3 perl572delta - what's new for perl v5.7.2
7 This document describes differences between the 5.7.1 release and the
10 (To view the differences between the 5.6.0 release and the 5.7.0
11 release, see L<perl570delta>. To view the differences between the
12 5.7.0 release and the 5.7.1 release, see L<perl571delta>.)
14 =head1 Security Vulnerability Closed
16 (This change was already made in 5.7.0 but bears repeating here.)
18 A security vulnerability affecting all Perl versions prior to 5.6.1
19 was found in August 2000. The vulnerability does not affect default
20 installations and as far as is known affects only the Linux platform.
22 You should upgrade your Perl to 5.6.1 as soon as possible. Patches
23 for earlier releases exist but using the patches require full
24 recompilation from the source code anyway, so 5.6.1 is your best
27 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
30 =head1 Incompatible Changes
32 =head2 64-bit platforms and malloc
34 If your pointers are 64 bits wide, the Perl malloc is no more being
35 used because it simply does not work with 8-byte pointers. Also,
36 usually the system malloc on such platforms are much better optimized
37 for such large memory models than the Perl malloc.
39 =head2 AIX Dynaloading
41 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
42 dlopen interface of AIX instead of the old emulated interface. This
43 change will probably break backward compatibility with compiled
44 modules. The change was made to make Perl more compliant with other
45 applications like modperl which are using the AIX native interface.
47 =head2 Socket Extension Dynamic in VMS
49 The Socket extension is now dynamically loaded instead of being
50 statically built in. This may or may not be a problem with ancient
51 TCP/IP stacks of VMS: we do not know since we weren't able to test
52 Perl in such configurations.
54 =head2 Different Definition of the Unicode Character Classes \p{In...}
56 As suggested by the Unicode consortium, the Unicode character classes
57 now prefer I<scripts> as opposed to I<blocks> (as defined by Unicode);
58 in Perl, when the C<\p{In....}> and the C<\p{In....}> regular expression
59 constructs are used. This has changed the definition of some of those
62 The difference between scripts and blocks is that scripts are the
63 glyphs used by a language or a group of languages, while the blocks
64 are more artificial groupings of 256 characters based on the Unicode
67 In general this change results in more inclusive Unicode character
68 classes, but changes to the other direction also do take place:
69 for example while the script C<Latin> includes all the Latin
70 characters and their various diacritic-adorned versions, it
71 does not include the various punctuation or digits (since they
72 are not solely C<Latin>).
74 Changes in the character class semantics may have happened if a script
75 and a block happen to have the same name, for example C<Hebrew>.
76 In such cases the script wins and C<\p{InHebrew}> now means the script
77 definition of Hebrew. The block definition in still available,
78 though, by appending C<Block> to the name: C<\p{InHebrewBlock}> means
79 what C<\p{InHebrew}> meant in perl 5.6.0. For the full list
80 of affected character classes, see L<perlunicode/Blocks>.
84 The current user-visible implementation of pseudo-hashes (the weird
85 use of the first array element) is deprecated starting from Perl 5.8.0
86 and will be removed in Perl 5.10.0, and the feature will be
87 implemented differently. Not only is the current interface rather
88 ugly, but the current implementation slows down normal array and hash
89 use quite noticeably. The C<fields> pragma interface will remain
92 The syntaxes C<@a->[...]> and C<@h->{...}> have now been deprecated.
94 The suidperl is also considered to be too much a risk to continue
95 maintaining and the suidperl code is likely to be removed in a future
98 The C<package;> syntax (C<package> without an argument has been
99 deprecated. Its semantics were never that clear and its
100 implementation even less so. If you have used that feature to
101 disallow all but fully qualified variables, C<use strict;> instead.
103 =head1 Core Enhancements
105 In general a lot of fixing has happened in the area of Perl's
106 understanding of numbers, both integer and floating point. Since in
107 many systems the standard number parsing functions like C<strtoul()>
108 and C<atof()> seem to have bugs, Perl tries to work around their
109 deficiencies. This results hopefully in more accurate numbers.
115 The rules for allowing underscores (underbars) in numeric constants
116 have been relaxed and simplified: now you can have an underscore
121 GMAGIC (right-hand side magic) could in many cases such as string
122 concatenation be invoked too many times.
126 Lexicals I: lexicals outside an eval "" weren't resolved
127 correctly inside a subroutine definition inside the eval "" if they
128 were not already referenced in the top level of the eval""ed code.
132 Lexicals II: lexicals leaked at file scope into subroutines that
133 were declared before the lexicals.
