NAME

perldelta - what is new for perl v5.8.0

DESCRIPTION

This document describes differences between the 5.6.0 release and the 5.8.0 release.

Many of the bug fixes in 5.8.0 were already seen in the 5.6.1 maintenance release since the two releases were kept closely coordinated.

If you are upgrading from Perl 5.005_03, you might also want to read perl56delta.

Highlights In 5.8.0

  • Better Unicode support

  • New Thread Implementation

  • Many New Modules

  • Better Numeric Accuracy

  • Safe Signals

  • More Extensive Regression Testing

Incompatible Changes

64-bit platforms and malloc

If your pointers are 64 bits wide, the Perl malloc is no longer being used because it does not work well with 8-byte pointers. Also, usually the system mallocs on such platforms are much better optimized for such large memory models than the Perl malloc. Some memory-hungry Perl applications like the PDL don't work well with Perl's malloc. Finally, other applications than Perl (like modperl) tend to prefer the system malloc. Such platforms include Alpha and 64-bit HPPA, MIPS, PPC, and Sparc.

AIX Dynaloading

The AIX dynaloading now uses in AIX releases 4.3 and newer the native dlopen interface of AIX instead of the old emulated interface. This change will probably break backward compatibility with compiled modules. The change was made to make Perl more compliant with other applications like modperl which are using the AIX native interface.

Attributes for my variables now handled at run-time.

The my EXPR : ATTRS syntax now applies variable attributes at run-time. (Subroutine and our variables still get attributes applied at compile-time.) See attributes for additional details. In particular, however, this allows variable attributes to be useful for tie interfaces, which was a deficiency of earlier releases. Note that the new semantics doesn't work with the Attribute::Handlers module (as of version 0.76).

Socket Extension Dynamic in VMS

The Socket extension is now dynamically loaded instead of being statically built in. This may or may not be a problem with ancient TCP/IP stacks of VMS: we do not know since we weren't able to test Perl in such configurations.

IEEE-format Floating Point Default on OpenVMS Alpha

Perl now uses IEEE format (T_FLOAT) as the default internal floating point format on OpenVMS Alpha, potentially breaking binary compatibility with external libraries or existing data. G_FLOAT is still available as a configuration option. The default on VAX (D_FLOAT) has not changed.

New Unicode Properties

Unicode scripts are now supported. Scripts are similar to (and superior to) Unicode blocks. The difference between scripts and blocks is that scripts are the glyphs used by a language or a group of languages, while the blocks are more artificial groupings of (mostly) 256 characters based on the Unicode numbering.

In general, scripts are more inclusive, but not universally so. For example, while the script Latin includes all the Latin characters and their various diacritic-adorned versions, it does not include the various punctuation or digits (since they are not solely Latin).

A number of other properties are now supported, including \p{L&}, \p{Any} \p{Assigned}, \p{Unassigned}, \p{Blank} and \p{SpacePerl} (along with their \P{...} versions, of course). See perlunicode for details, and more additions.

The In or Is prefix to names used with the \p{...} and \P{...} are now almost always optional. The only exception is that a In prefix is required to signify a Unicode block when a block name conflicts with a script name. For example, \p{Tibetan} refers to the script, while \p{InTibetan} refers to the block. When there is no name conflict, you can omit the In from the block name (e.g. \p{BraillePatterns}), but to be safe, it's probably best to always use the In).

Perl Parser Stress Tested

The Perl parser has been stress tested using both random input and Markov chain input and the few found crashes and lockups have been fixed.

REF(...) Instead Of SCALAR(...)

A reference to a reference now stringifies as "REF(0x81485ec)" instead of "SCALAR(0x81485ec)" in order to be more consistent with the return value of ref().

pack/unpack D/F recycled

The undocumented pack/unpack template letters D/F have been recycled for better use: now they stand for long double (if supported by the platform) and NV (Perl internal floating point type). (They used to be aliases for f/d, but you never knew that.)

Deprecations

  • The semantics of bless(REF, REF) were unclear and until someone proves it to make some sense, it is forbidden.

  • The obsolete chat2 library that should never have been allowed to escape the laboratory has been decommissioned.

  • The builtin dump() function has probably outlived most of its usefulness. The core-dumping functionality will remain in future available as an explicit call to CORE::dump(), but in future releases the behaviour of an unqualified dump() call may change.

  • The very dusty examples in the eg/ directory have been removed. Suggestions for new shiny examples welcome but the main issue is that the examples need to be documented, tested and (most importantly) maintained.

  • The (bogus) escape sequences \8 and \9 now give an optional warning ("Unrecognized escape passed through"). There is no need to \-escape any \w character.

  • The list of filenames from glob() (or <...>) is now by default sorted alphabetically to be csh-compliant (which is what happened before in most UNIX platforms). (bsd_glob() does still sort platform natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.)

  • Spurious syntax errors generated in certain situations, when glob() caused File::Glob to be loaded for the first time, have been fixed.

  • Although "you shouldn't do that", it was possible to write code that depends on Perl's hashed key order (Data::Dumper does this). The new algorithm "One-at-a-Time" produces a different hashed key order. More details are in "Performance Enhancements".

  • lstat(FILEHANDLE) now gives a warning because the operation makes no sense. In future releases this may become a fatal error.

  • The package; syntax (package without an argument) has been deprecated. Its semantics were never that clear and its implementation even less so. If you have used that feature to disallow all but fully qualified variables, use strict; instead.

  • The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still recognised but now cause fatal errors. The previous behaviour of ignoring them by default and warning if requested was unacceptable since it, in a way, falsely promised that the features could be used.

  • The current user-visible implementation of pseudo-hashes (the weird use of the first array element) is deprecated starting from Perl 5.8.0 and will be removed in Perl 5.10.0, and the feature will be implemented differently. Not only is the current interface rather ugly, but the current implementation slows down normal array and hash use quite noticeably. The fields pragma interface will remain available.

  • The syntaxes @a->[...] and %h->{...} have now been deprecated.

  • After years of trying the suidperl is considered to be too complex to ever be considered truly secure. The suidperl functionality is likely to be removed in a future release.

