Debian Codegraph @

upstream/debian-unstable/man1/perl5260delta.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 "PERL5260DELTA 1"
58 .TH PERL5260DELTA 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 perl5260delta \- what is new for perl v5.26.0
65 .SH DESCRIPTION
66 .IX Header "DESCRIPTION"
67 This document describes the differences between the 5.24.0 release and the
68 5.26.0 release.
69 .SH Notice
70 .IX Header "Notice"
71 This release includes three updates with widespread effects:
72 .IP \(bu 4
73 \&\f(CW"."\fR no longer in \f(CW@INC\fR
74 .Sp
75 For security reasons, the current directory (\f(CW"."\fR) is no longer included
76 by default at the end of the module search path (\f(CW@INC\fR). This may have
77 widespread implications for the building, testing and installing of
78 modules, and for the execution of scripts. See the section
79 "Removal of the current directory (\f(CW"."\fR) from \f(CW@INC\fR"
80 for the full details.
81 .IP \(bu 4
82 \&\f(CW\*(C`do\*(C'\fR may now warn
83 .Sp
84 \&\f(CW\*(C`do\*(C'\fR now gives a deprecation warning when it fails to load a file which
85 it would have loaded had \f(CW"."\fR been in \f(CW@INC\fR.
86 .IP \(bu 4
87 In regular expression patterns, a literal left brace \f(CW"{"\fR
88 should be escaped
89 .Sp
90 See "Unescaped literal \f(CW"{"\fR characters in regular expression patterns are no longer permissible".
91 .SH "Core Enhancements"
92 .IX Header "Core Enhancements"
93 .SS "Lexical subroutines are no longer experimental"
94 .IX Subsection "Lexical subroutines are no longer experimental"
95 Using the \f(CW\*(C`lexical_subs\*(C'\fR feature introduced in v5.18 no longer emits a warning. Existing
96 code that disables the \f(CW\*(C`experimental::lexical_subs\*(C'\fR warning category
97 that the feature previously used will continue to work. The
98 \&\f(CW\*(C`lexical_subs\*(C'\fR feature has no effect; all Perl code can use lexical
99 subroutines, regardless of what feature declarations are in scope.
100 .SS "Indented Here-documents"
101 .IX Subsection "Indented Here-documents"
102 This adds a new modifier \f(CW"~"\fR to here-docs that tells the parser
103 that it should look for \f(CW\*(C`/^\es*$DELIM\en/\*(C'\fR as the closing delimiter.
104 .PP
105 These syntaxes are all supported:
106 .PP
107 .Vb 8
108 \& <<~EOF;
109 \& <<~\eEOF;
110 \& <<~\*(AqEOF\*(Aq;
111 \& <<~"EOF";
112 \& <<~\`EOF\`;
113 \& <<~ \*(AqEOF\*(Aq;
114 \& <<~ "EOF";
115 \& <<~ \`EOF\`;
116 .Ve
117 .PP
118 The \f(CW"~"\fR modifier will strip, from each line in the here-doc, the
119 same whitespace that appears before the delimiter.
120 .PP
121 Newlines will be copied as-is, and lines that don't include the
122 proper beginning whitespace will cause perl to croak.
123 .PP
124 For example:
125 .PP
126 .Vb 5
127 \& if (1) {
128 \& print <<~EOF;
129 \& Hello there
130 \& EOF
131 \& }
132 .Ve
133 .PP
134 prints "Hello there\en" with no leading whitespace.
135 .ie n .SS "New regular expression modifier ""/xx"""
136 .el .SS "New regular expression modifier \f(CW/xx\fP"
137 .IX Subsection "New regular expression modifier /xx"
138 Specifying two \f(CW"x"\fR characters to modify a regular expression pattern
139 does everything that a single one does, but additionally TAB and SPACE
140 characters within a bracketed character class are generally ignored and
141 can be added to improve readability, like
142 \&\f(CW\*(C`/[\ ^\ A\-Z\ d\-f\ p\-x\ ]/xx\*(C'\fR. Details are at
143 "/x and /xx" in perlre.
144 .ie n .SS """@{^CAPTURE}"", ""%{^CAPTURE}"", and ""%{^CAPTURE_ALL}"""
145 .el .SS "\f(CW@{^CAPTURE}\fP, \f(CW%{^CAPTURE}\fP, and \f(CW%{^CAPTURE_ALL}\fP"
146 .IX Subsection "@{^CAPTURE}, %{^CAPTURE}, and %{^CAPTURE_ALL}"
147 \&\f(CW\*(C`@{^CAPTURE}\*(C'\fR exposes the capture buffers of the last match as an
148 array. So \f(CW$1\fR is \f(CW\*(C`${^CAPTURE}[0]\*(C'\fR. This is a more efficient equivalent
149 to code like \f(CW\*(C`substr($matched_string,$\-[0],$+[0]\-$\-[0])\*(C'\fR, and you don't
150 have to keep track of the \f(CW$matched_string\fR either. This variable has no
151 single character equivalent. Note that, like the other regex magic variables,
152 the contents of this variable is dynamic; if you wish to store it beyond
153 the lifetime of the match you must copy it to another array.
154 .PP
155 \&\f(CW\*(C`%{^CAPTURE}\*(C'\fR is equivalent to \f(CW\*(C`%+\*(C'\fR (\fIi.e.\fR, named captures). Other than
156 being more self-documenting there is no difference between the two forms.
157 .PP
158 \&\f(CW\*(C`%{^CAPTURE_ALL}\*(C'\fR is equivalent to \f(CW\*(C`%\-\*(C'\fR (\fIi.e.\fR, all named captures).
159 Other than being more self-documenting there is no difference between the
160 two forms.
161 .SS "Declaring a reference to a variable"
162 .IX Subsection "Declaring a reference to a variable"
163 As an experimental feature, Perl now allows the referencing operator to come
164 after \f(CWmy()\fR, \f(CWstate()\fR,
165 \&\f(CWour()\fR, or \f(CWlocal()\fR. This syntax must
166 be enabled with \f(CW\*(C`use feature \*(Aqdeclared_refs\*(Aq\*(C'\fR. It is experimental, and will
167 warn by default unless \f(CW\*(C`no warnings \*(Aqexperimental::refaliasing\*(Aq\*(C'\fR is in effect.
168 It is intended mainly for use in assignments to references. For example:
169 .PP
170 .Vb 2
171 \& use experimental \*(Aqrefaliasing\*(Aq, \*(Aqdeclared_refs\*(Aq;
172 \& my \e$a = \e$b;
173 .Ve
174 .PP
175 See "Assigning to References" in perlref for more details.
176 .SS "Unicode 9.0 is now supported"
177 .IX Subsection "Unicode 9.0 is now supported"
178 A list of changes is at <http://www.unicode.org/versions/Unicode9.0.0/>.
179 Modules that are shipped with core Perl but not maintained by p5p do not
180 necessarily support Unicode 9.0. Unicode::Normalize does work on 9.0.
181 .ie n .SS "Use of ""\ep{\fIscript\fP}"" uses the improved Script_Extensions property"
182 .el .SS "Use of \f(CW\ep{\fP\f(CIscript\fP\f(CW}\fP uses the improved Script_Extensions property"
183 .IX Subsection "Use of p{script} uses the improved Script_Extensions property"
184 Unicode 6.0 introduced an improved form of the Script (\f(CW\*(C`sc\*(C'\fR) property, and
185 called it Script_Extensions (\f(CW\*(C`scx\*(C'\fR). Perl now uses this improved
186 version when a property is specified as just \f(CW\*(C`\ep{\fR\f(CIscript\fR\f(CW}\*(C'\fR. This
187 should make programs more accurate when determining if a character is
188 used in a given script, but there is a slight chance of breakage for
189 programs that very specifically needed the old behavior. The meaning of
190 compound forms, like \f(CW\*(C`\ep{sc=\fR\f(CIscript\fR\f(CW}\*(C'\fR are unchanged. See
191 "Scripts" in perlunicode.
192 .SS "Perl can now do default collation in UTF\-8 locales on platforms that support it"
193 .IX Subsection "Perl can now do default collation in UTF-8 locales on platforms that support it"
194 Some platforms natively do a reasonable job of collating and sorting in
195 UTF\-8 locales. Perl now works with those. For portability and full
196 control, Unicode::Collate is still recommended, but now you may
197 not need to do anything special to get good-enough results, depending on
198 your application. See
199 "Category \f(CW\*(C`LC_COLLATE\*(C'\fR: Collation: Text Comparisons and Sorting" in perllocale.
200 .ie n .SS "Better locale collation of strings containing embedded ""NUL"" characters"
201 .el .SS "Better locale collation of strings containing embedded \f(CWNUL\fP characters"
202 .IX Subsection "Better locale collation of strings containing embedded NUL characters"
203 In locales that have multi-level character weights, \f(CW\*(C`NUL\*(C'\fRs are now
204 ignored at the higher priority ones. There are still some gotchas in
205 some strings, though. See
206 "Collation of strings containing embedded \f(CW\*(C`NUL\*(C'\fR characters" in perllocale.
207 .ie n .SS """CORE"" subroutines for hash and array functions callable via reference"
208 .el .SS "\f(CWCORE\fP subroutines for hash and array functions callable via reference"
209 .IX Subsection "CORE subroutines for hash and array functions callable via reference"
210 The hash and array functions in the \f(CW\*(C`CORE\*(C'\fR namespace (\f(CW\*(C`keys\*(C'\fR, \f(CW\*(C`each\*(C'\fR,
211 \&\f(CW\*(C`values\*(C'\fR, \f(CW\*(C`push\*(C'\fR, \f(CW\*(C`pop\*(C'\fR, \f(CW\*(C`shift\*(C'\fR, \f(CW\*(C`unshift\*(C'\fR and \f(CW\*(C`splice\*(C'\fR) can now
212 be called with ampersand syntax (\f(CW\*(C`&CORE::keys(\e%hash\*(C'\fR) and via reference
213 (\f(CW\*(C`my $k = \e&CORE::keys; $k\->(\e%hash)\*(C'\fR). Previously they could only be
214 used when inlined.
215 .SS "New Hash Function For 64\-bit Builds"
216 .IX Subsection "New Hash Function For 64-bit Builds"
217 We have switched to a hybrid hash function to better balance
218 performance for short and long keys.
219 .PP
220 For short keys, 16 bytes and under, we use an optimised variant of
221 One At A Time Hard, and for longer keys we use Siphash 1\-3. For very
222 long keys this is a big improvement in performance. For shorter keys
223 there is a modest improvement.
224 .SH Security
225 .IX Header "Security"
226 .ie n .SS "Removal of the current directory (""."") from @INC"
227 .el .SS "Removal of the current directory (\f(CW"".""\fP) from \f(CW@INC\fP"
228 .IX Subsection "Removal of the current directory (""."") from @INC"
229 The perl binary includes a default set of paths in \f(CW@INC\fR. Historically
230 it has also included the current directory (\f(CW"."\fR) as the final entry,
231 unless run with taint mode enabled (\f(CW\*(C`perl \-T\*(C'\fR). While convenient, this has
232 security implications: for example, where a script attempts to load an
233 optional module when its current directory is untrusted (such as \fI/tmp\fR),
234 it could load and execute code from under that directory.
235 .PP
236 Starting with v5.26, \f(CW"."\fR is always removed by default, not just under
237 tainting. This has major implications for installing modules and executing
238 scripts.
239 .PP
240 The following new features have been added to help ameliorate these
241 issues.
242 .IP \(bu 4
243 \&\fIConfigure \-Udefault_inc_excludes_dot\fR
244 .Sp
245 There is a new \fIConfigure\fR option, \f(CW\*(C`default_inc_excludes_dot\*(C'\fR (enabled
246 by default) which builds a perl executable without \f(CW"."\fR; unsetting this
247 option using \f(CW\*(C`\-U\*(C'\fR reverts perl to the old behaviour. This may fix your
248 path issues but will reintroduce all the security concerns, so don't
249 build a perl executable like this unless you're \fIreally\fR confident that
250 such issues are not a concern in your environment.
251 .IP \(bu 4
252 \&\f(CW\*(C`PERL_USE_UNSAFE_INC\*(C'\fR
253 .Sp
254 There is a new environment variable recognised by the perl interpreter.
255 If this variable has the value 1 when the perl interpreter starts up,
256 then \f(CW"."\fR will be automatically appended to \f(CW@INC\fR (except under tainting).
257 .Sp
258 This allows you restore the old perl interpreter behaviour on a
259 case-by-case basis. But note that this is intended to be a temporary crutch,
260 and this feature will likely be removed in some future perl version.
261 It is currently set by the \f(CW\*(C`cpan\*(C'\fR utility and \f(CW\*(C`Test::Harness\*(C'\fR to
262 ease installation of CPAN modules which have not been updated to handle the
263 lack of dot. Once again, don't use this unless you are sure that this
264 will not reintroduce any security concerns.
265 .IP \(bu 4
266 A new deprecation warning issued by \f(CW\*(C`do\*(C'\fR.
267 .Sp
268 While it is well-known that \f(CW\*(C`use\*(C'\fR and \f(CW\*(C`require\*(C'\fR use \f(CW@INC\fR to search
269 for the file to load, many people don't realise that \f(CW\*(C`do "file"\*(C'\fR also
270 searches \f(CW@INC\fR if the file is a relative path. With the removal of \f(CW"."\fR,
271 a simple \f(CW\*(C`do "file.pl"\*(C'\fR will fail to read in and execute \f(CW\*(C`file.pl\*(C'\fR from
272 the current directory. Since this is commonly expected behaviour, a new
273 deprecation warning is now issued whenever \f(CW\*(C`do\*(C'\fR fails to load a file which
274 it otherwise would have found if a dot had been in \f(CW@INC\fR.
275 .PP
276 Here are some things script and module authors may need to do to make
277 their software work in the new regime.
278 .IP \(bu 4
279 Script authors
280 .Sp
281 If the issue is within your own code (rather than within included
282 modules), then you have two main options. Firstly, if you are confident
283 that your script will only be run within a trusted directory (under which
284 you expect to find trusted files and modules), then add \f(CW"."\fR back into the
285 path; \fIe.g.\fR:
286 .Sp
287 .Vb 6
288 \& BEGIN {
289 \& my $dir = "/some/trusted/directory";
290 \& chdir $dir or die "Can\*(Aqt chdir to $dir: $!\en";
291 \& # safe now
292 \& push @INC, \*(Aq.\*(Aq;
293 \& }
294 \&
295 \& use "Foo::Bar"; # may load /some/trusted/directory/Foo/Bar.pm
296 \& do "config.pl"; # may load /some/trusted/directory/config.pl
297 .Ve
298 .Sp
299 On the other hand, if your script is intended to be run from within
300 untrusted directories (such as \fI/tmp\fR), then your script suddenly failing
301 to load files may be indicative of a security issue. You most likely want
302 to replace any relative paths with full paths; for example,
303 .Sp
304 .Vb 1
305 \& do "foo_config.pl"
306 .Ve
307 .Sp
308 might become
309 .Sp
310 .Vb 1
311 \& do "$ENV{HOME}/foo_config.pl"
312 .Ve
313 .Sp
314 If you are absolutely certain that you want your script to load and
315 execute a file from the current directory, then use a \f(CW\*(C`./\*(C'\fR prefix; for
316 example:
317 .Sp
318 .Vb 1
319 \& do "./foo_config.pl"
320 .Ve
321 .IP \(bu 4
322 Installing and using CPAN modules
323 .Sp
324 If you install a CPAN module using an automatic tool like \f(CW\*(C`cpan\*(C'\fR, then
325 this tool will itself set the \f(CW\*(C`PERL_USE_UNSAFE_INC\*(C'\fR environment variable
326 while building and testing the module, which may be sufficient to install
327 a distribution which hasn't been updated to be dot-aware. If you want to
328 install such a module manually, then you'll need to replace the
329 traditional invocation:
330 .Sp
331 .Vb 1
332 \& perl Makefile.PL && make && make test && make install
333 .Ve
334 .Sp
335 with something like
336 .Sp
337 .Vb 2
338 \& (export PERL_USE_UNSAFE_INC=1; \e
339 \& perl Makefile.PL && make && make test && make install)
340 .Ve
341 .Sp
342 Note that this only helps build and install an unfixed module. It's
343 possible for the tests to pass (since they were run under
344 \&\f(CW\*(C`PERL_USE_UNSAFE_INC=1\*(C'\fR), but for the module itself to fail to perform
345 correctly in production. In this case, you may have to temporarily modify
346 your script until a fixed version of the module is released.
347 For example:
348 .Sp
349 .Vb 6
350 \& use Foo::Bar;
351 \& {
352 \& local @INC = (@INC, \*(Aq.\*(Aq);
353 \& # assuming read_config() needs \*(Aq.\*(Aq in @INC
354 \& $config = Foo::Bar\->read_config();
355 \& }
356 .Ve
357 .Sp
358 This is only rarely expected to be necessary. Again, if doing this,
359 assess the resultant risks first.
