3 perldelta - what is new for perl v5.8.0
7 This document describes differences between the 5.6.0 release
10 Many of the bug fixes in 5.8.0 were already seen in the 5.6.1
11 maintenance release since the two releases were kept closely
14 If you are upgrading from Perl 5.005_03, you might also want
15 to read L<perl56delta>.
17 =head1 Highlights In 5.8.0
23 Better Unicode support
27 New Thread Implementation
35 Better Numeric Accuracy
43 More Extensive Regression Testing
47 =head1 Incompatible Changes
49 =head2 64-bit platforms and malloc
51 If your pointers are 64 bits wide, the Perl malloc is no longer being
52 used because it does not work well with 8-byte pointers. Also,
53 usually the system mallocs on such platforms are much better optimized
54 for such large memory models than the Perl malloc. Some memory-hungry
55 Perl applications like the PDL don't work well with Perl's malloc.
56 Finally, other applications than Perl (like modperl) tend to prefer
57 the system malloc. Such platforms include Alpha and 64-bit HPPA,
60 =head2 AIX Dynaloading
62 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
63 dlopen interface of AIX instead of the old emulated interface. This
64 change will probably break backward compatibility with compiled
65 modules. The change was made to make Perl more compliant with other
66 applications like modperl which are using the AIX native interface.
68 =head2 Attributes for C<my> variables now handled at run-time.
70 The C<my EXPR : ATTRS> syntax now applies variable attributes at
71 run-time. (Subroutine and C<our> variables still get attributes applied
72 at compile-time.) See L<attributes> for additional details. In particular,
73 however, this allows variable attributes to be useful for C<tie> interfaces,
74 which was a deficiency of earlier releases. Note that the new semantics
75 doesn't work with the Attribute::Handlers module (as of version 0.76).
77 =head2 Socket Extension Dynamic in VMS
79 The Socket extension is now dynamically loaded instead of being
80 statically built in. This may or may not be a problem with ancient
81 TCP/IP stacks of VMS: we do not know since we weren't able to test
82 Perl in such configurations.
84 =head2 IEEE-format Floating Point Default on OpenVMS Alpha
86 Perl now uses IEEE format (T_FLOAT) as the default internal floating
87 point format on OpenVMS Alpha, potentially breaking binary compatibility
88 with external libraries or existing data. G_FLOAT is still available as
89 a configuration option. The default on VAX (D_FLOAT) has not changed.
91 =head2 New Unicode Properties
93 Unicode I<scripts> are now supported. Scripts are similar to (and superior
94 to) Unicode I<blocks>. The difference between scripts and blocks is that
95 scripts are the glyphs used by a language or a group of languages, while
96 the blocks are more artificial groupings of (mostly) 256 characters based
97 on the Unicode numbering.
99 In general, scripts are more inclusive, but not universally so. For
100 example, while the script C<Latin> includes all the Latin characters and
101 their various diacritic-adorned versions, it does not include the various
102 punctuation or digits (since they are not solely C<Latin>).
104 A number of other properties are now supported, including C<\p{L&}>,
105 C<\p{Any}> C<\p{Assigned}>, C<\p{Unassigned}>, C<\p{Blank}> and
106 C<\p{SpacePerl}> (along with their C<\P{...}> versions, of course).
107 See L<perlunicode> for details, and more additions.
109 The C<In> or C<Is> prefix to names used with the C<\p{...}> and C<\P{...}>
110 are now almost always optional. The only exception is that a C<In> prefix
111 is required to signify a Unicode block when a block name conflicts with a
112 script name. For example, C<\p{Tibetan}> refers to the script, while
113 C<\p{InTibetan}> refers to the block. When there is no name conflict, you
114 can omit the C<In> from the block name (e.g. C<\p{BraillePatterns}>), but
115 to be safe, it's probably best to always use the C<In>).
117 =head2 Perl Parser Stress Tested
119 The Perl parser has been stress tested using both random input and
120 Markov chain input and the few found crashes and lockups have been
123 =head2 REF(...) Instead Of SCALAR(...)
125 A reference to a reference now stringifies as "REF(0x81485ec)" instead
126 of "SCALAR(0x81485ec)" in order to be more consistent with the return
129 =head2 pack/unpack D/F recycled
131 The undocumented pack/unpack template letters D/F have been recycled
132 for better use: now they stand for long double (if supported by the
133 platform) and NV (Perl internal floating point type). (They used
134 to be aliases for d/f, but you never knew that.)
142 The semantics of bless(REF, REF) were unclear and until someone proves
143 it to make some sense, it is forbidden.
147 The obsolete chat2 library that should never have been allowed
148 to escape the laboratory has been decommissioned.
152 The builtin dump() function has probably outlived most of its
153 usefulness. The core-dumping functionality will remain in future
154 available as an explicit call to C<CORE::dump()>, but in future
155 releases the behaviour of an unqualified C<dump()> call may change.
159 The very dusty examples in the eg/ directory have been removed.
160 Suggestions for new shiny examples welcome but the main issue is that
161 the examples need to be documented, tested and (most importantly)
166 The (bogus) escape sequences \8 and \9 now give an optional warning
167 ("Unrecognized escape passed through"). There is no need to \-escape
172 The list of filenames from glob() (or <...>) is now by default sorted
173 alphabetically to be csh-compliant (which is what happened before
174 in most UNIX platforms). (bsd_glob() does still sort platform
175 natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.)
179 Spurious syntax errors generated in certain situations, when glob()
180 caused File::Glob to be loaded for the first time, have been fixed.
184 Although "you shouldn't do that", it was possible to write code that
185 depends on Perl's hashed key order (Data::Dumper does this). The new
186 algorithm "One-at-a-Time" produces a different hashed key order.
187 More details are in L</"Performance Enhancements">.
191 lstat(FILEHANDLE) now gives a warning because the operation makes no sense.
192 In future releases this may become a fatal error.
196 The C<package;> syntax (C<package> without an argument) has been
197 deprecated. Its semantics were never that clear and its
198 implementation even less so. If you have used that feature to
199 disallow all but fully qualified variables, C<use strict;> instead.
203 The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still
204 recognised but now cause fatal errors. The previous behaviour of
205 ignoring them by default and warning if requested was unacceptable
206 since it, in a way, falsely promised that the features could be used.
210 The current user-visible implementation of pseudo-hashes (the weird
211 use of the first array element) is deprecated starting from Perl 5.8.0
212 and will be removed in Perl 5.10.0, and the feature will be
213 implemented differently. Not only is the current interface rather
214 ugly, but the current implementation slows down normal array and hash
215 use quite noticeably. The C<fields> pragma interface will remain
220 The syntaxes C<< @a->[...] >> and C<< %h->{...} >> have now been deprecated.
224 After years of trying the suidperl is considered to be too complex to
225 ever be considered truly secure. The suidperl functionality is likely
226 to be removed in a future release.
230 The long deprecated uppercase aliases for the string comparison
231 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
235 The tr///C and tr///U features have been removed and will not return;
236 the interface was a mistake. Sorry about that. For similar
237 functionality, see pack('U0', ...) and pack('C0', ...).
241 Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo (@)".
242 The prototypes are now checked at compile-time for invalid characters.
243 An optional warning is generated ("Illegal character in prototype...")
244 but this may be upgraded to a fatal error in a future release.
248 =head1 Core Enhancements
250 =head2 PerlIO is Now The Default
256 IO is now by default done via PerlIO rather than system's "stdio".
257 PerlIO allows "layers" to be "pushed" onto a file handle to alter the
258 handle's behaviour. Layers can be specified at open time via 3-arg
261 open($fh,'>:crlf :utf8', $path) || ...
263 or on already opened handles via extended C<binmode>:
265 binmode($fh,':encoding(iso-8859-7)');
267 The built-in layers are: unix (low level read/write), stdio (as in
268 previous Perls), perlio (re-implementation of stdio buffering in a
269 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
270 but available on any platform). A mmap layer may be available if
271 platform supports it (mostly UNIXes).
273 Layers to be applied by default may be specified via the 'open' pragma.
275 See L</"Installation and Configuration Improvements"> for the effects
276 of PerlIO on your architecture name.
280 File handles can be marked as accepting Perl's internal encoding of Unicode
281 (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :
283 open($fh,">:utf8","Uni.txt");
285 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named
286 for you since it's not UTF-8 what you will be getting but instead
287 UTF-EBCDIC. See L<perlunicode>, L<utf8>, and
288 http://www.unicode.org/unicode/reports/tr16/ for more information.
289 In future releases this naming may change.
293 File handles can translate character encodings from/to Perl's internal
294 Unicode form on read/write via the ":encoding()" layer.
298 File handles can be opened to "in memory" files held in Perl scalars via:
300 open($fh,'>', \$variable) || ...
304 Anonymous temporary files are available without need to
305 'use FileHandle' or other module via
307 open($fh,"+>", undef) || ...
