the wrong match after subsequent matches.
2007 .IP \(bu 4
2008 The code that parses regex backrefs (or ambiguous backref/octals) such as \e123
2009 did a simple \fBatoi()\fR, which could wrap round to negative values on long digit
2010 strings and cause segmentation faults. This has now been fixed. [perl
2011 #119505]
2012 .IP \(bu 4
2013 Assigning another typeglob to \f(CW\*(C`*^R\*(C'\fR no longer makes the regular expression
2014 engine crash.
2015 .IP \(bu 4
2016 The \f(CW\*(C`\eN\*(C'\fR regular expression escape, when used without the curly braces (to
2017 mean \f(CW\*(C`[^\en]\*(C'\fR), was ignoring a following \f(CW\*(C`*\*(C'\fR if followed by whitespace
2018 under /x. It had been this way since \f(CW\*(C`\eN\*(C'\fR to mean \f(CW\*(C`[^\en]\*(C'\fR was introduced
2019 in 5.12.0.
2020 .IP \(bu 4
2021 \&\f(CW\*(C`s///\*(C'\fR, \f(CW\*(C`tr///\*(C'\fR and \f(CW\*(C`y///\*(C'\fR now work when a wide character is used as the
2022 delimiter. [perl #120463]
2023 .IP \(bu 4
2024 Some cases of unterminated (?...) sequences in regular expressions (e.g.,
2025 \&\f(CW\*(C`/(?</\*(C'\fR) have been fixed to produce the proper error message instead of
2026 "panic: memory wrap". Other cases (e.g., \f(CW\*(C`/(?(/\*(C'\fR) have yet to be fixed.
2027 .IP \(bu 4
2028 When a reference to a reference to an overloaded object was returned from
2029 a regular expression \f(CW\*(C`(??{...})\*(C'\fR code block, an incorrect implicit
2030 dereference could take place if the inner reference had been returned by
2031 a code block previously.
2032 .IP \(bu 4
2033 A tied variable returned from \f(CW\*(C`(??{...})\*(C'\fR sees the inner values of match
2034 variables (i.e., the \f(CW$1\fR etc. from any matches inside the block) in its
2035 FETCH method. This was not the case if a reference to an overloaded object
2036 was the last thing assigned to the tied variable. Instead, the match
2037 variables referred to the outer pattern during the FETCH call.
2038 .IP \(bu 4
2039 Fix unexpected tainting via regexp using locale. Previously, under certain
2040 conditions, the use of character classes could cause tainting when it
2041 shouldn't. Some character classes are locale-dependent, but before this
2042 patch, sometimes tainting was happening even for character classes that
2043 don't depend on the locale. [perl #120675]
2044 .IP \(bu 4
2045 Under certain conditions, Perl would throw an error if in a lookbehind
2046 assertion in a regexp, the assertion referred to a named subpattern,
2047 complaining the lookbehind was variable when it wasn't. This has been
2048 fixed. [perl #120600], [perl #120618]. The current fix may be improved
2049 on in the future.
2050 .IP \(bu 4
2051 \&\f(CW$^R\fR wasn't available outside of the regular expression that
2052 initialized it. [perl #121070]
2053 .IP \(bu 4
2054 A large set of fixes and refactoring for \fBre_intuit_start()\fR was merged,
2055 the highlights are:
2056 .RS 4
2057 .IP \(bu 4
2058 Fixed a panic when compiling the regular expression
2059 \&\f(CW\*(C`/\ex{100}[xy]\ex{100}{2}/\*(C'\fR.
2060 .IP \(bu 4
2061 Fixed a performance regression when performing a global pattern match
2062 against a UTF\-8 string. [perl #120692]
2063 .IP \(bu 4
2064 Fixed another performance issue where matching a regular expression
2065 like \f(CW\*(C`/ab.{1,2}x/\*(C'\fR against a long UTF\-8 string would unnecessarily
2066 calculate byte offsets for a large portion of the string. [perl
2067 #120692]
2068 .RE
2069 .RS 4
2070 .RE
2071 .IP \(bu 4
2072 Fixed an alignment error when compiling regular expressions when built
2073 with GCC on HP-UX 64\-bit.
2074 .IP \(bu 4
2075 On 64\-bit platforms \f(CW\*(C`pos\*(C'\fR can now be set to a value higher than 2**31\-1.
2076 [perl #72766]
2077 .SS "Perl 5 Debugger and \-d"
2078 .IX Subsection "Perl 5 Debugger and -d"
2079 .IP \(bu 4
2080 The debugger's \f(CW\*(C`man\*(C'\fR command been fixed. It was broken in the v5.18.0
2081 release. The \f(CW\*(C`man\*(C'\fR command is aliased to the names \f(CW\*(C`doc\*(C'\fR and \f(CW\*(C`perldoc\*(C'\fR \-
2082 all now work again.
2083 .IP \(bu 4
2084 \&\f(CW@_\fR is now correctly visible in the debugger, fixing a regression
2085 introduced in v5.18.0's debugger. [RT #118169]
2086 .IP \(bu 4
2087 Under copy-on-write builds (the default as of 5.20.0) \f(CW\*(C`${\*(Aq_<\-e\*(Aq}[0]\*(C'\fR
2088 no longer gets mangled. This is the first line of input saved for the
2089 debugger's use for one-liners [perl #118627].
2090 .IP \(bu 4
2091 On non-threaded builds, setting \f(CW\*(C`${"_<filename"}\*(C'\fR to a reference or
2092 typeglob no longer causes \f(CW\*(C`_\|_FILE_\|_\*(C'\fR and some error messages to produce a
2093 corrupt string, and no longer prevents \f(CW\*(C`#line\*(C'\fR directives in string evals from
2094 providing the source lines to the debugger. Threaded builds were unaffected.
2095 .IP \(bu 4
2096 Starting with Perl 5.12, line numbers were off by one if the \fB\-d\fR switch was
2097 used on the #! line. Now they are correct.
2098 .IP \(bu 4
2099 \&\f(CW\*(C`*DB::DB = sub {} if 0\*(C'\fR no longer stops Perl's debugging mode from finding
2100 \&\f(CW\*(C`DB::DB\*(C'\fR subs declared thereafter.