360 .IP \(bu 4
361 Module Authors
362 .Sp
363 If you maintain a CPAN distribution, it may need updating to run in
364 a dotless environment. Although \f(CW\*(C`cpan\*(C'\fR and other such tools will
365 currently set the \f(CW\*(C`PERL_USE_UNSAFE_INC\*(C'\fR during module build, this is a
366 temporary workaround for the set of modules which rely on \f(CW"."\fR being in
367 \&\f(CW@INC\fR for installation and testing, and this may mask deeper issues. It
368 could result in a module which passes tests and installs, but which
369 fails at run time.
370 .Sp
371 During build, test, and install, it will normally be the case that any perl
372 processes will be executing directly within the root directory of the
373 untarred distribution, or a known subdirectory of that, such as \fIt/\fR. It
374 may well be that \fIMakefile.PL\fR or \fIt/foo.t\fR will attempt to include
375 local modules and configuration files using their direct relative
376 filenames, which will now fail.
377 .Sp
378 However, as described above, automatic tools like \fIcpan\fR will (for now)
379 set the \f(CW\*(C`PERL_USE_UNSAFE_INC\*(C'\fR environment variable, which introduces
380 dot during a build.
381 .Sp
382 This makes it likely that your existing build and test code will work, but
383 this may mask issues with your code which only manifest when used after
384 install. It is prudent to try and run your build process with that
385 variable explicitly disabled:
386 .Sp
387 .Vb 2
388 \& (export PERL_USE_UNSAFE_INC=0; \e
389 \& perl Makefile.PL && make && make test && make install)
390 .Ve
391 .Sp
392 This is more likely to show up any potential problems with your module's
393 build process, or even with the module itself. Fixing such issues will
394 ensure both that your module can again be installed manually, and that
395 it will still build once the \f(CW\*(C`PERL_USE_UNSAFE_INC\*(C'\fR crutch goes away.
396 .Sp
397 When fixing issues in tests due to the removal of dot from \f(CW@INC\fR,
398 reinsertion of dot into \f(CW@INC\fR should be performed with caution, for this
399 too may suppress real errors in your runtime code. You are encouraged
400 wherever possible to apply the aforementioned approaches with explicit
401 absolute/relative paths, or to relocate your needed files into a
402 subdirectory and insert that subdirectory into \f(CW@INC\fR instead.
403 .Sp
404 If your runtime code has problems under the dotless \f(CW@INC\fR, then the comments
405 above on how to fix for script authors will mostly apply here too. Bear in
406 mind though that it is considered bad form for a module to globally add a dot to
407 \&\f(CW@INC\fR, since it introduces both a security risk and hides issues of
408 accidentally requiring dot in \f(CW@INC\fR, as explained above.
409 .SS "Escaped colons and relative paths in PATH"
410 .IX Subsection "Escaped colons and relative paths in PATH"
411 On Unix systems, Perl treats any relative paths in the \f(CW\*(C`PATH\*(C'\fR environment
412 variable as tainted when starting a new process. Previously, it was
413 allowing a backslash to escape a colon (unlike the OS), consequently
414 allowing relative paths to be considered safe if the PATH was set to
415 something like \f(CW\*(C`/\e:.\*(C'\fR. The check has been fixed to treat \f(CW"."\fR as tainted
416 in that example.
417 .ie n .SS "New ""\-Di"" switch is now required for PerlIO debugging output"
418 .el .SS "New \f(CW\-Di\fP switch is now required for PerlIO debugging output"
419 .IX Subsection "New -Di switch is now required for PerlIO debugging output"
420 This is used for debugging of code within PerlIO to avoid recursive
421 calls. Previously this output would be sent to the file specified
422 by the \f(CW\*(C`PERLIO_DEBUG\*(C'\fR environment variable if perl wasn't running
423 setuid and the \f(CW\*(C`\-T\*(C'\fR or \f(CW\*(C`\-t\*(C'\fR switches hadn't been parsed yet.
424 .PP
425 If perl performed output at a point where it hadn't yet parsed its
426 switches this could result in perl creating or overwriting the file
427 named by \f(CW\*(C`PERLIO_DEBUG\*(C'\fR even when the \f(CW\*(C`\-T\*(C'\fR switch had been supplied.
428 .PP
429 Perl now requires the \f(CW\*(C`\-Di\*(C'\fR switch to be present before it will produce
430 PerlIO debugging
431 output. By default this is written to \f(CW\*(C`stderr\*(C'\fR, but can optionally
432 be redirected to a file by setting the \f(CW\*(C`PERLIO_DEBUG\*(C'\fR environment
433 variable.
434 .PP
435 If perl is running setuid or the \f(CW\*(C`\-T\*(C'\fR switch was supplied,
436 \&\f(CW\*(C`PERLIO_DEBUG\*(C'\fR is ignored and the debugging output is sent to
437 \&\f(CW\*(C`stderr\*(C'\fR as for any other \f(CW\*(C`\-D\*(C'\fR switch.
438 .SH "Incompatible Changes"
439 .IX Header "Incompatible Changes"
440 .ie n .SS "Unescaped literal ""{"" characters in regular expression patterns are no longer permissible"
441 .el .SS "Unescaped literal \f(CW""{""\fP characters in regular expression patterns are no longer permissible"
442 .IX Subsection "Unescaped literal ""{"" characters in regular expression patterns are no longer permissible"
443 You have to now say something like \f(CW"\e{"\fR or \f(CW"[{]"\fR to specify to
444 match a LEFT CURLY BRACKET; otherwise, it is a fatal pattern compilation
445 error. This change will allow future extensions to the language.
446 .PP
447 These have been deprecated since v5.16, with a deprecation message
448 raised for some uses starting in v5.22. Unfortunately, the code added
449 to raise the message was buggy and failed to warn in some cases where
450 it should have. Therefore, enforcement of this ban for these cases is
451 deferred until Perl 5.30, but the code has been fixed to raise a
452 default-on deprecation message for them in the meantime.
453 .PP
454 Some uses of literal \f(CW"{"\fR occur in contexts where we do not foresee
455 the meaning ever being anything but the literal, such as the very first
456 character in the pattern, or after a \f(CW"|"\fR meaning alternation. Thus
457 .PP
458 .Vb 1
459 \& qr/{fee|{fie/
460 .Ve
461 .PP
462 matches either of the strings \f(CW\*(C`{fee\*(C'\fR or \f(CW\*(C`{fie\*(C'\fR. To avoid forcing
463 unnecessary code changes, these uses do not need to be escaped, and no
464 warning is raised about them, and there are no current plans to change this.
465 .PP
466 But it is always correct to escape \f(CW"{"\fR, and the simple rule to
467 remember is to always do so.
468 .PP
469 See Unescaped left brace in regex is illegal here.
470 .ie n .SS "scalar(%hash) return signature changed"
471 .el .SS "\f(CWscalar(%hash)\fP return signature changed"
472 .IX Subsection "scalar(%hash) return signature changed"
473 The value returned for \f(CWscalar(%hash)\fR will no longer show information about
474 the buckets allocated in the hash. It will simply return the count of used
475 keys. It is thus equivalent to \f(CW\*(C`0+keys(%hash)\*(C'\fR.
476 .PP
477 A form of backward compatibility is provided via
478 \&\f(CWHash::Util::bucket_ratio()\fR which provides
479 the same behavior as
480 \&\f(CWscalar(%hash)\fR provided in Perl 5.24 and earlier.
481 .ie n .SS """keys"" returned from an lvalue subroutine"
482 .el .SS "\f(CWkeys\fP returned from an lvalue subroutine"
483 .IX Subsection "keys returned from an lvalue subroutine"
484 \&\f(CW\*(C`keys\*(C'\fR returned from an lvalue subroutine can no longer be assigned
485 to in list context.
486 .PP
487 .Vb 4
488 \& sub foo : lvalue { keys(%INC) }
489 \& (foo) = 3; # death
490 \& sub bar : lvalue { keys(@_) }
491 \& (bar) = 3; # also an error
492 .Ve
493 .PP
494 This makes the lvalue sub case consistent with \f(CW\*(C`(keys %hash) = ...\*(C'\fR and
495 \&\f(CW\*(C`(keys @_) = ...\*(C'\fR, which are also errors.
496 [GH #15339] <https://github.com/Perl/perl5/issues/15339>
497 .ie n .SS "The ""${^ENCODING}"" facility has been removed"
498 .el .SS "The \f(CW${^ENCODING}\fP facility has been removed"
499 .IX Subsection "The ${^ENCODING} facility has been removed"
500 The special behaviour associated with assigning a value to this variable
501 has been removed. As a consequence, the encoding pragma's default mode
502 is no longer supported. If
503 you still need to write your source code in encodings other than UTF\-8, use a
504 source filter such as Filter::Encoding on CPAN or encoding's \f(CW\*(C`Filter\*(C'\fR
505 option.
506 .ie n .SS "POSIX::tmpnam() has been removed"
507 .el .SS "\f(CWPOSIX::tmpnam()\fP has been removed"
508 .IX Subsection "POSIX::tmpnam() has been removed"
509 The fundamentally unsafe \f(CWtmpnam()\fR interface was deprecated in
510 Perl 5.22 and has now been removed. In its place, you can use,
511 for example, the File::Temp interfaces.
512 .SS "require ::Foo::Bar is now illegal"
513 .IX Subsection "require ::Foo::Bar is now illegal"
514 Formerly, \f(CW\*(C`require ::Foo::Bar\*(C'\fR would try to read \fI/Foo/Bar.pm\fR. Now any
515 bareword require which starts with a double colon dies instead.
516 .SS "Literal control character variable names are no longer permissible"
517 .IX Subsection "Literal control character variable names are no longer permissible"
518 A variable name may no longer contain a literal control character under
519 any circumstances. These previously were allowed in single-character
520 names on ASCII platforms, but have been deprecated there since Perl
521 5.20. This affects things like \f(CW\*(C`$\fR\f(CI\ecT\fR\f(CW\*(C'\fR, where \fI\ecT\fR is a literal
522 control (such as a \f(CW\*(C`NAK\*(C'\fR or \f(CW\*(C`NEGATIVE ACKNOWLEDGE\*(C'\fR character) in the
523 source code.
524 .ie n .SS """NBSP"" is no longer permissible in ""\eN{...}"""
525 .el .SS "\f(CWNBSP\fP is no longer permissible in \f(CW\eN{...}\fP"
526 .IX Subsection "NBSP is no longer permissible in N{...}"
527 The name of a character may no longer contain non-breaking spaces. It
528 has been deprecated to do so since Perl 5.22.
529 .SH Deprecations
530 .IX Header "Deprecations"
531 .SS "String delimiters that aren't stand-alone graphemes are now deprecated"
532 .IX Subsection "String delimiters that aren't stand-alone graphemes are now deprecated"
533 For Perl to eventually allow string delimiters to be Unicode
534 grapheme clusters (which look like a single character, but may be
535 a sequence of several ones), we have to stop allowing a single character
536 delimiter that isn't a grapheme by itself. These are unlikely to exist
537 in actual code, as they would typically display as attached to the
538 character in front of them.
539 .ie n .SS """\ec\fIX\fP"" that maps to a printable is no longer deprecated"
540 .el .SS "\f(CW\ec\fP\f(CIX\fP\f(CW\fP that maps to a printable is no longer deprecated"
541 .IX Subsection "cX that maps to a printable is no longer deprecated"
542 This means we have no plans to remove this feature. It still raises a
543 warning, but only if syntax warnings are enabled. The feature was
544 originally intended to be a way to express non-printable characters that
545 don't have a mnemonic (\f(CW\*(C`\et\*(C'\fR and \f(CW\*(C`\en\*(C'\fR are mnemonics for two
546 non-printable characters, but most non-printables don't have a
547 mnemonic.) But the feature can be used to specify a few printable
548 characters, though those are more clearly expressed as the printable
549 itself. See
550 <http://www.nntp.perl.org/group/perl.perl5.porters/2017/02/msg242944.html>.
551 .SH "Performance Enhancements"
552 .IX Header "Performance Enhancements"
553 .IP \(bu 4
554 A hash in boolean context is now sometimes faster, \fIe.g.\fR
555 .Sp
556 .Vb 1
557 \& if (!%h) { ... }
558 .Ve
559 .Sp
560 This was already special-cased, but some cases were missed (such as
561 \&\f(CW\*(C`grep %$_, @AoH\*(C'\fR), and even the ones which weren't have been improved.
562 .IP \(bu 4
563 New Faster Hash Function on 64 bit builds
564 .Sp
565 We use a different hash function for short and long keys. This should
566 improve performance and security, especially for long keys.
567 .IP \(bu 4
568 readline is faster
569 .Sp
570 Reading from a file line-by-line with \f(CWreadline()\fR or \f(CW\*(C`<>\*(C'\fR should
571 now typically be faster due to a better implementation of the code that
572 searches for the next newline character.
573 .IP \(bu 4
574 Assigning one reference to another, \fIe.g.\fR \f(CW\*(C`$ref1 = $ref2\*(C'\fR has been
575 optimized in some cases.
576 .IP \(bu 4
577 Remove some exceptions to creating Copy-on-Write strings. The string
578 buffer growth algorithm has been slightly altered so that you're less
579 likely to encounter a string which can't be COWed.
580 .IP \(bu 4
581 Better optimise array and hash assignment: where an array or hash appears
582 in the LHS of a list assignment, such as \f(CW\*(C`(..., @a) = (...);\*(C'\fR, it's
583 likely to be considerably faster, especially if it involves emptying the
584 array/hash. For example, this code runs about a third faster compared to
585 Perl 5.24.0:
586 .Sp
587 .Vb 5
588 \& my @a;
589 \& for my $i (1..10_000_000) {
590 \& @a = (1,2,3);
591 \& @a = ();
592 \& }
593 .Ve
594 .IP \(bu 4
595 Converting a single-digit string to a number is now substantially faster.
596 .IP \(bu 4
597 The \f(CW\*(C`split\*(C'\fR builtin is now slightly faster in many cases: in particular
598 for the two specially-handled forms
599 .Sp
600 .Vb 2
601 \& my @a = split ...;
602 \& local @a = split ...;
603 .Ve
604 .IP \(bu 4
605 The rather slow implementation for the experimental subroutine signatures
606 feature has been made much faster; it is now comparable in speed with the
607 traditional \f(CW\*(C`my ($a, $b, @c) = @_\*(C'\fR.
608 .IP \(bu 4
609 Bareword constant strings are now permitted to take part in constant
610 folding. They were originally exempted from constant folding in August 1999,
611 during the development of Perl 5.6, to ensure that \f(CW\*(C`use strict "subs"\*(C'\fR
612 would still apply to bareword constants. That has now been accomplished a
613 different way, so barewords, like other constants, now gain the performance
614 benefits of constant folding.
615 .Sp
616 This also means that void-context warnings on constant expressions of
617 barewords now report the folded constant operand, rather than the operation;
618 this matches the behaviour for non-bareword constants.
619 .SH "Modules and Pragmata"
620 .IX Header "Modules and Pragmata"
621 .SS "Updated Modules and Pragmata"
622 .IX Subsection "Updated Modules and Pragmata"
623 .IP \(bu 4
624 IO::Compress has been upgraded from version 2.069 to 2.074.
625 .IP \(bu 4
626 Archive::Tar has been upgraded from version 2.04 to 2.24.
627 .IP \(bu 4
628 arybase has been upgraded from version 0.11 to 0.12.
629 .IP \(bu 4
630 attributes has been upgraded from version 0.27 to 0.29.
631 .Sp
632 The deprecation message for the \f(CW\*(C`:unique\*(C'\fR and \f(CW\*(C`:locked\*(C'\fR attributes
633 now mention that they will disappear in Perl 5.28.
634 .IP \(bu 4
635 B has been upgraded from version 1.62 to 1.68.
636 .IP \(bu 4
637 B::Concise has been upgraded from version 0.996 to 0.999.
638 .Sp
639 Its output is now more descriptive for \f(CW\*(C`op_private\*(C'\fR flags.
640 .IP \(bu 4
641 B::Debug has been upgraded from version 1.23 to 1.24.
642 .IP \(bu 4
643 B::Deparse has been upgraded from version 1.37 to 1.40.
644 .IP \(bu 4
645 B::Xref has been upgraded from version 1.05 to 1.06.
646 .Sp
647 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
648 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
649 .IP \(bu 4
650 base has been upgraded from version 2.23 to 2.25.
651 .IP \(bu 4
652 bignum has been upgraded from version 0.42 to 0.47.
653 .IP \(bu 4
654 Carp has been upgraded from version 1.40 to 1.42.
655 .IP \(bu 4
656 charnames has been upgraded from version 1.43 to 1.44.
657 .IP \(bu 4
658 Compress::Raw::Bzip2 has been upgraded from version 2.069 to 2.074.
659 .IP \(bu 4
660 Compress::Raw::Zlib has been upgraded from version 2.069 to 2.074.
661 .IP \(bu 4
662 Config::Perl::V has been upgraded from version 0.25 to 0.28.
663 .IP \(bu 4
664 CPAN has been upgraded from version 2.11 to 2.18.
665 .IP \(bu 4
666 CPAN::Meta has been upgraded from version 2.150005 to 2.150010.
667 .IP \(bu 4
668 Data::Dumper has been upgraded from version 2.160 to 2.167.
669 .Sp
670 The XS implementation now supports Deparse.
671 .IP \(bu 4
672 DB_File has been upgraded from version 1.835 to 1.840.
673 .IP \(bu 4
674 Devel::Peek has been upgraded from version 1.23 to 1.26.
675 .IP \(bu 4
676 Devel::PPPort has been upgraded from version 3.32 to 3.35.
677 .IP \(bu 4
678 Devel::SelfStubber has been upgraded from version 1.05 to 1.06.
679 .Sp
680 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
681 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
682 .IP \(bu 4
683 diagnostics has been upgraded from version 1.34 to 1.36.
