3 perldelta - what is new for perl v5.8.0
7 This document describes differences between the 5.6.0 release and the
8 5.8.0 release. Changes that have been integrated into the 5.6.1
9 release are marked [561]. Changes that also appear in 5.6.1 but have
10 been improved since are marked [561+].
12 Many of the bug fixes in 5.8.0 were already seen in the 5.6.1
13 maintenance release since the two releases were kept closely
16 If you are upgrading from Perl 5.005_03, you might also want
17 to read L<perl56delta>.
19 =head1 Highlights In 5.8.0
25 Better Unicode support
29 New Thread Implementation
37 Better Numeric Accuracy
45 More Extensive Regression Testing
49 =head1 Incompatible Changes
51 =head2 Binary Incompatibility
53 B<Perl 5.8 is not binary compatible with earlier releases of Perl.>
55 B<You have to recompile your XS modules.>
57 (Pure Perl modules should continue to work.)
59 The major reason for the discontinuity is the new IO architecture
60 called PerlIO. PerlIO is the default configuration because without
61 it many new features of Perl 5.8 cannot be used. In other words:
62 you just have to recompile your modules containing XS code, sorry
65 In future releases of Perl, non-PerlIO aware XS modules may become
66 completely unsupported. This shouldn't be too difficult for module
67 authors, however: PerlIO has been designed as a drop-in replacement
68 (at the source code level) for the stdio interface.
70 Depending on your platform, there are also other reasons why
71 we decided to break binary compatibility, please read on.
73 =head2 64-bit platforms and malloc
75 If your pointers are 64 bits wide, the Perl malloc is no longer being
76 used because it does not work well with 8-byte pointers. Also,
77 usually the system mallocs on such platforms are much better optimized
78 for such large memory models than the Perl malloc. Some memory-hungry
79 Perl applications like the PDL don't work well with Perl's malloc.
80 Finally, other applications than Perl (such as mod_perl) tend to prefer
81 the system malloc. Such platforms include Alpha and 64-bit HPPA,
84 =head2 AIX Dynaloading
86 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
87 dlopen interface of AIX instead of the old emulated interface. This
88 change will probably break backward compatibility with compiled
89 modules. The change was made to make Perl more compliant with other
90 applications like mod_perl which are using the AIX native interface.
92 =head2 Attributes for C<my> variables now handled at run-time.
94 The C<my EXPR : ATTRS> syntax now applies variable attributes at
95 run-time. (Subroutine and C<our> variables still get attributes applied
96 at compile-time.) See L<attributes> for additional details. In particular,
97 however, this allows variable attributes to be useful for C<tie> interfaces,
98 which was a deficiency of earlier releases. Note that the new semantics
99 doesn't work with the Attribute::Handlers module (as of version 0.76).
101 =head2 Socket Extension Dynamic in VMS
103 The Socket extension is now dynamically loaded instead of being
104 statically built in. This may or may not be a problem with ancient
105 TCP/IP stacks of VMS: we do not know since we weren't able to test
106 Perl in such configurations.
108 =head2 IEEE-format Floating Point Default on OpenVMS Alpha
110 Perl now uses IEEE format (T_FLOAT) as the default internal floating
111 point format on OpenVMS Alpha, potentially breaking binary compatibility
112 with external libraries or existing data. G_FLOAT is still available as
113 a configuration option. The default on VAX (D_FLOAT) has not changed.
115 =head2 New Unicode Properties
117 Unicode I<scripts> are now supported. Scripts are similar to (and superior
118 to) Unicode I<blocks>. The difference between scripts and blocks is that
119 scripts are the glyphs used by a language or a group of languages, while
120 the blocks are more artificial groupings of (mostly) 256 characters based
121 on the Unicode numbering.
123 In general, scripts are more inclusive, but not universally so. For
124 example, while the script C<Latin> includes all the Latin characters and
125 their various diacritic-adorned versions, it does not include the various
126 punctuation or digits (since they are not solely C<Latin>).
128 A number of other properties are now supported, including C<\p{L&}>,
129 C<\p{Any}> C<\p{Assigned}>, C<\p{Unassigned}>, C<\p{Blank}> [561] and
130 C<\p{SpacePerl}> [561] (along with their C<\P{...}> versions, of course).
131 See L<perlunicode> for details, and more additions.
133 The C<In> or C<Is> prefix to names used with the C<\p{...}> and C<\P{...}>
134 are now almost always optional. The only exception is that a C<In> prefix
135 is required to signify a Unicode block when a block name conflicts with a
136 script name. For example, C<\p{Tibetan}> refers to the script, while
137 C<\p{InTibetan}> refers to the block. When there is no name conflict, you
138 can omit the C<In> from the block name (e.g. C<\p{BraillePatterns}>), but
139 to be safe, it's probably best to always use the C<In>).
141 =head2 REF(...) Instead Of SCALAR(...)
143 A reference to a reference now stringifies as "REF(0x81485ec)" instead
144 of "SCALAR(0x81485ec)" in order to be more consistent with the return
147 =head2 pack/unpack D/F recycled
149 The undocumented pack/unpack template letters D/F have been recycled
150 for better use: now they stand for long double (if supported by the
151 platform) and NV (Perl internal floating point type). (They used
152 to be aliases for d/f, but you never knew that.)
160 The semantics of bless(REF, REF) were unclear and until someone proves
161 it to make some sense, it is forbidden.
165 The obsolete chat2 library that should never have been allowed
166 to escape the laboratory has been decommissioned.
170 The builtin dump() function has probably outlived most of its
171 usefulness. The core-dumping functionality will remain in future
172 available as an explicit call to C<CORE::dump()>, but in future
173 releases the behaviour of an unqualified C<dump()> call may change.
177 The very dusty examples in the eg/ directory have been removed.
178 Suggestions for new shiny examples welcome but the main issue is that
179 the examples need to be documented, tested and (most importantly)
184 The (bogus) escape sequences \8 and \9 now give an optional warning
185 ("Unrecognized escape passed through"). There is no need to \-escape
190 The list of filenames from glob() (or <...>) is now by default sorted
191 alphabetically to be csh-compliant (which is what happened before
192 in most UNIX platforms). (bsd_glob() does still sort platform
193 natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.) [561]
197 Spurious syntax errors generated in certain situations, when glob()
198 caused File::Glob to be loaded for the first time, have been fixed. [561]
202 Although "you shouldn't do that", it was possible to write code that
203 depends on Perl's hashed key order (Data::Dumper does this). The new
204 algorithm "One-at-a-Time" produces a different hashed key order.
205 More details are in L</"Performance Enhancements">.
209 lstat(FILEHANDLE) now gives a warning because the operation makes no sense.
210 In future releases this may become a fatal error.
214 The C<package;> syntax (C<package> without an argument) has been
215 deprecated. Its semantics were never that clear and its
216 implementation even less so. If you have used that feature to
217 disallow all but fully qualified variables, C<use strict;> instead.
221 The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still
222 recognised but now cause fatal errors. The previous behaviour of
223 ignoring them by default and warning if requested was unacceptable
224 since it, in a way, falsely promised that the features could be used.
228 The current user-visible implementation of pseudo-hashes (the weird
229 use of the first array element) is deprecated starting from Perl 5.8.0
230 and will be removed in Perl 5.10.0, and the feature will be
231 implemented differently. Not only is the current interface rather
232 ugly, but the current implementation slows down normal array and hash
233 use quite noticeably. The C<fields> pragma interface will remain
234 available. The I<restricted hashes> interface is expected to
235 be the replacement interface (see L<Hash::Util>).
239 The syntaxes C<< @a->[...] >> and C<< %h->{...} >> have now been deprecated.
243 After years of trying, suidperl is considered to be too complex to
244 ever be considered truly secure. The suidperl functionality is likely
245 to be removed in a future release.
249 The 5.005 threads model (module C<Thread>) is deprecated and expected
250 to be removed in Perl 5.10. Multithreaded code should be migrated to
251 the new ithreads model (see L<threads>, L<threads::shared> and
256 The long deprecated uppercase aliases for the string comparison
257 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
261 The tr///C and tr///U features have been removed and will not return;
262 the interface was a mistake. Sorry about that. For similar
263 functionality, see pack('U0', ...) and pack('C0', ...). [561]
267 Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo (@)".
268 The prototypes are now checked better at compile-time for invalid
269 syntax. An optional warning is generated ("Illegal character in
270 prototype...") but this may be upgraded to a fatal error in a future
275 The existing behaviour when localising tied arrays and hashes is wrong,
276 and will be changed in a future release, so do not rely on the existing
277 behaviour. See L<"Localising Tied Arrays and Hashes Is Broken">.
281 =head1 Core Enhancements
283 =head2 PerlIO is Now The Default
289 IO is now by default done via PerlIO rather than system's "stdio".
290 PerlIO allows "layers" to be "pushed" onto a file handle to alter the
291 handle's behaviour. Layers can be specified at open time via 3-arg
294 open($fh,'>:crlf :utf8', $path) || ...
296 or on already opened handles via extended C<binmode>:
298 binmode($fh,':encoding(iso-8859-7)');
300 The built-in layers are: unix (low level read/write), stdio (as in
301 previous Perls), perlio (re-implementation of stdio buffering in a
302 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
303 but available on any platform). A mmap layer may be available if
304 platform supports it (mostly UNIXes).
306 Layers to be applied by default may be specified via the 'open' pragma.
308 See L</"Installation and Configuration Improvements"> for the effects
309 of PerlIO on your architecture name.
313 File handles can be marked as accepting Perl's internal encoding of Unicode
314 (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :
316 open($fh,">:utf8","Uni.txt");
318 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named
319 for you since it's not UTF-8 what you will be getting but instead
320 UTF-EBCDIC. See L<perlunicode>, L<utf8>, and
321 http://www.unicode.org/unicode/reports/tr16/ for more information.
322 In future releases this naming may change.
326 File handles can translate character encodings from/to Perl's internal
327 Unicode form on read/write via the ":encoding()" layer.
331 File handles can be opened to "in memory" files held in Perl scalars via:
333 open($fh,'>', \$variable) || ...
337 Anonymous temporary files are available without need to
338 'use FileHandle' or other module via
340 open($fh,"+>", undef) || ...
342 That is a literal undef, not an undefined value.
346 The list form of C<open> is now implemented for pipes (at least on UNIX):
348 open($fh,"-|", 'cat', '/etc/motd')
350 creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in
355 If your locale environment variables (LANGUAGE, LC_ALL, LC_CTYPE, LANG)
356 contain the strings 'UTF-8' or 'UTF8' (case-insensitive matching),
357 the default encoding of your STDIN, STDOUT, and STDERR, and of
358 B<any subsequent file open>, is UTF-8.
362 =head2 Restricted Hashes
364 A restricted hash is restricted to a certain set of keys, no keys
365 outside the set can be added. Also individual keys can be restricted
366 so that the key cannot be deleted and the value cannot be changed.
367 No new syntax is involved: the Hash::Util module is the interface.
371 Perl used to be fragile in that signals arriving at inopportune moments
372 could corrupt Perl's internal state. Now Perl postpones handling of
373 signals until it's safe (between opcodes).
