3 perldelta - what is new for perl v5.8.0
7 This document describes differences between the 5.6.0 release and
10 Many of the bug fixes in 5.8.0 were already seen in the 5.6.1
11 maintenance release since the two releases were kept closely
12 coordinated (while 5.8.0 was still called 5.7.something).
14 Changes that were integrated into the 5.6.1 release are marked C<[561]>.
15 Many of these changes have been further developed since 5.6.1 was released,
16 those are marked C<[561+]>.
18 You can see the list of changes in the 5.6.1 release (both from the
19 5.005_03 release and the 5.6.0 release) by reading L<perl561delta>.
21 =head1 Highlights In 5.8.0
27 Better Unicode support
35 New Thread Implementation
39 Better Numeric Accuracy
51 More Extensive Regression Testing
55 =head1 Incompatible Changes
57 =head2 Binary Incompatibility
59 B<Perl 5.8 is not binary compatible with earlier releases of Perl.>
61 B<You have to recompile your XS modules.>
63 (Pure Perl modules should continue to work.)
65 The major reason for the discontinuity is the new IO architecture
66 called PerlIO. PerlIO is the default configuration because without
67 it many new features of Perl 5.8 cannot be used. In other words:
68 you just have to recompile your modules containing XS code, sorry
71 In future releases of Perl, non-PerlIO aware XS modules may become
72 completely unsupported. This shouldn't be too difficult for module
73 authors, however: PerlIO has been designed as a drop-in replacement
74 (at the source code level) for the stdio interface.
76 Depending on your platform, there are also other reasons why
77 we decided to break binary compatibility, please read on.
79 =head2 64-bit platforms and malloc
81 If your pointers are 64 bits wide, the Perl malloc is no longer being
82 used because it does not work well with 8-byte pointers. Also,
83 usually the system mallocs on such platforms are much better optimized
84 for such large memory models than the Perl malloc. Some memory-hungry
85 Perl applications like the PDL don't work well with Perl's malloc.
86 Finally, other applications than Perl (such as mod_perl) tend to prefer
87 the system malloc. Such platforms include Alpha and 64-bit HPPA,
90 =head2 AIX Dynaloading
92 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
93 dlopen interface of AIX instead of the old emulated interface. This
94 change will probably break backward compatibility with compiled
95 modules. The change was made to make Perl more compliant with other
96 applications like mod_perl which are using the AIX native interface.
98 =head2 Attributes for C<my> variables now handled at run-time.
100 The C<my EXPR : ATTRS> syntax now applies variable attributes at
101 run-time. (Subroutine and C<our> variables still get attributes applied
102 at compile-time.) See L<attributes> for additional details. In particular,
103 however, this allows variable attributes to be useful for C<tie> interfaces,
104 which was a deficiency of earlier releases. Note that the new semantics
105 doesn't work with the Attribute::Handlers module (as of version 0.76).
107 =head2 Socket Extension Dynamic in VMS
109 The Socket extension is now dynamically loaded instead of being
110 statically built in. This may or may not be a problem with ancient
111 TCP/IP stacks of VMS: we do not know since we weren't able to test
112 Perl in such configurations.
114 =head2 IEEE-format Floating Point Default on OpenVMS Alpha
116 Perl now uses IEEE format (T_FLOAT) as the default internal floating
117 point format on OpenVMS Alpha, potentially breaking binary compatibility
118 with external libraries or existing data. G_FLOAT is still available as
119 a configuration option. The default on VAX (D_FLOAT) has not changed.
121 =head2 New Unicode Properties
123 Unicode I<scripts> are now supported. Scripts are similar to (and superior
124 to) Unicode I<blocks>. The difference between scripts and blocks is that
125 scripts are the glyphs used by a language or a group of languages, while
126 the blocks are more artificial groupings of (mostly) 256 characters based
127 on the Unicode numbering.
129 In general, scripts are more inclusive, but not universally so. For
130 example, while the script C<Latin> includes all the Latin characters and
131 their various diacritic-adorned versions, it does not include the various
132 punctuation or digits (since they are not solely C<Latin>).
134 A number of other properties are now supported, including C<\p{L&}>,
135 C<\p{Any}> C<\p{Assigned}>, C<\p{Unassigned}>, C<\p{Blank}> [561] and
136 C<\p{SpacePerl}> [561] (along with their C<\P{...}> versions, of course).
137 See L<perlunicode> for details, and more additions.
139 The C<In> or C<Is> prefix to names used with the C<\p{...}> and C<\P{...}>
140 are now almost always optional. The only exception is that a C<In> prefix
141 is required to signify a Unicode block when a block name conflicts with a
142 script name. For example, C<\p{Tibetan}> refers to the script, while
143 C<\p{InTibetan}> refers to the block. When there is no name conflict, you
144 can omit the C<In> from the block name (e.g. C<\p{BraillePatterns}>), but
145 to be safe, it's probably best to always use the C<In>).
147 =head2 REF(...) Instead Of SCALAR(...)
149 A reference to a reference now stringifies as "REF(0x81485ec)" instead
150 of "SCALAR(0x81485ec)" in order to be more consistent with the return
153 =head2 pack/unpack D/F recycled
155 The undocumented pack/unpack template letters D/F have been recycled
156 for better use: now they stand for long double (if supported by the
157 platform) and NV (Perl internal floating point type). (They used
158 to be aliases for d/f, but you never knew that.)
166 The semantics of bless(REF, REF) were unclear and until someone proves
167 it to make some sense, it is forbidden.
171 The obsolete chat2 library that should never have been allowed
172 to escape the laboratory has been decommissioned.
176 The builtin dump() function has probably outlived most of its
177 usefulness. The core-dumping functionality will remain in future
178 available as an explicit call to C<CORE::dump()>, but in future
179 releases the behaviour of an unqualified C<dump()> call may change.
183 The very dusty examples in the eg/ directory have been removed.
184 Suggestions for new shiny examples welcome but the main issue is that
185 the examples need to be documented, tested and (most importantly)
190 The (bogus) escape sequences \8 and \9 now give an optional warning
191 ("Unrecognized escape passed through"). There is no need to \-escape
196 The list of filenames from glob() (or <...>) is now by default sorted
197 alphabetically to be csh-compliant (which is what happened before
198 in most UNIX platforms). (bsd_glob() does still sort platform
199 natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.) [561]
203 Spurious syntax errors generated in certain situations, when glob()
204 caused File::Glob to be loaded for the first time, have been fixed. [561]
208 Although "you shouldn't do that", it was possible to write code that
209 depends on Perl's hashed key order (Data::Dumper does this). The new
210 algorithm "One-at-a-Time" produces a different hashed key order.
211 More details are in L</"Performance Enhancements">.
215 lstat(FILEHANDLE) now gives a warning because the operation makes no sense.
216 In future releases this may become a fatal error.
220 The C<package;> syntax (C<package> without an argument) has been
221 deprecated. Its semantics were never that clear and its
222 implementation even less so. If you have used that feature to
223 disallow all but fully qualified variables, C<use strict;> instead.
227 The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still
228 recognised but now cause fatal errors. The previous behaviour of
229 ignoring them by default and warning if requested was unacceptable
230 since it, in a way, falsely promised that the features could be used.
234 In future releases, non-PerlIO aware XS modules may become completely
235 unsupported. Since PerlIO is a drop-in replacement for stdio at the
236 source code level, this shouldn't be that drastic a change.
240 The PerlIO C<:raw> discipline (as described in Camel III) is deprecated
241 because its definition (as either the discipline version of C<binmode(FH)>
242 or as the opposite of C<:crlf>) didn't really work: most importantly
243 because turning off "clrfness" is not enough to make a stream truly
244 binary. Instead of C<:raw> use one of the following: C<open(..., ':bytes')>,
245 C<binmode(FH)>, C<sysopen()> + C<sysread()>.
249 The current user-visible implementation of pseudo-hashes (the weird
250 use of the first array element) is deprecated starting from Perl 5.8.0
251 and will be removed in Perl 5.10.0, and the feature will be
252 implemented differently. Not only is the current interface rather
253 ugly, but the current implementation slows down normal array and hash
254 use quite noticeably. The C<fields> pragma interface will remain
255 available. The I<restricted hashes> interface is expected to
256 be the replacement interface (see L<Hash::Util>).
260 The syntaxes C<< @a->[...] >> and C<< %h->{...} >> have now been deprecated.
264 After years of trying, suidperl is considered to be too complex to
265 ever be considered truly secure. The suidperl functionality is likely
266 to be removed in a future release.
270 The 5.005 threads model (module C<Thread>) is deprecated and expected
271 to be removed in Perl 5.10. Multithreaded code should be migrated to
272 the new ithreads model (see L<threads>, L<threads::shared> and
277 The long deprecated uppercase aliases for the string comparison
278 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
282 The tr///C and tr///U features have been removed and will not return;
283 the interface was a mistake. Sorry about that. For similar
284 functionality, see pack('U0', ...) and pack('C0', ...). [561]
288 Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo (@)".
289 The prototypes are now checked better at compile-time for invalid
290 syntax. An optional warning is generated ("Illegal character in
291 prototype...") but this may be upgraded to a fatal error in a future
296 The C<exec LIST> and C<system LIST> operations will produce fatal
297 errors on tainted data in some future release.
301 The existing behaviour when localising tied arrays and hashes is wrong,
302 and will be changed in a future release, so do not rely on the existing
303 behaviour. See L<"Localising Tied Arrays and Hashes Is Broken">.
307 =head1 Core Enhancements
309 =head2 PerlIO is Now The Default
315 IO is now by default done via PerlIO rather than system's "stdio".
316 PerlIO allows "layers" to be "pushed" onto a file handle to alter the
317 handle's behaviour. Layers can be specified at open time via 3-arg
320 open($fh,'>:crlf :utf8', $path) || ...
322 or on already opened handles via extended C<binmode>:
324 binmode($fh,':encoding(iso-8859-7)');
326 The built-in layers are: unix (low level read/write), stdio (as in
327 previous Perls), perlio (re-implementation of stdio buffering in a
328 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
329 but available on any platform). A mmap layer may be available if
330 platform supports it (mostly UNIXes).
332 Layers to be applied by default may be specified via the 'open' pragma.
334 See L</"Installation and Configuration Improvements"> for the effects
335 of PerlIO on your architecture name.
339 If your platform supports fork(), you can use the list form of C<open>
340 for pipes. For example:
342 open KID_PS, "-|", "ps", "aux" or die $!;
344 forks the ps(1) command (without spawning a shell, as there are more
345 than three arguments to open()), and reads its standard output via the
346 C<KID_PS> filehandle. See L<perlipc>.
350 File handles can be marked as accepting Perl's internal encoding of Unicode
351 (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :
353 open($fh,">:utf8","Uni.txt");
355 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named
356 for you since it's not UTF-8 what you will be getting but instead
357 UTF-EBCDIC. See L<perlunicode>, L<utf8>, and
358 http://www.unicode.org/unicode/reports/tr16/ for more information.
359 In future releases this naming may change. See L<perluniintro>
360 for more information about UTF-8.
364 If your environment variables (LC_ALL, LC_CTYPE, LANG, LANGUAGE) look
365 like you want to use UTF-8 (any of the the variables match C</utf-?8/i>),
366 your STDIN, STDOUT, STDERR handles and the default open discipline
367 (see L<open>) are marked as UTF-8. (This feature, like other new
368 features that combine Unicode and I/O, work only if you are using
369 PerlIO, but that's is the default.)
371 Note that after this Perl really does assume that everything is UTF-8:
372 for example if some input handle is not, Perl will probably very soon
373 complain about the input data like this "Malformed UTF-8 ..." since
374 any old eight-bit data is not legal UTF-8.
376 Note for code authors: if you want to enable your users to use UTF-8
377 as their default encoding but in your code still have eight-bit I/O streams
378 (such as images or zip files), you need to explicitly open() or binmode()
379 with C<:bytes> (see L<perlfunc/open> and L<perlfunc/binmode>), or you
380 can just use C<binmode(FH)> (nice for pre-5.8.0 backward compatibility).
384 File handles can translate character encodings from/to Perl's internal
385 Unicode form on read/write via the ":encoding()" layer.
389 File handles can be opened to "in memory" files held in Perl scalars via:
391 open($fh,'>', \$variable) || ...
395 Anonymous temporary files are available without need to
396 'use FileHandle' or other module via
398 open($fh,"+>", undef) || ...