684 .Sp
685 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
686 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
687 .IP \(bu 4
688 Digest has been upgraded from version 1.17 to 1.17_01.
689 .IP \(bu 4
690 Digest::MD5 has been upgraded from version 2.54 to 2.55.
691 .IP \(bu 4
692 Digest::SHA has been upgraded from version 5.95 to 5.96.
693 .IP \(bu 4
694 DynaLoader has been upgraded from version 1.38 to 1.42.
695 .IP \(bu 4
696 Encode has been upgraded from version 2.80 to 2.88.
697 .IP \(bu 4
698 encoding has been upgraded from version 2.17 to 2.19.
699 .Sp
700 This module's default mode is no longer supported. It now
701 dies when imported, unless the \f(CW\*(C`Filter\*(C'\fR option is being used.
702 .IP \(bu 4
703 encoding::warnings has been upgraded from version 0.12 to 0.13.
704 .Sp
705 This module is no longer supported. It emits a warning to
706 that effect and then does nothing.
707 .IP \(bu 4
708 Errno has been upgraded from version 1.25 to 1.28.
709 .Sp
710 It now documents that using \f(CW\*(C`%!\*(C'\fR automatically loads Errno for you.
711 .Sp
712 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
713 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
714 .IP \(bu 4
715 ExtUtils::Embed has been upgraded from version 1.33 to 1.34.
716 .Sp
717 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
718 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
719 .IP \(bu 4
720 ExtUtils::MakeMaker has been upgraded from version 7.10_01 to 7.24.
721 .IP \(bu 4
722 ExtUtils::Miniperl has been upgraded from version 1.05 to 1.06.
723 .IP \(bu 4
724 ExtUtils::ParseXS has been upgraded from version 3.31 to 3.34.
725 .IP \(bu 4
726 ExtUtils::Typemaps has been upgraded from version 3.31 to 3.34.
727 .IP \(bu 4
728 feature has been upgraded from version 1.42 to 1.47.
729 .IP \(bu 4
730 File::Copy has been upgraded from version 2.31 to 2.32.
731 .IP \(bu 4
732 File::Fetch has been upgraded from version 0.48 to 0.52.
733 .IP \(bu 4
734 File::Glob has been upgraded from version 1.26 to 1.28.
735 .Sp
736 It now Issues a deprecation message for \f(CWFile::Glob::glob()\fR.
737 .IP \(bu 4
738 File::Spec has been upgraded from version 3.63 to 3.67.
739 .IP \(bu 4
740 FileHandle has been upgraded from version 2.02 to 2.03.
741 .IP \(bu 4
742 Filter::Simple has been upgraded from version 0.92 to 0.93.
743 .Sp
744 It no longer treats \f(CW\*(C`no MyFilter\*(C'\fR immediately following \f(CW\*(C`use MyFilter\*(C'\fR as
745 end-of-file.
746 [GH #11853] <https://github.com/Perl/perl5/issues/11853>
747 .IP \(bu 4
748 Getopt::Long has been upgraded from version 2.48 to 2.49.
749 .IP \(bu 4
750 Getopt::Std has been upgraded from version 1.11 to 1.12.
751 .IP \(bu 4
752 Hash::Util has been upgraded from version 0.19 to 0.22.
753 .IP \(bu 4
754 HTTP::Tiny has been upgraded from version 0.056 to 0.070.
755 .Sp
756 Internal 599\-series errors now include the redirect history.
757 .IP \(bu 4
758 I18N::LangTags has been upgraded from version 0.40 to 0.42.
759 .Sp
760 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
761 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
762 .IP \(bu 4
763 IO has been upgraded from version 1.36 to 1.38.
764 .IP \(bu 4
765 IO::Socket::IP has been upgraded from version 0.37 to 0.38.
766 .IP \(bu 4
767 IPC::Cmd has been upgraded from version 0.92 to 0.96.
768 .IP \(bu 4
769 IPC::SysV has been upgraded from version 2.06_01 to 2.07.
770 .IP \(bu 4
771 JSON::PP has been upgraded from version 2.27300 to 2.27400_02.
772 .IP \(bu 4
773 lib has been upgraded from version 0.63 to 0.64.
774 .Sp
775 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
776 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
777 .IP \(bu 4
778 List::Util has been upgraded from version 1.42_02 to 1.46_02.
779 .IP \(bu 4
780 Locale::Codes has been upgraded from version 3.37 to 3.42.
781 .IP \(bu 4
782 Locale::Maketext has been upgraded from version 1.26 to 1.28.
783 .IP \(bu 4
784 Locale::Maketext::Simple has been upgraded from version 0.21 to 0.21_01.
785 .IP \(bu 4
786 Math::BigInt has been upgraded from version 1.999715 to 1.999806.
787 .IP \(bu 4
788 Math::BigInt::FastCalc has been upgraded from version 0.40 to 0.5005.
789 .IP \(bu 4
790 Math::BigRat has been upgraded from version 0.260802 to 0.2611.
791 .IP \(bu 4
792 Math::Complex has been upgraded from version 1.59 to 1.5901.
793 .IP \(bu 4
794 Memoize has been upgraded from version 1.03 to 1.03_01.
795 .IP \(bu 4
796 Module::CoreList has been upgraded from version 5.20170420 to 5.20170530.
797 .IP \(bu 4
798 Module::Load::Conditional has been upgraded from version 0.64 to 0.68.
799 .IP \(bu 4
800 Module::Metadata has been upgraded from version 1.000031 to 1.000033.
801 .IP \(bu 4
802 mro has been upgraded from version 1.18 to 1.20.
803 .IP \(bu 4
804 Net::Ping has been upgraded from version 2.43 to 2.55.
805 .Sp
806 IPv6 addresses and \f(CW\*(C`AF_INET6\*(C'\fR sockets are now supported, along with several
807 other enhancements.
808 .IP \(bu 4
809 NEXT has been upgraded from version 0.65 to 0.67.
810 .IP \(bu 4
811 Opcode has been upgraded from version 1.34 to 1.39.
812 .IP \(bu 4
813 open has been upgraded from version 1.10 to 1.11.
814 .IP \(bu 4
815 OS2::Process has been upgraded from version 1.11 to 1.12.
816 .Sp
817 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
818 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
819 .IP \(bu 4
820 overload has been upgraded from version 1.26 to 1.28.
821 .Sp
822 Its compilation speed has been improved slightly.
823 .IP \(bu 4
824 parent has been upgraded from version 0.234 to 0.236.
825 .IP \(bu 4
826 perl5db.pl has been upgraded from version 1.50 to 1.51.
827 .Sp
828 It now ignores \fI/dev/tty\fR on non-Unix systems.
829 [GH #12244] <https://github.com/Perl/perl5/issues/12244>
830 .IP \(bu 4
831 Perl::OSType has been upgraded from version 1.009 to 1.010.
832 .IP \(bu 4
833 perlfaq has been upgraded from version 5.021010 to 5.021011.
834 .IP \(bu 4
835 PerlIO has been upgraded from version 1.09 to 1.10.
836 .IP \(bu 4
837 PerlIO::encoding has been upgraded from version 0.24 to 0.25.
838 .IP \(bu 4
839 PerlIO::scalar has been upgraded from version 0.24 to 0.26.
840 .IP \(bu 4
841 Pod::Checker has been upgraded from version 1.60 to 1.73.
842 .IP \(bu 4
843 Pod::Functions has been upgraded from version 1.10 to 1.11.
844 .IP \(bu 4
845 Pod::Html has been upgraded from version 1.22 to 1.2202.
846 .IP \(bu 4
847 Pod::Perldoc has been upgraded from version 3.25_02 to 3.28.
848 .IP \(bu 4
849 Pod::Simple has been upgraded from version 3.32 to 3.35.
850 .IP \(bu 4
851 Pod::Usage has been upgraded from version 1.68 to 1.69.
852 .IP \(bu 4
853 POSIX has been upgraded from version 1.65 to 1.76.
854 .Sp
855 This remedies several defects in making its symbols exportable.
856 [GH #15260] <https://github.com/Perl/perl5/issues/15260>
857 .Sp
858 The \f(CWPOSIX::tmpnam()\fR interface has been removed,
859 see "\fBPOSIX::tmpnam()\fR has been removed".
860 .Sp
861 The following deprecated functions have been removed:
862 .Sp
863 .Vb 10
864 \& POSIX::isalnum
865 \& POSIX::isalpha
866 \& POSIX::iscntrl
867 \& POSIX::isdigit
868 \& POSIX::isgraph
869 \& POSIX::islower
870 \& POSIX::isprint
871 \& POSIX::ispunct
872 \& POSIX::isspace
873 \& POSIX::isupper
874 \& POSIX::isxdigit
875 \& POSIX::tolower
876 \& POSIX::toupper
877 .Ve
878 .Sp
879 Trying to import POSIX subs that have no real implementations
880 (like \f(CWPOSIX::atend()\fR) now fails at import time, instead of
881 waiting until runtime.
882 .IP \(bu 4
883 re has been upgraded from version 0.32 to 0.34
884 .Sp
885 This adds support for the new \f(CW\*(C`/xx\*(C'\fR
886 regular expression pattern modifier, and a change to the \f(CWuse\ re\ \*(Aqstrict\*(Aq\fR experimental feature. When \f(CWre\ \*(Aqstrict\*(Aq\fR is enabled, a warning now will be generated for all
887 unescaped uses of the two characters \f(CW"}"\fR and \f(CW"]"\fR in regular
888 expression patterns (outside bracketed character classes) that are taken
889 literally. This brings them more in line with the \f(CW")"\fR character which
890 is always a metacharacter unless escaped. Being a metacharacter only
891 sometimes, depending on an action at a distance, can lead to silently
892 having the pattern mean something quite different than was intended,
893 which the \f(CW\*(C`re\ \*(Aqstrict\*(Aq\*(C'\fR mode is intended to minimize.
894 .IP \(bu 4
895 Safe has been upgraded from version 2.39 to 2.40.
896 .IP \(bu 4
897 Scalar::Util has been upgraded from version 1.42_02 to 1.46_02.
898 .IP \(bu 4
899 Storable has been upgraded from version 2.56 to 2.62.
900 .Sp
901 Fixes
902 [GH #15714] <https://github.com/Perl/perl5/issues/15714>.
903 .IP \(bu 4
904 Symbol has been upgraded from version 1.07 to 1.08.
905 .IP \(bu 4
906 Sys::Syslog has been upgraded from version 0.33 to 0.35.
907 .IP \(bu 4
908 Term::ANSIColor has been upgraded from version 4.04 to 4.06.
909 .IP \(bu 4
910 Term::ReadLine has been upgraded from version 1.15 to 1.16.
911 .Sp
912 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
913 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
914 .IP \(bu 4
915 Test has been upgraded from version 1.28 to 1.30.
916 .Sp
917 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
918 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
919 .IP \(bu 4
920 Test::Harness has been upgraded from version 3.36 to 3.38.
921 .IP \(bu 4
922 Test::Simple has been upgraded from version 1.001014 to 1.302073.
923 .IP \(bu 4
924 Thread::Queue has been upgraded from version 3.09 to 3.12.
925 .IP \(bu 4
926 Thread::Semaphore has been upgraded from 2.12 to 2.13.
927 .Sp
928 Added the \f(CW\*(C`down_timed\*(C'\fR method.
929 .IP \(bu 4
930 threads has been upgraded from version 2.07 to 2.15.
931 .IP \(bu 4
932 threads::shared has been upgraded from version 1.51 to 1.56.
933 .IP \(bu 4
934 Tie::Hash::NamedCapture has been upgraded from version 0.09 to 0.10.
935 .IP \(bu 4
936 Time::HiRes has been upgraded from version 1.9733 to 1.9741.
937 .Sp
938 It now builds on systems with C++11 compilers (such as G++ 6 and Clang++
939 3.9).
940 .Sp
941 Now uses \f(CW\*(C`clockid_t\*(C'\fR.
942 .IP \(bu 4
943 Time::Local has been upgraded from version 1.2300 to 1.25.
944 .IP \(bu 4
945 Unicode::Collate has been upgraded from version 1.14 to 1.19.
946 .IP \(bu 4
947 Unicode::UCD has been upgraded from version 0.64 to 0.68.
948 .Sp
949 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
950 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
951 .IP \(bu 4
952 version has been upgraded from version 0.9916 to 0.9917.
953 .IP \(bu 4
954 VMS::DCLsym has been upgraded from version 1.06 to 1.08.
955 .Sp
956 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
957 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
958 .IP \(bu 4
959 warnings has been upgraded from version 1.36 to 1.37.
960 .IP \(bu 4
961 XS::Typemap has been upgraded from version 0.14 to 0.15.
962 .IP \(bu 4
963 XSLoader has been upgraded from version 0.21 to 0.27.
964 .Sp
965 Fixed a security hole in which binary files could be loaded from a path
966 outside of \f(CW@INC\fR.
967 .Sp
968 It now uses 3\-arg \f(CWopen()\fR instead of 2\-arg \f(CWopen()\fR.
969 [GH #15721] <https://github.com/Perl/perl5/issues/15721>
970 .SH Documentation
971 .IX Header "Documentation"
972 .SS "New Documentation"
973 .IX Subsection "New Documentation"
974 \fIperldeprecation\fR
975 .IX Subsection "perldeprecation"
976 .PP
977 This file documents all upcoming deprecations, and some of the deprecations
978 which already have been removed. The purpose of this documentation is
979 two-fold: document what will disappear, and by which version, and serve
980 as a guide for people dealing with code which has features that no longer
981 work after an upgrade of their perl.
982 .SS "Changes to Existing Documentation"
983 .IX Subsection "Changes to Existing Documentation"
984 We have attempted to update the documentation to reflect the changes
985 listed in this document. If you find any we have missed, send email to
986 [email protected] <mailto:[email protected]>.
987 .PP
988 Additionally, all references to Usenet have been removed, and the
989 following selected changes have been made:
990 .PP
991 \fIperlfunc\fR
992 .IX Subsection "perlfunc"
993 .IP \(bu 4
994 Removed obsolete text about \f(CWdefined()\fR
995 on aggregates that should have been deleted earlier, when the feature
996 was removed.
997 .IP \(bu 4
998 Corrected documentation of \f(CWeval()\fR,
999 and \f(CWevalbytes()\fR.
1000 .IP \(bu 4
1001 Clarified documentation of \f(CWseek()\fR,
1002 \&\f(CWtell()\fR and \f(CWsysseek()\fR
1003 emphasizing that positions are in bytes and not characters.
1004 [GH #15438] <https://github.com/Perl/perl5/issues/15438>
1005 .IP \(bu 4
1006 Clarified documentation of \f(CWsort()\fR concerning
1007 the variables \f(CW$a\fR and \f(CW$b\fR.
1008 .IP \(bu 4
1009 In \f(CWsplit()\fR noted that certain pattern modifiers are
1010 legal, and added a caution about its use in Perls before v5.11.
1011 .IP \(bu 4
1012 Removed obsolete documentation of \f(CWstudy()\fR, noting
1013 that it is now a no-op.
1014 .IP \(bu 4
1015 Noted that \f(CWvec()\fR doesn't work well when the string
1016 contains characters whose code points are above 255.
1017 .PP
1018 \fIperlguts\fR
1019 .IX Subsection "perlguts"
1020 .IP \(bu 4
1021 Added advice on
1022 formatted printing of operands of \f(CW\*(C`Size_t\*(C'\fR and \f(CW\*(C`SSize_t\*(C'\fR
1023 .PP
1024 \fIperlhack\fR
1025 .IX Subsection "perlhack"
1026 .IP \(bu 4
1027 Clarify what editor tab stop rules to use, and note that we are
1028 migrating away from using tabs, replacing them with sequences of SPACE
1029 characters.
1030 .PP
1031 \fIperlhacktips\fR
1032 .IX Subsection "perlhacktips"
1033 .IP \(bu 4
1034 Give another reason to use \f(CW\*(C`cBOOL\*(C'\fR to cast an expression to boolean.
1035 .IP \(bu 4
1036 Note that the macros \f(CW\*(C`TRUE\*(C'\fR and \f(CW\*(C`FALSE\*(C'\fR are available to express
1037 boolean values.
1038 .PP
1039 \fIperlinterp\fR
1040 .IX Subsection "perlinterp"
1041 .IP \(bu 4
1042 perlinterp has been expanded to give a more detailed example of how to
1043 hunt around in the parser for how a given operator is handled.
1044 .PP
1045 \fIperllocale\fR
1046 .IX Subsection "perllocale"
1047 .IP \(bu 4
1048 Some locales aren't compatible with Perl. Note that these can cause
1049 core dumps.