375 This change may have surprising side effects because signals no longer
376 interrupt Perl instantly. Perl will now first finish whatever it was
377 doing, like finishing an internal operation (like sort()) or an
378 external operation (like an I/O operation), and only then look at any
379 arrived signals (and before starting the next operation). No more corrupt
380 internal state since the current operation is always finished first,
381 but the signal may take more time to get heard. Note that breaking
382 out from potentially blocking operations should still work, though.
384 =head2 Unicode Overhaul
386 Unicode in general should be now much more usable than in Perl 5.6.0
387 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in
388 regular expressions should work now, Unicode in tr/// should work now,
389 Unicode in I/O should work now. See L<perluniintro> for introduction
390 and L<perlunicode> for details.
396 The Unicode Character Database coming with Perl has been upgraded
397 to Unicode 3.2.0. For more information, see http://www.unicode.org/ .
402 For developers interested in enhancing Perl's Unicode capabilities:
403 almost all the UCD files are included with the Perl distribution in
404 the F<lib/unicore> subdirectory. The most notable omission, for space
405 considerations, is the Unihan database.
409 The properties \p{Blank} and \p{SpacePerl} have been added. "Blank" is like
410 C isblank(), that is, it contains only "horizontal whitespace" (the space
411 character is, the newline isn't), and the "SpacePerl" is the Unicode
412 equivalent of C<\s> (\p{Space} isn't, since that includes the vertical
413 tabulator character, whereas C<\s> doesn't.)
415 See "New Unicode Properties" earlier in this document for additional
416 information on changes with Unicode properties.
420 =head2 Understanding of Numbers
422 In general a lot of fixing has happened in the area of Perl's
423 understanding of numbers, both integer and floating point. Since in
424 many systems the standard number parsing functions like C<strtoul()>
425 and C<atof()> seem to have bugs, Perl tries to work around their
426 deficiencies. This results hopefully in more accurate numbers.
428 Perl now tries internally to use integer values in numeric conversions
429 and basic arithmetics (+ - * /) if the arguments are integers, and
430 tries also to keep the results stored internally as integers.
431 This change leads to often slightly faster and always less lossy
432 arithmetics. (Previously Perl always preferred floating point numbers
435 =head2 Miscellaneous Changes
441 AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute
442 to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.
446 The $Config{byteorder} (and corresponding BYTEORDER in config.h) was
447 previously wrong in platforms if sizeof(long) was 4, but sizeof(IV)
448 was 8. The byteorder was only sizeof(long) bytes long (1234 or 4321),
449 but now it is correctly sizeof(IV) bytes long, (12345678 or 87654321).
450 (This problem didn't affect Windows platforms.)
452 Also, $Config{byteorder} is now computed dynamically--this is more
453 robust with "fat binaries" where an executable image contains binaries
454 for more than one binary platform, and when cross-compiling.
458 C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass
459 in multiple arguments.)
463 The builtin dump() now gives an optional warning
464 C<dump() better written as CORE::dump()>,
465 meaning that by default C<dump(...)> is resolved as the builtin
466 dump() which dumps core and aborts, not as (possibly) user-defined
467 C<sub dump>. To call the latter, qualify the call as C<&dump(...)>.
468 (The whole dump() feature is to considered deprecated, and possibly
469 removed/changed in future releases.)
473 chomp() and chop() are now overridable. Note, however, that their
474 prototype (as given by C<prototype("CORE::chomp")> is undefined,
475 because it cannot be expressed and therefore one cannot really write
476 replacements to override these builtins.
480 END blocks are now run even if you exit/die in a BEGIN block.
481 Internally, the execution of END blocks is now controlled by
482 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new
483 behaviour for Perl embedders. This will default in 5.10. See
488 Formats now support zero-padded decimal fields.
492 Lvalue subroutines can now return C<undef> in list context. However,
493 the lvalue subroutine feature still remains experimental. [561+]
497 A lost warning "Can't declare ... dereference in my" has been
498 restored (Perl had it earlier but it became lost in later releases.)
502 A new special regular expression variable has been introduced:
503 C<$^N>, which contains the most-recently closed group (submatch).
507 C<no Module;> does not produce an error even if Module does not have an
508 unimport() method. This parallels the behavior of C<use> vis-a-vis
513 The numerical comparison operators return C<undef> if either operand
514 is a NaN. Previously the behaviour was unspecified.
518 The following builtin functions are now overridable: each(), keys(),
519 pop(), push(), shift(), splice(), unshift(). [561]
523 C<pack() / unpack()> can now group template letters with C<()> and then
524 apply repetition/count modifiers on the groups.
528 C<pack() / unpack()> can now process the Perl internal numeric types:
529 IVs, UVs, NVs-- and also long doubles, if supported by the platform.
530 The template letters are C<j>, C<J>, C<F>, and C<D>.
534 C<pack('U0a*', ...)> can now be used to force a string to UTF8.
538 my __PACKAGE__ $obj now works. [561]
542 POSIX::sleep() now returns the number of I<unslept> seconds
543 (as the POSIX standard says), as opposed to CORE::sleep() which
544 returns the number of slept seconds.
548 The printf() and sprintf() now support parameter reordering using the
549 C<%\d+\$> and C<*\d+\$> syntaxes. For example
551 print "%2\$s %1\$s\n", "foo", "bar";
553 will print "bar foo\n". This feature helps in writing
554 internationalised software, and in general when the order
555 of the parameters can vary.
559 prototype(\&) is now available.
563 prototype(\[$@%&]) is now available to implicitly create references
564 (useful for example if you want to emulate the tie() interface).
568 A new command-line option, C<-t> is available. It is the
569 little brother of C<-T>: instead of dying on taint violations,
570 lexical warnings are given. B<This is only meant as a temporary
571 debugging aid while securing the code of old legacy applications.
572 This is not a substitute for -T.>
576 In other taint news, the C<exec LIST> and C<system LIST> have now been
577 considered too risky (think C<exec @ARGV>: it can start any program
578 with any arguments), and now the said forms cause a warning.
579 You should carefully launder the arguments to guarantee their
580 validity. In future releases of Perl the forms will become fatal
581 errors so consider starting laundering now.
585 Tied hash interfaces are now required to have the EXISTS and DELETE
586 methods (either own or inherited).
590 If tr/// is just counting characters, it doesn't attempt to
595 untie() will now call an UNTIE() hook if it exists. See L<perltie>
600 L<utime> now supports C<utime undef, undef, @files> to change the
601 file timestamps to the current time.
605 The rules for allowing underscores (underbars) in numeric constants
606 have been relaxed and simplified: now you can have an underscore
607 simply B<between digits>.
611 Rather than relying on C's argv[0] (which may not contain a full pathname)
612 where possible $^X is now set by asking the operating system.
613 (eg by reading F</proc/self/exe> on Linux, F</proc/curproc/file> on FreeBSD)
617 A new variable, C<${^TAINT}>, indicates whether taint mode is enabled.
621 You can now override the readline() builtin, and this overrides also
622 the <FILEHANDLE> angle bracket operator.
626 The command-line options -s and -F are now recognized on the shebang
631 Use of the C</c> match modifier without an accompanying C</g> modifier
632 elicits a new warning: C<Use of /c modifier is meaningless without /g>.
634 Use of C</c> in substitutions, even with C</g>, elicits
635 C<Use of /c modifier is meaningless in s///>.
637 Use of C</g> with C<split> elicits C<Use of /g modifier is meaningless
642 =head1 Modules and Pragmata
644 =head2 New Modules and Pragmata
650 C<Attribute::Handlers> allows a class to define attribute handlers.
653 use Attribute::Handlers;
654 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
656 # later, in some package using or inheriting from MyPack...
658 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
660 Both variables and routines can have attribute handlers. Handlers can
661 be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the
662 exact compilation phase (BEGIN, CHECK, INIT, or END).
663 See L<Attribute::Handlers>.
667 C<B::Concise>, by Stephen McCamant, is a new compiler backend for
668 walking the Perl syntax tree, printing concise info about ops.
669 The output is highly customisable. See L<B::Concise>. [561+]
673 The new bignum, bigint, and bigrat pragmas, by Tels, implement
674 transparent bignum support (using the Math::BigInt, Math::BigFloat,
675 and Math::BigRat backends).
679 C<Class::ISA>, by Sean Burke, is a module for reporting the search
680 path for a class's ISA tree. See L<Class::ISA>.
684 C<Cwd> now has a split personality: if possible, an XS extension is
685 used, (this will hopefully be faster, more secure, and more robust)
686 but if not possible, the familiar Perl implementation is used.
690 C<Devel::PPPort>, originally by Kenneth Albanowski and now
691 maintained by Paul Marquess, has been added. It is primarily used
692 by C<h2xs> to enhance portability of XS modules between different
693 versions of Perl. See L<Devel::PPPort>.
697 C<Digest>, frontend module for calculating digests (checksums), from
698 Gisle Aas, has been added. See L<Digest>.
702 C<Digest::MD5> for calculating MD5 digests (checksums) as defined in
703 RFC 1321, from Gisle Aas, has been added. See L<Digest::MD5>.
705 use Digest::MD5 'md5_hex';
707 $digest = md5_hex("Thirsty Camel");
709 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
711 NOTE: the C<MD5> backward compatibility module is deliberately not
712 included since its further use is discouraged.
716 C<Encode>, originally by Nick Ing-Simmons and now maintained by Dan
717 Kogai, provides a mechanism to translate between different character
718 encodings. Support for Unicode, ISO-8859-1, and ASCII are compiled in
719 to the module. Several other encodings (like the rest of the
720 ISO-8859, CP*/Win*, Mac, KOI8-R, three variants EBCDIC, Chinese,
721 Japanese, and Korean encodings) are included and can be loaded at
722 runtime. (For space considerations, the largest Chinese encodings
723 have been separated into their own CPAN module, Encode::HanExtra,
724 which Encode will use if available). See L<Encode>.
726 Any encoding supported by Encode module is also available to the
727 ":encoding()" layer if PerlIO is used.
731 C<Hash::Util> is the interface to the new I<restricted hashes>
732 feature. (Implemented by Jeffrey Friedl, Nick Ing-Simmons, and
733 Michael Schwern.) See L<Hash::Util>.
737 C<I18N::Langinfo> can be used to query locale information.
738 See L<I18N::Langinfo>.
742 C<I18N::LangTags>, by Sean Burke, has functions for dealing with
743 RFC3066-style language tags. See L<I18N::LangTags>.
747 C<ExtUtils::Constant>, by Nicholas Clark, is a new tool for extension
748 writers for generating XS code to import C header constants.
749 See L<ExtUtils::Constant>.
753 C<Filter::Simple>, by Damian Conway, is an easy-to-use frontend to
754 Filter::Util::Call. See L<Filter::Simple>.
760 use Filter::Simple sub {
761 while (my ($from, $to) = splice @_, 0, 2) {
770 use MyFilter qr/red/ => 'green';
772 print "red\n"; # this code is filtered, will print "green\n"
773 print "bored\n"; # this code is filtered, will print "bogreen\n"
777 print "red\n"; # this code is not filtered, will print "red\n"
781 C<File::Temp>, by Tim Jenness, allows one to create temporary files
782 and directories in an easy, portable, and secure way. See
783 L<File::Temp>. [561+]
787 C<Filter::Util::Call>, by Paul Marquess, provides you with the
788 framework to write I<source filters> in Perl. For most uses, the
789 frontend Filter::Simple is to be preferred. See L<Filter::Util::Call>.
