3 perldelta - what is new for perl v5.8.0
7 This document describes differences between the 5.6.0 release and
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
12 coordinated (while 5.8.0 was still called 5.7.something).
14 Changes that were integrated into the 5.6.1 release are marked C<[561]>.
16 You can see the list of changes in the 5.6.1 (from 5.6.0) by reading
19 If you are upgrading from Perl 5.005_03, you might also want
20 to read L<perl56delta>.
22 =head1 Highlights In 5.8.0
28 Better Unicode support
32 New Thread Implementation
40 Better Numeric Accuracy
48 More Extensive Regression Testing
52 =head1 Incompatible Changes
54 =head2 Binary Incompatibility
56 B<Perl 5.8 is not binary compatible with earlier releases of Perl.>
58 B<You have to recompile your XS modules.>
60 (Pure Perl modules should continue to work.)
62 The major reason for the discontinuity is the new IO architecture
63 called PerlIO. PerlIO is the default configuration because without
64 it many new features of Perl 5.8 cannot be used. In other words:
65 you just have to recompile your modules containing XS code, sorry
68 In future releases of Perl, non-PerlIO aware XS modules may become
69 completely unsupported. This shouldn't be too difficult for module
70 authors, however: PerlIO has been designed as a drop-in replacement
71 (at the source code level) for the stdio interface.
73 Depending on your platform, there are also other reasons why
74 we decided to break binary compatibility, please read on.
76 =head2 64-bit platforms and malloc
78 If your pointers are 64 bits wide, the Perl malloc is no longer being
79 used because it does not work well with 8-byte pointers. Also,
80 usually the system mallocs on such platforms are much better optimized
81 for such large memory models than the Perl malloc. Some memory-hungry
82 Perl applications like the PDL don't work well with Perl's malloc.
83 Finally, other applications than Perl (such as mod_perl) tend to prefer
84 the system malloc. Such platforms include Alpha and 64-bit HPPA,
87 =head2 AIX Dynaloading
89 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
90 dlopen interface of AIX instead of the old emulated interface. This
91 change will probably break backward compatibility with compiled
92 modules. The change was made to make Perl more compliant with other
93 applications like mod_perl which are using the AIX native interface.
95 =head2 Attributes for C<my> variables now handled at run-time.
97 The C<my EXPR : ATTRS> syntax now applies variable attributes at
98 run-time. (Subroutine and C<our> variables still get attributes applied
99 at compile-time.) See L<attributes> for additional details. In particular,
100 however, this allows variable attributes to be useful for C<tie> interfaces,
101 which was a deficiency of earlier releases. Note that the new semantics
102 doesn't work with the Attribute::Handlers module (as of version 0.76).
104 =head2 Socket Extension Dynamic in VMS
106 The Socket extension is now dynamically loaded instead of being
107 statically built in. This may or may not be a problem with ancient
108 TCP/IP stacks of VMS: we do not know since we weren't able to test
109 Perl in such configurations.
111 =head2 IEEE-format Floating Point Default on OpenVMS Alpha
113 Perl now uses IEEE format (T_FLOAT) as the default internal floating
114 point format on OpenVMS Alpha, potentially breaking binary compatibility
115 with external libraries or existing data. G_FLOAT is still available as
116 a configuration option. The default on VAX (D_FLOAT) has not changed.
118 =head2 New Unicode Properties
120 Unicode I<scripts> are now supported. Scripts are similar to (and superior
121 to) Unicode I<blocks>. The difference between scripts and blocks is that
122 scripts are the glyphs used by a language or a group of languages, while
123 the blocks are more artificial groupings of (mostly) 256 characters based
124 on the Unicode numbering.
126 In general, scripts are more inclusive, but not universally so. For
127 example, while the script C<Latin> includes all the Latin characters and
128 their various diacritic-adorned versions, it does not include the various
129 punctuation or digits (since they are not solely C<Latin>).
131 A number of other properties are now supported, including C<\p{L&}>,
132 C<\p{Any}> C<\p{Assigned}>, C<\p{Unassigned}>, C<\p{Blank}> [561] and
133 C<\p{SpacePerl}> [561] (along with their C<\P{...}> versions, of course).
134 See L<perlunicode> for details, and more additions.
136 The C<In> or C<Is> prefix to names used with the C<\p{...}> and C<\P{...}>
137 are now almost always optional. The only exception is that a C<In> prefix
138 is required to signify a Unicode block when a block name conflicts with a
139 script name. For example, C<\p{Tibetan}> refers to the script, while
140 C<\p{InTibetan}> refers to the block. When there is no name conflict, you
141 can omit the C<In> from the block name (e.g. C<\p{BraillePatterns}>), but
142 to be safe, it's probably best to always use the C<In>).
144 =head2 REF(...) Instead Of SCALAR(...)
146 A reference to a reference now stringifies as "REF(0x81485ec)" instead
147 of "SCALAR(0x81485ec)" in order to be more consistent with the return
150 =head2 pack/unpack D/F recycled
152 The undocumented pack/unpack template letters D/F have been recycled
153 for better use: now they stand for long double (if supported by the
154 platform) and NV (Perl internal floating point type). (They used
155 to be aliases for d/f, but you never knew that.)
163 The semantics of bless(REF, REF) were unclear and until someone proves
164 it to make some sense, it is forbidden.
168 The obsolete chat2 library that should never have been allowed
169 to escape the laboratory has been decommissioned.
173 The builtin dump() function has probably outlived most of its
174 usefulness. The core-dumping functionality will remain in future
175 available as an explicit call to C<CORE::dump()>, but in future
176 releases the behaviour of an unqualified C<dump()> call may change.
180 The very dusty examples in the eg/ directory have been removed.
181 Suggestions for new shiny examples welcome but the main issue is that
182 the examples need to be documented, tested and (most importantly)
187 The (bogus) escape sequences \8 and \9 now give an optional warning
188 ("Unrecognized escape passed through"). There is no need to \-escape
193 The list of filenames from glob() (or <...>) is now by default sorted
194 alphabetically to be csh-compliant (which is what happened before
195 in most UNIX platforms). (bsd_glob() does still sort platform
196 natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.) [561]
200 Spurious syntax errors generated in certain situations, when glob()
201 caused File::Glob to be loaded for the first time, have been fixed. [561]
205 Although "you shouldn't do that", it was possible to write code that
206 depends on Perl's hashed key order (Data::Dumper does this). The new
207 algorithm "One-at-a-Time" produces a different hashed key order.
208 More details are in L</"Performance Enhancements">.
212 lstat(FILEHANDLE) now gives a warning because the operation makes no sense.
213 In future releases this may become a fatal error.
217 The C<package;> syntax (C<package> without an argument) has been
218 deprecated. Its semantics were never that clear and its
219 implementation even less so. If you have used that feature to
220 disallow all but fully qualified variables, C<use strict;> instead.
224 The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still
225 recognised but now cause fatal errors. The previous behaviour of
226 ignoring them by default and warning if requested was unacceptable
227 since it, in a way, falsely promised that the features could be used.
231 The current user-visible implementation of pseudo-hashes (the weird
232 use of the first array element) is deprecated starting from Perl 5.8.0
233 and will be removed in Perl 5.10.0, and the feature will be
234 implemented differently. Not only is the current interface rather
235 ugly, but the current implementation slows down normal array and hash
236 use quite noticeably. The C<fields> pragma interface will remain
237 available. The I<restricted hashes> interface is expected to
238 be the replacement interface (see L<Hash::Util>).
242 The syntaxes C<< @a->[...] >> and C<< %h->{...} >> have now been deprecated.
246 After years of trying, suidperl is considered to be too complex to
247 ever be considered truly secure. The suidperl functionality is likely
248 to be removed in a future release.
252 The 5.005 threads model (module C<Thread>) is deprecated and expected
253 to be removed in Perl 5.10. Multithreaded code should be migrated to
254 the new ithreads model (see L<threads>, L<threads::shared> and
259 The long deprecated uppercase aliases for the string comparison
260 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
264 The tr///C and tr///U features have been removed and will not return;
265 the interface was a mistake. Sorry about that. For similar
266 functionality, see pack('U0', ...) and pack('C0', ...). [561]
270 Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo (@)".
271 The prototypes are now checked better at compile-time for invalid
272 syntax. An optional warning is generated ("Illegal character in
273 prototype...") but this may be upgraded to a fatal error in a future
278 The existing behaviour when localising tied arrays and hashes is wrong,
279 and will be changed in a future release, so do not rely on the existing
280 behaviour. See L<"Localising Tied Arrays and Hashes Is Broken">.
284 =head1 Core Enhancements
286 =head2 PerlIO is Now The Default
292 IO is now by default done via PerlIO rather than system's "stdio".
293 PerlIO allows "layers" to be "pushed" onto a file handle to alter the
294 handle's behaviour. Layers can be specified at open time via 3-arg
297 open($fh,'>:crlf :utf8', $path) || ...
299 or on already opened handles via extended C<binmode>:
301 binmode($fh,':encoding(iso-8859-7)');
303 The built-in layers are: unix (low level read/write), stdio (as in
304 previous Perls), perlio (re-implementation of stdio buffering in a
305 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
306 but available on any platform). A mmap layer may be available if
307 platform supports it (mostly UNIXes).
309 Layers to be applied by default may be specified via the 'open' pragma.
311 See L</"Installation and Configuration Improvements"> for the effects
312 of PerlIO on your architecture name.
316 File handles can be marked as accepting Perl's internal encoding of Unicode
317 (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :
319 open($fh,">:utf8","Uni.txt");
321 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named
322 for you since it's not UTF-8 what you will be getting but instead
323 UTF-EBCDIC. See L<perlunicode>, L<utf8>, and
324 http://www.unicode.org/unicode/reports/tr16/ for more information.
325 In future releases this naming may change.
329 File handles can translate character encodings from/to Perl's internal
330 Unicode form on read/write via the ":encoding()" layer.
334 File handles can be opened to "in memory" files held in Perl scalars via:
336 open($fh,'>', \$variable) || ...
340 Anonymous temporary files are available without need to
341 'use FileHandle' or other module via
343 open($fh,"+>", undef) || ...
345 That is a literal undef, not an undefined value.
349 The list form of C<open> is now implemented for pipes (at least on UNIX):
351 open($fh,"-|", 'cat', '/etc/motd')
353 creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in
358 If your locale environment variables (LANGUAGE, LC_ALL, LC_CTYPE, LANG)
359 contain the strings 'UTF-8' or 'UTF8' (case-insensitive matching),
360 the default encoding of your STDIN, STDOUT, and STDERR, and of
361 B<any subsequent file open>, is UTF-8.
365 =head2 Restricted Hashes
367 A restricted hash is restricted to a certain set of keys, no keys
368 outside the set can be added. Also individual keys can be restricted
369 so that the key cannot be deleted and the value cannot be changed.
370 No new syntax is involved: the Hash::Util module is the interface.
374 Perl used to be fragile in that signals arriving at inopportune moments
375 could corrupt Perl's internal state. Now Perl postpones handling of
376 signals until it's safe (between opcodes).