309 That is a literal undef, not an undefined value.
313 The list form of C<open> is now implemented for pipes (at least on UNIX):
315 open($fh,"-|", 'cat', '/etc/motd')
317 creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in
324 Perl used to be fragile in that signals arriving at inopportune moments
325 could corrupt Perl's internal state. Now Perl postpones handling of
326 signals until it's safe (between opcodes).
328 This change may have surprising side effects because signals no longer
329 interrupt Perl instantly. Perl will now first finish whatever it was
330 doing, like finishing an internal operation (like sort()) or an
331 external operation (like an I/O operation), and only then look at any
332 arrived signals (and before starting the next operation). No more corrupt
333 internal state since the current operation is always finished first,
334 but the signal may take more time to get heard. Note that breaking
335 out from potentially blocking operations should still work, though.
337 =head2 Unicode Overhaul
339 Unicode in general should be now much more usable than in Perl 5.6.0
340 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in
341 regular expressions should work now, Unicode in tr/// should work now,
342 Unicode in I/O should work now.
348 The Unicode Character Database coming with Perl has been upgraded
349 to Unicode 3.1.1. For more information, see http://www.unicode.org/.
353 For developers interested in enhancing Perl's Unicode capabilities:
354 almost all the UCD files are included with the Perl distribution in
355 the F<lib/unicore subdirectory>. The most notable omission, for space
356 considerations, is the Unihan database.
360 The properties \p{Blank} and \p{SpacePerl} have been added. "Blank" is like
361 C isblank(), that is, it contains only "horizontal whitespace" (the space
362 character is, the newline isn't), and the "SpacePerl" is the Unicode
363 equivalent of C<\s> (\p{Space} isn't, since that includes the vertical
364 tabulator character, whereas C<\s> doesn't.)
366 See "New Unicode Properties" earlier in this document for additional
367 information on changes with Unicode properties.
371 =head2 Understanding of Numbers
373 In general a lot of fixing has happened in the area of Perl's
374 understanding of numbers, both integer and floating point. Since in
375 many systems the standard number parsing functions like C<strtoul()>
376 and C<atof()> seem to have bugs, Perl tries to work around their
377 deficiencies. This results hopefully in more accurate numbers.
379 Perl now tries internally to use integer values in numeric conversions
380 and basic arithmetics (+ - * /) if the arguments are integers, and
381 tries also to keep the results stored internally as integers.
382 This change leads to often slightly faster and always less lossy
383 arithmetics. (Previously Perl always preferred floating point numbers
386 =head2 Miscellaneous Changes
392 AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute
393 to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.
397 C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass
398 in multiple arguments.)
402 The builtin dump() now gives an optional warning
403 C<dump() better written as CORE::dump()>,
404 meaning that by default C<dump(...)> is resolved as the builtin
405 dump() which dumps core and aborts, not as (possibly) user-defined
406 C<sub dump>. To call the latter, qualify the call as C<&dump(...)>.
407 (The whole dump() feature is to considered deprecated, and possibly
408 removed/changed in future releases.)
412 chomp() and chop() are now overridable. Note, however, that their
413 prototype (as given by C<prototype("CORE::chomp")> is undefined,
414 because it cannot be expressed and therefore one cannot really write
415 replacements to override these builtins.
419 END blocks are now run even if you exit/die in a BEGIN block.
420 Internally, the execution of END blocks is now controlled by
421 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new
422 behaviour for Perl embedders. This will default in 5.10. See
427 Formats now support zero-padded decimal fields.
431 Lvalue subroutines can now return C<undef> in list context.
432 However, the lvalue subroutine feature still remains experimental.
436 A lost warning "Can't declare ... dereference in my" has been
437 restored (Perl had it earlier but it became lost in later releases.)
441 A new special regular expression variable has been introduced:
442 C<$^N>, which contains the most-recently closed group (submatch).
446 C<no Module;> now works even if there is no "sub unimport" in the Module.
450 The numerical comparison operators return C<undef> if either operand
451 is a NaN. Previously the behaviour was unspecified.
455 The following builtin functions are now overridable: each(), keys(),
456 pop(), push(), shift(), splice(), unshift().
460 C<pack() / unpack()> now can group template letters with C<()> and then
461 apply repetition/count modifiers on the groups.
465 C<pack() / unpack()> can now process the Perl internal numeric types:
466 IVs, UVs, NVs-- and also long doubles, if supported by the platform.
467 The template letters are C<j>, C<J>, C<F>, and C<D>.
471 C<pack('U0a*', ...)> can now be used to force a string to UTF8.
475 my __PACKAGE__ $obj now works.
479 The printf() and sprintf() now support parameter reordering using the
480 C<%\d+\$> and C<*\d+\$> syntaxes. For example
482 print "%2\$s %1\$s\n", "foo", "bar";
484 will print "bar foo\n". This feature helps in writing
485 internationalised software, and in general when the order
486 of the parameters can vary.
490 prototype(\&) is now available.
494 prototype(\[$@%&]) is now available to implicitly create references
495 (useful for example if you want to emulate the tie() interface).
499 A new command-line option, C<-t> is available. It is the
500 little brother of C<-T>: instead of dieing on taint violations,
501 lexical warnings are given. B<This is only meant as a temporary
502 debugging aid while securing the code of old legacy applications.
503 This is not a substitute for -T.>
507 In other taint news, the C<exec LIST> and C<system LIST> have now been
508 considered too risky (think C<exec @ARGV>: it can start any program
509 with any arguments), and now the said forms cause a warning.
510 You should carefully launder the arguments to guarantee their
511 validity. In future releases of Perl the forms will become fatal
512 errors so consider starting laundering now.
516 If tr/// is just counting characters, it doesn't attempt to
521 untie() will now call an UNTIE() hook if it exists. See L<perltie>
526 L<utime> now supports C<utime undef, undef, @files> to change the
527 file timestamps to the current time.
531 The rules for allowing underscores (underbars) in numeric constants
532 have been relaxed and simplified: now you can have an underscore
533 simply B<between digits>.
537 Rather than relying on C's argv[0] (which may not contain a full pathname)
538 where possible $^X is now set by asking the operating system.
539 (eg by reading F</proc/self/exe> on Linux, F</proc/curproc/file> on FreeBSD)
543 =head1 Modules and Pragmata
545 =head2 New Modules and Pragmata
551 C<Attribute::Handlers> allows a class to define attribute handlers.
554 use Attribute::Handlers;
555 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
557 # later, in some package using or inheriting from MyPack...
559 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
561 Both variables and routines can have attribute handlers. Handlers can
562 be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the
563 exact compilation phase (BEGIN, CHECK, INIT, or END).
567 B<B::Concise> is a new compiler backend for walking the Perl syntax
568 tree, printing concise info about ops, from Stephen McCamant. The
569 output is highly customisable. See L<B::Concise>.
573 C<Class::ISA> for reporting the search path for a class's ISA tree,
574 by Sean Burke, has been added. See L<Class::ISA>.
578 C<Cwd> has now a split personality: if possible, an XS extension is
579 used, (this will hopefully be faster, more secure, and more robust)
580 but if not possible, the familiar Perl implementation is used.
584 C<Devel::PPPort>, originally from Kenneth Albanowski and now
585 maintained by Paul Marquess, has been added. It is primarily used
586 by C<h2xs> to enhance portability of XS modules between different
591 C<Digest>, frontend module for calculating digests (checksums), from
592 Gisle Aas, has been added. See L<Digest>.
596 C<Digest::MD5> for calculating MD5 digests (checksums) as defined in
597 RFC 1321, from Gisle Aas, has been added. See L<Digest::MD5>.
599 use Digest::MD5 'md5_hex';
601 $digest = md5_hex("Thirsty Camel");
603 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
605 NOTE: the C<MD5> backward compatibility module is deliberately not
606 included since its further use is discouraged.
610 C<Encode>, by Nick Ing-Simmons, provides a mechanism to translate
611 between different character encodings. Support for Unicode,
612 ISO-8859-*, ASCII, CP*, KOI8-R, and three variants of EBCDIC are
613 compiled in to the module. Several other encodings (like Japanese,
614 Chinese, and MacIntosh encodings) are included and will be loaded at
615 runtime. See L<Encode>.
617 Any encoding supported by Encode module is also available to the
618 ":encoding()" layer if PerlIO is used.
622 C<I18N::Langinfo> can be use to query locale information.
623 See L<I18N::Langinfo>.
627 C<I18N::LangTags> has functions for dealing with RFC3066-style
628 language tags, by Sean Burke. See L<I18N::LangTags>.
632 C<ExtUtils::Constant> is a new tool for extension writers for
633 generating XS code to import C header constants, by Nicholas Clark.
634 See L<ExtUtils::Constant>.
638 C<Filter::Simple> is an easy-to-use frontend to Filter::Util::Call,
639 from Damian Conway. See L<Filter::Simple>.
