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 Binary Incompatibility
51 B<Perl 5.8 is not binary compatible with earlier releases of Perl.>
53 B<You have to recompile your XS modules.>
55 (Pure Perl modules should continue to work.)
57 The major reason for the discontinuity is the new IO architecture
58 called PerlIO. PerlIO is the default configuration because without
59 it many new features of Perl 5.8 cannot be used. In other words:
60 you just have to recompile your modules containing XS code, sorry
63 In future releases of Perl, non-PerlIO aware XS modules may become
64 completely unsupported. This shouldn't be too difficult for module
65 authors, however: PerlIO has been designed as a drop-in replacement
66 (at the source code level) for the stdio interface.
68 Depending on your platform, there are also other reasons why
69 we decided to break binary compatibility, please read on.
71 =head2 64-bit platforms and malloc
73 If your pointers are 64 bits wide, the Perl malloc is no longer being
74 used because it does not work well with 8-byte pointers. Also,
75 usually the system mallocs on such platforms are much better optimized
76 for such large memory models than the Perl malloc. Some memory-hungry
77 Perl applications like the PDL don't work well with Perl's malloc.
78 Finally, other applications than Perl (such as mod_perl) tend to prefer
79 the system malloc. Such platforms include Alpha and 64-bit HPPA,
82 =head2 AIX Dynaloading
84 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
85 dlopen interface of AIX instead of the old emulated interface. This
86 change will probably break backward compatibility with compiled
87 modules. The change was made to make Perl more compliant with other
88 applications like mod_perl which are using the AIX native interface.
90 =head2 Attributes for C<my> variables now handled at run-time.
92 The C<my EXPR : ATTRS> syntax now applies variable attributes at
93 run-time. (Subroutine and C<our> variables still get attributes applied
94 at compile-time.) See L<attributes> for additional details. In particular,
95 however, this allows variable attributes to be useful for C<tie> interfaces,
96 which was a deficiency of earlier releases. Note that the new semantics
97 doesn't work with the Attribute::Handlers module (as of version 0.76).
99 =head2 Socket Extension Dynamic in VMS
101 The Socket extension is now dynamically loaded instead of being
102 statically built in. This may or may not be a problem with ancient
103 TCP/IP stacks of VMS: we do not know since we weren't able to test
104 Perl in such configurations.
106 =head2 IEEE-format Floating Point Default on OpenVMS Alpha
108 Perl now uses IEEE format (T_FLOAT) as the default internal floating
109 point format on OpenVMS Alpha, potentially breaking binary compatibility
110 with external libraries or existing data. G_FLOAT is still available as
111 a configuration option. The default on VAX (D_FLOAT) has not changed.
113 =head2 New Unicode Properties
115 Unicode I<scripts> are now supported. Scripts are similar to (and superior
116 to) Unicode I<blocks>. The difference between scripts and blocks is that
117 scripts are the glyphs used by a language or a group of languages, while
118 the blocks are more artificial groupings of (mostly) 256 characters based
119 on the Unicode numbering.
121 In general, scripts are more inclusive, but not universally so. For
122 example, while the script C<Latin> includes all the Latin characters and
123 their various diacritic-adorned versions, it does not include the various
124 punctuation or digits (since they are not solely C<Latin>).
126 A number of other properties are now supported, including C<\p{L&}>,
127 C<\p{Any}> C<\p{Assigned}>, C<\p{Unassigned}>, C<\p{Blank}> and
128 C<\p{SpacePerl}> (along with their C<\P{...}> versions, of course).
129 See L<perlunicode> for details, and more additions.
131 The C<In> or C<Is> prefix to names used with the C<\p{...}> and C<\P{...}>
132 are now almost always optional. The only exception is that a C<In> prefix
133 is required to signify a Unicode block when a block name conflicts with a
134 script name. For example, C<\p{Tibetan}> refers to the script, while
135 C<\p{InTibetan}> refers to the block. When there is no name conflict, you
136 can omit the C<In> from the block name (e.g. C<\p{BraillePatterns}>), but
137 to be safe, it's probably best to always use the C<In>).
139 =head2 REF(...) Instead Of SCALAR(...)
141 A reference to a reference now stringifies as "REF(0x81485ec)" instead
142 of "SCALAR(0x81485ec)" in order to be more consistent with the return
145 =head2 pack/unpack D/F recycled
147 The undocumented pack/unpack template letters D/F have been recycled
148 for better use: now they stand for long double (if supported by the
149 platform) and NV (Perl internal floating point type). (They used
150 to be aliases for d/f, but you never knew that.)
158 The semantics of bless(REF, REF) were unclear and until someone proves
159 it to make some sense, it is forbidden.
163 The obsolete chat2 library that should never have been allowed
164 to escape the laboratory has been decommissioned.
168 The builtin dump() function has probably outlived most of its
169 usefulness. The core-dumping functionality will remain in future
170 available as an explicit call to C<CORE::dump()>, but in future
171 releases the behaviour of an unqualified C<dump()> call may change.
175 The very dusty examples in the eg/ directory have been removed.
176 Suggestions for new shiny examples welcome but the main issue is that
177 the examples need to be documented, tested and (most importantly)
182 The (bogus) escape sequences \8 and \9 now give an optional warning
183 ("Unrecognized escape passed through"). There is no need to \-escape
188 The list of filenames from glob() (or <...>) is now by default sorted
189 alphabetically to be csh-compliant (which is what happened before
190 in most UNIX platforms). (bsd_glob() does still sort platform
191 natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.)
195 Spurious syntax errors generated in certain situations, when glob()
196 caused File::Glob to be loaded for the first time, have been fixed.
200 Although "you shouldn't do that", it was possible to write code that
201 depends on Perl's hashed key order (Data::Dumper does this). The new
202 algorithm "One-at-a-Time" produces a different hashed key order.
203 More details are in L</"Performance Enhancements">.
207 lstat(FILEHANDLE) now gives a warning because the operation makes no sense.
208 In future releases this may become a fatal error.
212 The C<package;> syntax (C<package> without an argument) has been
213 deprecated. Its semantics were never that clear and its
214 implementation even less so. If you have used that feature to
215 disallow all but fully qualified variables, C<use strict;> instead.
219 The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still
220 recognised but now cause fatal errors. The previous behaviour of
221 ignoring them by default and warning if requested was unacceptable
222 since it, in a way, falsely promised that the features could be used.
226 The current user-visible implementation of pseudo-hashes (the weird
227 use of the first array element) is deprecated starting from Perl 5.8.0
228 and will be removed in Perl 5.10.0, and the feature will be
229 implemented differently. Not only is the current interface rather
230 ugly, but the current implementation slows down normal array and hash
231 use quite noticeably. The C<fields> pragma interface will remain
232 available. The I<restricted hashes> interface is expected to
233 be the replacement interface (see L<Hash::Util>).
237 The syntaxes C<< @a->[...] >> and C<< %h->{...} >> have now been deprecated.
241 After years of trying, suidperl is considered to be too complex to
242 ever be considered truly secure. The suidperl functionality is likely
243 to be removed in a future release.
247 The 5.005 threads model (module C<Thread>) is deprecated and expected
248 to be removed in Perl 5.10. Multithreaded code should be migrated to
249 the new ithreads model (see L<threads> and L<threads::shared>).
253 The long deprecated uppercase aliases for the string comparison
254 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
258 The tr///C and tr///U features have been removed and will not return;
259 the interface was a mistake. Sorry about that. For similar
260 functionality, see pack('U0', ...) and pack('C0', ...).
264 Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo (@)".
265 The prototypes are now checked better at compile-time for invalid
266 syntax. An optional warning is generated ("Illegal character in
267 prototype...") but this may be upgraded to a fatal error in a future
272 The existing behaviour when localising tied arrays and hashes is wrong,
273 and will be changed in a future release, so do not rely on the existing
274 behaviour. See L<"Localising Tied Arrays and Hashes Is Broken">.
278 =head1 Core Enhancements
280 =head2 PerlIO is Now The Default
286 IO is now by default done via PerlIO rather than system's "stdio".
287 PerlIO allows "layers" to be "pushed" onto a file handle to alter the
288 handle's behaviour. Layers can be specified at open time via 3-arg
291 open($fh,'>:crlf :utf8', $path) || ...
293 or on already opened handles via extended C<binmode>:
295 binmode($fh,':encoding(iso-8859-7)');
297 The built-in layers are: unix (low level read/write), stdio (as in
298 previous Perls), perlio (re-implementation of stdio buffering in a
299 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
300 but available on any platform). A mmap layer may be available if
301 platform supports it (mostly UNIXes).
303 Layers to be applied by default may be specified via the 'open' pragma.
305 See L</"Installation and Configuration Improvements"> for the effects
306 of PerlIO on your architecture name.
310 File handles can be marked as accepting Perl's internal encoding of Unicode
311 (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :
313 open($fh,">:utf8","Uni.txt");
315 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named
316 for you since it's not UTF-8 what you will be getting but instead
317 UTF-EBCDIC. See L<perlunicode>, L<utf8>, and
318 http://www.unicode.org/unicode/reports/tr16/ for more information.
319 In future releases this naming may change.
323 File handles can translate character encodings from/to Perl's internal
324 Unicode form on read/write via the ":encoding()" layer.
328 File handles can be opened to "in memory" files held in Perl scalars via:
330 open($fh,'>', \$variable) || ...
334 Anonymous temporary files are available without need to
335 'use FileHandle' or other module via
337 open($fh,"+>", undef) || ...
339 That is a literal undef, not an undefined value.
343 The list form of C<open> is now implemented for pipes (at least on UNIX):
345 open($fh,"-|", 'cat', '/etc/motd')
347 creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in
352 If your locale environment variables (LANGUAGE, LC_ALL, LC_CTYPE, LANG)
353 contain the strings 'UTF-8' or 'UTF8' (case-insensitive matching),
354 the default encoding of your STDIN, STDOUT, and STDERR, and of
355 B<any subsequent file open>, is UTF-8.
359 =head2 Restricted Hashes
361 A restricted hash is restricted to a certain set of keys, no keys
362 outside the set can be added. Also individual keys can be restricted
363 so that the key cannot be deleted and the value cannot be changed.
364 No new syntax is involved: the Hash::Util module is the interface.
368 Perl used to be fragile in that signals arriving at inopportune moments
369 could corrupt Perl's internal state. Now Perl postpones handling of
370 signals until it's safe (between opcodes).
372 This change may have surprising side effects because signals no longer
373 interrupt Perl instantly. Perl will now first finish whatever it was
374 doing, like finishing an internal operation (like sort()) or an
375 external operation (like an I/O operation), and only then look at any
376 arrived signals (and before starting the next operation). No more corrupt
377 internal state since the current operation is always finished first,
378 but the signal may take more time to get heard. Note that breaking
379 out from potentially blocking operations should still work, though.
