3 perldelta - what is new for perl v5.8.0
7 This document describes differences between the 5.6.0 release and the
16 Better Unicode support
20 New Thread Implementation
28 Better Numeric Accuracy
36 More Extensive Regression Testing
40 =head1 Incompatible Changes
42 =head2 64-bit platforms and malloc
44 If your pointers are 64 bits wide, the Perl malloc is no longer being
45 used because it does not work well with 8-byte pointers. Also,
46 usually the system mallocs on such platforms are much better optimized
47 for such large memory models than the Perl malloc. Some memory-hungry
48 Perl applications like the PDL don't work well with Perl's malloc.
49 Finally, other applications than Perl (like modperl) tend to prefer
50 the system malloc. Such platforms include Alpha and 64-bit HPPA,
53 =head2 AIX Dynaloading
55 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
56 dlopen interface of AIX instead of the old emulated interface. This
57 change will probably break backward compatibility with compiled
58 modules. The change was made to make Perl more compliant with other
59 applications like modperl which are using the AIX native interface.
61 =head2 Socket Extension Dynamic in VMS
63 The Socket extension is now dynamically loaded instead of being
64 statically built in. This may or may not be a problem with ancient
65 TCP/IP stacks of VMS: we do not know since we weren't able to test
66 Perl in such configurations.
68 =head2 Different Definition of the Unicode Character Classes \p{In...}
70 As suggested by the Unicode consortium, the Unicode character classes
71 now prefer I<scripts> as opposed to I<blocks> (as defined by Unicode);
72 in Perl, when the C<\p{In....}> and the C<\p{In....}> regular expression
73 constructs are used. This has changed the definition of some of those
76 The difference between scripts and blocks is that scripts are the
77 glyphs used by a language or a group of languages, while the blocks
78 are more artificial groupings of 256 characters based on the Unicode
81 In general this change results in more inclusive Unicode character
82 classes, but changes to the other direction also do take place:
83 for example while the script C<Latin> includes all the Latin
84 characters and their various diacritic-adorned versions, it
85 does not include the various punctuation or digits (since they
86 are not solely C<Latin>).
88 Changes in the character class semantics may have happened if a script
89 and a block happen to have the same name, for example C<Hebrew>.
90 In such cases the script wins and C<\p{InHebrew}> now means the script
91 definition of Hebrew. The block definition in still available,
92 though, by appending C<Block> to the name: C<\p{InHebrewBlock}> means
93 what C<\p{InHebrew}> meant in perl 5.6.0. For the full list
94 of affected character classes, see L<perlunicode/Blocks>.
96 =head2 Perl Parser Stress Tested
98 The Perl parser has been stress tested using both random input and
99 Markov chain input and the few found crashes and lockups have been
102 =head2 REF(...) Instead Of SCALAR(...)
104 A reference to a reference now stringifies as "REF(0x81485ec)" instead
105 of "SCALAR(0x81485ec)" in order to be more consistent with the return
114 The semantics of bless(REF, REF) were unclear and until someone proves
115 it to make some sense, it is forbidden.
119 The obsolete chat2 library that should never have been allowed
120 to escape the laboratory has been decommissioned.
124 The very dusty examples in the eg/ directory have been removed.
125 Suggestions for new shiny examples welcome but the main issue is that
126 the examples need to be documented, tested and (most importantly)
131 The (bogus) escape sequences \8 and \9 now give an optional warning
132 ("Unrecognized escape passed through"). There is no need to \-escape
137 The list of filenames from glob() (or <...>) is now by default sorted
138 alphabetically to be csh-compliant. (bsd_glob() does still sort platform
139 natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.)
143 Although "you shouldn't do that", it was possible to write code that
144 depends on Perl's hashed key order (Data::Dumper does this). The new
145 algorithm "One-at-a-Time" produces a different hashed key order.
146 More details are in L</"Performance Enhancements">.
150 lstat(FILEHANDLE) now gives a warning because the operation makes no sense.
151 In future releases this may become a fatal error.
155 The C<package;> syntax (C<package> without an argument) has been
156 deprecated. Its semantics were never that clear and its
157 implementation even less so. If you have used that feature to
158 disallow all but fully qualified variables, C<use strict;> instead.
162 The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still
163 recognised but now cause fatal errors. The previous behaviour of
164 ignoring them by default and warning if requested was unacceptable
165 since it, in a way, falsely promised that the features could be used.
169 The current user-visible implementation of pseudo-hashes (the weird
170 use of the first array element) is deprecated starting from Perl 5.8.0
171 and will be removed in Perl 5.10.0, and the feature will be
172 implemented differently. Not only is the current interface rather
173 ugly, but the current implementation slows down normal array and hash
174 use quite noticeably. The C<fields> pragma interface will remain
179 The syntaxes C<< @a->[...] >> and C<< %h->{...} >> have now been deprecated.
183 After years of trying the suidperl is considered to be too complex to
184 ever be considered truly secure. The suidperl functionality is likely
185 to be removed in a future release.
189 The long deprecated uppercase aliases for the string comparison
190 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
194 The tr///C and tr///U features have been removed and will not return;
195 the interface was a mistake. Sorry about that. For similar
196 functionality, see pack('U0', ...) and pack('C0', ...).
200 =head1 Core Enhancements
202 =head2 PerlIO is Now The Default
208 IO is now by default done via PerlIO rather than system's "stdio".
209 PerlIO allows "layers" to be "pushed" onto a file handle to alter the
210 handle's behaviour. Layers can be specified at open time via 3-arg
213 open($fh,'>:crlf :utf8', $path) || ...
215 or on already opened handles via extended C<binmode>:
217 binmode($fh,':encoding(iso-8859-7)');
219 The built-in layers are: unix (low level read/write), stdio (as in
220 previous Perls), perlio (re-implementation of stdio buffering in a
221 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
222 but available on any platform). A mmap layer may be available if
223 platform supports it (mostly UNIXes).
225 Layers to be applied by default may be specified via the 'open' pragma.
227 See L</"Installation and Configuration Improvements"> for the effects
228 of PerlIO on your architecture name.
232 File handles can be marked as accepting Perl's internal encoding of Unicode
233 (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :
235 open($fh,">:utf8","Uni.txt");
237 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named
238 for you since it's not UTF-8 what you will be getting but instead
239 UTF-EBCDIC. See L<perlunicode>, L<utf8>, and
240 http://www.unicode.org/unicode/reports/tr16/ for more information.
241 In future releases this naming may change.
245 File handles can translate character encodings from/to Perl's internal
246 Unicode form on read/write via the ":encoding()" layer.
250 File handles can be opened to "in memory" files held in Perl scalars via:
252 open($fh,'>', \$variable) || ...
256 Anonymous temporary files are available without need to
257 'use FileHandle' or other module via
259 open($fh,"+>", undef) || ...
261 That is a literal undef, not an undefined value.
265 The list form of C<open> is now implemented for pipes (at least on UNIX):
267 open($fh,"-|", 'cat', '/etc/motd')
269 creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in
274 =head2 Signals Are Now Safe
276 Perl used to be fragile in that signals arriving at inopportune moments
277 could corrupt Perl's internal state. Now Perl postpones handling of
278 signals until it's safe.
280 =head2 Unicode Overhaul
282 Unicode in general should be now much more usable than in Perl 5.6.0
283 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in
284 regular expressions should work now, Unicode in tr/// should work now,
285 Unicode in I/O should work now.
291 The Unicode Character Database coming with Perl has been upgraded
292 to Unicode 3.1.1. For more information, see http://www.unicode.org/.