400 That is a literal undef, not an undefined value.
404 The list form of C<open> is now implemented for pipes (at least on UNIX):
406 open($fh,"-|", 'cat', '/etc/motd')
408 creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in
413 If your locale environment variables (LANGUAGE, LC_ALL, LC_CTYPE, LANG)
414 contain the strings 'UTF-8' or 'UTF8' (case-insensitive matching),
415 the default encoding of your STDIN, STDOUT, and STDERR, and of
416 B<any subsequent file open>, is UTF-8.
420 =head2 Restricted Hashes
422 A restricted hash is restricted to a certain set of keys, no keys
423 outside the set can be added. Also individual keys can be restricted
424 so that the key cannot be deleted and the value cannot be changed.
425 No new syntax is involved: the Hash::Util module is the interface.
429 Perl used to be fragile in that signals arriving at inopportune moments
430 could corrupt Perl's internal state. Now Perl postpones handling of
431 signals until it's safe (between opcodes).
433 This change may have surprising side effects because signals no longer
434 interrupt Perl instantly. Perl will now first finish whatever it was
435 doing, like finishing an internal operation (like sort()) or an
436 external operation (like an I/O operation), and only then look at any
437 arrived signals (and before starting the next operation). No more corrupt
438 internal state since the current operation is always finished first,
439 but the signal may take more time to get heard. Note that breaking
440 out from potentially blocking operations should still work, though.
442 =head2 Unicode Overhaul
444 Unicode in general should be now much more usable than in Perl 5.6.0
445 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in
446 regular expressions should work now, Unicode in tr/// should work now,
447 Unicode in I/O should work now. See L<perluniintro> for introduction
448 and L<perlunicode> for details.
454 The Unicode Character Database coming with Perl has been upgraded
455 to Unicode 3.2.0. For more information, see http://www.unicode.org/ .
456 [561+] (5.6.1 has UCD 3.0.1.)
460 For developers interested in enhancing Perl's Unicode capabilities:
461 almost all the UCD files are included with the Perl distribution in
462 the F<lib/unicore> subdirectory. The most notable omission, for space
463 considerations, is the Unihan database.
467 The properties \p{Blank} and \p{SpacePerl} have been added. "Blank" is like
468 C isblank(), that is, it contains only "horizontal whitespace" (the space
469 character is, the newline isn't), and the "SpacePerl" is the Unicode
470 equivalent of C<\s> (\p{Space} isn't, since that includes the vertical
471 tabulator character, whereas C<\s> doesn't.)
473 See "New Unicode Properties" earlier in this document for additional
474 information on changes with Unicode properties.
478 =head2 Understanding of Numbers
480 In general a lot of fixing has happened in the area of Perl's
481 understanding of numbers, both integer and floating point. Since in
482 many systems the standard number parsing functions like C<strtoul()>
483 and C<atof()> seem to have bugs, Perl tries to work around their
484 deficiencies. This results hopefully in more accurate numbers.
486 Perl now tries internally to use integer values in numeric conversions
487 and basic arithmetics (+ - * /) if the arguments are integers, and
488 tries also to keep the results stored internally as integers.
489 This change leads to often slightly faster and always less lossy
490 arithmetics. (Previously Perl always preferred floating point numbers
493 =head2 Arrays now always interpolate into double-quoted strings [561]
495 In double-quoted strings, arrays now interpolate, no matter what. The
496 behavior in earlier versions of perl 5 was that arrays would interpolate
497 into strings if the array had been mentioned before the string was
498 compiled, and otherwise Perl would raise a fatal compile-time error.
499 In versions 5.000 through 5.003, the error was
501 Literal @example now requires backslash
503 In versions 5.004_01 through 5.6.0, the error was
505 In string, @example now must be written as \@example
507 The idea here was to get people into the habit of writing
508 C<"fred\@example.com"> when they wanted a literal C<@> sign, just as
509 they have always written C<"Give me back my \$5"> when they wanted a
512 Starting with 5.6.1, when Perl now sees an C<@> sign in a
513 double-quoted string, it I<always> attempts to interpolate an array,
514 regardless of whether or not the array has been used or declared
515 already. The fatal error has been downgraded to an optional warning:
517 Possible unintended interpolation of @example in string
519 This warns you that C<"fred@example.com"> is going to turn into
520 C<fred.com> if you don't backslash the C<@>.
521 See http://www.plover.com/~mjd/perl/at-error.html for more details
522 about the history here.
524 =head2 Miscellaneous Changes
530 AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute
531 to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.
535 The $Config{byteorder} (and corresponding BYTEORDER in config.h) was
536 previously wrong in platforms if sizeof(long) was 4, but sizeof(IV)
537 was 8. The byteorder was only sizeof(long) bytes long (1234 or 4321),
538 but now it is correctly sizeof(IV) bytes long, (12345678 or 87654321).
539 (This problem didn't affect Windows platforms.)
541 Also, $Config{byteorder} is now computed dynamically--this is more
542 robust with "fat binaries" where an executable image contains binaries
543 for more than one binary platform, and when cross-compiling.
547 C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass
548 in multiple arguments.)
552 C<do> followed by a bareword now ensures that this bareword isn't
553 a keyword (to avoid a bug where C<do q(foo.pl)> tried to call a
554 subroutine called C<q>). This means that for example instead of
555 C<do format()> you must write C<do &format()>.
559 The builtin dump() now gives an optional warning
560 C<dump() better written as CORE::dump()>,
561 meaning that by default C<dump(...)> is resolved as the builtin
562 dump() which dumps core and aborts, not as (possibly) user-defined
563 C<sub dump>. To call the latter, qualify the call as C<&dump(...)>.
564 (The whole dump() feature is to considered deprecated, and possibly
565 removed/changed in future releases.)
569 chomp() and chop() are now overridable. Note, however, that their
570 prototype (as given by C<prototype("CORE::chomp")> is undefined,
571 because it cannot be expressed and therefore one cannot really write
572 replacements to override these builtins.
576 END blocks are now run even if you exit/die in a BEGIN block.
577 Internally, the execution of END blocks is now controlled by
578 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new
579 behaviour for Perl embedders. This will default in 5.10. See
584 Formats now support zero-padded decimal fields.
588 Lvalue subroutines can now return C<undef> in list context. However,
589 the lvalue subroutine feature still remains experimental. [561+]
593 A lost warning "Can't declare ... dereference in my" has been
594 restored (Perl had it earlier but it became lost in later releases.)
598 A new special regular expression variable has been introduced:
599 C<$^N>, which contains the most-recently closed group (submatch).
603 C<no Module;> does not produce an error even if Module does not have an
604 unimport() method. This parallels the behavior of C<use> vis-a-vis
609 The numerical comparison operators return C<undef> if either operand
610 is a NaN. Previously the behaviour was unspecified.
614 C<our> can now have an experimental optional attribute C<unique> that
615 affects how global variables are shared among multiple interpreters,
620 The following builtin functions are now overridable: each(), keys(),
621 pop(), push(), shift(), splice(), unshift(). [561]
625 C<pack() / unpack()> can now group template letters with C<()> and then
626 apply repetition/count modifiers on the groups.
630 C<pack() / unpack()> can now process the Perl internal numeric types:
631 IVs, UVs, NVs-- and also long doubles, if supported by the platform.
632 The template letters are C<j>, C<J>, C<F>, and C<D>.
636 C<pack('U0a*', ...)> can now be used to force a string to UTF8.
640 my __PACKAGE__ $obj now works. [561]
644 POSIX::sleep() now returns the number of I<unslept> seconds
645 (as the POSIX standard says), as opposed to CORE::sleep() which
646 returns the number of slept seconds.
650 The printf() and sprintf() now support parameter reordering using the
651 C<%\d+\$> and C<*\d+\$> syntaxes. For example
653 print "%2\$s %1\$s\n", "foo", "bar";
655 will print "bar foo\n". This feature helps in writing
656 internationalised software, and in general when the order
657 of the parameters can vary.
661 The (\&) prototype now works properly. [561]
665 prototype(\[$@%&]) is now available to implicitly create references
666 (useful for example if you want to emulate the tie() interface).
670 A new command-line option, C<-t> is available. It is the
671 little brother of C<-T>: instead of dying on taint violations,
672 lexical warnings are given. B<This is only meant as a temporary
673 debugging aid while securing the code of old legacy applications.
674 This is not a substitute for -T.>
678 In other taint news, the C<exec LIST> and C<system LIST> have now been
679 considered too risky (think C<exec @ARGV>: it can start any program
680 with any arguments), and now the said forms cause a warning under
681 lexical warnings. You should carefully launder the arguments to
682 guarantee their validity. In future releases of Perl the forms will
683 become fatal errors so consider starting laundering now.
687 Tied hash interfaces are now required to have the EXISTS and DELETE
688 methods (either own or inherited).
692 If tr/// is just counting characters, it doesn't attempt to
697 untie() will now call an UNTIE() hook if it exists. See L<perltie>
702 L<utime> now supports C<utime undef, undef, @files> to change the
703 file timestamps to the current time.
707 The rules for allowing underscores (underbars) in numeric constants
708 have been relaxed and simplified: now you can have an underscore
709 simply B<between digits>.
713 Rather than relying on C's argv[0] (which may not contain a full pathname)
714 where possible $^X is now set by asking the operating system.
715 (eg by reading F</proc/self/exe> on Linux, F</proc/curproc/file> on FreeBSD)
719 A new variable, C<${^TAINT}>, indicates whether taint mode is enabled.
723 You can now override the readline() builtin, and this overrides also
724 the <FILEHANDLE> angle bracket operator.
728 The command-line options -s and -F are now recognized on the shebang
733 Use of the C</c> match modifier without an accompanying C</g> modifier
734 elicits a new warning: C<Use of /c modifier is meaningless without /g>.
736 Use of C</c> in substitutions, even with C</g>, elicits
737 C<Use of /c modifier is meaningless in s///>.
739 Use of C</g> with C<split> elicits C<Use of /g modifier is meaningless
744 Support for the C<CLONE> special subroutine had been added.
745 With ithreads, when a new thread is created, all Perl data is cloned,
746 however non-Perl data cannot be cloned automatically. In C<CLONE> you
747 can do whatever you need to do, like for example handle the cloning of
748 non-Perl data, if necessary. C<CLONE> will be executed once for every
749 package that has it defined or inherited. It will be called in the
750 context of the new thread, so all modifications are made in the new area.
756 =head1 Modules and Pragmata
758 =head2 New Modules and Pragmata
764 C<Attribute::Handlers>, originally by Damian Conway and now maintained
765 by Arthur Bergman, allows a class to define attribute handlers.
768 use Attribute::Handlers;
769 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
771 # later, in some package using or inheriting from MyPack...
773 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
775 Both variables and routines can have attribute handlers. Handlers can
776 be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the
777 exact compilation phase (BEGIN, CHECK, INIT, or END).
778 See L<Attribute::Handlers>.
782 C<B::Concise>, by Stephen McCamant, is a new compiler backend for
783 walking the Perl syntax tree, printing concise info about ops.
784 The output is highly customisable. See L<B::Concise>. [561+]
788 The new bignum, bigint, and bigrat pragmas, by Tels, implement
789 transparent bignum support (using the Math::BigInt, Math::BigFloat,
790 and Math::BigRat backends).
794 C<Class::ISA>, by Sean Burke, is a module for reporting the search
795 path for a class's ISA tree. See L<Class::ISA>.
799 C<Cwd> now has a split personality: if possible, an XS extension is
800 used, (this will hopefully be faster, more secure, and more robust)
801 but if not possible, the familiar Perl implementation is used.
805 C<Devel::PPPort>, originally by Kenneth Albanowski and now
806 maintained by Paul Marquess, has been added. It is primarily used
807 by C<h2xs> to enhance portability of XS modules between different
808 versions of Perl. See L<Devel::PPPort>.
812 C<Digest>, frontend module for calculating digests (checksums), from
813 Gisle Aas, has been added. See L<Digest>.
817 C<Digest::MD5> for calculating MD5 digests (checksums) as defined in
818 RFC 1321, from Gisle Aas, has been added. See L<Digest::MD5>.
820 use Digest::MD5 'md5_hex';
822 $digest = md5_hex("Thirsty Camel");
824 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
826 NOTE: the C<MD5> backward compatibility module is deliberately not
827 included since its further use is discouraged.