793 C<if>, by Ilya Zakharevich, is a new pragma for conditional inclusion
798 L<libnet>, by Graham Barr, is a collection of perl5 modules related
799 to network programming. See L<Net::FTP>, L<Net::NNTP>, L<Net::Ping>
800 (not part of libnet, but related), L<Net::POP3>, L<Net::SMTP>,
803 Perl installation leaves libnet unconfigured; use F<libnetcfg>
808 C<List::Util>, by Graham Barr, is a selection of general-utility
809 list subroutines, such as sum(), min(), first(), and shuffle().
814 C<Locale::Constants>, C<Locale::Country>, C<Locale::Currency>
815 C<Locale::Language>, and L<Locale::Script>, by Neil Bowers, have
816 been added. They provide the codes for various locale standards, such
817 as "fr" for France, "usd" for US Dollar, and "ja" for Japanese.
821 $country = code2country('jp'); # $country gets 'Japan'
822 $code = country2code('Norway'); # $code gets 'no'
824 See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>,
825 and L<Locale::Language>.
829 C<Locale::Maketext>, by Sean Burke, is a localization framework. See
830 L<Locale::Maketext>, and L<Locale::Maketext::TPJ13>. The latter is an
831 article about software localization, originally published in The Perl
832 Journal #13, and republished here with kind permission.
836 C<Math::BigRat> for big rational numbers, to accompany Math::BigInt and
837 Math::BigFloat, from Tels. See L<Math::BigRat>.
841 C<Memoize> can make your functions faster by trading space for time,
842 from Mark-Jason Dominus. See L<Memoize>.
846 C<MIME::Base64>, by Gisle Aas, allows you to encode data in base64,
847 as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
852 $encoded = encode_base64('Aladdin:open sesame');
853 $decoded = decode_base64($encoded);
855 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
861 C<MIME::QuotedPrint>, by Gisle Aas, allows you to encode data
862 in quoted-printable encoding, as defined in RFC 2045 - I<MIME
863 (Multipurpose Internet Mail Extensions)>.
865 use MIME::QuotedPrint;
867 $encoded = encode_qp("Smiley in Unicode: \x{263a}");
868 $decoded = decode_qp($encoded);
870 print $encoded, "\n"; # "Smiley in Unicode: =263A"
872 MIME::QuotedPrint has been enhanced to provide the basic methods
873 necessary to use it with PerlIO::Via as in :
875 use MIME::QuotedPrint;
876 open($fh,">Via(MIME::QuotedPrint)",$path);
878 See L<MIME::QuotedPrint>.
882 C<NEXT>, by Damian Conway, is a pseudo-class for method redispatch.
887 C<open> is a new pragma for setting the default I/O disciplines
892 C<PerlIO::Scalar>, by Nick Ing-Simmons, provides the implementation
893 of IO to "in memory" Perl scalars as discussed above. It also serves
894 as an example of a loadable PerlIO layer. Other future possibilities
895 include PerlIO::Array and PerlIO::Code. See L<PerlIO::Scalar>.
899 C<PerlIO::Via>, by Nick Ing-Simmons, acts as a PerlIO layer and wraps
900 PerlIO layer functionality provided by a class (typically implemented
903 use MIME::QuotedPrint;
904 open($fh,">Via(MIME::QuotedPrint)",$path);
906 This will automatically convert everything output to C<$fh>
907 to Quoted-Printable. See L<PerlIO::Via>.
911 C<Pod::ParseLink>, by Russ Allbery, has been added,
912 to parse LZ<><> links in pods as described in the new
917 C<Pod::Text::Overstrike>, by Joe Smith, has been added.
918 It converts POD data to formatted overstrike text.
919 See L<Pod::Text::Overstrike>. [561+]
923 C<Scalar::Util> is a selection of general-utility scalar subroutines,
924 such as blessed(), reftype(), and tainted(). See L<Scalar::Util>.
928 C<sort> is a new pragma for controlling the behaviour of sort().
932 C<Storable> gives persistence to Perl data structures by allowing the
933 storage and retrieval of Perl data to and from files in a fast and
934 compact binary format. Because in effect Storable does serialisation
935 of Perl data structues, with it you can also clone deep, hierarchical
936 datastructures. Storable was originally created by Raphael Manfredi,
937 but it is now maintained by Abhijit Menon-Sen. Storable has been
938 enhanced to understand the two new hash features, Unicode keys and
939 restricted hashes. See L<Storable>.
943 C<Switch>, by Damian Conway, has been added. Just by saying
947 you have C<switch> and C<case> available in Perl.
953 case 1 { print "number 1" }
954 case "a" { print "string a" }
955 case [1..10,42] { print "number in list" }
956 case (@array) { print "number in list" }
957 case /\w+/ { print "pattern" }
958 case qr/\w+/ { print "pattern" }
959 case (%hash) { print "entry in hash" }
960 case (\%hash) { print "entry in hash" }
961 case (\&sub) { print "arg to subroutine" }
962 else { print "previous case not true" }
969 C<Test::More>, by Michael Schwern, is yet another framework for writing
970 test scripts, more extensive than Test::Simple. See L<Test::More>.
974 C<Test::Simple>, by Michael Schwern, has basic utilities for writing
975 tests. See L<Test::Simple>.
979 C<Text::Balanced>, by Damian Conway, has been added, for extracting
980 delimited text sequences from strings.
982 use Text::Balanced 'extract_delimited';
984 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
986 $a will be "'never say never'", $b will be ', he never said'.
988 In addition to extract_delimited(), there are also extract_bracketed(),
989 extract_quotelike(), extract_codeblock(), extract_variable(),
990 extract_tagged(), extract_multiple(), gen_delimited_pat(), and
991 gen_extract_tagged(). With these, you can implement rather advanced
992 parsing algorithms. See L<Text::Balanced>.
996 C<threads>, by Arthur Bergman, is an interface to interpreter threads.
997 Interpreter threads (ithreads) is the new thread model introduced in
998 Perl 5.6 but only available as an internal interface for extension
999 writers (and for Win32 Perl for C<fork()> emulation). See L<threads>,
1000 L<threads::shared>, and L<perlthrtut>.
1004 C<threads::shared>, by Arthur Bergman, allows data sharing for
1005 interpreter threads. In the ithreads model any data sharing between
1006 threads must be explicit, as opposed to the old 5.005 thread model
1007 where data sharing was implicit. See L<threads::shared>.
1011 C<Tie::File>, by Mark-Jason Dominus, associates a Perl array with the
1012 lines of a file. See L<Tie::File>.
1016 C<Tie::Memoize>, by Ilya Zakharevich, provides on-demand loaded hashes.
1017 See L<Tie::Memoize>.
1021 C<Tie::RefHash::Nestable>, by Edward Avis, allows storing hash
1022 references (unlike the standard Tie::RefHash) The module is contained
1023 within Tie::RefHash. See L<Tie::RefHash>.
1027 C<Time::HiRes>, by Douglas E. Wegscheid, provides high resolution
1028 timing (ualarm, usleep, and gettimeofday). See L<Time::HiRes>.
1032 C<Unicode::UCD> offers a querying interface to the Unicode Character
1033 Database. See L<Unicode::UCD>.
1037 C<Unicode::Collate>, by SADAHIRO Tomoyuki, implements the UCA
1038 (Unicode Collation Algorithm) for sorting Unicode strings.
1039 See L<Unicode::Collate>.
1043 C<Unicode::Normalize>, by SADAHIRO Tomoyuki, implements the various
1044 Unicode normalization forms. See L<Unicode::Normalize>.
1048 C<XS::Typemap>, by Tim Jenness, is a test extension that exercises
1049 XS typemaps. Nothing gets installed, but the code is worth studying
1050 for extension writers.
1054 =head2 Updated And Improved Modules and Pragmata
1060 The following independently supported modules have been updated to the
1061 newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp,
1062 Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle
1063 (Pod::Man, Pod::Text), Pod::LaTeX [561+], Pod::Parser, Storable,
1064 Term::ANSIColor, Test, Text-Tabs+Wrap.
1068 attributes::reftype() now works on tied arguments.
1072 AutoLoader can now be disabled with C<no AutoLoader;>.
1076 B::Deparse has been significantly enhanced by Robin Houston. It can
1077 now deparse almost all of the standard test suite (so that the tests
1078 still succeed). There is a make target "test.deparse" for trying this
1083 Carp now has better interface documentation, and the @CARP_NOT
1084 interface has been added to get optional control over where errors
1085 are reported independently of @ISA, by Ben Tilly.
1089 Class::Struct can now define the classes in compile time.
1093 Class::Struct now assigns the array/hash element if the accessor
1094 is called with an array/hash element as the B<sole> argument.
1098 The return value of Cwd::fastcwd() is now tainted.
1102 Data::Dumper now has an option to sort hashes.
1106 Data::Dumper now has an option to dump code references
1111 DB_File now supports newer Berkeley DB versions, among
1116 Devel::Peek now has an interface for the Perl memory statistics
1117 (this works only if you are using perl's malloc, and if you have
1118 compiled with debugging).
1122 The English module can now be used without the infamous performance
1125 use English '-no_match_vars';
1127 (Assuming, of course, that you don't need the troublesome variables
1128 C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and
1129 C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>.
1133 ExtUtils::MakeMaker now uses File::Spec internally, which hopefully
1134 leads to better portability.
1138 Fcntl, Socket, and Sys::Syslog have been rewritten by Nicholas Clark
1139 to use the new-style constant dispatch section (see L<ExtUtils::Constant>).
1140 This means that they will be more robust and hopefully faster.
1144 File::Find now chdir()s correctly when chasing symbolic links. [561]
1148 File::Find now has pre- and post-processing callbacks. It also
1149 correctly changes directories when chasing symbolic links. Callbacks
1150 (naughtily) exiting with "next;" instead of "return;" now work.
1154 File::Find is now (again) reentrant. It also has been made
1159 The warnings issued by File::Find now belong to their own category.
1160 You can enable/disable them with C<use/no warnings 'File::Find';>.
1164 File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid
1165 prototype mismatch with CORE::glob(). [561]
1169 File::Glob now supports C<GLOB_LIMIT> constant to limit the size of
1170 the returned list of filenames.
1174 IPC::Open3 now allows the use of numeric file descriptors.
1178 IO::Socket now has an atmark() method, which returns true if the socket
1179 is positioned at the out-of-band mark. The method is also exportable
1180 as a sockatmark() function.
1184 IO::Socket::INET has support for the ReusePort option (if your
1185 platform supports it). The Reuse option now has an alias, ReuseAddr.
1186 For clarity, you may want to prefer ReuseAddr.
1190 IO::Socket::INET now supports a value of zero for C<LocalPort>
1191 (usually meaning that the operating system will make one up.)