378 This change may have surprising side effects because signals no longer
379 interrupt Perl instantly. Perl will now first finish whatever it was
380 doing, like finishing an internal operation (like sort()) or an
381 external operation (like an I/O operation), and only then look at any
382 arrived signals (and before starting the next operation). No more corrupt
383 internal state since the current operation is always finished first,
384 but the signal may take more time to get heard. Note that breaking
385 out from potentially blocking operations should still work, though.
387 =head2 Unicode Overhaul
389 Unicode in general should be now much more usable than in Perl 5.6.0
390 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in
391 regular expressions should work now, Unicode in tr/// should work now,
392 Unicode in I/O should work now. See L<perluniintro> for introduction
393 and L<perlunicode> for details.
399 The Unicode Character Database coming with Perl has been upgraded
400 to Unicode 3.2.0. For more information, see http://www.unicode.org/ .
405 For developers interested in enhancing Perl's Unicode capabilities:
406 almost all the UCD files are included with the Perl distribution in
407 the F<lib/unicore> subdirectory. The most notable omission, for space
408 considerations, is the Unihan database.
412 The properties \p{Blank} and \p{SpacePerl} have been added. "Blank" is like
413 C isblank(), that is, it contains only "horizontal whitespace" (the space
414 character is, the newline isn't), and the "SpacePerl" is the Unicode
415 equivalent of C<\s> (\p{Space} isn't, since that includes the vertical
416 tabulator character, whereas C<\s> doesn't.)
418 See "New Unicode Properties" earlier in this document for additional
419 information on changes with Unicode properties.
423 =head2 Understanding of Numbers
425 In general a lot of fixing has happened in the area of Perl's
426 understanding of numbers, both integer and floating point. Since in
427 many systems the standard number parsing functions like C<strtoul()>
428 and C<atof()> seem to have bugs, Perl tries to work around their
429 deficiencies. This results hopefully in more accurate numbers.
431 Perl now tries internally to use integer values in numeric conversions
432 and basic arithmetics (+ - * /) if the arguments are integers, and
433 tries also to keep the results stored internally as integers.
434 This change leads to often slightly faster and always less lossy
435 arithmetics. (Previously Perl always preferred floating point numbers
438 =head2 Miscellaneous Changes
444 AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute
445 to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.
449 The $Config{byteorder} (and corresponding BYTEORDER in config.h) was
450 previously wrong in platforms if sizeof(long) was 4, but sizeof(IV)
451 was 8. The byteorder was only sizeof(long) bytes long (1234 or 4321),
452 but now it is correctly sizeof(IV) bytes long, (12345678 or 87654321).
453 (This problem didn't affect Windows platforms.)
455 Also, $Config{byteorder} is now computed dynamically--this is more
456 robust with "fat binaries" where an executable image contains binaries
457 for more than one binary platform, and when cross-compiling.
461 C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass
462 in multiple arguments.)
466 The builtin dump() now gives an optional warning
467 C<dump() better written as CORE::dump()>,
468 meaning that by default C<dump(...)> is resolved as the builtin
469 dump() which dumps core and aborts, not as (possibly) user-defined
470 C<sub dump>. To call the latter, qualify the call as C<&dump(...)>.
471 (The whole dump() feature is to considered deprecated, and possibly
472 removed/changed in future releases.)
476 chomp() and chop() are now overridable. Note, however, that their
477 prototype (as given by C<prototype("CORE::chomp")> is undefined,
478 because it cannot be expressed and therefore one cannot really write
479 replacements to override these builtins.
483 END blocks are now run even if you exit/die in a BEGIN block.
484 Internally, the execution of END blocks is now controlled by
485 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new
486 behaviour for Perl embedders. This will default in 5.10. See
491 Formats now support zero-padded decimal fields.
495 Lvalue subroutines can now return C<undef> in list context. However,
496 the lvalue subroutine feature still remains experimental.
500 A lost warning "Can't declare ... dereference in my" has been
501 restored (Perl had it earlier but it became lost in later releases.)
505 A new special regular expression variable has been introduced:
506 C<$^N>, which contains the most-recently closed group (submatch).
510 C<no Module;> does not produce an error even if Module does not have an
511 unimport() method. This parallels the behavior of C<use> vis-a-vis
516 The numerical comparison operators return C<undef> if either operand
517 is a NaN. Previously the behaviour was unspecified.
521 The following builtin functions are now overridable: each(), keys(),
522 pop(), push(), shift(), splice(), unshift(). [561]
526 C<pack() / unpack()> can now group template letters with C<()> and then
527 apply repetition/count modifiers on the groups.
531 C<pack() / unpack()> can now process the Perl internal numeric types:
532 IVs, UVs, NVs-- and also long doubles, if supported by the platform.
533 The template letters are C<j>, C<J>, C<F>, and C<D>.
537 C<pack('U0a*', ...)> can now be used to force a string to UTF8.
541 my __PACKAGE__ $obj now works. [561]
545 POSIX::sleep() now returns the number of I<unslept> seconds
546 (as the POSIX standard says), as opposed to CORE::sleep() which
547 returns the number of slept seconds.
551 The printf() and sprintf() now support parameter reordering using the
552 C<%\d+\$> and C<*\d+\$> syntaxes. For example
554 print "%2\$s %1\$s\n", "foo", "bar";
556 will print "bar foo\n". This feature helps in writing
557 internationalised software, and in general when the order
558 of the parameters can vary.
562 prototype(\&) is now available.
566 prototype(\[$@%&]) is now available to implicitly create references
567 (useful for example if you want to emulate the tie() interface).
571 A new command-line option, C<-t> is available. It is the
572 little brother of C<-T>: instead of dying on taint violations,
573 lexical warnings are given. B<This is only meant as a temporary
574 debugging aid while securing the code of old legacy applications.
575 This is not a substitute for -T.>
579 In other taint news, the C<exec LIST> and C<system LIST> have now been
580 considered too risky (think C<exec @ARGV>: it can start any program
581 with any arguments), and now the said forms cause a warning.
582 You should carefully launder the arguments to guarantee their
583 validity. In future releases of Perl the forms will become fatal
584 errors so consider starting laundering now.
588 Tied hash interfaces are now required to have the EXISTS and DELETE
589 methods (either own or inherited).
593 If tr/// is just counting characters, it doesn't attempt to
598 untie() will now call an UNTIE() hook if it exists. See L<perltie>
603 L<utime> now supports C<utime undef, undef, @files> to change the
604 file timestamps to the current time.
608 The rules for allowing underscores (underbars) in numeric constants
609 have been relaxed and simplified: now you can have an underscore
610 simply B<between digits>.
614 Rather than relying on C's argv[0] (which may not contain a full pathname)
615 where possible $^X is now set by asking the operating system.
616 (eg by reading F</proc/self/exe> on Linux, F</proc/curproc/file> on FreeBSD)
620 A new variable, C<${^TAINT}>, indicates whether taint mode is enabled.
624 You can now override the readline() builtin, and this overrides also
625 the <FILEHANDLE> angle bracket operator.
629 The command-line options -s and -F are now recognized on the shebang
634 Use of the C</c> match modifier without an accompanying C</g> modifier
635 elicits a new warning: C<Use of /c modifier is meaningless without /g>.
637 Use of C</c> in substitutions, even with C</g>, elicits
638 C<Use of /c modifier is meaningless in s///>.
640 Use of C</g> with C<split> elicits C<Use of /g modifier is meaningless
645 =head1 Modules and Pragmata
647 =head2 New Modules and Pragmata
653 C<Attribute::Handlers> allows a class to define attribute handlers.
656 use Attribute::Handlers;
657 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
659 # later, in some package using or inheriting from MyPack...
661 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
663 Both variables and routines can have attribute handlers. Handlers can
664 be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the
665 exact compilation phase (BEGIN, CHECK, INIT, or END).
666 See L<Attribute::Handlers>.
670 C<B::Concise>, by Stephen McCamant, is a new compiler backend for
671 walking the Perl syntax tree, printing concise info about ops.
672 The output is highly customisable. See L<B::Concise>.
676 The new bignum, bigint, and bigrat pragmas, by Tels, implement
677 transparent bignum support (using the Math::BigInt, Math::BigFloat,
678 and Math::BigRat backends).
682 C<Class::ISA>, by Sean Burke, is a module for reporting the search
683 path for a class's ISA tree. See L<Class::ISA>.
687 C<Cwd> now has a split personality: if possible, an XS extension is
688 used, (this will hopefully be faster, more secure, and more robust)
689 but if not possible, the familiar Perl implementation is used.
693 C<Devel::PPPort>, originally by Kenneth Albanowski and now
694 maintained by Paul Marquess, has been added. It is primarily used
695 by C<h2xs> to enhance portability of XS modules between different
696 versions of Perl. See L<Devel::PPPort>.
700 C<Digest>, frontend module for calculating digests (checksums), from
701 Gisle Aas, has been added. See L<Digest>.
705 C<Digest::MD5> for calculating MD5 digests (checksums) as defined in
706 RFC 1321, from Gisle Aas, has been added. See L<Digest::MD5>.
708 use Digest::MD5 'md5_hex';
710 $digest = md5_hex("Thirsty Camel");
712 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
714 NOTE: the C<MD5> backward compatibility module is deliberately not
715 included since its further use is discouraged.
719 C<Encode>, originally by Nick Ing-Simmons and now maintained by Dan
720 Kogai, provides a mechanism to translate between different character
721 encodings. Support for Unicode, ISO-8859-1, and ASCII are compiled in
722 to the module. Several other encodings (like the rest of the
723 ISO-8859, CP*/Win*, Mac, KOI8-R, three variants EBCDIC, Chinese,
724 Japanese, and Korean encodings) are included and can be loaded at
725 runtime. (For space considerations, the largest Chinese encodings
726 have been separated into their own CPAN module, Encode::HanExtra,
727 which Encode will use if available). See L<Encode>.
729 Any encoding supported by Encode module is also available to the
730 ":encoding()" layer if PerlIO is used.
734 C<Hash::Util> is the interface to the new I<restricted hashes>
735 feature. (Implemented by Jeffrey Friedl, Nick Ing-Simmons, and
736 Michael Schwern.) See L<Hash::Util>.
740 C<I18N::Langinfo> can be used to query locale information.
741 See L<I18N::Langinfo>.
745 C<I18N::LangTags>, by Sean Burke, has functions for dealing with
746 RFC3066-style language tags. See L<I18N::LangTags>.
750 C<ExtUtils::Constant>, by Nicholas Clark, is a new tool for extension
751 writers for generating XS code to import C header constants.
752 See L<ExtUtils::Constant>.
756 C<Filter::Simple>, by Damian Conway, is an easy-to-use frontend to
757 Filter::Util::Call. See L<Filter::Simple>.
763 use Filter::Simple sub {
764 while (my ($from, $to) = splice @_, 0, 2) {
773 use MyFilter qr/red/ => 'green';
775 print "red\n"; # this code is filtered, will print "green\n"
776 print "bored\n"; # this code is filtered, will print "bogreen\n"
780 print "red\n"; # this code is not filtered, will print "red\n"
784 C<File::Temp>, by Tim Jenness, allows one to create temporary files
785 and directories in an easy, portable, and secure way. See
790 C<Filter::Util::Call>, by Paul Marquess, provides you with the
791 framework to write I<source filters> in Perl. For most uses, the
792 frontend Filter::Simple is to be preferred. See L<Filter::Util::Call>.