  • The long deprecated uppercase aliases for the string comparison operators (EQ, NE, LT, LE, GE, GT) have now been removed.

  • The tr///C and tr///U features have been removed and will not return; the interface was a mistake. Sorry about that. For similar functionality, see pack('U0', ...) and pack('C0', ...).

  • Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo (@)". The prototypes are now checked at compile-time for invalid characters. An optional warning is generated ("Illegal character in prototype...") but this may be upgraded to a fatal error in a future release.

Core Enhancements

PerlIO is Now The Default

  • IO is now by default done via PerlIO rather than system's "stdio". PerlIO allows "layers" to be "pushed" onto a file handle to alter the handle's behaviour. Layers can be specified at open time via 3-arg form of open:

       open($fh,'>:crlf :utf8', $path) || ...

    or on already opened handles via extended binmode:

       binmode($fh,':encoding(iso-8859-7)');

    The built-in layers are: unix (low level read/write), stdio (as in previous Perls), perlio (re-implementation of stdio buffering in a portable manner), crlf (does CRLF <=> "\n" translation as on Win32, but available on any platform). A mmap layer may be available if platform supports it (mostly UNIXes).

    Layers to be applied by default may be specified via the 'open' pragma.

    See "Installation and Configuration Improvements" for the effects of PerlIO on your architecture name.

  • File handles can be marked as accepting Perl's internal encoding of Unicode (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" :

       open($fh,">:utf8","Uni.txt");

    Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named for you since it's not UTF-8 what you will be getting but instead UTF-EBCDIC. See perlunicode, utf8, and http://www.unicode.org/unicode/reports/tr16/ for more information. In future releases this naming may change.

  • File handles can translate character encodings from/to Perl's internal Unicode form on read/write via the ":encoding()" layer.

  • File handles can be opened to "in memory" files held in Perl scalars via:

       open($fh,'>', \$variable) || ...
  • Anonymous temporary files are available without need to 'use FileHandle' or other module via

       open($fh,"+>", undef) || ...

    That is a literal undef, not an undefined value.

  • The list form of open is now implemented for pipes (at least on UNIX):

       open($fh,"-|", 'cat', '/etc/motd')

    creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in the child process.

Safe Signals

Perl used to be fragile in that signals arriving at inopportune moments could corrupt Perl's internal state. Now Perl postpones handling of signals until it's safe (between opcodes).

This change may have surprising side effects because signals no longer interrupt Perl instantly. Perl will now first finish whatever it was doing, like finishing an internal operation (like sort()) or an external operation (like an I/O operation), and only then look at any arrived signals (and before starting the next operation). No more corrupt internal state since the current operation is always finished first, but the signal may take more time to get heard.

Unicode Overhaul

Unicode in general should be now much more usable than in Perl 5.6.0 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in regular expressions should work now, Unicode in tr/// should work now, Unicode in I/O should work now.

  • The Unicode Character Database coming with Perl has been upgraded to Unicode 3.1.1. For more information, see http://www.unicode.org/.

  • For developers interested in enhancing Perl's Unicode capabilities: almost all the UCD files are included with the Perl distribution in the lib/unicore subdirectory. The most notable omission, for space considerations, is the Unihan database.

  • The properties \p{Blank} and \p{SpacePerl} have been added. "Blank" is like C isblank(), that is, it contains only "horizontal whitespace" (the space character is, the newline isn't), and the "SpacePerl" is the Unicode equivalent of \s (\p{Space} isn't, since that includes the vertical tabulator character, whereas \s doesn't.)

    See "New Unicode Properties" earlier in this document for additional information on changes with Unicode properties.

Understanding of Numbers

In general a lot of fixing has happened in the area of Perl's understanding of numbers, both integer and floating point. Since in many systems the standard number parsing functions like strtoul() and atof() seem to have bugs, Perl tries to work around their deficiencies. This results hopefully in more accurate numbers.

Perl now tries internally to use integer values in numeric conversions and basic arithmetics (+ - * /) if the arguments are integers, and tries also to keep the results stored internally as integers. This change leads to often slightly faster and always less lossy arithmetics. (Previously Perl always preferred floating point numbers in its math.)

Miscellaneous Changes

  • AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value.

  • perl -d:Module=arg,arg,arg now works (previously one couldn't pass in multiple arguments.)

  • The builtin dump() now gives an optional warning dump() better written as CORE::dump(), meaning that by default dump(...) is resolved as the builtin dump() which dumps core and aborts, not as (possibly) user-defined sub dump. To call the latter, qualify the call as &dump(...). (The whole dump() feature is to considered deprecated, and possibly removed/changed in future releases.)

  • chomp() and chop() have been demoted back to not being overridable because they cannot really be overridden-- the problem is that their prototype cannot be expressed and therefore one really cannot write replacements to override these builtins.

  • END blocks are now run even if you exit/die in a BEGIN block. Internally, the execution of END blocks is now controlled by PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new behaviour for Perl embedders. This will default in 5.10. See perlembed.

  • Formats now support zero-padded decimal fields.

  • Lvalue subroutines can now return undef in list context. However, the lvalue subroutine feature still remains experimental.

  • A lost warning "Can't declare ... dereference in my" has been restored (Perl had it earlier but it became lost in later releases.)

  • A new special regular expression variable has been introduced: $^N, which contains the most-recently closed group (submatch).

  • no Module; now works even if there is no "sub unimport" in the Module.

  • The numerical comparison operators return undef if either operand is a NaN. Previously the behaviour was unspecified.

  • The following builtin functions are now overridable: each(), keys(), pop(), push(), shift(), splice(), unshift().

  • pack() / unpack() now can group template letters with () and then apply repetition/count modifiers on the groups.

  • pack() / unpack() can now process the Perl internal numeric types: IVs, UVs, NVs-- and also long doubles, if supported by the platform. The template letters are j, J, F, and D.

  • pack('U0a*', ...) can now be used to force a string to UTF8.

  • my __PACKAGE__ $obj now works.

  • The printf() and sprintf() now support parameter reordering using the %\d+\$ and *\d+\$ syntaxes. For example

        print "%2\$s %1\$s\n", "foo", "bar";

    will print "bar foo\n". This feature helps in writing internationalised software, and in general when the order of the parameters can vary.