645 use Filter::Simple sub {
646 while (my ($from, $to) = splice @_, 0, 2) {
655 use MyFilter qr/red/ => 'green';
657 print "red\n"; # this code is filtered, will print "green\n"
658 print "bored\n"; # this code is filtered, will print "bogreen\n"
662 print "red\n"; # this code is not filtered, will print "red\n"
666 C<File::Temp> allows one to create temporary files and directories in
667 an easy, portable, and secure way, by Tim Jenness. See L<File::Temp>.
671 C<Filter::Util::Call> provides you with the framework to write
672 I<Source Filters> in Perl, from Paul Marquess. For most uses the
673 frontend Filter::Simple is to be preferred. See L<Filter::Util::Call>.
677 C<if> is a new pragma for conditional inclusion of modules, from
682 L<libnet> is a collection of perl5 modules related to network
683 programming, from Graham Barr. See L<Net::FTP>, L<Net::NNTP>,
684 L<Net::Ping>, L<Net::POP3>, L<Net::SMTP>, and L<Net::Time>.
686 Perl installation leaves libnet unconfigured, use F<libnetcfg> to configure.
690 C<List::Util> is a selection of general-utility list subroutines, like
691 sum(), min(), first(), and shuffle(), by Graham Barr. See L<List::Util>.
695 C<Locale::Constants>, C<Locale::Country>, C<Locale::Currency>, and
696 C<Locale::Language>, from Neil Bowers, have been added. They provide the
697 codes for various locale standards, such as "fr" for France, "usd" for
698 US Dollar, and "jp" for Japanese.
702 $country = code2country('jp'); # $country gets 'Japan'
703 $code = country2code('Norway'); # $code gets 'no'
705 See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>,
706 and L<Locale::Language>.
710 C<Locale::Maketext> is localization framework from Sean Burke. See
711 L<Locale::Maketext>, and L<Locale::Maketext::TPJ13>. The latter is an
712 article about software localization, originally published in The Perl
713 Journal #13, republished here with kind permission.
717 C<Memoize> can make your functions faster by trading space for time,
718 from Mark-Jason Dominus. See L<Memoize>.
722 C<MIME::Base64> allows you to encode data in base64, from Gisle Aas,
723 as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
728 $encoded = encode_base64('Aladdin:open sesame');
729 $decoded = decode_base64($encoded);
731 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
737 C<MIME::QuotedPrint> allows you to encode data in quoted-printable
738 encoding, as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
739 Extensions)>, from Gisle Aas.
741 use MIME::QuotedPrint;
743 $encoded = encode_qp("Smiley in Unicode: \x{263a}");
744 $decoded = decode_qp($encoded);
746 print $encoded, "\n"; # "Smiley in Unicode: =263A"
748 MIME::QuotedPrint has been enhanced to provide the basic methods
749 necessary to use it with PerlIO::Via as in :
751 use MIME::QuotedPrint;
752 open($fh,">Via(MIME::QuotedPrint)",$path);
754 See L<MIME::QuotedPrint>.
758 C<NEXT> is pseudo-class for method redispatch, from Damian Conway.
763 C<open> is a new pragma for setting the default I/O disciplines
768 C<PerlIO::Scalar> provides the implementation of IO to "in memory"
769 Perl scalars as discussed above, from Nick Ing-Simmons. It also
770 serves as an example of a loadable PerlIO layer. Other future
771 possibilities include PerlIO::Array and PerlIO::Code.
772 See L<PerlIO::Scalar>.
776 C<PerlIO::Via> acts as a PerlIO layer and wraps PerlIO layer
777 functionality provided by a class (typically implemented in perl
778 code), from Nick Ing-Simmons.
780 use MIME::QuotedPrint;
781 open($fh,">Via(MIME::QuotedPrint)",$path);
783 This will automatically convert everything output to C<$fh>
784 to Quoted-Printable. See L<PerlIO::Via>.
788 C<Pod::ParseLink>, by Russ Allbery, has been added,
789 to parse LZ<><> links in pods as described in the new
794 C<Pod::Text::Overstrike>, by Joe Smith, has been added.
795 It converts POD data to formatted overstrike text.
796 See L<Pod::Text::Overstrike>.
800 C<Scalar::Util> is a selection of general-utility scalar subroutines,
801 like blessed(), reftype(), and tainted(). See L<Scalar::Util>.
805 C<sort> is a new pragma for controlling the behaviour of sort().
809 C<Storable> gives persistence to Perl data structures by allowing the
810 storage and retrieval of Perl data to and from files in a fast and
811 compact binary format, from Raphael Manfredi. See L<Storable>.
815 C<Switch>, from Damian Conway, has been added. Just by saying
819 you have C<switch> and C<case> available in Perl.
825 case 1 { print "number 1" }
826 case "a" { print "string a" }
827 case [1..10,42] { print "number in list" }
828 case (@array) { print "number in list" }
829 case /\w+/ { print "pattern" }
830 case qr/\w+/ { print "pattern" }
831 case (%hash) { print "entry in hash" }
832 case (\%hash) { print "entry in hash" }
833 case (\&sub) { print "arg to subroutine" }
834 else { print "previous case not true" }
841 C<Test::More> is yet another framework for writing test scripts,
842 more extensive than Test::Simple, by Michael Schwern. See L<Test::More>.
846 C<Test::Simple> has basic utilities for writing tests, by Michael
847 Schwern. See L<Test::Simple>.
851 C<Text::Balanced> has been added, for extracting delimited text
852 sequences from strings, from Damian Conway.
854 use Text::Balanced 'extract_delimited';
856 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
858 $a will be "'never say never'", $b will be ', he never said'.
860 In addition to extract_delimited() there are also extract_bracketed(),
861 extract_quotelike(), extract_codeblock(), extract_variable(),
862 extract_tagged(), extract_multiple(), gen_delimited_pat(), and
863 gen_extract_tagged(). With these you can implement rather advanced
864 parsing algorithms. See L<Text::Balanced>.
868 C<threads> is an interface to interpreter threads, by Arthur Bergman.
869 Interpreter threads (ithreads) is the new thread model introduced in
870 Perl 5.6 but only available as an internal interface for extension
871 writers (and for Win32 Perl for C<fork()> emulation). See L<threads>.
875 C<threads::shared> allows data sharing for interpreter threads, from
876 Arthur Bergman. In the ithreads model any data sharing between
877 threads must be explicit, as opposed to the old 5.005 thread model
878 where data sharing was implicit. See L<threads::shared>.
882 C<Tie::File>, by Mark-Jason Dominus, associates a Perl array with the
887 C<Tie::Memoize>, by Ilya Zakharevich, provides on-demand loaded hashes.
891 C<Tie::RefHash::Nestable>, by Edward Avis, allows storing hash
892 references (unlike the standard Tie::RefHash) The module is contained
893 within Tie::RefHash, see L<Tie::RefHash>.
897 C<Time::HiRes> provides high resolution timing (ualarm, usleep,
898 and gettimeofday), from Douglas E. Wegscheid. See L<Time::HiRes>.
902 C<Unicode::UCD> offers a querying interface to the Unicode Character
903 Database. See L<Unicode::UCD>.
907 C<Unicode::Collate> implements the UCA (Unicode Collation Algorithm)
908 for sorting Unicode strings, by SADAHIRO Tomoyuki. See L<Unicode::Collate>.
912 C<Unicode::Normalize> implements the various Unicode normalization
913 forms, by SADAHIRO Tomoyuki. See L<Unicode::Normalize>.
917 C<XS::Typemap>, by Tim Jenness, is a test extension that exercises XS
918 typemaps. Nothing gets installed but for extension writers the code
923 =head2 Updated And Improved Modules and Pragmata
929 The following independently supported modules have been updated to the
930 newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp,
931 Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle
932 (Pod::Man, Pod::Text), Pod::LaTeX, Pod::Parser, Storable,
933 Term::ANSIColor, Test, Text-Tabs+Wrap.
937 The attributes::reftype() now works on tied arguments.
941 AutoLoader can now be disabled with C<no AutoLoader;>.
945 B::Deparse has been significantly enhanced. It now can deparse almost
946 all of the standard test suite (so that the tests still succeed).
947 There is a make target "test.deparse" for trying this out.
951 Class::Struct can now define the classes in compile time.
955 Class::Struct now assigns the array/hash element if the accessor
956 is called with an array/hash element as the B<sole> argument.
960 Data::Dumper has now an option to sort hashes.
964 Data::Dumper has now an option to dump code references
969 DB_File now supports newer Berkeley DB versions, among
974 The English module can now be used without the infamous performance
977 use English '-no_match_vars';
979 (Assuming, of course, that one doesn't need the troublesome variables
980 C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and
981 C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>.
985 Fcntl, Socket, and Sys::Syslog have been rewritten to use the
986 new-style constant dispatch section (see L<ExtUtils::Constant>).