1050 .PP
1051 \fIperlmod\fR
1052 .IX Subsection "perlmod"
1053 .IP \(bu 4
1054 Various clarifications have been added.
1055 .PP
1056 \fIperlmodlib\fR
1057 .IX Subsection "perlmodlib"
1058 .IP \(bu 4
1059 Updated the site mirror list.
1060 .PP
1061 \fIperlobj\fR
1062 .IX Subsection "perlobj"
1063 .IP \(bu 4
1064 Added a section on calling methods using their fully qualified names.
1065 .IP \(bu 4
1066 Do not discourage manual \f(CW@ISA\fR.
1067 .PP
1068 \fIperlootut\fR
1069 .IX Subsection "perlootut"
1070 .IP \(bu 4
1071 Mention \f(CW\*(C`Moo\*(C'\fR more.
1072 .PP
1073 \fIperlop\fR
1074 .IX Subsection "perlop"
1075 .IP \(bu 4
1076 Note that white space must be used for quoting operators if the
1077 delimiter is a word character (\fIi.e.\fR, matches \f(CW\*(C`\ew\*(C'\fR).
1078 .IP \(bu 4
1079 Clarify that in regular expression patterns delimited by single quotes,
1080 no variable interpolation is done.
1081 .PP
1082 \fIperlre\fR
1083 .IX Subsection "perlre"
1084 .IP \(bu 4
1085 The first part was extensively rewritten to incorporate various basic
1086 points, that in earlier versions were mentioned in sort of an appendix
1087 on Version 8 regular expressions.
1088 .IP \(bu 4
1089 Note that it is common to have the \f(CW\*(C`/x\*(C'\fR modifier and forget that this
1090 means that \f(CW"#"\fR has to be escaped.
1091 .PP
1092 \fIperlretut\fR
1093 .IX Subsection "perlretut"
1094 .IP \(bu 4
1095 Add introductory material.
1096 .IP \(bu 4
1097 Note that a metacharacter occurring in a context where it can't mean
1098 that, silently loses its meta-ness and matches literally.
1099 \&\f(CW\*(C`use re \*(Aqstrict\*(Aq\*(C'\fR can catch some of these.
1100 .PP
1101 \fIperlunicode\fR
1102 .IX Subsection "perlunicode"
1103 .IP \(bu 4
1104 Corrected the text about Unicode BYTE ORDER MARK handling.
1105 .IP \(bu 4
1106 Updated the text to correspond with changes in Unicode UTS#18, concerning
1107 regular expressions, and Perl compatibility with what it says.
1108 .PP
1109 \fIperlvar\fR
1110 .IX Subsection "perlvar"
1111 .IP \(bu 4
1112 Document \f(CW@ISA\fR. It was documented in other places, but not in perlvar.
1113 .SH Diagnostics
1114 .IX Header "Diagnostics"
1115 .SS "New Diagnostics"
1116 .IX Subsection "New Diagnostics"
1117 \fINew Errors\fR
1118 .IX Subsection "New Errors"
1119 .IP \(bu 4
1120 A signature parameter must start with \f(CW\*(Aq$\*(Aq\fR, \f(CW\*(Aq@\*(Aq\fR or \f(CW\*(Aq%\*(Aq\fR
1121 .IP \(bu 4
1122 Bareword in require contains "%s"
1123 .IP \(bu 4
1124 Bareword in require maps to empty filename
1125 .IP \(bu 4
1126 Bareword in require maps to disallowed filename "%s"
1127 .IP \(bu 4
1128 Bareword in require must not start with a double-colon: "%s"
1129 .IP \(bu 4
1130 \&\f(CW%s:\fR command not found
1131 .Sp
1132 (A) You've accidentally run your script through \fBbash\fR or another shell
1133 instead of Perl. Check the \f(CW\*(C`#!\*(C'\fR line, or manually feed your script into
1134 Perl yourself. The \f(CW\*(C`#!\*(C'\fR line at the top of your file could look like:
1135 .Sp
1136 .Vb 1
1137 \& #!/usr/bin/perl
1138 .Ve
1139 .IP \(bu 4
1140 \&\f(CW%s:\fR command not found: \f(CW%s\fR
1141 .Sp
1142 (A) You've accidentally run your script through \fBzsh\fR or another shell
1143 instead of Perl. Check the \f(CW\*(C`#!\*(C'\fR line, or manually feed your script into
1144 Perl yourself. The \f(CW\*(C`#!\*(C'\fR line at the top of your file could look like:
1145 .Sp
1146 .Vb 1
1147 \& #!/usr/bin/perl
1148 .Ve
1149 .IP \(bu 4
1150 The experimental declared_refs feature is not enabled
1151 .Sp
1152 (F) To declare references to variables, as in \f(CW\*(C`my \e%x\*(C'\fR, you must first enable
1153 the feature:
1154 .Sp
1155 .Vb 2
1156 \& no warnings "experimental::declared_refs";
1157 \& use feature "declared_refs";
1158 .Ve
1159 .Sp
1160 See "Declaring a reference to a variable".
1161 .IP \(bu 4
1162 Illegal character following sigil in a subroutine signature
1163 .IP \(bu 4
1164 Indentation on line \f(CW%d\fR of here-doc doesn't match delimiter
1165 .IP \(bu 4
1166 Infinite recursion via empty pattern.
1167 .Sp
1168 Using the empty pattern (which re-executes the last successfully-matched
1169 pattern) inside a code block in another regex, as in \f(CW\*(C`/(?{ s!!new! })/\*(C'\fR, has
1170 always previously yielded a segfault. It now produces this error.
1171 .IP \(bu 4
1172 Malformed UTF\-8 string in "%s"
1173 .IP \(bu 4
1174 Multiple slurpy parameters not allowed
1175 .IP \(bu 4
1176 \&\f(CW\*(Aq#\*(Aq\fR not allowed immediately following a sigil in a subroutine signature
1177 .IP \(bu 4
1178 panic: unknown OA_*: \f(CW%x\fR
1179 .IP \(bu 4
1180 Unescaped left brace in regex is illegal here
1181 .Sp
1182 Unescaped left braces are now illegal in some contexts in regular expression
1183 patterns. In other contexts, they are still just deprecated; they will
1184 be illegal in Perl 5.30.
1185 .IP \(bu 4
1186 Version control conflict marker
1187 .Sp
1188 (F) The parser found a line starting with \f(CW\*(C`<<<<<<<\*(C'\fR,
1189 \&\f(CW\*(C`>>>>>>>\*(C'\fR, or \f(CW\*(C`=======\*(C'\fR. These may be left by a
1190 version control system to mark conflicts after a failed merge operation.
1191 .PP
1192 \fINew Warnings\fR
1193 .IX Subsection "New Warnings"
1194 .IP \(bu 4
1195 Can't determine class of operator \f(CW%s\fR, assuming \f(CW\*(C`BASEOP\*(C'\fR
1196 .IP \(bu 4
1197 Declaring references is experimental
1198 .Sp
1199 (S experimental::declared_refs) This warning is emitted if you use a reference
1200 constructor on the right-hand side of \f(CWmy()\fR, \f(CWstate()\fR, \f(CWour()\fR, or
1201 \&\f(CWlocal()\fR. Simply suppress the warning if you want to use the feature, but
1202 know that in doing so you are taking the risk of using an experimental feature
1203 which may change or be removed in a future Perl version:
1204 .Sp
1205 .Vb 3
1206 \& no warnings "experimental::declared_refs";
1207 \& use feature "declared_refs";
1208 \& $fooref = my \e$foo;
1209 .Ve
1210 .Sp
1211 See "Declaring a reference to a variable".
1212 .IP \(bu 4
1213 do "%s" failed, '.' is no longer in \f(CW@INC\fR
1214 .Sp
1215 Since \f(CW"."\fR is now removed from \f(CW@INC\fR by default, \f(CW\*(C`do\*(C'\fR will now trigger a warning recommending to fix the \f(CW\*(C`do\*(C'\fR statement.
1216 .IP \(bu 4
1217 \&\f(CWFile::Glob::glob()\fR will disappear in perl 5.30. Use \f(CWFile::Glob::bsd_glob()\fR instead.
1218 .IP \(bu 4
1219 Unescaped literal '%c' in regex; marked by <\-\- HERE in m/%s/
1220 .IP \(bu 4
1221 Use of unassigned code point or non-standalone grapheme for a delimiter will be a fatal error starting in Perl 5.30
1222 .Sp
1223 See "Deprecations"
1224 .SS "Changes to Existing Diagnostics"
1225 .IX Subsection "Changes to Existing Diagnostics"
1226 .IP \(bu 4
1227 When a \f(CW\*(C`require\*(C'\fR fails, we now do not provide \f(CW@INC\fR when the \f(CW\*(C`require\*(C'\fR
1228 is for a file instead of a module.
1229 .IP \(bu 4
1230 When \f(CW@INC\fR is not scanned for a \f(CW\*(C`require\*(C'\fR call, we no longer display
1231 \&\f(CW@INC\fR to avoid confusion.
1232 .IP \(bu 4
1233 Attribute "locked" is deprecated, and will disappear in Perl 5.28
1234 .Sp
1235 This existing warning has had the \fIand will disappear\fR text added in this
1236 release.
1237 .IP \(bu 4
1238 Attribute "unique" is deprecated, and will disappear in Perl 5.28
1239 .Sp
1240 This existing warning has had the \fIand will disappear\fR text added in this
1241 release.
1242 .IP \(bu 4
1243 Calling POSIX::%s() is deprecated
1244 .Sp
1245 This warning has been removed, as the deprecated functions have been
1246 removed from POSIX.
1247 .IP \(bu 4
1248 Constants from lexical variables potentially modified elsewhere are deprecated. This will not be allowed in Perl 5.32
1249 .Sp
1250 This existing warning has had the \fIthis will not be allowed\fR text added
1251 in this release.
1252 .IP \(bu 4
1253 Deprecated use of \f(CWmy()\fR in false conditional. This will be a fatal error in Perl 5.30
1254 .Sp
1255 This existing warning has had the \fIthis will be a fatal error\fR text added
1256 in this release.
1257 .IP \(bu 4
1258 \&\f(CWdump()\fR better written as \f(CWCORE::dump()\fR. \f(CWdump()\fR will no longer be available in Perl 5.30
1259 .Sp
1260 This existing warning has had the \fIno longer be available\fR text added in
1261 this release.
1262 .IP \(bu 4
1263 Experimental \f(CW%s\fR on scalar is now forbidden
1264 .Sp
1265 This message is now followed by more helpful text.
1266 [GH #15291] <https://github.com/Perl/perl5/issues/15291>
1267 .IP \(bu 4
1268 Experimental "%s" subs not enabled
1269 .Sp
1270 This warning was been removed, as lexical subs are no longer experimental.
1271 .IP \(bu 4
1272 Having more than one /%c regexp modifier is deprecated
1273 .Sp
1274 This deprecation warning has been removed, since \f(CW\*(C`/xx\*(C'\fR now has a new
1275 meaning.
1276 .IP \(bu 4
1277 %s() is deprecated on \f(CW\*(C`:utf8\*(C'\fR handles. This will be a fatal error in Perl 5.30
1278 \&.
1279 .Sp
1280 where "%s" is one of \f(CW\*(C`sysread\*(C'\fR, \f(CW\*(C`recv\*(C'\fR, \f(CW\*(C`syswrite\*(C'\fR, or \f(CW\*(C`send\*(C'\fR.
1281 .Sp
1282 This existing warning has had the \fIthis will be a fatal error\fR text added
1283 in this release.
1284 .Sp
1285 This warning is now enabled by default, as all \f(CW\*(C`deprecated\*(C'\fR category
1286 warnings should be.
1287 .IP \(bu 4
1288 \&\f(CW$*\fR is no longer supported. Its use will be fatal in Perl 5.30
1289 .Sp
1290 This existing warning has had the \fIits use will be fatal\fR text added in
1291 this release.
1292 .IP \(bu 4
1293 \&\f(CW$#\fR is no longer supported. Its use will be fatal in Perl 5.30
1294 .Sp
1295 This existing warning has had the \fIits use will be fatal\fR text added in
1296 this release.
1297 .IP \(bu 4
1298 Malformed UTF\-8 character%s
1299 .Sp
1300 Details as to the exact problem have been added at the end of this
1301 message
1302 .IP \(bu 4
1303 Missing or undefined argument to \f(CW%s\fR
1304 .Sp
1305 This warning used to warn about \f(CW\*(C`require\*(C'\fR, even if it was actually \f(CW\*(C`do\*(C'\fR
1306 which being executed. It now gets the operation name right.
1307 .IP \(bu 4
1308 NO-BREAK SPACE in a charnames alias definition is deprecated
1309 .Sp
1310 This warning has been removed as the behavior is now an error.
1311 .IP \(bu 4
1312 Odd name/value argument for subroutine '%s'
1313 .Sp
1314 This warning now includes the name of the offending subroutine.
1315 .IP \(bu 4
1316 Opening dirhandle \f(CW%s\fR also as a file. This will be a fatal error in Perl 5.28
1317 .Sp
1318 This existing warning has had the \fIthis will be a fatal error\fR text added
1319 in this release.
1320 .IP \(bu 4
1321 Opening filehandle \f(CW%s\fR also as a directory. This will be a fatal error in Perl 5.28
1322 .Sp
1323 This existing warning has had the \fIthis will be a fatal error\fR text added
1324 in this release.
1325 .IP \(bu 4
1326 panic: ck_split, type=%u
1327 .Sp
1328 panic: pp_split, pm=%p, s=%p
1329 .Sp
1330 These panic errors have been removed.
1331 .IP \(bu 4
1332 Passing malformed UTF\-8 to "%s" is deprecated
1333 .Sp
1334 This warning has been changed to the fatal
1335 Malformed UTF\-8 string in "%s"
1336 .IP \(bu 4
1337 Setting \f(CW$/\fR to a reference to \f(CW%s\fR as a form of slurp is deprecated, treating as undef. This will be fatal in Perl 5.28
1338 .Sp
1339 This existing warning has had the \fIthis will be fatal\fR text added in
1340 this release.
1341 .IP \(bu 4
1342 \&\f(CW\*(C`${^ENCODING}\*(C'\fR is no longer supported. Its use will be fatal in Perl 5.28
1343 .Sp
1344 This warning used to be: "Setting \f(CW\*(C`${^ENCODING}\*(C'\fR is deprecated".
1345 .Sp
1346 The special action of the variable \f(CW\*(C`${^ENCODING}\*(C'\fR was formerly used to
1347 implement the \f(CW\*(C`encoding\*(C'\fR pragma. As of Perl 5.26, rather than being
1348 deprecated, assigning to this variable now has no effect except to issue
1349 the warning.
1350 .IP \(bu 4
1351 Too few arguments for subroutine '%s'
1352 .Sp
1353 This warning now includes the name of the offending subroutine.
1354 .IP \(bu 4
1355 Too many arguments for subroutine '%s'
1356 .Sp
1357 This warning now includes the name of the offending subroutine.
1358 .IP \(bu 4
1359 Unescaped left brace in regex is deprecated here (and will be fatal in Perl 5.30), passed through in regex; marked by <\-\-\ HERE in m/%s/
1360 .Sp
1361 This existing warning has had the \fIhere (and will be fatal...)\fR text
1362 added in this release.
1363 .IP \(bu 4
1364 Unknown charname '' is deprecated. Its use will be fatal in Perl 5.28
1365 .Sp
1366 This existing warning has had the \fIits use will be fatal\fR text added in
1367 this release.
1368 .IP \(bu 4
1369 Use of bare << to mean <<"" is deprecated. Its use will be fatal in Perl 5.28
1370 .Sp
1371 This existing warning has had the \fIits use will be fatal\fR text added in
1372 this release.
1373 .IP \(bu 4
1374 Use of code point 0x%s is deprecated; the permissible max is 0x%s. This will be fatal in Perl 5.28
1375 .Sp
1376 This existing warning has had the \fIthis will be fatal\fR text added in
1377 this release.
1378 .IP \(bu 4
1379 Use of comma-less variable list is deprecated. Its use will be fatal in Perl 5.28
1380 .Sp
1381 This existing warning has had the \fIits use will be fatal\fR text added in
1382 this release.
1383 .IP \(bu 4
1384 Use of inherited \f(CW\*(C`AUTOLOAD\*(C'\fR for non-method %s() is deprecated. This will be fatal in Perl 5.28
1385 .Sp
1386 This existing warning has had the \fIthis will be fatal\fR text added in
1387 this release.
1388 .IP \(bu 4
1389 Use of strings with code points over 0xFF as arguments to \f(CW%s\fR operator is deprecated. This will be a fatal error in Perl 5.28
1390 .Sp
1391 This existing warning has had the \fIthis will be a fatal error\fR text added in
1392 this release.
1393 .SH "Utility Changes"
1394 .IX Header "Utility Changes"
1395 .SS "\fIc2ph\fP and \fIpstruct\fP"
1396 .IX Subsection "c2ph and pstruct"
1397 .IP \(bu 4
1398 These old utilities have long since superceded by h2xs, and are
1399 now gone from the distribution.