  • prototype(\&) is now available.

  • prototype(\[$@%&]) is now available to implicitly create references (useful for example if you want to emulate the tie() interface).

  • A new command-line option, -t is available. It is the little brother of -T: instead of dieing on taint violations, lexical warnings are given. This is only meant as a temporary debugging aid while securing the code of old legacy applications. This is not a substitute for -T.

  • In other taint news, the exec LIST and system LIST have now been considered too risky (think exec @ARGV: it can start any program with any arguments), and now the said forms cause a warning. You should carefully launder the arguments to guarantee their validity. In future releases of Perl the forms will become fatal errors so consider starting laundering now.

  • If tr/// is just counting characters, it doesn't attempt to modify its target.

  • untie() will now call an UNTIE() hook if it exists. See perltie for details.

  • utime now supports utime undef, undef, @files to change the file timestamps to the current time.

  • The rules for allowing underscores (underbars) in numeric constants have been relaxed and simplified: now you can have an underscore simply between digits.

  • Rather than relying on C's argv[0] (which may not contain a full pathname) where possible $^X is now set by asking the operating system. (eg by reading /proc/self/exe on Linux, /proc/curproc/file on FreeBSD)

Modules and Pragmata

New Modules and Pragmata

  • Attribute::Handlers allows a class to define attribute handlers.

        package MyPack;
        use Attribute::Handlers;
        sub Wolf :ATTR(SCALAR) { print "howl!\n" }
    
        # later, in some package using or inheriting from MyPack...
    
        my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called

    Both variables and routines can have attribute handlers. Handlers can be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific to the exact compilation phase (BEGIN, CHECK, INIT, or END).

  • B::Concise is a new compiler backend for walking the Perl syntax tree, printing concise info about ops, from Stephen McCamant. The output is highly customisable. See B::Concise.

  • Class::ISA for reporting the search path for a class's ISA tree, by Sean Burke, has been added. See Class::ISA.

  • Cwd has now a split personality: if possible, an XS extension is used, (this will hopefully be faster, more secure, and more robust) but if not possible, the familiar Perl implementation is used.

  • Devel::PPPort, originally from Kenneth Albanowski and now maintained by Paul Marquess, has been added. It is primarily used by h2xs to enhance portability of XS modules between different versions of Perl.

  • Digest, frontend module for calculating digests (checksums), from Gisle Aas, has been added. See Digest.

  • Digest::MD5 for calculating MD5 digests (checksums) as defined in RFC 1321, from Gisle Aas, has been added. See Digest::MD5.

        use Digest::MD5 'md5_hex';
    
        $digest = md5_hex("Thirsty Camel");
    
        print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1

    NOTE: the MD5 backward compatibility module is deliberately not included since its further use is discouraged.

  • Encode, by Nick Ing-Simmons, provides a mechanism to translate between different character encodings. Support for Unicode, ISO-8859-*, ASCII, CP*, KOI8-R, and three variants of EBCDIC are compiled in to the module. Several other encodings (like Japanese, Chinese, and MacIntosh encodings) are included and will be loaded at runtime. See Encode.

    Any encoding supported by Encode module is also available to the ":encoding()" layer if PerlIO is used.

  • I18N::Langinfo can be use to query locale information. See I18N::Langinfo.

  • I18N::LangTags has functions for dealing with RFC3066-style language tags, by Sean Burke. See I18N::LangTags.

  • ExtUtils::Constant is a new tool for extension writers for generating XS code to import C header constants, by Nicholas Clark. See ExtUtils::Constant.

  • Filter::Simple is an easy-to-use frontend to Filter::Util::Call, from Damian Conway. See Filter::Simple.

        # in MyFilter.pm:
    
        package MyFilter;
    
        use Filter::Simple sub {
            while (my ($from, $to) = splice @_, 0, 2) {
                    s/$from/$to/g;
            }
        };
    
        1;
    
        # in user's code:
    
        use MyFilter qr/red/ => 'green';
    
        print "red\n";   # this code is filtered, will print "green\n"
        print "bored\n"; # this code is filtered, will print "bogreen\n"
    
        no MyFilter;
    
        print "red\n";   # this code is not filtered, will print "red\n"
  • File::Temp allows one to create temporary files and directories in an easy, portable, and secure way, by Tim Jenness. See File::Temp.

  • Filter::Util::Call provides you with the framework to write Source Filters in Perl, from Paul Marquess. For most uses the frontend Filter::Simple is to be preferred. See Filter::Util::Call.

  • if is a new pragma for conditional inclusion of modules, from Ilya Zakharevich.

  • libnet is a collection of perl5 modules related to network programming, from Graham Barr. See Net::FTP, Net::NNTP, Net::Ping, Net::POP3, Net::SMTP, and Net::Time.

    Perl installation leaves libnet unconfigured, use libnetcfg to configure.

  • List::Util is a selection of general-utility list subroutines, like sum(), min(), first(), and shuffle(), by Graham Barr. See List::Util.

  • Locale::Constants, Locale::Country, Locale::Currency, and Locale::Language, from Neil Bowers, have been added. They provide the codes for various locale standards, such as "fr" for France, "usd" for US Dollar, and "jp" for Japanese.

        use Locale::Country;
    
        $country = code2country('jp');               # $country gets 'Japan'
        $code    = country2code('Norway');           # $code gets 'no'

    See Locale::Constants, Locale::Country, Locale::Currency, and Locale::Language.

  • Locale::Maketext is localization framework from Sean Burke. See Locale::Maketext, and Locale::Maketext::TPJ13. The latter is an article about software localization, originally published in The Perl Journal #13, republished here with kind permission.

  • Memoize can make your functions faster by trading space for time, from Mark-Jason Dominus. See Memoize.

  • MIME::Base64 allows you to encode data in base64, from Gisle Aas, as defined in RFC 2045 - MIME (Multipurpose Internet Mail Extensions).

        use MIME::Base64;
    
        $encoded = encode_base64('Aladdin:open sesame');
        $decoded = decode_base64($encoded);
    
        print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="

    See MIME::Base64.