796 C<if>, by Ilya Zakharevich, is a new pragma for conditional inclusion
801 L<libnet>, by Graham Barr, is a collection of perl5 modules related
802 to network programming. See L<Net::FTP>, L<Net::NNTP>, L<Net::Ping>
803 (not part of libnet, but related), L<Net::POP3>, L<Net::SMTP>,
806 Perl installation leaves libnet unconfigured; use F<libnetcfg>
811 C<List::Util>, by Graham Barr, is a selection of general-utility
812 list subroutines, such as sum(), min(), first(), and shuffle().
817 C<Locale::Constants>, C<Locale::Country>, C<Locale::Currency>
818 C<Locale::Language>, and L<Locale::Script>, by Neil Bowers, have
819 been added. They provide the codes for various locale standards, such
820 as "fr" for France, "usd" for US Dollar, and "ja" for Japanese.
824 $country = code2country('jp'); # $country gets 'Japan'
825 $code = country2code('Norway'); # $code gets 'no'
827 See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>,
828 and L<Locale::Language>.
832 C<Locale::Maketext>, by Sean Burke, is a localization framework. See
833 L<Locale::Maketext>, and L<Locale::Maketext::TPJ13>. The latter is an
834 article about software localization, originally published in The Perl
835 Journal #13, and republished here with kind permission.
839 C<Math::BigRat> for big rational numbers, to accompany Math::BigInt and
840 Math::BigFloat, from Tels. See L<Math::BigRat>.
844 C<Memoize> can make your functions faster by trading space for time,
845 from Mark-Jason Dominus. See L<Memoize>.
849 C<MIME::Base64>, by Gisle Aas, allows you to encode data in base64,
850 as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
855 $encoded = encode_base64('Aladdin:open sesame');
856 $decoded = decode_base64($encoded);
858 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
864 C<MIME::QuotedPrint>, by Gisle Aas, allows you to encode data
865 in quoted-printable encoding, as defined in RFC 2045 - I<MIME
866 (Multipurpose Internet Mail Extensions)>.
868 use MIME::QuotedPrint;
870 $encoded = encode_qp("Smiley in Unicode: \x{263a}");
871 $decoded = decode_qp($encoded);
873 print $encoded, "\n"; # "Smiley in Unicode: =263A"
875 MIME::QuotedPrint has been enhanced to provide the basic methods
876 necessary to use it with PerlIO::Via as in :
878 use MIME::QuotedPrint;
879 open($fh,">Via(MIME::QuotedPrint)",$path);
881 See L<MIME::QuotedPrint>.
885 C<NEXT>, by Damian Conway, is a pseudo-class for method redispatch.
890 C<open> is a new pragma for setting the default I/O disciplines
895 C<PerlIO::Scalar>, by Nick Ing-Simmons, provides the implementation
896 of IO to "in memory" Perl scalars as discussed above. It also serves
897 as an example of a loadable PerlIO layer. Other future possibilities
898 include PerlIO::Array and PerlIO::Code. See L<PerlIO::Scalar>.
902 C<PerlIO::Via>, by Nick Ing-Simmons, acts as a PerlIO layer and wraps
903 PerlIO layer functionality provided by a class (typically implemented
906 use MIME::QuotedPrint;
907 open($fh,">Via(MIME::QuotedPrint)",$path);
909 This will automatically convert everything output to C<$fh>
910 to Quoted-Printable. See L<PerlIO::Via>.
914 C<Pod::ParseLink>, by Russ Allbery, has been added,
915 to parse LZ<><> links in pods as described in the new
920 C<Pod::Text::Overstrike>, by Joe Smith, has been added.
921 It converts POD data to formatted overstrike text.
922 See L<Pod::Text::Overstrike>.
926 C<Scalar::Util> is a selection of general-utility scalar subroutines,
927 such as blessed(), reftype(), and tainted(). See L<Scalar::Util>.
931 C<sort> is a new pragma for controlling the behaviour of sort().
935 C<Storable> gives persistence to Perl data structures by allowing the
936 storage and retrieval of Perl data to and from files in a fast and
937 compact binary format. Because in effect Storable does serialisation
938 of Perl data structues, with it you can also clone deep, hierarchical
939 datastructures. Storable was originally created by Raphael Manfredi,
940 but it is now maintained by Abhijit Menon-Sen. Storable has been
941 enhanced to understand the two new hash features, Unicode keys and
942 restricted hashes. See L<Storable>.
946 C<Switch>, by Damian Conway, has been added. Just by saying
950 you have C<switch> and C<case> available in Perl.
956 case 1 { print "number 1" }
957 case "a" { print "string a" }
958 case [1..10,42] { print "number in list" }
959 case (@array) { print "number in list" }
960 case /\w+/ { print "pattern" }
961 case qr/\w+/ { print "pattern" }
962 case (%hash) { print "entry in hash" }
963 case (\%hash) { print "entry in hash" }
964 case (\&sub) { print "arg to subroutine" }
965 else { print "previous case not true" }
972 C<Test::More>, by Michael Schwern, is yet another framework for writing
973 test scripts, more extensive than Test::Simple. See L<Test::More>.
977 C<Test::Simple>, by Michael Schwern, has basic utilities for writing
978 tests. See L<Test::Simple>.
982 C<Text::Balanced>, by Damian Conway, has been added, for extracting
983 delimited text sequences from strings.
985 use Text::Balanced 'extract_delimited';
987 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
989 $a will be "'never say never'", $b will be ', he never said'.
991 In addition to extract_delimited(), there are also extract_bracketed(),
992 extract_quotelike(), extract_codeblock(), extract_variable(),
993 extract_tagged(), extract_multiple(), gen_delimited_pat(), and
994 gen_extract_tagged(). With these, you can implement rather advanced
995 parsing algorithms. See L<Text::Balanced>.
999 C<threads>, by Arthur Bergman, is an interface to interpreter threads.
1000 Interpreter threads (ithreads) is the new thread model introduced in
1001 Perl 5.6 but only available as an internal interface for extension
1002 writers (and for Win32 Perl for C<fork()> emulation). See L<threads>,
1003 L<threads::shared>, and L<perlthrtut>.
1007 C<threads::shared>, by Arthur Bergman, allows data sharing for
1008 interpreter threads. In the ithreads model any data sharing between
1009 threads must be explicit, as opposed to the old 5.005 thread model
1010 where data sharing was implicit. See L<threads::shared>.
1014 C<Tie::File>, by Mark-Jason Dominus, associates a Perl array with the
1015 lines of a file. See L<Tie::File>.
1019 C<Tie::Memoize>, by Ilya Zakharevich, provides on-demand loaded hashes.
1020 See L<Tie::Memoize>.
1024 C<Tie::RefHash::Nestable>, by Edward Avis, allows storing hash
1025 references (unlike the standard Tie::RefHash) The module is contained
1026 within Tie::RefHash. See L<Tie::RefHash>.
1030 C<Time::HiRes>, by Douglas E. Wegscheid, provides high resolution
1031 timing (ualarm, usleep, and gettimeofday). See L<Time::HiRes>.
1035 C<Unicode::UCD> offers a querying interface to the Unicode Character
1036 Database. See L<Unicode::UCD>.
1040 C<Unicode::Collate>, by SADAHIRO Tomoyuki, implements the UCA
1041 (Unicode Collation Algorithm) for sorting Unicode strings.
1042 See L<Unicode::Collate>.
1046 C<Unicode::Normalize>, by SADAHIRO Tomoyuki, implements the various
1047 Unicode normalization forms. See L<Unicode::Normalize>.
1051 C<XS::Typemap>, by Tim Jenness, is a test extension that exercises
1052 XS typemaps. Nothing gets installed, but the code is worth studying
1053 for extension writers.
1057 =head2 Updated And Improved Modules and Pragmata
1063 The following independently supported modules have been updated to the
1064 newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp,
1065 Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle
1066 (Pod::Man, Pod::Text), Pod::LaTeX, Pod::Parser, Storable,
1067 Term::ANSIColor, Test, Text-Tabs+Wrap.
1071 attributes::reftype() now works on tied arguments.
1075 AutoLoader can now be disabled with C<no AutoLoader;>.
1079 B::Deparse has been significantly enhanced by Robin Houston. It can
1080 now deparse almost all of the standard test suite (so that the tests
1081 still succeed). There is a make target "test.deparse" for trying this
1086 Carp now has better interface documentation, and the @CARP_NOT
1087 interface has been added to get optional control over where errors
1088 are reported independently of @ISA, by Ben Tilly.
1092 Class::Struct can now define the classes in compile time.
1096 Class::Struct now assigns the array/hash element if the accessor
1097 is called with an array/hash element as the B<sole> argument.
1101 The return value of Cwd::fastcwd() is now tainted.
1105 Data::Dumper now has an option to sort hashes.
1109 Data::Dumper now has an option to dump code references
1114 DB_File now supports newer Berkeley DB versions, among
1119 Devel::Peek now has an interface for the Perl memory statistics
1120 (this works only if you are using perl's malloc, and if you have
1121 compiled with debugging).
1125 The English module can now be used without the infamous performance
1128 use English '-no_match_vars';
1130 (Assuming, of course, that you don't need the troublesome variables
1131 C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and
1132 C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>.
1136 ExtUtils::MakeMaker now uses File::Spec internally, which hopefully
1137 leads to better portability.
1141 Fcntl, Socket, and Sys::Syslog have been rewritten by Nicholas Clark
1142 to use the new-style constant dispatch section (see L<ExtUtils::Constant>).
1143 This means that they will be more robust and hopefully faster.
1147 File::Find now chdir()s correctly when chasing symbolic links. [561]
1151 File::Find now has pre- and post-processing callbacks. It also
1152 correctly changes directories when chasing symbolic links. Callbacks
1153 (naughtily) exiting with "next;" instead of "return;" now work.
1157 File::Find is now (again) reentrant. It also has been made
1162 The warnings issued by File::Find now belong to their own category.
1163 You can enable/disable them with C<use/no warnings 'File::Find';>.
1167 File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid
1168 prototype mismatch with CORE::glob(). [561]
1172 File::Glob now supports C<GLOB_LIMIT> constant to limit the size of
1173 the returned list of filenames.
1177 IPC::Open3 now allows the use of numeric file descriptors.
1181 IO::Socket now has an atmark() method, which returns true if the socket
1182 is positioned at the out-of-band mark. The method is also exportable
1183 as a sockatmark() function.
1187 IO::Socket::INET has support for the ReusePort option (if your
1188 platform supports it). The Reuse option now has an alias, ReuseAddr.
1189 For clarity, you may want to prefer ReuseAddr.
1193 IO::Socket::INET now supports a value of zero for C<LocalPort>
1194 (usually meaning that the operating system will make one up.)