296 For developers interested in enhancing Perl's Unicode capabilities:
297 almost all the UCD files are included with the Perl distribution in
298 the lib/unicore subdirectory. The most notable omission, for space
299 considerations, is the Unihan database.
303 The Unicode character classes \p{Blank} and \p{SpacePerl} have been
304 added. "Blank" is like C isblank(), that is, it contains only
305 "horizontal whitespace" (the space character is, the newline isn't),
306 and the "SpacePerl" is the Unicode equivalent of C<\s> (\p{Space}
307 isn't, since that includes the vertical tabulator character, whereas
312 =head2 Understanding of Numbers
314 In general a lot of fixing has happened in the area of Perl's
315 understanding of numbers, both integer and floating point. Since in
316 many systems the standard number parsing functions like C<strtoul()>
317 and C<atof()> seem to have bugs, Perl tries to work around their
318 deficiencies. This results hopefully in more accurate numbers.
320 Perl now tries internally to use integer values in numeric conversions
321 and basic arithmetics (+ - * /) if the arguments are integers, and
322 tries also to keep the results stored internally as integers.
323 This change leads to often slightly faster and always less lossy
324 arithmetics. (Previously Perl always preferred floating point numbers
327 =head2 Miscellaneous Enhancements
333 AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute
334 to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.
338 C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass
339 in multiple arguments.)
343 END blocks are now run even if you exit/die in a BEGIN block.
344 Internally, the execution of END blocks is now controlled by
345 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new
346 behaviour for Perl embedders. This will default in 5.10. See
351 Formats now support zero-padded decimal fields.
355 Lvalue subroutines can now return C<undef> in list context.
359 A new special regular expression variable has been introduced:
360 C<$^N>, which contains the most-recently closed group (submatch).
364 C<no Module;> now works even if there is no "sub unimport" in the Module.
368 The numerical comparison operators return C<undef> if either operand
369 is a NaN. Previously the behaviour was unspecified.
373 The following builtin functions are now overridable: each(), keys(),
374 pop(), push(), shift(), splice(), unshift().
378 C<pack('U0a*', ...)> can now be used to force a string to UTF8.
382 my __PACKAGE__ $obj now works.
386 The printf() and sprintf() now support parameter reordering using the
387 C<%\d+\$> and C<*\d+\$> syntaxes. For example
389 print "%2\$s %1\$s\n", "foo", "bar";
391 will print "bar foo\n"; This feature helps in writing
392 internationalised software.
396 prototype(\&) is now available.
400 prototype(\[$@%&]) is now available to implicitly create references
401 (useful for example if you want to emulate the tie() interface).
405 UNTIE method is now recognised.
409 L<utime> now supports C<utime undef, undef, @files> to change the
410 file timestamps to the current time.
414 The rules for allowing underscores (underbars) in numeric constants
415 have been relaxed and simplified: now you can have an underscore
416 simply B<between digits>.
420 =head1 Modules and Pragmata
422 =head2 New Modules and Pragmata
428 C<Attribute::Handlers> allows a class to define attribute handlers.
431 use Attribute::Handlers;
432 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
434 # later, in some package using or inheriting from MyPack...
436 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
438 Both variables and routines can have attribute handlers. Handlers can
439 be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the
440 exact compilation phase (BEGIN, CHECK, INIT, or END).
444 B<B::Concise> is a new compiler backend for walking the Perl syntax
445 tree, printing concise info about ops, from Stephen McCamant. The
446 output is highly customisable. See L<B::Concise>.
450 C<Class::ISA> for reporting the search path for a class's ISA tree,
451 by Sean Burke, has been added. See L<Class::ISA>.
455 C<Cwd> has now a split personality: if possible, an XS extension is
456 used, (this will hopefully be faster, more secure, and more robust)
457 but if not possible, the familiar Perl implementation is used.
461 C<Devel::PPPort>, originally from Kenneth Albanowski and now
462 maintained by Paul Marquess, has been added. It is primarily used
463 by C<h2xs> to enhance portability of of XS modules between different
468 C<Digest>, frontend module for calculating digests (checksums), from
469 Gisle Aas, has been added. See L<Digest>.
473 C<Digest::MD5> for calculating MD5 digests (checksums) as defined in
474 RFC 1321, from Gisle Aas, has been added. See L<Digest::MD5>.
476 use Digest::MD5 'md5_hex';
478 $digest = md5_hex("Thirsty Camel");
480 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
482 NOTE: the C<MD5> backward compatibility module is deliberately not
483 included since its further use is discouraged.
487 C<Encode>, by Nick Ing-Simmons, provides a mechanism to translate
488 between different character encodings. Support for Unicode,
489 ISO-8859-*, ASCII, CP*, KOI8-R, and three variants of EBCDIC are
490 compiled in to the module. Several other encodings (like Japanese,
491 Chinese, and MacIntosh encodings) are included and will be loaded at
492 runtime. See L<Encode>.
494 Any encoding supported by Encode module is also available to the
495 ":encoding()" layer if PerlIO is used.
499 C<I18N::Langinfo> can be use to query locale information.
500 See L<I18N::Langinfo>.
504 C<I18N::LangTags> has functions for dealing with RFC3066-style
505 language tags, by Sean Burke. See L<I18N::LangTags>.
509 C<ExtUtils::Constant> is a new tool for extension writers for
510 generating XS code to import C header constants, by Nicholas Clark.
511 See L<ExtUtils::Constant>.
515 C<Filter::Simple> is an easy-to-use frontend to Filter::Util::Call,
516 from Damian Conway. See L<Filter::Simple>.
522 use Filter::Simple sub {
523 while (my ($from, $to) = splice @_, 0, 2) {
532 use MyFilter qr/red/ => 'green';
534 print "red\n"; # this code is filtered, will print "green\n"
535 print "bored\n"; # this code is filtered, will print "bogreen\n"
539 print "red\n"; # this code is not filtered, will print "red\n"
543 C<File::Temp> allows one to create temporary files and directories in
544 an easy, portable, and secure way, by Tim Jenness. See L<File::Temp>.
548 C<Filter::Util::Call> provides you with the framework to write
549 I<Source Filters> in Perl, from Paul Marquess. For most uses the
550 frontend Filter::Simple is to be preferred. See L<Filter::Util::Call>.
554 L<libnet> is a collection of perl5 modules related to network
555 programming, from Graham Barr. See L<Net::FTP>, L<Net::NNTP>,
556 L<Net::Ping>, L<Net::POP3>, L<Net::SMTP>, and L<Net::Time>.
558 Perl installation leaves libnet unconfigured, use F<libnetcfg> to configure.
562 C<List::Util> is a selection of general-utility list subroutines, like
563 sum(), min(), first(), and shuffle(), by Graham Barr. See L<List::Util>.
567 C<Locale::Constants>, C<Locale::Country>, C<Locale::Currency>, and
568 C<Locale::Language>, from Neil Bowers, have been added. They provide the
569 codes for various locale standards, such as "fr" for France, "usd" for
570 US Dollar, and "jp" for Japanese.
574 $country = code2country('jp'); # $country gets 'Japan'
575 $code = country2code('Norway'); # $code gets 'no'
577 See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>,
578 and L<Locale::Language>.
582 C<Locale::Maketext> is localization framework from Sean Burke. See
583 L<Locale::Maketext>, and L<Locale::Maketext::TPJ13>. The latter is an
584 article about software localization, originally published in The Perl
585 Journal #13, republished here with kind permission.
589 C<Memoize> can make your functions faster by trading space for time,
590 from Mark-Jason Dominus. See L<Memoize>.