137 Lvalue subroutines can now return C<undef> in list context.
141 The C<op_clear> and C<op_null> are now exported.
145 A new special regular expression variable has been introduced:
146 C<$^N>, which contains the most-recently closed group (submatch).
150 L<utime> now supports C<utime undef, undef, @files> to change the
151 file timestamps to the current time.
155 The Perl parser has been stress tested using both random input and
160 C<eval "v200"> now works.
164 VMS now works under PerlIO.
168 =head1 Modules and Pragmata
170 =head2 New Modules and Distributions
176 L<Attribute::Handlers> - Simpler definition of attribute handlers
180 L<ExtUtils::Constant> - generate XS code to import C header constants
184 L<I18N::Langinfo> - query locale information
188 L<I18N::LangTags> - functions for dealing with RFC3066-style language tags
192 L<libnet> - a collection of perl5 modules related to network programming
194 Perl installation leaves libnet unconfigured, use F<libnetcfg> to configure.
198 L<List::Util> - selection of general-utility list subroutines
202 L<Locale::Maketext> - framework for localization
206 L<Memoize> - Make your functions faster by trading space for time
210 L<NEXT> - pseudo-class for method redispatch
214 L<Scalar::Util> - selection of general-utility scalar subroutines
218 L<Test::More> - yet another framework for writing test scripts
222 L<Test::Simple> - Basic utilities for writing tests
226 L<Time::HiRes> - high resolution ualarm, usleep, and gettimeofday
230 L<Time::Piece> - Object Oriented time objects
232 (Previously known as L<Time::Object>.)
236 L<Time::Seconds> - a simple API to convert seconds to other date values
240 L<UnicodeCD> - Unicode Character Database
244 =head2 Updated And Improved Modules and Pragmata
250 L<B::Deparse> module has been significantly enhanced. It now
251 can deparse almost all of the standard test suite (so that the
252 tests still succeed). There is a make target "test.deparse"
257 L<Class::Struct> now assigns the array/hash element if the accessor
258 is called with an array/hash element as the B<sole> argument.
262 L<Cwd> extension is now (even) faster.
266 L<DB_File> extension has been updated to version 1.77.
270 L<Fcntl>, L<Socket>, and L<Sys::Syslog> have been rewritten to use the
271 new-style constant dispatch section (see L<ExtUtils::Constant>).
275 L<File::Find> is now (again) reentrant. It also has been made
280 L<File::Glob> now supports C<GLOB_LIMIT> constant to limit the
281 size of the returned list of filenames.
285 L<IO::Socket::INET> now supports C<LocalPort> of zero (usually meaning
286 that the operating system will make one up.)
290 The L<vars> pragma now supports declaring fully qualified variables.
291 (Something that C<our()> does not and will not support.)
295 =head1 Utility Changes
301 The F<emacs/e2ctags.pl> is now much faster.
305 L<h2ph> now supports C trigraphs.
309 L<h2xs> uses the new L<ExtUtils::Constant> module which will affect
310 newly created extensions that define constants. Since the new code is
311 more correct (if you have two constants where the first one is a
312 prefix of the second one, the first constant B<never> gets defined),
313 less lossy (it uses integers for integer constant, as opposed to the
314 old code that used floating point numbers even for integer constants),
315 and slightly faster, you might want to consider regenerating your
316 extension code (the new scheme makes regenerating easy).
317 L<h2xs> now also supports C trigraphs.
321 L<libnetcfg> has been added to configure the libnet.
325 The F<Pod::Html> (and thusly L<pod2html>) now allows specifying
330 =head1 New Documentation
336 L<Locale::Maketext::TPJ13> is an article about software localization,
337 originally published in The Perl Journal #13, republished here with
342 More README.$PLATFORM files have been converted into pod, which also
343 means that they also be installed as perl$PLATFORM documentation
344 files. The new files are L<perlapollo>, L<perlbeos>, L<perldgux>,
345 L<perlhurd>, L<perlmint>, L<perlnetware>, L<perlplan9>, L<perlqnx>,
350 The F<Todo> and F<Todo-5.6> files have been merged into L<perltodo>.
354 Use of the F<gprof> tool to profile Perl has been documented in
355 L<perlhack>. There is a make target "perl.gprof" for generating a
356 gprofiled Perl executable.
360 =head1 Installation and Configuration Improvements
362 =head2 New Or Improved Platforms
368 AIX should now work better with gcc, threads, and 64-bitness. Also the
369 long doubles support in AIX should be better now. See L<perlaix>.