381 =head2 Unicode Overhaul
383 Unicode in general should be now much more usable than in Perl 5.6.0
384 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in
385 regular expressions should work now, Unicode in tr/// should work now,
386 Unicode in I/O should work now. See L<perluniintro> for introduction
387 and L<perlunicode> for details.
393 The Unicode Character Database coming with Perl has been upgraded
394 to Unicode 3.2.0. For more information, see http://www.unicode.org/ .
398 For developers interested in enhancing Perl's Unicode capabilities:
399 almost all the UCD files are included with the Perl distribution in
400 the F<lib/unicore> subdirectory. The most notable omission, for space
401 considerations, is the Unihan database.
405 The properties \p{Blank} and \p{SpacePerl} have been added. "Blank" is like
406 C isblank(), that is, it contains only "horizontal whitespace" (the space
407 character is, the newline isn't), and the "SpacePerl" is the Unicode
408 equivalent of C<\s> (\p{Space} isn't, since that includes the vertical
409 tabulator character, whereas C<\s> doesn't.)
411 See "New Unicode Properties" earlier in this document for additional
412 information on changes with Unicode properties.
416 =head2 Understanding of Numbers
418 In general a lot of fixing has happened in the area of Perl's
419 understanding of numbers, both integer and floating point. Since in
420 many systems the standard number parsing functions like C<strtoul()>
421 and C<atof()> seem to have bugs, Perl tries to work around their
422 deficiencies. This results hopefully in more accurate numbers.
424 Perl now tries internally to use integer values in numeric conversions
425 and basic arithmetics (+ - * /) if the arguments are integers, and
426 tries also to keep the results stored internally as integers.
427 This change leads to often slightly faster and always less lossy
428 arithmetics. (Previously Perl always preferred floating point numbers
431 =head2 Miscellaneous Changes
437 AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute
438 to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.
442 The $Config{byteorder} (and corresponding BYTEORDER in config.h) was
443 previously wrong in platforms if sizeof(long) was 4, but sizeof(IV)
444 was 8. The byteorder was only sizeof(long) bytes long (1234 or 4321),
445 but now it is correctly sizeof(IV) bytes long, (12345678 or 87654321).
446 (This problem didn't affect Windows platforms.)
448 Also, $Config{byteorder} is now computed dynamically--this is more
449 robust with "fat binaries" where an executable image contains binaries
450 for more than one binary platform, and when cross-compiling.
454 C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass
455 in multiple arguments.)
459 The builtin dump() now gives an optional warning
460 C<dump() better written as CORE::dump()>,
461 meaning that by default C<dump(...)> is resolved as the builtin
462 dump() which dumps core and aborts, not as (possibly) user-defined
463 C<sub dump>. To call the latter, qualify the call as C<&dump(...)>.
464 (The whole dump() feature is to considered deprecated, and possibly
465 removed/changed in future releases.)
469 chomp() and chop() are now overridable. Note, however, that their
470 prototype (as given by C<prototype("CORE::chomp")> is undefined,
471 because it cannot be expressed and therefore one cannot really write
472 replacements to override these builtins.
476 END blocks are now run even if you exit/die in a BEGIN block.
477 Internally, the execution of END blocks is now controlled by
478 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new
479 behaviour for Perl embedders. This will default in 5.10. See
484 Formats now support zero-padded decimal fields.
488 Lvalue subroutines can now return C<undef> in list context.
489 However, the lvalue subroutine feature still remains experimental.
493 A lost warning "Can't declare ... dereference in my" has been
494 restored (Perl had it earlier but it became lost in later releases.)
498 A new special regular expression variable has been introduced:
499 C<$^N>, which contains the most-recently closed group (submatch).
503 C<no Module;> now works even if there is no "sub unimport" in the Module.
507 The numerical comparison operators return C<undef> if either operand
508 is a NaN. Previously the behaviour was unspecified.
512 The following builtin functions are now overridable: each(), keys(),
513 pop(), push(), shift(), splice(), unshift().
517 C<pack() / unpack()> now can group template letters with C<()> and then
518 apply repetition/count modifiers on the groups.
522 C<pack() / unpack()> can now process the Perl internal numeric types:
523 IVs, UVs, NVs-- and also long doubles, if supported by the platform.
524 The template letters are C<j>, C<J>, C<F>, and C<D>.
528 C<pack('U0a*', ...)> can now be used to force a string to UTF8.
532 my __PACKAGE__ $obj now works.
536 POSIX::sleep() now returns the number of I<unslept> seconds
537 (as the POSIX standard says), as opposed to CORE::sleep() which
538 returns the number of slept seconds.
542 The printf() and sprintf() now support parameter reordering using the
543 C<%\d+\$> and C<*\d+\$> syntaxes. For example
545 print "%2\$s %1\$s\n", "foo", "bar";
547 will print "bar foo\n". This feature helps in writing
548 internationalised software, and in general when the order
549 of the parameters can vary.
553 prototype(\&) is now available.
557 prototype(\[$@%&]) is now available to implicitly create references
558 (useful for example if you want to emulate the tie() interface).
562 A new command-line option, C<-t> is available. It is the
563 little brother of C<-T>: instead of dying on taint violations,
564 lexical warnings are given. B<This is only meant as a temporary
565 debugging aid while securing the code of old legacy applications.
566 This is not a substitute for -T.>
570 In other taint news, the C<exec LIST> and C<system LIST> have now been
571 considered too risky (think C<exec @ARGV>: it can start any program
572 with any arguments), and now the said forms cause a warning.
573 You should carefully launder the arguments to guarantee their
574 validity. In future releases of Perl the forms will become fatal
575 errors so consider starting laundering now.
579 Tied hash interfaces are now required to have the EXISTS and DELETE
580 methods (either own or inherited).
584 If tr/// is just counting characters, it doesn't attempt to
589 untie() will now call an UNTIE() hook if it exists. See L<perltie>
594 L<utime> now supports C<utime undef, undef, @files> to change the
595 file timestamps to the current time.
599 The rules for allowing underscores (underbars) in numeric constants
600 have been relaxed and simplified: now you can have an underscore
601 simply B<between digits>.
605 Rather than relying on C's argv[0] (which may not contain a full pathname)
606 where possible $^X is now set by asking the operating system.
607 (eg by reading F</proc/self/exe> on Linux, F</proc/curproc/file> on FreeBSD)
611 A new variable, C<${^TAINT}>, indicates whether taint mode is enabled.
615 You can now override the readline() builtin, and this overrides also
616 the <FILEHANDLE> angle bracket operator.
620 The command-line options -s and -F are now recognized on the shebang
625 Use of the C</c> match modifier without an accompanying C</g> modifier
626 elicits a new warning: C<Use of /c modifier is meaningless without /g>.
628 Use of C</c> in substitutions, even with C</g>, elicits
629 C<Use of /c modifier is meaningless in s///>.
631 Use of C</g> with C<split> elicits C<Use of /g modifier is meaningless
636 =head1 Modules and Pragmata
638 =head2 New Modules and Pragmata
644 C<Attribute::Handlers> allows a class to define attribute handlers.
647 use Attribute::Handlers;
648 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
650 # later, in some package using or inheriting from MyPack...
652 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
654 Both variables and routines can have attribute handlers. Handlers can
655 be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the
656 exact compilation phase (BEGIN, CHECK, INIT, or END).
657 See L<Attribute::Handlers>.
661 C<B::Concise>, by Stephen McCamant, is a new compiler backend for
662 walking the Perl syntax tree, printing concise info about ops.
663 The output is highly customisable. See L<B::Concise>.
667 The new bignum, bigint, and bigrat pragmas, by Tels, implement
668 transparent bignum support (using the Math::BigInt, Math::BigFloat,
669 and Math::BigRat backends).
673 C<Class::ISA>, by Sean Burke, is a module for reporting the search
674 path for a class's ISA tree. See L<Class::ISA>.
678 C<Cwd> now has a split personality: if possible, an XS extension is
679 used, (this will hopefully be faster, more secure, and more robust)
680 but if not possible, the familiar Perl implementation is used.
684 C<Devel::PPPort>, originally by Kenneth Albanowski and now
685 maintained by Paul Marquess, has been added. It is primarily used
686 by C<h2xs> to enhance portability of XS modules between different
687 versions of Perl. See L<Devel::PPPort>.
691 C<Digest>, frontend module for calculating digests (checksums), from
692 Gisle Aas, has been added. See L<Digest>.
696 C<Digest::MD5> for calculating MD5 digests (checksums) as defined in
697 RFC 1321, from Gisle Aas, has been added. See L<Digest::MD5>.
699 use Digest::MD5 'md5_hex';
701 $digest = md5_hex("Thirsty Camel");
703 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
705 NOTE: the C<MD5> backward compatibility module is deliberately not
706 included since its further use is discouraged.
710 C<Encode>, originally by Nick Ing-Simmons and now maintained by Dan
711 Kogai, provides a mechanism to translate between different character
712 encodings. Support for Unicode, ISO-8859-1, and ASCII are compiled in
713 to the module. Several other encodings (like the rest of the
714 ISO-8859, CP*/Win*, Mac, KOI8-R, three variants EBCDIC, Chinese,
715 Japanese, and Korean encodings) are included and can be loaded at
716 runtime. (For space considerations, the largest Chinese encodings
717 have been separated into their own CPAN module, Encode::HanExtra,
718 which Encode will use if available). See L<Encode>.
720 Any encoding supported by Encode module is also available to the
721 ":encoding()" layer if PerlIO is used.
725 C<Hash::Util> is the interface to the new I<restricted hashes>
726 feature. (Implemented by Jeffrey Friedl, Nick Ing-Simmons, and
727 Michael Schwern.) See L<Hash::Util>.
731 C<I18N::Langinfo> can be used to query locale information.
732 See L<I18N::Langinfo>.
736 C<I18N::LangTags>, by Sean Burke, has functions for dealing with
737 RFC3066-style language tags. See L<I18N::LangTags>.
741 C<ExtUtils::Constant>, by Nicholas Clark, is a new tool for extension
742 writers for generating XS code to import C header constants.
743 See L<ExtUtils::Constant>.
747 C<Filter::Simple>, by Damian Conway, is an easy-to-use frontend to
748 Filter::Util::Call. See L<Filter::Simple>.
754 use Filter::Simple sub {
755 while (my ($from, $to) = splice @_, 0, 2) {
764 use MyFilter qr/red/ => 'green';
766 print "red\n"; # this code is filtered, will print "green\n"
767 print "bored\n"; # this code is filtered, will print "bogreen\n"
771 print "red\n"; # this code is not filtered, will print "red\n"
775 C<File::Temp>, by Tim Jenness, allows one to create temporary files and
776 directories in an easy, portable, and secure way. See L<File::Temp>.