2101 .IP \(bu 4
2102 \&\f(CW\*(C`%{\*(Aq_<...\*(Aq}\*(C'\fR hashes now set breakpoints on the corresponding \f(CW\*(C`@{\*(Aq_<...\*(Aq}\*(C'\fR
2103 rather than whichever array \f(CW@DB::dbline\fR is aliased to. [perl #119799]
2104 .IP \(bu 4
2105 Call set-magic when setting \f(CW$DB::sub\fR. [perl #121255]
2106 .IP \(bu 4
2107 The debugger's "n" command now respects lvalue subroutines and steps over
2108 them [perl #118839].
2109 .SS "Lexical Subroutines"
2110 .IX Subsection "Lexical Subroutines"
2111 .IP \(bu 4
2112 Lexical constants (\f(CW\*(C`my sub a() { 42 }\*(C'\fR) no longer crash when inlined.
2113 .IP \(bu 4
2114 Parameter prototypes attached to lexical subroutines are now respected when
2115 compiling sub calls without parentheses. Previously, the prototypes were
2116 honoured only for calls \fIwith\fR parentheses. [RT #116735]
2117 .IP \(bu 4
2118 Syntax errors in lexical subroutines in combination with calls to the same
2119 subroutines no longer cause crashes at compile time.
2120 .IP \(bu 4
2121 Deep recursion warnings no longer crash lexical subroutines. [RT #118521]
2122 .IP \(bu 4
2123 The dtrace sub-entry probe now works with lexical subs, instead of
2124 crashing [perl #118305].
2125 .IP \(bu 4
2126 Undefining an inlinable lexical subroutine (\f(CWmy sub foo() { 42 } undef
2127 &foo\fR) would result in a crash if warnings were turned on.
2128 .IP \(bu 4
2129 An undefined lexical sub used as an inherited method no longer crashes.
2130 .IP \(bu 4
2131 The presence of a lexical sub named "CORE" no longer stops the CORE::
2132 prefix from working.
2133 .SS "Everything Else"
2134 .IX Subsection "Everything Else"
2135 .IP \(bu 4
2136 The OP allocation code now returns correctly aligned memory in all cases
2137 for \f(CW\*(C`struct pmop\*(C'\fR. Previously it could return memory only aligned to a
2138 4\-byte boundary, which is not correct for an ithreads build with 64 bit IVs
2139 on some 32 bit platforms. Notably, this caused the build to fail completely
2140 on sparc GNU/Linux. [RT #118055]
2141 .IP \(bu 4
2142 Evaluating large hashes in scalar context is now much faster, as the number
2143 of used chains in the hash is now cached for larger hashes. Smaller hashes
2144 continue not to store it and calculate it when needed, as this saves one IV.
2145 That would be 1 IV overhead for every object built from a hash. [RT #114576]
2146 .IP \(bu 4
2147 Perl v5.16 inadvertently introduced a bug whereby calls to XSUBs that were
2148 not visible at compile time were treated as lvalues and could be assigned
2149 to, even when the subroutine was not an lvalue sub. This has been fixed.
2150 [RT #117947]
2151 .IP \(bu 4
2152 In Perl v5.18.0 dualvars that had an empty string for the string part but a
2153 non-zero number for the number part starting being treated as true. In
2154 previous versions they were treated as false, the string representation
2155 taking precedence. The old behaviour has been restored. [RT #118159]
2156 .IP \(bu 4
2157 Since Perl v5.12, inlining of constants that override built-in keywords of
2158 the same name had countermanded \f(CW\*(C`use subs\*(C'\fR, causing subsequent mentions of
2159 the constant to use the built-in keyword instead. This has been fixed.
2160 .IP \(bu 4
2161 The warning produced by \f(CW\*(C`\-l $handle\*(C'\fR now applies to IO refs and globs, not
2162 just to glob refs. That warning is also now UTF8\-clean. [RT #117595]
2163 .IP \(bu 4
2164 \&\f(CW\*(C`delete local $ENV{nonexistent_env_var}\*(C'\fR no longer leaks memory.
2165 .IP \(bu 4
2166 \&\f(CW\*(C`sort\*(C'\fR and \f(CW\*(C`require\*(C'\fR followed by a keyword prefixed with \f(CW\*(C`CORE::\*(C'\fR now
2167 treat it as a keyword, and not as a subroutine or module name. [RT #24482]
2168 .IP \(bu 4
2169 Through certain conundrums, it is possible to cause the current package to
2170 be freed. Certain operators (\f(CW\*(C`bless\*(C'\fR, \f(CW\*(C`reset\*(C'\fR, \f(CW\*(C`open\*(C'\fR, \f(CW\*(C`eval\*(C'\fR) could
2171 not cope and would crash. They have been made more resilient. [RT #117941]
2172 .IP \(bu 4
2173 Aliasing filehandles through glob-to-glob assignment would not update
2174 internal method caches properly if a package of the same name as the
2175 filehandle existed, resulting in filehandle method calls going to the
2176 package instead. This has been fixed.
2177 .IP \(bu 4
2178 \&\f(CW\*(C`./Configure \-de \-Dusevendorprefix\*(C'\fR didn't default. [RT #64126]
2179 .IP \(bu 4
2180 The \f(CW\*(C`Statement unlikely to be reached\*(C'\fR warning was listed in
2181 perldiag as an \f(CW\*(C`exec\*(C'\fR\-category warning, but was enabled and disabled
2182 by the \f(CW\*(C`syntax\*(C'\fR category. On the other hand, the \f(CW\*(C`exec\*(C'\fR category
2183 controlled its fatal-ness. It is now entirely handled by the \f(CW\*(C`exec\*(C'\fR
2184 category.
2185 .IP \(bu 4
2186 The "Replacement list is longer that search list" warning for \f(CW\*(C`tr///\*(C'\fR and
2187 \&\f(CW\*(C`y///\*(C'\fR no longer occurs in the presence of the \f(CW\*(C`/c\*(C'\fR flag. [RT #118047]
2188 .IP \(bu 4
2189 Stringification of NVs are not cached so that the lexical locale controls
2190 stringification of the decimal point. [perl #108378] [perl #115800]
2191 .IP \(bu 4
2192 There have been several fixes related to Perl's handling of locales. perl
2193 #38193 was described above in "Internal Changes".