1195 'use lib' now works identically to @INC. Removing directories
1196 with 'no lib' now works.
1200 Math::BigFloat and Math::BigInt have undergone a full rewrite by Tels.
1201 They are now magnitudes faster, and they support various bignum
1202 libraries such as GMP and PARI as their backends.
1206 Math::Complex handles inf, NaN etc., better.
1210 Net::Ping has been considerably enhanced by Rob Brown: multihoming is
1211 now supported, Win32 functionality is better, there is now time
1212 measuring functionality (optionally high-resolution using
1213 Time::HiRes), and there is now "external" protocol which uses
1214 Net::Ping::External module which runs your external ping utility and
1215 parses the output. A version of Net::Ping::External is available in
1218 Note that some of the Net::Ping tests are disabled when running
1219 under the Perl distribution since one cannot assume one or more
1220 of the following: enabled echo port at localhost, full Internet
1221 connectivity, or sympathetic firewalls. You can set the environment
1222 variable PERL_TEST_Net_Ping to "1" (one) before running the Perl test
1223 suite to enable all the Net::Ping tests.
1227 POSIX::sigaction() is now much more flexible and robust.
1228 You can now install coderef handlers, 'DEFAULT', and 'IGNORE'
1229 handlers, installing new handlers was not atomic.
1233 In Safe, C<%INC> is now localised in a Safe compartment so that
1238 In SDBM_File on dosish platforms, some keys went missing because of
1239 lack of support for files with "holes". A workaround for the problem
1244 In Search::Dict one can now have a pre-processing hook for the
1245 lines being searched.
1249 The Shell module now has an OO interface.
1253 In Sys::Syslog there is now a failover mechanism that will go
1254 through alternative connection mechanisms until the message
1255 is successfully logged.
1259 The Test module has been significantly enhanced.
1263 Time::Local::timelocal() does not handle fractional seconds anymore.
1264 The rationale is that neither does localtime(), and timelocal() and
1265 localtime() are supposed to be inverses of each other.
1269 The vars pragma now supports declaring fully qualified variables.
1270 (Something that C<our()> does not and will not support.)
1274 The C<utf8::> name space (as in the pragma) provides various
1275 Perl-callable functions to provide low level access to Perl's
1276 internal Unicode representation. At the moment only length()
1277 has been implemented.
1281 =head1 Utility Changes
1287 Emacs perl mode (emacs/cperl-mode.el) has been updated to version
1292 F<emacs/e2ctags.pl> is now much faster.
1296 C<enc2xs> is a tool for people adding their own encodings to the
1301 C<h2ph> now supports C trigraphs.
1305 C<h2xs> now produces a template README.
1309 C<h2xs> now uses C<Devel::PPPort> for better portability between
1310 different versions of Perl.
1314 C<h2xs> uses the new L<ExtUtils::Constant|ExtUtils::Constant> module
1315 which will affect newly created extensions that define constants.
1316 Since the new code is more correct (if you have two constants where the
1317 first one is a prefix of the second one, the first constant B<never>
1318 got defined), less lossy (it uses integers for integer constant,
1319 as opposed to the old code that used floating point numbers even for
1320 integer constants), and slightly faster, you might want to consider
1321 regenerating your extension code (the new scheme makes regenerating
1322 easy). L<h2xs> now also supports C trigraphs.
1326 C<libnetcfg> has been added to configure libnet.
1330 C<perlbug> is now much more robust. It also sends the bug report to
1331 perl.org, not perl.com.
1335 C<perlcc> has been rewritten and its user interface (that is,
1336 command line) is much more like that of the UNIX C compiler, cc.
1337 (The perlbc tools has been removed. Use C<perlcc -B> instead.)
1338 B<Note that perlcc is still considered very experimental and
1343 C<perlivp> is a new Installation Verification Procedure utility
1344 for running any time after installing Perl.
1348 C<piconv> is an implementation of the character conversion utility
1349 C<iconv>, demonstrating the new Encode module.
1353 C<pod2html> now allows specifying a cache directory.
1357 C<pod2html> now produces XHTML 1.0.
1361 C<pod2html> now understands POD written using different line endings
1362 (PC-like CRLF versus UNIX-like LF versus MacClassic-like CR).
1366 C<s2p> has been completely rewritten in Perl. (It is in fact a full
1367 implementation of sed in Perl: you can use the sed functionality by
1368 using the C<psed> utility.)
1372 C<xsubpp> now understands POD documentation embedded in the *.xs
1377 C<xsubpp> now supports the OUT keyword.
1381 =head1 New Documentation
1387 perl56delta details the changes between the 5.005 release and the
1392 perlclib documents the internal replacements for standard C library
1393 functions. (Interesting only for extension writers and Perl core
1398 perldebtut is a Perl debugging tutorial. [561+]
1402 perlebcdic contains considerations for running Perl on EBCDIC
1407 perlintro is a gentle introduction to Perl.
1411 perliol documents the internals of PerlIO with layers.
1415 perlmodstyle is a style guide for writing modules.
1419 perlnewmod tells about writing and submitting a new module. [561+]
1423 perlpacktut is a pack() tutorial.
1427 perlpod has been rewritten to be clearer and to record the best
1428 practices gathered over the years.
1432 perlpodspec is a more formal specification of the pod format,
1433 mainly of interest for writers of pod applications, not to
1434 people writing in pod.
1438 perlretut is a regular expression tutorial. [561+]
1442 perlrequick is a regular expressions quick-start guide.
1443 Yes, much quicker than perlretut. [561]
1447 perltodo has been updated.
1451 perltootc has been renamed as perltooc (to not to conflict
1452 with perltoot in filesystems restricted to "8.3" names).
1456 perluniintro is an introduction to using Unicode in Perl.
1457 (perlunicode is more of a detailed reference and background
1462 perlutil explains the command line utilities packaged with the Perl
1463 distribution. [561+]
1467 The following platform-specific documents are available before
1468 the installation as README.I<platform>, and after the installation
1471 perlaix perlamiga perlapollo perlbeos perlbs2000
1472 perlce perlcygwin perldgux perldos perlepoc perlhpux
1473 perlhurd perlmachten perlmacos perlmint perlmpeix
1474 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1475 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1477 Eastern Asian Perl users are now welcomed in their own languages:
1478 README.jp (Japanese), README.ko (Korean), README.cn (simplified
1479 Chinese) and README.tw (traditional Chinese), which are written in
1480 normal pod but encoded in EUC-JP, EUC-KR, EUC-CN and Big5. These
1481 will get installed as
1483 perljp perlko perlcn perltw
1489 The documentation for the POSIX-BC platform is called "BS2000", to avoid
1490 confusion with the Perl POSIX module.
1494 The documentation for the WinCE platform is called perlce (README.ce
1495 in the source code kit), to avoid confusion with the perlwin32
1496 documentation on 8.3-restricted filesystems.
1500 =head1 Performance Enhancements
1506 map() could get pathologically slow when the result list it generates
1507 is larger than the source list. The performance has been improved for
1508 common scenarios. [561]
1512 sort() has been changed to use primarily mergesort internally as
1513 opposed to the earlier quicksort. For very small lists this may
1514 result in slightly slower sorting times, but in general the speedup
1515 should be at least 20%. Additional bonuses are that the worst case
1516 behaviour of sort() is now better (in computer science terms it now
1517 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1518 worst-case run time behaviour), and that sort() is now stable
1519 (meaning that elements with identical keys will stay ordered as they
1520 were before the sort). See the C<sort> pragma for information.
1522 The story in more detail: suppose you want to serve yourself a little
1525 @digits = ( 3,1,4,1,5,9 );
1527 A numerical sort of the digits will yield (1,1,3,4,5,9), as expected.
1528 Which C<1> comes first is hard to know, since one C<1> looks pretty
1529 much like any other. You can regard this as totally trivial,
1530 or somewhat profound. However, if you just want to sort the even
1531 digits ahead of the odd ones, then what will
1533 sort { ($a % 2) <=> ($b % 2) } @digits;
1535 yield? The only even digit, C<4>, will come first. But how about
1536 the odd numbers, which all compare equal? With the quicksort algorithm
1537 used to implement Perl 5.6 and earlier, the order of ties is left up
1538 to the sort. So, as you add more and more digits of Pi, the order
1539 in which the sorted even and odd digits appear will change.
1540 and, for sufficiently large slices of Pi, the quicksort algorithm
1541 in Perl 5.8 won't return the same results even if reinvoked with the
1542 same input. The justification for this rests with quicksort's
1543 worst case behavior. If you run
1545 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1547 (something you might approximate if you wanted to merge two sorted
1548 arrays using sort), doubling $N doesn't just double the quicksort time,
1549 it I<quadruples> it. Quicksort has a worst case run time that can
1550 grow like N**2, so-called I<quadratic> behaviour, and it can happen
1551 on patterns that may well arise in normal use. You won't notice this
1552 for small arrays, but you I<will> notice it with larger arrays,
1553 and you may not live long enough for the sort to complete on arrays
1554 of a million elements. So the 5.8 quicksort scrambles large arrays
1555 before sorting them, as a statistical defence against quadratic behaviour.
1556 But that means if you sort the same large array twice, ties may be
1557 broken in different ways.
1559 Because of the unpredictability of tie-breaking order, and the quadratic
1560 worst-case behaviour, quicksort was I<almost> replaced completely with
1561 a stable mergesort. I<Stable> means that ties are broken to preserve
1562 the original order of appearance in the input array. So
1564 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1566 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1567 appear in the output in the same order they appeared in the input.
1568 Mergesort has worst case O(N log N) behaviour, the best value
1569 attainable. And, ironically, this mergesort does particularly
1570 well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N)
1571 in O(N) time. But quicksort was rescued at the last moment because
1572 it is faster than mergesort on certain inputs and platforms.
1573 For example, if you really I<don't> care about the order of even
1574 and odd digits, quicksort will run in O(N) time; it's very good
1575 at sorting many repetitions of a small number of distinct elements.
1576 The quicksort divide and conquer strategy works well on platforms
1577 with relatively small, very fast, caches. Eventually, the problem gets
1578 whittled down to one that fits in the cache, from which point it
1579 benefits from the increased memory speed.
1581 Quicksort was rescued by implementing a sort pragma to control aspects
1582 of the sort. The B<stable> subpragma forces stable behaviour,
1583 regardless of algorithm. The B<_quicksort> and B<_mergesort>
1584 subpragmas are heavy-handed ways to select the underlying implementation.
1585 The leading C<_> is a reminder that these subpragmas may not survive
1586 beyond 5.8. More appropriate mechanisms for selecting the implementation
1587 exist, but they wouldn't have arrived in time to save quicksort.
1591 Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm
1592 ( http://burtleburtle.net/bob/hash/doobs.html ). This algorithm is
1593 reasonably fast while producing a much better spread of values than
1594 the old hashing algorithm (originally by Chris Torek, later tweaked by
1595 Ilya Zakharevich). Hash values output from the algorithm on a hash of
1596 all 3-char printable ASCII keys comes much closer to passing the
1597 DIEHARD random number generation tests. According to perlbench, this
1598 change has not affected the overall speed of Perl.