987 This means that they will be more robust and hopefully faster.
991 File::Find now chdir()s correctly when chasing symbolic links.
995 File::Find now has pre- and post-processing callbacks. It also
996 correctly changes directories when chasing symbolic links. Callbacks
997 (naughtily) exiting with "next;" instead of "return;" now work.
1001 File::Find is now (again) reentrant. It also has been made
1006 File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid
1007 prototype mismatch with CORE::glob().
1011 File::Glob now supports C<GLOB_LIMIT> constant to limit the size of
1012 the returned list of filenames.
1016 Devel::Peek now has an interface for the Perl memory statistics
1017 (this works only if you are using perl's malloc, and if you have
1018 compiled with debugging).
1022 IPC::Open3 now allows the use of numeric file descriptors.
1026 IO::Socket has now atmark() method, which returns true if the socket
1027 is positioned at the out-of-band mark. The method is also exportable
1028 as a sockatmark() function.
1032 IO::Socket::INET has support for ReusePort option (if your platform
1033 supports it). The Reuse option now has an alias, ReuseAddr. For clarity
1034 you may want to prefer ReuseAddr.
1038 IO::Socket::INET now supports C<LocalPort> of zero (usually meaning
1039 that the operating system will make one up.)
1043 use lib now works identically to @INC. Removing directories
1044 with 'no lib' now works.
1048 ExtUtils::MakeMaker now uses File::Spec internally, which hopefully
1049 leads into better portability.
1053 Math::BigFloat and Math::BigInt have undergone a full rewrite.
1054 They are now magnitudes faster, and they support various
1055 bignum libraries such as GMP and PARI as their backends.
1059 Math::Complex handles inf, NaN etc., better.
1063 Net::Ping has been muchly enhanced. Multihoming is now supported.
1064 There is now "external" protocol which uses Net::Ping::External module
1065 which runs external ping(1) and parses the output. A version of
1066 Net::Ping::External is available in CPAN.
1070 POSIX::sigaction() is now much more flexible and robust.
1071 You can now install coderef handlers, 'DEFAULT', and 'IGNORE'
1072 handlers, installing new handlers was not atomic.
1076 In Safe the C<%INC> now localised in a Safe compartment so that
1081 In SDBM_File on dosish platforms, some keys went missing because of
1082 lack of support for files with "holes". A workaround for the problem
1087 In Search::Dict one can now have a pre-processing hook for the
1088 lines being searched.
1092 The Shell module now has an OO interface.
1096 The Test module has been significantly enhanced.
1100 The vars pragma now supports declaring fully qualified variables.
1101 (Something that C<our()> does not and will not support.)
1105 The C<utf8::> name space (as in the pragma) provides various
1106 Perl-callable functions to provide low level access to Perl's
1107 internal Unicode representation. At the moment only length()
1108 has been implemented.
1112 =head1 Utility Changes
1118 Emacs perl mode (emacs/cperl-mode.el) has been updated to version
1123 F<emacs/e2ctags.pl> is now much faster.
1127 C<h2ph> now supports C trigraphs.
1131 C<h2xs> now produces a template README.
1135 C<h2xs> now uses C<Devel::PPort> for better portability between
1136 different versions of Perl.
1140 C<h2xs> uses the new L<ExtUtils::Constant> module which will affect
1141 newly created extensions that define constants. Since the new code is
1142 more correct (if you have two constants where the first one is a
1143 prefix of the second one, the first constant B<never> gets defined),
1144 less lossy (it uses integers for integer constant, as opposed to the
1145 old code that used floating point numbers even for integer constants),
1146 and slightly faster, you might want to consider regenerating your
1147 extension code (the new scheme makes regenerating easy).
1148 L<h2xs> now also supports C trigraphs.
1152 C<libnetcfg> has been added to configure the libnet.
1156 C<perlbug> is now much more robust. It also sends the bug report to
1157 perl.org, not perl.com.
1161 C<perlcc> has been rewritten and its user interface (that is,
1162 command line) is much more like that of the UNIX C compiler, cc.
1163 (The perlbc tools has been removed. Use C<perlcc -B> instead.)
1167 C<perlivp> is a new Installation Verification Procedure utility
1168 for running any time after installing Perl.
1172 C<pod2html> now allows specifying a cache directory.
1176 C<s2p> has been completely rewritten in Perl. (It is in fact a full
1177 implementation of sed in Perl: you can use the sed functionality by
1178 using the C<psed> utility.)
1182 C<xsubpp> now understands POD documentation embedded in the *.xs files.
1186 C<xsubpp> now supports OUT keyword.
1190 =head1 New Documentation
1196 perl56delta details the changes between the 5.005 release and the
1201 perlclib documents the internal replacements for standard C library
1202 functions. (Interesting only for extension writers and Perl core
1207 perldebtut is a Perl debugging tutorial.
1211 perlebcdic contains considerations for running Perl on EBCDIC platforms.
1215 perlintro is a gentle introduction to Perl.
1219 perliol documents the internals of PerlIO with layers.
1223 perlmodstyle is a style guide for writing modules.
1227 perlnewmod tells about writing and submitting a new module.
1231 perlpacktut is a pack() tutorial.
1235 perlpod has been rewritten to be clearer and to record the best
1236 practices gathered over the years.
1240 perlpodspec is a more formal specification of the pod format,
1241 mainly of interest for writers of pod applications, not to
1242 people writing in pod.
1246 perlretut is a regular expression tutorial.
1250 perlrequick is a regular expressions quick-start guide.
1251 Yes, much quicker than perlretut.
1255 perltodo has been updated.
1259 perltootc has been renamed as perltooc (to not to conflict
1260 with perltoot in filesystems restricted to "8.3" names)
1264 perluniintro is an introduction to using Unicode in Perl.
1265 (perlunicode is more of a detailed reference and background
1270 perlutil explains the command line utilities packaged with the Perl
1275 The following platform-specific documents are available before
1276 the installation as README.I<platform>, and after the installation
1279 perlaix perlamiga perlapollo perlbeos perlbs2000
1280 perlce perlcygwin perldgux perldos perlepoc perlhpux
1281 perlhurd perlmachten perlmacos perlmint perlmpeix
1282 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1283 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1289 The documentation for the POSIX-BC platform is called "BS2000", to avoid
1290 confusion with the Perl POSIX module.
1294 The documentation for the WinCE platform is called "CE", to avoid
1295 confusion with the perlwin32 documentation on 8.3-restricted filesystems.
1299 =head1 Performance Enhancements
1305 map() could get pathologically slow when the result list it generates
1306 is larger than the source list. The performance has been improved for
1311 sort() has been changed to use primarily mergesort internally as
1312 opposed to the earlier quicksort. For very small lists this may
1313 result in slightly slower sorting times, but in general the speedup
1314 should be at least 20%. Additional bonuses are that the worst case
1315 behaviour of sort() is now better (in computer science terms it now
1316 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1317 worst-case run time behaviour), and that sort() is now stable
1318 (meaning that elements with identical keys will stay ordered as they
1319 were before the sort). See the C<sort> pragma for information.
1321 The story in more detail: suppose you want to serve yourself a little
1324 @digits = ( 3,1,4,1,5,9 );
1326 A numerical sort of the digits will yield (1,1,3,4,5,9), as expected.
1327 Which C<1> comes first is hard to know, since one C<1> looks pretty
1328 much like any other. You can regard this as totally trivial,
1329 or somewhat profound. However, if you just want to sort the even
1330 digits ahead of the odd ones, then what will
1332 sort { ($a % 2) <=> ($b % 2) } @digits;
1334 yield? The only even digit, C<4>, will come first. But how about
1335 the odd numbers, which all compare equal? With the quicksort algorithm
1336 used to implement Perl 5.6 and earlier, the order of ties is left up
1337 to the sort. So, as you add more and more digits of Pi, the order
1338 in which the sorted even and odd digits appear will change.
1339 and, for sufficiently large slices of Pi, the quicksort algorithm
1340 in Perl 5.8 won't return the same results even if reinvoked with the
1341 same input. The justification for this rests with quicksort's
1342 worst case behavior. If you run
1344 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1346 (something you might approximate if you wanted to merge two sorted
1347 arrays using sort), doubling $N doesn't just double the quicksort time,
1348 it I<quadruples> it. Quicksort has a worst case run time that can
1349 grow like N**2, so-called I<quadratic> behaviour, and it can happen
1350 on patterns that may well arise in normal use. You won't notice this
1351 for small arrays, but you I<will> notice it with larger arrays,
1352 and you may not live long enough for the sort to complete on arrays
1353 of a million elements. So the 5.8 quicksort scrambles large arrays
1354 before sorting them, as a statistical defence against quadratic behaviour.
1355 But that means if you sort the same large array twice, ties may be
1356 broken in different ways.
1358 Because of the unpredictability of tie-breaking order, and the quadratic
1359 worst-case behaviour, quicksort was I<almost> replaced completely with
1360 a stable mergesort. I<Stable> means that ties are broken to preserve
1361 the original order of appearance in the input array. So
1363 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1365 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1366 appear in the output in the same order they appeared in the input.