594 C<MIME::Base64> allows you to encode data in base64, from Gisle Aas,
595 as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
600 $encoded = encode_base64('Aladdin:open sesame');
601 $decoded = decode_base64($encoded);
603 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
609 C<MIME::QuotedPrint> allows you to encode data in quoted-printable
610 encoding, as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
611 Extensions)>, from Gisle Aas.
613 use MIME::QuotedPrint;
615 $encoded = encode_qp("Smiley in Unicode: \x{263a}");
616 $decoded = decode_qp($encoded);
618 print $encoded, "\n"; # "Smiley in Unicode: =263A"
620 MIME::QuotedPrint has been enhanced to provide the basic methods
621 necessary to use it with PerlIO::Via as in :
623 use MIME::QuotedPrint;
624 open($fh,">Via(MIME::QuotedPrint)",$path);
626 See L<MIME::QuotedPrint>.
630 C<NEXT> is pseudo-class for method redispatch, from Damian Conway.
635 C<open> is a new pragma for setting the default I/O disciplines
640 C<PerlIO::Scalar> provides the implementation of IO to "in memory"
641 Perl scalars as discussed above, from Nick Ing-Simmons. It also
642 serves as an example of a loadable PerlIO layer. Other future
643 possibilities include PerlIO::Array and PerlIO::Code.
644 See L<PerlIO::Scalar>.
648 C<PerlIO::Via> acts as a PerlIO layer and wraps PerlIO layer
649 functionality provided by a class (typically implemented in perl
650 code), from Nick Ing-Simmons.
652 use MIME::QuotedPrint;
653 open($fh,">Via(MIME::QuotedPrint)",$path);
655 This will automatically convert everything output to C<$fh>
656 to Quoted-Printable. See L<PerlIO::Via>.
660 C<Pod::ParseLink>, by Russ Allbery, has been added,
661 to parse L<> links in pods as described in the new
666 C<Pod::Text::Overstrike>, by Joe Smith, has been added.
667 It converts POD data to formatted overstrike text.
668 See L<Pod::Text::Overstrike>.
672 C<Scalar::Util> is a selection of general-utility scalar subroutines,
673 like blessed(), reftype(), and tainted(). See L<Scalar::Util>.
677 C<sort> is a new pragma for controlling the behaviour of sort().
681 C<Storable> gives persistence to Perl data structures by allowing the
682 storage and retrieval of Perl data to and from files in a fast and
683 compact binary format, from Raphael Manfredi. See L<Storable>.
687 C<Switch>, from Damian Conway, has been added. Just by saying
691 you have C<switch> and C<case> available in Perl.
697 case 1 { print "number 1" }
698 case "a" { print "string a" }
699 case [1..10,42] { print "number in list" }
700 case (@array) { print "number in list" }
701 case /\w+/ { print "pattern" }
702 case qr/\w+/ { print "pattern" }
703 case (%hash) { print "entry in hash" }
704 case (\%hash) { print "entry in hash" }
705 case (\&sub) { print "arg to subroutine" }
706 else { print "previous case not true" }
713 C<Test::More> is yet another framework for writing test scripts,
714 more extensive than Test::Simple, by Michael Schwern. See L<Test::More>.
718 C<Test::Simple> has basic utilities for writing tests, by Michael
719 Schwern. See L<Test::Simple>.
723 C<Text::Balanced> has been added, for extracting delimited text
724 sequences from strings, from Damian Conway.
726 use Text::Balanced 'extract_delimited';
728 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
730 $a will be "'never say never'", $b will be ', he never said'.
732 In addition to extract_delimited() there are also extract_bracketed(),
733 extract_quotelike(), extract_codeblock(), extract_variable(),
734 extract_tagged(), extract_multiple(), gen_delimited_pat(), and
735 gen_extract_tagged(). With these you can implement rather advanced
736 parsing algorithms. See L<Text::Balanced>.
740 C<threads> is an interface to interpreter threads, by Arthur Bergman.
741 Interpreter threads (ithreads) is the new thread model introduced in
742 Perl 5.6 but only available as an internal interface for extension
743 writers (and for Win32 Perl for C<fork()> emulation). See L<threads>.
747 C<threads::shared> allows data sharing for interpreter threads, from
748 Arthur Bergman. In the ithreads model any data sharing between
749 threads must be explicit, as opposed to the old 5.005 thread model
750 where data sharing was implicit. See L<threads::shared>.
754 C<Tie::RefHash::Nestable>, by Edward Avis, allows storing hash
755 references (unlike the standard Tie::RefHash) The module is contained
756 within Tie::RefHash, see L<Tie::RefHash>.
760 C<Time::HiRes> provides high resolution timing (ualarm, usleep,
761 and gettimeofday), from Douglas E. Wegscheid. See L<Time::HiRes>.
765 C<Unicode::UCD> offers a querying interface to the Unicode Character
766 Database. See L<Unicode::UCD>.
770 C<Unicode::Collate> implements the UCA (Unicode Collation Algorithm)
771 for sorting Unicode strings, by SADAHIRO Tomoyuki. See L<Unicode::Collate>.
775 C<Unicode::Normalize> implements the various Unicode normalization
776 forms, by SADAHIRO Tomoyuki. See L<Unicode::Normalize>.
780 C<XS::Typemap>, by Tim Jenness, is a test extension that exercises XS
781 typemaps. Nothing gets installed but for extension writers the code
786 =head2 Updated And Improved Modules and Pragmata
792 The following independently supported modules have been updated to the
793 newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp,
794 Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle
795 (Pod::Man, Pod::Text), Pod::LaTeX, Pod::Parser, Storable,
796 Term::ANSIColor, Test, Text-Tabs+Wrap.
800 The attributes::reftype() now works on tied arguments.
804 AutoLoader can now be disabled with C<no AutoLoader;>.
808 B::Deparse has been significantly enhanced. It now can deparse almost
809 all of the standard test suite (so that the tests still succeed).
810 There is a make target "test.deparse" for trying this out.
814 Class::Struct can now define the classes in compile time.
818 Class::Struct now assigns the array/hash element if the accessor
819 is called with an array/hash element as the B<sole> argument.
823 Data::Dumper has now an option to sort hashes.
827 Data::Dumper has now an option to dump code references
832 The English module can now be used without the infamous performance
835 use English '-no_performance_hit';
837 (Assuming, of course, that one doesn't need the troublesome variables
838 C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and
839 C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>.
843 Fcntl, Socket, and Sys::Syslog have been rewritten to use the
844 new-style constant dispatch section (see L<ExtUtils::Constant>).
845 This means that they will be more robust and hopefully faster.
849 File::Find now has pre- and post-processing callbacks. It also
850 correctly changes directories when chasing symbolic links. Callbacks
851 (naughtily) exiting with "next;" instead of "return;" now work.
855 File::Find is now (again) reentrant. It also has been made
860 File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid
861 prototype mismatch with CORE::glob().
865 File::Glob now supports C<GLOB_LIMIT> constant to limit the size of
866 the returned list of filenames.
870 Devel::Peek now has an interface for the Perl memory statistics
871 (this works only if you are using perl's malloc, and if you have
872 compiled with debugging).
876 IPC::Open3 now allows the use of numeric file descriptors.
880 IO::Socket has now atmark() method, which returns true if the socket
881 is positioned at the out-of-band mark. The method is also exportable
882 as a sockatmark() function.
886 IO::Socket::INET has support for ReusePort option (if your platform
887 supports it). The Reuse option now has an alias, ReuseAddr. For clarity
888 you may want to prefer ReuseAddr.
892 IO::Socket::INET now supports C<LocalPort> of zero (usually meaning
893 that the operating system will make one up.)
897 use lib now works identically to @INC. Removing directories
898 with 'no lib' now works.