2194 Also fixed is
2195 #118197, where the radix (decimal point) character had to be an ASCII
2196 character (which doesn't work for some non-Western languages);
2197 and #115808, in which \f(CWPOSIX::setlocale()\fR on failure returned an
2198 \&\f(CW\*(C`undef\*(C'\fR which didn't warn about not being defined even if those
2199 warnings were enabled.
2200 .IP \(bu 4
2201 Compiling a \f(CW\*(C`split\*(C'\fR operator whose third argument is a named constant
2202 evaluating to 0 no longer causes the constant's value to change.
2203 .IP \(bu 4
2204 A named constant used as the second argument to \f(CW\*(C`index\*(C'\fR no longer gets
2205 coerced to a string if it is a reference, regular expression, dualvar, etc.
2206 .IP \(bu 4
2207 A named constant evaluating to the undefined value used as the second
2208 argument to \f(CW\*(C`index\*(C'\fR no longer produces "uninitialized" warnings at compile
2209 time. It will still produce them at run time.
2210 .IP \(bu 4
2211 When a scalar was returned from a subroutine in \f(CW@INC\fR, the referenced scalar
2212 was magically converted into an IO thingy, possibly resulting in "Bizarre
2213 copy" errors if that scalar continued to be used elsewhere. Now Perl uses
2214 an internal copy of the scalar instead.
2215 .IP \(bu 4
2216 Certain uses of the \f(CW\*(C`sort\*(C'\fR operator are optimised to modify an array in
2217 place, such as \f(CW\*(C`@a = sort @a\*(C'\fR. During the sorting, the array is made
2218 read-only. If a sort block should happen to die, then the array remained
2219 read-only even outside the \f(CW\*(C`sort\*(C'\fR. This has been fixed.
2220 .IP \(bu 4
2221 \&\f(CW$a\fR and \f(CW$b\fR inside a sort block are aliased to the actual arguments to
2222 \&\f(CW\*(C`sort\*(C'\fR, so they can be modified through those two variables. This did not
2223 always work, e.g., for lvalue subs and \f(CW$#ary\fR, and probably many other
2224 operators. It works now.
2225 .IP \(bu 4
2226 The arguments to \f(CW\*(C`sort\*(C'\fR are now all in list context. If the \f(CW\*(C`sort\*(C'\fR
2227 itself were called in void or scalar context, then \fIsome\fR, but not all, of
2228 the arguments used to be in void or scalar context.
2229 .IP \(bu 4
2230 Subroutine prototypes with Unicode characters above U+00FF were getting
2231 mangled during closure cloning. This would happen with subroutines closing
2232 over lexical variables declared outside, and with lexical subs.
2233 .IP \(bu 4
2234 \&\f(CW\*(C`UNIVERSAL::can\*(C'\fR now treats its first argument the same way that method
2235 calls do: Typeglobs and glob references with non-empty IO slots are treated
2236 as handles, and strings are treated as filehandles, rather than packages,
2237 if a handle with that name exists [perl #113932].
2238 .IP \(bu 4
2239 Method calls on typeglobs (e.g., \f(CW\*(C`*ARGV\->getline\*(C'\fR) used to stringify
2240 the typeglob and then look it up again. Combined with changes in Perl
2241 5.18.0, this allowed \f(CW\*(C`*foo\->bar\*(C'\fR to call methods on the "foo" package
2242 (like \f(CW\*(C`foo\->bar\*(C'\fR). In some cases it could cause the method to be
2243 called on the wrong handle. Now a typeglob argument is treated as a
2244 handle (just like \f(CW\*(C`(\e*foo)\->bar\*(C'\fR), or, if its IO slot is empty, an
2245 error is raised.
2246 .IP \(bu 4
2247 Assigning a vstring to a tied variable or to a subroutine argument aliased
2248 to a nonexistent hash or array element now works, without flattening the
2249 vstring into a regular string.
2250 .IP \(bu 4
2251 \&\f(CW\*(C`pos\*(C'\fR, \f(CW\*(C`tie\*(C'\fR, \f(CW\*(C`tied\*(C'\fR and \f(CW\*(C`untie\*(C'\fR did not work
2252 properly on subroutine arguments aliased to nonexistent
2253 hash and array elements [perl #77814, #27010].
2254 .IP \(bu 4
2255 The \f(CW\*(C`=>\*(C'\fR fat arrow operator can now quote built-in keywords even if it
2256 occurs on the next line, making it consistent with how it treats other
2257 barewords.
2258 .IP \(bu 4
2259 Autovivifying a subroutine stub via \f(CW\*(C`\e&$glob\*(C'\fR started causing crashes in Perl
2260 5.18.0 if the \f(CW$glob\fR was merely a copy of a real glob, i.e., a scalar that had
2261 had a glob assigned to it. This has been fixed. [perl #119051]
2262 .IP \(bu 4
2263 Perl used to leak an implementation detail when it came to referencing the
2264 return values of certain operators. \f(CW\*(C`for ($a+$b) { warn \e$_; warn \e$_ }\*(C'\fR used
2265 to display two different memory addresses, because the \f(CW\*(C`\e\*(C'\fR operator was
2266 copying the variable. Under threaded builds, it would also happen for
2267 constants (\f(CW\*(C`for(1) { ... }\*(C'\fR). This has been fixed. [perl #21979, #78194,
2268 #89188, #109746, #114838, #115388]
2269 .IP \(bu 4
2270 The range operator \f(CW\*(C`..\*(C'\fR was returning the same modifiable scalars with each
2271 call, unless it was the only thing in a \f(CW\*(C`foreach\*(C'\fR loop header. This meant
2272 that changes to values within the list returned would be visible the next time
2273 the operator was executed. [perl #3105]
2274 .IP \(bu 4
2275 Constant folding and subroutine inlining no longer cause operations that would
2276 normally return new modifiable scalars to return read-only values instead.
2277 .IP \(bu 4
2278 Closures of the form \f(CW\*(C`sub () { $some_variable }\*(C'\fR are no longer inlined,
2279 causing changes to the variable to be ignored by callers of the subroutine.