1367 Mergesort has worst case O(NlogN) behaviour, the best value
1368 attainable. And, ironically, this mergesort does particularly
1369 well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N)
1370 in O(N) time. But quicksort was rescued at the last moment because
1371 it is faster than mergesort on certain inputs and platforms.
1372 For example, if you really I<don't> care about the order of even
1373 and odd digits, quicksort will run in O(N) time; it's very good
1374 at sorting many repetitions of a small number of distinct elements.
1375 The quicksort divide and conquer strategy works well on platforms
1376 with relatively small, very fast, caches. Eventually, the problem gets
1377 whittled down to one that fits in the cache, from which point it
1378 benefits from the increased memory speed.
1380 Quicksort was rescued by implementing a sort pragma to control aspects
1381 of the sort. The B<stable> subpragma forces stable behaviour,
1382 regardless of algorithm. The B<_quicksort> and B<_mergesort>
1383 subpragmas are heavy-handed ways to select the underlying implementation.
1384 The leading C<_> is a reminder that these subpragmas may not survive
1385 beyond 5.8. More appropriate mechanisms for selecting the implementation
1386 exist, but they wouldn't have arrived in time to save quicksort.
1390 Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm
1391 (http://burtleburtle.net/bob/hash/doobs.html). This algorithm is
1392 reasonably fast while producing a much better spread of values than
1393 the old hashing algorithm (originally by Chris Torek, later tweaked by
1394 Ilya Zakharevich). Hash values output from the algorithm on a hash of
1395 all 3-char printable ASCII keys comes much closer to passing the
1396 DIEHARD random number generation tests. According to perlbench, this
1397 change has not affected the overall speed of Perl.
1401 unshift() should now be noticeably faster.
1405 =head1 Installation and Configuration Improvements
1407 =head2 Generic Improvements
1413 INSTALL now explains how you can configure Perl to use 64-bit
1414 integers even on non-64-bit platforms.
1418 Policy.sh policy change: if you are reusing a Policy.sh file
1419 (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1420 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1421 them will now be changed to the new prefix, /foo/bar. (Previously
1422 only $prefix changed.) If you do not like this new behaviour,
1423 specify prefix, siteprefix, and vendorprefix explicitly.
1427 A new optional location for Perl libraries, otherlibdirs, is available.
1428 It can be used for example for vendor add-ons without disturbing Perl's
1429 own library directories.
1433 In many platforms the vendor-supplied 'cc' is too stripped-down to
1434 build Perl (basically, 'cc' doesn't do ANSI C). If this seems
1435 to be the case and 'cc' does not seem to be the GNU C compiler
1436 'gcc', an automatic attempt is made to find and use 'gcc' instead.
1440 gcc needs to closely track the operating system release to avoid
1441 build problems. If Configure finds that gcc was built for a different
1442 operating system release than is running, it now gives a clearly visible
1443 warning that there may be trouble ahead.
1447 If binary compatibility with the 5.005 release is not wanted, Configure
1448 no longer suggests including the 5.005 modules in @INC.
1452 Configure C<-S> can now run non-interactively.
1456 Configure support for pdp11-style memory models has been removed due
1461 configure.gnu now works with options with whitespace in them.
1465 installperl now outputs everything to STDERR.
1469 $Config{byteorder} is now computed dynamically (this is more robust
1470 with "fat binaries" where an executable image contains binaries for
1471 more than one binary platform.)
1475 Because PerlIO is now the default on most platforms, "-perlio" doesn't
1476 get appended to the $Config{archname} (also known as $^O) anymore.
1477 Instead, if you explicitly choose not to use perlio (Configure command
1478 line option -Uuseperlio), you will get "-stdio" appended.
1482 Another change related to the architecture name is that "-64all"
1483 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1484 pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)
1488 In AFS installations one can configure the root of the AFS to be
1489 somewhere else than the default F</afs> by using the Configure
1490 parameter C<-Dafsroot=/some/where/else>.
1494 APPLLIB_EXP, a less-know configuration-time definition, has been
1495 documented. It can be used to prepend site-specific directories
1496 to Perl's default search path (@INC), see INSTALL for information.
1500 The version of Berkeley DB used when the Perl (and, presumably, the
1501 DB_File extension) was built is now available as
1502 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
1503 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
1504 DB_VERSION_PATCH_CFG> from C.
1508 Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM
1509 has been documented in INSTALL.
1513 If you have CPAN access (either network or a local copy such as a
1514 CD-ROM) you can during specify extra modules to Configure to build and
1515 install with Perl using the -Dextras=... option. See INSTALL for
1520 In addition to config.over a new override file, config.arch, is
1521 available. That is supposed to be used by hints file writers for
1522 architecture-wide changes (as opposed to config.over which is for
1527 If your file system supports symbolic links you can build Perl outside
1528 of the source directory by
1530 mkdir /tmp/perl/build/directory
1531 cd /tmp/perl/build/directory
1532 sh /path/to/perl/source/Configure -Dmksymlinks ...
1534 This will create in /tmp/perl/build/directory a tree of symbolic links
1535 pointing to files in /path/to/perl/source. The original files are left
1536 unaffected. After Configure has finished you can just say
1540 and Perl will be built and tested, all in /tmp/perl/build/directory.
1544 For Perl developers several new make targets for profiling
1545 and debugging have been added, see L<perlhack>.
1551 Use of the F<gprof> tool to profile Perl has been documented in
1552 L<perlhack>. There is a make target called "perl.gprof" for
1553 generating a gprofiled Perl executable.
1557 If you have GCC 3, there is a make target called "perl.gcov" for
1558 creating a gcoved Perl executable for coverage analysis. See
1563 If you are on IRIX or Tru64 platforms, new profiling/debugging options
1564 have been added, see L<perlhack> for more information about pixie and
1571 Guidelines of how to construct minimal Perl installations have
1572 been added to INSTALL.
1576 The Thread extension is now not built at all under ithreads
1577 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
1578 Thread extension requires being Configured with C<-Duse5005threads>).
1580 But note that the Thread.pm interface is now shared by both
1585 =head2 New Or Improved Platforms
1587 For the list of platforms known to support Perl,
1588 see L<perlport/"Supported Platforms">.
1594 AIX dynamic loading should be now better supported.
1598 AIX should now work better with gcc, threads, and 64-bitness. Also the
1599 long doubles support in AIX should be better now. See L<perlaix>.
1603 After a long pause, AmigaOS has been verified to be happy with Perl.
1607 AtheOS (http://www.atheos.cx/) is a new platform.
1611 BeOS has been reclaimed.
1615 DG/UX platform now supports the 5.005-style threads. See L<perldgux>.
1619 DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near osvers 4.5.2.
1623 EBCDIC platforms (z/OS, also known as OS/390, POSIX-BC, and VM/ESA)
1624 have been regained. Many test suite tests still fail and the
1625 co-existence of Unicode and EBCDIC isn't quite settled, but the
1626 situation is much better than with Perl 5.6. See L<perlos390>,
1627 L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information.
1631 Building perl with -Duseithreads or -Duse5005threads now works under
1632 HP-UX 10.20 (previously it only worked under 10.30 or later). You will
1633 need a thread library package installed. See README.hpux.
1637 MacOS Classic (MacPerl has of course been available since
1638 perl 5.004 but now the source code bases of standard Perl
1639 and MacPerl have been synchronised)
1643 MacOS X (or Darwin) should now be able to build Perl even on HFS+
1644 filesystems. (The case-insensitivity confused the Perl build process.)
1648 NCR MP-RAS is now supported.
1652 All the NetBSD specific patches (except for the installation
1653 specific ones) have been merged back to the main distribution.
1657 NetWare from Novell is now supported. See L<perlnetware>.
1661 NonStop-UX is now supported.
1665 NEC SUPER-UX is now supported.
1669 All the OpenBSD specific patches (except for the installation
1670 specific ones) have been merged back to the main distribution.
1674 Perl has been tested with the GNU pth userlevel thread package
1675 ( http://www.gnu.org/software/pth/pth.html ) . All but one thread
1676 test worked, and that one failure was because of test results arriving
1677 in unexpected order.
1681 Amdahl UTS UNIX mainframe platform is now supported.
1685 WinCE is now supported. See L<perlce>.
1689 z/OS (formerly known as OS/390, formerly known as MVS OE) has now
1690 support for dynamic loading. This is not selected by default,
1691 however, you must specify -Dusedl in the arguments of Configure.
1695 =head1 Selected Bug Fixes
1697 Numerous memory leaks and uninitialized memory accesses have been
1698 hunted down. Most importantly anonymous subs used to leak quite
1705 The autouse pragma didn't work for Multi::Part::Function::Names.
1709 caller() could cause core dumps in certain situations. Carp was sometimes
1710 affected by this problem.