1400 .SS \fIPorting/pod_lib.pl\fP
1401 .IX Subsection "Porting/pod_lib.pl"
1402 .IP \(bu 4
1403 Removed spurious executable bit.
1404 .IP \(bu 4
1405 Account for the possibility of DOS file endings.
1406 .SS \fIPorting/sync\-with\-cpan\fP
1407 .IX Subsection "Porting/sync-with-cpan"
1408 .IP \(bu 4
1409 Many improvements.
1410 .SS \fIperf/benchmarks\fP
1411 .IX Subsection "perf/benchmarks"
1412 .IP \(bu 4
1413 Tidy file, rename some symbols.
1414 .SS \fIPorting/checkAUTHORS.pl\fP
1415 .IX Subsection "Porting/checkAUTHORS.pl"
1416 .IP \(bu 4
1417 Replace obscure character range with \f(CW\*(C`\ew\*(C'\fR.
1418 .SS \fIt/porting/regen.t\fP
1419 .IX Subsection "t/porting/regen.t"
1420 .IP \(bu 4
1421 Try to be more helpful when tests fail.
1422 .SS \fIutils/h2xs.PL\fP
1423 .IX Subsection "utils/h2xs.PL"
1424 .IP \(bu 4
1425 Avoid infinite loop for enums.
1426 .SS perlbug
1427 .IX Subsection "perlbug"
1428 .IP \(bu 4
1429 Long lines in the message body are now wrapped at 900 characters, to stay
1430 well within the 1000\-character limit imposed by SMTP mail transfer agents.
1431 This is particularly likely to be important for the list of arguments to
1432 \&\fIConfigure\fR, which can readily exceed the limit if, for example, it names
1433 several non-default installation paths. This change also adds the first unit
1434 tests for perlbug.
1435 [perl #128020] <https://rt.perl.org/Public/Bug/Display.html?id=128020>
1436 .SH "Configuration and Compilation"
1437 .IX Header "Configuration and Compilation"
1438 .IP \(bu 4
1439 \&\f(CW\*(C`\-Ddefault_inc_excludes_dot\*(C'\fR has added, and enabled by default.
1440 .IP \(bu 4
1441 The \f(CW\*(C`dtrace\*(C'\fR build process has further changes
1442 [GH #15718] <https://github.com/Perl/perl5/issues/15718>:
1443 .RS 4
1444 .IP \(bu 4
1445 If the \f(CW\*(C`\-xnolibs\*(C'\fR is available, use that so a \fIdtrace\fR perl can be
1446 built within a FreeBSD jail.
1447 .IP \(bu 4
1448 On systems that build a \fIdtrace\fR object file (FreeBSD, Solaris, and
1449 SystemTap's dtrace emulation), copy the input objects to a separate
1450 directory and process them there, and use those objects in the link,
1451 since \f(CW\*(C`dtrace \-G\*(C'\fR also modifies these objects.
1452 .IP \(bu 4
1453 Add \fIlibelf\fR to the build on FreeBSD 10.x, since \fIdtrace\fR adds
1454 references to \fIlibelf\fR symbols.
1455 .IP \(bu 4
1456 Generate a dummy \fIdtrace_main.o\fR if \f(CW\*(C`dtrace \-G\*(C'\fR fails to build it. A
1457 default build on Solaris generates probes from the unused inline
1458 functions, while they don't on FreeBSD, which causes \f(CW\*(C`dtrace \-G\*(C'\fR to
1459 fail.
1460 .RE
1461 .RS 4
1462 .RE
1463 .IP \(bu 4
1464 You can now disable perl's use of the \f(CW\*(C`PERL_HASH_SEED\*(C'\fR and
1465 \&\f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR environment variables by configuring perl with
1466 \&\f(CW\*(C`\-Accflags=NO_PERL_HASH_ENV\*(C'\fR.
1467 .IP \(bu 4
1468 You can now disable perl's use of the \f(CW\*(C`PERL_HASH_SEED_DEBUG\*(C'\fR environment
1469 variable by configuring perl with
1470 \&\f(CW\*(C`\-Accflags=\-DNO_PERL_HASH_SEED_DEBUG\*(C'\fR.
1471 .IP \(bu 4
1472 \&\fIConfigure\fR now zeroes out the alignment bytes when calculating the bytes
1473 for 80\-bit \f(CW\*(C`NaN\*(C'\fR and \f(CW\*(C`Inf\*(C'\fR to make builds more reproducible.
1474 [GH #15725] <https://github.com/Perl/perl5/issues/15725>
1475 .IP \(bu 4
1476 Since v5.18, for testing purposes we have included support for
1477 building perl with a variety of non-standard, and non-recommended
1478 hash functions. Since we do not recommend the use of these functions,
1479 we have removed them and their corresponding build options. Specifically
1480 this includes the following build options:
1481 .Sp
1482 .Vb 8
1483 \& PERL_HASH_FUNC_SDBM
1484 \& PERL_HASH_FUNC_DJB2
1485 \& PERL_HASH_FUNC_SUPERFAST
1486 \& PERL_HASH_FUNC_MURMUR3
1487 \& PERL_HASH_FUNC_ONE_AT_A_TIME
1488 \& PERL_HASH_FUNC_ONE_AT_A_TIME_OLD
1489 \& PERL_HASH_FUNC_MURMUR_HASH_64A
1490 \& PERL_HASH_FUNC_MURMUR_HASH_64B
1491 .Ve
1492 .IP \(bu 4
1493 Remove "Warning: perl appears in your path"
1494 .Sp
1495 This install warning is more or less obsolete, since most platforms already
1496 \&\fBwill\fR have a \fI/usr/bin/perl\fR or similar provided by the OS.
1497 .IP \(bu 4
1498 Reduce verbosity of \f(CW\*(C`make install.man\*(C'\fR
1499 .Sp
1500 Previously, two progress messages were emitted for each manpage: one by
1501 installman itself, and one by the function in \fIinstall_lib.pl\fR that it calls to
1502 actually install the file. Disabling the second of those in each case saves
1503 over 750 lines of unhelpful output.
1504 .IP \(bu 4
1505 Cleanup for \f(CW\*(C`clang \-Weverything\*(C'\fR support.
1506 [GH #15683] <https://github.com/Perl/perl5/issues/15683>
1507 .IP \(bu 4
1508 \&\fIConfigure\fR: signbit scan was assuming too much, stop assuming negative 0.
1509 .IP \(bu 4
1510 Various compiler warnings have been silenced.
1511 .IP \(bu 4
1512 Several smaller changes have been made to remove impediments to compiling
1513 under C++11.
1514 .IP \(bu 4
1515 Builds using \f(CW\*(C`USE_PAD_RESET\*(C'\fR now work again; this configuration had
1516 bit-rotted.
1517 .IP \(bu 4
1518 A probe for \f(CW\*(C`gai_strerror\*(C'\fR was added to \fIConfigure\fR that checks if
1519 the \f(CWgai_strerror()\fR routine is available and can be used to
1520 translate error codes returned by \f(CWgetaddrinfo()\fR into human
1521 readable strings.
1522 .IP \(bu 4
1523 \&\fIConfigure\fR now aborts if both \f(CW\*(C`\-Duselongdouble\*(C'\fR and \f(CW\*(C`\-Dusequadmath\*(C'\fR are
1524 requested.
1525 [GH #14944] <https://github.com/Perl/perl5/issues/14944>
1526 .IP \(bu 4
1527 Fixed a bug in which \fIConfigure\fR could append \f(CW\*(C`\-quadmath\*(C'\fR to the
1528 archname even if it was already present.
1529 [GH #15423] <https://github.com/Perl/perl5/issues/15423>
1530 .IP \(bu 4
1531 Clang builds with \f(CW\*(C`\-DPERL_GLOBAL_STRUCT\*(C'\fR or
1532 \&\f(CW\*(C`\-DPERL_GLOBAL_STRUCT_PRIVATE\*(C'\fR have
1533 been fixed (by disabling Thread Safety Analysis for these configurations).
1534 .IP \(bu 4
1535 \&\fImake_ext.pl\fR no longer updates a module's \fIpm_to_blib\fR file when no
1536 files require updates. This could cause dependencies, \fIperlmain.c\fR
1537 in particular, to be rebuilt unnecessarily.
1538 [GH #15060] <https://github.com/Perl/perl5/issues/15060>
1539 .IP \(bu 4
1540 The output of \f(CW\*(C`perl \-V\*(C'\fR has been reformatted so that each configuration
1541 and compile-time option is now listed one per line, to improve
1542 readability.
1543 .IP \(bu 4
1544 \&\fIConfigure\fR now builds \f(CW\*(C`miniperl\*(C'\fR and \f(CW\*(C`generate_uudmap\*(C'\fR if you
1545 invoke it with \f(CW\*(C`\-Dusecrosscompiler\*(C'\fR but not \f(CW\*(C`\-Dtargethost=somehost\*(C'\fR.
1546 This means you can supply your target platform \f(CW\*(C`config.sh\*(C'\fR, generate
1547 the headers and proceed to build your cross-target perl.
1548 [GH #15126] <https://github.com/Perl/perl5/issues/15126>
1549 .IP \(bu 4
1550 Perl built with \f(CW\*(C`\-Accflags=\-DPERL_TRACE_OPS\*(C'\fR now only dumps the operator
1551 counts when the environment variable \f(CW\*(C`PERL_TRACE_OPS\*(C'\fR is set to a
1552 non-zero integer. This allows \f(CW\*(C`make test\*(C'\fR to pass on such a build.
1553 .IP \(bu 4
1554 When building with GCC 6 and link-time optimization (the \f(CW\*(C`\-flto\*(C'\fR option to
1555 \&\f(CW\*(C`gcc\*(C'\fR), \fIConfigure\fR was treating all probed symbols as present on the
1556 system, regardless of whether they actually exist. This has been fixed.
1557 [GH #15322] <https://github.com/Perl/perl5/issues/15322>
1558 .IP \(bu 4
1559 The \fIt/test.pl\fR library is used for internal testing of Perl itself, and
1560 also copied by several CPAN modules. Some of those modules must work on
1561 older versions of Perl, so \fIt/test.pl\fR must in turn avoid newer Perl
1562 features. Compatibility with Perl 5.8 was inadvertently removed some time
1563 ago; it has now been restored.
1564 [GH #15302] <https://github.com/Perl/perl5/issues/15302>
1565 .IP \(bu 4
1566 The build process no longer emits an extra blank line before building each
1567 "simple" extension (those with only \fI*.pm\fR and \fI*.pod\fR files).
1568 .SH Testing
1569 .IX Header "Testing"
1570 Tests were added and changed to reflect the other additions and changes
1571 in this release. Furthermore, these substantive changes were made:
1572 .IP \(bu 4
1573 A new test script, \fIcomp/parser_run.t\fR, has been added that is like
1574 \&\fIcomp/parser.t\fR but with \fItest.pl\fR included so that \f(CWrunperl()\fR and the
1575 like are available for use.
1576 .IP \(bu 4
1577 Tests for locales were erroneously using locales incompatible with Perl.
1578 .IP \(bu 4
1579 Some parts of the test suite that try to exhaustively test edge cases in the
1580 regex implementation have been restricted to running for a maximum of five
1581 minutes. On slow systems they could otherwise take several hours, without
1582 significantly improving our understanding of the correctness of the code
1583 under test.
1584 .IP \(bu 4
1585 A new internal facility allows analysing the time taken by the individual
1586 tests in Perl's own test suite; see \fIPorting/harness\-timer\-report.pl\fR.
1587 .IP \(bu 4
1588 \&\fIt/re/regexp_nonull.t\fR has been added to test that the regular expression
1589 engine can handle scalars that do not have a null byte just past the end of
1590 the string.
1591 .IP \(bu 4
1592 A new test script, \fIt/op/decl\-refs.t\fR, has been added to test the new feature
1593 "Declaring a reference to a variable".
1594 .IP \(bu 4
1595 A new test script, \fIt/re/keep_tabs.t\fR has been added to contain tests
1596 where \f(CW\*(C`\et\*(C'\fR characters should not be expanded into spaces.
1597 .IP \(bu 4
1598 A new test script, \fIt/re/anyof.t\fR, has been added to test that the ANYOF nodes
1599 generated by bracketed character classes are as expected.
1600 .IP \(bu 4
1601 There is now more extensive testing of the Unicode-related API macros
1602 and functions.
1603 .IP \(bu 4
1604 Several of the longer running API test files have been split into
1605 multiple test files so that they can be run in parallel.
1606 .IP \(bu 4
1607 \&\fIt/harness\fR now tries really hard not to run tests which are located
1608 outside of the Perl source tree.
1609 [GH #14578] <https://github.com/Perl/perl5/issues/14578>
1610 .IP \(bu 4
1611 Prevent debugger tests (\fIlib/perl5db.t\fR) from failing due to the contents
1612 of \f(CW$ENV{PERLDB_OPTS}\fR.
1613 [GH #15782] <https://github.com/Perl/perl5/issues/15782>
1614 .SH "Platform Support"
1615 .IX Header "Platform Support"
1616 .SS "New Platforms"
1617 .IX Subsection "New Platforms"
1618 .IP NetBSD/VAX 4
1619 .IX Item "NetBSD/VAX"
1620 Perl now compiles under NetBSD on VAX machines. However, it's not
1621 possible for that platform to implement floating-point infinities and
1622 NaNs compatible with most modern systems, which implement the IEEE\-754
1623 floating point standard. The hexadecimal floating point (\f(CW\*(C`0x...p[+\-]n\*(C'\fR
1624 literals, \f(CW\*(C`printf %a\*(C'\fR) is not implemented, either.
1625 The \f(CW\*(C`make test\*(C'\fR passes 98% of tests.
1626 .RS 4
1627 .IP \(bu 4
1628 Test fixes and minor updates.
1629 .IP \(bu 4
1630 Account for lack of \f(CW\*(C`inf\*(C'\fR, \f(CW\*(C`nan\*(C'\fR, and \f(CW\-0.0\fR support.
1631 .RE
1632 .RS 4
1633 .RE
1634 .SS "Platform-Specific Notes"
1635 .IX Subsection "Platform-Specific Notes"
1636 .IP Darwin 4
1637 .IX Item "Darwin"
1638 .RS 4
1639 .PD 0
1640 .IP \(bu 4
1641 .PD
1642 Don't treat \f(CW\*(C`\-Dprefix=/usr\*(C'\fR as special: instead require an extra option
1643 \&\f(CW\*(C`\-Ddarwin_distribution\*(C'\fR to produce the same results.
1644 .IP \(bu 4
1645 OS X El Capitan doesn't implement the \f(CWclock_gettime()\fR or
1646 \&\f(CWclock_getres()\fR APIs; emulate them as necessary.
1647 .IP \(bu 4
1648 Deprecated \f(CWsyscall(2)\fR on macOS 10.12.
1649 .RE
1650 .RS 4
1651 .RE
1652 .IP EBCDIC 4
1653 .IX Item "EBCDIC"
1654 Several tests have been updated to work (or be skipped) on EBCDIC platforms.
1655 .IP HP-UX 4
1656 .IX Item "HP-UX"
1657 The Net::Ping UDP test is now skipped on HP-UX.
1658 .IP Hurd 4
1659 .IX Item "Hurd"
1660 The hints for Hurd have been improved, enabling malloc wrap and reporting the
1661 GNU libc used (previously it was an empty string when reported).
1662 .IP VAX 4
1663 .IX Item "VAX"
1664 VAX floating point formats are now supported on NetBSD.
1665 .IP VMS 4
1666 .IX Item "VMS"
1667 .RS 4
1668 .PD 0
1669 .IP \(bu 4
1670 .PD
1671 The path separator for the \f(CW\*(C`PERL5LIB\*(C'\fR and \f(CW\*(C`PERLLIB\*(C'\fR environment entries is
1672 now a colon (\f(CW":"\fR) when running under a Unix shell. There is no change when
1673 running under DCL (it's still \f(CW"|"\fR).
1674 .IP \(bu 4
1675 \&\fIconfigure.com\fR now recognizes the VSI-branded C compiler and no longer
1676 recognizes the "DEC"\-branded C compiler (as there hasn't been such a thing for
1677 15 or more years).
1678 .RE
1679 .RS 4
1680 .RE
1681 .IP Windows 4
1682 .IX Item "Windows"
1683 .RS 4
1684 .PD 0
1685 .IP \(bu 4
1686 .PD
1687 Support for compiling perl on Windows using Microsoft Visual Studio 2015
1688 (containing Visual C++ 14.0) has been added.
1689 .Sp
1690 This version of VC++ includes a completely rewritten C run-time library, some
1691 of the changes in which mean that work done to resolve a socket
1692 \&\f(CWclose()\fR bug in
1693 perl #120091 and perl #118059 is not workable in its current state with this
1694 version of VC++. Therefore, we have effectively reverted that bug fix for
1695 VS2015 onwards on the basis that being able to build with VS2015 onwards is
1696 more important than keeping the bug fix. We may revisit this in the future to
1697 attempt to fix the bug again in a way that is compatible with VS2015.
1698 .Sp
1699 These changes do not affect compilation with GCC or with Visual Studio versions
1700 up to and including VS2013, \fIi.e.\fR, the bug fix is retained (unchanged) for those
1701 compilers.