  • MIME::QuotedPrint allows you to encode data in quoted-printable encoding, as defined in RFC 2045 - MIME (Multipurpose Internet Mail Extensions), from Gisle Aas.

        use MIME::QuotedPrint;
    
        $encoded = encode_qp("Smiley in Unicode: \x{263a}");
        $decoded = decode_qp($encoded);
    
        print $encoded, "\n"; # "Smiley in Unicode: =263A"

    MIME::QuotedPrint has been enhanced to provide the basic methods necessary to use it with PerlIO::Via as in :

        use MIME::QuotedPrint;
        open($fh,">Via(MIME::QuotedPrint)",$path);

    See MIME::QuotedPrint.

  • NEXT is pseudo-class for method redispatch, from Damian Conway. See NEXT.

  • open is a new pragma for setting the default I/O disciplines for open().

  • PerlIO::Scalar provides the implementation of IO to "in memory" Perl scalars as discussed above, from Nick Ing-Simmons. It also serves as an example of a loadable PerlIO layer. Other future possibilities include PerlIO::Array and PerlIO::Code. See PerlIO::Scalar.

  • PerlIO::Via acts as a PerlIO layer and wraps PerlIO layer functionality provided by a class (typically implemented in perl code), from Nick Ing-Simmons.

        use MIME::QuotedPrint;
        open($fh,">Via(MIME::QuotedPrint)",$path);

    This will automatically convert everything output to $fh to Quoted-Printable. See PerlIO::Via.

  • Pod::ParseLink, by Russ Allbery, has been added, to parse L<> links in pods as described in the new perlpodspec.

  • Pod::Text::Overstrike, by Joe Smith, has been added. It converts POD data to formatted overstrike text. See Pod::Text::Overstrike.

  • Scalar::Util is a selection of general-utility scalar subroutines, like blessed(), reftype(), and tainted(). See Scalar::Util.

  • sort is a new pragma for controlling the behaviour of sort().

  • Storable gives persistence to Perl data structures by allowing the storage and retrieval of Perl data to and from files in a fast and compact binary format, from Raphael Manfredi. See Storable.

  • Switch, from Damian Conway, has been added. Just by saying

        use Switch;

    you have switch and case available in Perl.

        use Switch;
    
        switch ($val) {
    
                    case 1          { print "number 1" }
                    case "a"        { print "string a" }
                    case [1..10,42] { print "number in list" }
                    case (@array)   { print "number in list" }
                    case /\w+/      { print "pattern" }
                    case qr/\w+/    { print "pattern" }
                    case (%hash)    { print "entry in hash" }
                    case (\%hash)   { print "entry in hash" }
                    case (\&sub)    { print "arg to subroutine" }
                    else            { print "previous case not true" }
        }

    See Switch.

  • Test::More is yet another framework for writing test scripts, more extensive than Test::Simple, by Michael Schwern. See Test::More.

  • Test::Simple has basic utilities for writing tests, by Michael Schwern. See Test::Simple.

  • Text::Balanced has been added, for extracting delimited text sequences from strings, from Damian Conway.

        use Text::Balanced 'extract_delimited';
    
        ($a, $b) = extract_delimited("'never say never', he never said", "'", '');

    $a will be "'never say never'", $b will be ', he never said'.

    In addition to extract_delimited() there are also extract_bracketed(), extract_quotelike(), extract_codeblock(), extract_variable(), extract_tagged(), extract_multiple(), gen_delimited_pat(), and gen_extract_tagged(). With these you can implement rather advanced parsing algorithms. See Text::Balanced.

  • threads is an interface to interpreter threads, by Arthur Bergman. Interpreter threads (ithreads) is the new thread model introduced in Perl 5.6 but only available as an internal interface for extension writers (and for Win32 Perl for fork() emulation). See threads.

  • threads::shared allows data sharing for interpreter threads, from Arthur Bergman. In the ithreads model any data sharing between threads must be explicit, as opposed to the old 5.005 thread model where data sharing was implicit. See threads::shared.

  • Tie::File, by Mark-Jason Dominus, associates a Perl array with the lines of a file.

  • Tie::Memoize, by Ilya Zakharevich, provides on-demand loaded hashes.

  • Tie::RefHash::Nestable, by Edward Avis, allows storing hash references (unlike the standard Tie::RefHash) The module is contained within Tie::RefHash, see Tie::RefHash.

  • Time::HiRes provides high resolution timing (ualarm, usleep, and gettimeofday), from Douglas E. Wegscheid. See Time::HiRes.

  • Unicode::UCD offers a querying interface to the Unicode Character Database. See Unicode::UCD.

  • Unicode::Collate implements the UCA (Unicode Collation Algorithm) for sorting Unicode strings, by SADAHIRO Tomoyuki. See Unicode::Collate.

  • Unicode::Normalize implements the various Unicode normalization forms, by SADAHIRO Tomoyuki. See Unicode::Normalize.

  • XS::Typemap, by Tim Jenness, is a test extension that exercises XS typemaps. Nothing gets installed but for extension writers the code is worth studying.

Updated And Improved Modules and Pragmata

  • The following independently supported modules have been updated to the newest versions from CPAN: CGI, CPAN, DB_File, File::Spec, File::Temp, Getopt::Long, Math::BigFloat, Math::BigInt, the podlators bundle (Pod::Man, Pod::Text), Pod::LaTeX, Pod::Parser, Storable, Term::ANSIColor, Test, Text-Tabs+Wrap.

  • The attributes::reftype() now works on tied arguments.

  • AutoLoader can now be disabled with no AutoLoader;.

  • B::Deparse has been significantly enhanced. It now can deparse almost all of the standard test suite (so that the tests still succeed). There is a make target "test.deparse" for trying this out.

  • Class::Struct can now define the classes in compile time.

  • Class::Struct now assigns the array/hash element if the accessor is called with an array/hash element as the sole argument.

  • Data::Dumper has now an option to sort hashes.

  • Data::Dumper has now an option to dump code references using B::Deparse.

  • DB_File now supports newer Berkeley DB versions, among other improvements.