1602 unshift() should now be noticeably faster.
1606 =head1 Installation and Configuration Improvements
1608 =head2 Generic Improvements
1614 INSTALL now explains how you can configure Perl to use 64-bit
1615 integers even on non-64-bit platforms.
1619 Policy.sh policy change: if you are reusing a Policy.sh file
1620 (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1621 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1622 them will now be changed to the new prefix, /foo/bar. (Previously
1623 only $prefix changed.) If you do not like this new behaviour,
1624 specify prefix, siteprefix, and vendorprefix explicitly.
1628 A new optional location for Perl libraries, otherlibdirs, is available.
1629 It can be used for example for vendor add-ons without disturbing Perl's
1630 own library directories.
1634 In many platforms, the vendor-supplied 'cc' is too stripped-down to
1635 build Perl (basically, 'cc' doesn't do ANSI C). If this seems
1636 to be the case and 'cc' does not seem to be the GNU C compiler
1637 'gcc', an automatic attempt is made to find and use 'gcc' instead.
1641 gcc needs to closely track the operating system release to avoid
1642 build problems. If Configure finds that gcc was built for a different
1643 operating system release than is running, it now gives a clearly visible
1644 warning that there may be trouble ahead.
1648 Since Perl 5.8 is not binary-compatible with previous releases
1649 of Perl, Configure no longer suggests including the 5.005
1654 Configure C<-S> can now run non-interactively.
1658 Configure support for pdp11-style memory models has been removed due
1659 to obsolescence. [561]
1663 configure.gnu now works with options with whitespace in them.
1667 installperl now outputs everything to STDERR.
1671 Because PerlIO is now the default on most platforms, "-perlio" doesn't
1672 get appended to the $Config{archname} (also known as $^O) anymore.
1673 Instead, if you explicitly choose not to use perlio (Configure command
1674 line option -Uuseperlio), you will get "-stdio" appended.
1678 Another change related to the architecture name is that "-64all"
1679 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1680 pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)
1684 In AFS installations, one can configure the root of the AFS to be
1685 somewhere else than the default F</afs> by using the Configure
1686 parameter C<-Dafsroot=/some/where/else>.
1690 APPLLIB_EXP, a lesser-known configuration-time definition, has been
1691 documented. It can be used to prepend site-specific directories
1692 to Perl's default search path (@INC); see INSTALL for information.
1696 The version of Berkeley DB used when the Perl (and, presumably, the
1697 DB_File extension) was built is now available as
1698 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
1699 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
1700 DB_VERSION_PATCH_CFG> from C.
1704 Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM
1705 has been documented in INSTALL.
1709 If you have CPAN access (either network or a local copy such as a
1710 CD-ROM) you can during specify extra modules to Configure to build and
1711 install with Perl using the -Dextras=... option. See INSTALL for
1716 In addition to config.over, a new override file, config.arch, is
1717 available. This file is supposed to be used by hints file writers
1718 for architecture-wide changes (as opposed to config.over which is
1719 for site-wide changes).
1723 If your file system supports symbolic links, you can build Perl outside
1724 of the source directory by
1726 mkdir /tmp/perl/build/directory
1727 cd /tmp/perl/build/directory
1728 sh /path/to/perl/source/Configure -Dmksymlinks ...
1730 This will create in /tmp/perl/build/directory a tree of symbolic links
1731 pointing to files in /path/to/perl/source. The original files are left
1732 unaffected. After Configure has finished, you can just say
1736 and Perl will be built and tested, all in /tmp/perl/build/directory.
1741 For Perl developers, several new make targets for profiling
1742 and debugging have been added; see L<perlhack>.
1748 Use of the F<gprof> tool to profile Perl has been documented in
1749 L<perlhack>. There is a make target called "perl.gprof" for
1750 generating a gprofiled Perl executable.
1754 If you have GCC 3, there is a make target called "perl.gcov" for
1755 creating a gcoved Perl executable for coverage analysis. See
1760 If you are on IRIX or Tru64 platforms, new profiling/debugging options
1761 have been added; see L<perlhack> for more information about pixie and
1768 Guidelines of how to construct minimal Perl installations have
1769 been added to INSTALL.
1773 The Thread extension is now not built at all under ithreads
1774 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
1775 Thread extension requires being Configured with C<-Duse5005threads>).
1777 But note that the Thread.pm interface is now shared by both
1782 The Gconvert macro ($Config{d_Gconvert}) used by perl for stringifying
1783 floating-point numbers is now more picky about using sprintf %.*g
1784 rules for the conversion. Some platforms that used to use gcvt may
1785 now resort to the slower sprintf.
1789 The obsolete method of making a special (e.g., debugging) flavor
1792 make LIBPERL=libperld.a
1794 has been removed. Use -DDEBUGGING instead.
1798 =head2 New Or Improved Platforms
1800 For the list of platforms known to support Perl,
1801 see L<perlport/"Supported Platforms">.
1807 AIX dynamic loading should be now better supported.
1811 AIX should now work better with gcc, threads, and 64-bitness. Also the
1812 long doubles support in AIX should be better now. See L<perlaix>.
1816 AtheOS ( http://www.atheos.cx/ ) is a new platform.
1820 BeOS has been reclaimed.
1824 The DG/UX platform now supports 5.005-style threads.
1829 The DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near
1834 EBCDIC platforms (z/OS (also known as OS/390), POSIX-BC, and VM/ESA)
1835 have been regained. Many test suite tests still fail and the
1836 co-existence of Unicode and EBCDIC isn't quite settled, but the
1837 situation is much better than with Perl 5.6. See L<perlos390>,
1838 L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information.
1842 Building perl with -Duseithreads or -Duse5005threads now works under
1843 HP-UX 10.20 (previously it only worked under 10.30 or later). You will
1844 need a thread library package installed. See README.hpux. [561]
1848 Mac OS Classic (MacPerl has of course been available since
1849 perl 5.004 but now the source code bases of standard Perl
1850 and MacPerl have been synchronised)
1854 Mac OS X (or Darwin) should now be able to build Perl even on HFS+
1855 filesystems. (The case-insensitivity used to confuse the Perl build
1860 NCR MP-RAS is now supported. [561]
1864 All the NetBSD specific patches (except for the installation
1865 specific ones) have been merged back to the main distribution.
1869 NetWare from Novell is now supported. See L<perlnetware>.
1873 NonStop-UX is now supported. [561]
1877 NEC SUPER-UX is now supported.
1881 All the OpenBSD specific patches (except for the installation
1882 specific ones) have been merged back to the main distribution.
1886 Perl has been tested with the GNU pth userlevel thread package
1887 ( http://www.gnu.org/software/pth/pth.html ) . All but one thread
1888 test worked, and that one failure was because of test results arriving
1889 in unexpected order.
1893 Stratus VOS is now supported using Perl's native build method
1894 (Configure). This is the recommended method to build Perl on
1895 VOS. The older methods, which build miniperl, are still
1896 available. See L<perlvos>. [561+]
1900 The Amdahl UTS UNIX mainframe platform is now supported. [561]
1904 WinCE is now supported. See L<perlce>.
1908 z/OS (formerly known as OS/390, formerly known as MVS OE) now has
1909 support for dynamic loading. This is not selected by default,
1910 however, you must specify -Dusedl in the arguments of Configure. [561]
1914 =head1 Selected Bug Fixes
1916 Numerous memory leaks and uninitialized memory accesses have been
1917 hunted down. Most importantly, anonymous subs used to leak quite
1924 The autouse pragma didn't work for Multi::Part::Function::Names.
1928 caller() could cause core dumps in certain situations. Carp was sometimes
1929 affected by this problem. In particular, caller() now returns a
1930 subroutine name of C<(unknown)> for subroutines that have been removed
1931 from the symbol table. [561+]
1935 chop(@list) in list context returned the characters chopped in
1936 reverse order. This has been reversed to be in the right order. [561]
1940 Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm)
1941 when building the Perl binary. The only exception to this is SunOS 4.x,
1942 which needs them. [561]
1946 The behaviour of non-decimal but numeric string constants such as
1947 "0x23" was platform-dependent: in some platforms that was seen as 35,
1948 in some as 0, in some as a floating point number (don't ask). This
1949 was caused by Perl's using the operating system libraries in a situation
1950 where the result of the string to number conversion is undefined: now
1951 Perl consistently handles such strings as zero in numeric contexts.
1955 The order of DESTROYs has been made more predictable.
1959 Several debugger fixes: exit code now reflects the script exit code,
1960 condition C<"0"> now treated correctly, the C<d> command now checks
1961 line number, C<$.> no longer gets corrupted, and all debugger output
1962 now goes correctly to the socket if RemotePort is set. [561]
1966 Perl 5.6.0 could emit spurious warnings about redefinition of dl_error()
1967 when statically building extensions into perl. This has been corrected. [561]
1971 L<dprofpp> -R didn't work.
1975 C<*foo{FORMAT}> now works.
1979 Infinity is now recognized as a number.
1983 UNIVERSAL::isa no longer caches methods incorrectly. (This broke
1984 the Tk extension with 5.6.0.) [561]
1988 Lexicals I: lexicals outside an eval "" weren't resolved
1989 correctly inside a subroutine definition inside the eval "" if they
1990 were not already referenced in the top level of the eval""ed code.
1994 Lexicals II: lexicals leaked at file scope into subroutines that
1995 were declared before the lexicals.
1999 Lexical warnings now propagating correctly between scopes
2000 and into C<eval "...">. [561+]
2004 C<use warnings qw(FATAL all)> did not work as intended. This has been
2009 warnings::enabled() now reports the state of $^W correctly if the caller
2010 isn't using lexical warnings. [561]
2014 Line renumbering with eval and C<#line> now works. [561]
2018 Fixed numerous memory leaks, especially in eval "".
2022 Localised tied variables no longer leak memory
2025 tie my %tied_hash => 'Tie::StdHash';
2029 # Used to leak memory every time local() was called;
2030 # in a loop, this added up.
2031 local($tied_hash{Foo}) = 1;
2035 Localised hash elements (and %ENV) are correctly unlocalised to not
2036 exist, if they didn't before they were localised.
2040 tie my %tied_hash => 'Tie::StdHash';
2044 # Nothing has set the FOO element so far
2046 { local $tied_hash{FOO} = 'Bar' }
2048 # This used to print, but not now.
2049 print "exists!\n" if exists $tied_hash{FOO};
2051 As a side effect of this fix, tied hash interfaces B<must> define
2052 the EXISTS and DELETE methods.
2056 mkdir() now ignores trailing slashes in the directory name,
2057 as mandated by POSIX.
2061 Some versions of glibc have a broken modfl(). This affects builds
2062 with C<-Duselongdouble>. This version of Perl detects this brokenness
2063 and has a workaround for it. The glibc release 2.2.2 is known to have
2064 fixed the modfl() bug.