1702 .Sp
1703 Note that you may experience compatibility problems if you mix a perl built
1704 with GCC or VS <= VS2013 with XS modules built with VS2015, or if you mix a
1705 perl built with VS2015 with XS modules built with GCC or VS <= VS2013.
1706 Some incompatibility may arise because of the bug fix that has been reverted
1707 for VS2015 builds of perl, but there may well be incompatibility anyway because
1708 of the rewritten CRT in VS2015 (\fIe.g.\fR, see discussion at
1709 <http://stackoverflow.com/questions/30412951>).
1710 .IP \(bu 4
1711 It now automatically detects GCC versus Visual C and sets the VC version
1712 number on Win32.
1713 .RE
1714 .RS 4
1715 .RE
1716 .IP Linux 4
1717 .IX Item "Linux"
1718 Drop support for Linux \fIa.out\fR executable format. Linux has used ELF for
1719 over twenty years.
1720 .IP "OpenBSD 6" 4
1721 .IX Item "OpenBSD 6"
1722 OpenBSD 6 still does not support returning \f(CW\*(C`pid\*(C'\fR, \f(CW\*(C`gid\*(C'\fR, or \f(CW\*(C`uid\*(C'\fR with
1723 \&\f(CW\*(C`SA_SIGINFO\*(C'\fR. Make sure to account for it.
1724 .IP FreeBSD 4
1725 .IX Item "FreeBSD"
1726 \&\fIt/uni/overload.t\fR: Skip hanging test on FreeBSD.
1727 .IP "DragonFly BSD" 4
1728 .IX Item "DragonFly BSD"
1729 DragonFly BSD now has support for \f(CWsetproctitle()\fR.
1730 [GH #15703] <https://github.com/Perl/perl5/issues/15703>.
1731 .SH "Internal Changes"
1732 .IX Header "Internal Changes"
1733 .IP \(bu 4
1734 A new API function \f(CWsv_setpv_bufsize()\fR
1735 allows simultaneously setting the
1736 length and the allocated size of the buffer in an \f(CW\*(C`SV\*(C'\fR, growing the
1737 buffer if necessary.
1738 .IP \(bu 4
1739 A new API macro \f(CWSvPVCLEAR()\fR sets its \f(CW\*(C`SV\*(C'\fR
1740 argument to an empty string,
1741 like Perl-space \f(CW\*(C`$x = \*(Aq\*(Aq\*(C'\fR, but with several optimisations.
1742 .IP \(bu 4
1743 Several new macros and functions for dealing with Unicode and
1744 UTF\-8\-encoded strings have been added to the API, as well as some
1745 changes in the
1746 functionality of existing functions (see "Unicode Support" in perlapi for
1747 more details):
1748 .RS 4
1749 .IP \(bu 4
1750 New versions of the API macros like \f(CW\*(C`isALPHA_utf8\*(C'\fR and \f(CW\*(C`toLOWER_utf8\*(C'\fR
1751 have been added, each with the suffix \f(CW\*(C`_safe\*(C'\fR, like
1752 \&\f(CW\*(C`isSPACE_utf8_safe\*(C'\fR. These take an extra
1753 parameter, giving an upper
1754 limit of how far into the string it is safe to read. Using the old
1755 versions could cause attempts to read beyond the end of the input buffer
1756 if the UTF\-8 is not well-formed, and their use now raises a deprecation
1757 warning. Details are at "Character classification" in perlapi.
1758 .IP \(bu 4
1759 Macros like \f(CW\*(C`isALPHA_utf8\*(C'\fR and
1760 \&\f(CW\*(C`toLOWER_utf8\*(C'\fR now die if they detect
1761 that their input UTF\-8 is malformed. A deprecation warning had been
1762 issued since Perl 5.18.
1763 .IP \(bu 4
1764 Several new macros for analysing the validity of utf8 sequences. These
1765 are:
1766 .Sp
1767 \&\f(CW\*(C`UTF8_GOT_ABOVE_31_BIT\*(C'\fR
1768 \&\f(CW\*(C`UTF8_GOT_CONTINUATION\*(C'\fR
1769 \&\f(CW\*(C`UTF8_GOT_EMPTY\*(C'\fR
1770 \&\f(CW\*(C`UTF8_GOT_LONG\*(C'\fR
1771 \&\f(CW\*(C`UTF8_GOT_NONCHAR\*(C'\fR
1772 \&\f(CW\*(C`UTF8_GOT_NON_CONTINUATION\*(C'\fR
1773 \&\f(CW\*(C`UTF8_GOT_OVERFLOW\*(C'\fR
1774 \&\f(CW\*(C`UTF8_GOT_SHORT\*(C'\fR
1775 \&\f(CW\*(C`UTF8_GOT_SUPER\*(C'\fR
1776 \&\f(CW\*(C`UTF8_GOT_SURROGATE\*(C'\fR
1777 \&\f(CW\*(C`UTF8_IS_INVARIANT\*(C'\fR
1778 \&\f(CW\*(C`UTF8_IS_NONCHAR\*(C'\fR
1779 \&\f(CW\*(C`UTF8_IS_SUPER\*(C'\fR
1780 \&\f(CW\*(C`UTF8_IS_SURROGATE\*(C'\fR
1781 \&\f(CW\*(C`UVCHR_IS_INVARIANT\*(C'\fR
1782 \&\f(CW\*(C`isUTF8_CHAR_flags\*(C'\fR
1783 \&\f(CW\*(C`isSTRICT_UTF8_CHAR\*(C'\fR
1784 \&\f(CW\*(C`isC9_STRICT_UTF8_CHAR\*(C'\fR
1785 .IP \(bu 4
1786 Functions that are all extensions of the \f(CW\*(C`is_utf8_string_\fR\f(CI*\fR\f(CW()\*(C'\fR functions,
1787 that apply various restrictions to the UTF\-8 recognized as valid:
1788 .Sp
1789 \&\f(CW\*(C`is_strict_utf8_string\*(C'\fR,
1790 \&\f(CW\*(C`is_strict_utf8_string_loc\*(C'\fR,
1791 \&\f(CW\*(C`is_strict_utf8_string_loclen\*(C'\fR,
1792 .Sp
1793 \&\f(CW\*(C`is_c9strict_utf8_string\*(C'\fR,
1794 \&\f(CW\*(C`is_c9strict_utf8_string_loc\*(C'\fR,
1795 \&\f(CW\*(C`is_c9strict_utf8_string_loclen\*(C'\fR,
1796 .Sp
1797 \&\f(CW\*(C`is_utf8_string_flags\*(C'\fR,
1798 \&\f(CW\*(C`is_utf8_string_loc_flags\*(C'\fR,
1799 \&\f(CW\*(C`is_utf8_string_loclen_flags\*(C'\fR,
1800 .Sp
1801 \&\f(CW\*(C`is_utf8_fixed_width_buf_flags\*(C'\fR,
1802 \&\f(CW\*(C`is_utf8_fixed_width_buf_loc_flags\*(C'\fR,
1803 \&\f(CW\*(C`is_utf8_fixed_width_buf_loclen_flags\*(C'\fR.
1804 .Sp
1805 \&\f(CW\*(C`is_utf8_invariant_string\*(C'\fR.
1806 \&\f(CW\*(C`is_utf8_valid_partial_char\*(C'\fR.
1807 \&\f(CW\*(C`is_utf8_valid_partial_char_flags\*(C'\fR.
1808 .IP \(bu 4
1809 The functions \f(CW\*(C`utf8n_to_uvchr\*(C'\fR and its
1810 derivatives have had several changes of behaviour.
1811 .Sp
1812 Calling them, while passing a string length of 0 is now asserted against
1813 in DEBUGGING builds, and otherwise, returns the Unicode REPLACEMENT
1814 CHARACTER. If you have nothing to decode, you shouldn't call the decode
1815 function.
1816 .Sp
1817 They now return the Unicode REPLACEMENT CHARACTER if called with UTF\-8
1818 that has the overlong malformation and that malformation is allowed by
1819 the input parameters. This malformation is where the UTF\-8 looks valid
1820 syntactically, but there is a shorter sequence that yields the same code
1821 point. This has been forbidden since Unicode version 3.1.
1822 .Sp
1823 They now accept an input
1824 flag to allow the overflow malformation. This malformation is when the
1825 UTF\-8 may be syntactically valid, but the code point it represents is
1826 not capable of being represented in the word length on the platform.
1827 What "allowed" means, in this case, is that the function doesn't return an
1828 error, and it advances the parse pointer to beyond the UTF\-8 in
1829 question, but it returns the Unicode REPLACEMENT CHARACTER as the value
1830 of the code point (since the real value is not representable).
1831 .Sp
1832 They no longer abandon searching for other malformations when the first
1833 one is encountered. A call to one of these functions thus can generate
1834 multiple diagnostics, instead of just one.
1835 .IP \(bu 4
1836 \&\f(CWvalid_utf8_to_uvchr()\fR has been added
1837 to the API (although it was
1838 present in core earlier). Like \f(CWutf8_to_uvchr_buf()\fR, but assumes that
1839 the next character is well-formed. Use with caution.
1840 .IP \(bu 4
1841 A new function, \f(CW\*(C`utf8n_to_uvchr_error\*(C'\fR,
1842 has been added for
1843 use by modules that need to know the details of UTF\-8 malformations
1844 beyond pass/fail. Previously, the only ways to know why a sequence was
1845 ill-formed was to capture and parse the generated diagnostics or to do
1846 your own analysis.
1847 .IP \(bu 4
1848 There is now a safer version of \fButf8_hop()\fR, called
1849 \&\f(CWutf8_hop_safe()\fR.
1850 Unlike \fButf8_hop()\fR, \fButf8_hop_safe()\fR won't navigate before the beginning or
1851 after the end of the supplied buffer.
1852 .IP \(bu 4
1853 Two new functions, \f(CWutf8_hop_forward()\fR and
1854 \&\f(CWutf8_hop_back()\fR are
1855 similar to \f(CWutf8_hop_safe()\fR but are for when you know which direction
1856 you wish to travel.
1857 .IP \(bu 4
1858 Two new macros which return useful utf8 byte sequences:
1859 .Sp
1860 \&\f(CW\*(C`BOM_UTF8\*(C'\fR
1861 .Sp
1862 \&\f(CW\*(C`REPLACEMENT_CHARACTER_UTF8\*(C'\fR
1863 .RE
1864 .RS 4
1865 .RE
1866 .IP \(bu 4
1867 Perl is now built with the \f(CW\*(C`PERL_OP_PARENT\*(C'\fR compiler define enabled by
1868 default. To disable it, use the \f(CW\*(C`PERL_NO_OP_PARENT\*(C'\fR compiler define.
1869 This flag alters how the \f(CW\*(C`op_sibling\*(C'\fR field is used in \f(CW\*(C`OP\*(C'\fR structures,
1870 and has been available optionally since perl 5.22.
1871 .Sp
1872 See "Internal Changes" in perl5220delta for more details of what this
1873 build option does.
1874 .IP \(bu 4
1875 Three new ops, \f(CW\*(C`OP_ARGELEM\*(C'\fR, \f(CW\*(C`OP_ARGDEFELEM\*(C'\fR, and \f(CW\*(C`OP_ARGCHECK\*(C'\fR have
1876 been added. These are intended principally to implement the individual
1877 elements of a subroutine signature, plus any overall checking required.
1878 .IP \(bu 4
1879 The \f(CW\*(C`OP_PUSHRE\*(C'\fR op has been eliminated and the \f(CW\*(C`OP_SPLIT\*(C'\fR op has been
1880 changed from class \f(CW\*(C`LISTOP\*(C'\fR to \f(CW\*(C`PMOP\*(C'\fR.
1881 .Sp
1882 Formerly the first child of a split would be a \f(CW\*(C`pushre\*(C'\fR, which would have the
1883 \&\f(CW\*(C`split\*(C'\fR's regex attached to it. Now the regex is attached directly to the
1884 \&\f(CW\*(C`split\*(C'\fR op, and the \f(CW\*(C`pushre\*(C'\fR has been eliminated.
1885 .IP \(bu 4
1886 The \f(CWop_class()\fR API function has been added. This
1887 is like the existing
1888 \&\f(CWOP_CLASS()\fR macro, but can more accurately determine what struct an op
1889 has been allocated as. For example \f(CWOP_CLASS()\fR might return
1890 \&\f(CW\*(C`OA_BASEOP_OR_UNOP\*(C'\fR indicating that ops of this type are usually
1891 allocated as an \f(CW\*(C`OP\*(C'\fR or \f(CW\*(C`UNOP\*(C'\fR; while \f(CWop_class()\fR will return
1892 \&\f(CW\*(C`OPclass_BASEOP\*(C'\fR or \f(CW\*(C`OPclass_UNOP\*(C'\fR as appropriate.
1893 .IP \(bu 4
1894 All parts of the internals now agree that the \f(CW\*(C`sassign\*(C'\fR op is a \f(CW\*(C`BINOP\*(C'\fR;
1895 previously it was listed as a \f(CW\*(C`BASEOP\*(C'\fR in \fIregen/opcodes\fR, which meant
1896 that several parts of the internals had to be special-cased to accommodate
1897 it. This oddity's original motivation was to handle code like \f(CW\*(C`$x ||= 1\*(C'\fR;
1898 that is now handled in a simpler way.
1899 .IP \(bu 4
1900 The output format of the \f(CWop_dump()\fR function (as
1901 used by \f(CW\*(C`perl \-Dx\*(C'\fR)
1902 has changed: it now displays an "ASCII-art" tree structure, and shows more
1903 low-level details about each op, such as its address and class.
1904 .IP \(bu 4
1905 The \f(CW\*(C`PADOFFSET\*(C'\fR type has changed from being unsigned to signed, and
1906 several pad-related variables such as \f(CW\*(C`PL_padix\*(C'\fR have changed from being
1907 of type \f(CW\*(C`I32\*(C'\fR to type \f(CW\*(C`PADOFFSET\*(C'\fR.
1908 .IP \(bu 4
1909 The \f(CW\*(C`DEBUGGING\*(C'\fR\-mode output for regex compilation and execution has been
1910 enhanced.
1911 .IP \(bu 4
1912 Several obscure SV flags have been eliminated, sometimes along with the
1913 macros which manipulate them: \f(CW\*(C`SVpbm_VALID\*(C'\fR, \f(CW\*(C`SVpbm_TAIL\*(C'\fR, \f(CW\*(C`SvTAIL_on\*(C'\fR,
1914 \&\f(CW\*(C`SvTAIL_off\*(C'\fR, \f(CW\*(C`SVrepl_EVAL\*(C'\fR, \f(CW\*(C`SvEVALED\*(C'\fR.
1915 .IP \(bu 4
1916 An OP \f(CW\*(C`op_private\*(C'\fR flag has been eliminated: \f(CW\*(C`OPpRUNTIME\*(C'\fR. This used to
1917 often get set on \f(CW\*(C`PMOP\*(C'\fR ops, but had become meaningless over time.
1918 .SH "Selected Bug Fixes"
1919 .IX Header "Selected Bug Fixes"
1920 .IP \(bu 4
1921 Perl no longer panics when switching into some locales on machines with
1922 buggy \f(CWstrxfrm()\fR implementations in their \fIlibc\fR.
1923 [GH #13768] <https://github.com/Perl/perl5/issues/13768>
1924 .IP \(bu 4
1925 \&\f(CW\*(C` $\-{$name} \*(C'\fR would leak an \f(CW\*(C`AV\*(C'\fR on each access if the regular
1926 expression had no named captures. The same applies to access to any
1927 hash tied with Tie::Hash::NamedCapture and \f(CW\*(C`all => 1\*(C'\fR.
1928 [GH #15882] <https://github.com/Perl/perl5/issues/15882>
1929 .IP \(bu 4
1930 Attempting to use the deprecated variable \f(CW$#\fR as the object in an
1931 indirect object method call could cause a heap use after free or
1932 buffer overflow.
1933 [GH #15599] <https://github.com/Perl/perl5/issues/15599>
1934 .IP \(bu 4
1935 When checking for an indirect object method call, in some rare cases
1936 the parser could reallocate the line buffer but then continue to use
1937 pointers to the old buffer.
1938 [GH #15585] <https://github.com/Perl/perl5/issues/15585>
1939 .IP \(bu 4
1940 Supplying a glob as the format argument to
1941 \&\f(CW\*(C`formline\*(C'\fR would
1942 cause an assertion failure.
1943 [GH #15862] <https://github.com/Perl/perl5/issues/15862>
1944 .IP \(bu 4
1945 Code like \f(CW\*(C` $value1 =~ qr/.../ ~~ $value2 \*(C'\fR would have the match
1946 converted into a \f(CW\*(C`qr//\*(C'\fR operator, leaving extra elements on the stack to
1947 confuse any surrounding expression.
1948 [GH #15859] <https://github.com/Perl/perl5/issues/15859>
1949 .IP \(bu 4
1950 Since v5.24 in some obscure cases, a regex which included code blocks
1951 from multiple sources (\fIe.g.\fR, via embedded via \f(CW\*(C`qr//\*(C'\fR objects) could end up
1952 with the wrong current pad and crash or give weird results.
1953 [GH #15657] <https://github.com/Perl/perl5/issues/15657>
1954 .IP \(bu 4
1955 Occasionally \f(CWlocal()\fRs in a code block within a patterns weren't being
1956 undone when the pattern matching backtracked over the code block.