1198 'use lib' now works identically to @INC. Removing directories
1199 with 'no lib' now works.
1203 Math::BigFloat and Math::BigInt have undergone a full rewrite by Tels.
1204 They are now magnitudes faster, and they support various bignum
1205 libraries such as GMP and PARI as their backends.
1209 Math::Complex handles inf, NaN etc., better.
1213 Net::Ping has been considerably enhanced by Rob Brown: multihoming is
1214 now supported, Win32 functionality is better, there is now time
1215 measuring functionality (optionally high-resolution using
1216 Time::HiRes), and there is now "external" protocol which uses
1217 Net::Ping::External module which runs your external ping utility and
1218 parses the output. A version of Net::Ping::External is available in
1221 Note that some of the Net::Ping tests are disabled when running
1222 under the Perl distribution since one cannot assume one or more
1223 of the following: enabled echo port at localhost, full Internet
1224 connectivity, or sympathetic firewalls. You can set the environment
1225 variable PERL_TEST_Net_Ping to "1" (one) before running the Perl test
1226 suite to enable all the Net::Ping tests.
1230 POSIX::sigaction() is now much more flexible and robust.
1231 You can now install coderef handlers, 'DEFAULT', and 'IGNORE'
1232 handlers, installing new handlers was not atomic.
1236 In Safe, C<%INC> is now localised in a Safe compartment so that
1241 In SDBM_File on dosish platforms, some keys went missing because of
1242 lack of support for files with "holes". A workaround for the problem
1247 In Search::Dict one can now have a pre-processing hook for the
1248 lines being searched.
1252 The Shell module now has an OO interface.
1256 In Sys::Syslog there is now a failover mechanism that will go
1257 through alternative connection mechanisms until the message
1258 is successfully logged.
1262 The Test module has been significantly enhanced.
1266 Time::Local::timelocal() does not handle fractional seconds anymore.
1267 The rationale is that neither does localtime(), and timelocal() and
1268 localtime() are supposed to be inverses of each other.
1272 The vars pragma now supports declaring fully qualified variables.
1273 (Something that C<our()> does not and will not support.)
1277 The C<utf8::> name space (as in the pragma) provides various
1278 Perl-callable functions to provide low level access to Perl's
1279 internal Unicode representation. At the moment only length()
1280 has been implemented.
1284 =head1 Utility Changes
1290 Emacs perl mode (emacs/cperl-mode.el) has been updated to version
1295 F<emacs/e2ctags.pl> is now much faster.
1299 C<enc2xs> is a tool for people adding their own encodings to the
1304 C<h2ph> now supports C trigraphs.
1308 C<h2xs> now produces a template README.
1312 C<h2xs> now uses C<Devel::PPPort> for better portability between
1313 different versions of Perl.
1317 C<h2xs> uses the new L<ExtUtils::Constant|ExtUtils::Constant> module
1318 which will affect newly created extensions that define constants.
1319 Since the new code is more correct (if you have two constants where the
1320 first one is a prefix of the second one, the first constant B<never>
1321 got defined), less lossy (it uses integers for integer constant,
1322 as opposed to the old code that used floating point numbers even for
1323 integer constants), and slightly faster, you might want to consider
1324 regenerating your extension code (the new scheme makes regenerating
1325 easy). L<h2xs> now also supports C trigraphs.
1329 C<libnetcfg> has been added to configure libnet.
1333 C<perlbug> is now much more robust. It also sends the bug report to
1334 perl.org, not perl.com.
1338 C<perlcc> has been rewritten and its user interface (that is,
1339 command line) is much more like that of the UNIX C compiler, cc.
1340 (The perlbc tools has been removed. Use C<perlcc -B> instead.)
1341 B<Note that perlcc is still considered very experimental and
1346 C<perlivp> is a new Installation Verification Procedure utility
1347 for running any time after installing Perl.
1351 C<piconv> is an implementation of the character conversion utility
1352 C<iconv>, demonstrating the new Encode module.
1356 C<pod2html> now allows specifying a cache directory.
1360 C<pod2html> now produces XHTML 1.0.
1364 C<pod2html> now understands POD written using different line endings
1365 (PC-like CRLF versus UNIX-like LF versus MacClassic-like CR).
1369 C<s2p> has been completely rewritten in Perl. (It is in fact a full
1370 implementation of sed in Perl: you can use the sed functionality by
1371 using the C<psed> utility.)
1375 C<xsubpp> now understands POD documentation embedded in the *.xs
1380 C<xsubpp> now supports the OUT keyword.
1384 =head1 New Documentation
1390 perl56delta details the changes between the 5.005 release and the
1395 perlclib documents the internal replacements for standard C library
1396 functions. (Interesting only for extension writers and Perl core
1401 perldebtut is a Perl debugging tutorial.
1405 perlebcdic contains considerations for running Perl on EBCDIC
1410 perlintro is a gentle introduction to Perl.
1414 perliol documents the internals of PerlIO with layers.
1418 perlmodstyle is a style guide for writing modules.
1422 perlnewmod tells about writing and submitting a new module.
1426 perlpacktut is a pack() tutorial.
1430 perlpod has been rewritten to be clearer and to record the best
1431 practices gathered over the years.
1435 perlpodspec is a more formal specification of the pod format,
1436 mainly of interest for writers of pod applications, not to
1437 people writing in pod.
1441 perlretut is a regular expression tutorial.
1445 perlrequick is a regular expressions quick-start guide.
1446 Yes, much quicker than perlretut. [561]
1450 perltodo has been updated.
1454 perltootc has been renamed as perltooc (to not to conflict
1455 with perltoot in filesystems restricted to "8.3" names).
1459 perluniintro is an introduction to using Unicode in Perl.
1460 (perlunicode is more of a detailed reference and background
1465 perlutil explains the command line utilities packaged with the Perl
1470 The following platform-specific documents are available before
1471 the installation as README.I<platform>, and after the installation
1474 perlaix perlamiga perlapollo perlbeos perlbs2000
1475 perlce perlcygwin perldgux perldos perlepoc perlhpux
1476 perlhurd perlmachten perlmacos perlmint perlmpeix
1477 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1478 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1480 Eastern Asian Perl users are now welcomed in their own languages:
1481 README.jp (Japanese), README.ko (Korean), README.cn (simplified
1482 Chinese) and README.tw (traditional Chinese), which are written in
1483 normal pod but encoded in EUC-JP, EUC-KR, EUC-CN and Big5. These
1484 will get installed as
1486 perljp perlko perlcn perltw
1492 The documentation for the POSIX-BC platform is called "BS2000", to avoid
1493 confusion with the Perl POSIX module.
1497 The documentation for the WinCE platform is called perlce (README.ce
1498 in the source code kit), to avoid confusion with the perlwin32
1499 documentation on 8.3-restricted filesystems.
1503 =head1 Performance Enhancements
1509 map() could get pathologically slow when the result list it generates
1510 is larger than the source list. The performance has been improved for
1511 common scenarios. [561]
1515 sort() has been changed to use primarily mergesort internally as
1516 opposed to the earlier quicksort. For very small lists this may
1517 result in slightly slower sorting times, but in general the speedup
1518 should be at least 20%. Additional bonuses are that the worst case
1519 behaviour of sort() is now better (in computer science terms it now
1520 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1521 worst-case run time behaviour), and that sort() is now stable
1522 (meaning that elements with identical keys will stay ordered as they
1523 were before the sort). See the C<sort> pragma for information.
1525 The story in more detail: suppose you want to serve yourself a little
1528 @digits = ( 3,1,4,1,5,9 );
1530 A numerical sort of the digits will yield (1,1,3,4,5,9), as expected.
1531 Which C<1> comes first is hard to know, since one C<1> looks pretty
1532 much like any other. You can regard this as totally trivial,
1533 or somewhat profound. However, if you just want to sort the even
1534 digits ahead of the odd ones, then what will
1536 sort { ($a % 2) <=> ($b % 2) } @digits;
1538 yield? The only even digit, C<4>, will come first. But how about
1539 the odd numbers, which all compare equal? With the quicksort algorithm
1540 used to implement Perl 5.6 and earlier, the order of ties is left up
1541 to the sort. So, as you add more and more digits of Pi, the order
1542 in which the sorted even and odd digits appear will change.
1543 and, for sufficiently large slices of Pi, the quicksort algorithm
1544 in Perl 5.8 won't return the same results even if reinvoked with the
1545 same input. The justification for this rests with quicksort's
1546 worst case behavior. If you run
1548 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1550 (something you might approximate if you wanted to merge two sorted
1551 arrays using sort), doubling $N doesn't just double the quicksort time,
1552 it I<quadruples> it. Quicksort has a worst case run time that can
1553 grow like N**2, so-called I<quadratic> behaviour, and it can happen
1554 on patterns that may well arise in normal use. You won't notice this
1555 for small arrays, but you I<will> notice it with larger arrays,
1556 and you may not live long enough for the sort to complete on arrays
1557 of a million elements. So the 5.8 quicksort scrambles large arrays
1558 before sorting them, as a statistical defence against quadratic behaviour.
1559 But that means if you sort the same large array twice, ties may be
1560 broken in different ways.
1562 Because of the unpredictability of tie-breaking order, and the quadratic
1563 worst-case behaviour, quicksort was I<almost> replaced completely with
1564 a stable mergesort. I<Stable> means that ties are broken to preserve
1565 the original order of appearance in the input array. So
1567 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1569 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1570 appear in the output in the same order they appeared in the input.
1571 Mergesort has worst case O(N log N) behaviour, the best value
1572 attainable. And, ironically, this mergesort does particularly
1573 well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N)
1574 in O(N) time. But quicksort was rescued at the last moment because
1575 it is faster than mergesort on certain inputs and platforms.
1576 For example, if you really I<don't> care about the order of even
1577 and odd digits, quicksort will run in O(N) time; it's very good
1578 at sorting many repetitions of a small number of distinct elements.
1579 The quicksort divide and conquer strategy works well on platforms
1580 with relatively small, very fast, caches. Eventually, the problem gets
1581 whittled down to one that fits in the cache, from which point it
1582 benefits from the increased memory speed.
1584 Quicksort was rescued by implementing a sort pragma to control aspects
1585 of the sort. The B<stable> subpragma forces stable behaviour,
1586 regardless of algorithm. The B<_quicksort> and B<_mergesort>
1587 subpragmas are heavy-handed ways to select the underlying implementation.
1588 The leading C<_> is a reminder that these subpragmas may not survive
1589 beyond 5.8. More appropriate mechanisms for selecting the implementation
1590 exist, but they wouldn't have arrived in time to save quicksort.
1594 Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm
1595 ( http://burtleburtle.net/bob/hash/doobs.html ). This algorithm is
1596 reasonably fast while producing a much better spread of values than
1597 the old hashing algorithm (originally by Chris Torek, later tweaked by
1598 Ilya Zakharevich). Hash values output from the algorithm on a hash of
1599 all 3-char printable ASCII keys comes much closer to passing the
1600 DIEHARD random number generation tests. According to perlbench, this
1601 change has not affected the overall speed of Perl.
1605 unshift() should now be noticeably faster.