2280 [perl #79908]
2281 .IP \(bu 4
2282 Return values of certain operators such as \f(CW\*(C`ref\*(C'\fR would sometimes be shared
2283 between recursive calls to the same subroutine, causing the inner call to
2284 modify the value returned by \f(CW\*(C`ref\*(C'\fR in the outer call. This has been fixed.
2285 .IP \(bu 4
2286 \&\f(CW\*(C`_\|_PACKAGE_\|_\*(C'\fR and constants returning a package name or hash key are now
2287 consistently read-only. In various previous Perl releases, they have become
2288 mutable under certain circumstances.
2289 .IP \(bu 4
2290 Enabling "used once" warnings no longer causes crashes on stash circularities
2291 created at compile time (\f(CW\*(C`*Foo::Bar::Foo:: = *Foo::\*(C'\fR).
2292 .IP \(bu 4
2293 Undef constants used in hash keys (\f(CW\*(C`use constant u => undef; $h{+u}\*(C'\fR) no
2294 longer produce "uninitialized" warnings at compile time.
2295 .IP \(bu 4
2296 Modifying a substitution target inside the substitution replacement no longer
2297 causes crashes.
2298 .IP \(bu 4
2299 The first statement inside a string eval used to use the wrong pragma setting
2300 sometimes during constant folding. \f(CW\*(C`eval \*(Aquc chr 0xe0\*(Aq\*(C'\fR would randomly choose
2301 between Unicode, byte, and locale semantics. This has been fixed.
2302 .IP \(bu 4
2303 The handling of return values of \f(CW@INC\fR filters (subroutines returned by
2304 subroutines in \f(CW@INC\fR) has been fixed in various ways. Previously tied variables
2305 were mishandled, and setting \f(CW$_\fR to a reference or typeglob could result in
2306 crashes.
2307 .IP \(bu 4
2308 The \f(CW\*(C`SvPVbyte\*(C'\fR XS function has been fixed to work with tied scalars returning
2309 something other than a string. It used to return utf8 in those cases where
2310 \&\f(CW\*(C`SvPV\*(C'\fR would.
2311 .IP \(bu 4
2312 Perl 5.18.0 inadvertently made \f(CW\*(C`\-\-\*(C'\fR and \f(CW\*(C`++\*(C'\fR crash on dereferenced regular
2313 expressions, and stopped \f(CW\*(C`++\*(C'\fR from flattening vstrings.
2314 .IP \(bu 4
2315 \&\f(CW\*(C`bless\*(C'\fR no longer dies with "Can't bless non-reference value" if its first
2316 argument is a tied reference.
2317 .IP \(bu 4
2318 \&\f(CW\*(C`reset\*(C'\fR with an argument no longer skips copy-on-write scalars, regular
2319 expressions, typeglob copies, and vstrings. Also, when encountering those or
2320 read-only values, it no longer skips any array or hash with the same name.
2321 .IP \(bu 4
2322 \&\f(CW\*(C`reset\*(C'\fR with an argument now skips scalars aliased to typeglobs
2323 (\f(CW\*(C`for $z (*foo) { reset "z" }\*(C'\fR). Previously it would corrupt memory or crash.
2324 .IP \(bu 4
2325 \&\f(CW\*(C`ucfirst\*(C'\fR and \f(CW\*(C`lcfirst\*(C'\fR were not respecting the bytes pragma. This was a
2326 regression from Perl 5.12. [perl #117355]
2327 .IP \(bu 4
2328 Changes to \f(CW\*(C`UNIVERSAL::DESTROY\*(C'\fR now update DESTROY caches in all classes,
2329 instead of causing classes that have already had objects destroyed to continue
2330 using the old sub. This was a regression in Perl 5.18. [perl #114864]
2331 .IP \(bu 4
2332 All known false-positive occurrences of the deprecation warning "Useless use of
2333 \&'\e'; doesn't escape metacharacter '%c'", added in Perl 5.18.0, have been
2334 removed. [perl #119101]
2335 .IP \(bu 4
2336 The value of $^E is now saved across signal handlers on Windows. [perl #85104]
2337 .IP \(bu 4
2338 A lexical filehandle (as in \f(CW\*(C`open my $fh...\*(C'\fR) is usually given a name based on
2339 the current package and the name of the variable, e.g. "main::$fh". Under
2340 recursion, the filehandle was losing the "$fh" part of the name. This has been
2341 fixed.
2342 .IP \(bu 4
2343 Uninitialized values returned by XSUBs are no longer exempt from uninitialized
2344 warnings. [perl #118693]
2345 .IP \(bu 4
2346 \&\f(CW\*(C`elsif ("")\*(C'\fR no longer erroneously produces a warning about void context.
2347 [perl #118753]
2348 .IP \(bu 4
2349 Passing \f(CW\*(C`undef\*(C'\fR to a subroutine now causes \f(CW@_\fR to contain the same read-only
2350 undefined scalar that \f(CW\*(C`undef\*(C'\fR returns. Furthermore, \f(CW\*(C`exists $_[0]\*(C'\fR will now
2351 return true if \f(CW\*(C`undef\*(C'\fR was the first argument. [perl #7508, #109726]
2352 .IP \(bu 4
2353 Passing a non-existent array element to a subroutine does not usually
2354 autovivify it unless the subroutine modifies its argument. This did not work
2355 correctly with negative indices and with non-existent elements within the
2356 array. The element would be vivified immediately. The delayed vivification
2357 has been extended to work with those. [perl #118691]
2358 .IP \(bu 4
2359 Assigning references or globs to the scalar returned by $#foo after the \f(CW@foo\fR
2360 array has been freed no longer causes assertion failures on debugging builds
2361 and memory leaks on regular builds.
2362 .IP \(bu 4
2363 On 64\-bit platforms, large ranges like 1..1000000000000 no longer crash, but
2364 eat up all your memory instead. [perl #119161]
2365 .IP \(bu 4
2366 \&\f(CW\*(C`_\|_DATA_\|_\*(C'\fR now puts the \f(CW\*(C`DATA\*(C'\fR handle in the right package, even if the
2367 current package has been renamed through glob assignment.