2068 Modulus of unsigned numbers now works (4063328477 % 65535 used to
2069 return 27406, instead of 27047). [561]
2073 Some "not a number" warnings introduced in 5.6.0 eliminated to be
2074 more compatible with 5.005. Infinity is now recognised as a number. [561]
2078 Numeric conversions did not recognize changes in the string value
2079 properly in certain circumstances. [561]
2083 Attributes (such as :shared) didn't work with our().
2087 our() variables will not cause "will not stay shared" warnings.
2091 "our" variables of the same name declared in two sibling blocks
2092 resulted in bogus warnings about "redeclaration" of the variables.
2093 The problem has been corrected. [561]
2097 pack "Z" now correctly terminates the string with "\0".
2101 Fix password routines which in some shadow password platforms
2102 (e.g. HP-UX) caused getpwent() to return every other entry.
2106 The PERL5OPT environment variable (for passing command line arguments
2107 to Perl) didn't work for more than a single group of options. [561]
2111 PERL5OPT with embedded spaces didn't work.
2115 printf() no longer resets the numeric locale to "C".
2119 C<qw(a\\b)> now parses correctly as C<'a\\b'>: that is, as three
2120 characters, not four. [561]
2124 pos() did not return the correct value within s///ge in earlier
2125 versions. This is now handled correctly. [561]
2129 Printing quads (64-bit integers) with printf/sprintf now works
2130 without the q L ll prefixes (assuming you are on a quad-capable platform).
2134 Regular expressions on references and overloaded scalars now work. [561+]
2138 Right-hand side magic (GMAGIC) could in many cases such as string
2139 concatenation be invoked too many times.
2143 scalar() now forces scalar context even when used in void context.
2147 SOCKS support is now much more robust.
2151 sort() arguments are now compiled in the right wantarray context
2152 (they were accidentally using the context of the sort() itself).
2153 The comparison block is now run in scalar context, and the arguments
2154 to be sorted are always provided list context. [561]
2158 Changed the POSIX character class C<[[:space:]]> to include the (very
2159 rarely used) vertical tab character. Added a new POSIX-ish character
2160 class C<[[:blank:]]> which stands for horizontal whitespace
2161 (currently, the space and the tab).
2165 The tainting behaviour of sprintf() has been rationalized. It does
2166 not taint the result of floating point formats anymore, making the
2167 behaviour consistent with that of string interpolation. [561]
2171 Some cases of inconsistent taint propagation (such as within hash
2172 values) have been fixed.
2176 The RE engine found in Perl 5.6.0 accidentally pessimised certain kinds
2177 of simple pattern matches. These are now handled better. [561]
2181 Regular expression debug output (whether through C<use re 'debug'>
2182 or via C<-Dr>) now looks better. [561]
2186 Multi-line matches like C<"a\nxb\n" =~ /(?!\A)x/m> were flawed. The
2187 bug has been fixed. [561]
2191 Use of $& could trigger a core dump under some situations. This
2192 is now avoided. [561]
2196 The regular expression captured submatches ($1, $2, ...) are now
2197 more consistently unset if the match fails, instead of leaving false
2198 data lying around in them. [561]
2202 readline() on files opened in "slurp" mode could return an extra
2203 "" (blank line) at the end in certain situations. This has been
2208 Autovivification of symbolic references of special variables described
2209 in L<perlvar> (as in C<${$num}>) was accidentally disabled. This works
2214 Sys::Syslog ignored the C<LOG_AUTH> constant.
2218 All but the first argument of the IO syswrite() method are now optional.
2222 $AUTOLOAD, sort(), lock(), and spawning subprocesses
2223 in multiple threads simultaneously are now thread-safe.
2227 Tie::Array's SPLICE method was broken.
2231 Allow a read-only string on the left-hand side of a non-modifying tr///.
2235 If C<STDERR> is tied, warnings caused by C<warn> and C<die> now
2236 correctly pass to it.
2240 Several Unicode fixes.
2246 BOMs (byte order marks) at the beginning of Perl files
2247 (scripts, modules) should now be transparently skipped.
2248 UTF-16 and UCS-2 encoded Perl files should now be read correctly.
2252 The character tables have been updated to Unicode 3.2.0.
2256 Comparing with utf8 data does not magically upgrade non-utf8 data
2257 into utf8. (This was a problem for example if you were mixing data
2258 from I/O and Unicode data: your output might have got magically encoded
2263 Generating illegal Unicode code points such as U+FFFE, or the UTF-16
2264 surrogates, now also generates an optional warning.
2268 C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase.
2272 Concatenation with the C<.> operator or via variable interpolation,
2273 C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator,
2274 substitution with C<s///>, single-quoted UTF8, should now work.
2278 The C<tr///> operator now works. Note that the C<tr///CU>
2279 functionality has been removed (but see pack('U0', ...)).
2283 C<eval "v200"> now works.
2287 Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spurious warnings.
2288 This has been corrected. [561]
2292 Zero entries were missing from the Unicode classes such as C<IsDigit>.
2298 Large unsigned numbers (those above 2**31) could sometimes lose their
2299 unsignedness, causing bogus results in arithmetic operations. [561]
2303 The Perl parser has been stress tested using both random input and
2304 Markov chain input and the few found crashes and lockups have been
2309 =head2 Platform Specific Changes and Fixes
2317 Perl now works on post-4.0 BSD/OSes.
2323 Setting C<$0> now works (as much as possible; see L<perlvar> for details).
2329 Numerous updates; currently synchronised with Cygwin 1.3.10.
2333 Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.
2339 EPOC update after Perl 5.6.0. See README.epoc. [561]
2345 Perl now works on post-3.0 FreeBSDs.
2351 README.hpux updated; C<Configure -Duse64bitall> now works;
2352 now uses HP-UX malloc instead of Perl malloc.
2358 Numerous compilation flag and hint enhancements; accidental mixing
2359 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
2369 Long doubles should now work (see INSTALL). [561]
2373 Linux previously had problems related to sockaddrlen when using
2374 accept(), recvfrom() (in Perl: recv()), getpeername(), and
2383 Compilation of the standard Perl distribution in Mac OS Classic should
2384 now work if you have the Metrowerks development environment and the
2385 missing Mac-specific toolkit bits. Contact the macperl mailing list
2392 MPE/iX update after Perl 5.6.0. See README.mpeix. [561]
2396 NetBSD/threads: try installing the GNU pth (should be in the
2397 packages collection, or http://www.gnu.org/software/pth/),
2398 and Configure with -Duseithreads.
2404 Perl now works on NetBSD/sparc.
2410 Now works with usethreads (see INSTALL). [561]
2416 64-bitness using the Sun Workshop compiler now works.
2422 The native build method requires at least VOS Release 14.5.0
2423 and GNU C++/GNU Tools 2.0.1 or later. The Perl pack function
2424 now maps overflowed values to +infinity and underflowed values
2429 Tru64 (aka Digital UNIX, aka DEC OSF/1)
2431 The operating system version letter now recorded in $Config{osvers}.
2432 Allow compiling with gcc (previously explicitly forbidden). Compiling
2433 with gcc still not recommended because buggy code results, even with
2440 Fixed various alignment problems that lead into core dumps either
2441 during build or later; no longer dies on math errors at runtime;
2442 now using full quad integers (64 bits), previously was using
2443 only 46 bit integers for speed.
2449 chdir() now works better despite a CRT bug; now works with MULTIPLICITY
2450 (see INSTALL); now works with Perl's malloc.
2452 The tainting of C<%ENV> elements via C<keys> or C<values> was previously
2453 unimplemented. It now works as documented.
2455 The C<waitpid> emulation has been improved. The worst bug (now fixed)
2456 was that a pid of -1 would cause a wildcard search of all processes on
2459 POSIX-style signals are now emulated much better on VMS versions prior
2462 The C<system> function and backticks operator have improved
2463 functionality and better error handling. [561]
2465 File access tests now use current process privileges rather than the
2466 user's default privileges, which could sometimes result in a mismatch
2467 between reported access and actual access.
2469 There is a new C<kill> implementation based on C<sys$sigprc> that allows
2470 older VMS systems (pre-7.0) to use C<kill> to send signals rather than
2471 simply force exit. This implementation also allows later systems to
2472 call C<kill> from within a signal handler.
2474 Iterative logical name translations are now limited to 10 iterations in
2475 imitation of SHOW LOGICAL and other OpenVMS facilities.
2485 accept() no longer leaks memory. [561]
2489 Borland C++ v5.5 is now a supported compiler that can build Perl.
2490 However, the generated binaries continue to be incompatible with those
2491 generated by the other supported compilers (GCC and Visual C++). [561]
2495 Better chdir() return value for a non-existent directory.
2499 Duping socket handles with open(F, ">&MYSOCK") now works under Windows
2504 New %ENV entries now propagate to subprocesses. [561]
2508 Current directory entries in %ENV are now correctly propagated to child
2513 $ENV{LIB} now used to search for libs under Visual C.
2517 fork() emulation has been improved in various ways, but still continues
2518 to be experimental. See L<perlfork> for known bugs and caveats. [561+]
2522 A failed (pseudo)fork now returns undef and sets errno to EAGAIN.
2526 Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root.
2527 Other bugs in chdir() and Cwd::cwd() have also been fixed. [561]
2531 HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html
2535 The makefiles now provide a single switch to bulk-enable all the
2536 features enabled in ActiveState ActivePerl (a popular Win32 binary
2537 distribution). [561]
2541 Allow REG_EXPAND_SZ keys in the registry.
2545 Can now send() from all threads, not just the first one. [561]
2549 Fake signal handling reenabled, bugs and all.
2553 %SIG has been enabled under USE_ITHREADS, but its use is completely
2554 unsupported under all configurations. [561]
2558 Less stack reserved per thread so that more threads can run
2559 concurrently. (Still 16M per thread.)
2563 C<< File::Spec->tmpdir() >> now prefers C:/temp over /tmp
2564 (works better when perl is running as service).
2568 Better UNC path handling under ithreads. [561]
2572 wait(), waitpid(), and backticks now return the correct exit status
2573 under Windows 9x. [561]
2577 Win64 compilation is now supported.
2581 winsock handle leak fixed. [561]
2587 =head1 New or Changed Diagnostics
2593 The lexical warnings category "deprecated" is no longer a sub-category
2594 of the "syntax" category. It is now a top-level category in its own
2599 All regular expression compilation error messages are now hopefully
2600 easier to understand both because the error message now comes before
2601 the failed regex and because the point of failure is now clearly
2602 marked by a C<E<lt>-- HERE> marker.
2606 The various "opened only for", "on closed", "never opened" warnings
2607 drop the C<main::> prefix for filehandles in the C<main> package,
2608 for example C<STDIN> instead of C<main::STDIN>.
2612 The "Unrecognized escape" warning has been extended to include C<\8>,
2613 C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters.
2617 Two new debugging options have been added: if you have compiled your
2618 Perl with debugging, you can use the -DT [561] and -DR options to trace
2619 tokenising and to add reference counts to displaying variables,
2624 The debugger (perl5db.pl) has been modified to present a more
2625 consistent commands interface, via (CommandSet=580). perl5db.t was
2626 also added to test the changes, and as a placeholder for further tests.