1609 =head1 Installation and Configuration Improvements
1611 =head2 Generic Improvements
1617 INSTALL now explains how you can configure Perl to use 64-bit
1618 integers even on non-64-bit platforms.
1622 Policy.sh policy change: if you are reusing a Policy.sh file
1623 (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1624 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1625 them will now be changed to the new prefix, /foo/bar. (Previously
1626 only $prefix changed.) If you do not like this new behaviour,
1627 specify prefix, siteprefix, and vendorprefix explicitly.
1631 A new optional location for Perl libraries, otherlibdirs, is available.
1632 It can be used for example for vendor add-ons without disturbing Perl's
1633 own library directories.
1637 In many platforms, the vendor-supplied 'cc' is too stripped-down to
1638 build Perl (basically, 'cc' doesn't do ANSI C). If this seems
1639 to be the case and 'cc' does not seem to be the GNU C compiler
1640 'gcc', an automatic attempt is made to find and use 'gcc' instead.
1644 gcc needs to closely track the operating system release to avoid
1645 build problems. If Configure finds that gcc was built for a different
1646 operating system release than is running, it now gives a clearly visible
1647 warning that there may be trouble ahead.
1651 Since Perl 5.8 is not binary-compatible with previous releases
1652 of Perl, Configure no longer suggests including the 5.005
1657 Configure C<-S> can now run non-interactively.
1661 Configure support for pdp11-style memory models has been removed due
1662 to obsolescence. [561]
1666 configure.gnu now works with options with whitespace in them.
1670 installperl now outputs everything to STDERR.
1674 Because PerlIO is now the default on most platforms, "-perlio" doesn't
1675 get appended to the $Config{archname} (also known as $^O) anymore.
1676 Instead, if you explicitly choose not to use perlio (Configure command
1677 line option -Uuseperlio), you will get "-stdio" appended.
1681 Another change related to the architecture name is that "-64all"
1682 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1683 pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)
1687 In AFS installations, one can configure the root of the AFS to be
1688 somewhere else than the default F</afs> by using the Configure
1689 parameter C<-Dafsroot=/some/where/else>.
1693 APPLLIB_EXP, a lesser-known configuration-time definition, has been
1694 documented. It can be used to prepend site-specific directories
1695 to Perl's default search path (@INC); see INSTALL for information.
1699 The version of Berkeley DB used when the Perl (and, presumably, the
1700 DB_File extension) was built is now available as
1701 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
1702 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
1703 DB_VERSION_PATCH_CFG> from C.
1707 Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM
1708 has been documented in INSTALL.
1712 If you have CPAN access (either network or a local copy such as a
1713 CD-ROM) you can during specify extra modules to Configure to build and
1714 install with Perl using the -Dextras=... option. See INSTALL for
1719 In addition to config.over, a new override file, config.arch, is
1720 available. This file is supposed to be used by hints file writers
1721 for architecture-wide changes (as opposed to config.over which is
1722 for site-wide changes).
1726 If your file system supports symbolic links, you can build Perl outside
1727 of the source directory by
1729 mkdir /tmp/perl/build/directory
1730 cd /tmp/perl/build/directory
1731 sh /path/to/perl/source/Configure -Dmksymlinks ...
1733 This will create in /tmp/perl/build/directory a tree of symbolic links
1734 pointing to files in /path/to/perl/source. The original files are left
1735 unaffected. After Configure has finished, you can just say
1739 and Perl will be built and tested, all in /tmp/perl/build/directory.
1744 For Perl developers, several new make targets for profiling
1745 and debugging have been added; see L<perlhack>.
1751 Use of the F<gprof> tool to profile Perl has been documented in
1752 L<perlhack>. There is a make target called "perl.gprof" for
1753 generating a gprofiled Perl executable.
1757 If you have GCC 3, there is a make target called "perl.gcov" for
1758 creating a gcoved Perl executable for coverage analysis. See
1763 If you are on IRIX or Tru64 platforms, new profiling/debugging options
1764 have been added; see L<perlhack> for more information about pixie and
1771 Guidelines of how to construct minimal Perl installations have
1772 been added to INSTALL.
1776 The Thread extension is now not built at all under ithreads
1777 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
1778 Thread extension requires being Configured with C<-Duse5005threads>).
1780 But note that the Thread.pm interface is now shared by both
1785 The Gconvert macro ($Config{d_Gconvert}) used by perl for stringifying
1786 floating-point numbers is now more picky about using sprintf %.*g
1787 rules for the conversion. Some platforms that used to use gcvt may
1788 now resort to the slower sprintf.
1792 The obsolete method of making a special (e.g., debugging) flavor
1795 make LIBPERL=libperld.a
1797 has been removed. Use -DDEBUGGING instead.
1801 =head2 New Or Improved Platforms
1803 For the list of platforms known to support Perl,
1804 see L<perlport/"Supported Platforms">.
1810 AIX dynamic loading should be now better supported.
1814 AIX should now work better with gcc, threads, and 64-bitness. Also the
1815 long doubles support in AIX should be better now. See L<perlaix>.
1819 AtheOS ( http://www.atheos.cx/ ) is a new platform.
1823 BeOS has been reclaimed.
1827 The DG/UX platform now supports 5.005-style threads.
1832 The DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near
1837 EBCDIC platforms (z/OS (also known as OS/390), POSIX-BC, and VM/ESA)
1838 have been regained. Many test suite tests still fail and the
1839 co-existence of Unicode and EBCDIC isn't quite settled, but the
1840 situation is much better than with Perl 5.6. See L<perlos390>,
1841 L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information.
1845 Building perl with -Duseithreads or -Duse5005threads now works under
1846 HP-UX 10.20 (previously it only worked under 10.30 or later). You will
1847 need a thread library package installed. See README.hpux. [561]
1851 Mac OS Classic (MacPerl has of course been available since
1852 perl 5.004 but now the source code bases of standard Perl
1853 and MacPerl have been synchronised)
1857 Mac OS X (or Darwin) should now be able to build Perl even on HFS+
1858 filesystems. (The case-insensitivity used to confuse the Perl build
1863 NCR MP-RAS is now supported. [561]
1867 All the NetBSD specific patches (except for the installation
1868 specific ones) have been merged back to the main distribution.
1872 NetWare from Novell is now supported. See L<perlnetware>.
1876 NonStop-UX is now supported. [561]
1880 NEC SUPER-UX is now supported.
1884 All the OpenBSD specific patches (except for the installation
1885 specific ones) have been merged back to the main distribution.
1889 Perl has been tested with the GNU pth userlevel thread package
1890 ( http://www.gnu.org/software/pth/pth.html ) . All but one thread
1891 test worked, and that one failure was because of test results arriving
1892 in unexpected order.
1896 Stratus VOS is now supported using Perl's native build method
1897 (Configure). This is the recommended method to build Perl on
1898 VOS. The older methods, which build miniperl, are still
1899 available. See L<perlvos>.
1903 The Amdahl UTS UNIX mainframe platform is now supported. [561]
1907 WinCE is now supported. See L<perlce>.
1911 z/OS (formerly known as OS/390, formerly known as MVS OE) now has
1912 support for dynamic loading. This is not selected by default,
1913 however, you must specify -Dusedl in the arguments of Configure. [561]
1917 =head1 Selected Bug Fixes
1919 Numerous memory leaks and uninitialized memory accesses have been
1920 hunted down. Most importantly, anonymous subs used to leak quite
1927 The autouse pragma didn't work for Multi::Part::Function::Names.
1931 caller() could cause core dumps in certain situations. Carp was sometimes
1932 affected by this problem. In particular, caller() now returns a
1933 subroutine name of C<(unknown)> for subroutines that have been removed
1934 from the symbol table.
1938 chop(@list) in list context returned the characters chopped in
1939 reverse order. This has been reversed to be in the right order. [561]
1943 Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm)
1944 when building the Perl binary. The only exception to this is SunOS 4.x,
1945 which needs them. [561]
1949 The behaviour of non-decimal but numeric string constants such as
1950 "0x23" was platform-dependent: in some platforms that was seen as 35,
1951 in some as 0, in some as a floating point number (don't ask). This
1952 was caused by Perl's using the operating system libraries in a situation
1953 where the result of the string to number conversion is undefined: now
1954 Perl consistently handles such strings as zero in numeric contexts.
1958 The order of DESTROYs has been made more predictable.
1962 Several debugger fixes: exit code now reflects the script exit code,
1963 condition C<"0"> now treated correctly, the C<d> command now checks
1964 line number, C<$.> no longer gets corrupted, and all debugger output
1965 now goes correctly to the socket if RemotePort is set. [561]
1969 Perl 5.6.0 could emit spurious warnings about redefinition of dl_error()
1970 when statically building extensions into perl. This has been corrected. [561]
1974 L<dprofpp> -R didn't work.
1978 C<*foo{FORMAT}> now works.
1982 Infinity is now recognized as a number.
1986 UNIVERSAL::isa no longer caches methods incorrectly. (This broke
1987 the Tk extension with 5.6.0.) [561]
1991 Lexicals I: lexicals outside an eval "" weren't resolved
1992 correctly inside a subroutine definition inside the eval "" if they
1993 were not already referenced in the top level of the eval""ed code.
1997 Lexicals II: lexicals leaked at file scope into subroutines that
1998 were declared before the lexicals.
2002 Lexical warnings now propagating correctly between scopes
2003 and into C<eval "...">.
2007 C<use warnings qw(FATAL all)> did not work as intended. This has been
2012 warnings::enabled() now reports the state of $^W correctly if the caller
2013 isn't using lexical warnings. [561]
2017 Line renumbering with eval and C<#line> now works. [561]
2021 Fixed numerous memory leaks, especially in eval "".
2025 Localised tied variables no longer leak memory
2028 tie my %tied_hash => 'Tie::StdHash';
2032 # Used to leak memory every time local() was called;
2033 # in a loop, this added up.
2034 local($tied_hash{Foo}) = 1;
2038 Localised hash elements (and %ENV) are correctly unlocalised to not
2039 exist, if they didn't before they were localised.
2043 tie my %tied_hash => 'Tie::StdHash';
2047 # Nothing has set the FOO element so far
2049 { local $tied_hash{FOO} = 'Bar' }
2051 # This used to print, but not now.
2052 print "exists!\n" if exists $tied_hash{FOO};
2054 As a side effect of this fix, tied hash interfaces B<must> define
2055 the EXISTS and DELETE methods.
2059 mkdir() now ignores trailing slashes in the directory name,
2060 as mandated by POSIX.
2064 Some versions of glibc have a broken modfl(). This affects builds
2065 with C<-Duselongdouble>. This version of Perl detects this brokenness
2066 and has a workaround for it. The glibc release 2.2.2 is known to have
2067 fixed the modfl() bug.
2071 Modulus of unsigned numbers now works (4063328477 % 65535 used to
2072 return 27406, instead of 27047). [561]
2076 Some "not a number" warnings introduced in 5.6.0 eliminated to be
2077 more compatible with 5.005. Infinity is now recognised as a number. [561]
2081 Numeric conversions did not recognize changes in the string value
2082 properly in certain circumstances. [561]
2086 Attributes (such as :shared) didn't work with our().
2090 our() variables will not cause "will not stay shared" warnings.