831 C<Encode>, originally by Nick Ing-Simmons and now maintained by Dan
832 Kogai, provides a mechanism to translate between different character
833 encodings. Support for Unicode, ISO-8859-1, and ASCII are compiled in
834 to the module. Several other encodings (like the rest of the
835 ISO-8859, CP*/Win*, Mac, KOI8-R, three variants EBCDIC, Chinese,
836 Japanese, and Korean encodings) are included and can be loaded at
837 runtime. (For space considerations, the largest Chinese encodings
838 have been separated into their own CPAN module, Encode::HanExtra,
839 which Encode will use if available). See L<Encode>.
841 Any encoding supported by Encode module is also available to the
842 ":encoding()" layer if PerlIO is used.
846 C<Hash::Util> is the interface to the new I<restricted hashes>
847 feature. (Implemented by Jeffrey Friedl, Nick Ing-Simmons, and
848 Michael Schwern.) See L<Hash::Util>.
852 C<I18N::Langinfo> can be used to query locale information.
853 See L<I18N::Langinfo>.
857 C<I18N::LangTags>, by Sean Burke, has functions for dealing with
858 RFC3066-style language tags. See L<I18N::LangTags>.
862 C<ExtUtils::Constant>, by Nicholas Clark, is a new tool for extension
863 writers for generating XS code to import C header constants.
864 See L<ExtUtils::Constant>.
868 C<Filter::Simple>, by Damian Conway, is an easy-to-use frontend to
869 Filter::Util::Call. See L<Filter::Simple>.
875 use Filter::Simple sub {
876 while (my ($from, $to) = splice @_, 0, 2) {
885 use MyFilter qr/red/ => 'green';
887 print "red\n"; # this code is filtered, will print "green\n"
888 print "bored\n"; # this code is filtered, will print "bogreen\n"
892 print "red\n"; # this code is not filtered, will print "red\n"
896 C<File::Temp>, by Tim Jenness, allows one to create temporary files
897 and directories in an easy, portable, and secure way. See L<File::Temp>.
902 C<Filter::Util::Call>, by Paul Marquess, provides you with the
903 framework to write I<source filters> in Perl. For most uses, the
904 frontend Filter::Simple is to be preferred. See L<Filter::Util::Call>.
908 C<if>, by Ilya Zakharevich, is a new pragma for conditional inclusion
913 L<libnet>, by Graham Barr, is a collection of perl5 modules related
914 to network programming. See L<Net::FTP>, L<Net::NNTP>, L<Net::Ping>
915 (not part of libnet, but related), L<Net::POP3>, L<Net::SMTP>,
918 Perl installation leaves libnet unconfigured; use F<libnetcfg>
923 C<List::Util>, by Graham Barr, is a selection of general-utility
924 list subroutines, such as sum(), min(), first(), and shuffle().
929 C<Locale::Constants>, C<Locale::Country>, C<Locale::Currency>
930 C<Locale::Language>, and L<Locale::Script>, by Neil Bowers, have
931 been added. They provide the codes for various locale standards, such
932 as "fr" for France, "usd" for US Dollar, and "ja" for Japanese.
936 $country = code2country('jp'); # $country gets 'Japan'
937 $code = country2code('Norway'); # $code gets 'no'
939 See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>,
940 and L<Locale::Language>.
944 C<Locale::Maketext>, by Sean Burke, is a localization framework. See
945 L<Locale::Maketext>, and L<Locale::Maketext::TPJ13>. The latter is an
946 article about software localization, originally published in The Perl
947 Journal #13, and republished here with kind permission.
951 C<Math::BigRat> for big rational numbers, to accompany Math::BigInt and
952 Math::BigFloat, from Tels. See L<Math::BigRat>.
956 C<Memoize> can make your functions faster by trading space for time,
957 from Mark-Jason Dominus. See L<Memoize>.
961 C<MIME::Base64>, by Gisle Aas, allows you to encode data in base64,
962 as defined in RFC 2045 - I<MIME (Multipurpose Internet Mail
967 $encoded = encode_base64('Aladdin:open sesame');
968 $decoded = decode_base64($encoded);
970 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
976 C<MIME::QuotedPrint>, by Gisle Aas, allows you to encode data
977 in quoted-printable encoding, as defined in RFC 2045 - I<MIME
978 (Multipurpose Internet Mail Extensions)>.
980 use MIME::QuotedPrint;
982 $encoded = encode_qp("Smiley in Unicode: \x{263a}");
983 $decoded = decode_qp($encoded);
985 print $encoded, "\n"; # "Smiley in Unicode: =263A"
987 MIME::QuotedPrint has been enhanced to provide the basic methods
988 necessary to use it with PerlIO::Via as in :
990 use MIME::QuotedPrint;
991 open($fh,">Via(MIME::QuotedPrint)",$path);
993 See L<MIME::QuotedPrint>.
997 C<NEXT>, by Damian Conway, is a pseudo-class for method redispatch.
1002 C<open> is a new pragma for setting the default I/O disciplines
1007 C<PerlIO::Scalar>, by Nick Ing-Simmons, provides the implementation
1008 of IO to "in memory" Perl scalars as discussed above. It also serves
1009 as an example of a loadable PerlIO layer. Other future possibilities
1010 include PerlIO::Array and PerlIO::Code. See L<PerlIO::Scalar>.
1014 C<PerlIO::Via>, by Nick Ing-Simmons, acts as a PerlIO layer and wraps
1015 PerlIO layer functionality provided by a class (typically implemented
1018 use MIME::QuotedPrint;
1019 open($fh,">Via(MIME::QuotedPrint)",$path);
1021 This will automatically convert everything output to C<$fh>
1022 to Quoted-Printable. See L<PerlIO::Via>.
1026 C<Pod::ParseLink>, by Russ Allbery, has been added,
1027 to parse LZ<><> links in pods as described in the new
1032 C<Pod::Text::Overstrike>, by Joe Smith, has been added.
1033 It converts POD data to formatted overstrike text.
1034 See L<Pod::Text::Overstrike>. [561+]
1038 C<Scalar::Util> is a selection of general-utility scalar subroutines,
1039 such as blessed(), reftype(), and tainted(). See L<Scalar::Util>.
1043 C<sort> is a new pragma for controlling the behaviour of sort().
1047 C<Storable> gives persistence to Perl data structures by allowing the
1048 storage and retrieval of Perl data to and from files in a fast and
1049 compact binary format. Because in effect Storable does serialisation
1050 of Perl data structues, with it you can also clone deep, hierarchical
1051 datastructures. Storable was originally created by Raphael Manfredi,
1052 but it is now maintained by Abhijit Menon-Sen. Storable has been
1053 enhanced to understand the two new hash features, Unicode keys and
1054 restricted hashes. See L<Storable>.
1058 C<Switch>, by Damian Conway, has been added. Just by saying
1062 you have C<switch> and C<case> available in Perl.
1068 case 1 { print "number 1" }
1069 case "a" { print "string a" }
1070 case [1..10,42] { print "number in list" }
1071 case (@array) { print "number in list" }
1072 case /\w+/ { print "pattern" }
1073 case qr/\w+/ { print "pattern" }
1074 case (%hash) { print "entry in hash" }
1075 case (\%hash) { print "entry in hash" }
1076 case (\&sub) { print "arg to subroutine" }
1077 else { print "previous case not true" }
1084 C<Test::More>, by Michael Schwern, is yet another framework for writing
1085 test scripts, more extensive than Test::Simple. See L<Test::More>.
1089 C<Test::Simple>, by Michael Schwern, has basic utilities for writing
1090 tests. See L<Test::Simple>.
1094 C<Text::Balanced>, by Damian Conway, has been added, for extracting
1095 delimited text sequences from strings.
1097 use Text::Balanced 'extract_delimited';
1099 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
1101 $a will be "'never say never'", $b will be ', he never said'.
1103 In addition to extract_delimited(), there are also extract_bracketed(),
1104 extract_quotelike(), extract_codeblock(), extract_variable(),
1105 extract_tagged(), extract_multiple(), gen_delimited_pat(), and
1106 gen_extract_tagged(). With these, you can implement rather advanced
1107 parsing algorithms. See L<Text::Balanced>.
1111 C<threads>, by Arthur Bergman, is an interface to interpreter threads.
1112 Interpreter threads (ithreads) is the new thread model introduced in
1113 Perl 5.6 but only available as an internal interface for extension
1114 writers (and for Win32 Perl for C<fork()> emulation). See L<threads>,
1115 L<threads::shared>, and L<perlthrtut>.
1119 C<threads::shared>, by Arthur Bergman, allows data sharing for
1120 interpreter threads. In the ithreads model any data sharing between
1121 threads must be explicit, as opposed to the old 5.005 thread model
1122 where data sharing was implicit. See L<threads::shared>.
1126 C<Tie::File>, by Mark-Jason Dominus, associates a Perl array with the
1127 lines of a file. See L<Tie::File>.
1131 C<Tie::Memoize>, by Ilya Zakharevich, provides on-demand loaded hashes.
1132 See L<Tie::Memoize>.
1136 C<Tie::RefHash::Nestable>, by Edward Avis, allows storing hash
1137 references (unlike the standard Tie::RefHash) The module is contained
1138 within Tie::RefHash. See L<Tie::RefHash>.
1142 C<Time::HiRes>, by Douglas E. Wegscheid, provides high resolution
1143 timing (ualarm, usleep, and gettimeofday). See L<Time::HiRes>.
1147 C<Unicode::UCD> offers a querying interface to the Unicode Character
1148 Database. See L<Unicode::UCD>.
1152 C<Unicode::Collate>, by SADAHIRO Tomoyuki, implements the UCA
1153 (Unicode Collation Algorithm) for sorting Unicode strings.
1154 See L<Unicode::Collate>.
1158 C<Unicode::Normalize>, by SADAHIRO Tomoyuki, implements the various
1159 Unicode normalization forms. See L<Unicode::Normalize>.
1163 C<XS::APItest>, by Tim Jenness, is a test extension that exercises XS
1164 APIs. Currently only C<printf()> is tested: how to output various
1165 basic data types from XS.
1169 C<XS::Typemap>, by Tim Jenness, is a test extension that exercises
1170 XS typemaps. Nothing gets installed, but the code is worth studying
1171 for extension writers.
1175 =head2 Updated And Improved Modules and Pragmata
1181 The following independently supported modules have been updated to the
1182 newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp,
1183 Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle
1184 (Pod::Man, Pod::Text), Pod::LaTeX [561+], Pod::Parser, Storable,
1185 Term::ANSIColor, Test, Text-Tabs+Wrap.
1189 attributes::reftype() now works on tied arguments.
1193 AutoLoader can now be disabled with C<no AutoLoader;>.
1197 B::Deparse has been significantly enhanced by Robin Houston. It can
1198 now deparse almost all of the standard test suite (so that the tests
1199 still succeed). There is a make target "test.deparse" for trying this
1204 Carp now has better interface documentation, and the @CARP_NOT
1205 interface has been added to get optional control over where errors
1206 are reported independently of @ISA, by Ben Tilly.
1210 Class::Struct can now define the classes in compile time.
1214 Class::Struct now assigns the array/hash element if the accessor
1215 is called with an array/hash element as the B<sole> argument.
1219 The return value of Cwd::fastcwd() is now tainted.
1223 Data::Dumper now has an option to sort hashes.
1227 Data::Dumper now has an option to dump code references
1232 DB_File now supports newer Berkeley DB versions, among
1237 Devel::Peek now has an interface for the Perl memory statistics
1238 (this works only if you are using perl's malloc, and if you have
1239 compiled with debugging).
1243 The English module can now be used without the infamous performance
1246 use English '-no_match_vars';
1248 (Assuming, of course, that you don't need the troublesome variables
1249 C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and
1250 C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>.
1254 ExtUtils::MakeMaker has been significantly cleaned up and fixed.
1255 The enhanced version has also been backported to earlier releases
1256 of Perl and submitted to CPAN so that the earlier releases can
1261 The arguments of WriteMakefile() in Makefile.PL are now checked
1262 for sanity much more carefully than before. This may cause new
1263 warnings when modules are being insalled. See L<ExtUtils::MakeMaker>
1268 ExtUtils::MakeMaker now uses File::Spec internally, which hopefully
1269 leads to better portability.
1273 Fcntl, Socket, and Sys::Syslog have been rewritten by Nicholas Clark
1274 to use the new-style constant dispatch section (see L<ExtUtils::Constant>).