373 AtheOS (http://www.atheos.cx/) is a new platform.
377 DG/UX platform now supports the 5.005-style threads. See L<perldgux>.
381 DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near osvers 4.5.2.
385 Several MacOS (Classic) portability patches have been applied. We
386 hope to get a fully working port by 5.8.0. (The remaining problems
387 relate to the changed IO model of Perl.) See L<perlmacos>.
391 MacOS X (or Darwin) should now be able to build Perl even on HFS+
392 filesystems. (The case-insensitivity confused the Perl build process.)
396 NetWare from Novell is now supported. See L<perlnetware>.
400 The Amdahl UTS UNIX mainframe platform is now supported.
404 =head2 Generic Improvements
410 In AFS installations one can configure the root of the AFS to be
411 somewhere else than the default F</afs> by using the Configure
412 parameter C<-Dafsroot=/some/where/else>.
416 The version of Berkeley DB used when the Perl (and, presumably, the
417 DB_File extension) was built is now available as
418 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
419 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
420 DB_VERSION_PATCH_CFG> from C.
424 The Thread extension is now not built at all under ithreads
425 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
426 Thread extension requires being Configured with C<-Duse5005threads>).
430 The C<B::Deparse> compiler backend has been so significantly improved
431 that almost the whole Perl test suite passes after being deparsed. A
432 make target has been added to help in further testing: C<make test.deparse>.
436 =head1 Selected Bug Fixes
442 The autouse pragma didn't work for Multi::Part::Function::Names.
446 The behaviour of non-decimal but numeric string constants such as
447 "0x23" was platform-dependent: in some platforms that was seen as 35,
448 in some as 0, in some as a floating point number (don't ask). This
449 was caused by Perl using the operating system libraries in a situation
450 where the result of the string to number conversion is undefined: now
451 Perl consistently handles such strings as zero in numeric contexts.
455 L<dprofpp> -R didn't work.
459 PERL5OPT with embedded spaces didn't work.
463 L<Sys::Syslog> ignored the C<LOG_AUTH> constant.
467 =head2 Platform Specific Changes and Fixes
473 Some versions of glibc have a broken modfl(). This affects builds
474 with C<-Duselongdouble>. This version of Perl detects this brokenness
475 and has a workaround for it. The glibc release 2.2.2 is known to have
476 fixed the modfl() bug.
480 =head1 New or Changed Diagnostics
486 In the regular expression diagnostics the C<E<lt>E<lt> HERE> marker
487 introduced in 5.7.0 has been changed to be C<E<lt>-- HERE> since too
488 many people found the C<E<lt>E<lt>> to be too similar to here-document
493 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
494 using the C<"C"> format you will get an optional warning. Similarly
495 for the C<"c"> format and a number less than -128 or more than 127.
499 Certain regex modifiers such as C<(?o)> make sense only if applied to
500 the entire regex. You will an optional warning if you try to do otherwise.
504 Using arrays or hashes as references (e.g. C<%foo->{bar}> has been
505 deprecated for a while. Now you will get an optional warning.
509 =head1 Source Code Enhancements
511 =head2 MAGIC constants
513 The MAGIC constants (e.g. C<'P'>) have been macrofied
514 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
517 =head2 Better commented code
519 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
521 =head2 Regex pre-/post-compilation items matched up
523 The regex compiler now maintains a structure that identifies nodes in
524 the compiled bytecode with the corresponding syntactic features of the
525 original regex expression. The information is attached to the new
526 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
527 complete information.
531 The C code has been made much more C<gcc -Wall> clean. Some warning
532 messages still remain, though, so if you are compiling with gcc you
533 will see some warnings about dubious practices. The warnings are
538 Several new tests have been added, especially for the F<lib> subsection.
540 The tests are now reported in a different order than in earlier Perls.
541 (This happens because the test scripts from under t/lib have been moved
542 to be closer to the library/extension they are testing.)
544 =head1 Known Problems
546 Note that unlike other sections in this document (which describe
547 changes since 5.7.0) this section is cumulative containing known
548 problems for all the 5.7 releases.
556 In AIX 4.2 Perl extensions that use C++ functions that use statics
557 may have problems in that the statics are not getting initialized.
558 In newer AIX releases this has been solved by linking Perl with
559 the libC_r library, but unfortunately in AIX 4.2 the said library
560 has an obscure bug where the various functions related to time
561 (such as time() and gettimeofday()) return broken values, and
562 therefore in AIX 4.2 Perl is not linked against the libC_r.
566 vac 5.0.0.0 May Produce Buggy Code For Perl
568 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
569 resulting in few random tests failing, but when the failing tests
570 are run by hand, they succeed. We suggest upgrading to at least
571 vac version 5.0.1.0, that has been known to compile Perl correctly.