2094 "our" variables of the same name declared in two sibling blocks
2095 resulted in bogus warnings about "redeclaration" of the variables.
2096 The problem has been corrected. [561]
2100 pack "Z" now correctly terminates the string with "\0".
2104 Fix password routines which in some shadow password platforms
2105 (e.g. HP-UX) caused getpwent() to return every other entry.
2109 The PERL5OPT environment variable (for passing command line arguments
2110 to Perl) didn't work for more than a single group of options. [561]
2114 PERL5OPT with embedded spaces didn't work.
2118 printf() no longer resets the numeric locale to "C".
2122 C<qw(a\\b)> now parses correctly as C<'a\\b'>: that is, as three
2123 characters, not four. [561]
2127 pos() did not return the correct value within s///ge in earlier
2128 versions. This is now handled correctly. [561]
2132 Printing quads (64-bit integers) with printf/sprintf now works
2133 without the q L ll prefixes (assuming you are on a quad-capable platform).
2137 Regular expressions on references and overloaded scalars now work.
2141 Right-hand side magic (GMAGIC) could in many cases such as string
2142 concatenation be invoked too many times.
2146 scalar() now forces scalar context even when used in void context.
2150 SOCKS support is now much more robust.
2154 sort() arguments are now compiled in the right wantarray context
2155 (they were accidentally using the context of the sort() itself).
2156 The comparison block is now run in scalar context, and the arguments
2157 to be sorted are always provided list context. [561]
2161 Changed the POSIX character class C<[[:space:]]> to include the (very
2162 rarely used) vertical tab character. Added a new POSIX-ish character
2163 class C<[[:blank:]]> which stands for horizontal whitespace
2164 (currently, the space and the tab).
2168 The tainting behaviour of sprintf() has been rationalized. It does
2169 not taint the result of floating point formats anymore, making the
2170 behaviour consistent with that of string interpolation. [561]
2174 Some cases of inconsistent taint propagation (such as within hash
2175 values) have been fixed.
2179 The RE engine found in Perl 5.6.0 accidentally pessimised certain kinds
2180 of simple pattern matches. These are now handled better. [561]
2184 Regular expression debug output (whether through C<use re 'debug'>
2185 or via C<-Dr>) now looks better. [561]
2189 Multi-line matches like C<"a\nxb\n" =~ /(?!\A)x/m> were flawed. The
2190 bug has been fixed. [561]
2194 Use of $& could trigger a core dump under some situations. This
2195 is now avoided. [561]
2199 The regular expression captured submatches ($1, $2, ...) are now
2200 more consistently unset if the match fails, instead of leaving false
2201 data lying around in them. [561]
2205 readline() on files opened in "slurp" mode could return an extra
2206 "" (blank line) at the end in certain situations. This has been
2211 Autovivification of symbolic references of special variables described
2212 in L<perlvar> (as in C<${$num}>) was accidentally disabled. This works
2217 Sys::Syslog ignored the C<LOG_AUTH> constant.
2221 All but the first argument of the IO syswrite() method are now optional.
2225 $AUTOLOAD, sort(), lock(), and spawning subprocesses
2226 in multiple threads simultaneously are now thread-safe.
2230 Tie::Array's SPLICE method was broken.
2234 Allow a read-only string on the left-hand side of a non-modifying tr///.
2238 If C<STDERR> is tied, warnings caused by C<warn> and C<die> now
2239 correctly pass to it.
2243 Several Unicode fixes.
2249 BOMs (byte order marks) at the beginning of Perl files
2250 (scripts, modules) should now be transparently skipped.
2251 UTF-16 and UCS-2 encoded Perl files should now be read correctly.
2255 The character tables have been updated to Unicode 3.2.0.
2259 Comparing with utf8 data does not magically upgrade non-utf8 data
2260 into utf8. (This was a problem for example if you were mixing data
2261 from I/O and Unicode data: your output might have got magically encoded
2266 Generating illegal Unicode code points such as U+FFFE, or the UTF-16
2267 surrogates, now also generates an optional warning.
2271 C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase.
2275 Concatenation with the C<.> operator or via variable interpolation,
2276 C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator,
2277 substitution with C<s///>, single-quoted UTF8, should now work.
2281 The C<tr///> operator now works. Note that the C<tr///CU>
2282 functionality has been removed (but see pack('U0', ...)).
2286 C<eval "v200"> now works.
2290 Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spurious warnings.
2291 This has been corrected. [561]
2295 Zero entries were missing from the Unicode classes such as C<IsDigit>.
2301 Large unsigned numbers (those above 2**31) could sometimes lose their
2302 unsignedness, causing bogus results in arithmetic operations. [561]
2306 The Perl parser has been stress tested using both random input and
2307 Markov chain input and the few found crashes and lockups have been
2312 =head2 Platform Specific Changes and Fixes
2320 Perl now works on post-4.0 BSD/OSes.
2326 Setting C<$0> now works (as much as possible; see L<perlvar> for details).
2332 Numerous updates; currently synchronised with Cygwin 1.3.10.
2336 Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.
2342 EPOC now better supported. See README.epoc. [561]
2348 Perl now works on post-3.0 FreeBSDs.
2354 README.hpux updated; C<Configure -Duse64bitall> now works;
2355 now uses HP-UX malloc instead of Perl malloc.
2361 Numerous compilation flag and hint enhancements; accidental mixing
2362 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
2372 Long doubles should now work (see INSTALL). [561]
2376 Linux previously had problems related to sockaddrlen when using
2377 accept(), recvfrom() (in Perl: recv()), getpeername(), and
2386 Compilation of the standard Perl distribution in Mac OS Classic should
2387 now work if you have the Metrowerks development environment and the
2388 missing Mac-specific toolkit bits. Contact the macperl mailing list
2395 MPE/iX update after Perl 5.6.0. See README.mpeix. [561]
2399 NetBSD/threads: try installing the GNU pth (should be in the
2400 packages collection, or http://www.gnu.org/software/pth/),
2401 and Configure with -Duseithreads.
2407 Perl now works on NetBSD/sparc.
2413 Now works with usethreads (see INSTALL). [561]
2419 64-bitness using the Sun Workshop compiler now works.
2425 The native build method requires at least VOS Release 14.5.0
2426 and GNU C++/GNU Tools 2.0.1 or later. The Perl pack function
2427 now maps overflowed values to +infinity and underflowed values
2432 Tru64 (aka Digital UNIX, aka DEC OSF/1)
2434 The operating system version letter now recorded in $Config{osvers}.
2435 Allow compiling with gcc (previously explicitly forbidden). Compiling
2436 with gcc still not recommended because buggy code results, even with
2443 Fixed various alignment problems that lead into core dumps either
2444 during build or later; no longer dies on math errors at runtime;
2445 now using full quad integers (64 bits), previously was using
2446 only 46 bit integers for speed.
2452 chdir() now works better despite a CRT bug; now works with MULTIPLICITY
2453 (see INSTALL); now works with Perl's malloc.
2455 The tainting of C<%ENV> elements via C<keys> or C<values> was previously
2456 unimplemented. It now works as documented.
2458 The C<waitpid> emulation has been improved. The worst bug (now fixed)
2459 was that a pid of -1 would cause a wildcard search of all processes on
2462 POSIX-style signals are now emulated much better on VMS versions prior
2465 The C<system> function and backticks operator have improved
2466 functionality and better error handling. [561]
2468 File access tests now use current process privileges rather than the
2469 user's default privileges, which could sometimes result in a mismatch
2470 between reported access and actual access.
2472 There is a new C<kill> implementation based on C<sys$sigprc> that allows
2473 older VMS systems (pre-7.0) to use C<kill> to send signals rather than
2474 simply force exit. This implementation also allows later systems to
2475 call C<kill> from within a signal handler.
2477 Iterative logical name translations are now limited to 10 iterations in
2478 imitation of SHOW LOGICAL and other OpenVMS facilities.
2488 accept() no longer leaks memory. [561]
2492 Borland C++ v5.5 is now a supported compiler that can build Perl.
2493 However, the generated binaries continue to be incompatible with those
2494 generated by the other supported compilers (GCC and Visual C++). [561]
2498 Better chdir() return value for a non-existent directory.
2502 Duping socket handles with open(F, ">&MYSOCK") now works under Windows
2507 New %ENV entries now propagate to subprocesses. [561]
2511 Current directory entries in %ENV are now correctly propagated to child
2516 $ENV{LIB} now used to search for libs under Visual C.
2520 fork() emulation has been improved in various ways, but still continues
2521 to be experimental. See L<perlfork> for known bugs and caveats.
2525 A failed (pseudo)fork now returns undef and sets errno to EAGAIN.
2529 Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root.
2530 Other bugs in chdir() and Cwd::cwd() have also been fixed. [561]
2534 HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html
2538 The makefiles now provide a single switch to bulk-enable all the
2539 features enabled in ActiveState ActivePerl (a popular Win32 binary
2540 distribution). [561]
2544 Allow REG_EXPAND_SZ keys in the registry.
2548 Can now send() from all threads, not just the first one. [561]
2552 Fake signal handling reenabled, bugs and all.
2556 %SIG has been enabled under USE_ITHREADS, but its use is completely
2557 unsupported under all configurations. [561]
2561 Less stack reserved per thread so that more threads can run
2562 concurrently. (Still 16M per thread.)
2566 C<< File::Spec->tmpdir() >> now prefers C:/temp over /tmp
2567 (works better when perl is running as service).
2571 Better UNC path handling under ithreads. [561]
2575 wait(), waitpid(), and backticks now return the correct exit status
2576 under Windows 9x. [561]
2580 Win64 compilation is now supported.
2584 winsock handle leak fixed. [561]
2590 =head1 New or Changed Diagnostics
2596 The lexical warnings category "deprecated" is no longer a sub-category
2597 of the "syntax" category. It is now a top-level category in its own
2602 All regular expression compilation error messages are now hopefully
2603 easier to understand both because the error message now comes before
2604 the failed regex and because the point of failure is now clearly
2605 marked by a C<E<lt>-- HERE> marker.
2609 The various "opened only for", "on closed", "never opened" warnings
2610 drop the C<main::> prefix for filehandles in the C<main> package,
2611 for example C<STDIN> instead of C<main::STDIN>.
2615 The "Unrecognized escape" warning has been extended to include C<\8>,
2616 C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters.
2620 Two new debugging options have been added: if you have compiled your
2621 Perl with debugging, you can use the -DT [561] and -DR options to trace
2622 tokenising and to add reference counts to displaying variables,
2627 The debugger (perl5db.pl) has been modified to present a more
2628 consistent commands interface, via (CommandSet=580). perl5db.t was
2629 also added to test the changes, and as a placeholder for further tests.
2635 The debugger has a new C<dumpDepth> option to control the maximum
2636 depth to which nested structures are dumped. The C<x> command has
2637 been extended so that C<x N EXPR> dumps out the value of I<EXPR> to a
2638 depth of at most I<N> levels.