  • The English module can now be used without the infamous performance hit by saying

            use English '-no_match_vars';

    (Assuming, of course, that one doesn't need the troublesome variables $`, $&, or $'.) Also, introduced @LAST_MATCH_START and @LAST_MATCH_END English aliases for @- and @+.

  • Fcntl, Socket, and Sys::Syslog have been rewritten to use the new-style constant dispatch section (see ExtUtils::Constant). This means that they will be more robust and hopefully faster.

  • File::Find now chdir()s correctly when chasing symbolic links.

  • File::Find now has pre- and post-processing callbacks. It also correctly changes directories when chasing symbolic links. Callbacks (naughtily) exiting with "next;" instead of "return;" now work.

  • File::Find is now (again) reentrant. It also has been made more portable.

  • File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid prototype mismatch with CORE::glob().

  • File::Glob now supports GLOB_LIMIT constant to limit the size of the returned list of filenames.

  • Devel::Peek now has an interface for the Perl memory statistics (this works only if you are using perl's malloc, and if you have compiled with debugging).

  • IPC::Open3 now allows the use of numeric file descriptors.

  • IO::Socket has now atmark() method, which returns true if the socket is positioned at the out-of-band mark. The method is also exportable as a sockatmark() function.

  • IO::Socket::INET has support for ReusePort option (if your platform supports it). The Reuse option now has an alias, ReuseAddr. For clarity you may want to prefer ReuseAddr.

  • IO::Socket::INET now supports LocalPort of zero (usually meaning that the operating system will make one up.)

  • use lib now works identically to @INC. Removing directories with 'no lib' now works.

  • ExtUtils::MakeMaker now uses File::Spec internally, which hopefully leads into better portability.

  • Math::BigFloat and Math::BigInt have undergone a full rewrite. They are now magnitudes faster, and they support various bignum libraries such as GMP and PARI as their backends.

  • Math::Complex handles inf, NaN etc., better.

  • Net::Ping has been muchly enhanced. Multihoming is now supported. There is now "external" protocol which uses Net::Ping::External module which runs external ping(1) and parses the output. A version of Net::Ping::External is available in CPAN.

  • POSIX::sigaction() is now much more flexible and robust. You can now install coderef handlers, 'DEFAULT', and 'IGNORE' handlers, installing new handlers was not atomic.

  • In Safe the %INC now localised in a Safe compartment so that use/require work.

  • In SDBM_File on dosish platforms, some keys went missing because of lack of support for files with "holes". A workaround for the problem has been added.

  • In Search::Dict one can now have a pre-processing hook for the lines being searched.

  • The Shell module now has an OO interface.

  • The Test module has been significantly enhanced.

  • The vars pragma now supports declaring fully qualified variables. (Something that our() does not and will not support.)

  • The utf8:: name space (as in the pragma) provides various Perl-callable functions to provide low level access to Perl's internal Unicode representation. At the moment only length() has been implemented.

Utility Changes

  • Emacs perl mode (emacs/cperl-mode.el) has been updated to version 4.31.

  • emacs/e2ctags.pl is now much faster.

  • h2ph now supports C trigraphs.

  • h2xs now produces a template README.

  • h2xs now uses Devel::PPort for better portability between different versions of Perl.

  • h2xs uses the new ExtUtils::Constant module which will affect newly created extensions that define constants. Since the new code is more correct (if you have two constants where the first one is a prefix of the second one, the first constant never gets defined), less lossy (it uses integers for integer constant, as opposed to the old code that used floating point numbers even for integer constants), and slightly faster, you might want to consider regenerating your extension code (the new scheme makes regenerating easy). h2xs now also supports C trigraphs.

  • libnetcfg has been added to configure the libnet.

  • perlbug is now much more robust. It also sends the bug report to perl.org, not perl.com.

  • perlcc has been rewritten and its user interface (that is, command line) is much more like that of the UNIX C compiler, cc. (The perlbc tools has been removed. Use perlcc -B instead.)

  • perlivp is a new Installation Verification Procedure utility for running any time after installing Perl.

  • pod2html now allows specifying a cache directory.

  • s2p has been completely rewritten in Perl. (It is in fact a full implementation of sed in Perl: you can use the sed functionality by using the psed utility.)

  • xsubpp now understands POD documentation embedded in the *.xs files.

  • xsubpp now supports OUT keyword.

New Documentation

  • perl56delta details the changes between the 5.005 release and the 5.6.0 release.

  • perlclib documents the internal replacements for standard C library functions. (Interesting only for extension writers and Perl core hackers.)

  • perldebtut is a Perl debugging tutorial.

  • perlebcdic contains considerations for running Perl on EBCDIC platforms.

  • perlintro is a gentle introduction to Perl.

  • perliol documents the internals of PerlIO with layers.

  • perlmodstyle is a style guide for writing modules.

  • perlnewmod tells about writing and submitting a new module.

  • perlpacktut is a pack() tutorial.

  • perlpod has been rewritten to be clearer and to record the best practices gathered over the years.

  • perlpodspec is a more formal specification of the pod format, mainly of interest for writers of pod applications, not to people writing in pod.

  • perlretut is a regular expression tutorial.

  • perlrequick is a regular expressions quick-start guide. Yes, much quicker than perlretut.

  • perltodo has been updated.

  • perltootc has been renamed as perltooc (to not to conflict with perltoot in filesystems restricted to "8.3" names)

  • perluniintro is an introduction to using Unicode in Perl. (perlunicode is more of a detailed reference and background information)

  • perlutil explains the command line utilities packaged with the Perl distribution.

The following platform-specific documents are available before the installation as README.platform, and after the installation as perlplatform:

    perlaix perlamiga perlapollo perlbeos perlbs2000
    perlce perlcygwin perldgux perldos perlepoc perlhpux
    perlhurd perlmachten perlmacos perlmint perlmpeix
    perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
    perltru64 perluts perlvmesa perlvms perlvos perlwin32
  • The documentation for the POSIX-BC platform is called "BS2000", to avoid confusion with the Perl POSIX module.