1275 This means that they will be more robust and hopefully faster.
1279 File::Find now chdir()s correctly when chasing symbolic links. [561]
1283 File::Find now has pre- and post-processing callbacks. It also
1284 correctly changes directories when chasing symbolic links. Callbacks
1285 (naughtily) exiting with "next;" instead of "return;" now work.
1289 File::Find is now (again) reentrant. It also has been made
1294 The warnings issued by File::Find now belong to their own category.
1295 You can enable/disable them with C<use/no warnings 'File::Find';>.
1299 File::Glob::glob() has been renamed to File::Glob::bsd_glob()
1300 because the name clashes with the builtin glob(). The older
1301 name is still available for compatibility, but is deprecated. [561]
1305 File::Glob now supports C<GLOB_LIMIT> constant to limit the size of
1306 the returned list of filenames.
1310 IPC::Open3 now allows the use of numeric file descriptors.
1314 IO::Socket now has an atmark() method, which returns true if the socket
1315 is positioned at the out-of-band mark. The method is also exportable
1316 as a sockatmark() function.
1320 IO::Socket::INET failed to open the specified port if the service name
1321 was not known. It now correctly uses the supplied port number as is. [561]
1325 IO::Socket::INET has support for the ReusePort option (if your
1326 platform supports it). The Reuse option now has an alias, ReuseAddr.
1327 For clarity, you may want to prefer ReuseAddr.
1331 IO::Socket::INET now supports a value of zero for C<LocalPort>
1332 (usually meaning that the operating system will make one up.)
1336 'use lib' now works identically to @INC. Removing directories
1337 with 'no lib' now works.
1341 Math::BigFloat and Math::BigInt have undergone a full rewrite by Tels.
1342 They are now magnitudes faster, and they support various bignum
1343 libraries such as GMP and PARI as their backends.
1347 Math::Complex handles inf, NaN etc., better.
1351 Net::Ping has been considerably enhanced by Rob Brown: multihoming is
1352 now supported, Win32 functionality is better, there is now time
1353 measuring functionality (optionally high-resolution using
1354 Time::HiRes), and there is now "external" protocol which uses
1355 Net::Ping::External module which runs your external ping utility and
1356 parses the output. A version of Net::Ping::External is available in
1359 Note that some of the Net::Ping tests are disabled when running
1360 under the Perl distribution since one cannot assume one or more
1361 of the following: enabled echo port at localhost, full Internet
1362 connectivity, or sympathetic firewalls. You can set the environment
1363 variable PERL_TEST_Net_Ping to "1" (one) before running the Perl test
1364 suite to enable all the Net::Ping tests.
1368 POSIX::sigaction() is now much more flexible and robust.
1369 You can now install coderef handlers, 'DEFAULT', and 'IGNORE'
1370 handlers, installing new handlers was not atomic.
1374 In Safe, C<%INC> is now localised in a Safe compartment so that
1379 In SDBM_File on dosish platforms, some keys went missing because of
1380 lack of support for files with "holes". A workaround for the problem
1385 In Search::Dict one can now have a pre-processing hook for the
1386 lines being searched.
1390 The Shell module now has an OO interface.
1394 In Sys::Syslog there is now a failover mechanism that will go
1395 through alternative connection mechanisms until the message
1396 is successfully logged.
1400 The Test module has been significantly enhanced.
1404 Time::Local::timelocal() does not handle fractional seconds anymore.
1405 The rationale is that neither does localtime(), and timelocal() and
1406 localtime() are supposed to be inverses of each other.
1410 The vars pragma now supports declaring fully qualified variables.
1411 (Something that C<our()> does not and will not support.)
1415 The C<utf8::> name space (as in the pragma) provides various
1416 Perl-callable functions to provide low level access to Perl's
1417 internal Unicode representation. At the moment only length()
1418 has been implemented.
1422 =head1 Utility Changes
1428 Emacs perl mode (emacs/cperl-mode.el) has been updated to version
1433 F<emacs/e2ctags.pl> is now much faster.
1437 C<enc2xs> is a tool for people adding their own encodings to the
1442 C<h2ph> now supports C trigraphs.
1446 C<h2xs> now produces a template README.
1450 C<h2xs> now uses C<Devel::PPPort> for better portability between
1451 different versions of Perl.
1455 C<h2xs> uses the new L<ExtUtils::Constant|ExtUtils::Constant> module
1456 which will affect newly created extensions that define constants.
1457 Since the new code is more correct (if you have two constants where the
1458 first one is a prefix of the second one, the first constant B<never>
1459 got defined), less lossy (it uses integers for integer constant,
1460 as opposed to the old code that used floating point numbers even for
1461 integer constants), and slightly faster, you might want to consider
1462 regenerating your extension code (the new scheme makes regenerating
1463 easy). L<h2xs> now also supports C trigraphs.
1467 C<libnetcfg> has been added to configure libnet.
1471 C<perlbug> is now much more robust. It also sends the bug report to
1472 perl.org, not perl.com.
1476 C<perlcc> has been rewritten and its user interface (that is,
1477 command line) is much more like that of the UNIX C compiler, cc.
1478 (The perlbc tools has been removed. Use C<perlcc -B> instead.)
1479 B<Note that perlcc is still considered very experimental and
1484 C<perlivp> is a new Installation Verification Procedure utility
1485 for running any time after installing Perl.
1489 C<piconv> is an implementation of the character conversion utility
1490 C<iconv>, demonstrating the new Encode module.
1494 C<pod2html> now allows specifying a cache directory.
1498 C<pod2html> now produces XHTML 1.0.
1502 C<pod2html> now understands POD written using different line endings
1503 (PC-like CRLF versus UNIX-like LF versus MacClassic-like CR).
1507 C<s2p> has been completely rewritten in Perl. (It is in fact a full
1508 implementation of sed in Perl: you can use the sed functionality by
1509 using the C<psed> utility.)
1513 C<xsubpp> now understands POD documentation embedded in the *.xs
1518 C<xsubpp> now supports the OUT keyword.
1522 =head1 New Documentation
1528 perl56delta details the changes between the 5.005 release and the
1533 perlclib documents the internal replacements for standard C library
1534 functions. (Interesting only for extension writers and Perl core
1539 perldebtut is a Perl debugging tutorial. [561+]
1543 perlebcdic contains considerations for running Perl on EBCDIC
1548 perlintro is a gentle introduction to Perl.
1552 perliol documents the internals of PerlIO with layers.
1556 perlmodstyle is a style guide for writing modules.
1560 perlnewmod tells about writing and submitting a new module. [561+]
1564 perlpacktut is a pack() tutorial.
1568 perlpod has been rewritten to be clearer and to record the best
1569 practices gathered over the years.
1573 perlpodspec is a more formal specification of the pod format,
1574 mainly of interest for writers of pod applications, not to
1575 people writing in pod.
1579 perlretut is a regular expression tutorial. [561+]
1583 perlrequick is a regular expressions quick-start guide.
1584 Yes, much quicker than perlretut. [561]
1588 perltodo has been updated.
1592 perltootc has been renamed as perltooc (to not to conflict
1593 with perltoot in filesystems restricted to "8.3" names).
1597 perluniintro is an introduction to using Unicode in Perl.
1598 (perlunicode is more of a detailed reference and background
1603 perlutil explains the command line utilities packaged with the Perl
1604 distribution. [561+]
1608 The following platform-specific documents are available before
1609 the installation as README.I<platform>, and after the installation
1612 perlaix perlamiga perlapollo perlbeos perlbs2000
1613 perlce perlcygwin perldgux perldos perlepoc perlfreebsd perlhpux
1614 perlhurd perlirix perlmachten perlmacos perlmint perlmpeix
1615 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1616 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1618 These documents usually detail one or more of the following subjects:
1619 configuring, building, testing, installing, and sometimes also using
1620 Perl on the said platform.
1622 Eastern Asian Perl users are now welcomed in their own languages:
1623 README.jp (Japanese), README.ko (Korean), README.cn (simplified
1624 Chinese) and README.tw (traditional Chinese), which are written in
1625 normal pod but encoded in EUC-JP, EUC-KR, EUC-CN and Big5. These
1626 will get installed as
1628 perljp perlko perlcn perltw
1634 The documentation for the POSIX-BC platform is called "BS2000", to avoid
1635 confusion with the Perl POSIX module.
1639 The documentation for the WinCE platform is called perlce (README.ce
1640 in the source code kit), to avoid confusion with the perlwin32
1641 documentation on 8.3-restricted filesystems.
1645 =head1 Performance Enhancements
1651 map() could get pathologically slow when the result list it generates
1652 is larger than the source list. The performance has been improved for
1653 common scenarios. [561]
1657 sort() is also fully reentrant, in the sense that the sort function
1658 can itself call sort(). This did not work reliably in previous
1663 sort() has been changed to use primarily mergesort internally as
1664 opposed to the earlier quicksort. For very small lists this may
1665 result in slightly slower sorting times, but in general the speedup
1666 should be at least 20%. Additional bonuses are that the worst case
1667 behaviour of sort() is now better (in computer science terms it now
1668 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1669 worst-case run time behaviour), and that sort() is now stable
1670 (meaning that elements with identical keys will stay ordered as they
1671 were before the sort). See the C<sort> pragma for information.
1673 The story in more detail: suppose you want to serve yourself a little
1676 @digits = ( 3,1,4,1,5,9 );
1678 A numerical sort of the digits will yield (1,1,3,4,5,9), as expected.
1679 Which C<1> comes first is hard to know, since one C<1> looks pretty
1680 much like any other. You can regard this as totally trivial,
1681 or somewhat profound. However, if you just want to sort the even
1682 digits ahead of the odd ones, then what will
1684 sort { ($a % 2) <=> ($b % 2) } @digits;
1686 yield? The only even digit, C<4>, will come first. But how about
1687 the odd numbers, which all compare equal? With the quicksort algorithm
1688 used to implement Perl 5.6 and earlier, the order of ties is left up
1689 to the sort. So, as you add more and more digits of Pi, the order
1690 in which the sorted even and odd digits appear will change.
1691 and, for sufficiently large slices of Pi, the quicksort algorithm
1692 in Perl 5.8 won't return the same results even if reinvoked with the
1693 same input. The justification for this rests with quicksort's
1694 worst case behavior. If you run
1696 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1698 (something you might approximate if you wanted to merge two sorted
1699 arrays using sort), doubling $N doesn't just double the quicksort time,
1700 it I<quadruples> it. Quicksort has a worst case run time that can
1701 grow like N**2, so-called I<quadratic> behaviour, and it can happen
1702 on patterns that may well arise in normal use. You won't notice this
1703 for small arrays, but you I<will> notice it with larger arrays,
1704 and you may not live long enough for the sort to complete on arrays
1705 of a million elements. So the 5.8 quicksort scrambles large arrays
1706 before sorting them, as a statistical defence against quadratic behaviour.
1707 But that means if you sort the same large array twice, ties may be
1708 broken in different ways.
1710 Because of the unpredictability of tie-breaking order, and the quadratic
1711 worst-case behaviour, quicksort was I<almost> replaced completely with
1712 a stable mergesort. I<Stable> means that ties are broken to preserve
1713 the original order of appearance in the input array. So
1715 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1717 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1718 appear in the output in the same order they appeared in the input.
1719 Mergesort has worst case O(N log N) behaviour, the best value
1720 attainable. And, ironically, this mergesort does particularly
1721 well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N)
1722 in O(N) time. But quicksort was rescued at the last moment because
1723 it is faster than mergesort on certain inputs and platforms.
1724 For example, if you really I<don't> care about the order of even
1725 and odd digits, quicksort will run in O(N) time; it's very good
1726 at sorting many repetitions of a small number of distinct elements.
1727 The quicksort divide and conquer strategy works well on platforms
1728 with relatively small, very fast, caches. Eventually, the problem gets
1729 whittled down to one that fits in the cache, from which point it
1730 benefits from the increased memory speed.
1732 Quicksort was rescued by implementing a sort pragma to control aspects
1733 of the sort. The B<stable> subpragma forces stable behaviour,
1734 regardless of algorithm. The B<_quicksort> and B<_mergesort>
1735 subpragmas are heavy-handed ways to select the underlying implementation.
1736 The leading C<_> is a reminder that these subpragmas may not survive
1737 beyond 5.8. More appropriate mechanisms for selecting the implementation
1738 exist, but they wouldn't have arrived in time to save quicksort.