2642 The debugger can now show lexical variables if you have the CPAN
2643 module PadWalker installed.
2647 If an attempt to use a (non-blessed) reference as an array index
2648 is made, a warning is given.
2652 C<push @a;> and C<unshift @a;> (with no values to push or unshift)
2653 now give a warning. This may be a problem for generated and evaled
2658 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
2659 using the C<"C"> format you will get an optional warning. Similarly
2660 for the C<"c"> format and a number less than -128 or more than 127.
2664 Certain regex modifiers such as C<(?o)> make sense only if applied to
2665 the entire regex. You will get an optional warning if you try to do
2670 Using arrays or hashes as references (e.g. C<< %foo->{bar} >>
2671 has been deprecated for a while. Now you will get an optional warning.
2675 Using C<sort> in scalar context now issues an optional warning.
2676 This didn't do anything useful, as the sort was not performed.
2680 =head1 Changed Internals
2686 perlapi.pod (a companion to perlguts) now attempts to document the
2691 You can now build a really minimal perl called microperl.
2692 Building microperl does not require even running Configure;
2693 C<make -f Makefile.micro> should be enough. Beware: microperl makes
2694 many assumptions, some of which may be too bold; the resulting
2695 executable may crash or otherwise misbehave in wondrous ways.
2696 For careful hackers only.
2700 Added rsignal(), whichsig(), do_join(), op_clear, op_null,
2701 ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2702 interfaces to the publicised API. For the full list of the available
2703 APIs see L<perlapi>.
2707 Made possible to propagate customised exceptions via croak()ing.
2711 Now xsubs can have attributes just like subs. (Well, at least the
2712 built-in attributes.)
2716 dTHR and djSP have been obsoleted; the former removed (because it's
2717 a no-op) and the latter replaced with dSP.
2721 PERL_OBJECT has been completely removed.
2725 The MAGIC constants (e.g. C<'P'>) have been macrofied
2726 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
2727 and maintainability.
2731 The regex compiler now maintains a structure that identifies nodes in
2732 the compiled bytecode with the corresponding syntactic features of the
2733 original regex expression. The information is attached to the new
2734 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
2735 complete information.
2739 The C code has been made much more C<gcc -Wall> clean. Some warning
2740 messages still remain in some platforms, so if you are compiling with
2741 gcc you may see some warnings about dubious practices. The warnings
2742 are being worked on.
2746 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
2750 Documentation on how to use the Perl source repository has been added
2751 to F<Porting/repository.pod>.
2755 There are now several profiling make targets.
2759 =head1 Security Vulnerability Closed [561]
2761 (This change was already made in 5.7.0 but bears repeating here.)
2763 A potential security vulnerability in the optional suidperl component
2764 of Perl was identified in August 2000. suidperl is neither built nor
2765 installed by default. As of November 2001 the only known vulnerable
2766 platform is Linux, most likely all Linux distributions. CERT and
2767 various vendors and distributors have been alerted about the vulnerability.
2768 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2769 for more information.
2771 The problem was caused by Perl trying to report a suspected security
2772 exploit attempt using an external program, /bin/mail. On Linux
2773 platforms the /bin/mail program had an undocumented feature which
2774 when combined with suidperl gave access to a root shell, resulting in
2775 a serious compromise instead of reporting the exploit attempt. If you
2776 don't have /bin/mail, or if you have 'safe setuid scripts', or if
2777 suidperl is not installed, you are safe.
2779 The exploit attempt reporting feature has been completely removed from
2780 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2781 from all the Perl 5.7 releases), so that particular vulnerability
2782 isn't there anymore. However, further security vulnerabilities are,
2783 unfortunately, always possible. The suidperl functionality is most
2784 probably going to be removed in Perl 5.10. In any case, suidperl
2785 should only be used by security experts who know exactly what they are
2786 doing and why they are using suidperl instead of some other solution
2787 such as sudo ( see http://www.courtesan.com/sudo/ ).
2791 Several new tests have been added, especially for the F<lib> and F<ext>
2792 subsections. There are now about 65 000 individual tests (spread over
2793 about 700 test scripts), in the regression suite (5.6.1 has about
2794 11700 tests, in 258 test scripts) Many of the new tests are of course
2795 introduced by the new modules, but still in general Perl is now more
2798 Because of the large number of tests, running the regression suite
2799 will take considerably longer time than it used to: expect the suite
2800 to take up to 4-5 times longer to run than in perl 5.6. On a really
2801 fast machine you can hope to finish the suite in about 6-8 minutes
2804 The tests are now reported in a different order than in earlier Perls.
2805 (This happens because the test scripts from under t/lib have been moved
2806 to be closer to the library/extension they are testing.)
2808 =head1 Known Problems
2816 If using the AIX native make command, instead of just "make" issue
2817 "make all". In some setups the former has been known to spuriously
2818 also try to run "make install". Alternatively, you may want to use
2823 In AIX 4.2, Perl extensions that use C++ functions that use statics
2824 may have problems in that the statics are not getting initialized.
2825 In newer AIX releases, this has been solved by linking Perl with
2826 the libC_r library, but unfortunately in AIX 4.2 the said library
2827 has an obscure bug where the various functions related to time
2828 (such as time() and gettimeofday()) return broken values, and
2829 therefore in AIX 4.2 Perl is not linked against libC_r.
2833 vac 5.0.0.0 May Produce Buggy Code For Perl
2835 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
2836 resulting in a few random tests failing when run as part of "make
2837 test", but when the failing tests are run by hand, they succeed.
2838 We suggest upgrading to at least vac version 5.0.1.0, that has been
2839 known to compile Perl correctly. "lslpp -L|grep vac.C" will tell
2840 you the vac version. See README.aix.
2844 If building threaded Perl, you may get compilation warning from pp_sys.c:
2846 "pp_sys.c", line 4651.39: 1506-280 (W) Function argument assignment between types "unsigned char*" and "const void*" is not allowed.
2848 This is harmless; it is caused by the getnetbyaddr() and getnetbyaddr_r()
2849 having slightly different types for their first argument.
2853 =head2 Alpha systems with old gccs fail several tests
2855 If you see op/pack, op/pat, op/regexp, or ext/Storable tests failing
2856 in a Linux/alpha or *BSD/Alpha, it's probably time to upgrade your gcc.
2857 gccs prior to 2.95.3 are definitely not good enough, and gcc 3.1 may
2858 be even better. (RedHat Linux/alpha with gcc 3.1 reported no problems,
2859 as did Linux 2.4.18 with gcc 2.95.4.) (In Tru64, it is preferable to
2860 use the bundled C compiler.)
2864 Perl 5.8.0 doesn't build in AmigaOS. It broke at some point during
2865 the ithreads work and we could not find Amiga experts to unbreak the
2866 problems. Perl 5.6.1 still works for AmigaOS (as does the the 5.7.2
2867 development release).
2871 The following tests fail on 5.8.0 Perl in BeOS Personal 5.03:
2873 t/op/lfs............................FAILED at test 17
2874 t/op/magic..........................FAILED at test 24
2875 ext/POSIX/t/sigaction...............FAILED at test 13
2876 ext/POSIX/t/waitpid.................FAILED at test 1
2878 See L<perlbeos> (README.beos) for more details.
2880 =head2 Cygwin "unable to remap"
2882 For example when building the Tk extension for Cygwin,
2883 you may get an error message saying "unable to remap".
2884 This is known problem with Cygwin, and a workaround is
2885 detailed in here: http://sources.redhat.com/ml/cygwin/2001-12/msg00894.html
2887 =head2 ext/threads/t/libc
2889 If this test fails, it indicates that your libc (C library) is not
2890 threadsafe. This particular test stress tests the localtime() call to
2891 find out whether it is threadsafe. See L<perlthrtut> for more information.
2893 =head2 FreeBSD Failing locale Test 117 For ISO8859-15 Locales
2895 The ISO8859-15 locales may fail the locale test 117 in FreeBSD.
2896 This is caused by the characters \xFF (y with diaeresis) and \xBE
2897 (Y with diaeresis) not behaving correctly when being matched
2900 =head2 IRIX fails ext/List/Util/t/shuffle.t
2902 IRIX with MIPSpro 7.3.1.3m compiler may fail the said List::Util test
2903 by dumping core. This seems to be a compiler error since if compiled
2904 with gcc no core dump ensues, and no failures on the said test on any
2907 =head2 Modifying $_ Inside for(..)
2911 works without complaint. It shouldn't. (You should be able to
2912 modify only lvalue elements inside the loops.) You can see the
2913 correct behaviour by replacing the 1..5 with 1, 2, 3, 4, 5.
2915 =head2 mod_perl 1.26 Doesn't Build With Threaded Perl
2917 Use mod_perl 1.27 or higher.
2919 =head2 lib/ftmp-security tests warn 'system possibly insecure'
2921 Don't panic. Read the 'make test' section of INSTALL instead.
2923 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
2925 If perl is configured with -Duse64bitall, the successful result of the
2926 subtest 10 of lib/posix may arrive before the successful result of the
2927 subtest 9, which confuses the test harness so much that it thinks the
2930 =head2 Linux with glibc 2.2.5 fails t/op/int subtest #6 with -Duse64bitint
2932 This is a known bug in the glibc 2.2.5 with long long integers.
2933 ( http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=65612 )
2935 =head2 Linux With Sfio Fails op/misc Test 48
2939 =head2 libwww-perl (LWP) fails base/date #51
2941 Use libwww-perl 5.65 or later.
2945 Please remember to set your environment variable LC_ALL to "C"
2946 (setenv LC_ALL C) before running "make test" to avoid a lot of
2947 warnings about the broken locales of Mac OS X.
2949 The following tests are known to fail in Mac OS X 10.1.5 because of
2950 buggy (old) implementations of Berkeley DB included in Mac OS X:
2952 Failed Test Stat Wstat Total Fail Failed List of Failed
2953 -------------------------------------------------------------------------
2954 ../ext/DB_File/t/db-btree.t 0 11 ?? ?? % ??
2955 ../ext/DB_File/t/db-recno.t 149 3 2.01% 61 63 65
2957 If you are building on a UFS partition, you will also probably see
2958 t/op/stat.t subtest #9 fail. This is caused by Darwin's UFS not
2959 supporting inode change time.
2961 Also the ext/POSIX/t/posix.t subtest #10 fails but it is skipped for
2962 now because the failure is Apple's fault, not Perl's (blocked signals
2965 If you Configure with ithreads, ext/threads/t/libc.t will fail. Again,
2966 this is not Perl's fault-- the libc of Mac OS X is not threadsafe
2967 (in this particular test, the localtime() call is found to be
2970 =head2 op/sprintf tests 91, 129, and 130
2972 The op/sprintf tests 91, 129, and 130 are known to fail on some platforms.
2973 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
2975 Test 91 is known to fail on QNX6 (nto), because C<sprintf '%e',0>
2976 incorrectly produces C<0.000000e+0> instead of C<0.000000e+00>.
2978 For tests 129 and 130, the failing platforms do not comply with
2979 the ANSI C Standard: lines 19ff on page 134 of ANSI X3.159 1989, to
2980 be exact. (They produce something other than "1" and "-1" when
2981 formatting 0.6 and -0.6 using the printf format "%.0f"; most often,
2982 they produce "0" and "-0".)