780 C<Filter::Util::Call>, by Paul Marquess, provides you with the
781 framework to write I<source filters> in Perl. For most uses, the
782 frontend Filter::Simple is to be preferred. See L<Filter::Util::Call>.
786 C<if>, by Ilya Zakharevich, is a new pragma for conditional inclusion
791 L<libnet>, by Graham Barr, is a collection of perl5 modules related
792 to network programming. See L<Net::FTP>, L<Net::NNTP>, L<Net::Ping>
793 (not part of libnet, but related), L<Net::POP3>, L<Net::SMTP>,
796 Perl installation leaves libnet unconfigured; use F<libnetcfg>
801 C<List::Util>, by Graham Barr, is a selection of general-utility
802 list subroutines, such as sum(), min(), first(), and shuffle().
807 C<Locale::Constants>, C<Locale::Country>, C<Locale::Currency>
808 C<Locale::Language>, and L<Locale::Script>, by Neil Bowers, have
809 been added. They provide the codes for various locale standards, such
810 as "fr" for France, "usd" for US Dollar, and "ja" for Japanese.
814 $country = code2country('jp'); # $country gets 'Japan'
815 $code = country2code('Norway'); # $code gets 'no'
817 See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>,
818 and L<Locale::Language>.
822 C<Locale::Maketext>, by Sean Burke, is a localization framework. See
823 L<Locale::Maketext>, and L<Locale::Maketext::TPJ13>. The latter is an
824 article about software localization, originally published in The Perl
825 Journal #13, and republished here with kind permission.
829 C<Math::BigRat> for big rational numbers, to accompany Math::BigInt and
830 Math::BigFloat, from Tels. See L<Math::BigRat>.
834 C<Memoize> can make your functions faster by trading space for time,
835 from Mark-Jason Dominus. See L<Memoize>.
839 C<MIME::Base64>, by Gisle Aas, allows you to encode data in base64,
840 as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
845 $encoded = encode_base64('Aladdin:open sesame');
846 $decoded = decode_base64($encoded);
848 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
854 C<MIME::QuotedPrint>, by Gisle Aas, allows you to encode data
855 in quoted-printable encoding, as defined in RFC 2045 - I<MIME
856 (Multipurpose Internet Mail Extensions)>.
858 use MIME::QuotedPrint;
860 $encoded = encode_qp("Smiley in Unicode: \x{263a}");
861 $decoded = decode_qp($encoded);
863 print $encoded, "\n"; # "Smiley in Unicode: =263A"
865 MIME::QuotedPrint has been enhanced to provide the basic methods
866 necessary to use it with PerlIO::Via as in :
868 use MIME::QuotedPrint;
869 open($fh,">Via(MIME::QuotedPrint)",$path);
871 See L<MIME::QuotedPrint>.
875 C<NEXT>, by Damian Conway, is a pseudo-class for method redispatch.
880 C<open> is a new pragma for setting the default I/O disciplines
885 C<PerlIO::Scalar>, by Nick Ing-Simmons, provides the implementation
886 of IO to "in memory" Perl scalars as discussed above. It also serves
887 as an example of a loadable PerlIO layer. Other future possibilities
888 include PerlIO::Array and PerlIO::Code. See L<PerlIO::Scalar>.
892 C<PerlIO::Via>, by Nick Ing-Simmons, acts as a PerlIO layer and wraps
893 PerlIO layer functionality provided by a class (typically implemented
896 use MIME::QuotedPrint;
897 open($fh,">Via(MIME::QuotedPrint)",$path);
899 This will automatically convert everything output to C<$fh>
900 to Quoted-Printable. See L<PerlIO::Via>.
904 C<Pod::ParseLink>, by Russ Allbery, has been added,
905 to parse LZ<><> links in pods as described in the new
910 C<Pod::Text::Overstrike>, by Joe Smith, has been added.
911 It converts POD data to formatted overstrike text.
912 See L<Pod::Text::Overstrike>.
916 C<Scalar::Util> is a selection of general-utility scalar subroutines,
917 such as blessed(), reftype(), and tainted(). See L<Scalar::Util>.
921 C<sort> is a new pragma for controlling the behaviour of sort().
925 C<Storable> gives persistence to Perl data structures by allowing the
926 storage and retrieval of Perl data to and from files in a fast and
927 compact binary format. Because in effect Storable does serialisation
928 of Perl data structues, with it you can also clone deep, hierarchical
929 datastructures. Storable was originally created by Raphael Manfredi,
930 but it is now maintained by Abhijit Menon-Sen. Storable has been
931 enhanced to understand the two new hash features, Unicode keys and
932 restricted hashes. See L<Storable>.
936 C<Switch>, by Damian Conway, has been added. Just by saying
940 you have C<switch> and C<case> available in Perl.
946 case 1 { print "number 1" }
947 case "a" { print "string a" }
948 case [1..10,42] { print "number in list" }
949 case (@array) { print "number in list" }
950 case /\w+/ { print "pattern" }
951 case qr/\w+/ { print "pattern" }
952 case (%hash) { print "entry in hash" }
953 case (\%hash) { print "entry in hash" }
954 case (\&sub) { print "arg to subroutine" }
955 else { print "previous case not true" }
962 C<Test::More>, by Michael Schwern, is yet another framework for writing
963 test scripts, more extensive than Test::Simple. See L<Test::More>.
967 C<Test::Simple>, by Michael Schwern, has basic utilities for writing
968 tests. See L<Test::Simple>.
972 C<Text::Balanced>, by Damian Conway, has been added, for extracting
973 delimited text sequences from strings.
975 use Text::Balanced 'extract_delimited';
977 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
979 $a will be "'never say never'", $b will be ', he never said'.
981 In addition to extract_delimited(), there are also extract_bracketed(),
982 extract_quotelike(), extract_codeblock(), extract_variable(),
983 extract_tagged(), extract_multiple(), gen_delimited_pat(), and
984 gen_extract_tagged(). With these, you can implement rather advanced
985 parsing algorithms. See L<Text::Balanced>.
989 C<threads>, by Arthur Bergman, is an interface to interpreter threads.
990 Interpreter threads (ithreads) is the new thread model introduced in
991 Perl 5.6 but only available as an internal interface for extension
992 writers (and for Win32 Perl for C<fork()> emulation). See L<threads>.
996 C<threads::shared>, by Arthur Bergman, allows data sharing for
997 interpreter threads. In the ithreads model any data sharing between
998 threads must be explicit, as opposed to the old 5.005 thread model
999 where data sharing was implicit. See L<threads::shared>.
1003 C<Tie::File>, by Mark-Jason Dominus, associates a Perl array with the
1004 lines of a file. See L<Tie::File>.
1008 C<Tie::Memoize>, by Ilya Zakharevich, provides on-demand loaded hashes.
1009 See L<Tie::Memoize>.
1013 C<Tie::RefHash::Nestable>, by Edward Avis, allows storing hash
1014 references (unlike the standard Tie::RefHash) The module is contained
1015 within Tie::RefHash. See L<Tie::RefHash>.
1019 C<Time::HiRes>, by Douglas E. Wegscheid, provides high resolution
1020 timing (ualarm, usleep, and gettimeofday). See L<Time::HiRes>.
1024 C<Unicode::UCD> offers a querying interface to the Unicode Character
1025 Database. See L<Unicode::UCD>.
1029 C<Unicode::Collate>, by SADAHIRO Tomoyuki, implements the UCA
1030 (Unicode Collation Algorithm) for sorting Unicode strings.
1031 See L<Unicode::Collate>.
1035 C<Unicode::Normalize>, by SADAHIRO Tomoyuki, implements the various
1036 Unicode normalization forms. See L<Unicode::Normalize>.
1040 C<XS::Typemap>, by Tim Jenness, is a test extension that exercises
1041 XS typemaps. Nothing gets installed, but the code is worth studying
1042 for extension writers.
1046 =head2 Updated And Improved Modules and Pragmata
1052 The following independently supported modules have been updated to the
1053 newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp,
1054 Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle
1055 (Pod::Man, Pod::Text), Pod::LaTeX, Pod::Parser, Storable,
1056 Term::ANSIColor, Test, Text-Tabs+Wrap.
1060 attributes::reftype() now works on tied arguments.
1064 AutoLoader can now be disabled with C<no AutoLoader;>.
1068 B::Deparse has been significantly enhanced. It now can deparse almost
1069 all of the standard test suite (so that the tests still succeed).
1070 There is a make target "test.deparse" for trying this out.
1074 Class::Struct can now define the classes in compile time.
1078 Class::Struct now assigns the array/hash element if the accessor
1079 is called with an array/hash element as the B<sole> argument.
1083 The return value of Cwd::fastcwd() is now tainted.
1087 Data::Dumper now has an option to sort hashes.
1091 Data::Dumper now has an option to dump code references
1096 DB_File now supports newer Berkeley DB versions, among
1101 Devel::Peek now has an interface for the Perl memory statistics
1102 (this works only if you are using perl's malloc, and if you have
1103 compiled with debugging).
1107 The English module can now be used without the infamous performance
1110 use English '-no_match_vars';
1112 (Assuming, of course, that you don't need the troublesome variables
1113 C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and
1114 C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>.
1118 ExtUtils::MakeMaker now uses File::Spec internally, which hopefully
1119 leads to better portability.
1123 Fcntl, Socket, and Sys::Syslog have been rewritten to use the
1124 new-style constant dispatch section (see L<ExtUtils::Constant>).
1125 This means that they will be more robust and hopefully faster.
1129 File::Find now chdir()s correctly when chasing symbolic links.
1133 File::Find now has pre- and post-processing callbacks. It also
1134 correctly changes directories when chasing symbolic links. Callbacks
1135 (naughtily) exiting with "next;" instead of "return;" now work.
1139 File::Find is now (again) reentrant. It also has been made
1144 The warnings issued by File::Find now belong to their own category.
1145 You can enable/disable them with C<use/no warnings 'File::Find';>.
1149 File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid
1150 prototype mismatch with CORE::glob().
1154 File::Glob now supports C<GLOB_LIMIT> constant to limit the size of
1155 the returned list of filenames.
1159 IPC::Open3 now allows the use of numeric file descriptors.
1163 IO::Socket now has an atmark() method, which returns true if the socket
1164 is positioned at the out-of-band mark. The method is also exportable
1165 as a sockatmark() function.
1169 IO::Socket::INET has support for the ReusePort option (if your
1170 platform supports it). The Reuse option now has an alias, ReuseAddr.