1742 Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm
1743 ( http://burtleburtle.net/bob/hash/doobs.html ). This algorithm is
1744 reasonably fast while producing a much better spread of values than
1745 the old hashing algorithm (originally by Chris Torek, later tweaked by
1746 Ilya Zakharevich). Hash values output from the algorithm on a hash of
1747 all 3-char printable ASCII keys comes much closer to passing the
1748 DIEHARD random number generation tests. According to perlbench, this
1749 change has not affected the overall speed of Perl.
1753 unshift() should now be noticeably faster.
1757 =head1 Installation and Configuration Improvements
1759 =head2 Generic Improvements
1765 INSTALL now explains how you can configure Perl to use 64-bit
1766 integers even on non-64-bit platforms.
1770 Policy.sh policy change: if you are reusing a Policy.sh file
1771 (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1772 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1773 them will now be changed to the new prefix, /foo/bar. (Previously
1774 only $prefix changed.) If you do not like this new behaviour,
1775 specify prefix, siteprefix, and vendorprefix explicitly.
1779 A new optional location for Perl libraries, otherlibdirs, is available.
1780 It can be used for example for vendor add-ons without disturbing Perl's
1781 own library directories.
1785 In many platforms, the vendor-supplied 'cc' is too stripped-down to
1786 build Perl (basically, 'cc' doesn't do ANSI C). If this seems
1787 to be the case and 'cc' does not seem to be the GNU C compiler
1788 'gcc', an automatic attempt is made to find and use 'gcc' instead.
1792 gcc needs to closely track the operating system release to avoid
1793 build problems. If Configure finds that gcc was built for a different
1794 operating system release than is running, it now gives a clearly visible
1795 warning that there may be trouble ahead.
1799 Since Perl 5.8 is not binary-compatible with previous releases
1800 of Perl, Configure no longer suggests including the 5.005
1805 Configure C<-S> can now run non-interactively. [561]
1809 Configure support for pdp11-style memory models has been removed due
1810 to obsolescence. [561]
1814 configure.gnu now works with options with whitespace in them.
1818 installperl now outputs everything to STDERR.
1822 Because PerlIO is now the default on most platforms, "-perlio" doesn't
1823 get appended to the $Config{archname} (also known as $^O) anymore.
1824 Instead, if you explicitly choose not to use perlio (Configure command
1825 line option -Uuseperlio), you will get "-stdio" appended.
1829 Another change related to the architecture name is that "-64all"
1830 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1831 pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)
1835 In AFS installations, one can configure the root of the AFS to be
1836 somewhere else than the default F</afs> by using the Configure
1837 parameter C<-Dafsroot=/some/where/else>.
1841 APPLLIB_EXP, a lesser-known configuration-time definition, has been
1842 documented. It can be used to prepend site-specific directories
1843 to Perl's default search path (@INC); see INSTALL for information.
1847 The version of Berkeley DB used when the Perl (and, presumably, the
1848 DB_File extension) was built is now available as
1849 C<@Config{qw(db_version_major db_version_minor db_version_patch)}>
1850 from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG
1851 DB_VERSION_PATCH_CFG> from C.
1855 Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM
1856 has been documented in INSTALL.
1860 If you have CPAN access (either network or a local copy such as a
1861 CD-ROM) you can during specify extra modules to Configure to build and
1862 install with Perl using the -Dextras=... option. See INSTALL for
1867 In addition to config.over, a new override file, config.arch, is
1868 available. This file is supposed to be used by hints file writers
1869 for architecture-wide changes (as opposed to config.over which is
1870 for site-wide changes).
1874 If your file system supports symbolic links, you can build Perl outside
1875 of the source directory by
1877 mkdir /tmp/perl/build/directory
1878 cd /tmp/perl/build/directory
1879 sh /path/to/perl/source/Configure -Dmksymlinks ...
1881 This will create in /tmp/perl/build/directory a tree of symbolic links
1882 pointing to files in /path/to/perl/source. The original files are left
1883 unaffected. After Configure has finished, you can just say
1887 and Perl will be built and tested, all in /tmp/perl/build/directory.
1892 For Perl developers, several new make targets for profiling
1893 and debugging have been added; see L<perlhack>.
1899 Use of the F<gprof> tool to profile Perl has been documented in
1900 L<perlhack>. There is a make target called "perl.gprof" for
1901 generating a gprofiled Perl executable.
1905 If you have GCC 3, there is a make target called "perl.gcov" for
1906 creating a gcoved Perl executable for coverage analysis. See
1911 If you are on IRIX or Tru64 platforms, new profiling/debugging options
1912 have been added; see L<perlhack> for more information about pixie and
1919 Guidelines of how to construct minimal Perl installations have
1920 been added to INSTALL.
1924 The Thread extension is now not built at all under ithreads
1925 (C<Configure -Duseithreads>) because it wouldn't work anyway (the
1926 Thread extension requires being Configured with C<-Duse5005threads>).
1928 But note that the Thread.pm interface is now shared by both
1933 The Gconvert macro ($Config{d_Gconvert}) used by perl for stringifying
1934 floating-point numbers is now more picky about using sprintf %.*g
1935 rules for the conversion. Some platforms that used to use gcvt may
1936 now resort to the slower sprintf.
1940 The obsolete method of making a special (e.g., debugging) flavor
1943 make LIBPERL=libperld.a
1945 has been removed. Use -DDEBUGGING instead.
1949 =head2 New Or Improved Platforms
1951 For the list of platforms known to support Perl,
1952 see L<perlport/"Supported Platforms">.
1958 AIX dynamic loading should be now better supported.
1962 AIX should now work better with gcc, threads, and 64-bitness. Also the
1963 long doubles support in AIX should be better now. See L<perlaix>.
1967 AtheOS ( http://www.atheos.cx/ ) is a new platform.
1971 BeOS has been reclaimed.
1975 The DG/UX platform now supports 5.005-style threads.
1980 The DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near
1985 EBCDIC platforms (z/OS (also known as OS/390), POSIX-BC, and VM/ESA)
1986 have been regained. Many test suite tests still fail and the
1987 co-existence of Unicode and EBCDIC isn't quite settled, but the
1988 situation is much better than with Perl 5.6. See L<perlos390>,
1989 L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information.
1993 Building perl with -Duseithreads or -Duse5005threads now works under
1994 HP-UX 10.20 (previously it only worked under 10.30 or later). You will
1995 need a thread library package installed. See README.hpux. [561]
1999 Mac OS Classic is now supported in the mainstream source package
2000 (MacPerl has of course been available since perl 5.004 but now the
2001 source code bases of standard Perl and MacPerl have been synchronised)
2006 Mac OS X (or Darwin) should now be able to build Perl even on HFS+
2007 filesystems. (The case-insensitivity used to confuse the Perl build
2012 NCR MP-RAS is now supported. [561]
2016 All the NetBSD specific patches (except for the installation
2017 specific ones) have been merged back to the main distribution.
2021 NetWare from Novell is now supported. See L<perlnetware>.
2025 NonStop-UX is now supported. [561]
2029 NEC SUPER-UX is now supported.
2033 All the OpenBSD specific patches (except for the installation
2034 specific ones) have been merged back to the main distribution.
2038 Perl has been tested with the GNU pth userlevel thread package
2039 ( http://www.gnu.org/software/pth/pth.html ). All thread tests
2040 of Perl now work, but not without adding some yield()s to the tests,
2041 so while pth (and other userlevel thread implementations) can be
2042 considered to be "working" with Perl ithreads, keep in mind the
2043 possible non-preemptability of the underlying thread implementation.
2047 Stratus VOS is now supported using Perl's native build method
2048 (Configure). This is the recommended method to build Perl on
2049 VOS. The older methods, which build miniperl, are still
2050 available. See L<perlvos>. [561+]
2054 The Amdahl UTS UNIX mainframe platform is now supported. [561]
2058 WinCE is now supported. See L<perlce>.
2062 z/OS (formerly known as OS/390, formerly known as MVS OE) now has
2063 support for dynamic loading. This is not selected by default,
2064 however, you must specify -Dusedl in the arguments of Configure. [561]
2068 =head1 Selected Bug Fixes
2070 Numerous memory leaks and uninitialized memory accesses have been
2071 hunted down. Most importantly, anonymous subs used to leak quite
2078 The autouse pragma didn't work for Multi::Part::Function::Names.
2082 caller() could cause core dumps in certain situations. Carp was
2083 sometimes affected by this problem. In particular, caller() now
2084 returns a subroutine name of C<(unknown)> for subroutines that have
2085 been removed from the symbol table.
2089 chop(@list) in list context returned the characters chopped in
2090 reverse order. This has been reversed to be in the right order. [561]
2094 Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm)
2095 when building the Perl binary. The only exception to this is SunOS 4.x,
2096 which needs them. [561]
2100 The behaviour of non-decimal but numeric string constants such as
2101 "0x23" was platform-dependent: in some platforms that was seen as 35,
2102 in some as 0, in some as a floating point number (don't ask). This
2103 was caused by Perl's using the operating system libraries in a situation
2104 where the result of the string to number conversion is undefined: now
2105 Perl consistently handles such strings as zero in numeric contexts.
2109 The order of DESTROYs has been made more predictable.
2113 Perl 5.6.0 could emit spurious warnings about redefinition of
2114 dl_error() when statically building extensions into perl.
2115 This has been corrected. [561]
2119 L<dprofpp> -R didn't work.
2123 C<*foo{FORMAT}> now works.
2127 Infinity is now recognized as a number.
2131 UNIVERSAL::isa no longer caches methods incorrectly. (This broke
2132 the Tk extension with 5.6.0.) [561]
2136 Lexicals I: lexicals outside an eval "" weren't resolved
2137 correctly inside a subroutine definition inside the eval "" if they
2138 were not already referenced in the top level of the eval""ed code.
2142 Lexicals II: lexicals leaked at file scope into subroutines that
2143 were declared before the lexicals.
2147 Lexical warnings now propagating correctly between scopes
2148 and into C<eval "...">.
2152 C<use warnings qw(FATAL all)> did not work as intended. This has been
2157 warnings::enabled() now reports the state of $^W correctly if the caller
2158 isn't using lexical warnings. [561]
2162 Line renumbering with eval and C<#line> now works. [561]
2166 Fixed numerous memory leaks, especially in eval "".
2170 Localised tied variables no longer leak memory
2173 tie my %tied_hash => 'Tie::StdHash';
2177 # Used to leak memory every time local() was called;
2178 # in a loop, this added up.
2179 local($tied_hash{Foo}) = 1;
2183 Localised hash elements (and %ENV) are correctly unlocalised to not
2184 exist, if they didn't before they were localised.
2188 tie my %tied_hash => 'Tie::StdHash';
2192 # Nothing has set the FOO element so far
2194 { local $tied_hash{FOO} = 'Bar' }
2196 # This used to print, but not now.
2197 print "exists!\n" if exists $tied_hash{FOO};
2199 As a side effect of this fix, tied hash interfaces B<must> define
2200 the EXISTS and DELETE methods.
2204 mkdir() now ignores trailing slashes in the directory name,
2205 as mandated by POSIX.
2209 Some versions of glibc have a broken modfl(). This affects builds
2210 with C<-Duselongdouble>. This version of Perl detects this brokenness
2211 and has a workaround for it. The glibc release 2.2.2 is known to have
2212 fixed the modfl() bug.
2216 Modulus of unsigned numbers now works (4063328477 % 65535 used to
2217 return 27406, instead of 27047). [561]
2221 Some "not a number" warnings introduced in 5.6.0 eliminated to be
2222 more compatible with 5.005. Infinity is now recognised as a number. [561]
2226 Numeric conversions did not recognize changes in the string value
2227 properly in certain circumstances. [561]
2231 Attributes (such as :shared) didn't work with our().
2235 our() variables will not cause bogus "Variable will not stay shared"
2240 "our" variables of the same name declared in two sibling blocks
2241 resulted in bogus warnings about "redeclaration" of the variables.
2242 The problem has been corrected. [561]
2246 pack "Z" now correctly terminates the string with "\0".
2250 Fix password routines which in some shadow password platforms
2251 (e.g. HP-UX) caused getpwent() to return every other entry.
2255 The PERL5OPT environment variable (for passing command line arguments
2256 to Perl) didn't work for more than a single group of options. [561]
2260 PERL5OPT with embedded spaces didn't work.
2264 printf() no longer resets the numeric locale to "C".