  • The documentation for the WinCE platform is called "CE", to avoid confusion with the perlwin32 documentation on 8.3-restricted filesystems.

Performance Enhancements

  • map() could get pathologically slow when the result list it generates is larger than the source list. The performance has been improved for common scenarios.

  • sort() has been changed to use primarily mergesort internally as opposed to the earlier quicksort. For very small lists this may result in slightly slower sorting times, but in general the speedup should be at least 20%. Additional bonuses are that the worst case behaviour of sort() is now better (in computer science terms it now runs in time O(N log N), as opposed to quicksort's Theta(N**2) worst-case run time behaviour), and that sort() is now stable (meaning that elements with identical keys will stay ordered as they were before the sort). See the sort pragma for information.

    The story in more detail: suppose you want to serve yourself a little slice of Pi.

        @digits = ( 3,1,4,1,5,9 );

    A numerical sort of the digits will yield (1,1,3,4,5,9), as expected. Which 1 comes first is hard to know, since one 1 looks pretty much like any other. You can regard this as totally trivial, or somewhat profound. However, if you just want to sort the even digits ahead of the odd ones, then what will

        sort { ($a % 2) <=> ($b % 2) } @digits;

    yield? The only even digit, 4, will come first. But how about the odd numbers, which all compare equal? With the quicksort algorithm used to implement Perl 5.6 and earlier, the order of ties is left up to the sort. So, as you add more and more digits of Pi, the order in which the sorted even and odd digits appear will change. and, for sufficiently large slices of Pi, the quicksort algorithm in Perl 5.8 won't return the same results even if reinvoked with the same input. The justification for this rests with quicksort's worst case behavior. If you run

       sort { $a <=> $b } ( 1 .. $N , 1 .. $N );

    (something you might approximate if you wanted to merge two sorted arrays using sort), doubling $N doesn't just double the quicksort time, it quadruples it. Quicksort has a worst case run time that can grow like N**2, so-called quadratic behaviour, and it can happen on patterns that may well arise in normal use. You won't notice this for small arrays, but you will notice it with larger arrays, and you may not live long enough for the sort to complete on arrays of a million elements. So the 5.8 quicksort scrambles large arrays before sorting them, as a statistical defence against quadratic behaviour. But that means if you sort the same large array twice, ties may be broken in different ways.

    Because of the unpredictability of tie-breaking order, and the quadratic worst-case behaviour, quicksort was almost replaced completely with a stable mergesort. Stable means that ties are broken to preserve the original order of appearance in the input array. So

        sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);

    will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers appear in the output in the same order they appeared in the input. Mergesort has worst case O(NlogN) behaviour, the best value attainable. And, ironically, this mergesort does particularly well where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N) in O(N) time. But quicksort was rescued at the last moment because it is faster than mergesort on certain inputs and platforms. For example, if you really don't care about the order of even and odd digits, quicksort will run in O(N) time; it's very good at sorting many repetitions of a small number of distinct elements. The quicksort divide and conquer strategy works well on platforms with relatively small, very fast, caches. Eventually, the problem gets whittled down to one that fits in the cache, from which point it benefits from the increased memory speed.

    Quicksort was rescued by implementing a sort pragma to control aspects of the sort. The stable subpragma forces stable behaviour, regardless of algorithm. The _quicksort and _mergesort subpragmas are heavy-handed ways to select the underlying implementation. The leading _ is a reminder that these subpragmas may not survive beyond 5.8. More appropriate mechanisms for selecting the implementation exist, but they wouldn't have arrived in time to save quicksort.

  • Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm (http://burtleburtle.net/bob/hash/doobs.html). This algorithm is reasonably fast while producing a much better spread of values than the old hashing algorithm (originally by Chris Torek, later tweaked by Ilya Zakharevich). Hash values output from the algorithm on a hash of all 3-char printable ASCII keys comes much closer to passing the DIEHARD random number generation tests. According to perlbench, this change has not affected the overall speed of Perl.

  • unshift() should now be noticeably faster.

Installation and Configuration Improvements

Generic Improvements

  • INSTALL now explains how you can configure Perl to use 64-bit integers even on non-64-bit platforms.

  • Policy.sh policy change: if you are reusing a Policy.sh file (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of them will now be changed to the new prefix, /foo/bar. (Previously only $prefix changed.) If you do not like this new behaviour, specify prefix, siteprefix, and vendorprefix explicitly.

  • A new optional location for Perl libraries, otherlibdirs, is available. It can be used for example for vendor add-ons without disturbing Perl's own library directories.

  • In many platforms the vendor-supplied 'cc' is too stripped-down to build Perl (basically, 'cc' doesn't do ANSI C). If this seems to be the case and 'cc' does not seem to be the GNU C compiler 'gcc', an automatic attempt is made to find and use 'gcc' instead.

  • gcc needs to closely track the operating system release to avoid build problems. If Configure finds that gcc was built for a different operating system release than is running, it now gives a clearly visible warning that there may be trouble ahead.

  • If binary compatibility with the 5.005 release is not wanted, Configure no longer suggests including the 5.005 modules in @INC.

  • Configure -S can now run non-interactively.

  • Configure support for pdp11-style memory models has been removed due to obsolescence.

  • configure.gnu now works with options with whitespace in them.

  • installperl now outputs everything to STDERR.

  • $Config{byteorder} is now computed dynamically (this is more robust with "fat binaries" where an executable image contains binaries for more than one binary platform.)

  • Because PerlIO is now the default on most platforms, "-perlio" doesn't get appended to the $Config{archname} (also known as $^O) anymore. Instead, if you explicitly choose not to use perlio (Configure command line option -Uuseperlio), you will get "-stdio" appended.

  • Another change related to the architecture name is that "-64all" (-Duse64bitall, or "maximally 64-bit") is appended only if your pointers are 64 bits wide. (To be exact, the use64bitall is ignored.)

  • In AFS installations one can configure the root of the AFS to be somewhere else than the default /afs by using the Configure parameter -Dafsroot=/some/where/else.

  • APPLLIB_EXP, a less-know configuration-time definition, has been documented. It can be used to prepend site-specific directories to Perl's default search path (@INC), see INSTALL for information.