1171 For clarity, you may want to prefer ReuseAddr.
1175 IO::Socket::INET now supports a value of zero for C<LocalPort>
1176 (usually meaning that the operating system will make one up.)
1180 'use lib' now works identically to @INC. Removing directories
1181 with 'no lib' now works.
1185 Math::BigFloat and Math::BigInt have undergone a full rewrite.
1186 They are now magnitudes faster, and they support various
1187 bignum libraries such as GMP and PARI as their backends.
1191 Math::Complex handles inf, NaN etc., better.
1195 Net::Ping has been considerably enhanced: multihoming is now supported,
1196 Win32 functionality is better, there is now time measuring
1197 functionality (optionally high-resolution using Time::HiRes),
1198 and there is now "external" protocol which uses Net::Ping::External
1199 module which runs your external ping utility and parses the output.
1200 A version of Net::Ping::External is available in CPAN.
1202 Note that some of the Net::Ping tests are disabled when running
1203 under the Perl distribution since one cannot assume one or more
1204 of the following: enabled echo port at localhost, full Internet
1205 connectivity, or sympathetic firewalls. You can set the environment
1206 variable PERL_TEST_Net_Ping to "1" (one) before running the Perl test
1207 suite to enable all the Net::Ping tests.
1211 POSIX::sigaction() is now much more flexible and robust.
1212 You can now install coderef handlers, 'DEFAULT', and 'IGNORE'
1213 handlers, installing new handlers was not atomic.
1217 In Safe, C<%INC> is now localised in a Safe compartment so that
1222 In SDBM_File on dosish platforms, some keys went missing because of
1223 lack of support for files with "holes". A workaround for the problem
1228 In Search::Dict one can now have a pre-processing hook for the
1229 lines being searched.
1233 The Shell module now has an OO interface.
1237 In Sys::Syslog there is now a failover mechanism that will go
1238 through alternative connection mechanisms until the message
1239 is successfully logged.
1243 The Test module has been significantly enhanced.
1247 Time::Local::timelocal() does not handle fractional seconds anymore.
1248 The rationale is that neither does localtime(), and timelocal() and
1249 localtime() are supposed to be inverses of each other.
1253 The vars pragma now supports declaring fully qualified variables.
1254 (Something that C<our()> does not and will not support.)
1258 The C<utf8::> name space (as in the pragma) provides various
1259 Perl-callable functions to provide low level access to Perl's
1260 internal Unicode representation. At the moment only length()
1261 has been implemented.
1265 =head1 Utility Changes
1271 Emacs perl mode (emacs/cperl-mode.el) has been updated to version
1276 F<emacs/e2ctags.pl> is now much faster.
1280 C<enc2xs> is a tool for people adding their own encodings to the
1285 C<h2ph> now supports C trigraphs.
1289 C<h2xs> now produces a template README.
1293 C<h2xs> now uses C<Devel::PPPort> for better portability between
1294 different versions of Perl.
1298 C<h2xs> uses the new L<ExtUtils::Constant|ExtUtils::Constant> module
1299 which will affect newly created extensions that define constants.
1300 Since the new code is more correct (if you have two constants where the
1301 first one is a prefix of the second one, the first constant B<never>
1302 got defined), less lossy (it uses integers for integer constant,
1303 as opposed to the old code that used floating point numbers even for
1304 integer constants), and slightly faster, you might want to consider
1305 regenerating your extension code (the new scheme makes regenerating
1306 easy). L<h2xs> now also supports C trigraphs.
1310 C<libnetcfg> has been added to configure libnet.
1314 C<perlbug> is now much more robust. It also sends the bug report to
1315 perl.org, not perl.com.
1319 C<perlcc> has been rewritten and its user interface (that is,
1320 command line) is much more like that of the UNIX C compiler, cc.
1321 (The perlbc tools has been removed. Use C<perlcc -B> instead.)
1322 B<Note that perlcc is still considered very experimental and
1327 C<perlivp> is a new Installation Verification Procedure utility
1328 for running any time after installing Perl.
1332 C<piconv> is an implementation of the character conversion utility
1333 C<iconv>, demonstrating the new Encode module.
1337 C<pod2html> now allows specifying a cache directory.
1341 C<pod2html> now produces XHTML 1.0.
1345 C<pod2html> now understands POD written using different line endings
1346 (PC-like CRLF versus UNIX-like LF versus MacClassic-like CR).
1350 C<s2p> has been completely rewritten in Perl. (It is in fact a full
1351 implementation of sed in Perl: you can use the sed functionality by
1352 using the C<psed> utility.)
1356 C<xsubpp> now understands POD documentation embedded in the *.xs files.
1360 C<xsubpp> now supports the OUT keyword.
1364 =head1 New Documentation
1370 perl56delta details the changes between the 5.005 release and the
1375 perlclib documents the internal replacements for standard C library
1376 functions. (Interesting only for extension writers and Perl core
1381 perldebtut is a Perl debugging tutorial.
1385 perlebcdic contains considerations for running Perl on EBCDIC platforms.
1389 perlintro is a gentle introduction to Perl.
1393 perliol documents the internals of PerlIO with layers.
1397 perlmodstyle is a style guide for writing modules.
1401 perlnewmod tells about writing and submitting a new module.
1405 perlpacktut is a pack() tutorial.
1409 perlpod has been rewritten to be clearer and to record the best
1410 practices gathered over the years.
1414 perlpodspec is a more formal specification of the pod format,
1415 mainly of interest for writers of pod applications, not to
1416 people writing in pod.
1420 perlretut is a regular expression tutorial.
1424 perlrequick is a regular expressions quick-start guide.
1425 Yes, much quicker than perlretut.
1429 perltodo has been updated.
1433 perltootc has been renamed as perltooc (to not to conflict
1434 with perltoot in filesystems restricted to "8.3" names).
1438 perluniintro is an introduction to using Unicode in Perl.
1439 (perlunicode is more of a detailed reference and background
1444 perlutil explains the command line utilities packaged with the Perl
1449 The following platform-specific documents are available before
1450 the installation as README.I<platform>, and after the installation
1453 perlaix perlamiga perlapollo perlbeos perlbs2000
1454 perlce perlcygwin perldgux perldos perlepoc perlhpux
1455 perlhurd perlmachten perlmacos perlmint perlmpeix
1456 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1457 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1459 Eastern Asian Perl users are now welcomed in their own languages:
1460 README.jp (Japanese), README.ko (Korean), README.cn (simplified
1461 Chinese) and README.tw (traditional Chinese), which are written in
1462 normal pod but encoded in EUC-JP, EUC-KR, EUC-CN and Big5. These
1463 will get installed as
1465 perljp perlko perlcn perltw
1471 The documentation for the POSIX-BC platform is called "BS2000", to avoid
1472 confusion with the Perl POSIX module.
1476 The documentation for the WinCE platform is called perlce (README.ce
1477 in the source code kit), to avoid confusion with the perlwin32
1478 documentation on 8.3-restricted filesystems.
1482 =head1 Performance Enhancements
1488 map() could get pathologically slow when the result list it generates
1489 is larger than the source list. The performance has been improved for
1494 sort() has been changed to use primarily mergesort internally as
1495 opposed to the earlier quicksort. For very small lists this may
1496 result in slightly slower sorting times, but in general the speedup
1497 should be at least 20%. Additional bonuses are that the worst case
1498 behaviour of sort() is now better (in computer science terms it now
1499 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1500 worst-case run time behaviour), and that sort() is now stable
1501 (meaning that elements with identical keys will stay ordered as they
1502 were before the sort). See the C<sort> pragma for information.
1504 The story in more detail: suppose you want to serve yourself a little
1507 @digits = ( 3,1,4,1,5,9 );
1509 A numerical sort of the digits will yield (1,1,3,4,5,9), as expected.
1510 Which C<1> comes first is hard to know, since one C<1> looks pretty
1511 much like any other. You can regard this as totally trivial,
1512 or somewhat profound. However, if you just want to sort the even
1513 digits ahead of the odd ones, then what will
1515 sort { ($a % 2) <=> ($b % 2) } @digits;
1517 yield? The only even digit, C<4>, will come first. But how about
1518 the odd numbers, which all compare equal? With the quicksort algorithm
1519 used to implement Perl 5.6 and earlier, the order of ties is left up
1520 to the sort. So, as you add more and more digits of Pi, the order
1521 in which the sorted even and odd digits appear will change.
1522 and, for sufficiently large slices of Pi, the quicksort algorithm
1523 in Perl 5.8 won't return the same results even if reinvoked with the
1524 same input. The justification for this rests with quicksort's
1525 worst case behavior. If you run
1527 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1529 (something you might approximate if you wanted to merge two sorted
1530 arrays using sort), doubling $N doesn't just double the quicksort time,
1531 it I<quadruples> it. Quicksort has a worst case run time that can
1532 grow like N**2, so-called I<quadratic> behaviour, and it can happen
1533 on patterns that may well arise in normal use. You won't notice this
1534 for small arrays, but you I<will> notice it with larger arrays,
1535 and you may not live long enough for the sort to complete on arrays
1536 of a million elements. So the 5.8 quicksort scrambles large arrays
1537 before sorting them, as a statistical defence against quadratic behaviour.
1538 But that means if you sort the same large array twice, ties may be
1539 broken in different ways.
1541 Because of the unpredictability of tie-breaking order, and the quadratic
1542 worst-case behaviour, quicksort was I<almost> replaced completely with
1543 a stable mergesort. I<Stable> means that ties are broken to preserve
1544 the original order of appearance in the input array. So
1546 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1548 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1549 appear in the output in the same order they appeared in the input.
1550 Mergesort has worst case O(N log N) behaviour, the best value
1551 attainable. And, ironically, this mergesort does particularly
1552 well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N)
1553 in O(N) time. But quicksort was rescued at the last moment because
1554 it is faster than mergesort on certain inputs and platforms.
1555 For example, if you really I<don't> care about the order of even
1556 and odd digits, quicksort will run in O(N) time; it's very good
1557 at sorting many repetitions of a small number of distinct elements.
1558 The quicksort divide and conquer strategy works well on platforms
1559 with relatively small, very fast, caches. Eventually, the problem gets
1560 whittled down to one that fits in the cache, from which point it
1561 benefits from the increased memory speed.
1563 Quicksort was rescued by implementing a sort pragma to control aspects
1564 of the sort. The B<stable> subpragma forces stable behaviour,
1565 regardless of algorithm. The B<_quicksort> and B<_mergesort>
1566 subpragmas are heavy-handed ways to select the underlying implementation.