2632 The debugger has a new C<dumpDepth> option to control the maximum
2633 depth to which nested structures are dumped. The C<x> command has
2634 been extended so that C<x N EXPR> dumps out the value of I<EXPR> to a
2635 depth of at most I<N> levels.
2639 The debugger can now show lexical variables if you have the CPAN
2640 module PadWalker installed.
2644 If an attempt to use a (non-blessed) reference as an array index
2645 is made, a warning is given.
2649 C<push @a;> and C<unshift @a;> (with no values to push or unshift)
2650 now give a warning. This may be a problem for generated and evaled
2655 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
2656 using the C<"C"> format you will get an optional warning. Similarly
2657 for the C<"c"> format and a number less than -128 or more than 127.
2661 Certain regex modifiers such as C<(?o)> make sense only if applied to
2662 the entire regex. You will get an optional warning if you try to do
2667 Using arrays or hashes as references (e.g. C<< %foo->{bar} >>
2668 has been deprecated for a while. Now you will get an optional warning.
2672 Using C<sort> in scalar context now issues an optional warning.
2673 This didn't do anything useful, as the sort was not performed.
2677 =head1 Changed Internals
2683 perlapi.pod (a companion to perlguts) now attempts to document the
2688 You can now build a really minimal perl called microperl.
2689 Building microperl does not require even running Configure;
2690 C<make -f Makefile.micro> should be enough. Beware: microperl makes
2691 many assumptions, some of which may be too bold; the resulting
2692 executable may crash or otherwise misbehave in wondrous ways.
2693 For careful hackers only.
2697 Added rsignal(), whichsig(), do_join(), op_clear, op_null,
2698 ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2699 interfaces to the publicised API. For the full list of the available
2700 APIs see L<perlapi>.
2704 Made possible to propagate customised exceptions via croak()ing.
2708 Now xsubs can have attributes just like subs. (Well, at least the
2709 built-in attributes.)
2713 dTHR and djSP have been obsoleted; the former removed (because it's
2714 a no-op) and the latter replaced with dSP.
2718 PERL_OBJECT has been completely removed.
2722 The MAGIC constants (e.g. C<'P'>) have been macrofied
2723 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
2724 and maintainability.
2728 The regex compiler now maintains a structure that identifies nodes in
2729 the compiled bytecode with the corresponding syntactic features of the
2730 original regex expression. The information is attached to the new
2731 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
2732 complete information.
2736 The C code has been made much more C<gcc -Wall> clean. Some warning
2737 messages still remain in some platforms, so if you are compiling with
2738 gcc you may see some warnings about dubious practices. The warnings
2739 are being worked on.
2743 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
2747 Documentation on how to use the Perl source repository has been added
2748 to F<Porting/repository.pod>.
2752 There are now several profiling make targets.
2756 =head1 Security Vulnerability Closed [561]
2758 (This change was already made in 5.7.0 but bears repeating here.)
2760 A potential security vulnerability in the optional suidperl component
2761 of Perl was identified in August 2000. suidperl is neither built nor
2762 installed by default. As of November 2001 the only known vulnerable
2763 platform is Linux, most likely all Linux distributions. CERT and
2764 various vendors and distributors have been alerted about the vulnerability.
2765 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2766 for more information.
2768 The problem was caused by Perl trying to report a suspected security
2769 exploit attempt using an external program, /bin/mail. On Linux
2770 platforms the /bin/mail program had an undocumented feature which
2771 when combined with suidperl gave access to a root shell, resulting in
2772 a serious compromise instead of reporting the exploit attempt. If you
2773 don't have /bin/mail, or if you have 'safe setuid scripts', or if
2774 suidperl is not installed, you are safe.
2776 The exploit attempt reporting feature has been completely removed from
2777 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2778 from all the Perl 5.7 releases), so that particular vulnerability
2779 isn't there anymore. However, further security vulnerabilities are,
2780 unfortunately, always possible. The suidperl functionality is most
2781 probably going to be removed in Perl 5.10. In any case, suidperl
2782 should only be used by security experts who know exactly what they are
2783 doing and why they are using suidperl instead of some other solution
2784 such as sudo ( see http://www.courtesan.com/sudo/ ).
2786 =head1 New Tests [561+]
2788 Several new tests have been added, especially for the F<lib> and F<ext>
2789 subsections. There are now about 65 000 individual tests (spread over
2790 about 700 test scripts), in the regression suite (5.6.1 has about
2791 11700 tests, in 258 test scripts) Many of the new tests are of course
2792 introduced by the new modules, but still in general Perl is now more
2795 Because of the large number of tests, running the regression suite
2796 will take considerably longer time than it used to: expect the suite
2797 to take up to 4-5 times longer to run than in perl 5.6. On a really
2798 fast machine you can hope to finish the suite in about 6-8 minutes
2801 The tests are now reported in a different order than in earlier Perls.
2802 (This happens because the test scripts from under t/lib have been moved
2803 to be closer to the library/extension they are testing.)
2805 =head1 Known Problems
2813 If using the AIX native make command, instead of just "make" issue
2814 "make all". In some setups the former has been known to spuriously
2815 also try to run "make install". Alternatively, you may want to use
2820 In AIX 4.2, Perl extensions that use C++ functions that use statics
2821 may have problems in that the statics are not getting initialized.
2822 In newer AIX releases, this has been solved by linking Perl with
2823 the libC_r library, but unfortunately in AIX 4.2 the said library
2824 has an obscure bug where the various functions related to time
2825 (such as time() and gettimeofday()) return broken values, and
2826 therefore in AIX 4.2 Perl is not linked against libC_r.
2830 vac 5.0.0.0 May Produce Buggy Code For Perl
2832 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
2833 resulting in a few random tests failing when run as part of "make
2834 test", but when the failing tests are run by hand, they succeed.
2835 We suggest upgrading to at least vac version 5.0.1.0, that has been
2836 known to compile Perl correctly. "lslpp -L|grep vac.C" will tell
2837 you the vac version. See README.aix.
2841 If building threaded Perl, you may get compilation warning from pp_sys.c:
2843 "pp_sys.c", line 4651.39: 1506-280 (W) Function argument assignment between types "unsigned char*" and "const void*" is not allowed.
2845 This is harmless; it is caused by the getnetbyaddr() and getnetbyaddr_r()
2846 having slightly different types for their first argument.
2850 =head2 Alpha systems with old gccs fail several tests
2852 If you see op/pack, op/pat, op/regexp, or ext/Storable tests failing
2853 in a Linux/alpha or *BSD/Alpha, it's probably time to upgrade your gcc.
2854 gccs prior to 2.95.3 are definitely not good enough, and gcc 3.1 may
2855 be even better. (RedHat Linux/alpha with gcc 3.1 reported no problems,
2856 as did Linux 2.4.18 with gcc 2.95.4.) (In Tru64, it is preferable to
2857 use the bundled C compiler.)
2861 Perl 5.8.0 doesn't build in AmigaOS. It broke at some point
2862 during the ithreads work and we could not find Amiga experts
2863 to unbreak the problems. [561]
2867 The following tests fail on 5.8.0 Perl in BeOS Personal 5.03:
2869 t/op/lfs............................FAILED at test 17
2870 t/op/magic..........................FAILED at test 24
2871 ext/POSIX/t/sigaction...............FAILED at test 13
2872 ext/POSIX/t/waitpid.................FAILED at test 1
2874 See L<perlbeos> (README.beos) for more details.
2876 =head2 Cygwin "unable to remap"
2878 For example when building the Tk extension for Cygwin,
2879 you may get an error message saying "unable to remap".
2880 This is known problem with Cygwin, and a workaround is
2881 detailed in here: http://sources.redhat.com/ml/cygwin/2001-12/msg00894.html
2883 =head2 ext/threads/t/libc
2885 If this test fails, it indicates that your libc (C library) is not
2886 threadsafe. This particular test stress tests the localtime() call to
2887 find out whether it is threadsafe. See L<perlthrtut> for more information.
2889 =head2 FreeBSD Failing locale Test 117 For ISO8859-15 Locales
2891 The ISO8859-15 locales may fail the locale test 117 in FreeBSD.
2892 This is caused by the characters \xFF (y with diaeresis) and \xBE
2893 (Y with diaeresis) not behaving correctly when being matched
2896 =head2 IRIX fails ext/List/Util/t/shuffle.t
2898 IRIX with MIPSpro 7.3.1.3m compiler may fail the said List::Util test
2899 by dumping core. This seems to be a compiler error since if compiled
2900 with gcc no core dump ensues, and no failures on the said test on any
2903 =head2 Modifying $_ Inside for(..)
2907 works without complaint. It shouldn't. (You should be able to
2908 modify only lvalue elements inside the loops.) You can see the
2909 correct behaviour by replacing the 1..5 with 1, 2, 3, 4, 5.
2911 =head2 mod_perl 1.26 Doesn't Build With Threaded Perl
2913 Use mod_perl 1.27 or higher.
2915 =head2 lib/ftmp-security tests warn 'system possibly insecure'
2917 Don't panic. Read the 'make test' section of INSTALL instead.
2919 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
2921 If perl is configured with -Duse64bitall, the successful result of the
2922 subtest 10 of lib/posix may arrive before the successful result of the
2923 subtest 9, which confuses the test harness so much that it thinks the
2926 =head2 Linux with glibc 2.2.5 fails t/op/int subtest #6 with -Duse64bitint
2928 This is a known bug in the glibc 2.2.5 with long long integers.
2929 ( http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=65612 )
2931 =head2 Linux With Sfio Fails op/misc Test 48
2935 =head2 libwww-perl (LWP) fails base/date #51
2937 Use libwww-perl 5.65 or later.
2941 Please remember to set your environment variable LC_ALL to "C"
2942 (setenv LC_ALL C) before running "make test" to avoid a lot of
2943 warnings about the broken locales of Mac OS X.
2945 The following tests are known to fail in Mac OS X 10.1.5 because of
2946 buggy (old) implementations of Berkeley DB included in Mac OS X:
2948 Failed Test Stat Wstat Total Fail Failed List of Failed
2949 -------------------------------------------------------------------------
2950 ../ext/DB_File/t/db-btree.t 0 11 ?? ?? % ??
2951 ../ext/DB_File/t/db-recno.t 149 3 2.01% 61 63 65
2953 If you are building on a UFS partition, you will also probably see
2954 t/op/stat.t subtest #9 fail. This is caused by Darwin's UFS not
2955 supporting inode change time.
2957 Also the ext/POSIX/t/posix.t subtest #10 fails but it is skipped for
2958 now because the failure is Apple's fault, not Perl's (blocked signals
2961 If you Configure with ithreads, ext/threads/t/libc.t will fail. Again,
2962 this is not Perl's fault-- the libc of Mac OS X is not threadsafe
2963 (in this particular test, the localtime() call is found to be
2966 =head2 op/sprintf tests 91, 129, and 130
2968 The op/sprintf tests 91, 129, and 130 are known to fail on some platforms.
2969 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
2971 Test 91 is known to fail on QNX6 (nto), because C<sprintf '%e',0>
2972 incorrectly produces C<0.000000e+0> instead of C<0.000000e+00>.