  • The version of Berkeley DB used when the Perl (and, presumably, the DB_File extension) was built is now available as @Config{qw(db_version_major db_version_minor db_version_patch)} from Perl and as DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG DB_VERSION_PATCH_CFG from C.

  • Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM has been documented in INSTALL.

  • If you have CPAN access (either network or a local copy such as a CD-ROM) you can during specify extra modules to Configure to build and install with Perl using the -Dextras=... option. See INSTALL for more details.

  • In addition to config.over a new override file, config.arch, is available. That is supposed to be used by hints file writers for architecture-wide changes (as opposed to config.over which is for site-wide changes).

  • If your file system supports symbolic links you can build Perl outside of the source directory by

            mkdir /tmp/perl/build/directory
            cd /tmp/perl/build/directory
            sh /path/to/perl/source/Configure -Dmksymlinks ...

    This will create in /tmp/perl/build/directory a tree of symbolic links pointing to files in /path/to/perl/source. The original files are left unaffected. After Configure has finished you can just say

            make all test

    and Perl will be built and tested, all in /tmp/perl/build/directory.

  • For Perl developers several new make targets for profiling and debugging have been added, see perlhack.

    • Use of the gprof tool to profile Perl has been documented in perlhack. There is a make target called "perl.gprof" for generating a gprofiled Perl executable.

    • If you have GCC 3, there is a make target called "perl.gcov" for creating a gcoved Perl executable for coverage analysis. See perlhack.

    • If you are on IRIX or Tru64 platforms, new profiling/debugging options have been added, see perlhack for more information about pixie and Third Degree.

  • Guidelines of how to construct minimal Perl installations have been added to INSTALL.

  • The Thread extension is now not built at all under ithreads (Configure -Duseithreads) because it wouldn't work anyway (the Thread extension requires being Configured with -Duse5005threads).

    But note that the Thread.pm interface is now shared by both thread models.

New Or Improved Platforms

For the list of platforms known to support Perl, see "Supported Platforms" in perlport.

  • AIX dynamic loading should be now better supported.

  • AIX should now work better with gcc, threads, and 64-bitness. Also the long doubles support in AIX should be better now. See perlaix.

  • After a long pause, AmigaOS has been verified to be happy with Perl.

  • AtheOS (http://www.atheos.cx/) is a new platform.

  • BeOS has been reclaimed.

  • DG/UX platform now supports the 5.005-style threads. See perldgux.

  • DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near osvers 4.5.2.

  • EBCDIC platforms (z/OS, also known as OS/390, POSIX-BC, and VM/ESA) have been regained. Many test suite tests still fail and the co-existence of Unicode and EBCDIC isn't quite settled, but the situation is much better than with Perl 5.6. See perlos390, perlbs2000 (for POSIX-BC), and perlvmesa for more information.

  • Building perl with -Duseithreads or -Duse5005threads now works under HP-UX 10.20 (previously it only worked under 10.30 or later). You will need a thread library package installed. See README.hpux.

  • MacOS Classic (MacPerl has of course been available since perl 5.004 but now the source code bases of standard Perl and MacPerl have been synchronised)

  • MacOS X (or Darwin) should now be able to build Perl even on HFS+ filesystems. (The case-insensitivity confused the Perl build process.)

  • NCR MP-RAS is now supported.

  • All the NetBSD specific patches (except for the installation specific ones) have been merged back to the main distribution.

  • NetWare from Novell is now supported. See perlnetware.

  • NonStop-UX is now supported.

  • NEC SUPER-UX is now supported.

  • All the OpenBSD specific patches (except for the installation specific ones) have been merged back to the main distribution.

  • Perl has been tested with the GNU pth userlevel thread package ( http://www.gnu.org/software/pth/pth.html ) . All but one thread test worked, and that one failure was because of test results arriving in unexpected order.

  • Amdahl UTS UNIX mainframe platform is now supported.

  • WinCE is now supported. See perlce.

  • z/OS (formerly known as OS/390, formerly known as MVS OE) has now support for dynamic loading. This is not selected by default, however, you must specify -Dusedl in the arguments of Configure.

Selected Bug Fixes

Numerous memory leaks and uninitialized memory accesses have been hunted down. Most importantly anonymous subs used to leak quite a bit.

  • The autouse pragma didn't work for Multi::Part::Function::Names.

  • caller() could cause core dumps in certain situations. Carp was sometimes affected by this problem.

  • chop(@list) in list context returned the characters chopped in reverse order. This has been reversed to be in the right order.

  • Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm) when building the Perl binary. The only exception to this is SunOS 4.x, which needs them.

  • The behaviour of non-decimal but numeric string constants such as "0x23" was platform-dependent: in some platforms that was seen as 35, in some as 0, in some as a floating point number (don't ask). This was caused by Perl using the operating system libraries in a situation where the result of the string to number conversion is undefined: now Perl consistently handles such strings as zero in numeric contexts.

  • The order of DESTROYs has been made more predictable.

  • Several debugger fixes: exit code now reflects the script exit code, condition "0" now treated correctly, the d command now checks line number, the $. no longer gets corrupted, all debugger output now goes correctly to the socket if RemotePort is set.

  • Perl 5.6.0 could emit spurious warnings about redefinition of dl_error() when statically building extensions into perl. This has been corrected.

  • dprofpp -R didn't work.

  • *foo{FORMAT} now works. =item *

    Infinity is now recognized as a number.

  • UNIVERSAL::isa no longer caches methods incorrectly. (This broke the Tk extension with 5.6.0.)

  • Lexicals I: lexicals outside an eval "" weren't resolved correctly inside a subroutine definition inside the eval "" if they were not already referenced in the top level of the eval""ed code.

  • Lexicals II: lexicals leaked at file scope into subroutines that were declared before the lexicals.

  • Lexical warnings now propagating correctly between scopes and into eval "...".

  • use warnings qw(FATAL all) did not work as intended. This has been corrected.

  • warnings::enabled() now reports the state of $^W correctly if the caller isn't using lexical warnings.

  • Line renumbering with eval and #line now works.

  • Fixed numerous memory leaks, especially in eval "".