1567 The leading C<_> is a reminder that these subpragmas may not survive
1568 beyond 5.8. More appropriate mechanisms for selecting the implementation
1569 exist, but they wouldn't have arrived in time to save quicksort.
1573 Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm
1574 ( http://burtleburtle.net/bob/hash/doobs.html ). This algorithm is
1575 reasonably fast while producing a much better spread of values than
1576 the old hashing algorithm (originally by Chris Torek, later tweaked by
1577 Ilya Zakharevich). Hash values output from the algorithm on a hash of
1578 all 3-char printable ASCII keys comes much closer to passing the
1579 DIEHARD random number generation tests. According to perlbench, this
1580 change has not affected the overall speed of Perl.
1584 unshift() should now be noticeably faster.
1588 =head1 Installation and Configuration Improvements
1590 =head2 Generic Improvements
1596 INSTALL now explains how you can configure Perl to use 64-bit
1597 integers even on non-64-bit platforms.
1601 Policy.sh policy change: if you are reusing a Policy.sh file
1602 (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1603 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1604 them will now be changed to the new prefix, /foo/bar. (Previously
1605 only $prefix changed.) If you do not like this new behaviour,
1606 specify prefix, siteprefix, and vendorprefix explicitly.
1610 A new optional location for Perl libraries, otherlibdirs, is available.
1611 It can be used for example for vendor add-ons without disturbing Perl's
1612 own library directories.
1616 In many platforms, the vendor-supplied 'cc' is too stripped-down to
1617 build Perl (basically, 'cc' doesn't do ANSI C). If this seems
1618 to be the case and 'cc' does not seem to be the GNU C compiler
1619 'gcc', an automatic attempt is made to find and use 'gcc' instead.
1623 gcc needs to closely track the operating system release to avoid
1624 build problems. If Configure finds that gcc was built for a different
1625 operating system release than is running, it now gives a clearly visible
1626 warning that there may be trouble ahead.
1630 Since Perl 5.8 is not binary-compatible with previous releases
1631 of Perl, Configure no longer suggests including the 5.005
1636 Configure C<-S> can now run non-interactively.
1640 Configure support for pdp11-style memory models has been removed due
1645 configure.gnu now works with options with whitespace in them.
1649 installperl now outputs everything to STDERR.
1653 Because PerlIO is now the default on most platforms, "-perlio" doesn't
1654 get appended to the $Config{archname} (also known as $^O) anymore.
1655 Instead, if you explicitly choose not to use perlio (Configure command
1656 line option -Uuseperlio), you will get "-stdio" appended.
1660 Another change related to the architecture name is that "-64all"
1661 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1662 pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)
1666 In AFS installations, one can configure the root of the AFS to be
1667 somewhere else than the default F</afs> by using the Configure
1668 parameter C<-Dafsroot=/some/where/else>.
1672 APPLLIB_EXP, a lesser-known configuration-time definition, has been
1673 documented. It can be used to prepend site-specific directories
1674 to Perl's default search path (@INC); see INSTALL for information.
1678 The version of Berkeley DB used when the Perl (and, presumably, the
1679 DB_File extension) was built is now available as
1680 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
1681 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
1682 DB_VERSION_PATCH_CFG> from C.
1686 Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM
1687 has been documented in INSTALL.
1691 If you have CPAN access (either network or a local copy such as a
1692 CD-ROM) you can during specify extra modules to Configure to build and
1693 install with Perl using the -Dextras=... option. See INSTALL for
1698 In addition to config.over, a new override file, config.arch, is
1699 available. This file is supposed to be used by hints file writers
1700 for architecture-wide changes (as opposed to config.over which is
1701 for site-wide changes).
1705 If your file system supports symbolic links, you can build Perl outside
1706 of the source directory by
1708 mkdir /tmp/perl/build/directory
1709 cd /tmp/perl/build/directory
1710 sh /path/to/perl/source/Configure -Dmksymlinks ...
1712 This will create in /tmp/perl/build/directory a tree of symbolic links
1713 pointing to files in /path/to/perl/source. The original files are left
1714 unaffected. After Configure has finished, you can just say
1718 and Perl will be built and tested, all in /tmp/perl/build/directory.
1722 For Perl developers, several new make targets for profiling
1723 and debugging have been added; see L<perlhack>.
1729 Use of the F<gprof> tool to profile Perl has been documented in
1730 L<perlhack>. There is a make target called "perl.gprof" for
1731 generating a gprofiled Perl executable.
1735 If you have GCC 3, there is a make target called "perl.gcov" for
1736 creating a gcoved Perl executable for coverage analysis. See
1741 If you are on IRIX or Tru64 platforms, new profiling/debugging options
1742 have been added; see L<perlhack> for more information about pixie and
1749 Guidelines of how to construct minimal Perl installations have
1750 been added to INSTALL.
1754 The Thread extension is now not built at all under ithreads
1755 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
1756 Thread extension requires being Configured with C<-Duse5005threads>).
1758 But note that the Thread.pm interface is now shared by both
1763 The Gconvert macro ($Config{d_Gconvert}) used by perl for stringifying
1764 floating-point numbers is now more picky about using sprintf %.*g
1765 rules for the conversion. Some platforms that used to use gcvt may
1766 now resort to the slower sprintf.
1770 The obsolete method of making a special (e.g., debugging) flavor
1773 make LIBPERL=libperld.a
1775 has been removed. Use -DDEBUGGING instead.
1779 =head2 New Or Improved Platforms
1781 For the list of platforms known to support Perl,
1782 see L<perlport/"Supported Platforms">.
1788 AIX dynamic loading should be now better supported.
1792 AIX should now work better with gcc, threads, and 64-bitness. Also the
1793 long doubles support in AIX should be better now. See L<perlaix>.
1797 AtheOS ( http://www.atheos.cx/ ) is a new platform.
1801 BeOS has been reclaimed.
1805 The DG/UX platform now supports 5.005-style threads.
1810 The DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near
1815 EBCDIC platforms (z/OS (also known as OS/390), POSIX-BC, and VM/ESA)
1816 have been regained. Many test suite tests still fail and the
1817 co-existence of Unicode and EBCDIC isn't quite settled, but the
1818 situation is much better than with Perl 5.6. See L<perlos390>,
1819 L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information.
1823 Building perl with -Duseithreads or -Duse5005threads now works under
1824 HP-UX 10.20 (previously it only worked under 10.30 or later). You will
1825 need a thread library package installed. See README.hpux.
1829 MacOS Classic (MacPerl has of course been available since
1830 perl 5.004 but now the source code bases of standard Perl
1831 and MacPerl have been synchronised)
1835 MacOS X (or Darwin) should now be able to build Perl even on HFS+
1836 filesystems. (The case-insensitivity used to confuse the Perl build
1841 NCR MP-RAS is now supported.
1845 All the NetBSD specific patches (except for the installation
1846 specific ones) have been merged back to the main distribution.
1850 NetWare from Novell is now supported. See L<perlnetware>.
1854 NonStop-UX is now supported.
1858 NEC SUPER-UX is now supported.
1862 All the OpenBSD specific patches (except for the installation
1863 specific ones) have been merged back to the main distribution.
1867 Perl has been tested with the GNU pth userlevel thread package
1868 ( http://www.gnu.org/software/pth/pth.html ) . All but one thread
1869 test worked, and that one failure was because of test results arriving
1870 in unexpected order.
1874 Stratus VOS is now supported using Perl's native build method
1875 (Configure). This is the recommended method to build Perl on
1876 VOS. The older methods, which build miniperl, are still
1877 available. See L<perlvos>.
1881 The Amdahl UTS UNIX mainframe platform is now supported.
1885 WinCE is now supported. See L<perlce>.
1889 z/OS (formerly known as OS/390, formerly known as MVS OE) now has
1890 support for dynamic loading. This is not selected by default,
1891 however, you must specify -Dusedl in the arguments of Configure.
1895 =head1 Selected Bug Fixes
1897 Numerous memory leaks and uninitialized memory accesses have been
1898 hunted down. Most importantly, anonymous subs used to leak quite
1905 The autouse pragma didn't work for Multi::Part::Function::Names.
1909 caller() could cause core dumps in certain situations. Carp was sometimes
1910 affected by this problem. In particular, caller() now returns a
1911 subroutine name of C<(unknown)> for subroutines that have been removed
1912 from the symbol table.
1916 chop(@list) in list context returned the characters chopped in
1917 reverse order. This has been reversed to be in the right order.
1921 Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm)
1922 when building the Perl binary. The only exception to this is SunOS 4.x,
1927 The behaviour of non-decimal but numeric string constants such as
1928 "0x23" was platform-dependent: in some platforms that was seen as 35,
1929 in some as 0, in some as a floating point number (don't ask). This
1930 was caused by Perl's using the operating system libraries in a situation
1931 where the result of the string to number conversion is undefined: now
1932 Perl consistently handles such strings as zero in numeric contexts.
1936 The order of DESTROYs has been made more predictable.
1940 Several debugger fixes: exit code now reflects the script exit code,
1941 condition C<"0"> now treated correctly, the C<d> command now checks
1942 line number, C<$.> no longer gets corrupted, and all debugger output
1943 now goes correctly to the socket if RemotePort is set.
1947 Perl 5.6.0 could emit spurious warnings about redefinition of dl_error()
1948 when statically building extensions into perl. This has been corrected.
1952 L<dprofpp> -R didn't work.
1956 C<*foo{FORMAT}> now works.
1960 Infinity is now recognized as a number.
1964 UNIVERSAL::isa no longer caches methods incorrectly. (This broke
1965 the Tk extension with 5.6.0.)
1969 Lexicals I: lexicals outside an eval "" weren't resolved
1970 correctly inside a subroutine definition inside the eval "" if they
1971 were not already referenced in the top level of the eval""ed code.
1975 Lexicals II: lexicals leaked at file scope into subroutines that
1976 were declared before the lexicals.
1980 Lexical warnings now propagating correctly between scopes
1981 and into C<eval "...">.
1985 C<use warnings qw(FATAL all)> did not work as intended. This has been
1990 warnings::enabled() now reports the state of $^W correctly if the caller
1991 isn't using lexical warnings.
1995 Line renumbering with eval and C<#line> now works.
1999 Fixed numerous memory leaks, especially in eval "".
2003 Localised tied variables no longer leak memory
2006 tie my %tied_hash => 'Tie::StdHash';
2010 # Used to leak memory every time local() was called;
2011 # in a loop, this added up.
2012 local($tied_hash{Foo}) = 1;
2016 Localised hash elements (and %ENV) are correctly unlocalised to not
2017 exist, if they didn't before they were localised.
2021 tie my %tied_hash => 'Tie::StdHash';
2025 # Nothing has set the FOO element so far
2027 { local $tied_hash{FOO} = 'Bar' }
2029 # This used to print, but not now.