572 "lslpp -L|grep vac.C" will tell you the vac version.
576 =head2 Amiga Perl Invoking Mystery
578 One cannot call Perl using the C<volume:> syntax, that is, C<perl -v>
579 works, but for example C<bin:perl -v> doesn't. The exact reason is
580 known but the current suspect is the F<ixemul> library.
582 =head2 lib/ftmp-security tests warn 'system possibly insecure'
584 Don't panic. Read INSTALL 'make test' section instead.
586 =head2 Cygwin intermittent failures of lib/Memoize/t/expire_file 11 and 12
588 The subtests 11 and 12 sometimes fail and sometimes work.
590 =head2 HP-UX lib/io_multihomed Fails When LP64-Configur
592 The lib/io_multihomed test may hang in HP-UX if Perl has been
593 configured to be 64-bit. Because other 64-bit platforms do not hang in
594 this test, HP-UX is suspect. All other tests pass in 64-bit HP-UX. The
595 test attempts to create and connect to "multihomed" sockets (sockets
596 which have multiple IP addresses).
598 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
600 If perl is configured with -Duse64bitall, the successful result of the
601 subtest 10 of lib/posix may arrive before the successful result of the
602 subtest 9, which confuses the test harness so much that it thinks the
605 =head2 Linux With Sfio Fails op/misc Test 48
611 OS/390 has rather many test failures but the situation is actually
612 better than it was in 5.6.0, it's just that so many new modules and
613 tests have been added.
615 Failed Test Stat Wstat Total Fail Failed List of Failed
616 -----------------------------------------------------------------------------
617 ../ext/B/Deparse.t 14 1 7.14% 14
618 ../ext/B/Showlex.t 1 1 100.00% 1
619 ../ext/Encode/Encode/Tcl.t 610 13 2.13% 592 594 596 598
621 ../ext/IO/lib/IO/t/io_unix.t 113 28928 5 3 60.00% 3-5
622 ../ext/POSIX/POSIX.t 29 1 3.45% 14
623 ../ext/Storable/t/lock.t 255 65280 5 3 60.00% 3-5
624 ../lib/locale.t 129 33024 117 19 16.24% 99-117
625 ../lib/warnings.t 434 1 0.23% 75
626 ../lib/ExtUtils.t 27 1 3.70% 25
627 ../lib/Math/BigInt/t/bigintpm.t 1190 1 0.08% 1145
628 ../lib/Unicode/UCD.t 81 48 59.26% 1-16 49-64 66-81
629 ../lib/User/pwent.t 9 1 11.11% 4
630 op/pat.t 660 6 0.91% 242-243 424-425
632 op/split.t 0 9 ?? ?? % ??
633 op/taint.t 174 3 1.72% 156 162 168
634 op/tr.t 70 3 4.29% 50 58-59
635 Failed 16/422 test scripts, 96.21% okay. 105/23251 subtests failed, 99.55% okay.
637 =head2 op/sprintf tests 129 and 130
639 The op/sprintf tests 129 and 130 are known to fail on some platforms.
640 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
641 The failing platforms do not comply with the ANSI C Standard, line
642 19ff on page 134 of ANSI X3.159 1989 to be exact. (They produce
643 something other than "1" and "-1" when formatting 0.6 and -0.6 using
644 the printf format "%.0f", most often they produce "0" and "-0".)
646 =head2 Failure of Thread tests
648 B<Note that support for 5.005-style threading remains experimental.>
650 The following tests are known to fail due to fundamental problems in
651 the 5.005 threading implementation. These are not new failures--Perl
652 5.005_0x has the same bugs, but didn't have these tests.
655 t/lib/thr5005.t 19-20
663 ext/POSIX/sigaction subtests 6 and 13 may fail.
667 lib/ExtUtils may spuriously claim that subtest 28 failed,
668 which is interesting since the test only has 27 tests.
672 Numerous numerical test failures
674 op/numconvert 209,210,217,218
676 ext/Time/HiRes/HiRes 9
677 lib/Math/BigInt/t/bigintpm 1145
680 These tests fail because of yet unresolved floating point inaccuracies.
686 There are a few known test failures, see L<perluts>.
690 Rather many tests are failing in VMS but that actually more tests
691 succeed in VMS than they used to, it's just that there are many,
692 many more tests than there used to be.