902 Math::BigFloat and Math::BigInt have undergone a full rewrite.
903 They are now magnitudes faster, and they support various
904 bignum libraries such as GMP and PARI as their backends.
908 Net::Ping has been enhanced. There is now "external" protocol which
909 uses Net::Ping::External module which runs external ping(1) and parses
910 the output. An alpha version of Net::Ping::External is available in
911 CPAN and in 5.7.2 the Net::Ping::External may be integrated to Perl.
915 C<POSIX::sigaction()> is now much more flexible and robust.
916 You can now install coderef handlers, 'DEFAULT', and 'IGNORE'
917 handlers, installing new handlers was not atomic.
921 In C<Safe> the C<%INC> now localised in a Safe compartment so that
926 In C<Search::Dict> one can now have a pre-processing hook for the
927 lines being searched.
931 The Shell module now has an OO interface.
935 The Test module has been significantly enhanced.
939 The C<vars> pragma now supports declaring fully qualified variables.
940 (Something that C<our()> does not and will not support.)
944 The utf8:: name space (as in the pragma) provides various
945 Perl-callable functions to provide low level access to Perl's
946 internal Unicode representation. At the moment only length()
947 has been implemented.
951 =head1 Utility Changes
957 Emacs perl mode (emacs/cperl-mode.el) has been updated to version
962 F<emacs/e2ctags.pl> is now much faster.
966 C<h2ph> now supports C trigraphs.
970 C<h2xs> now produces a template README.
974 C<h2xs> now uses C<Devel::PPort> for better portability between
975 different versions of Perl.
979 C<h2xs> uses the new L<ExtUtils::Constant> module which will affect
980 newly created extensions that define constants. Since the new code is
981 more correct (if you have two constants where the first one is a
982 prefix of the second one, the first constant B<never> gets defined),
983 less lossy (it uses integers for integer constant, as opposed to the
984 old code that used floating point numbers even for integer constants),
985 and slightly faster, you might want to consider regenerating your
986 extension code (the new scheme makes regenerating easy).
987 L<h2xs> now also supports C trigraphs.
991 C<libnetcfg> has been added to configure the libnet.
995 C<perlbug> is now much more robust. It also sends the bug report to
996 perl.org, not perl.com.
1000 C<perlcc> has been rewritten and its user interface (that is,
1001 command line) is much more like that of the UNIX C compiler, cc.
1005 C<perlivp> is a new Installation Verification Procedure utility
1006 for running any time after installing Perl.
1010 C<pod2html> now allows specifying a cache directory.
1014 C<s2p> has been completely rewritten in Perl. (It is in fact a full
1015 implementation of sed in Perl: you can use the sed functionality by
1016 using the C<psed> utility.)
1020 C<xsubpp> now understands POD documentation embedded in the *.xs files.
1024 C<xsubpp> now supports OUT keyword.
1028 =head1 New Documentation
1034 perl56delta details the changes between the 5.005 release and the
1039 perlclib documents the internal replacements for standard C library
1040 functions. (Interesting only for extension writers and Perl core
1045 perldebtut is a Perl debugging tutorial.
1049 perlebcdic contains considerations for running Perl on EBCDIC platforms.
1053 perlintro is a gentle introduction to Perl.
1057 perliol documents the internals of PerlIO with layers.
1061 perlmodstyle is a style guide for writing modules.
1065 perlnewmod tells about writing and submitting a new module.
1069 perlpacktut is a pack() tutorial.
1073 perlpod has been rewritten to be clearer and to record the best
1074 practices gathered over the years.
1078 perlpodspec is a more formal specification of the pod format,
1079 mainly of interest for writers of pod applications, not to
1080 people writing in pod.
1084 perlretut is a regular expression tutorial.
1088 perlrequick is a regular expressions quick-start guide.
1089 Yes, much quicker than perlretut.
1093 perltodo has been updated.
1097 perltootc has been renamed as perltooc (to not to conflict
1098 with perltoot in filesystems restricted to "8.3" names)
1102 perluniintro is an introduction to using Unicode in Perl
1103 (perlunicode is more of a reference)
1107 perlutil explains the command line utilities packaged with the Perl
1112 The following platform-specific documents are available before
1113 the installation as README.I<platform>, and after the installation
1116 perlaix perlamiga perlapollo perlbeos perlbs2000
1117 perlce perlcygwin perldgux perldos perlepoc perlhpux
1118 perlhurd perlmachten perlmacos perlmint perlmpeix
1119 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1120 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1126 The documentation for the POSIX-BC platform is called "BS2000", to avoid
1127 confusion with the Perl POSIX module.
1131 The documentation for the WinCE platform is called "CE", to avoid
1132 confusion with the perlwin32 documentation on 8.3-restricted filesystems.
1136 =head1 Performance Enhancements
1142 map() that changes the size of the list should now work faster.
1146 sort() has been changed to use primarily mergesort internally as
1147 opposed to the earlier quicksort. For very small lists this may
1148 result in slightly slower sorting times, but in general the speedup
1149 should be at least 20%. Additional bonuses are that the worst case
1150 behaviour of sort() is now better (in computer science terms it now
1151 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1152 worst-case run time behaviour), and that sort() is now stable
1153 (meaning that elements with identical keys will stay ordered as they
1154 were before the sort). See the C<sort> pragma for information.
1156 The story in more detail: suppose you want to serve yourself a little
1159 @digits = ( 3,1,4,1,5,9 );
1161 A numerical sort of the digits will yield (1,1,3,4,5,9), as expected.
1162 Which C<1> comes first is hard to know, since one C<1> looks pretty
1163 much like any other. You can regard this as totally trivial,
1164 or somewhat profound. However, if you just want to sort the even
1165 digits ahead of the odd ones, then what will
1167 sort { ($a % 2) <=> ($b % 2) } @digits;
1169 yield? The only even digit, C<4>, will come first. But how about
1170 the odd numbers, which all compare equal? With the quicksort algorithm
1171 used to implement Perl 5.6 and earlier, the order of ties is left up
1172 to the sort. So, as you add more and more digits of Pi, the order
1173 in which the sorted even and odd digits appear will change.
1174 and, for sufficiently large slices of Pi, the quicksort algorithm
1175 in Perl 5.8 won't return the same results even if reinvoked with the
1176 same input. The justification for this rests with quicksort's
1177 worst case behavior. If you run
1179 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1181 (something you might approximate if you wanted to merge two sorted
1182 arrays using sort), doubling $N doesn't just double the quicksort time,
1183 it I<quadruples> it. Quicksort has a worst case run time that can
1184 grow like N**2, so-called I<quadratic> behaviour, and it can happen
1185 on patterns that may well arise in normal use. You won't notice this
1186 for small arrays, but you I<will> notice it with larger arrays,
1187 and you may not live long enough for the sort to complete on arrays
1188 of a million elements. So the 5.8 quicksort scrambles large arrays
1189 before sorting them, as a statistical defence against quadratic behaviour.
1190 But that means if you sort the same large array twice, ties may be
1191 broken in different ways.
1193 Because of the unpredictability of tie-breaking order, and the quadratic
1194 worst-case behaviour, quicksort was I<almost> replaced completely with
1195 a stable mergesort. I<Stable> means that ties are broken to preserve
1196 the original order of appearance in the input array. So
1198 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1200 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1201 appear in the output in the same order they appeared in the input.
1202 Mergesort has worst case O(NlogN) behaviour, the best value
1203 attainable. And, ironically, this mergesort does particularly
1204 well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N)
1205 in O(N) time. But quicksort was rescued at the last moment because
1206 it is faster than mergesort on certain inputs and platforms.
1207 For example, if you really I<don't> care about the order of even
1208 and odd digits, quicksort will run in O(N) time; it's very good
1209 at sorting many repetitions of a small number of distinct elements.