2986 In case you are still using Solaris 2.5 (aka SunOS 5.5), you may
2987 experience failures (the test core dumping) in lib/locale.t.
2988 The suggested cure is to upgrade your Solaris.
2990 =head2 SUPER-UX (NEC SX)
2992 The following tests are known to fail on SUPER-UX:
2994 op/64bitint...........................FAILED tests 29-30, 32-33, 35-36
2995 op/arith..............................FAILED tests 128-130
2996 op/pack...............................FAILED tests 25-5625
2997 op/pow................................
2998 op/taint..............................# msgsnd failed
2999 ../ext/IO/lib/IO/t/io_poll............FAILED tests 3-4
3000 ../ext/IPC/SysV/ipcsysv...............FAILED tests 2, 5-6
3001 ../ext/IPC/SysV/t/msg.................FAILED tests 2, 4-6
3002 ../ext/Socket/socketpair..............FAILED tests 12
3003 ../lib/IPC/SysV.......................FAILED tests 2, 5-6
3004 ../lib/warnings.......................FAILED tests 115-116, 118-119
3006 The op/pack failure ("Cannot compress negative numbers at op/pack.t line 126")
3007 is serious but as of yet unsolved. It points at some problems with the
3008 signedness handling of the C compiler, as do the 64bitint, arith, and pow
3009 failures. Most of the rest point at problems with SysV IPC.
3011 =head2 Term::ReadKey not working on Win32
3013 Use Term::ReadKey 2.20 or later.
3015 =head2 Failure of Thread (5.005-style) tests
3017 B<Note that support for 5.005-style threading is deprecated,
3018 experimental and practically unsupported. In 5.10, it is expected
3021 The following tests are known to fail due to fundamental problems in
3022 the 5.005 threading implementation. These are not new failures--Perl
3023 5.005_0x has the same bugs, but didn't have these tests.
3025 ../ext/B/t/xref.t 255 65280 14 12 85.71% 3-14
3026 ../ext/List/Util/t/first.t 255 65280 7 4 57.14% 2 5-7
3027 ../lib/English.t 2 512 54 2 3.70% 2-3
3028 ../lib/ExtUtils/t/basic.t 1 256 17 1 5.88% 14
3029 ../lib/FileCache.t 5 1 20.00% 5
3030 ../lib/Filter/Simple/t/data.t 6 3 50.00% 1-3
3031 ../lib/Filter/Simple/t/filter_onl 9 3 33.33% 1-2 5
3032 ../lib/Tie/File/t/31_autodefer.t 255 65280 65 32 49.23% 34-65
3033 ../lib/autouse.t 10 1 10.00% 4
3034 op/flip.t 15 1 6.67% 15
3036 These failures are unlikely to get fixed as 5.005-style threads
3037 are considered fundamentally broken. (Basically what happens is that
3038 competing threads can corrupt shared global state.)
3040 =head2 Timing problems
3042 The following tests may fail intermittently because of timing
3043 problems, for example if the system is heavily loaded.
3046 ext/Time/HiRes/HiRes.t
3048 lib/Memoize/t/expmod_t.t
3049 lib/Memoize/t/speed.t
3051 In case of failure please try running them manually, for example
3053 ./perl -Ilib ext/Time/HiRes/HiRes.t
3057 ../lib/Math/Trig.t 26 1 3.85% 25
3058 ../lib/warnings.t 470 1 0.21% 429
3060 The Trig.t failure is caused by the slighly differing (from IEEE)
3061 floating point implementation of UNICOS. The warnings.t failure is
3062 also related: the test assumes a certain floating point output format;
3063 this assumption fails in UNICOS.
3067 =head2 VOS (Stratus)
3069 When Perl is built using the native build process on VOS Release
3070 14.5.0 and GNU C++/GNU Tools 2.0.1, all attempted tests either
3071 pass or result in TODO (ignored) failures.
3077 During Configure, the test
3079 Guessing which symbols your C compiler and preprocessor define...
3081 will probably fail with error messages like
3083 CC-20 cc: ERROR File = try.c, Line = 3
3084 The identifier "bad" is undefined.
3086 bad switch yylook 79bad switch yylook 79bad switch yylook 79bad switch yylook 79#ifdef A29K
3089 CC-65 cc: ERROR File = try.c, Line = 3
3090 A semicolon is expected at this point.
3092 This is caused by a bug in the awk utility of UNICOS/mk. You can ignore
3093 the error, but it does cause a slight problem: you cannot fully
3094 benefit from the h2ph utility (see L<h2ph>) that can be used to
3095 convert C headers to Perl libraries, mainly used to be able to access
3096 from Perl the constants defined using C preprocessor, cpp. Because of
3097 the above error, parts of the converted headers will be invisible.
3098 Luckily, these days the need for h2ph is rare.
3102 If building Perl with interpreter threads (ithreads), the
3103 getgrent(), getgrnam(), and getgrgid() functions cannot return the
3104 list of the group members due to a bug in the multithreaded support of
3105 UNICOS/mk. What this means is that in list context the functions will
3106 return only three values, not four.
3112 There are a few known test failures, see L<perluts> (README.uts).
3116 There should be no reported test failures with a default configuration,
3117 though there are a number of tests marked TODO that point to areas
3118 needing further debugging and/or porting work.
3122 In multi-CPU boxes, there are some problems with the I/O buffering:
3123 some output may appear twice.
3125 =head2 XML::Parser not working
3127 Use XML::Parser 2.31 or later.
3129 =head2 z/OS (OS/390)
3131 z/OS has rather many test failures but the situation is actually
3132 better than it was in 5.6.0; it's just that so many new modules and
3133 tests have been added.
3135 Failed Test Stat Wstat Total Fail Failed List of Failed
3136 ---------------------------------------------------------------------------
3137 ../ext/Data/Dumper/t/dumper.t 357 8 2.24% 311 314 325 327
3139 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5
3140 ../ext/Storable/t/downgrade.t 12 3072 169 12 7.10% 14-15 46-47 78-79
3142 ../lib/ExtUtils/t/Constant.t 121 30976 48 48 100.00% 1-48
3143 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9
3144 op/pat.t 910 7 0.77% 665 776 785 832-
3146 op/sprintf.t 224 3 1.34% 98 100 136
3147 op/tr.t 97 5 5.15% 63 71-74
3148 uni/fold.t 780 6 0.77% 61 169 196 661
3151 The failures in dumper.t and downgrade.t are problems in the tests,
3152 those in io_unix and sprintf are problems in the USS (UDP sockets
3153 and printf formats). The pat, tr, and fold failures are genuine Perl
3154 problems caused by EBCDIC (and in the pat and fold cases, combining
3155 that with Unicode). The Constant and Embed are probably problems
3156 in the tests (since they test Perl's ability to build extensions,
3157 and that seems to be working reasonably well.)
3159 =head2 Localising Tied Arrays and Hashes Is Broken
3163 doesn't work as one would expect: the old value is restored
3164 incorrectly. This will be changed in a future release, but we don't
3165 know yet what the new semantics will exactly be. In any case, the
3166 change will break existing code that relies on the current
3167 (ill-defined) semantics, so just avoid doing this in general.
3169 =head2 Self-tying Problems
3171 Self-tying of arrays and hashes is broken in rather deep and
3172 hard-to-fix ways. As a stop-gap measure to avoid people from getting
3173 frustrated at the mysterious results (core dumps, most often), it is
3174 forbidden for now (you will get a fatal error even from an attempt).
3176 A change to self-tying of globs has caused them to be recursively
3177 referenced (see: L<perlobj/"Two-Phased Garbage Collection">). You
3178 will now need an explicit untie to destroy a self-tied glob. This
3179 behaviour may be fixed at a later date.
3181 Self-tying of scalars and IO thingies works.
3183 =head2 Building Extensions Can Fail Because Of Largefiles
3185 Some extensions like mod_perl are known to have issues with
3186 `largefiles', a change brought by Perl 5.6.0 in which file offsets
3187 default to 64 bits wide, where supported. Modules may fail to compile
3188 at all, or they may compile and work incorrectly. Currently, there
3189 is no good solution for the problem, but Configure now provides
3190 appropriate non-largefile ccflags, ldflags, libswanted, and libs
3191 in the %Config hash (e.g., $Config{ccflags_nolargefiles}) so the
3192 extensions that are having problems can try configuring themselves
3193 without the largefileness. This is admittedly not a clean solution,
3194 and the solution may not even work at all. One potential failure is
3195 whether one can (or, if one can, whether it's a good idea to) link
3196 together at all binaries with different ideas about file offsets;
3197 all this is platform-dependent.
3199 =head2 Unicode Support on EBCDIC Still Spotty
3201 Though mostly working, Unicode support still has problem spots on
3202 EBCDIC platforms. One such known spot are the C<\p{}> and C<\P{}>
3203 regular expression constructs for code points less than 256: the
3204 C<pP> are testing for Unicode code points, not knowing about EBCDIC.
3206 =head2 The Compiler Suite Is Still Very Experimental
3208 The compiler suite is slowly getting better but it continues to be
3209 highly experimental. Use in production environments is discouraged.
3211 =head2 The Long Double Support Is Still Experimental
3213 The ability to configure Perl's numbers to use "long doubles",
3214 floating point numbers of hopefully better accuracy, is still
3215 experimental. The implementations of long doubles are not yet
3216 widespread and the existing implementations are not quite mature
3217 or standardised, therefore trying to support them is a rare
3218 and moving target. The gain of more precision may also be offset
3219 by slowdown in computations (more bits to move around, and the
3220 operations are more likely to be executed by less optimised
3223 =head2 Seen In Perl 5.7 But Gone Now
3225 C<Time::Piece> (previously known as C<Time::Object>) was removed
3226 because it was felt that it didn't have enough value in it to be a
3227 core module. It is still a useful module, though, and is available
3230 Perl 5.8 unfortunately does not build anymore on AmigaOS; this broke
3231 accidentally at some point. Since there are not that many Amiga
3232 developers available, we could not get this fixed and tested in time
3233 for 5.8.0. Perl 5.6.1 still works for AmigaOS (as does the the 5.7.2
3234 development release).
3236 =head1 Reporting Bugs
3238 If you find what you think is a bug, you might check the articles
3239 recently posted to the comp.lang.perl.misc newsgroup and the perl
3240 bug database at http://bugs.perl.org/ . There may also be
3241 information at http://www.perl.com/ , the Perl Home Page.
3243 If you believe you have an unreported bug, please run the B<perlbug>
3244 program included with your release. Be sure to trim your bug down
3245 to a tiny but sufficient test case. Your bug report, along with the
3246 output of C<perl -V>, will be sent off to perlbug@perl.org to be
3247 analysed by the Perl porting team.
3251 The F<Changes> file for exhaustive details on what changed.
3253 The F<INSTALL> file for how to build Perl.
3255 The F<README> file for general stuff.
3257 The F<Artistic> and F<Copying> files for copyright information.
3261 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>.