2030 print "exists!\n" if exists $tied_hash{FOO};
2032 As a side effect of this fix, tied hash interfaces B<must> define
2033 the EXISTS and DELETE methods.
2037 mkdir() now ignores trailing slashes in the directory name,
2038 as mandated by POSIX.
2042 Some versions of glibc have a broken modfl(). This affects builds
2043 with C<-Duselongdouble>. This version of Perl detects this brokenness
2044 and has a workaround for it. The glibc release 2.2.2 is known to have
2045 fixed the modfl() bug.
2049 Modulus of unsigned numbers now works (4063328477 % 65535 used to
2050 return 27406, instead of 27047).
2054 Some "not a number" warnings introduced in 5.6.0 eliminated to be
2055 more compatible with 5.005. Infinity is now recognised as a number.
2059 Numeric conversions did not recognize changes in the string value
2060 properly in certain circumstances.
2064 Attributes (such as :shared) didn't work with our().
2068 our() variables will not cause "will not stay shared" warnings.
2072 "our" variables of the same name declared in two sibling blocks
2073 resulted in bogus warnings about "redeclaration" of the variables.
2074 The problem has been corrected.
2078 pack "Z" now correctly terminates the string with "\0".
2082 Fix password routines which in some shadow password platforms
2083 (e.g. HP-UX) caused getpwent() to return every other entry.
2087 The PERL5OPT environment variable (for passing command line arguments
2088 to Perl) didn't work for more than a single group of options.
2092 PERL5OPT with embedded spaces didn't work.
2096 printf() no longer resets the numeric locale to "C".
2100 C<qw(a\\b)> now parses correctly as C<'a\\b'>: that is, as three
2101 characters, not four.
2105 pos() did not return the correct value within s///ge in earlier
2106 versions. This is now handled correctly.
2110 Printing quads (64-bit integers) with printf/sprintf now works
2111 without the q L ll prefixes (assuming you are on a quad-capable platform).
2115 Regular expressions on references and overloaded scalars now work.
2119 Right-hand side magic (GMAGIC) could in many cases such as string
2120 concatenation be invoked too many times.
2124 scalar() now forces scalar context even when used in void context.
2128 SOCKS support is now much more robust.
2132 sort() arguments are now compiled in the right wantarray context
2133 (they were accidentally using the context of the sort() itself).
2134 The comparison block is now run in scalar context, and the arguments
2135 to be sorted are always provided list context.
2139 Changed the POSIX character class C<[[:space:]]> to include the (very
2140 rarely used) vertical tab character. Added a new POSIX-ish character
2141 class C<[[:blank:]]> which stands for horizontal whitespace
2142 (currently, the space and the tab).
2146 The tainting behaviour of sprintf() has been rationalized. It does
2147 not taint the result of floating point formats anymore, making the
2148 behaviour consistent with that of string interpolation.
2152 Some cases of inconsistent taint propagation (such as within hash
2153 values) have been fixed.
2157 The RE engine found in Perl 5.6.0 accidentally pessimised certain kinds
2158 of simple pattern matches. These are now handled better.
2162 Regular expression debug output (whether through C<use re 'debug'>
2163 or via C<-Dr>) now looks better.
2167 Multi-line matches like C<"a\nxb\n" =~ /(?!\A)x/m> were flawed. The
2172 Use of $& could trigger a core dump under some situations. This
2177 The regular expression captured submatches ($1, $2, ...) are now
2178 more consistently unset if the match fails, instead of leaving false
2179 data lying around in them.
2183 readline() on files opened in "slurp" mode could return an extra
2184 "" (blank line) at the end in certain situations. This has been
2189 Autovivification of symbolic references of special variables described
2190 in L<perlvar> (as in C<${$num}>) was accidentally disabled. This works
2195 Sys::Syslog ignored the C<LOG_AUTH> constant.
2199 All but the first argument of the IO syswrite() method are now optional.
2203 $AUTOLOAD, sort(), lock(), and spawning subprocesses
2204 in multiple threads simultaneously are now thread-safe.
2208 Tie::Array's SPLICE method was broken.
2212 Allow a read-only string on the left-hand side of a non-modifying tr///.
2216 If C<STDERR> is tied, warnings caused by C<warn> and C<die> now
2217 correctly pass to it.
2221 Several Unicode fixes.
2227 BOMs (byte order marks) at the beginning of Perl files
2228 (scripts, modules) should now be transparently skipped.
2229 UTF-16 and UCS-2 encoded Perl files should now be read correctly.
2233 The character tables have been updated to Unicode 3.2.0.
2237 Comparing with utf8 data does not magically upgrade non-utf8 data
2238 into utf8. (This was a problem for example if you were mixing data
2239 from I/O and Unicode data: your output might have got magically encoded
2244 Generating illegal Unicode code points such as U+FFFE, or the UTF-16
2245 surrogates, now also generates an optional warning.
2249 C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase.
2253 Concatenation with the C<.> operator or via variable interpolation,
2254 C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator,
2255 substitution with C<s///>, single-quoted UTF8, should now work.
2259 The C<tr///> operator now works. Note that the C<tr///CU>
2260 functionality has been removed (but see pack('U0', ...)).
2264 C<eval "v200"> now works.
2268 Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spurious warnings.
2269 This has been corrected.
2273 Zero entries were missing from the Unicode classes such as C<IsDigit>.
2279 Large unsigned numbers (those above 2**31) could sometimes lose their
2280 unsignedness, causing bogus results in arithmetic operations.
2284 The Perl parser has been stress tested using both random input and
2285 Markov chain input and the few found crashes and lockups have been
2290 =head2 Platform Specific Changes and Fixes
2298 Perl now works on post-4.0 BSD/OSes.
2304 Setting C<$0> now works (as much as possible; see L<perlvar> for details).
2310 Numerous updates; currently synchronised with Cygwin 1.3.10.
2314 Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.
2320 EPOC update after Perl 5.6.0. See README.epoc.
2326 Perl now works on post-3.0 FreeBSDs.
2332 README.hpux updated; C<Configure -Duse64bitall> now works;
2333 now uses HP-UX malloc instead of Perl malloc.
2339 Numerous compilation flag and hint enhancements; accidental mixing
2340 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
2350 Long doubles should now work (see INSTALL).
2354 Linux previously had problems related to sockaddrlen when using
2355 accept(), recvfrom() (in Perl: recv()), getpeername(), and
2364 Compilation of the standard Perl distribution in MacOS Classic should
2365 now work if you have the Metrowerks development environment and
2366 the missing Mac-specific toolkit bits. Contact the macperl mailing
2373 MPE/iX update after Perl 5.6.0. See README.mpeix.
2377 NetBSD/threads: try installing the GNU pth (should be in the
2378 packages collection, or http://www.gnu.org/software/pth/),
2379 and Configure with -Duseithreads.
2385 Perl now works on NetBSD/sparc.
2391 Now works with usethreads (see INSTALL).
2397 64-bitness using the Sun Workshop compiler now works.
2403 The native build method requires at least VOS Release 14.5.0
2404 and GNU C++/GNU Tools 2.0.1 or later. The Perl pack function
2405 now maps overflowed values to +infinity and underflowed values
2410 Tru64 (aka Digital UNIX, aka DEC OSF/1)
2412 The operating system version letter now recorded in $Config{osvers}.
2413 Allow compiling with gcc (previously explicitly forbidden). Compiling
2414 with gcc still not recommended because buggy code results, even with
2421 Fixed various alignment problems that lead into core dumps either
2422 during build or later; no longer dies on math errors at runtime;
2423 now using full quad integers (64 bits), previously was using
2424 only 46 bit integers for speed.
2430 chdir() now works better despite a CRT bug; now works with MULTIPLICITY
2431 (see INSTALL); now works with Perl's malloc.
2433 The tainting of C<%ENV> elements via C<keys> or C<values> was previously
2434 unimplemented. It now works as documented.
2436 The C<waitpid> emulation has been improved. The worst bug (now fixed)
2437 was that a pid of -1 would cause a wildcard search of all processes on
2440 POSIX-style signals are now emulated much better on VMS versions prior
2443 The C<system> function and backticks operator have improved
2444 functionality and better error handling.
2446 File access tests now use current process privileges rather than the
2447 user's default privileges, which could sometimes result in a mismatch
2448 between reported access and actual access.
2450 There is a new C<kill> implementation based on C<sys$sigprc> that allows
2451 older VMS systems (pre-7.0) to use C<kill> to send signals rather than
2452 simply force exit. This implementation also allows later systems to
2453 call C<kill> from within a signal handler.
2455 Iterative logical name translations are now limited to 10 iterations in
2456 imitation of SHOW LOGICAL and other OpenVMS facilities.
2466 accept() no longer leaks memory.
2470 Borland C++ v5.5 is now a supported compiler that can build Perl.
2471 However, the generated binaries continue to be incompatible with those
2472 generated by the other supported compilers (GCC and Visual C++).
2476 Better chdir() return value for a non-existent directory.
2480 Duping socket handles with open(F, ">&MYSOCK") now works under Windows 9x.
2484 New %ENV entries now propagate to subprocesses.
2488 Current directory entries in %ENV are now correctly propagated to child
2493 $ENV{LIB} now used to search for libs under Visual C.
2497 fork() emulation has been improved in various ways, but still continues
2498 to be experimental. See L<perlfork> for known bugs and caveats.
2502 A failed (pseudo)fork now returns undef and sets errno to EAGAIN.
2506 Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root.
2507 Other bugs in chdir() and Cwd::cwd() have also been fixed.
2511 HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html
2515 The makefiles now provide a single switch to bulk-enable all the features
2516 enabled in ActiveState ActivePerl (a popular Win32 binary distribution).
2520 Allow REG_EXPAND_SZ keys in the registry.
2524 Can now send() from all threads, not just the first one.
2528 Fake signal handling reenabled, bugs and all.
2532 %SIG has been enabled under USE_ITHREADS, but its use is completely
2533 unsupported under all configurations.
2537 Less stack reserved per thread so that more threads can run
2538 concurrently. (Still 16M per thread.)
2542 C<< File::Spec->tmpdir() >> now prefers C:/temp over /tmp
2543 (works better when perl is running as service).
2547 Better UNC path handling under ithreads.
2551 wait(), waitpid(), and backticks now return the correct exit status
2556 Win64 compilation is now supported.
2560 winsock handle leak fixed.
2566 =head1 New or Changed Diagnostics
2572 The lexical warnings category "deprecated" is no longer a sub-category
2573 of the "syntax" category. It is now a top-level category in its own
2578 All regular expression compilation error messages are now hopefully
2579 easier to understand both because the error message now comes before
2580 the failed regex and because the point of failure is now clearly
2581 marked by a C<E<lt>-- HERE> marker.