1210 The quicksort divide and conquer strategy works well on platforms
1211 with relatively small, very fast, caches. Eventually, the problem gets
1212 whittled down to one that fits in the cache, from which point it
1213 benefits from the increased memory speed.
1215 Quicksort was rescued by implementing a sort pragma to control aspects
1216 of the sort. The B<stable> subpragma forces stable behaviour,
1217 regardless of algorithm. The B<_quicksort> and B<_mergesort>
1218 subpragmas are heavy-handed ways to select the underlying implementation.
1219 The leading C<_> is a reminder that these subpragmas may not survive
1220 beyond 5.8. More appropriate mechanisms for selecting the implementation
1221 exist, but they wouldn't have arrived in time to save quicksort.
1225 Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm
1226 (http://burtleburtle.net/bob/hash/doobs.html). This algorithm is
1227 reasonably fast while producing a much better spread of values than
1228 the old hashing algorithm (originally by Chris Torek, later tweaked by
1229 Ilya Zakharevich). Hash values output from the algorithm on a hash of
1230 all 3-char printable ASCII keys comes much closer to passing the
1231 DIEHARD random number generation tests. According to perlbench, this
1232 change has not affected the overall speed of Perl.
1236 unshift() should now be noticeably faster.
1240 =head1 Installation and Configuration Improvements
1242 =head2 Generic Improvements
1248 INSTALL now explains how you can configure Perl to use 64-bit
1249 integers even on non-64-bit platforms.
1253 Policy.sh policy change: if you are reusing a Policy.sh file
1254 (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1255 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1256 them will now be changed to the new prefix, /foo/bar. (Previously
1257 only $prefix changed.) If you do not like this new behaviour,
1258 specify prefix, siteprefix, and vendorprefix explicitly.
1262 A new optional location for Perl libraries, otherlibdirs, is available.
1263 It can be used for example for vendor add-ons without disturbing Perl's
1264 own library directories.
1268 In many platforms the vendor-supplied 'cc' is too stripped-down to
1269 build Perl (basically, 'cc' doesn't do ANSI C). If this seems
1270 to be the case and 'cc' does not seem to be the GNU C compiler
1271 'gcc', an automatic attempt is made to find and use 'gcc' instead.
1275 gcc needs to closely track the operating system release to avoid
1276 build problems. If Configure finds that gcc was built for a different
1277 operating system release than is running, it now gives a clearly visible
1278 warning that there may be trouble ahead.
1282 If binary compatibility with the 5.005 release is not wanted, Configure
1283 no longer suggests including the 5.005 modules in @INC.
1287 Configure C<-S> can now run non-interactively.
1291 configure.gnu now works with options with whitespace in them.
1295 installperl now outputs everything to STDERR.
1299 $Config{byteorder} is now computed dynamically (this is more robust
1300 with "fat binaries" where an executable image contains binaries for
1301 more than one binary platform.)
1305 Because PerlIO is now the default on most platforms, "-perlio" doesn't
1306 get appended to the $Config{archname} (also known as $^O) anymore.
1307 Instead, if you explicitly choose not to use perlio (Configure command
1308 line option -Uuseperlio), you will get "-stdio" appended.
1312 Another change related to the architecture name is that "-64all"
1313 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1314 pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)
1318 In AFS installations one can configure the root of the AFS to be
1319 somewhere else than the default F</afs> by using the Configure
1320 parameter C<-Dafsroot=/some/where/else>.
1324 APPLLIB_EXP, a less-know configuration-time definition, has been
1325 documented. It can be used to prepend site-specific directories
1326 to Perl's default search path (@INC), see INSTALL for information.
1330 The version of Berkeley DB used when the Perl (and, presumably, the
1331 DB_File extension) was built is now available as
1332 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
1333 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
1334 DB_VERSION_PATCH_CFG> from C.
1338 Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM
1339 has been documented in INSTALL.
1343 If you have CPAN access (either network or a local copy such as a
1344 CD-ROM) you can during specify extra modules to Configure to build and
1345 install with Perl using the -Dextras=... option. See INSTALL for
1350 In addition to config.over a new override file, config.arch, is
1351 available. That is supposed to be used by hints file writers for
1352 architecture-wide changes (as opposed to config.over which is for
1357 If your file system supports symbolic links you can build Perl outside
1358 of the source directory by
1360 mkdir /tmp/perl/build/directory
1361 cd /tmp/perl/build/directory
1362 sh /path/to/perl/source/Configure -Dmksymlinks ...
1364 This will create in /tmp/perl/build/directory a tree of symbolic links
1365 pointing to files in /path/to/perl/source. The original files are left
1366 unaffected. After Configure has finished you can just say
1370 and Perl will be built and tested, all in /tmp/perl/build/directory.
1374 For Perl developers several new make targets for profiling
1375 and debugging have been added, see L<perlhack>.
1381 Use of the F<gprof> tool to profile Perl has been documented in
1382 L<perlhack>. There is a make target called "perl.gprof" for
1383 generating a gprofiled Perl executable.
1387 If you have GCC 3, there is a make target called "perl.gcov" for
1388 creating a gcoved Perl executable for coverage analysis. See
1393 If you are on IRIX or Tru64 platforms, new profiling/debugging options
1394 have been added, see L<perlhack> for more information about pixie and
1401 Guidelines of how to construct minimal Perl installations have
1402 been added to INSTALL.
1406 The Thread extension is now not built at all under ithreads
1407 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
1408 Thread extension requires being Configured with C<-Duse5005threads>).
1410 But note that the Thread.pm interface is now shared by both
1415 =head2 New Or Improved Platforms
1417 For the list of platforms known to support Perl,
1418 see L<perlport/"Supported Platforms">.
1424 AIX dynamic loading should be now better supported.
1428 AIX should now work better with gcc, threads, and 64-bitness. Also the
1429 long doubles support in AIX should be better now. See L<perlaix>.
1433 After a long pause, AmigaOS has been verified to be happy with Perl.
1437 AtheOS (http://www.atheos.cx/) is a new platform.
1441 DG/UX platform now supports the 5.005-style threads. See L<perldgux>.
1445 DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near osvers 4.5.2.
1449 EBCDIC platforms (z/OS, also known as OS/390, POSIX-BC, and VM/ESA)
1450 have been regained. Many test suite tests still fail and the
1451 co-existence of Unicode and EBCDIC isn't quite settled, but the
1452 situation is much better than with Perl 5.6. See L<perlos390>,
1453 L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information.
1457 Building perl with -Duseithreads or -Duse5005threads now works under
1458 HP-UX 10.20 (previously it only worked under 10.30 or later). You will
1459 need a thread library package installed. See README.hpux.
1463 MacOS Classic (MacPerl has of course been available since
1464 perl 5.004 but now the source code bases of standard Perl
1465 and MacPerl have been synchronised)
1469 MacOS X (or Darwin) should now be able to build Perl even on HFS+
1470 filesystems. (The case-insensitivity confused the Perl build process.)
1474 NCR MP-RAS is now supported.
1478 NetWare from Novell is now supported. See L<perlnetware>.
1482 NonStop-UX is now supported.
1486 Amdahl UTS UNIX mainframe platform is now supported.
1490 WinCE is now supported. See L<perlce>.
1494 z/OS (formerly known as OS/390, formerly known as MVS OE) has now
1495 support for dynamic loading. This is not selected by default,
1496 however, you must specify -Dusedl in the arguments of Configure.