2368 .IP \(bu 4
2369 When \f(CW\*(C`die\*(C'\fR, \f(CW\*(C`last\*(C'\fR, \f(CW\*(C`next\*(C'\fR, \f(CW\*(C`redo\*(C'\fR, \f(CW\*(C`goto\*(C'\fR and \f(CW\*(C`exit\*(C'\fR unwind the scope,
2370 it is possible for \f(CW\*(C`DESTROY\*(C'\fR recursively to call a subroutine or format that
2371 is currently being exited. It that case, sometimes the lexical variables
2372 inside the sub would start out having values from the outer call, instead of
2373 being undefined as they should. This has been fixed. [perl #119311]
2374 .IP \(bu 4
2375 ${^MPEN} is no longer treated as a synonym for ${^MATCH}.
2376 .IP \(bu 4
2377 Perl now tries a little harder to return the correct line number in
2378 \&\f(CW\*(C`(caller)[2]\*(C'\fR. [perl #115768]
2379 .IP \(bu 4
2380 Line numbers inside multiline quote-like operators are now reported correctly.
2381 [perl #3643]
2382 .IP \(bu 4
2383 \&\f(CW\*(C`#line\*(C'\fR directives inside code embedded in quote-like operators are now
2384 respected.
2385 .IP \(bu 4
2386 Line numbers are now correct inside the second here-doc when two here-doc
2387 markers occur on the same line.
2388 .IP \(bu 4
2389 An optimization in Perl 5.18 made incorrect assumptions causing a bad
2390 interaction with the Devel::CallParser CPAN module. If the module was
2391 loaded then lexical variables declared in separate statements following a
2392 \&\f(CWmy(...)\fR list might fail to be cleared on scope exit.
2393 .IP \(bu 4
2394 \&\f(CW&xsub\fR and \f(CW\*(C`goto &xsub\*(C'\fR calls now allow the called subroutine to autovivify
2395 elements of \f(CW@_\fR.
2396 .IP \(bu 4
2397 \&\f(CW&xsub\fR and \f(CW\*(C`goto &xsub\*(C'\fR no longer crash if *_ has been undefined and has no
2398 ARRAY entry (i.e. \f(CW@_\fR does not exist).
2399 .IP \(bu 4
2400 \&\f(CW&xsub\fR and \f(CW\*(C`goto &xsub\*(C'\fR now work with tied \f(CW@_\fR.
2401 .IP \(bu 4
2402 Overlong identifiers no longer cause a buffer overflow (and a crash). They
2403 started doing so in Perl 5.18.
2404 .IP \(bu 4
2405 The warning "Scalar value \f(CW@hash\fR{foo} better written as \f(CW$hash\fR{foo}" now produces
2406 far fewer false positives. In particular, \f(CW@hash{+function_returning_a_list}\fR
2407 and \f(CW@hash{ qw "foo bar baz" }\fR no longer warn. The same applies to array
2408 slices. [perl #28380, #114024]
2409 .IP \(bu 4
2410 \&\f(CW\*(C`$! = EINVAL; waitpid(0, WNOHANG);\*(C'\fR no longer goes into an internal infinite
2411 loop. [perl #85228]
2412 .IP \(bu 4
2413 A possible segmentation fault in filehandle duplication has been fixed.
2414 .IP \(bu 4
2415 A subroutine in \f(CW@INC\fR can return a reference to a scalar containing the initial
2416 contents of the file. However, that scalar was freed prematurely if not
2417 referenced elsewhere, giving random results.
2418 .IP \(bu 4
2419 \&\f(CW\*(C`last\*(C'\fR no longer returns values that the same statement has accumulated so
2420 far, fixing amongst other things the long-standing bug that \f(CW\*(C`push @a, last\*(C'\fR
2421 would try to return the \f(CW@a\fR, copying it like a scalar in the process and
2422 resulting in the error, "Bizarre copy of ARRAY in last." [perl #3112]
2423 .IP \(bu 4
2424 In some cases, closing file handles opened to pipe to or from a process, which
2425 had been duplicated into a standard handle, would call perl's internal waitpid
2426 wrapper with a pid of zero. With the fix for [perl #85228] this zero pid was
2427 passed to \f(CW\*(C`waitpid\*(C'\fR, possibly blocking the process. This wait for process
2428 zero no longer occurs. [perl #119893]
2429 .IP \(bu 4
2430 \&\f(CW\*(C`select\*(C'\fR used to ignore magic on the fourth (timeout) argument, leading to
2431 effects such as \f(CW\*(C`select\*(C'\fR blocking indefinitely rather than the expected sleep
2432 time. This has now been fixed. [perl #120102]
2433 .IP \(bu 4
2434 The class name in \f(CW\*(C`for my class $foo\*(C'\fR is now parsed correctly. In the case of
2435 the second character of the class name being followed by a digit (e.g. 'a1b')
2436 this used to give the error "Missing $ on loop variable". [perl #120112]
2437 .IP \(bu 4
2438 Perl 5.18.0 accidentally disallowed \f(CW\*(C`\-bareword\*(C'\fR under \f(CW\*(C`use strict\*(C'\fR and
2439 \&\f(CW\*(C`use integer\*(C'\fR. This has been fixed. [perl #120288]
2440 .IP \(bu 4
2441 \&\f(CW\*(C`\-a\*(C'\fR at the start of a line (or a hyphen with any single letter that is
2442 not a filetest operator) no longer produces an erroneous 'Use of "\-a"
2443 without parentheses is ambiguous' warning. [perl #120288]
2444 .IP \(bu 4
2445 Lvalue context is now properly propagated into bare blocks and \f(CW\*(C`if\*(C'\fR and
2446 \&\f(CW\*(C`else\*(C'\fR blocks in lvalue subroutines. Previously, arrays and hashes would
2447 sometimes incorrectly be flattened when returned in lvalue list context, or
2448 "Bizarre copy" errors could occur. [perl #119797]
2449 .IP \(bu 4
2450 Lvalue context is now propagated to the branches of \f(CW\*(C`||\*(C'\fR and \f(CW\*(C`&&\*(C'\fR (and
2451 their alphabetic equivalents, \f(CW\*(C`or\*(C'\fR and \f(CW\*(C`and\*(C'\fR). This means
2452 \&\f(CW\*(C`foreach (pos $x || pos $y) {...}\*(C'\fR now allows \f(CW\*(C`pos\*(C'\fR to be modified
2453 through \f(CW$_\fR.