694 Here are the known failures from some compiler/platform combinations.
696 DEC C V5.3-006 on OpenVMS VAX V6.2
698 [-.ext.list.util.t]tainted..............FAILED on test 3
699 [-.ext.posix]sigaction..................FAILED on test 7
700 [-.ext.time.hires]hires.................FAILED on test 14
701 [-.lib.file.find]taint..................FAILED on test 17
702 [-.lib.math.bigint.t]bigintpm...........FAILED on test 1183
703 [-.lib.test.simple.t]exit...............FAILED on test 1
704 [.lib]vmsish............................FAILED on test 13
705 [.op]sprintf............................FAILED on test 12
706 Failed 8/399 tests, 91.23% okay.
708 DEC C V6.0-001 on OpenVMS Alpha V7.2-1 and
709 Compaq C V6.2-008 on OpenVMS Alpha V7.1
711 [-.ext.list.util.t]tainted..............FAILED on test 3
712 [-.lib.file.find]taint..................FAILED on test 17
713 [-.lib.test.simple.t]exit...............FAILED on test 1
714 [.lib]vmsish............................FAILED on test 13
715 Failed 4/399 tests, 92.48% okay.
717 Compac C V6.4-005 on OpenVMS Alpha 7.2.1
719 [-.ext.b]showlex........................FAILED on test 1
720 [-.ext.list.util.t]tainted..............FAILED on test 3
721 [-.lib.file.find]taint..................FAILED on test 17
722 [-.lib.test.simple.t]exit...............FAILED on test 1
723 [.lib]vmsish............................FAILED on test 13
724 [.op]misc...............................FAILED on test 49
725 Failed 6/401 tests, 92.77% okay.
729 In multi-CPU boxes there are some problems with the I/O buffering:
730 some output may appear twice.
732 =head2 Localising a Tied Variable Leaks Memory
735 tie my %tie_hash => 'Tie::StdHash';
739 local($tie_hash{Foo}) = 1; # leaks
741 Code like the above is known to leak memory every time the local()
744 =head2 Self-tying of Arrays and Hashes Is Forbidden
746 Self-tying of arrays and hashes is broken in rather deep and
747 hard-to-fix ways. As a stop-gap measure to avoid people from getting
748 frustrated at the mysterious results (core dumps, most often) it is
749 for now forbidden (you will get a fatal error even from an attempt).
751 =head2 Variable Attributes are not Currently Usable for Tieing
753 This limitation will hopefully be fixed in future. (Subroutine
754 attributes work fine for tieing, see L<Attribute::Handlers>).
756 =head2 Building Extensions Can Fail Because Of Largefiles
758 Some extensions like mod_perl are known to have issues with
759 `largefiles', a change brought by Perl 5.6.0 in which file offsets
760 default to 64 bits wide, where supported. Modules may fail to compile
761 at all or compile and work incorrectly. Currently there is no good
762 solution for the problem, but Configure now provides appropriate
763 non-largefile ccflags, ldflags, libswanted, and libs in the %Config
764 hash (e.g., $Config{ccflags_nolargefiles}) so the extensions that are
765 having problems can try configuring themselves without the
766 largefileness. This is admittedly not a clean solution, and the
767 solution may not even work at all. One potential failure is whether
768 one can (or, if one can, whether it's a good idea) link together at
769 all binaries with different ideas about file offsets, all this is
772 =head2 The Compiler Suite Is Still Experimental
774 The compiler suite is slowly getting better but is nowhere near
777 =head2 The Long Double Support is Still Experimental
779 The ability to configure Perl's numbers to use "long doubles",
780 floating point numbers of hopefully better accuracy, is still
781 experimental. The implementations of long doubles are not yet
782 widespread and the existing implementations are not quite mature
783 or standardised, therefore trying to support them is a rare
784 and moving target. The gain of more precision may also be offset
785 by slowdown in computations (more bits to move around, and the
786 operations are more likely to be executed by less optimised
789 =head1 Reporting Bugs
791 If you find what you think is a bug, you might check the articles
792 recently posted to the comp.lang.perl.misc newsgroup and the perl
793 bug database at http://bugs.perl.org. There may also be
794 information at http://www.perl.com/perl/, the Perl Home Page.
796 If you believe you have an unreported bug, please run the B<perlbug>
797 program included with your release. Be sure to trim your bug down
798 to a tiny but sufficient test case. Your bug report, along with the
799 output of C<perl -V>, will be sent off to perlbug@perl.org to be
800 analysed by the Perl porting team.
804 The F<Changes> file for exhaustive details on what changed.
806 The F<INSTALL> file for how to build Perl.
808 The F<README> file for general stuff.
810 The F<Artistic> and F<Copying> files for copyright information.
814 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>, with many contributions
815 from The Perl Porters and Perl Users submitting feedback and patches.
817 Send omissions or corrections to <F<perlbug@perl.org>>.