2268 C<qw(a\\b)> now parses correctly as C<'a\\b'>: that is, as three
2269 characters, not four. [561]
2273 pos() did not return the correct value within s///ge in earlier
2274 versions. This is now handled correctly. [561]
2278 Printing quads (64-bit integers) with printf/sprintf now works
2279 without the q L ll prefixes (assuming you are on a quad-capable platform).
2283 Regular expressions on references and overloaded scalars now work. [561+]
2287 Right-hand side magic (GMAGIC) could in many cases such as string
2288 concatenation be invoked too many times.
2292 scalar() now forces scalar context even when used in void context.
2296 SOCKS support is now much more robust.
2300 sort() arguments are now compiled in the right wantarray context
2301 (they were accidentally using the context of the sort() itself).
2302 The comparison block is now run in scalar context, and the arguments
2303 to be sorted are always provided list context. [561]
2307 Changed the POSIX character class C<[[:space:]]> to include the (very
2308 rarely used) vertical tab character. Added a new POSIX-ish character
2309 class C<[[:blank:]]> which stands for horizontal whitespace
2310 (currently, the space and the tab).
2314 The tainting behaviour of sprintf() has been rationalized. It does
2315 not taint the result of floating point formats anymore, making the
2316 behaviour consistent with that of string interpolation. [561]
2320 Some cases of inconsistent taint propagation (such as within hash
2321 values) have been fixed.
2325 The RE engine found in Perl 5.6.0 accidentally pessimised certain kinds
2326 of simple pattern matches. These are now handled better. [561]
2330 Regular expression debug output (whether through C<use re 'debug'>
2331 or via C<-Dr>) now looks better. [561]
2335 Multi-line matches like C<"a\nxb\n" =~ /(?!\A)x/m> were flawed. The
2336 bug has been fixed. [561]
2340 Use of $& could trigger a core dump under some situations. This
2341 is now avoided. [561]
2345 The regular expression captured submatches ($1, $2, ...) are now
2346 more consistently unset if the match fails, instead of leaving false
2347 data lying around in them. [561]
2351 readline() on files opened in "slurp" mode could return an extra
2352 "" (blank line) at the end in certain situations. This has been
2357 Autovivification of symbolic references of special variables described
2358 in L<perlvar> (as in C<${$num}>) was accidentally disabled. This works
2363 Sys::Syslog ignored the C<LOG_AUTH> constant.
2367 $AUTOLOAD, sort(), lock(), and spawning subprocesses
2368 in multiple threads simultaneously are now thread-safe.
2372 Tie::Array's SPLICE method was broken.
2376 Allow a read-only string on the left-hand side of a non-modifying tr///.
2380 If C<STDERR> is tied, warnings caused by C<warn> and C<die> now
2381 correctly pass to it.
2385 Several Unicode fixes.
2391 BOMs (byte order marks) at the beginning of Perl files
2392 (scripts, modules) should now be transparently skipped.
2393 UTF-16 and UCS-2 encoded Perl files should now be read correctly.
2397 The character tables have been updated to Unicode 3.2.0.
2401 Comparing with utf8 data does not magically upgrade non-utf8 data
2402 into utf8. (This was a problem for example if you were mixing data
2403 from I/O and Unicode data: your output might have got magically encoded
2408 Generating illegal Unicode code points such as U+FFFE, or the UTF-16
2409 surrogates, now also generates an optional warning.
2413 C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase.
2417 Concatenation with the C<.> operator or via variable interpolation,
2418 C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator,
2419 substitution with C<s///>, single-quoted UTF8, should now work.
2423 The C<tr///> operator now works. Note that the C<tr///CU>
2424 functionality has been removed (but see pack('U0', ...)).
2428 C<eval "v200"> now works.
2432 Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spurious warnings.
2433 This has been corrected. [561]
2437 Zero entries were missing from the Unicode classes such as C<IsDigit>.
2443 Large unsigned numbers (those above 2**31) could sometimes lose their
2444 unsignedness, causing bogus results in arithmetic operations. [561]
2448 The Perl parser has been stress tested using both random input and
2449 Markov chain input and the few found crashes and lockups have been
2454 =head2 Platform Specific Changes and Fixes
2462 Perl now works on post-4.0 BSD/OSes.
2468 Setting C<$0> now works (as much as possible; see L<perlvar> for details).
2474 Numerous updates; currently synchronised with Cygwin 1.3.10.
2478 Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.
2484 EPOC now better supported. See README.epoc. [561]
2490 Perl now works on post-3.0 FreeBSDs.
2496 README.hpux updated; C<Configure -Duse64bitall> now works;
2497 now uses HP-UX malloc instead of Perl malloc.
2503 Numerous compilation flag and hint enhancements; accidental mixing
2504 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
2514 Long doubles should now work (see INSTALL). [561]
2518 Linux previously had problems related to sockaddrlen when using
2519 accept(), recvfrom() (in Perl: recv()), getpeername(), and
2528 Compilation of the standard Perl distribution in Mac OS Classic should
2529 now work if you have the Metrowerks development environment and the
2530 missing Mac-specific toolkit bits. Contact the macperl mailing list
2537 MPE/iX update after Perl 5.6.0. See README.mpeix. [561]
2541 NetBSD/threads: try installing the GNU pth (should be in the
2542 packages collection, or http://www.gnu.org/software/pth/),
2543 and Configure with -Duseithreads.
2549 Perl now works on NetBSD/sparc.
2555 Now works with usethreads (see INSTALL). [561]
2561 64-bitness using the Sun Workshop compiler now works.
2567 The native build method requires at least VOS Release 14.5.0
2568 and GNU C++/GNU Tools 2.0.1 or later. The Perl pack function
2569 now maps overflowed values to +infinity and underflowed values
2574 Tru64 (aka Digital UNIX, aka DEC OSF/1)
2576 The operating system version letter now recorded in $Config{osvers}.
2577 Allow compiling with gcc (previously explicitly forbidden). Compiling
2578 with gcc still not recommended because buggy code results, even with
2585 Fixed various alignment problems that lead into core dumps either
2586 during build or later; no longer dies on math errors at runtime;
2587 now using full quad integers (64 bits), previously was using
2588 only 46 bit integers for speed.
2594 See L</"Socket Extension Dynamic in VMS"> and L</"IEEE-format Floating Point
2595 Default on OpenVMS Alpha"> for important changes not otherwise listed here.
2597 chdir() now works better despite a CRT bug; now works with MULTIPLICITY
2598 (see INSTALL); now works with Perl's malloc.
2600 The tainting of C<%ENV> elements via C<keys> or C<values> was previously
2601 unimplemented. It now works as documented.
2603 The C<waitpid> emulation has been improved. The worst bug (now fixed)
2604 was that a pid of -1 would cause a wildcard search of all processes on
2607 POSIX-style signals are now emulated much better on VMS versions prior
2610 The C<system> function and backticks operator have improved
2611 functionality and better error handling. [561]
2613 File access tests now use current process privileges rather than the
2614 user's default privileges, which could sometimes result in a mismatch
2615 between reported access and actual access. This improvement is only
2616 available on VMS v6.0 and later.
2618 There is a new C<kill> implementation based on C<sys$sigprc> that allows
2619 older VMS systems (pre-7.0) to use C<kill> to send signals rather than
2620 simply force exit. This implementation also allows later systems to
2621 call C<kill> from within a signal handler.
2623 Iterative logical name translations are now limited to 10 iterations in
2624 imitation of SHOW LOGICAL and other OpenVMS facilities.
2634 Signal handling now works better than it used to. It is now implemented
2635 using a Windows message loop, and is therefore less prone to random
2640 fork() emulation is now more robust, but still continues to have a few
2641 esoteric bugs and caveats. See L<perlfork> for details. [561+]
2645 A failed (pseudo)fork now returns undef and sets errno to EAGAIN. [561]
2649 The following modules now work on Windows:
2651 ExtUtils::Embed [561]
2658 IO::File::new_tmpfile() is no longer limited to 32767 invocations
2663 Better chdir() return value for a non-existent directory.
2667 Compiling perl using the 64-bit Platform SDK tools is now supported.
2671 The Win32::SetChildShowWindow() builtin can be used to control the
2672 visibility of windows created by child processes. See L<Win32> for
2677 Non-blocking waits for child processes (or pseudo-processes) are
2678 supported via C<waitpid($pid, &POSIX::WNOHANG)>.
2682 The behavior of system() with multiple arguments has been rationalized.
2683 Each unquoted argument will be automatically quoted to protect whitespace,
2684 and any existing whitespace in the arguments will be preserved. This
2685 improves the portability of system(@args) by avoiding the need for
2686 Windows C<cmd> shell specific quoting in perl programs.
2688 Note that this means that some scripts that may have relied on earlier
2689 buggy behavior may no longer work correctly. For example,
2690 C<system("nmake /nologo", @args)> will now attempt to run the file
2691 C<nmake /nologo> and will fail when such a file isn't found.
2692 On the other hand, perl will now execute code such as
2693 C<system("c:/Program Files/MyApp/foo.exe", @args)> correctly.
2697 The perl header files no longer suppress common warnings from the
2698 Microsoft Visual C++ compiler. This means that additional warnings may
2699 now show up when compiling XS code.
2703 Borland C++ v5.5 is now a supported compiler that can build Perl.
2704 However, the generated binaries continue to be incompatible with those
2705 generated by the other supported compilers (GCC and Visual C++). [561]
2709 Duping socket handles with open(F, ">&MYSOCK") now works under Windows 9x.
2714 Current directory entries in %ENV are now correctly propagated to child
2719 New %ENV entries now propagate to subprocesses. [561]
2723 Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root.
2724 Other bugs in chdir() and Cwd::cwd() have also been fixed. [561]
2728 The makefiles now default to the features enabled in ActiveState ActivePerl
2729 (a popular Win32 binary distribution). [561]
2733 HTML files will now be installed in c:\perl\html instead of
2734 c:\perl\lib\pod\html
2738 REG_EXPAND_SZ keys are now allowed in registry settings used by perl. [561]
2742 Can now send() from all threads, not just the first one. [561]
2746 ExtUtils::MakeMaker now uses $ENV{LIB} to search for libraries. [561]
2750 Less stack reserved per thread so that more threads can run
2751 concurrently. (Still 16M per thread.) [561]
2755 C<< File::Spec->tmpdir() >> now prefers C:/temp over /tmp
2756 (works better when perl is running as service).
2760 Better UNC path handling under ithreads. [561]
2764 wait(), waitpid(), and backticks now return the correct exit status
2765 under Windows 9x. [561]
2769 A socket handle leak in accept() has been fixed. [561]
2775 =head1 New or Changed Diagnostics
2781 The lexical warnings category "deprecated" is no longer a sub-category
2782 of the "syntax" category. It is now a top-level category in its own
2787 All regular expression compilation error messages are now hopefully
2788 easier to understand both because the error message now comes before
2789 the failed regex and because the point of failure is now clearly
2790 marked by a C<E<lt>-- HERE> marker.
2794 The various "opened only for", "on closed", "never opened" warnings
2795 drop the C<main::> prefix for filehandles in the C<main> package,
2796 for example C<STDIN> instead of C<main::STDIN>.
2800 The "Unrecognized escape" warning has been extended to include C<\8>,
2801 C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters.
2805 Two new debugging options have been added: if you have compiled your
2806 Perl with debugging, you can use the -DT [561] and -DR options to trace
2807 tokenising and to add reference counts to displaying variables,
2812 Several debugger fixes: exit code now reflects the script exit code,
2813 condition C<"0"> now treated correctly, the C<d> command now checks
2814 line number, C<$.> no longer gets corrupted, and all debugger output
2815 now goes correctly to the socket if RemotePort is set. [561]
2819 The debugger (perl5db.pl) has been modified to present a more
2820 consistent commands interface, via (CommandSet=580). perl5db.t was
2821 also added to test the changes, and as a placeholder for further tests.
2827 The debugger has a new C<dumpDepth> option to control the maximum
2828 depth to which nested structures are dumped. The C<x> command has
2829 been extended so that C<x N EXPR> dumps out the value of I<EXPR> to a
2830 depth of at most I<N> levels.
2834 The debugger can now show lexical variables if you have the CPAN
2835 module PadWalker installed.
2839 If an attempt to use a (non-blessed) reference as an array index
2840 is made, a warning is given.