1500 =head1 Selected Bug Fixes
1502 Numerous memory leaks and uninitialized memory accesses have been
1503 hunted down. Most importantly anonymous subs used to leak quite
1510 The autouse pragma didn't work for Multi::Part::Function::Names.
1514 chop(@list) in list context returned the characters chopped in
1515 reverse order. This has been reversed to be in the right order.
1519 Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm)
1520 when building the Perl binary. The only exception to this is SunOS 4.x,
1525 The behaviour of non-decimal but numeric string constants such as
1526 "0x23" was platform-dependent: in some platforms that was seen as 35,
1527 in some as 0, in some as a floating point number (don't ask). This
1528 was caused by Perl using the operating system libraries in a situation
1529 where the result of the string to number conversion is undefined: now
1530 Perl consistently handles such strings as zero in numeric contexts.
1534 The order of DESTROYs has been made more predictable.
1538 Several debugger fixes: exit code now reflects the script exit code,
1539 condition C<"0"> now treated correctly, the C<d> command now checks
1540 line number, the C<$.> no longer gets corrupted, all debugger output now
1541 goes correctly to the socket if RemotePort is set.
1545 L<dprofpp> -R didn't work.
1549 C<*foo{FORMAT}> now works.
1553 UNIVERSAL::isa no longer caches methods incorrectly. (This broke
1554 the Tk extension with 5.6.0.)
1558 Lexicals I: lexicals outside an eval "" weren't resolved
1559 correctly inside a subroutine definition inside the eval "" if they
1560 were not already referenced in the top level of the eval""ed code.
1564 Lexicals II: lexicals leaked at file scope into subroutines that
1565 were declared before the lexicals.
1569 Lexical warnings now propagating correctly between scopes.
1573 Line renumbering with eval and C<#line> now works.
1577 Fixed numerous memory leaks, especially in eval "".
1581 mkdir() now ignores trailing slashes in the directory name,
1582 as mandated by POSIX.
1586 Some versions of glibc have a broken modfl(). This affects builds
1587 with C<-Duselongdouble>. This version of Perl detects this brokenness
1588 and has a workaround for it. The glibc release 2.2.2 is known to have
1589 fixed the modfl() bug.
1593 Modulus of unsigned numbers now works (4063328477 % 65535 used to
1594 return 27406, instead of 27047).
1598 Some "not a number" warnings introduced in 5.6.0 eliminated to be
1599 more compatible with 5.005. Infinity is now recognised as a number.
1603 Attributes (like :shared) didn't work with our().
1607 our() variables will not cause "will not stay shared" warnings.
1611 pack "Z" now correctly terminates the string with "\0".
1615 Fix password routines which in some shadow password platforms
1616 (e.g. HP-UX) caused getpwent() to return every other entry.
1620 The PERL5OPT environment variable (for passing command line arguments
1621 to Perl) didn't work for more than a single group of options.
1625 PERL5OPT with embedded spaces didn't work.
1629 printf() no longer resets the numeric locale to "C".
1633 C<q(a\\b)> now parses correctly as C<'a\\b'>.
1637 Printing quads (64-bit integers) with printf/sprintf now works
1638 without the q L ll prefixes (assuming you are on a quad-capable platform).
1642 Regular expressions on references and overloaded scalars now work.
1646 Right-hand side magic (GMAGIC) could in many cases such as string
1647 concatenation be invoked too many times.
1651 scalar() now forces scalar context even when used in void context.
1655 SOCKS support is now much more robust.
1659 sort() arguments are now compiled in the right wantarray context
1660 (they were accidentally using the context of the sort() itself).
1664 Changed the POSIX character class C<[[:space:]]> to include the (very
1665 rarely used) vertical tab character. Added a new POSIX-ish character
1666 class C<[[:blank:]]> which stands for horizontal whitespace
1667 (currently, the space and the tab).
1671 The tainting behaviour of sprintf() has been rationalized. It does
1672 not taint the result of floating point formats anymore, making the
1673 behaviour consistent with that of string interpolation.
1677 The regular expression captured submatches ($1, $2, ...) are now
1678 more consistently unset if the match fails, instead of leaving false
1679 data lying around in them.
1683 C<Sys::Syslog> ignored the C<LOG_AUTH> constant.
1687 All but the first argument of the IO syswrite() method are now optional.
1691 $AUTOLOAD, sort(), lock(), and spawning subprocesses
1692 in multiple threads simultaneously are now thread-safe.
1696 Tie::ARRAY SPLICE method was broken.
1700 Allow read-only string on left hand side of non-modifying tr///.
1704 Several Unicode fixes.
1710 BOMs (byte order marks) in the beginning of Perl files
1711 (scripts, modules) should now be transparently skipped.
1712 UTF-16 (UCS-2) encoded Perl files should now be read correctly.
1716 The character tables have been updated to Unicode 3.1.1.
1720 Comparing with utf8 data does not magically upgrade non-utf8 data
1725 C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase.
1729 Concatenation with the C<.> operator or via variable interpolation,
1730 C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator,
1731 substitution with C<s///>, single-quoted UTF8, should now work.
1735 The C<tr///> operator now works. Note that the C<tr///CU>
1736 functionality has been removed (but see pack('U0', ...)).
1740 C<eval "v200"> now works.
1744 Zero entries were missing from the Unicode classes like C<IsDigit>.
1750 =head2 Platform Specific Changes and Fixes
1758 Perl now works on post-4.0 BSD/OSes.
1764 Setting C<$0> now works (as much as possible; see L<perlvar> for details).
1770 Numerous updates; currently synchronised with Cygwin 1.1.4.
1774 Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.
1780 EPOC update after Perl 5.6.0. See README.epoc.
1786 Perl now works on post-3.0 FreeBSDs.
1792 README.hpux updated; C<Configure -Duse64bitall> now almost works.
1798 Numerous compilation flag and hint enhancements; accidental mixing
1799 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
1809 Long doubles should now work (see INSTALL).
1813 Linux previously had problems related to sockaddrlen when using
1814 accept(), revcfrom() (in Perl: recv()), getpeername(), and getsockname().
1822 Compilation of the standard Perl distribution in MacOS Classic should
1823 now work if you have the Metrowerks development environment and
1824 the missing Mac-specific toolkit bits. Contact the macperl mailing
1831 MPE/iX update after Perl 5.6.0. See README.mpeix.
1837 Perl now works on NetBSD/sparc.
1843 Now works with usethreads (see INSTALL).
1849 64-bitness using the Sun Workshop compiler now works.
1853 Tru64 (aka Digital UNIX, aka DEC OSF/1)
1855 The operating system version letter now recorded in $Config{osvers}.
1856 Allow compiling with gcc (previously explicitly forbidden). Compiling
1857 with gcc still not recommended because buggy code results, even with
1864 Fixed various alignment problems that lead into core dumps either
1865 during build or later; no longer dies on math errors at runtime;
1866 now using full quad integers (64 bits), previously was using
1867 only 46 bit integers for speed.
1873 chdir() now works better despite a CRT bug; now works with MULTIPLICITY
1874 (see INSTALL); now works with Perl's malloc.
1884 accept() no longer leaks memory.
1888 Borland C++ v5.5 is now a supported compiler that can build Perl.
1889 However, the generated binaries continue to be incompatible with those
1890 generated by the other supported compilers (GCC and Visual C++).
1894 Better chdir() return value for a non-existent directory.
1898 Duping socket handles with open(F, ">&MYSOCK") now works under Windows 9x.
1902 New %ENV entries now propagate to subprocesses.
1906 $ENV{LIB} now used to search for libs under Visual C.
1910 Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root.
1911 Other bugs in chdir() and Cwd::cwd() have also been fixed.
1915 A failed (pseudo)fork now returns undef and sets errno to EAGAIN.
1919 HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html
1923 The makefiles now provide a single switch to bulk-enable all the features
1924 enabled in ActiveState ActivePerl (a popular Win32 binary distribution).