2454 .IP \(bu 4
2455 \&\f(CW\*(C`stat\*(C'\fR and \f(CW\*(C`readline\*(C'\fR remember the last handle used; the former
2456 for the special \f(CW\*(C`_\*(C'\fR filehandle, the latter for \f(CW\*(C`${^LAST_FH}\*(C'\fR.
2457 \&\f(CW\*(C`eval "*foo if 0"\*(C'\fR where *foo was the last handle passed to \f(CW\*(C`stat\*(C'\fR
2458 or \f(CW\*(C`readline\*(C'\fR could cause that handle to be forgotten if the
2459 handle were not opened yet. This has been fixed.
2460 .IP \(bu 4
2461 Various cases of \f(CW\*(C`delete $::{a}\*(C'\fR, \f(CW\*(C`delete $::{ENV}\*(C'\fR etc. causing a crash
2462 have been fixed. [perl #54044]
2463 .IP \(bu 4
2464 Setting \f(CW$!\fR to EACCESS before calling \f(CW\*(C`require\*(C'\fR could affect
2465 \&\f(CW\*(C`require\*(C'\fR's behaviour. This has been fixed.
2466 .IP \(bu 4
2467 The "Can't use \e1 to mean \f(CW$1\fR in expression" warning message now only occurs
2468 on the right-hand (replacement) part of a substitution. Formerly it could
2469 happen in code embedded in the left-hand side, or in any other quote-like
2470 operator.
2471 .IP \(bu 4
2472 Blessing into a reference (\f(CW\*(C`bless $thisref, $thatref\*(C'\fR) has long been
2473 disallowed, but magical scalars for the second like \f(CW$/\fR and those tied
2474 were exempt. They no longer are. [perl #119809]
2475 .IP \(bu 4
2476 Blessing into a reference was accidentally allowed in 5.18 if the class
2477 argument were a blessed reference with stale method caches (i.e., whose
2478 class had had subs defined since the last method call). They are
2479 disallowed once more, as in 5.16.
2480 .IP \(bu 4
2481 \&\f(CW\*(C`$x\->{key}\*(C'\fR where \f(CW$x\fR was declared as \f(CW\*(C`my Class $x\*(C'\fR no longer crashes
2482 if a Class::FIELDS subroutine stub has been declared.
2483 .IP \(bu 4
2484 \&\f(CW@$obj{\*(Aqkey\*(Aq}\fR and \f(CW\*(C`${$obj}{key}\*(C'\fR used to be exempt from compile-time
2485 field checking ("No such class field"; see fields) but no longer are.
2486 .IP \(bu 4
2487 A nonexistent array element with a large index passed to a subroutine that
2488 ties the array and then tries to access the element no longer results in a
2489 crash.
2490 .IP \(bu 4
2491 Declaring a subroutine stub named NEGATIVE_INDICES no longer makes negative
2492 array indices crash when the current package is a tied array class.
2493 .IP \(bu 4
2494 Declaring a \f(CW\*(C`require\*(C'\fR, \f(CW\*(C`glob\*(C'\fR, or \f(CW\*(C`do\*(C'\fR subroutine stub in the
2495 CORE::GLOBAL:: package no longer makes compilation of calls to the
2496 corresponding functions crash.
2497 .IP \(bu 4
2498 Aliasing CORE::GLOBAL:: functions to constants stopped working in Perl 5.10
2499 but has now been fixed.
2500 .IP \(bu 4
2501 When \f(CW\`...\`\fR or \f(CW\*(C`qx/.../\*(C'\fR calls a \f(CW\*(C`readpipe\*(C'\fR override, double-quotish
2502 interpolation now happens, as is the case when there is no override.
2503 Previously, the presence of an override would make these quote-like
2504 operators act like \f(CW\*(C`q{}\*(C'\fR, suppressing interpolation. [perl #115330]
2505 .IP \(bu 4
2506 \&\f(CW\*(C`<<<\`...\`\*(C'\fR here-docs (with backticks as the delimiters) now call
2507 \&\f(CW\*(C`readpipe\*(C'\fR overrides. [perl #119827]
2508 .IP \(bu 4
2509 \&\f(CW&CORE::exit()\fR and \f(CW&CORE::die()\fR now respect vmsish hints.
2510 .IP \(bu 4
2511 Undefining a glob that triggers a DESTROY method that undefines the same
2512 glob is now safe. It used to produce "Attempt to free unreferenced glob
2513 pointer" warnings and leak memory.
2514 .IP \(bu 4
2515 If subroutine redefinition (\f(CW\*(C`eval \*(Aqsub foo{}\*(Aq\*(C'\fR or \f(CW\*(C`newXS\*(C'\fR for XS code)
2516 triggers a DESTROY method on the sub that is being redefined, and that
2517 method assigns a subroutine to the same slot (\f(CW\*(C`*foo = sub {}\*(C'\fR), \f(CW$_[0]\fR
2518 is no longer left pointing to a freed scalar. Now DESTROY is delayed until
2519 the new subroutine has been installed.
2520 .IP \(bu 4
2521 On Windows, perl no longer calls \fBCloseHandle()\fR on a socket handle. This makes
2522 debugging easier on Windows by removing certain irrelevant bad handle
2523 exceptions. It also fixes a race condition that made socket functions randomly
2524 fail in a Perl process with multiple OS threads, and possible test failures in
2525 \&\fIdist/IO/t/cachepropagate\-tcp.t\fR. [perl #120091/118059]
2526 .IP \(bu 4
2527 Formats involving UTF\-8 encoded strings, or strange vars like ties,
2528 overloads, or stringified refs (and in recent
2529 perls, pure NOK vars) would generally do the wrong thing in formats
2530 when the var is treated as a string and repeatedly chopped, as in
2531 \&\f(CW\*(C`^<<<~~\*(C'\fR and similar. This has now been resolved.