2585 The various "opened only for", "on closed", "never opened" warnings
2586 drop the C<main::> prefix for filehandles in the C<main> package,
2587 for example C<STDIN> instead of C<main::STDIN>.
2591 The "Unrecognized escape" warning has been extended to include C<\8>,
2592 C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters.
2596 Two new debugging options have been added: if you have compiled your
2597 Perl with debugging, you can use the -DT and -DR options to trace
2598 tokenising and to add reference counts to displaying variables,
2603 The debugger (perl5db.pl) has been modified to present a more
2604 consistent commands interface, via (CommandSet=580). perl5db.t was
2605 also added to test the changes, and as a placeholder for further tests.
2611 The debugger has a new C<dumpDepth> option to control the maximum
2612 depth to which nested structures are dumped. The C<x> command has
2613 been extended so that C<x N EXPR> dumps out the value of I<EXPR> to a
2614 depth of at most I<N> levels.
2618 The debugger can now show lexical variables if you have the CPAN
2619 module PadWalker installed.
2623 If an attempt to use a (non-blessed) reference as an array index
2624 is made, a warning is given.
2628 C<push @a;> and C<unshift @a;> (with no values to push or unshift)
2629 now give a warning. This may be a problem for generated and evaled
2634 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
2635 using the C<"C"> format you will get an optional warning. Similarly
2636 for the C<"c"> format and a number less than -128 or more than 127.
2640 Certain regex modifiers such as C<(?o)> make sense only if applied to
2641 the entire regex. You will get an optional warning if you try to do
2646 Using arrays or hashes as references (e.g. C<< %foo->{bar} >>
2647 has been deprecated for a while. Now you will get an optional warning.
2651 Using C<sort> in scalar context now issues an optional warning.
2652 This didn't do anything useful, as the sort was not performed.
2656 =head1 Changed Internals
2662 perlapi.pod (a companion to perlguts) now attempts to document the
2667 You can now build a really minimal perl called microperl.
2668 Building microperl does not require even running Configure;
2669 C<make -f Makefile.micro> should be enough. Beware: microperl makes
2670 many assumptions, some of which may be too bold; the resulting
2671 executable may crash or otherwise misbehave in wondrous ways.
2672 For careful hackers only.
2676 Added rsignal(), whichsig(), do_join(), op_clear, op_null,
2677 ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2678 interfaces to the publicised API. For the full list of the available
2679 APIs see L<perlapi>.
2683 Made possible to propagate customised exceptions via croak()ing.
2687 Now xsubs can have attributes just like subs. (Well, at least the
2688 built-in attributes.)
2692 dTHR and djSP have been obsoleted; the former removed (because it's
2693 a no-op) and the latter replaced with dSP.
2697 PERL_OBJECT has been completely removed.
2701 The MAGIC constants (e.g. C<'P'>) have been macrofied
2702 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
2703 and maintainability.
2707 The regex compiler now maintains a structure that identifies nodes in
2708 the compiled bytecode with the corresponding syntactic features of the
2709 original regex expression. The information is attached to the new
2710 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
2711 complete information.
2715 The C code has been made much more C<gcc -Wall> clean. Some warning
2716 messages still remain in some platforms, so if you are compiling with
2717 gcc you may see some warnings about dubious practices. The warnings
2718 are being worked on.
2722 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
2726 Documentation on how to use the Perl source repository has been added
2727 to F<Porting/repository.pod>.
2731 There are now several profiling make targets.
2735 =head1 Security Vulnerability Closed
2737 (This change was already made in 5.7.0 but bears repeating here.)
2739 A potential security vulnerability in the optional suidperl component
2740 of Perl was identified in August 2000. suidperl is neither built nor
2741 installed by default. As of November 2001 the only known vulnerable
2742 platform is Linux, most likely all Linux distributions. CERT and
2743 various vendors and distributors have been alerted about the vulnerability.
2744 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2745 for more information.
2747 The problem was caused by Perl trying to report a suspected security
2748 exploit attempt using an external program, /bin/mail. On Linux
2749 platforms the /bin/mail program had an undocumented feature which
2750 when combined with suidperl gave access to a root shell, resulting in
2751 a serious compromise instead of reporting the exploit attempt. If you
2752 don't have /bin/mail, or if you have 'safe setuid scripts', or if
2753 suidperl is not installed, you are safe.
2755 The exploit attempt reporting feature has been completely removed from
2756 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2757 from all the Perl 5.7 releases), so that particular vulnerability
2758 isn't there anymore. However, further security vulnerabilities are,
2759 unfortunately, always possible. The suidperl functionality is most
2760 probably going to be removed in Perl 5.10. In any case, suidperl
2761 should only be used by security experts who know exactly what they are
2762 doing and why they are using suidperl instead of some other solution
2763 such as sudo ( see http://www.courtesan.com/sudo/ ).
2767 Several new tests have been added, especially for the F<lib> and F<ext>
2768 subsections. There are now about 65 000 individual tests (spread over
2769 about 700 test scripts), in the regression suite (5.6.1 has about
2770 11700 tests, in 258 test scripts) Many of the new tests are of course
2771 introduced by the new modules, but still in general Perl is now more
2774 Because of the large number of tests, running the regression suite
2775 will take considerably longer time than it used to: expect the suite
2776 to take up to 4-5 times longer to run than in perl 5.6. On a really
2777 fast machine you can hope to finish the suite in about 6-8 minutes
2780 The tests are now reported in a different order than in earlier Perls.
2781 (This happens because the test scripts from under t/lib have been moved
2782 to be closer to the library/extension they are testing.)
2784 =head1 Known Problems
2792 In AIX 4.2, Perl extensions that use C++ functions that use statics
2793 may have problems in that the statics are not getting initialized.
2794 In newer AIX releases, this has been solved by linking Perl with
2795 the libC_r library, but unfortunately in AIX 4.2 the said library
2796 has an obscure bug where the various functions related to time
2797 (such as time() and gettimeofday()) return broken values, and
2798 therefore in AIX 4.2 Perl is not linked against libC_r.
2802 vac 5.0.0.0 May Produce Buggy Code For Perl
2804 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
2805 resulting in a few random tests failing when run as part of "make
2806 test", but when the failing tests are run by hand, they succeed.
2807 We suggest upgrading to at least vac version 5.0.1.0, that has been
2808 known to compile Perl correctly. "lslpp -L|grep vac.C" will tell
2809 you the vac version. See README.aix.
2813 If building threaded Perl, you may get compilation warning from pp_sys.c:
2815 "pp_sys.c", line 4651.39: 1506-280 (W) Function argument assignment between types "unsigned char*" and "const void*" is not allowed.
2817 This is harmless; it is caused by the getnetbyaddr() and getnetbyaddr_r()
2818 having slightly different types for their first argument.
2822 =head2 Alpha systems with old gccs fail several tests
2824 If you see op/pack, op/pat, op/regexp, or ext/Storable tests failing
2825 in a Linux/alpha or *BSD/Alpha, it's probably time to upgrade your gcc.
2826 gccs prior to 2.95.3 are definitely not good enough, and gcc 3.1 may
2827 be even better. (RedHat Linux/alpha with gcc 3.1 reported no problems,
2828 as did Linux 2.4.18 with gcc 2.95.4.) (In Tru64, it is preferable to
2829 use the bundled C compiler.)
2833 Perl 5.8.0 doesn't build in AmigaOS. It broke at some point
2834 during the ithreads work and we could not find Amiga experts
2835 to unbreak the problems.
2839 The following tests fail on 5.8.0 Perl in BeOS Personal 5.03:
2841 t/op/lfs............................FAILED at test 17
2842 t/op/magic..........................FAILED at test 24
2843 ext/POSIX/t/sigaction...............FAILED at test 13
2844 ext/POSIX/t/waitpid.................FAILED at test 1
2846 See L<perlbeos> (README.beos) for more details.
2848 =head2 Cygwin "unable to remap"
2850 For example when building the Tk extension for Cygwin,
2851 you may get an error message saying "unable to remap".
2852 This is known problem with Cygwin, and a workaround is
2853 detailed in here: http://sources.redhat.com/ml/cygwin/2001-12/msg00894.html
2855 =head2 ext/threads/t/libc
2857 If this test fails, it indicates that your libc (C library) is not
2858 threadsafe. This particular test stress tests the localtime() call to
2859 find out whether it is threadsafe. See L<perlthrtut> for more information.
2861 =head2 FreeBSD Failing locale Test 117 For ISO8859-15 Locales
2863 The ISO8859-15 locales may fail the locale test 117 in FreeBSD.
2864 This is caused by the characters \xFF (y with diaeresis) and \xBE
2865 (Y with diaeresis) not behaving correctly when being matched
2868 =head2 IRIX fails ext/List/Util/t/shuffle.t
2870 IRIX with MIPSpro 7.3.1.3m compiler may fail the said List::Util test
2871 by dumping core. This seems to be a compiler error since if compiled
2872 with gcc no core dump ensues, and no failures on the said test on any
2875 =head2 Modifying $_ Inside for(..)
2879 works without complaint. It shouldn't. (You should be able to
2880 modify only lvalue elements inside the loops.) You can see the
2881 correct behaviour by replacing the 1..5 with 1, 2, 3, 4, 5.
2883 =head2 mod_perl 1.26 Doesn't Build With Threaded Perl
2885 Use mod_perl 1.27 or higher.
2887 =head2 lib/ftmp-security tests warn 'system possibly insecure'
2889 Don't panic. Read the 'make test' section of INSTALL instead.
2891 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
2893 If perl is configured with -Duse64bitall, the successful result of the
2894 subtest 10 of lib/posix may arrive before the successful result of the
2895 subtest 9, which confuses the test harness so much that it thinks the
2898 =head2 Linux with glibc 2.2.5 fails t/op/int subtest #6 with -Duse64bitint
2900 This is a known bug in the glibc 2.2.5 with long long integers.
2901 ( http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=65612 )
2903 =head2 Linux With Sfio Fails op/misc Test 48
2907 =head2 libwww-perl (LWP) fails base/date #51
2909 Use libwww-perl 5.65 or later.
2913 Please remember to set your environment variable LC_ALL to "C"
2914 (setenv LC_ALL C) before running "make test" to avoid a lot of
2915 warnings about the broken locales of Mac OS X.
2917 The following tests are known to fail in Mac OS X 10.1.4 because of
2918 buggy (old) implementations of Berkeley DB included in Mac OS X:
2920 Failed Test Stat Wstat Total Fail Failed List of Failed
2921 -------------------------------------------------------------------------
2922 ../ext/DB_File/t/db-btree.t 0 11 ?? ?? % ??