1928 Allow REG_EXPAND_SZ keys in the registry.
1932 Can now send() from all threads, not just the first one.
1936 Fake signal handling reenabled, bugs and all.
1940 Less stack reserved per thread so that more threads can run
1941 concurrently. (Still 16M per thread.)
1945 C<File::Spec->tmpdir()> now prefers C:/temp over /tmp
1946 (works better when perl is running as service).
1950 Better UNC path handling under ithreads.
1954 wait() and waitpid() now work much better.
1958 winsock handle leak fixed.
1964 =head1 New or Changed Diagnostics
1970 All regular expression compilation error messages are now hopefully
1971 easier to understand both because the error message now comes before
1972 the failed regex and because the point of failure is now clearly
1973 marked by a C<E<lt>-- HERE> marker.
1977 The various "opened only for", "on closed", "never opened" warnings
1978 drop the C<main::> prefix for filehandles in the C<main> package,
1979 for example C<STDIN> instead of C<main::STDIN>.
1983 The "Unrecognized escape" warning has been extended to include C<\8>,
1984 C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters.
1988 Two new debugging options have been added: if you have compiled your
1989 Perl with debugging, you can use the -DT and -DR options to trace
1990 tokenising and to add reference counts to displaying variables,
1995 If an attempt to use a (non-blessed) reference as an array index
1996 is made, a warning is given.
2000 C<push @a;> and C<unshift @a;> (with no values to push or unshift)
2001 now give a warning. This may be a problem for generated and evaled
2006 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
2007 using the C<"C"> format you will get an optional warning. Similarly
2008 for the C<"c"> format and a number less than -128 or more than 127.
2012 Certain regex modifiers such as C<(?o)> make sense only if applied to
2013 the entire regex. You will an optional warning if you try to do otherwise.
2017 Using arrays or hashes as references (e.g. C<< %foo->{bar} >>
2018 has been deprecated for a while. Now you will get an optional warning.
2022 =head1 Changed Internals
2028 perlapi.pod (a companion to perlguts) now attempts to document the
2033 You can now build a really minimal perl called microperl.
2034 Building microperl does not require even running Configure;
2035 C<make -f Makefile.micro> should be enough. Beware: microperl makes
2036 many assumptions, some of which may be too bold; the resulting
2037 executable may crash or otherwise misbehave in wondrous ways.
2038 For careful hackers only.
2042 Added rsignal(), whichsig(), do_join(), op_clear, op_null,
2043 ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2044 interfaces to the publicised API. For the full list of the available
2045 APIs see L<perlapi>.
2049 Made possible to propagate customised exceptions via croak()ing.
2053 Now xsubs can have attributes just like subs.
2057 dTHR and djSP have been obsoleted; the former removed (because it's
2058 a no-op) and the latter replaced with dSP.
2062 PERL_OBJECT has been completely removed.
2066 The MAGIC constants (e.g. C<'P'>) have been macrofied
2067 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
2068 and maintainability.
2072 The regex compiler now maintains a structure that identifies nodes in
2073 the compiled bytecode with the corresponding syntactic features of the
2074 original regex expression. The information is attached to the new
2075 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
2076 complete information.
2080 The C code has been made much more C<gcc -Wall> clean. Some warning
2081 messages still remain in some platforms, so if you are compiling with
2082 gcc you may see some warnings about dubious practices. The warnings
2083 are being worked on.
2087 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
2091 Documentation on how to use the Perl source repository has been added
2092 to F<Porting/repository.pod>.
2096 There are now several profiling make targets.
2100 =head1 Security Vulnerability Closed
2102 (This change was already made in 5.7.0 but bears repeating here.)
2104 A potential security vulnerability in the optional suidperl component
2105 of Perl was identified in August 2000. suidperl is neither built nor
2106 installed by default. As of November 2001 the only known vulnerable
2107 platform is Linux, most likely all Linux distributions. CERT and
2108 various vendors and distributors have been alerted about the vulnerability.
2109 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2110 for more information.
2112 The problem was caused by Perl trying to report a suspected security
2113 exploit attempt using an external program, /bin/mail. On Linux
2114 platforms the /bin/mail program had an undocumented feature which
2115 when combined with suidperl gave access to a root shell, resulting in
2116 a serious compromise instead of reporting the exploit attempt. If you
2117 don't have /bin/mail, or if you have 'safe setuid scripts', or if
2118 suidperl is not installed, you are safe.
2120 The exploit attempt reporting feature has been completely removed from
2121 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2122 from all the Perl 5.7 releases), so that particular vulnerability
2123 isn't there anymore. However, further security vulnerabilities are,
2124 unfortunately, always possible. The suidperl functionality is most
2125 probably going to be removed in Perl 5.10. In any case, suidperl
2126 should only be used by security experts who know exactly what they are
2127 doing and why they are using suidperl instead of some other solution
2128 such as sudo (see http://www.courtesan.com/sudo/).
2132 Several new tests have been added, especially for the F<lib>
2133 subsection. There are now about 34 000 individual tests (spread over
2134 about 530 test scripts), in the regression suite (5.6.1 has about
2135 11700 tests, in 258 test scripts) Many of the new tests are introduced
2136 by the new modules, but still in general Perl is now more thoroughly
2139 Because of the large number of tests, running the regression suite
2140 will take considerably longer time than it used to: expect the suite
2141 to take up to 4-5 times longer to run than in perl 5.6. In a really
2142 fast machine you can hope to finish the suite in about 5 minutes
2145 The tests are now reported in a different order than in earlier Perls.
2146 (This happens because the test scripts from under t/lib have been moved
2147 to be closer to the library/extension they are testing.)
2149 =head1 Known Problems
2151 Note that unlike other sections in this document (which describe
2152 changes since 5.7.0) this section is cumulative containing known
2153 problems for all the 5.7 releases.
2161 In AIX 4.2 Perl extensions that use C++ functions that use statics
2162 may have problems in that the statics are not getting initialized.
2163 In newer AIX releases this has been solved by linking Perl with
2164 the libC_r library, but unfortunately in AIX 4.2 the said library
2165 has an obscure bug where the various functions related to time
2166 (such as time() and gettimeofday()) return broken values, and
2167 therefore in AIX 4.2 Perl is not linked against the libC_r.
2171 vac 5.0.0.0 May Produce Buggy Code For Perl
2173 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
2174 resulting in few random tests failing, but when the failing tests
2175 are run by hand, they succeed. We suggest upgrading to at least
2176 vac version 5.0.1.0, that has been known to compile Perl correctly.
2177 "lslpp -L|grep vac.C" will tell you the vac version.
2181 =head2 Amiga Perl Invoking Mystery
2183 One cannot call Perl using the C<volume:> syntax, that is, C<perl -v>
2184 works, but for example C<bin:perl -v> doesn't. The exact reason isn't
2185 known but the current suspect is the F<ixemul> library.
2187 =head2 lib/ftmp-security tests warn 'system possibly insecure'
2189 Don't panic. Read INSTALL 'make test' section instead.
2191 =head2 Cygwin intermittent failures of lib/Memoize/t/expire_file 11 and 12
2193 The subtests 11 and 12 sometimes fail and sometimes work.
2195 =head2 HP-UX lib/io_multihomed Fails When LP64-Configured
2197 The lib/io_multihomed test may hang in HP-UX if Perl has been
2198 configured to be 64-bit. Because other 64-bit platforms do not hang in
2199 this test, HP-UX is suspect. All other tests pass in 64-bit HP-UX. The
2200 test attempts to create and connect to "multihomed" sockets (sockets
2201 which have multiple IP addresses).