2974 For tests 129 and 130, the failing platforms do not comply with
2975 the ANSI C Standard: lines 19ff on page 134 of ANSI X3.159 1989, to
2976 be exact. (They produce something other than "1" and "-1" when
2977 formatting 0.6 and -0.6 using the printf format "%.0f"; most often,
2978 they produce "0" and "-0".)
2982 In case you are still using Solaris 2.5 (aka SunOS 5.5), you may
2983 experience failures (the test core dumping) in lib/locale.t.
2984 The suggested cure is to upgrade your Solaris.
2986 =head2 SUPER-UX (NEC SX)
2988 The following tests are known to fail on SUPER-UX:
2990 op/64bitint...........................FAILED tests 29-30, 32-33, 35-36
2991 op/arith..............................FAILED tests 128-130
2992 op/pack...............................FAILED tests 25-5625
2993 op/pow................................
2994 op/taint..............................# msgsnd failed
2995 ../ext/IO/lib/IO/t/io_poll............FAILED tests 3-4
2996 ../ext/IPC/SysV/ipcsysv...............FAILED tests 2, 5-6
2997 ../ext/IPC/SysV/t/msg.................FAILED tests 2, 4-6
2998 ../ext/Socket/socketpair..............FAILED tests 12
2999 ../lib/IPC/SysV.......................FAILED tests 2, 5-6
3000 ../lib/warnings.......................FAILED tests 115-116, 118-119
3002 The op/pack failure ("Cannot compress negative numbers at op/pack.t line 126")
3003 is serious but as of yet unsolved. It points at some problems with the
3004 signedness handling of the C compiler, as do the 64bitint, arith, and pow
3005 failures. Most of the rest point at problems with SysV IPC.
3007 =head2 Term::ReadKey not working on Win32
3009 Use Term::ReadKey 2.20 or later.
3011 =head2 Failure of Thread (5.005-style) tests
3013 B<Note that support for 5.005-style threading is deprecated,
3014 experimental and practically unsupported. In 5.10, it is expected
3017 The following tests are known to fail due to fundamental problems in
3018 the 5.005 threading implementation. These are not new failures--Perl
3019 5.005_0x has the same bugs, but didn't have these tests.
3021 ../ext/B/t/xref.t 255 65280 14 12 85.71% 3-14
3022 ../ext/List/Util/t/first.t 255 65280 7 4 57.14% 2 5-7
3023 ../lib/English.t 2 512 54 2 3.70% 2-3
3024 ../lib/ExtUtils/t/basic.t 1 256 17 1 5.88% 14
3025 ../lib/FileCache.t 5 1 20.00% 5
3026 ../lib/Filter/Simple/t/data.t 6 3 50.00% 1-3
3027 ../lib/Filter/Simple/t/filter_onl 9 3 33.33% 1-2 5
3028 ../lib/Tie/File/t/31_autodefer.t 255 65280 65 32 49.23% 34-65
3029 ../lib/autouse.t 10 1 10.00% 4
3030 op/flip.t 15 1 6.67% 15
3032 These failures are unlikely to get fixed as 5.005-style threads
3033 are considered fundamentally broken. (Basically what happens is that
3034 competing threads can corrupt shared global state.)
3036 =head2 Timing problems
3038 The following tests may fail intermittently because of timing
3039 problems, for example if the system is heavily loaded.
3042 ext/Time/HiRes/HiRes.t
3044 lib/Memoize/t/expmod_t.t
3045 lib/Memoize/t/speed.t
3047 In case of failure please try running them manually, for example
3049 ./perl -Ilib ext/Time/HiRes/HiRes.t
3053 ../lib/Math/Trig.t 26 1 3.85% 25
3054 ../lib/warnings.t 470 1 0.21% 429
3056 The Trig.t failure is caused by the slighly differing (from IEEE)
3057 floating point implementation of UNICOS. The warnings.t failure is
3058 also related: the test assumes a certain floating point output format;
3059 this assumption fails in UNICOS.
3063 =head2 VOS (Stratus)
3065 When Perl is built using the native build process on VOS Release
3066 14.5.0 and GNU C++/GNU Tools 2.0.1, all attempted tests either
3067 pass or result in TODO (ignored) failures.
3073 During Configure, the test
3075 Guessing which symbols your C compiler and preprocessor define...
3077 will probably fail with error messages like
3079 CC-20 cc: ERROR File = try.c, Line = 3
3080 The identifier "bad" is undefined.
3082 bad switch yylook 79bad switch yylook 79bad switch yylook 79bad switch yylook 79#ifdef A29K
3085 CC-65 cc: ERROR File = try.c, Line = 3
3086 A semicolon is expected at this point.
3088 This is caused by a bug in the awk utility of UNICOS/mk. You can ignore
3089 the error, but it does cause a slight problem: you cannot fully
3090 benefit from the h2ph utility (see L<h2ph>) that can be used to
3091 convert C headers to Perl libraries, mainly used to be able to access
3092 from Perl the constants defined using C preprocessor, cpp. Because of
3093 the above error, parts of the converted headers will be invisible.
3094 Luckily, these days the need for h2ph is rare.
3098 If building Perl with interpreter threads (ithreads), the
3099 getgrent(), getgrnam(), and getgrgid() functions cannot return the
3100 list of the group members due to a bug in the multithreaded support of
3101 UNICOS/mk. What this means is that in list context the functions will
3102 return only three values, not four.
3108 There are a few known test failures, see L<perluts> (README.uts).
3112 There should be no reported test failures with a default configuration,
3113 though there are a number of tests marked TODO that point to areas
3114 needing further debugging and/or porting work.
3118 In multi-CPU boxes, there are some problems with the I/O buffering:
3119 some output may appear twice.
3121 =head2 XML::Parser not working
3123 Use XML::Parser 2.31 or later.
3125 =head2 z/OS (OS/390)
3127 z/OS has rather many test failures but the situation is actually
3128 better than it was in 5.6.0; it's just that so many new modules and
3129 tests have been added.
3131 Failed Test Stat Wstat Total Fail Failed List of Failed
3132 ---------------------------------------------------------------------------
3133 ../ext/Data/Dumper/t/dumper.t 357 8 2.24% 311 314 325 327
3135 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5
3136 ../ext/Storable/t/downgrade.t 12 3072 169 12 7.10% 14-15 46-47 78-79
3138 ../lib/ExtUtils/t/Constant.t 121 30976 48 48 100.00% 1-48
3139 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9
3140 op/pat.t 910 7 0.77% 665 776 785 832-
3142 op/sprintf.t 224 3 1.34% 98 100 136
3143 op/tr.t 97 5 5.15% 63 71-74
3144 uni/fold.t 780 6 0.77% 61 169 196 661
3147 The failures in dumper.t and downgrade.t are problems in the tests,
3148 those in io_unix and sprintf are problems in the USS (UDP sockets
3149 and printf formats). The pat, tr, and fold failures are genuine Perl
3150 problems caused by EBCDIC (and in the pat and fold cases, combining
3151 that with Unicode). The Constant and Embed are probably problems
3152 in the tests (since they test Perl's ability to build extensions,
3153 and that seems to be working reasonably well.)
3155 =head2 Localising Tied Arrays and Hashes Is Broken
3159 doesn't work as one would expect: the old value is restored
3160 incorrectly. This will be changed in a future release, but we don't
3161 know yet what the new semantics will exactly be. In any case, the
3162 change will break existing code that relies on the current
3163 (ill-defined) semantics, so just avoid doing this in general.
3165 =head2 Self-tying Problems
3167 Self-tying of arrays and hashes is broken in rather deep and
3168 hard-to-fix ways. As a stop-gap measure to avoid people from getting
3169 frustrated at the mysterious results (core dumps, most often), it is
3170 forbidden for now (you will get a fatal error even from an attempt).
3172 A change to self-tying of globs has caused them to be recursively
3173 referenced (see: L<perlobj/"Two-Phased Garbage Collection">). You
3174 will now need an explicit untie to destroy a self-tied glob. This
3175 behaviour may be fixed at a later date.
3177 Self-tying of scalars and IO thingies works.
3179 =head2 Building Extensions Can Fail Because Of Largefiles
3181 Some extensions like mod_perl are known to have issues with
3182 `largefiles', a change brought by Perl 5.6.0 in which file offsets
3183 default to 64 bits wide, where supported. Modules may fail to compile
3184 at all, or they may compile and work incorrectly. Currently, there
3185 is no good solution for the problem, but Configure now provides
3186 appropriate non-largefile ccflags, ldflags, libswanted, and libs
3187 in the %Config hash (e.g., $Config{ccflags_nolargefiles}) so the
3188 extensions that are having problems can try configuring themselves
3189 without the largefileness. This is admittedly not a clean solution,
3190 and the solution may not even work at all. One potential failure is
3191 whether one can (or, if one can, whether it's a good idea to) link
3192 together at all binaries with different ideas about file offsets;
3193 all this is platform-dependent.
3195 =head2 Unicode Support on EBCDIC Still Spotty
3197 Though mostly working, Unicode support still has problem spots on
3198 EBCDIC platforms. One such known spot are the C<\p{}> and C<\P{}>
3199 regular expression constructs for code points less than 256: the
3200 C<pP> are testing for Unicode code points, not knowing about EBCDIC.
3202 =head2 The Compiler Suite Is Still Very Experimental
3204 The compiler suite is slowly getting better but it continues to be
3205 highly experimental. Use in production environments is discouraged.
3207 =head2 The Long Double Support Is Still Experimental
3209 The ability to configure Perl's numbers to use "long doubles",
3210 floating point numbers of hopefully better accuracy, is still
3211 experimental. The implementations of long doubles are not yet
3212 widespread and the existing implementations are not quite mature
3213 or standardised, therefore trying to support them is a rare
3214 and moving target. The gain of more precision may also be offset
3215 by slowdown in computations (more bits to move around, and the
3216 operations are more likely to be executed by less optimised
3219 =head2 Seen In Perl 5.7 But Gone Now
3221 C<Time::Piece> (previously known as C<Time::Object>) was removed
3222 because it was felt that it didn't have enough value in it to be a
3223 core module. It is still a useful module, though, and is available
3226 Perl 5.8 unfortunately does not build anymore on AmigaOS;
3227 this broke accidentally at some point. Since there are not that many
3228 Amiga developers available, we could not get this fixed and tested in
3231 =head1 Reporting Bugs
3233 If you find what you think is a bug, you might check the articles
3234 recently posted to the comp.lang.perl.misc newsgroup and the perl
3235 bug database at http://bugs.perl.org/ . There may also be
3236 information at http://www.perl.com/ , the Perl Home Page.
3238 If you believe you have an unreported bug, please run the B<perlbug>
3239 program included with your release. Be sure to trim your bug down
3240 to a tiny but sufficient test case. Your bug report, along with the
3241 output of C<perl -V>, will be sent off to perlbug@perl.org to be
3242 analysed by the Perl porting team.
3246 The F<Changes> file for exhaustive details on what changed.
3248 The F<INSTALL> file for how to build Perl.
3250 The F<README> file for general stuff.
3252 The F<Artistic> and F<Copying> files for copyright information.
3256 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>.