1714 chop(@list) in list context returned the characters chopped in
1715 reverse order. This has been reversed to be in the right order.
1719 Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm)
1720 when building the Perl binary. The only exception to this is SunOS 4.x,
1725 The behaviour of non-decimal but numeric string constants such as
1726 "0x23" was platform-dependent: in some platforms that was seen as 35,
1727 in some as 0, in some as a floating point number (don't ask). This
1728 was caused by Perl using the operating system libraries in a situation
1729 where the result of the string to number conversion is undefined: now
1730 Perl consistently handles such strings as zero in numeric contexts.
1734 The order of DESTROYs has been made more predictable.
1738 Several debugger fixes: exit code now reflects the script exit code,
1739 condition C<"0"> now treated correctly, the C<d> command now checks
1740 line number, the C<$.> no longer gets corrupted, all debugger output
1741 now goes correctly to the socket if RemotePort is set.
1745 Perl 5.6.0 could emit spurious warnings about redefinition of dl_error()
1746 when statically building extensions into perl. This has been corrected.
1750 L<dprofpp> -R didn't work.
1754 C<*foo{FORMAT}> now works.
1757 Infinity is now recognized as a number.
1761 UNIVERSAL::isa no longer caches methods incorrectly. (This broke
1762 the Tk extension with 5.6.0.)
1766 Lexicals I: lexicals outside an eval "" weren't resolved
1767 correctly inside a subroutine definition inside the eval "" if they
1768 were not already referenced in the top level of the eval""ed code.
1772 Lexicals II: lexicals leaked at file scope into subroutines that
1773 were declared before the lexicals.
1777 Lexical warnings now propagating correctly between scopes
1778 and into C<eval "...">.
1782 C<use warnings qw(FATAL all)> did not work as intended. This has been
1787 warnings::enabled() now reports the state of $^W correctly if the caller
1788 isn't using lexical warnings.
1792 Line renumbering with eval and C<#line> now works.
1796 Fixed numerous memory leaks, especially in eval "".
1800 mkdir() now ignores trailing slashes in the directory name,
1801 as mandated by POSIX.
1805 Some versions of glibc have a broken modfl(). This affects builds
1806 with C<-Duselongdouble>. This version of Perl detects this brokenness
1807 and has a workaround for it. The glibc release 2.2.2 is known to have
1808 fixed the modfl() bug.
1812 Modulus of unsigned numbers now works (4063328477 % 65535 used to
1813 return 27406, instead of 27047).
1817 Some "not a number" warnings introduced in 5.6.0 eliminated to be
1818 more compatible with 5.005. Infinity is now recognised as a number.
1822 Numeric conversions did not recognize changes in the string value
1823 properly in certain circumstances.
1827 Attributes (like :shared) didn't work with our().
1831 our() variables will not cause "will not stay shared" warnings.
1835 "our" variables of the same name declared in two sibling blocks
1836 resulted in bogus warnings about "redeclaration" of the variables.
1837 The problem has been corrected.
1841 pack "Z" now correctly terminates the string with "\0".
1845 Fix password routines which in some shadow password platforms
1846 (e.g. HP-UX) caused getpwent() to return every other entry.
1850 The PERL5OPT environment variable (for passing command line arguments
1851 to Perl) didn't work for more than a single group of options.
1855 PERL5OPT with embedded spaces didn't work.
1859 printf() no longer resets the numeric locale to "C".
1863 C<qw(a\\b)> now parses correctly as C<'a\\b'>.
1867 pos() did not return the correct value within s///ge in earlier
1868 versions. This is now handled correctly.
1872 Printing quads (64-bit integers) with printf/sprintf now works
1873 without the q L ll prefixes (assuming you are on a quad-capable platform).
1877 Regular expressions on references and overloaded scalars now work.
1881 Right-hand side magic (GMAGIC) could in many cases such as string
1882 concatenation be invoked too many times.
1886 scalar() now forces scalar context even when used in void context.
1890 SOCKS support is now much more robust.
1894 sort() arguments are now compiled in the right wantarray context
1895 (they were accidentally using the context of the sort() itself).
1896 The comparison block is now run in scalar context, and the arguments
1897 to be sorted are always provided list context.
1901 Changed the POSIX character class C<[[:space:]]> to include the (very
1902 rarely used) vertical tab character. Added a new POSIX-ish character
1903 class C<[[:blank:]]> which stands for horizontal whitespace
1904 (currently, the space and the tab).
1908 The tainting behaviour of sprintf() has been rationalized. It does
1909 not taint the result of floating point formats anymore, making the
1910 behaviour consistent with that of string interpolation.
1914 Some cases of inconsistent taint propagation (such as within hash
1915 values) have been fixed.
1919 The RE engine found in Perl 5.6.0 accidentally pessimised certain kinds
1920 of simple pattern matches. These are now handled better.
1924 Regular expression debug output (whether through C<use re 'debug'>
1925 or via C<-Dr>) now looks better.
1929 Multi-line matches like C<"a\nxb\n" =~ /(?!\A)x/m> were flawed. The
1934 Use of $& could trigger a core dump under some situations. This
1939 The regular expression captured submatches ($1, $2, ...) are now
1940 more consistently unset if the match fails, instead of leaving false
1941 data lying around in them.
1945 readline() on files opened in "slurp" mode could return an extra "" at
1946 the end in certain situations. This has been corrected.
1950 Autovivification of symbolic references of special variables described
1951 in L<perlvar> (as in C<${$num}>) was accidentally disabled. This works
1956 Sys::Syslog ignored the C<LOG_AUTH> constant.
1960 All but the first argument of the IO syswrite() method are now optional.
1964 $AUTOLOAD, sort(), lock(), and spawning subprocesses
1965 in multiple threads simultaneously are now thread-safe.
1969 Tie::ARRAY SPLICE method was broken.
1973 Allow read-only string on left hand side of non-modifying tr///.
1977 Several Unicode fixes.
1983 BOMs (byte order marks) in the beginning of Perl files
1984 (scripts, modules) should now be transparently skipped.
1985 UTF-16 (UCS-2) encoded Perl files should now be read correctly.
1989 The character tables have been updated to Unicode 3.1.1.
1993 Comparing with utf8 data does not magically upgrade non-utf8 data
1994 into utf8. (This was a problem for example if you were mixing data
1995 from I/O and Unicode data: your output might have got magically encoded
2000 Generating illegal Unicode code points like U+FFFE, or the UTF-16
2001 surrogates, now also generates an optional warning.
2005 C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase.
2009 Concatenation with the C<.> operator or via variable interpolation,
2010 C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator,
2011 substitution with C<s///>, single-quoted UTF8, should now work.
2015 The C<tr///> operator now works. Note that the C<tr///CU>
2016 functionality has been removed (but see pack('U0', ...)).
2020 C<eval "v200"> now works.
2024 Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spurious warnings.
2025 This has been corrected.
2029 Zero entries were missing from the Unicode classes like C<IsDigit>.
2035 Large unsigned numbers (those above 2**31) could sometimes lose their
2036 unsignedness, causing bogus results in arithmetic operations.
2040 =head2 Platform Specific Changes and Fixes
2048 Perl now works on post-4.0 BSD/OSes.
2054 Setting C<$0> now works (as much as possible; see L<perlvar> for details).
2060 Numerous updates; currently synchronised with Cygwin 1.3.10.
2064 Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.
2070 EPOC update after Perl 5.6.0. See README.epoc.
2076 Perl now works on post-3.0 FreeBSDs.
2082 README.hpux updated; C<Configure -Duse64bitall> now works.
2088 Numerous compilation flag and hint enhancements; accidental mixing
2089 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
2099 Long doubles should now work (see INSTALL).
2103 Linux previously had problems related to sockaddrlen when using
2104 accept(), revcfrom() (in Perl: recv()), getpeername(), and getsockname().
2112 Compilation of the standard Perl distribution in MacOS Classic should
2113 now work if you have the Metrowerks development environment and
2114 the missing Mac-specific toolkit bits. Contact the macperl mailing
2121 MPE/iX update after Perl 5.6.0. See README.mpeix.
2127 Perl now works on NetBSD/sparc.
2133 Now works with usethreads (see INSTALL).
2139 64-bitness using the Sun Workshop compiler now works.
2143 Tru64 (aka Digital UNIX, aka DEC OSF/1)
2145 The operating system version letter now recorded in $Config{osvers}.
2146 Allow compiling with gcc (previously explicitly forbidden). Compiling
2147 with gcc still not recommended because buggy code results, even with
2154 Fixed various alignment problems that lead into core dumps either
2155 during build or later; no longer dies on math errors at runtime;
2156 now using full quad integers (64 bits), previously was using
2157 only 46 bit integers for speed.
2163 chdir() now works better despite a CRT bug; now works with MULTIPLICITY
2164 (see INSTALL); now works with Perl's malloc.