2203 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
2205 If perl is configured with -Duse64bitall, the successful result of the
2206 subtest 10 of lib/posix may arrive before the successful result of the
2207 subtest 9, which confuses the test harness so much that it thinks the
2210 =head2 Linux With Sfio Fails op/misc Test 48
2216 OS/390 has rather many test failures but the situation is actually
2217 better than it was in 5.6.0, it's just that so many new modules and
2218 tests have been added.
2220 Failed Test Stat Wstat Total Fail Failed List of Failed
2221 -----------------------------------------------------------------------------
2222 ../ext/B/Deparse.t 14 1 7.14% 14
2223 ../ext/B/Showlex.t 1 1 100.00% 1
2224 ../ext/Encode/Encode/Tcl.t 610 13 2.13% 592 594 596 598
2226 ../ext/IO/lib/IO/t/io_unix.t 113 28928 5 3 60.00% 3-5
2227 ../ext/POSIX/POSIX.t 29 1 3.45% 14
2228 ../ext/Storable/t/lock.t 255 65280 5 3 60.00% 3-5
2229 ../lib/locale.t 129 33024 117 19 16.24% 99-117
2230 ../lib/warnings.t 434 1 0.23% 75
2231 ../lib/ExtUtils.t 27 1 3.70% 25
2232 ../lib/Math/BigInt/t/bigintpm.t 1190 1 0.08% 1145
2233 ../lib/Unicode/UCD.t 81 48 59.26% 1-16 49-64 66-81
2234 ../lib/User/pwent.t 9 1 11.11% 4
2235 op/pat.t 660 6 0.91% 242-243 424-425
2237 op/split.t 0 9 ?? ?? % ??
2238 op/taint.t 174 3 1.72% 156 162 168
2239 op/tr.t 70 3 4.29% 50 58-59
2240 Failed 16/422 test scripts, 96.21% okay. 105/23251 subtests failed, 99.55% okay.
2242 =head2 op/sprintf tests 129 and 130
2244 The op/sprintf tests 129 and 130 are known to fail on some platforms.
2245 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
2246 The failing platforms do not comply with the ANSI C Standard, line
2247 19ff on page 134 of ANSI X3.159 1989 to be exact. (They produce
2248 something other than "1" and "-1" when formatting 0.6 and -0.6 using
2249 the printf format "%.0f", most often they produce "0" and "-0".)
2251 =head2 Failure of Thread tests
2253 B<Note that support for 5.005-style threading remains experimental.>
2255 The following tests are known to fail due to fundamental problems in
2256 the 5.005 threading implementation. These are not new failures--Perl
2257 5.005_0x has the same bugs, but didn't have these tests.
2260 t/lib/thr5005.t 19-20
2268 ext/POSIX/sigaction subtests 6 and 13 may fail.
2272 lib/ExtUtils may spuriously claim that subtest 28 failed,
2273 which is interesting since the test only has 27 tests.
2277 Numerous numerical test failures
2279 op/numconvert 209,210,217,218
2281 ext/Time/HiRes/HiRes 9
2282 lib/Math/BigInt/t/bigintpm 1145
2285 These tests fail because of yet unresolved floating point inaccuracies.
2291 There are a few known test failures, see L<perluts>.
2295 Rather a lot of tests are failing in VMS, but actually more tests
2296 succeed in VMS than they used to; it's just that there are many,
2297 many more tests than there used to be.
2299 Here are the known failures from some compiler/platform combinations.
2301 Compaq C V6.2-009 on OpenVMS Alpha V7.3
2303 [.run]switches..........................FAILED on test 1
2304 [-.ext.posix.t]posix....................FAILED on test 10
2305 [-.ext.time.hires]hires.................FAILED on test 17
2306 [-.lib]db...............................FAILED on test 24
2307 [-.lib.net]hostent......................FAILED on test 5
2308 [-.lib.pod.t]basic......................FAILED on test 10
2312 In multi-CPU boxes there are some problems with the I/O buffering:
2313 some output may appear twice.
2315 =head2 Localising a Tied Variable Leaks Memory
2318 tie my %tie_hash => 'Tie::StdHash';
2322 local($tie_hash{Foo}) = 1; # leaks
2324 Code like the above is known to leak memory every time the local()
2327 =head2 Localising Tied Arrays and Hashes Is Broken
2331 doesn't work as one would expect: the old value is restored
2334 =head2 Self-tying of Arrays and Hashes Is Forbidden
2336 Self-tying of arrays and hashes is broken in rather deep and
2337 hard-to-fix ways. As a stop-gap measure to avoid people from getting
2338 frustrated at the mysterious results (core dumps, most often) it is
2339 for now forbidden (you will get a fatal error even from an attempt).
2341 =head2 Variable Attributes are not Currently Usable for Tieing
2343 This limitation will hopefully be fixed in future. (Subroutine
2344 attributes work fine for tieing, see L<Attribute::Handlers>).
2346 One way to run into this limitation is to have a loop variable with
2347 attributes within a loop: the tie is called only once, not for each
2348 iteration of the loop.
2350 =head2 Building Extensions Can Fail Because Of Largefiles
2352 Some extensions like mod_perl are known to have issues with
2353 `largefiles', a change brought by Perl 5.6.0 in which file offsets
2354 default to 64 bits wide, where supported. Modules may fail to compile
2355 at all or compile and work incorrectly. Currently there is no good
2356 solution for the problem, but Configure now provides appropriate
2357 non-largefile ccflags, ldflags, libswanted, and libs in the %Config
2358 hash (e.g., $Config{ccflags_nolargefiles}) so the extensions that are
2359 having problems can try configuring themselves without the
2360 largefileness. This is admittedly not a clean solution, and the
2361 solution may not even work at all. One potential failure is whether
2362 one can (or, if one can, whether it's a good idea) link together at
2363 all binaries with different ideas about file offsets, all this is
2366 =head2 Unicode Support on EBCDIC Still Spotty
2368 Though mostly working, Unicode support still has problem spots on
2369 EBCDIC platforms. One such known spot are the C<\p{}> and C<\P{}>
2370 regular expression constructs for code points less than 256: the
2371 pP are testing for Unicode code points, not knowing about EBCDIC.
2373 =head2 The Compiler Suite Is Still Experimental
2375 The compiler suite is slowly getting better but is nowhere near
2378 =head2 The Long Double Support is Still Experimental
2380 The ability to configure Perl's numbers to use "long doubles",
2381 floating point numbers of hopefully better accuracy, is still
2382 experimental. The implementations of long doubles are not yet
2383 widespread and the existing implementations are not quite mature
2384 or standardised, therefore trying to support them is a rare
2385 and moving target. The gain of more precision may also be offset
2386 by slowdown in computations (more bits to move around, and the
2387 operations are more likely to be executed by less optimised
2390 =head1 Reporting Bugs
2392 If you find what you think is a bug, you might check the articles
2393 recently posted to the comp.lang.perl.misc newsgroup and the perl
2394 bug database at http://bugs.perl.org. There may also be
2395 information at http://www.perl.com/perl/, the Perl Home Page.
2397 If you believe you have an unreported bug, please run the B<perlbug>
2398 program included with your release. Be sure to trim your bug down
2399 to a tiny but sufficient test case. Your bug report, along with the
2400 output of C<perl -V>, will be sent off to perlbug@perl.org to be
2401 analysed by the Perl porting team.
2405 The F<Changes> file for exhaustive details on what changed.
2407 The F<INSTALL> file for how to build Perl.
2409 The F<README> file for general stuff.
2411 The F<Artistic> and F<Copying> files for copyright information.
2415 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>.