2532 [perl #33832/45325/113868/119847/119849/119851]
2533 .IP \(bu 4
2534 \&\f(CW\*(C`semctl(..., SETVAL, ...)\*(C'\fR would set the semaphore to the top
2535 32\-bits of the supplied integer instead of the bottom 32\-bits on
2536 64\-bit big-endian systems. [perl #120635]
2537 .IP \(bu 4
2538 \&\f(CWreaddir()\fR now only sets \f(CW$!\fR on error. \f(CW$!\fR is no longer set
2539 to \f(CW\*(C`EBADF\*(C'\fR when then terminating \f(CW\*(C`undef\*(C'\fR is read from the directory
2540 unless the system call sets \f(CW$!\fR. [perl #118651]
2541 .IP \(bu 4
2542 \&\f(CW&CORE::glob\fR no longer causes an intermittent crash due to perl's stack
2543 getting corrupted. [perl #119993]
2544 .IP \(bu 4
2545 \&\f(CW\*(C`open\*(C'\fR with layers that load modules (e.g., "<:encoding(utf8)") no longer
2546 runs the risk of crashing due to stack corruption.
2547 .IP \(bu 4
2548 Perl 5.18 broke autoloading via \f(CW\*(C`\->SUPER::foo\*(C'\fR method calls by looking
2549 up AUTOLOAD from the current package rather than the current package's
2550 superclass. This has been fixed. [perl #120694]
2551 .IP \(bu 4
2552 A longstanding bug causing \f(CW\*(C`do {} until CONSTANT\*(C'\fR, where the constant
2553 holds a true value, to read unallocated memory has been resolved. This
2554 would usually happen after a syntax error. In past versions of Perl it has
2555 crashed intermittently. [perl #72406]
2556 .IP \(bu 4
2557 Fix HP-UX \f(CW$!\fR failure. HP-UX \fBstrerror()\fR returns an empty string for an
2558 unknown error code. This caused an assertion to fail under DEBUGGING
2559 builds. Now instead, the returned string for \f(CW"$!"\fR contains text
2560 indicating the code is for an unknown error.
2561 .IP \(bu 4
2562 Individually-tied elements of \f(CW@INC\fR (as in \f(CW\*(C`tie $INC[0]...\*(C'\fR) are now
2563 handled correctly. Formerly, whether a sub returned by such a tied element
2564 would be treated as a sub depended on whether a FETCH had occurred
2565 previously.
2566 .IP \(bu 4
2567 \&\f(CW\*(C`getc\*(C'\fR on a byte-sized handle after the same \f(CW\*(C`getc\*(C'\fR operator had been
2568 used on a utf8 handle used to treat the bytes as utf8, resulting in erratic
2569 behavior (e.g., malformed UTF\-8 warnings).
2570 .IP \(bu 4
2571 An initial \f(CW\*(C`{\*(C'\fR at the beginning of a format argument line was always
2572 interpreted as the beginning of a block prior to v5.18. In Perl v5.18, it
2573 started being treated as an ambiguous token. The parser would guess
2574 whether it was supposed to be an anonymous hash constructor or a block
2575 based on the contents. Now the previous behaviour has been restored.
2576 [perl #119973]
2577 .IP \(bu 4
2578 In Perl v5.18 \f(CW\*(C`undef *_; goto &sub\*(C'\fR and \f(CW\*(C`local *_; goto &sub\*(C'\fR started
2579 crashing. This has been fixed. [perl #119949]
2580 .IP \(bu 4
2581 Backticks (\f(CW \`\` \fR or \f(CW\*(C` qx// \*(C'\fR) combined with multiple threads on
2582 Win32 could result in output sent to stdout on one thread being
2583 captured by backticks of an external command in another thread.
2584 .Sp
2585 This could occur for pseudo-forked processes too, as Win32's
2586 pseudo-fork is implemented in terms of threads. [perl #77672]
2587 .IP \(bu 4
2588 \&\f(CW\*(C`open $fh, ">+", undef\*(C'\fR no longer leaks memory when TMPDIR is set
2589 but points to a directory a temporary file cannot be created in. [perl
2590 #120951]
2591 .IP \(bu 4
2592 \&\f(CW\*(C` for ( $h{k} || \*(Aq\*(Aq ) \*(C'\fR no longer auto-vivifies \f(CW$h{k}\fR. [perl
2593 #120374]
2594 .IP \(bu 4
2595 On Windows machines, Perl now emulates the POSIX use of the environment
2596 for locale initialization. Previously, the environment was ignored.
2597 See "ENVIRONMENT" in perllocale.
2598 .IP \(bu 4
2599 Fixed a crash when destroying a self-referencing GLOB. [perl #121242]
2600 .SH "Known Problems"
2601 .IX Header "Known Problems"
2602 .IP \(bu 4
2603 IO::Socket is known to fail tests on AIX 5.3. There is
2604 a patch <https://github.com/Perl/perl5/issues/13484> in the request
2605 tracker, #120835, which may be applied to future releases.
2606 .IP \(bu 4
2607 The following modules are known to have test failures with this version of
2608 Perl. Patches have been submitted, so there will hopefully be new releases
2609 soon:
2610 .RS 4
2611 .IP \(bu 4
2612 Data::Structure::Util version 0.15
2613 .IP \(bu 4
2614 HTML::StripScripts version 1.05
2615 .IP \(bu 4
2616 List::Gather version 0.08.
2617 .RE
2618 .RS 4
2619 .RE
2620 .SH Obituary
2621 .IX Header "Obituary"
2622 Diana Rosa, 27, of Rio de Janeiro, went to her long rest on May 10,
2623 2014, along with the plush camel she kept hanging on her computer screen
2624 all the time. She was a passionate Perl hacker who loved the language and its
2625 community, and who never missed a Rio.pm event. She was a true artist, an
2626 enthusiast about writing code, singing arias and graffiting walls. We'll never
2627 forget you.
2628 .PP
2629 Greg McCarroll died on August 28, 2013.
2630 .PP
2631 Greg was well known for many good reasons. He was one of the organisers of
2632 the first YAPC::Europe, which concluded with an unscheduled auction where he
2633 frantically tried to raise extra money to avoid the conference making a
2634 loss. It was Greg who mistakenly arrived for a london.pm meeting a week
2635 late; some years later he was the one who sold the choice of official
2636 meeting date at a YAPC::Europe auction, and eventually as glorious leader of
2637 london.pm he got to inherit the irreverent confusion that he had created.