2166 The tainting of C<%ENV> elements via C<keys> or C<values> was previously
2167 unimplemented. It now works as documented.
2169 The C<waitpid> emulation has been improved. The worst bug (now fixed)
2170 was that a pid of -1 would cause a wildcard search of all processes on
2171 the system. The most significant enhancement is that we can now
2172 usually get the completion status of a terminated process.
2174 POSIX-style signals are now emulated much better on VMS versions prior
2177 The C<system> function and backticks operator have improved
2178 functionality and better error handling.
2180 File access tests now use current process privileges rather than the
2181 user's default privileges, which could sometimes result in a mismatch
2182 between reported access and actual access.
2192 accept() no longer leaks memory.
2196 Borland C++ v5.5 is now a supported compiler that can build Perl.
2197 However, the generated binaries continue to be incompatible with those
2198 generated by the other supported compilers (GCC and Visual C++).
2202 Better chdir() return value for a non-existent directory.
2206 Duping socket handles with open(F, ">&MYSOCK") now works under Windows 9x.
2210 New %ENV entries now propagate to subprocesses.
2214 Current directory entries in %ENV are now correctly propagated to child
2219 $ENV{LIB} now used to search for libs under Visual C.
2223 fork() emulation has been improved in various ways, but still continues
2224 to be experimental. See L<perlfork> for known bugs and caveats.
2228 A failed (pseudo)fork now returns undef and sets errno to EAGAIN.
2232 Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root.
2233 Other bugs in chdir() and Cwd::cwd() have also been fixed.
2237 HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html
2241 The makefiles now provide a single switch to bulk-enable all the features
2242 enabled in ActiveState ActivePerl (a popular Win32 binary distribution).
2246 Allow REG_EXPAND_SZ keys in the registry.
2250 Can now send() from all threads, not just the first one.
2254 Fake signal handling reenabled, bugs and all.
2258 %SIG has been enabled under USE_ITHREADS, but its use is completely
2259 unsupported under all configurations.
2263 Less stack reserved per thread so that more threads can run
2264 concurrently. (Still 16M per thread.)
2268 C<File::Spec->tmpdir()> now prefers C:/temp over /tmp
2269 (works better when perl is running as service).
2273 Better UNC path handling under ithreads.
2277 wait(), waitpid() and backticks now return the correct exit status under
2282 winsock handle leak fixed.
2288 =head1 New or Changed Diagnostics
2294 The lexical warnings category "deprecated" is no longer a sub-category
2295 of the "syntax" category. It is now a top-level category in its own
2300 All regular expression compilation error messages are now hopefully
2301 easier to understand both because the error message now comes before
2302 the failed regex and because the point of failure is now clearly
2303 marked by a C<E<lt>-- HERE> marker.
2307 The various "opened only for", "on closed", "never opened" warnings
2308 drop the C<main::> prefix for filehandles in the C<main> package,
2309 for example C<STDIN> instead of C<main::STDIN>.
2313 The "Unrecognized escape" warning has been extended to include C<\8>,
2314 C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters.
2318 Two new debugging options have been added: if you have compiled your
2319 Perl with debugging, you can use the -DT and -DR options to trace
2320 tokenising and to add reference counts to displaying variables,
2325 perl5db.pl has been modified to present a more consistent commands
2326 interface, via (CommandSet=580). perl5db.t was also added to test the
2327 changes, and as a placeholder for further tests.
2333 The debugger has a new C<dumpDepth> option to control the maximum
2334 depth to which nested structures are dumped. The C<x> command has
2335 been extended so that C<x N EXPR> dumps out the value of I<EXPR> to a
2336 depth of at most I<N> levels.
2340 If an attempt to use a (non-blessed) reference as an array index
2341 is made, a warning is given.
2345 C<push @a;> and C<unshift @a;> (with no values to push or unshift)
2346 now give a warning. This may be a problem for generated and evaled
2351 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
2352 using the C<"C"> format you will get an optional warning. Similarly
2353 for the C<"c"> format and a number less than -128 or more than 127.
2357 Certain regex modifiers such as C<(?o)> make sense only if applied to
2358 the entire regex. You will an optional warning if you try to do otherwise.
2362 Using arrays or hashes as references (e.g. C<< %foo->{bar} >>
2363 has been deprecated for a while. Now you will get an optional warning.
2367 =head1 Changed Internals
2373 perlapi.pod (a companion to perlguts) now attempts to document the
2378 You can now build a really minimal perl called microperl.
2379 Building microperl does not require even running Configure;
2380 C<make -f Makefile.micro> should be enough. Beware: microperl makes
2381 many assumptions, some of which may be too bold; the resulting
2382 executable may crash or otherwise misbehave in wondrous ways.
2383 For careful hackers only.
2387 Added rsignal(), whichsig(), do_join(), op_clear, op_null,
2388 ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2389 interfaces to the publicised API. For the full list of the available
2390 APIs see L<perlapi>.
2394 Made possible to propagate customised exceptions via croak()ing.
2398 Now xsubs can have attributes just like subs. (Well, at least the
2399 built-in attributes.)
2403 dTHR and djSP have been obsoleted; the former removed (because it's
2404 a no-op) and the latter replaced with dSP.
2408 PERL_OBJECT has been completely removed.
2412 The MAGIC constants (e.g. C<'P'>) have been macrofied
2413 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
2414 and maintainability.
2418 The regex compiler now maintains a structure that identifies nodes in
2419 the compiled bytecode with the corresponding syntactic features of the
2420 original regex expression. The information is attached to the new
2421 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
2422 complete information.
2426 The C code has been made much more C<gcc -Wall> clean. Some warning
2427 messages still remain in some platforms, so if you are compiling with
2428 gcc you may see some warnings about dubious practices. The warnings
2429 are being worked on.
2433 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
2437 Documentation on how to use the Perl source repository has been added
2438 to F<Porting/repository.pod>.
2442 There are now several profiling make targets.
2446 =head1 Security Vulnerability Closed
2448 (This change was already made in 5.7.0 but bears repeating here.)
2450 A potential security vulnerability in the optional suidperl component
2451 of Perl was identified in August 2000. suidperl is neither built nor
2452 installed by default. As of November 2001 the only known vulnerable
2453 platform is Linux, most likely all Linux distributions. CERT and
2454 various vendors and distributors have been alerted about the vulnerability.
2455 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2456 for more information.
2458 The problem was caused by Perl trying to report a suspected security
2459 exploit attempt using an external program, /bin/mail. On Linux
2460 platforms the /bin/mail program had an undocumented feature which
2461 when combined with suidperl gave access to a root shell, resulting in
2462 a serious compromise instead of reporting the exploit attempt. If you
2463 don't have /bin/mail, or if you have 'safe setuid scripts', or if
2464 suidperl is not installed, you are safe.
2466 The exploit attempt reporting feature has been completely removed from
2467 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2468 from all the Perl 5.7 releases), so that particular vulnerability
2469 isn't there anymore. However, further security vulnerabilities are,
2470 unfortunately, always possible. The suidperl functionality is most
2471 probably going to be removed in Perl 5.10. In any case, suidperl
2472 should only be used by security experts who know exactly what they are
2473 doing and why they are using suidperl instead of some other solution
2474 such as sudo (see http://www.courtesan.com/sudo/).
2478 Several new tests have been added, especially for the F<lib>
2479 subsection. There are now about 34 000 individual tests (spread over
2480 about 530 test scripts), in the regression suite (5.6.1 has about
2481 11700 tests, in 258 test scripts) Many of the new tests are introduced
2482 by the new modules, but still in general Perl is now more thoroughly
2485 Because of the large number of tests, running the regression suite
2486 will take considerably longer time than it used to: expect the suite
2487 to take up to 4-5 times longer to run than in perl 5.6. In a really
2488 fast machine you can hope to finish the suite in about 5 minutes
2491 The tests are now reported in a different order than in earlier Perls.
2492 (This happens because the test scripts from under t/lib have been moved
2493 to be closer to the library/extension they are testing.)
2495 =head1 Known Problems
2503 In AIX 4.2 Perl extensions that use C++ functions that use statics
2504 may have problems in that the statics are not getting initialized.
2505 In newer AIX releases this has been solved by linking Perl with
2506 the libC_r library, but unfortunately in AIX 4.2 the said library
2507 has an obscure bug where the various functions related to time
2508 (such as time() and gettimeofday()) return broken values, and
2509 therefore in AIX 4.2 Perl is not linked against the libC_r.
2513 vac 5.0.0.0 May Produce Buggy Code For Perl
2515 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
2516 resulting in few random tests failing, but when the failing tests
2517 are run by hand, they succeed. We suggest upgrading to at least
2518 vac version 5.0.1.0, that has been known to compile Perl correctly.
2519 "lslpp -L|grep vac.C" will tell you the vac version. See README.aix.