2923 ../ext/DB_File/t/db-recno.t 149 3 2.01% 61 63 65
2925 If you are building on a UFS partition, you will also probably see
2926 t/op/stat.t subtest #9 fail. This is caused by Darwin's UFS not
2927 supporting inode change time.
2929 Also the ext/POSIX/t/posix.t subtest #10 fails but it is skipped for
2930 now because the failure is Apple's fault, not Perl's (blocked signals
2933 If you Configure with ithreads, ext/threads/t/libc.t will fail. Again,
2934 this is not Perl's fault-- the libc of Mac OS X is not threadsafe
2935 (in this particular test, the localtime() call is found to be
2938 =head2 op/sprintf tests 91, 129, and 130
2940 The op/sprintf tests 91, 129, and 130 are known to fail on some platforms.
2941 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
2943 Test 91 is known to fail on QNX6 (nto), because C<sprintf '%e',0>
2944 incorrectly produces C<0.000000e+0> instead of C<0.000000e+00>.
2946 For tests 129 and 130, the failing platforms do not comply with
2947 the ANSI C Standard: lines 19ff on page 134 of ANSI X3.159 1989, to
2948 be exact. (They produce something other than "1" and "-1" when
2949 formatting 0.6 and -0.6 using the printf format "%.0f"; most often,
2950 they produce "0" and "-0".)
2954 In case you are still using Solaris 2.5 (aka SunOS 5.5), you may
2955 experience failures (the test core dumping) in lib/locale.t.
2956 The suggested cure is to upgrade your Solaris.
2960 When Perl is built using the native build process on VOS Release
2961 14.5.0 and GNU C++/GNU Tools 2.0.1, all attempted tests either
2962 pass or result in TODO (ignored) failures.
2964 =head2 Term::ReadKey not working on Win32
2966 Use Term::ReadKey 2.20 or later.
2968 =head2 Failure of Thread (5.005-style) tests
2970 B<Note that support for 5.005-style threading is deprecated,
2971 experimental and practically unsupported. In 5.10, it is expected
2974 The following tests are known to fail due to fundamental problems in
2975 the 5.005 threading implementation. These are not new failures--Perl
2976 5.005_0x has the same bugs, but didn't have these tests.
2978 ../ext/List/Util/t/first.t 255 65280 7 4 57.14% 2 5-7
2979 ../lib/English.t 2 512 54 2 3.70% 2-3
2980 ../lib/Filter/Simple/t/data.t 6 3 50.00% 1-3
2981 ../lib/Filter/Simple/t/filter_only 9 3 33.33% 1-2 5
2982 ../lib/autouse.t 10 1 10.00% 4
2983 op/flip.t 15 1 6.67% 15
2985 These failures are unlikely to get fixed as 5.005-style threads
2986 are considered fundamentally broken. (Basically what happens is that
2987 competing threads can corrupt shared global state.)
2989 =head2 Timing problems
2991 The following tests may fail intermittently because of timing
2992 problems, for example if the system is heavily loaded.
2995 ext/Time/HiRes/HiRes.t
2997 lib/Memoize/t/expmod_t.t
2998 lib/Memoize/t/speed.t
3000 In case of failure please try running them manually, for example
3002 ./perl -Ilib ext/Time/HiRes/HiRes.t
3006 ../lib/Math/Trig.t 26 1 3.85% 25
3007 ../lib/warnings.t 470 1 0.21% 429
3009 The Trig.t failure is caused by the slighly differing (from IEEE)
3010 floating point implementation of UNICOS. The warnings.t failure is
3011 also related: the test assumes a certain floating point output format;
3012 this assumption fails in UNICOS.
3020 During Configure, the test
3022 Guessing which symbols your C compiler and preprocessor define...
3024 will probably fail with error messages like
3026 CC-20 cc: ERROR File = try.c, Line = 3
3027 The identifier "bad" is undefined.
3029 bad switch yylook 79bad switch yylook 79bad switch yylook 79bad switch yylook 79#ifdef A29K
3032 CC-65 cc: ERROR File = try.c, Line = 3
3033 A semicolon is expected at this point.
3035 This is caused by a bug in the awk utility of UNICOS/mk. You can ignore
3036 the error, but it does cause a slight problem: you cannot fully
3037 benefit from the h2ph utility (see L<h2ph>) that can be used to
3038 convert C headers to Perl libraries, mainly used to be able to access
3039 from Perl the constants defined using C preprocessor, cpp. Because of
3040 the above error, parts of the converted headers will be invisible.
3041 Luckily, these days the need for h2ph is rare.
3045 If building Perl with interpreter threads (ithreads), the
3046 getgrent(), getgrnam(), and getgrgid() functions cannot return the
3047 list of the group members due to a bug in the multithreaded support of
3048 UNICOS/mk. What this means is that in list context the functions will
3049 return only three values, not four.
3055 There are a few known test failures, see L<perluts> (README.uts).
3059 There should be no reported test failures with a default configuration,
3060 though there are a number of tests marked TODO that point to areas
3061 needing further debugging and/or porting work.
3065 In multi-CPU boxes, there are some problems with the I/O buffering:
3066 some output may appear twice.
3068 =head2 XML::Parser not working
3070 Use XML::Parser 2.31 or later.
3072 =head2 z/OS (OS/390)
3074 z/OS has rather many test failures but the situation is actually
3075 better than it was in 5.6.0; it's just that so many new modules and
3076 tests have been added.
3078 Failed Test Stat Wstat Total Fail Failed List of Failed
3079 ---------------------------------------------------------------------------
3080 ../ext/Data/Dumper/t/dumper.t 357 8 2.24% 311 314 325 327
3082 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5
3083 ../ext/Storable/t/downgrade.t 12 3072 169 12 7.10% 14-15 46-47 78-79
3085 ../lib/ExtUtils/t/Constant.t 121 30976 48 48 100.00% 1-48
3086 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9
3087 op/pat.t 910 7 0.77% 665 776 785 832-
3089 op/sprintf.t 224 3 1.34% 98 100 136
3090 op/tr.t 97 5 5.15% 63 71-74
3091 uni/fold.t 780 6 0.77% 61 169 196 661
3094 The failures in dumper.t and downgrade.t are problems in the tests,
3095 those in io_unix and sprintf are problems in the USS (UDP sockets
3096 and printf formats). The pat, tr, and fold failures are genuine Perl
3097 problems caused by EBCDIC (and in the pat and fold cases, combining
3098 that with Unicode). The Constant and Embed are probably problems
3099 in the tests (since they test Perl's ability to build extensions,
3100 and that seems to be working reasonably well.)
3102 =head2 Localising Tied Arrays and Hashes Is Broken
3106 doesn't work as one would expect: the old value is restored
3107 incorrectly. This will be changed in a future release, but we don't
3108 know yet what the new semantics will exactly be. In any case, the
3109 change will break existing code that relies on the current
3110 (ill-defined) semantics, so just avoid doing this in general.
3112 =head2 Self-tying Problems
3114 Self-tying of arrays and hashes is broken in rather deep and
3115 hard-to-fix ways. As a stop-gap measure to avoid people from getting
3116 frustrated at the mysterious results (core dumps, most often), it is
3117 forbidden for now (you will get a fatal error even from an attempt).
3119 A change to self-tying of globs has caused them to be recursively
3120 referenced (see: L<perlobj/"Two-Phased Garbage Collection">). You
3121 will now need an explicit untie to destroy a self-tied glob. This
3122 behaviour may be fixed at a later date.
3124 Self-tying of scalars and IO thingies works.
3126 =head2 Building Extensions Can Fail Because Of Largefiles
3128 Some extensions like mod_perl are known to have issues with
3129 `largefiles', a change brought by Perl 5.6.0 in which file offsets
3130 default to 64 bits wide, where supported. Modules may fail to compile
3131 at all, or they may compile and work incorrectly. Currently, there
3132 is no good solution for the problem, but Configure now provides
3133 appropriate non-largefile ccflags, ldflags, libswanted, and libs
3134 in the %Config hash (e.g., $Config{ccflags_nolargefiles}) so the
3135 extensions that are having problems can try configuring themselves
3136 without the largefileness. This is admittedly not a clean solution,
3137 and the solution may not even work at all. One potential failure is
3138 whether one can (or, if one can, whether it's a good idea to) link
3139 together at all binaries with different ideas about file offsets;
3140 all this is platform-dependent.
3142 =head2 Unicode Support on EBCDIC Still Spotty
3144 Though mostly working, Unicode support still has problem spots on
3145 EBCDIC platforms. One such known spot are the C<\p{}> and C<\P{}>
3146 regular expression constructs for code points less than 256: the
3147 C<pP> are testing for Unicode code points, not knowing about EBCDIC.
3149 =head2 The Compiler Suite Is Still Very Experimental
3151 The compiler suite is slowly getting better but it continues to be
3152 highly experimental. Use in production environments is discouraged.
3154 =head2 The Long Double Support Is Still Experimental
3156 The ability to configure Perl's numbers to use "long doubles",
3157 floating point numbers of hopefully better accuracy, is still
3158 experimental. The implementations of long doubles are not yet
3159 widespread and the existing implementations are not quite mature
3160 or standardised, therefore trying to support them is a rare
3161 and moving target. The gain of more precision may also be offset
3162 by slowdown in computations (more bits to move around, and the
3163 operations are more likely to be executed by less optimised
3166 =head2 Seen In Perl 5.7 But Gone Now
3168 C<Time::Piece> (previously known as C<Time::Object>) was removed
3169 because it was felt that it didn't have enough value in it to be a
3170 core module. It is still a useful module, though, and is available
3173 Perl 5.8 unfortunately does not build anymore on AmigaOS;
3174 this broke accidentally at some point. Since there are not that many
3175 Amiga developers available, we could not get this fixed and tested in
3178 =head1 Reporting Bugs
3180 If you find what you think is a bug, you might check the articles
3181 recently posted to the comp.lang.perl.misc newsgroup and the perl
3182 bug database at http://bugs.perl.org/ . There may also be
3183 information at http://www.perl.com/ , the Perl Home Page.
3185 If you believe you have an unreported bug, please run the B<perlbug>
3186 program included with your release. Be sure to trim your bug down
3187 to a tiny but sufficient test case. Your bug report, along with the
3188 output of C<perl -V>, will be sent off to perlbug@perl.org to be
3189 analysed by the Perl porting team.
3193 The F<Changes> file for exhaustive details on what changed.
3195 The F<INSTALL> file for how to build Perl.
3197 The F<README> file for general stuff.
3199 The F<Artistic> and F<Copying> files for copyright information.
3203 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>.