2844 C<push @a;> and C<unshift @a;> (with no values to push or unshift)
2845 now give a warning. This may be a problem for generated and evaled
2850 If you try to L<perlfunc/pack> a number less than 0 or larger than 255
2851 using the C<"C"> format you will get an optional warning. Similarly
2852 for the C<"c"> format and a number less than -128 or more than 127.
2856 Certain regex modifiers such as C<(?o)> make sense only if applied to
2857 the entire regex. You will get an optional warning if you try to do
2862 Using arrays or hashes as references (e.g. C<< %foo->{bar} >>
2863 has been deprecated for a while. Now you will get an optional warning.
2867 Using C<sort> in scalar context now issues an optional warning.
2868 This didn't do anything useful, as the sort was not performed.
2872 =head1 Changed Internals
2878 perlapi.pod (a companion to perlguts) now attempts to document the
2883 You can now build a really minimal perl called microperl.
2884 Building microperl does not require even running Configure;
2885 C<make -f Makefile.micro> should be enough. Beware: microperl makes
2886 many assumptions, some of which may be too bold; the resulting
2887 executable may crash or otherwise misbehave in wondrous ways.
2888 For careful hackers only.
2892 Added rsignal(), whichsig(), do_join(), op_clear, op_null,
2893 ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2894 interfaces to the publicised API. For the full list of the available
2895 APIs see L<perlapi>.
2899 Made possible to propagate customised exceptions via croak()ing.
2903 Now xsubs can have attributes just like subs. (Well, at least the
2904 built-in attributes.)
2908 dTHR and djSP have been obsoleted; the former removed (because it's
2909 a no-op) and the latter replaced with dSP.
2913 PERL_OBJECT has been completely removed.
2917 The MAGIC constants (e.g. C<'P'>) have been macrofied
2918 (e.g. C<PERL_MAGIC_TIED>) for better source code readability
2919 and maintainability.
2923 The regex compiler now maintains a structure that identifies nodes in
2924 the compiled bytecode with the corresponding syntactic features of the
2925 original regex expression. The information is attached to the new
2926 C<offsets> member of the C<struct regexp>. See L<perldebguts> for more
2927 complete information.
2931 The C code has been made much more C<gcc -Wall> clean. Some warning
2932 messages still remain in some platforms, so if you are compiling with
2933 gcc you may see some warnings about dubious practices. The warnings
2934 are being worked on.
2938 F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented.
2942 Documentation on how to use the Perl source repository has been added
2943 to F<Porting/repository.pod>.
2947 There are now several profiling make targets.
2951 =head1 Security Vulnerability Closed [561]
2953 (This change was already made in 5.7.0 but bears repeating here.)
2954 (5.7.0 came out before 5.6.1: the development branch 5.7 released
2955 sooner than the maintenance branch 5.6)
2957 A potential security vulnerability in the optional suidperl component
2958 of Perl was identified in August 2000. suidperl is neither built nor
2959 installed by default. As of November 2001 the only known vulnerable
2960 platform is Linux, most likely all Linux distributions. CERT and
2961 various vendors and distributors have been alerted about the vulnerability.
2962 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2963 for more information.
2965 The problem was caused by Perl trying to report a suspected security
2966 exploit attempt using an external program, /bin/mail. On Linux
2967 platforms the /bin/mail program had an undocumented feature which
2968 when combined with suidperl gave access to a root shell, resulting in
2969 a serious compromise instead of reporting the exploit attempt. If you
2970 don't have /bin/mail, or if you have 'safe setuid scripts', or if
2971 suidperl is not installed, you are safe.
2973 The exploit attempt reporting feature has been completely removed from
2974 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2975 from all the Perl 5.7 releases), so that particular vulnerability
2976 isn't there anymore. However, further security vulnerabilities are,
2977 unfortunately, always possible. The suidperl functionality is most
2978 probably going to be removed in Perl 5.10. In any case, suidperl
2979 should only be used by security experts who know exactly what they are
2980 doing and why they are using suidperl instead of some other solution
2981 such as sudo ( see http://www.courtesan.com/sudo/ ).
2985 Several new tests have been added, especially for the F<lib> and F<ext>
2986 subsections. There are now about 65 000 individual tests (spread over
2987 about 700 test scripts), in the regression suite (5.6.1 has about
2988 11700 tests, in 258 test scripts) Many of the new tests are of course
2989 introduced by the new modules, but still in general Perl is now more
2992 Because of the large number of tests, running the regression suite
2993 will take considerably longer time than it used to: expect the suite
2994 to take up to 4-5 times longer to run than in perl 5.6. On a really
2995 fast machine you can hope to finish the suite in about 6-8 minutes
2998 The tests are now reported in a different order than in earlier Perls.
2999 (This happens because the test scripts from under t/lib have been moved
3000 to be closer to the library/extension they are testing.)
3002 =head1 Known Problems
3010 If using the AIX native make command, instead of just "make" issue
3011 "make all". In some setups the former has been known to spuriously
3012 also try to run "make install". Alternatively, you may want to use
3017 In AIX 4.2, Perl extensions that use C++ functions that use statics
3018 may have problems in that the statics are not getting initialized.
3019 In newer AIX releases, this has been solved by linking Perl with
3020 the libC_r library, but unfortunately in AIX 4.2 the said library
3021 has an obscure bug where the various functions related to time
3022 (such as time() and gettimeofday()) return broken values, and
3023 therefore in AIX 4.2 Perl is not linked against libC_r.
3027 vac 5.0.0.0 May Produce Buggy Code For Perl
3029 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
3030 resulting in a few random tests failing when run as part of "make
3031 test", but when the failing tests are run by hand, they succeed.
3032 We suggest upgrading to at least vac version 5.0.1.0, that has been
3033 known to compile Perl correctly. "lslpp -L|grep vac.C" will tell
3034 you the vac version. See README.aix.
3038 If building threaded Perl, you may get compilation warning from pp_sys.c:
3040 "pp_sys.c", line 4651.39: 1506-280 (W) Function argument assignment between types "unsigned char*" and "const void*" is not allowed.
3042 This is harmless; it is caused by the getnetbyaddr() and getnetbyaddr_r()
3043 having slightly different types for their first argument.
3047 =head2 Alpha systems with old gccs fail several tests
3049 If you see op/pack, op/pat, op/regexp, or ext/Storable tests failing
3050 in a Linux/alpha or *BSD/Alpha, it's probably time to upgrade your gcc.
3051 gccs prior to 2.95.3 are definitely not good enough, and gcc 3.1 may
3052 be even better. (RedHat Linux/alpha with gcc 3.1 reported no problems,
3053 as did Linux 2.4.18 with gcc 2.95.4.) (In Tru64, it is preferable to
3054 use the bundled C compiler.)
3058 Perl 5.8.0 doesn't build in AmigaOS. It broke at some point during
3059 the ithreads work and we could not find Amiga experts to unbreak the
3060 problems. Perl 5.6.1 still works for AmigaOS (as does the the 5.7.2
3061 development release).
3065 The following tests fail on 5.8.0 Perl in BeOS Personal 5.03:
3067 t/op/lfs............................FAILED at test 17
3068 t/op/magic..........................FAILED at test 24
3069 ext/POSIX/t/sigaction...............FAILED at test 13
3070 ext/POSIX/t/waitpid.................FAILED at test 1
3072 See L<perlbeos> (README.beos) for more details.
3074 =head2 Cygwin "unable to remap"
3076 For example when building the Tk extension for Cygwin,
3077 you may get an error message saying "unable to remap".
3078 This is known problem with Cygwin, and a workaround is
3079 detailed in here: http://sources.redhat.com/ml/cygwin/2001-12/msg00894.html
3081 =head2 ext/threads/t/libc
3083 If this test fails, it indicates that your libc (C library) is not
3084 threadsafe. This particular test stress tests the localtime() call to
3085 find out whether it is threadsafe. See L<perlthrtut> for more information.
3087 =head2 FreeBSD built with ithreads coredumps reading large directories
3089 This is a known bug in FreeBSD's readdir_r() (see L<perlfreebsd>
3090 (README.freebsd)), which hopefully will be fixed in FreeBSD 4.6.
3092 =head2 FreeBSD Failing locale Test 117 For ISO 8859-15 Locales
3094 The ISO 8859-15 locales may fail the locale test 117 in FreeBSD.
3095 This is caused by the characters \xFF (y with diaeresis) and \xBE
3096 (Y with diaeresis) not behaving correctly when being matched
3097 case-insensitively. Apparently this problem has been fixed in
3098 the latest FreeBSD releases.
3099 ( http://www.freebsd.org/cgi/query-pr.cgi?pr=34308 )
3101 =head2 IRIX fails ext/List/Util/t/shuffle.t
3103 IRIX with MIPSpro 7.3.1.3m compiler may fail the said List::Util test
3104 by dumping core. This seems to be a compiler error since if compiled
3105 with gcc no core dump ensues, and no failures on the said test on any
3108 =head2 Modifying $_ Inside for(..)
3112 works without complaint. It shouldn't. (You should be able to
3113 modify only lvalue elements inside the loops.) You can see the
3114 correct behaviour by replacing the 1..5 with 1, 2, 3, 4, 5.
3116 =head2 mod_perl 1.26 Doesn't Build With Threaded Perl
3118 Use mod_perl 1.27 or higher.
3120 =head2 lib/ftmp-security tests warn 'system possibly insecure'
3122 Don't panic. Read the 'make test' section of INSTALL instead.
3124 =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
3126 If perl is configured with -Duse64bitall, the successful result of the
3127 subtest 10 of lib/posix may arrive before the successful result of the
3128 subtest 9, which confuses the test harness so much that it thinks the
3131 =head2 Linux with glibc 2.2.5 fails t/op/int subtest #6 with -Duse64bitint
3133 This is a known bug in the glibc 2.2.5 with long long integers.
3134 ( http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=65612 )
3136 =head2 Linux With Sfio Fails op/misc Test 48
3140 =head2 libwww-perl (LWP) fails base/date #51
3142 Use libwww-perl 5.65 or later.
3146 Please remember to set your environment variable LC_ALL to "C"
3147 (setenv LC_ALL C) before running "make test" to avoid a lot of
3148 warnings about the broken locales of Mac OS X.
3150 The following tests are known to fail in Mac OS X 10.1.5 because of
3151 buggy (old) implementations of Berkeley DB included in Mac OS X:
3153 Failed Test Stat Wstat Total Fail Failed List of Failed
3154 -------------------------------------------------------------------------
3155 ../ext/DB_File/t/db-btree.t 0 11 ?? ?? % ??
3156 ../ext/DB_File/t/db-recno.t 149 3 2.01% 61 63 65
3158 If you are building on a UFS partition, you will also probably see
3159 t/op/stat.t subtest #9 fail. This is caused by Darwin's UFS not
3160 supporting inode change time.
3162 Also the ext/POSIX/t/posix.t subtest #10 fails but it is skipped for
3163 now because the failure is Apple's fault, not Perl's (blocked signals
3166 If you Configure with ithreads, ext/threads/t/libc.t will fail. Again,
3167 this is not Perl's fault-- the libc of Mac OS X is not threadsafe
3168 (in this particular test, the localtime() call is found to be
3171 =head2 OS/2 Test Failures
3173 The following tests are known to fail on OS/2 (for clarity
3174 only the failures are shown, not the full error messages):
3176 t/io/utf8............................FAILED at test 19
3177 t/op/grent...........................FAILED at test 2
3178 t/op/pwent...........................FAILED at test 1
3179 t/lib/os2_base.......................FAILED at test 13
3180 t/lib/os2_process....................FAILED at test 10
3181 t/lib/os2_process_kid................FAILED at test 10
3182 t/lib/rx_cmprt.......................FAILED at test 16
3183 ext/DB_File/t/db-btree...............FAILED at test 0
3184 ext/DB_File/t/db-hash................FAILED at test 0
3185 ext/DB_File/t/db-recno...............FAILED at test 0
3186 lib/ExtUtils/t/basic.................FAILED at test 14
3187 lib/ExtUtils/t/Constant..............FAILED at test 4
3188 lib/Memoize/t/errors.................FAILED at test 4
3190 =head2 op/sprintf tests 91, 129, and 130
3192 The op/sprintf tests 91, 129, and 130 are known to fail on some platforms.
3193 Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX.
3195 Test 91 is known to fail on QNX6 (nto), because C<sprintf '%e',0>
3196 incorrectly produces C<0.000000e+0> instead of C<0.000000e+00>.