1957 [GH #15056] <https://github.com/Perl/perl5/issues/15056>
1958 .IP \(bu 4
1959 Using \f(CWsubstr()\fR to modify a magic variable could access freed memory
1960 in some cases.
1961 [GH #15871] <https://github.com/Perl/perl5/issues/15871>
1962 .IP \(bu 4
1963 Under \f(CW\*(C`use utf8\*(C'\fR, the entire source code is now checked for being UTF\-8
1964 well formed, not just quoted strings as before.
1965 [GH #14973] <https://github.com/Perl/perl5/issues/14973>.
1966 .IP \(bu 4
1967 The range operator \f(CW".."\fR on strings now handles its arguments correctly when in
1968 the scope of the \f(CW\*(C`unicode_strings\*(C'\fR
1969 feature. The previous behaviour was sufficiently unexpected that we believe no
1970 correct program could have made use of it.
1971 .IP \(bu 4
1972 The \f(CW\*(C`split\*(C'\fR operator did not ensure enough space was allocated for
1973 its return value in scalar context. It could then write a single
1974 pointer immediately beyond the end of the memory block allocated for
1975 the stack.
1976 [GH #15749] <https://github.com/Perl/perl5/issues/15749>
1977 .IP \(bu 4
1978 Using a large code point with the \f(CW"W"\fR pack template character with
1979 the current output position aligned at just the right point could
1980 cause a write of a single zero byte immediately beyond the end of an
1981 allocated buffer.
1982 [GH #15572] <https://github.com/Perl/perl5/issues/15572>
1983 .IP \(bu 4
1984 Supplying a format's picture argument as part of the format argument list
1985 where the picture specifies modifying the argument could cause an
1986 access to the new freed compiled format.
1987 [GH #15566] <https://github.com/Perl/perl5/issues/15566>
1988 .IP \(bu 4
1989 The \fBsort()\fR operator's built-in numeric comparison
1990 function didn't handle large integers that weren't exactly
1991 representable by a double. This now uses the same code used to
1992 implement the \f(CW\*(C`<=>\*(C'\fR operator.
1993 [GH #15768] <https://github.com/Perl/perl5/issues/15768>
1994 .IP \(bu 4
1995 Fix issues with \f(CW\*(C`/(?{ ... <<EOF })/\*(C'\fR that broke
1996 Method::Signatures.
1997 [GH #15779] <https://github.com/Perl/perl5/issues/15779>
1998 .IP \(bu 4
1999 Fixed an assertion failure with \f(CW\*(C`chop\*(C'\fR and \f(CW\*(C`chomp\*(C'\fR, which
2000 could be triggered by \f(CW\*(C`chop(@x =~ tr/1/1/)\*(C'\fR.
2001 [GH #15738] <https://github.com/Perl/perl5/issues/15738>.
2002 .IP \(bu 4
2003 Fixed a comment skipping error in patterns under \f(CW\*(C`/x\*(C'\fR; it could stop
2004 skipping a byte early, which could be in the middle of a UTF\-8
2005 character.
2006 [GH #15790] <https://github.com/Perl/perl5/issues/15790>.
2007 .IP \(bu 4
2008 \&\fIperldb\fR now ignores \fI/dev/tty\fR on non-Unix systems.
2009 [GH #12244] <https://github.com/Perl/perl5/issues/12244>;
2010 .IP \(bu 4
2011 Fix assertion failure for \f(CW\*(C`{}\->$x\*(C'\fR when \f(CW$x\fR isn't defined.
2012 [GH #15791] <https://github.com/Perl/perl5/issues/15791>.
2013 .IP \(bu 4
2014 Fix an assertion error which could be triggered when a lookahead string
2015 in patterns exceeded a minimum length.
2016 [GH #15796] <https://github.com/Perl/perl5/issues/15796>.
2017 .IP \(bu 4
2018 Only warn once per literal number about a misplaced \f(CW"_"\fR.
2019 [GH #9989] <https://github.com/Perl/perl5/issues/9989>.
2020 .IP \(bu 4
2021 The \f(CW\*(C`tr///\*(C'\fR parse code could be looking at uninitialized data after a
2022 perse error.
2023 [GH #15624] <https://github.com/Perl/perl5/issues/15624>.
2024 .IP \(bu 4
2025 In a pattern match, a back-reference (\f(CW\*(C`\e1\*(C'\fR) to an unmatched capture could
2026 read back beyond the start of the string being matched.
2027 [GH #15634] <https://github.com/Perl/perl5/issues/15634>.
2028 .IP \(bu 4
2029 \&\f(CW\*(C`use re \*(Aqstrict\*(Aq\*(C'\fR is supposed to warn if you use a range (such as
2030 \&\f(CW\*(C`/(?[ [ X\-Y ] ])/\*(C'\fR) whose start and end digit aren't from the same group
2031 of 10. It didn't do that for five groups of mathematical digits starting
2032 at \f(CW\*(C`U+1D7E\*(C'\fR.
2033 .IP \(bu 4
2034 A sub containing a "forward" declaration with the same name (\fIe.g.\fR,
2035 \&\f(CW\*(C`sub c { sub c; }\*(C'\fR) could sometimes crash or loop infinitely.
2036 [GH #15557] <https://github.com/Perl/perl5/issues/15557>
2037 .IP \(bu 4
2038 A crash in executing a regex with a non-anchored UTF\-8 substring against a
2039 target string that also used UTF\-8 has been fixed.
2040 [GH #15628] <https://github.com/Perl/perl5/issues/15628>
2041 .IP \(bu 4
2042 Previously, a shebang line like \f(CW\*(C`#!perl \-i u\*(C'\fR could be erroneously
2043 interpreted as requesting the \f(CW\*(C`\-u\*(C'\fR option. This has been fixed.
2044 [GH #15623] <https://github.com/Perl/perl5/issues/15623>
2045 .IP \(bu 4
2046 The regex engine was previously producing incorrect results in some rare
2047 situations when backtracking past an alternation that matches only one
2048 thing; this
2049 showed up as capture buffers (\f(CW$1\fR, \f(CW$2\fR, \fIetc.\fR) erroneously containing data
2050 from regex execution paths that weren't actually executed for the final
2051 match.
2052 [GH #15666] <https://github.com/Perl/perl5/issues/15666>
2053 .IP \(bu 4
2054 Certain regexes making use of the experimental \f(CW\*(C`regex_sets\*(C'\fR feature could
2055 trigger an assertion failure. This has been fixed.
2056 [GH #15620] <https://github.com/Perl/perl5/issues/15620>
2057 .IP \(bu 4
2058 Invalid assignments to a reference constructor (\fIe.g.\fR, \f(CW\*(C`\eeval=time\*(C'\fR) could
2059 sometimes crash in addition to giving a syntax error.
2060 [GH #14815] <https://github.com/Perl/perl5/issues/14815>
2061 .IP \(bu 4
2062 The parser could sometimes crash if a bareword came after \f(CW\*(C`evalbytes\*(C'\fR.
2063 [GH #15586] <https://github.com/Perl/perl5/issues/15586>
2064 .IP \(bu 4
2065 Autoloading via a method call would warn erroneously ("Use of inherited
2066 AUTOLOAD for non-method") if there was a stub present in the package into
2067 which the invocant had been blessed. The warning is no longer emitted in
2068 such circumstances.
2069 [GH #9094] <https://github.com/Perl/perl5/issues/9094>
2070 .IP \(bu 4
2071 The use of \f(CW\*(C`splice\*(C'\fR on arrays with non-existent elements could cause other
2072 operators to crash.
2073 [GH #15577] <https://github.com/Perl/perl5/issues/15577>
2074 .IP \(bu 4
2075 A possible buffer overrun when a pattern contains a fixed utf8 substring.
2076 [GH #15534] <https://github.com/Perl/perl5/issues/15534>
2077 .IP \(bu 4
2078 Fixed two possible use-after-free bugs in perl's lexer.
2079 [GH #15549] <https://github.com/Perl/perl5/issues/15549>
2080 .IP \(bu 4
2081 Fixed a crash with \f(CW\*(C`s///l\*(C'\fR where it thought it was dealing with UTF\-8
2082 when it wasn't.
2083 [GH #15543] <https://github.com/Perl/perl5/issues/15543>
2084 .IP \(bu 4
2085 Fixed a place where the regex parser was not setting the syntax error
2086 correctly on a syntactically incorrect pattern.
2087 [GH #15565] <https://github.com/Perl/perl5/issues/15565>
2088 .IP \(bu 4
2089 The \f(CW\*(C`&.\*(C'\fR operator (and the \f(CW"&"\fR operator, when it treats its arguments as
2090 strings) were failing to append a trailing null byte if at least one string
2091 was marked as utf8 internally. Many code paths (system calls, regexp
2092 compilation) still expect there to be a null byte in the string buffer
2093 just past the end of the logical string. An assertion failure was the
2094 result.
2095 [GH #15606] <https://github.com/Perl/perl5/issues/15606>
2096 .IP \(bu 4
2097 Avoid a heap-after-use error in the parser when creating an error messge
2098 for a syntactically invalid heredoc.
2099 [GH #15527] <https://github.com/Perl/perl5/issues/15527>
2100 .IP \(bu 4
2101 Fix a segfault when run with \f(CW\*(C`\-DC\*(C'\fR options on DEBUGGING builds.
2102 [GH #15563] <https://github.com/Perl/perl5/issues/15563>
2103 .IP \(bu 4
2104 Fixed the parser error handling in subroutine attributes for an
2105 \&'\f(CW\*(C`:attr(foo\*(C'\fR' that does not have an ending '\f(CW")"\fR'.
2106 .IP \(bu 4
2107 Fix the perl lexer to correctly handle a backslash as the last char in
2108 quoted-string context. This actually fixed two bugs,
2109 [GH #15546] <https://github.com/Perl/perl5/issues/15546> and
2110 [GH #15582] <https://github.com/Perl/perl5/issues/15582>.
2111 .IP \(bu 4
2112 In the API function \f(CW\*(C`gv_fetchmethod_pvn_flags\*(C'\fR, rework separator parsing
2113 to prevent possible string overrun with an invalid \f(CW\*(C`len\*(C'\fR argument.
2114 [GH #15598] <https://github.com/Perl/perl5/issues/15598>
2115 .IP \(bu 4
2116 Problems with in-place array sorts: code like \f(CW\*(C`@a = sort { ... } @a\*(C'\fR,
2117 where the source and destination of the sort are the same plain array, are
2118 optimised to do less copying around. Two side-effects of this optimisation
2119 were that the contents of \f(CW@a\fR as seen by sort routines were
2120 partially sorted; and under some circumstances accessing \f(CW@a\fR during the
2121 sort could crash the interpreter. Both these issues have been fixed, and
2122 Sort functions see the original value of \f(CW@a\fR.
2123 [GH #15387] <https://github.com/Perl/perl5/issues/15387>
2124 .IP \(bu 4
2125 Non-ASCII string delimiters are now reported correctly in error messages
2126 for unterminated strings.
2127 [GH #15469] <https://github.com/Perl/perl5/issues/15469>
2128 .IP \(bu 4
2129 \&\f(CW\*(C`pack("p", ...)\*(C'\fR used to emit its warning ("Attempt to pack pointer to
2130 temporary value") erroneously in some cases, but has been fixed.
2131 .IP \(bu 4
2132 \&\f(CW@DB::args\fR is now exempt from "used once" warnings. The warnings only
2133 occurred under \fB\-w\fR, because \fIwarnings.pm\fR itself uses \f(CW@DB::args\fR
2134 multiple times.
2135 .IP \(bu 4
2136 The use of built-in arrays or hash slices in a double-quoted string no
2137 longer issues a warning ("Possible unintended interpolation...") if the
2138 variable has not been mentioned before. This affected code like
2139 \&\f(CW\*(C`qq|@DB::args|\*(C'\fR and \f(CW\*(C`qq|@SIG{\*(AqCHLD\*(Aq, \*(AqHUP\*(Aq}|\*(C'\fR. (The special variables
2140 \&\f(CW\*(C`@\-\*(C'\fR and \f(CW\*(C`@+\*(C'\fR were already exempt from the warning.)
2141 .IP \(bu 4
2142 \&\f(CW\*(C`gethostent\*(C'\fR and similar functions now perform a null check internally, to
2143 avoid crashing with the torsocks library. This was a regression from v5.22.
2144 [GH #15478] <https://github.com/Perl/perl5/issues/15478>
2145 .IP \(bu 4
2146 \&\f(CW\*(C`defined *{\*(Aq!\*(Aq}\*(C'\fR, \f(CW\*(C`defined *{\*(Aq[\*(Aq}\*(C'\fR, and \f(CW\*(C`defined *{\*(Aq\-\*(Aq}\*(C'\fR no longer leak
2147 memory if the typeglob in question has never been accessed before.
2148 .IP \(bu 4
2149 Mentioning the same constant twice in a row (which is a syntax error) no
2150 longer fails an assertion under debugging builds. This was a regression
2151 from v5.20.
2152 [GH #15017] <https://github.com/Perl/perl5/issues/15017>
2153 .IP \(bu 4
2154 Many issues relating to \f(CW\*(C`printf "%a"\*(C'\fR of hexadecimal floating point
2155 were fixed. In addition, the "subnormals" (formerly known as "denormals")
2156 floating point numbers are now supported both with the plain IEEE 754
2157 floating point numbers (64\-bit or 128\-bit) and the x86 80\-bit
2158 "extended precision". Note that subnormal hexadecimal floating
2159 point literals will give a warning about "exponent underflow".
2160 [GH #15495] <https://github.com/Perl/perl5/issues/15495>
2161 [GH #15503] <https://github.com/Perl/perl5/issues/15503>
2162 [GH #15504] <https://github.com/Perl/perl5/issues/15504>
2163 [GH #15505] <https://github.com/Perl/perl5/issues/15505>
2164 [GH #15510] <https://github.com/Perl/perl5/issues/15510>
2165 [GH #15512] <https://github.com/Perl/perl5/issues/15512>
2166 .IP \(bu 4
2167 A regression in v5.24 with \f(CW\*(C`tr/\eN{U+...}/foo/\*(C'\fR when the code point was between
2168 128 and 255 has been fixed.
2169 [GH #15475] <https://github.com/Perl/perl5/issues/15475>.
2170 .IP \(bu 4
2171 Use of a string delimiter whose code point is above 2**31 now works
2172 correctly on platforms that allow this. Previously, certain characters,
2173 due to truncation, would be confused with other delimiter characters
2174 with special meaning (such as \f(CW"?"\fR in \f(CW\*(C`m?...?\*(C'\fR), resulting
2175 in inconsistent behaviour. Note that this is non-portable,
2176 and is based on Perl's extension to UTF\-8, and is probably not
2177 displayable nor enterable by any editor.
2178 [GH #15477] <https://github.com/Perl/perl5/issues/15477>
2179 .IP \(bu 4
2180 \&\f(CW\*(C`@{x\*(C'\fR followed by a newline where \f(CW"x"\fR represents a control or non-ASCII
2181 character no longer produces a garbled syntax error message or a crash.
2182 [GH #15518] <https://github.com/Perl/perl5/issues/15518>
2183 .IP \(bu 4
2184 An assertion failure with \f(CW\*(C`%: = 0\*(C'\fR has been fixed.
2185 [GH #15358] <https://github.com/Perl/perl5/issues/15358>
2186 .IP \(bu 4
2187 In Perl 5.18, the parsing of \f(CW"$foo::$bar"\fR was accidentally changed, such
2188 that it would be treated as \f(CW\*(C`$foo."::".$bar\*(C'\fR. The previous behavior, which
2189 was to parse it as \f(CW\*(C`$foo:: . $bar\*(C'\fR, has been restored.
2190 [GH #15408] <https://github.com/Perl/perl5/issues/15408>
2191 .IP \(bu 4
2192 Since Perl 5.20, line numbers have been off by one when perl is invoked with
2193 the \fB\-x\fR switch. This has been fixed.
2194 [GH #15413] <https://github.com/Perl/perl5/issues/15413>
2195 .IP \(bu 4
2196 Vivifying a subroutine stub in a deleted stash (\fIe.g.\fR,
2197 \&\f(CW\*(C`delete $My::{"Foo::"}; \e&My::Foo::foo\*(C'\fR) no longer crashes. It had begun
2198 crashing in Perl 5.18.
2199 [GH #15420] <https://github.com/Perl/perl5/issues/15420>
2200 .IP \(bu 4
2201 Some obscure cases of subroutines and file handles being freed at the same time
2202 could result in crashes, but have been fixed. The crash was introduced in Perl
2203 5.22.
2204 [GH #15435] <https://github.com/Perl/perl5/issues/15435>
2205 .IP \(bu 4
2206 Code that looks for a variable name associated with an uninitialized value
2207 could cause an assertion failure in cases where magic is involved, such as
2208 \&\f(CW\*(C`$ISA[0][0]\*(C'\fR. This has now been fixed.