2638 .PP
2639 Always helpful, friendly and cheerfully optimistic, you will be missed, but
2640 never forgotten.
2641 .SH Acknowledgements
2642 .IX Header "Acknowledgements"
2643 Perl 5.20.0 represents approximately 12 months of development since Perl 5.18.0
2644 and contains approximately 470,000 lines of changes across 2,900 files from 124
2645 authors.
2646 .PP
2647 Excluding auto-generated files, documentation and release tools, there were
2648 approximately 280,000 lines of changes to 1,800 .pm, .t, .c and .h files.
2649 .PP
2650 Perl continues to flourish into its third decade thanks to a vibrant community
2651 of users and developers. The following people are known to have contributed the
2652 improvements that became Perl 5.20.0:
2653 .PP
2654 Aaron Crane, Abhijit Menon-Sen, Abigail, Abir Viqar, Alan Haggai Alavi, Alan
2655 Hourihane, Alexander Voronov, Alexandr Ciornii, Andy Dougherty, Anno Siegel,
2656 Aristotle Pagaltzis, Arthur Axel 'fREW' Schmidt, Brad Gilbert, Brendan Byrd,
2657 Brian Childs, Brian Fraser, Brian Gottreu, Chris 'BinGOs' Williams, Christian
2658 Millour, Colin Kuskie, Craig A. Berry, Dabrien 'Dabe' Murphy, Dagfinn Ilmari
2659 Mannsåker, Daniel Dragan, Darin McBride, David Golden, David Leadbeater, David
2660 Mitchell, David Nicol, David Steinbrunner, Dennis Kaarsemaker, Dominic
2661 Hargreaves, Ed Avis, Eric Brine, Evan Zacks, Father Chrysostomos, Florian
2662 Ragwitz, François Perrad, Gavin Shelley, Gideon Israel Dsouza, Gisle Aas,
2663 Graham Knop, H.Merijn Brand, Hauke D, Heiko Eissfeldt, Hiroo Hayashi, Hojung
2664 Youn, James E Keenan, Jarkko Hietaniemi, Jerry D. Hedden, Jess Robinson, Jesse
2665 Luehrs, Johan Vromans, John Gardiner Myers, John Goodyear, John P. Linderman,
2666 John Peacock, kafka, Kang-min Liu, Karen Etheridge, Karl Williamson, Keedi Kim,
2667 Kent Fredric, kevin dawson, Kevin Falcone, Kevin Ryde, Leon Timmermans, Lukas
2668 Mai, Marc Simpson, Marcel Grünauer, Marco Peereboom, Marcus Holland-Moritz,
2669 Mark Jason Dominus, Martin McGrath, Matthew Horsfall, Max Maischein, Mike
2670 Doherty, Moritz Lenz, Nathan Glenn, Nathan Trapuzzano, Neil Bowers, Neil
2671 Williams, Nicholas Clark, Niels Thykier, Niko Tyni, Olivier Mengué, Owain G.
2672 Ainsworth, Paul Green, Paul Johnson, Peter John Acklam, Peter Martini, Peter
2673 Rabbitson, Petr Písař, Philip Boulain, Philip Guenther, Piotr Roszatycki,
2674 Rafael Garcia-Suarez, Reini Urban, Reuben Thomas, Ricardo Signes, Ruslan
2675 Zakirov, Sergey Alekseev, Shirakata Kentaro, Shlomi Fish, Slaven Rezic,
2676 Smylers, Steffen Müller, Steve Hay, Sullivan Beck, Thomas Sibley, Tobias
2677 Leich, Toby Inkster, Tokuhiro Matsuno, Tom Christiansen, Tom Hukins, Tony Cook,
2678 Victor Efimov, Viktor Turskyi, Vladimir Timofeev, YAMASHINA Hio, Yves Orton,
2679 Zefram, Zsbán Ambrus, Ævar Arnfjörð Bjarmason.
2680 .PP
2681 The list above is almost certainly incomplete as it is automatically generated
2682 from version control history. In particular, it does not include the names of
2683 the (very much appreciated) contributors who reported issues to the Perl bug
2684 tracker.
2685 .PP
2686 Many of the changes included in this version originated in the CPAN modules
2687 included in Perl's core. We're grateful to the entire CPAN community for
2688 helping Perl to flourish.
2689 .PP
2690 For a more complete list of all of Perl's historical contributors, please see
2691 the \fIAUTHORS\fR file in the Perl source distribution.
2692 .SH "Reporting Bugs"
2693 .IX Header "Reporting Bugs"
2694 If you find what you think is a bug, you might check the articles recently
2695 posted to the comp.lang.perl.misc newsgroup and the perl bug database at
2696 http://rt.perl.org/perlbug/ . There may also be information at
2697 http://www.perl.org/ , the Perl Home Page.
2698 .PP
2699 If you believe you have an unreported bug, please run the perlbug program
2700 included with your release. Be sure to trim your bug down to a tiny but
2701 sufficient test case. Your bug report, along with the output of \f(CW\*(C`perl \-V\*(C'\fR,
2702 will be sent off to [email protected] to be analysed by the Perl porting team.
2703 .PP
2704 If the bug you are reporting has security implications, which make it
2705 inappropriate to send to a publicly archived mailing list, then please send it
2706 to perl5\-security\[email protected]. This points to a closed subscription
2707 unarchived mailing list, which includes all the core committers, who will be
2708 able to help assess the impact of issues, figure out a resolution, and help
2709 co-ordinate the release of patches to mitigate or fix the problem across all
2710 platforms on which Perl is supported. Please only use this address for
2711 security issues in the Perl core, not for modules independently distributed on
2712 CPAN.
2713 .SH "SEE ALSO"
2714 .IX Header "SEE ALSO"
2715 The \fIChanges\fR file for an explanation of how to view exhaustive details on
2716 what changed.
2717 .PP
2718 The \fIINSTALL\fR file for how to build Perl.
2719 .PP
2720 The \fIREADME\fR file for general stuff.
2721 .PP
2722 The \fIArtistic\fR and \fICopying\fR files for copyright information.