2523 =head2 Amiga Perl Invoking Mystery
2525 One cannot call Perl using the C<volume:> syntax, that is, C<perl -v>
2526 works, but for example C<bin:perl -v> doesn't. The exact reason isn't
2527 known but the current suspect is the F<ixemul> library.
2529 =head2 lib/ftmp-security tests warn 'system possibly insecure'
2531 Don't panic. Read INSTALL 'make test' section instead.
2533 =head2 FreeBSD 4.3, 4.4, 4.5 fail lib/File/Spec/t/rel2abs2rel.t
2535 F<lib/File/Spec/t/rel2abs2rel.t> tests that "`` works" by running a a perl 1
2536 liner in backticks, using "$^X" as the path to perl. It is known to be
2537 failing on FreeBSD 4.3, 4.4 and 4.5, but only when run as part of make test.
2538 This seems to be a kernel problem rather than perl - reading the symlink
2539 F</proc/curproc/file> returns "unknown" rather than the path to perl, and a
2540 kernel debugger reveals that variable C<numfullpathfail2> in
2541 F</usr/src/sys/kern/vfs_cache.c> is being incremented whenever
2542 F</proc/curproc/file> fails to return the perl executable's path.
2543 [If you find that if fails on other versions of FreeBSD, please use perlbug
2544 to report them to us. If you are able to fix the bug, even better.]
2546 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
2548 If perl is configured with -Duse64bitall, the successful result of the
2549 subtest 10 of lib/posix may arrive before the successful result of the
2550 subtest 9, which confuses the test harness so much that it thinks the
2553 =head2 Linux With Sfio Fails op/misc Test 48
2559 The following tests are known to fail:
2561 Failed Test Stat Wstat Total Fail Failed List of Failed
2562 -------------------------------------------------------------------------
2563 ../ext/DB_File/t/db-btree.t 0 11 ?? ?? % ??
2564 ../ext/DB_File/t/db-recno.t 149 3 2.01% 61 63 65
2565 ../ext/POSIX/t/posix.t 31 1 3.23% 10
2569 OS/390 has rather many test failures but the situation is actually
2570 better than it was in 5.6.0, it's just that so many new modules and
2571 tests have been added.
2573 ../ext/B/t/deparse.t 17 1 5.88% 14
2574 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5
2575 ../lib/utf8.t 94 13 13.83% 27 30-31 43 46 73
2578 ../lib/Benchmark.t 1 256 159 1 0.63% 75
2579 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9
2580 ../lib/ExtUtils/t/ExtUtils.t 27 19 70.37% 5-23
2581 op/pat.t 858 9 1.05% 242-243 665 776 785
2583 op/sprintf.t 224 3 1.34% 98 100 136
2584 op/tr.t 97 5 5.15% 63 71-74
2585 uni/fold.t 767 8 1.04% 25-26 62 169 196
2587 57 tests and 377 subtests skipped.
2589 =head2 op/sprintf tests 129 and 130
2591 The op/sprintf tests 129 and 130 are known to fail on some platforms.
2592 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
2593 The failing platforms do not comply with the ANSI C Standard, line
2594 19ff on page 134 of ANSI X3.159 1989 to be exact. (They produce
2595 something other than "1" and "-1" when formatting 0.6 and -0.6 using
2596 the printf format "%.0f", most often they produce "0" and "-0".)
2598 =head2 Failure of Thread tests
2600 B<Note that support for 5.005-style threading remains experimental
2601 and practically unsupported.>
2603 The following tests are known to fail due to fundamental problems in
2604 the 5.005 threading implementation. These are not new failures--Perl
2605 5.005_0x has the same bugs, but didn't have these tests.
2607 ../ext/List/Util/t/first.t 255 65280 7 4 57.14% 2 5-7
2608 ../lib/English.t 2 512 54 2 3.70% 2-3
2609 ../lib/Filter/Simple/t/data.t 6 3 50.00% 1-3
2610 ../lib/Filter/Simple/t/filter_onl 9 3 33.33% 1-2 5
2611 ../lib/autouse.t 10 1 10.00% 4
2612 op/flip.t 15 1 6.67% 15
2614 These failures are unlikely to get fixed as the 5.005-style
2615 threads are considered fundamentally broken.
2619 ../ext/Socket/socketpair.t 1 256 45 1 2.22% 12
2620 ../lib/Math/Trig.t 26 1 3.85% 25
2621 ../lib/warnings.t 460 1 0.22% 425
2622 io/fs.t 36 1 2.78% 31
2623 op/numconvert.t 1440 13 0.90% 208 509-510
2624 657-658 665-666 829-830 989-990 1149-1150
2626 =head2 UNICOS and UNICOS/mk
2628 The io/fs test #31 is failing because in UNICOS and UNICOS/mk
2629 truncate() cannot be used to grow the size of filehandles, only
2630 to reduce the size. The workaround is to truncate files instead
2635 There are a few known test failures, see L<perluts>.
2639 There should be no reported test failures with a default configuration,
2640 though there are a number of tests marked TODO that point to areas
2641 needing further debugging and/or porting work.
2645 In multi-CPU boxes there are some problems with the I/O buffering:
2646 some output may appear twice. The Win32 following failures are known
2649 ..\ext/Encode/t/JP.t 4 1024 22 4 18.18% 9 14 18 21
2650 ..\ext/threads/t/end.t 6 4 66.67% 3-6
2651 ..\lib/blib.t 3 768 7 3 42.86% 1 4-5
2653 =head2 Localising a Tied Variable Leaks Memory
2656 tie my %tie_hash => 'Tie::StdHash';
2660 local($tie_hash{Foo}) = 1; # leaks
2662 Code like the above is known to leak memory every time the local()
2665 =head2 Localising Tied Arrays and Hashes Is Broken
2669 doesn't work as one would expect: the old value is restored
2672 =head2 Self-tying of Arrays and Hashes Is Forbidden
2674 Self-tying of arrays and hashes is broken in rather deep and
2675 hard-to-fix ways. As a stop-gap measure to avoid people from getting
2676 frustrated at the mysterious results (core dumps, most often) it is
2677 for now forbidden (you will get a fatal error even from an attempt).
2679 =head2 Building Extensions Can Fail Because Of Largefiles
2681 Some extensions like mod_perl are known to have issues with
2682 `largefiles', a change brought by Perl 5.6.0 in which file offsets
2683 default to 64 bits wide, where supported. Modules may fail to compile
2684 at all or compile and work incorrectly. Currently there is no good
2685 solution for the problem, but Configure now provides appropriate
2686 non-largefile ccflags, ldflags, libswanted, and libs in the %Config
2687 hash (e.g., $Config{ccflags_nolargefiles}) so the extensions that are
2688 having problems can try configuring themselves without the
2689 largefileness. This is admittedly not a clean solution, and the
2690 solution may not even work at all. One potential failure is whether
2691 one can (or, if one can, whether it's a good idea) link together at
2692 all binaries with different ideas about file offsets, all this is
2695 =head2 Unicode Support on EBCDIC Still Spotty
2697 Though mostly working, Unicode support still has problem spots on
2698 EBCDIC platforms. One such known spot are the C<\p{}> and C<\P{}>
2699 regular expression constructs for code points less than 256: the
2700 pP are testing for Unicode code points, not knowing about EBCDIC.
2702 =head2 The Compiler Suite Is Still Experimental
2704 The compiler suite is slowly getting better but it continues to be
2705 highly experimental. Use in production environments is discouraged.
2707 =head2 The Long Double Support Is Still Experimental
2709 The ability to configure Perl's numbers to use "long doubles",
2710 floating point numbers of hopefully better accuracy, is still
2711 experimental. The implementations of long doubles are not yet
2712 widespread and the existing implementations are not quite mature
2713 or standardised, therefore trying to support them is a rare
2714 and moving target. The gain of more precision may also be offset
2715 by slowdown in computations (more bits to move around, and the
2716 operations are more likely to be executed by less optimised
2719 =head2 Seen In Perl 5.7 But Gone Now
2721 C<Time::Piece> (previously known as C<Time::Object>) was removed
2722 because it was felt that it didn't have enough value in it to be a
2723 core module. It is still a useful module, though, and is available
2726 =head1 Reporting Bugs
2728 If you find what you think is a bug, you might check the articles
2729 recently posted to the comp.lang.perl.misc newsgroup and the perl
2730 bug database at http://bugs.perl.org. There may also be
2731 information at http://www.perl.com/, the Perl Home Page.
2733 If you believe you have an unreported bug, please run the B<perlbug>
2734 program included with your release. Be sure to trim your bug down
2735 to a tiny but sufficient test case. Your bug report, along with the
2736 output of C<perl -V>, will be sent off to perlbug@perl.org to be
2737 analysed by the Perl porting team.
2741 The F<Changes> file for exhaustive details on what changed.
2743 The F<INSTALL> file for how to build Perl.
2745 The F<README> file for general stuff.
2747 The F<Artistic> and F<Copying> files for copyright information.
2751 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>.