2209 [GH #15364] <https://github.com/Perl/perl5/issues/15364>
2210 .IP \(bu 4
2211 A crash caused by code generating the warning "Subroutine STASH::NAME
2212 redefined" in cases such as \f(CW\*(C`sub P::f{} undef *P::; *P::f =sub{};\*(C'\fR has been
2213 fixed. In these cases, where the STASH is missing, the warning will now appear
2214 as "Subroutine NAME redefined".
2215 [GH #15368] <https://github.com/Perl/perl5/issues/15368>
2216 .IP \(bu 4
2217 Fixed an assertion triggered by some code that handles deprecated behavior in
2218 formats, \fIe.g.\fR, in cases like this:
2219 .Sp
2220 .Vb 3
2221 \& format STDOUT =
2222 \& @
2223 \& 0"$x"
2224 .Ve
2225 .Sp
2226 [GH #15366] <https://github.com/Perl/perl5/issues/15366>
2227 .IP \(bu 4
2228 A possible divide by zero in string transformation code on Windows has been
2229 avoided, fixing a crash when collating an empty string.
2230 [GH #15439] <https://github.com/Perl/perl5/issues/15439>
2231 .IP \(bu 4
2232 Some regular expression parsing glitches could lead to assertion failures with
2233 regular expressions such as \f(CW\*(C`/(?<=/\*(C'\fR and \f(CW\*(C`/(?<!/\*(C'\fR. This has now been fixed.
2234 [GH #15332] <https://github.com/Perl/perl5/issues/15332>
2235 .IP \(bu 4
2236 \&\f(CW\*(C` until ($x = 1) { ... } \*(C'\fR and \f(CW\*(C` ... until $x = 1 \*(C'\fR now properly
2237 warn when syntax warnings are enabled.
2238 [GH #15138] <https://github.com/Perl/perl5/issues/15138>
2239 .IP \(bu 4
2240 \&\fBsocket()\fR now leaves the error code returned by the system in \f(CW$!\fR on
2241 failure.
2242 [GH #15383] <https://github.com/Perl/perl5/issues/15383>
2243 .IP \(bu 4
2244 Assignment variants of any bitwise ops under the \f(CW\*(C`bitwise\*(C'\fR feature would
2245 crash if the left-hand side was an array or hash.
2246 [GH #15346] <https://github.com/Perl/perl5/issues/15346>
2247 .IP \(bu 4
2248 \&\f(CW\*(C`require\*(C'\fR followed by a single colon (as in \f(CW\*(C`foo() ? require : ...\*(C'\fR is
2249 now parsed correctly as \f(CW\*(C`require\*(C'\fR with implicit \f(CW$_\fR, rather than
2250 \&\f(CW\*(C`require ""\*(C'\fR.
2251 [GH #15380] <https://github.com/Perl/perl5/issues/15380>
2252 .IP \(bu 4
2253 Scalar \f(CW\*(C`keys %hash\*(C'\fR can now be assigned to consistently in all scalar
2254 lvalue contexts. Previously it worked for some contexts but not others.
2255 .IP \(bu 4
2256 List assignment to \f(CW\*(C`vec\*(C'\fR or \f(CW\*(C`substr\*(C'\fR with an array or hash for its first
2257 argument used to result in crashes or "Can't coerce" error messages at run
2258 time, unlike scalar assignment, which would give an error at compile time.
2259 List assignment now gives a compile-time error, too.
2260 [GH #15370] <https://github.com/Perl/perl5/issues/15370>
2261 .IP \(bu 4
2262 Expressions containing an \f(CW\*(C`&&\*(C'\fR or \f(CW\*(C`||\*(C'\fR operator (or their synonyms \f(CW\*(C`and\*(C'\fR
2263 and \f(CW\*(C`or\*(C'\fR) were being compiled incorrectly in some cases. If the left-hand
2264 side consisted of either a negated bareword constant or a negated \f(CW\*(C`do {}\*(C'\fR
2265 block containing a constant expression, and the right-hand side consisted of
2266 a negated non-foldable expression, one of the negations was effectively
2267 ignored. The same was true of \f(CW\*(C`if\*(C'\fR and \f(CW\*(C`unless\*(C'\fR statement modifiers,
2268 though with the left-hand and right-hand sides swapped. This long-standing
2269 bug has now been fixed.
2270 [GH #15285] <https://github.com/Perl/perl5/issues/15285>
2271 .IP \(bu 4
2272 \&\f(CW\*(C`reset\*(C'\fR with an argument no longer crashes when encountering stash entries
2273 other than globs.
2274 [GH #15314] <https://github.com/Perl/perl5/issues/15314>
2275 .IP \(bu 4
2276 Assignment of hashes to, and deletion of, typeglobs named \f(CW*::::::\fR no
2277 longer causes crashes.
2278 [GH #15307] <https://github.com/Perl/perl5/issues/15307>
2279 .IP \(bu 4
2280 Perl wasn't correctly handling true/false values in the LHS of a list
2281 assign; specifically the truth values returned by boolean operators.
2282 This could trigger an assertion failure in something like the following:
2283 .Sp
2284 .Vb 3
2285 \& for ($x > $y) {
2286 \& ($_, ...) = (...); # here $_ is aliased to a truth value
2287 \& }
2288 .Ve
2289 .Sp
2290 This was a regression from v5.24.
2291 [GH #15690] <https://github.com/Perl/perl5/issues/15690>
2292 .IP \(bu 4
2293 Assertion failure with user-defined Unicode-like properties.
2294 [GH #15696] <https://github.com/Perl/perl5/issues/15696>
2295 .IP \(bu 4
2296 Fix error message for unclosed \f(CW\*(C`\eN{\*(C'\fR in a regex. An unclosed \f(CW\*(C`\eN{\*(C'\fR
2297 could give the wrong error message:
2298 \&\f(CW"\eN{NAME} must be resolved by the lexer"\fR.
2299 .IP \(bu 4
2300 List assignment in list context where the LHS contained aggregates and
2301 where there were not enough RHS elements, used to skip scalar lvalues.
2302 Previously, \f(CW\*(C`(($a,$b,@c,$d) = (1))\*(C'\fR in list context returned \f(CW\*(C`($a)\*(C'\fR; now
2303 it returns \f(CW\*(C`($a,$b,$d)\*(C'\fR. \f(CW\*(C`(($a,$b,$c) = (1))\*(C'\fR is unchanged: it still
2304 returns \f(CW\*(C`($a,$b,$c)\*(C'\fR. This can be seen in the following:
2305 .Sp
2306 .Vb 2
2307 \& sub inc { $_++ for @_ }
2308 \& inc(($a,$b,@c,$d) = (10))
2309 .Ve
2310 .Sp
2311 Formerly, the values of \f(CW\*(C`($a,$b,$d)\*(C'\fR would be left as \f(CW\*(C`(11,undef,undef)\*(C'\fR;
2312 now they are \f(CW\*(C`(11,1,1)\*(C'\fR.
2313 .IP \(bu 4
2314 Code like this: \f(CW\*(C`/(?{ s!!! })/\*(C'\fR could trigger infinite recursion on the C
2315 stack (not the normal perl stack) when the last successful pattern in
2316 scope is itself. We avoid the segfault by simply forbidding the use of
2317 the empty pattern when it would resolve to the currently executing
2318 pattern.
2319 [GH #15669] <https://github.com/Perl/perl5/issues/15669>
2320 .IP \(bu 4
2321 Avoid reading beyond the end of the line buffer in perl's lexer when
2322 there's a short UTF\-8 character at the end.
2323 [GH #15531] <https://github.com/Perl/perl5/issues/15531>
2324 .IP \(bu 4
2325 Alternations in regular expressions were sometimes failing to match
2326 a utf8 string against a utf8 alternate.
2327 [GH #15680] <https://github.com/Perl/perl5/issues/15680>
2328 .IP \(bu 4
2329 Make \f(CW\*(C`do "a\e0b"\*(C'\fR fail silently (and return \f(CW\*(C`undef\*(C'\fR and set \f(CW$!\fR)
2330 instead of throwing an error.
2331 [GH #15676] <https://github.com/Perl/perl5/issues/15676>
2332 .IP \(bu 4
2333 \&\f(CW\*(C`chdir\*(C'\fR with no argument didn't ensure that there was stack space
2334 available for returning its result.
2335 [GH #15569] <https://github.com/Perl/perl5/issues/15569>
2336 .IP \(bu 4
2337 All error messages related to \f(CW\*(C`do\*(C'\fR now refer to \f(CW\*(C`do\*(C'\fR; some formerly
2338 claimed to be from \f(CW\*(C`require\*(C'\fR instead.
2339 .IP \(bu 4
2340 Executing \f(CW\*(C`undef $x\*(C'\fR where \f(CW$x\fR is tied or magical no longer incorrectly
2341 blames the variable for an uninitialized-value warning encountered by the
2342 tied/magical code.
2343 .IP \(bu 4
2344 Code like \f(CW\*(C`$x = $x . "a"\*(C'\fR was incorrectly failing to yield a
2345 use of uninitialized value
2346 warning when \f(CW$x\fR was a lexical variable with an undefined value. That has
2347 now been fixed.
2348 [GH #15269] <https://github.com/Perl/perl5/issues/15269>
2349 .IP \(bu 4
2350 \&\f(CW\*(C`undef *_; shift\*(C'\fR or \f(CW\*(C`undef *_; pop\*(C'\fR inside a subroutine, with no
2351 argument to \f(CW\*(C`shift\*(C'\fR or \f(CW\*(C`pop\*(C'\fR, began crashing in Perl 5.14, but has now
2352 been fixed.
2353 .IP \(bu 4
2354 \&\f(CW"string$scalar\->$*"\fR now correctly prefers concatenation
2355 overloading to string overloading if \f(CW\*(C`$scalar\->$*\*(C'\fR returns an
2356 overloaded object, bringing it into consistency with \f(CW$$scalar\fR.
2357 .IP \(bu 4
2358 \&\f(CW\*(C`/@0{0*\->@*/*0\*(C'\fR and similar contortions used to crash, but no longer
2359 do, but merely produce a syntax error.
2360 [GH #15333] <https://github.com/Perl/perl5/issues/15333>
2361 .IP \(bu 4
2362 \&\f(CW\*(C`do\*(C'\fR or \f(CW\*(C`require\*(C'\fR with an argument which is a reference or typeglob
2363 which, when stringified,
2364 contains a null character, started crashing in Perl 5.20, but has now been
2365 fixed.
2366 [GH #15337] <https://github.com/Perl/perl5/issues/15337>
2367 .IP \(bu 4
2368 Improve the error message for a missing \f(CWtie()\fR package/method. This
2369 brings the error messages in line with the ones used for normal method
2370 calls.
2371 .IP \(bu 4
2372 Parsing bad POSIX charclasses no longer leaks memory.
2373 [GH #15382] <https://github.com/Perl/perl5/issues/15382>
2374 .SH "Known Problems"
2375 .IX Header "Known Problems"
2376 .IP \(bu 4
2377 G++ 6 handles subnormal (denormal) floating point values differently
2378 than gcc 6 or g++ 5 resulting in "flush-to-zero". The end result is
2379 that if you specify very small values using the hexadecimal floating
2380 point format, like \f(CW\*(C`0x1.fffffffffffffp\-1022\*(C'\fR, they become zeros.
2381 [GH #15990] <https://github.com/Perl/perl5/issues/15990>
2382 .SH "Errata From Previous Releases"
2383 .IX Header "Errata From Previous Releases"
2384 .IP \(bu 4
2385 Fixed issues with recursive regexes. The behavior was fixed in Perl 5.24.
2386 [GH #14935] <https://github.com/Perl/perl5/issues/14935>
2387 .SH Obituary
2388 .IX Header "Obituary"
2389 Jon Portnoy (AVENJ), a prolific Perl author and admired Gentoo community
2390 member, has passed away on August 10, 2016. He will be remembered and
2391 missed by all those who he came in contact with, and enriched with his
2392 intellect, wit, and spirit.
2393 .PP
2394 It is with great sadness that we also note Kip Hampton's passing. Probably
2395 best known as the author of the Perl & XML column on XML.com, he was a
2396 core contributor to AxKit, an XML server platform that became an Apache
2397 Foundation project. He was a frequent speaker in the early days at
2398 OSCON, and most recently at YAPC::NA in Madison. He was frequently on
2399 irc.perl.org as ubu, generally in the #axkit\-dahut community, the
2400 group responsible for YAPC::NA Asheville in 2011.
2401 .PP
2402 Kip and his constant contributions to the community will be greatly
2403 missed.
2404 .SH Acknowledgements
2405 .IX Header "Acknowledgements"
2406 Perl 5.26.0 represents approximately 13 months of development since Perl 5.24.0
2407 and contains approximately 360,000 lines of changes across 2,600 files from 86
2408 authors.
2409 .PP
2410 Excluding auto-generated files, documentation and release tools, there were
2411 approximately 230,000 lines of changes to 1,800 .pm, .t, .c and .h files.
2412 .PP
2413 Perl continues to flourish into its third decade thanks to a vibrant community
2414 of users and developers. The following people are known to have contributed the
2415 improvements that became Perl 5.26.0:
2416 .PP
2417 Aaron Crane, Abigail, Ævar Arnfjörð Bjarmason, Alex Vandiver, Andreas
2418 König, Andreas Voegele, Andrew Fresh, Andy Lester, Aristotle Pagaltzis, Chad
2419 Granum, Chase Whitener, Chris 'BinGOs' Williams, Chris Lamb, Christian Hansen,
2420 Christian Millour, Colin Newell, Craig A. Berry, Dagfinn Ilmari Mannsåker, Dan
2421 Collins, Daniel Dragan, Dave Cross, Dave Rolsky, David Golden, David H.
2422 Gutteridge, David Mitchell, Dominic Hargreaves, Doug Bell, E. Choroba, Ed Avis,
2423 Father Chrysostomos, François Perrad, Hauke D, H.Merijn Brand, Hugo van der
2424 Sanden, Ivan Pozdeev, James E Keenan, James Raspass, Jarkko Hietaniemi, Jerry
2425 D. Hedden, Jim Cromie, J. Nick Koston, John Lightsey, Karen Etheridge, Karl
2426 Williamson, Leon Timmermans, Lukas Mai, Matthew Horsfall, Maxwell Carey, Misty
2427 De Meo, Neil Bowers, Nicholas Clark, Nicolas R., Niko Tyni, Pali, Paul
2428 Marquess, Peter Avalos, Petr Písař, Pino Toscano, Rafael Garcia-Suarez, Reini
2429 Urban, Renee Baecker, Ricardo Signes, Richard Levitte, Rick Delaney, Salvador
2430 Fandiño, Samuel Thibault, Sawyer X, Sébastien Aperghis-Tramoni, Sergey
2431 Aleynikov, Shlomi Fish, Smylers, Stefan Seifert, Steffen Müller, Stevan
2432 Little, Steve Hay, Steven Humphrey, Sullivan Beck, Theo Buehler, Thomas Sibley,
2433 Todd Rinaldo, Tomasz Konojacki, Tony Cook, Unicode Consortium, Yaroslav Kuzmin,
2434 Yves Orton, Zefram.
2435 .PP
2436 The list above is almost certainly incomplete as it is automatically generated
2437 from version control history. In particular, it does not include the names of
2438 the (very much appreciated) contributors who reported issues to the Perl bug
2439 tracker.
2440 .PP
2441 Many of the changes included in this version originated in the CPAN modules
2442 included in Perl's core. We're grateful to the entire CPAN community for
2443 helping Perl to flourish.
2444 .PP
2445 For a more complete list of all of Perl's historical contributors, please see
2446 the \fIAUTHORS\fR file in the Perl source distribution.
2447 .SH "Reporting Bugs"
2448 .IX Header "Reporting Bugs"
2449 If you find what you think is a bug, you might check the perl bug database at
2450 <https://rt.perl.org/>. There may also be information at
2451 <http://www.perl.org/>, the Perl Home Page.
2452 .PP
2453 If you believe you have an unreported bug, please run the perlbug program
2454 included with your release. Be sure to trim your bug down to a tiny but
2455 sufficient test case. Your bug report, along with the output of \f(CW\*(C`perl \-V\*(C'\fR,
2456 will be sent off to \f(CW\*(C`[email protected]\*(C'\fR to be analysed by the Perl porting team.
2457 .PP
2458 If the bug you are reporting has security implications which make it
2459 inappropriate to send to a publicly archived mailing list, then see
2460 "SECURITY VULNERABILITY CONTACT INFORMATION" in perlsec
2461 for details of how to report the issue.
2462 .SH "Give Thanks"
2463 .IX Header "Give Thanks"
2464 If you wish to thank the Perl 5 Porters for the work we had done in Perl 5,
2465 you can do so by running the \f(CW\*(C`perlthanks\*(C'\fR program:
2466 .PP
2467 .Vb 1
2468 \& perlthanks
2469 .Ve
2470 .PP
2471 This will send an email to the Perl 5 Porters list with your show of thanks.
2472 .SH "SEE ALSO"
2473 .IX Header "SEE ALSO"
2474 The \fIChanges\fR file for an explanation of how to view exhaustive details on
2475 what changed.
2476 .PP
2477 The \fIINSTALL\fR file for how to build Perl.
2478 .PP
2479 The \fIREADME\fR file for general stuff.
2480 .PP
2481 The \fIArtistic\fR and \fICopying\fR files for copyright information.