3198 For tests 129 and 130, the failing platforms do not comply with
3199 the ANSI C Standard: lines 19ff on page 134 of ANSI X3.159 1989, to
3200 be exact. (They produce something other than "1" and "-1" when
3201 formatting 0.6 and -0.6 using the printf format "%.0f"; most often,
3202 they produce "0" and "-0".)
3206 In case you are still using Solaris 2.5 (aka SunOS 5.5), you may
3207 experience failures (the test core dumping) in lib/locale.t.
3208 The suggested cure is to upgrade your Solaris.
3210 =head2 Solaris x86 Fails Tests With -Duse64bitint
3212 The following tests are known to fail in Solaris x86 with Perl
3213 configured to use 64 bit integers:
3215 ext/Data/Dumper/t/dumper.............FAILED at test 268
3216 ext/Devel/Peek/Peek..................FAILED at test 7
3218 =head2 SUPER-UX (NEC SX)
3220 The following tests are known to fail on SUPER-UX:
3222 op/64bitint...........................FAILED tests 29-30, 32-33, 35-36
3223 op/arith..............................FAILED tests 128-130
3224 op/pack...............................FAILED tests 25-5625
3225 op/pow................................
3226 op/taint..............................# msgsnd failed
3227 ../ext/IO/lib/IO/t/io_poll............FAILED tests 3-4
3228 ../ext/IPC/SysV/ipcsysv...............FAILED tests 2, 5-6
3229 ../ext/IPC/SysV/t/msg.................FAILED tests 2, 4-6
3230 ../ext/Socket/socketpair..............FAILED tests 12
3231 ../lib/IPC/SysV.......................FAILED tests 2, 5-6
3232 ../lib/warnings.......................FAILED tests 115-116, 118-119
3234 The op/pack failure ("Cannot compress negative numbers at op/pack.t line 126")
3235 is serious but as of yet unsolved. It points at some problems with the
3236 signedness handling of the C compiler, as do the 64bitint, arith, and pow
3237 failures. Most of the rest point at problems with SysV IPC.
3239 =head2 PDL failing some tests
3241 Use PDL 2.3.4 or later.
3243 =head2 Term::ReadKey not working on Win32
3245 Use Term::ReadKey 2.20 or later.
3247 =head2 Failure of Thread (5.005-style) tests
3249 B<Note that support for 5.005-style threading is deprecated,
3250 experimental and practically unsupported. In 5.10, it is expected
3253 The following tests are known to fail due to fundamental problems in
3254 the 5.005 threading implementation. These are not new failures--Perl
3255 5.005_0x has the same bugs, but didn't have these tests.
3257 ../ext/B/t/xref.t 255 65280 14 12 85.71% 3-14
3258 ../ext/List/Util/t/first.t 255 65280 7 4 57.14% 2 5-7
3259 ../lib/English.t 2 512 54 2 3.70% 2-3
3260 ../lib/ExtUtils/t/basic.t 1 256 17 1 5.88% 14
3261 ../lib/FileCache.t 5 1 20.00% 5
3262 ../lib/Filter/Simple/t/data.t 6 3 50.00% 1-3
3263 ../lib/Filter/Simple/t/filter_onl 9 3 33.33% 1-2 5
3264 ../lib/Tie/File/t/31_autodefer.t 255 65280 65 32 49.23% 34-65
3265 ../lib/autouse.t 10 1 10.00% 4
3266 op/flip.t 15 1 6.67% 15
3268 These failures are unlikely to get fixed as 5.005-style threads
3269 are considered fundamentally broken. (Basically what happens is that
3270 competing threads can corrupt shared global state.)
3272 =head2 Timing problems
3274 The following tests may fail intermittently because of timing
3275 problems, for example if the system is heavily loaded.
3278 ext/Time/HiRes/HiRes.t
3280 lib/Memoize/t/expmod_t.t
3281 lib/Memoize/t/speed.t
3283 In case of failure please try running them manually, for example
3285 ./perl -Ilib ext/Time/HiRes/HiRes.t
3287 =head2 Unicode in package/class and subroutine names does not work
3289 One can have Unicode in identifier names, but not in package/class or
3290 subroutine names. While some limited functionality towards this does
3291 exist as of Perl 5.8.0, that is more accidental than designed; use of
3292 Unicode for the said purposes is unsupported.
3294 One reason of this unfinishedness is its (currently) inherent
3295 unportability: since both package names and subroutine names may
3296 need to be mapped to file and directory names, the Unicode capability
3297 of the filesystem becomes important-- and there unfortunately aren't
3306 During Configure, the test
3308 Guessing which symbols your C compiler and preprocessor define...
3310 will probably fail with error messages like
3312 CC-20 cc: ERROR File = try.c, Line = 3
3313 The identifier "bad" is undefined.
3315 bad switch yylook 79bad switch yylook 79bad switch yylook 79bad switch yylook 79#ifdef A29K
3318 CC-65 cc: ERROR File = try.c, Line = 3
3319 A semicolon is expected at this point.
3321 This is caused by a bug in the awk utility of UNICOS/mk. You can ignore
3322 the error, but it does cause a slight problem: you cannot fully
3323 benefit from the h2ph utility (see L<h2ph>) that can be used to
3324 convert C headers to Perl libraries, mainly used to be able to access
3325 from Perl the constants defined using C preprocessor, cpp. Because of
3326 the above error, parts of the converted headers will be invisible.
3327 Luckily, these days the need for h2ph is rare.
3331 If building Perl with interpreter threads (ithreads), the
3332 getgrent(), getgrnam(), and getgrgid() functions cannot return the
3333 list of the group members due to a bug in the multithreaded support of
3334 UNICOS/mk. What this means is that in list context the functions will
3335 return only three values, not four.
3341 There are a few known test failures, see L<perluts> (README.uts).
3343 =head2 VOS (Stratus)
3345 When Perl is built using the native build process on VOS Release
3346 14.5.0 and GNU C++/GNU Tools 2.0.1, all attempted tests either
3347 pass or result in TODO (ignored) failures.
3351 There should be no reported test failures with a default configuration,
3352 though there are a number of tests marked TODO that point to areas
3353 needing further debugging and/or porting work.
3357 In multi-CPU boxes, there are some problems with the I/O buffering:
3358 some output may appear twice.
3360 =head2 XML::Parser not working
3362 Use XML::Parser 2.31 or later.
3364 =head2 z/OS (OS/390)
3366 z/OS has rather many test failures but the situation is actually
3367 better than it was in 5.6.0; it's just that so many new modules and
3368 tests have been added.
3370 Failed Test Stat Wstat Total Fail Failed List of Failed
3371 ---------------------------------------------------------------------------
3372 ../ext/Data/Dumper/t/dumper.t 357 8 2.24% 311 314 325 327
3374 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5
3375 ../ext/Storable/t/downgrade.t 12 3072 169 12 7.10% 14-15 46-47 78-79
3377 ../lib/ExtUtils/t/Constant.t 121 30976 48 48 100.00% 1-48
3378 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9
3379 op/pat.t 910 7 0.77% 665 776 785 832-
3381 op/sprintf.t 224 3 1.34% 98 100 136
3382 op/tr.t 97 5 5.15% 63 71-74
3383 uni/fold.t 780 6 0.77% 61 169 196 661
3386 The failures in dumper.t and downgrade.t are problems in the tests,
3387 those in io_unix and sprintf are problems in the USS (UDP sockets
3388 and printf formats). The pat, tr, and fold failures are genuine Perl
3389 problems caused by EBCDIC (and in the pat and fold cases, combining
3390 that with Unicode). The Constant and Embed are probably problems
3391 in the tests (since they test Perl's ability to build extensions,
3392 and that seems to be working reasonably well.)
3394 =head2 Localising Tied Arrays and Hashes Is Broken
3398 doesn't work as one would expect: the old value is restored
3399 incorrectly. This will be changed in a future release, but we don't
3400 know yet what the new semantics will exactly be. In any case, the
3401 change will break existing code that relies on the current
3402 (ill-defined) semantics, so just avoid doing this in general.
3404 =head2 Self-tying Problems
3406 Self-tying of arrays and hashes is broken in rather deep and
3407 hard-to-fix ways. As a stop-gap measure to avoid people from getting
3408 frustrated at the mysterious results (core dumps, most often), it is
3409 forbidden for now (you will get a fatal error even from an attempt).
3411 A change to self-tying of globs has caused them to be recursively
3412 referenced (see: L<perlobj/"Two-Phased Garbage Collection">). You
3413 will now need an explicit untie to destroy a self-tied glob. This
3414 behaviour may be fixed at a later date.
3416 Self-tying of scalars and IO thingies works.
3418 =head2 Tied/Magical Array/Hash Elements Do Not Autovivify
3420 For normal arrays C<$foo = \$bar[1]> will assign C<undef> to
3421 C<$bar[1]> (assuming that it didn't exist before), but for
3422 tied/magical arrays and hashes such autovivification does not happen
3423 because there is currently no way to catch the reference creation.
3424 The same problem affects slicing over non-existent indices/keys of
3425 a tied/magical array/hash.
3427 =head2 Building Extensions Can Fail Because Of Largefiles
3429 Some extensions like mod_perl are known to have issues with
3430 `largefiles', a change brought by Perl 5.6.0 in which file offsets
3431 default to 64 bits wide, where supported. Modules may fail to compile
3432 at all, or they may compile and work incorrectly. Currently, there
3433 is no good solution for the problem, but Configure now provides
3434 appropriate non-largefile ccflags, ldflags, libswanted, and libs
3435 in the %Config hash (e.g., $Config{ccflags_nolargefiles}) so the
3436 extensions that are having problems can try configuring themselves
3437 without the largefileness. This is admittedly not a clean solution,
3438 and the solution may not even work at all. One potential failure is
3439 whether one can (or, if one can, whether it's a good idea to) link
3440 together at all binaries with different ideas about file offsets;
3441 all this is platform-dependent.
3443 =head2 Unicode Support on EBCDIC Still Spotty
3445 Though mostly working, Unicode support still has problem spots on
3446 EBCDIC platforms. One such known spot are the C<\p{}> and C<\P{}>
3447 regular expression constructs for code points less than 256: the
3448 C<pP> are testing for Unicode code points, not knowing about EBCDIC.
3450 =head2 The Compiler Suite Is Still Very Experimental
3452 The compiler suite is slowly getting better but it continues to be
3453 highly experimental. Use in production environments is discouraged.
3455 =head2 The Long Double Support Is Still Experimental
3457 The ability to configure Perl's numbers to use "long doubles",
3458 floating point numbers of hopefully better accuracy, is still
3459 experimental. The implementations of long doubles are not yet
3460 widespread and the existing implementations are not quite mature
3461 or standardised, therefore trying to support them is a rare
3462 and moving target. The gain of more precision may also be offset
3463 by slowdown in computations (more bits to move around, and the
3464 operations are more likely to be executed by less optimised
3467 =head2 Seen In Perl 5.7 But Gone Now
3469 C<Time::Piece> (previously known as C<Time::Object>) was removed
3470 because it was felt that it didn't have enough value in it to be a
3471 core module. It is still a useful module, though, and is available
3474 Perl 5.8 unfortunately does not build anymore on AmigaOS; this broke
3475 accidentally at some point. Since there are not that many Amiga
3476 developers available, we could not get this fixed and tested in time
3477 for 5.8.0. Perl 5.6.1 still works for AmigaOS (as does the the 5.7.2
3478 development release).
3480 =head1 Reporting Bugs
3482 If you find what you think is a bug, you might check the articles
3483 recently posted to the comp.lang.perl.misc newsgroup and the perl
3484 bug database at http://bugs.perl.org/ . There may also be
3485 information at http://www.perl.com/ , the Perl Home Page.
3487 If you believe you have an unreported bug, please run the B<perlbug>
3488 program included with your release. Be sure to trim your bug down
3489 to a tiny but sufficient test case. Your bug report, along with the
3490 output of C<perl -V>, will be sent off to perlbug@perl.org to be
3491 analysed by the Perl porting team.
3495 The F<Changes> file for exhaustive details on what changed.
3497 The F<INSTALL> file for how to build Perl.
3499 The F<README> file for general stuff.
3501 The F<Artistic> and F<Copying> files for copyright information.
3505 Written by Jarkko Hietaniemi <F<jhi@iki.fi>>.