  • mkdir() now ignores trailing slashes in the directory name, as mandated by POSIX.

  • Some versions of glibc have a broken modfl(). This affects builds with -Duselongdouble. This version of Perl detects this brokenness and has a workaround for it. The glibc release 2.2.2 is known to have fixed the modfl() bug.

  • Modulus of unsigned numbers now works (4063328477 % 65535 used to return 27406, instead of 27047).

  • Some "not a number" warnings introduced in 5.6.0 eliminated to be more compatible with 5.005. Infinity is now recognised as a number.

  • Numeric conversions did not recognize changes in the string value properly in certain circumstances.

  • Attributes (like :shared) didn't work with our().

  • our() variables will not cause "will not stay shared" warnings.

  • "our" variables of the same name declared in two sibling blocks resulted in bogus warnings about "redeclaration" of the variables. The problem has been corrected.

  • pack "Z" now correctly terminates the string with "\0".

  • Fix password routines which in some shadow password platforms (e.g. HP-UX) caused getpwent() to return every other entry.

  • The PERL5OPT environment variable (for passing command line arguments to Perl) didn't work for more than a single group of options.

  • PERL5OPT with embedded spaces didn't work.

  • printf() no longer resets the numeric locale to "C".

  • qw(a\\b) now parses correctly as 'a\\b'.

  • pos() did not return the correct value within s///ge in earlier versions. This is now handled correctly.

  • Printing quads (64-bit integers) with printf/sprintf now works without the q L ll prefixes (assuming you are on a quad-capable platform).

  • Regular expressions on references and overloaded scalars now work.

  • Right-hand side magic (GMAGIC) could in many cases such as string concatenation be invoked too many times.

  • scalar() now forces scalar context even when used in void context.

  • SOCKS support is now much more robust.

  • sort() arguments are now compiled in the right wantarray context (they were accidentally using the context of the sort() itself). The comparison block is now run in scalar context, and the arguments to be sorted are always provided list context.

  • Changed the POSIX character class [[:space:]] to include the (very rarely used) vertical tab character. Added a new POSIX-ish character class [[:blank:]] which stands for horizontal whitespace (currently, the space and the tab).

  • The tainting behaviour of sprintf() has been rationalized. It does not taint the result of floating point formats anymore, making the behaviour consistent with that of string interpolation.

  • Some cases of inconsistent taint propagation (such as within hash values) have been fixed.

  • The RE engine found in Perl 5.6.0 accidentally pessimised certain kinds of simple pattern matches. These are now handled better.

  • Regular expression debug output (whether through use re 'debug' or via -Dr) now looks better.

  • Multi-line matches like "a\nxb\n" =~ /(?!\A)x/m were flawed. The bug has been fixed.

  • Use of $& could trigger a core dump under some situations. This is now avoided.

  • The regular expression captured submatches ($1, $2, ...) are now more consistently unset if the match fails, instead of leaving false data lying around in them.

  • readline() on files opened in "slurp" mode could return an extra "" at the end in certain situations. This has been corrected.

  • Autovivification of symbolic references of special variables described in perlvar (as in ${$num}) was accidentally disabled. This works again now.

  • Sys::Syslog ignored the LOG_AUTH constant.

  • All but the first argument of the IO syswrite() method are now optional.

  • $AUTOLOAD, sort(), lock(), and spawning subprocesses in multiple threads simultaneously are now thread-safe.

  • Tie::ARRAY SPLICE method was broken.

  • Allow read-only string on left hand side of non-modifying tr///.

  • Several Unicode fixes.

    • BOMs (byte order marks) in the beginning of Perl files (scripts, modules) should now be transparently skipped. UTF-16 (UCS-2) encoded Perl files should now be read correctly.

    • The character tables have been updated to Unicode 3.1.1.

    • Comparing with utf8 data does not magically upgrade non-utf8 data into utf8. (This was a problem for example if you were mixing data from I/O and Unicode data: your output might have got magically encoded as UTF-8.)

    • Generating illegal Unicode code points like U+FFFE, or the UTF-16 surrogates, now also generates an optional warning.

    • IsAlnum, IsAlpha, and IsWord now match titlecase.

    • Concatenation with the . operator or via variable interpolation, eq, substr, reverse, quotemeta, the x operator, substitution with s///, single-quoted UTF8, should now work.

    • The tr/// operator now works. Note that the tr///CU functionality has been removed (but see pack('U0', ...)).

    • eval "v200" now works.

    • Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spurious warnings. This has been corrected.

    • Zero entries were missing from the Unicode classes like IsDigit.

  • Large unsigned numbers (those above 2**31) could sometimes lose their unsignedness, causing bogus results in arithmetic operations.

Platform Specific Changes and Fixes

  • BSDI 4.*

    Perl now works on post-4.0 BSD/OSes.

  • All BSDs

    Setting $0 now works (as much as possible; see perlvar for details).

  • Cygwin

    Numerous updates; currently synchronised with Cygwin 1.1.4.

  • Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O.

  • EPOC

    EPOC update after Perl 5.6.0. See README.epoc.

  • FreeBSD 3.*

    Perl now works on post-3.0 FreeBSDs.

  • HP-UX

    README.hpux updated; Configure -Duse64bitall now almost works.

  • IRIX

    Numerous compilation flag and hint enhancements; accidental mixing of 32-bit and 64-bit libraries (a doomed attempt) made much harder.

  • Linux

    • Long doubles should now work (see INSTALL).

    • Linux previously had problems related to sockaddrlen when using accept(), revcfrom() (in Perl: recv()), getpeername(), and getsockname().

  • MacOS Classic

    Compilation of the standard Perl distribution in MacOS Classic should now work if you have the Metrowerks development environment and the missing Mac-specific toolkit bits. Contact the macperl mailing list for details.

  • MPE/iX

    MPE/iX update after Perl 5.6.0. See README.mpeix.

  • NetBSD/sparc

    Perl now works on NetBSD/sparc.

  • OS/2

    Now works with usethreads (see INSTALL).

  • Solaris

    64-bitness using the Sun Workshop compiler now works.

  • Tru64 (aka Digital UNIX, aka DEC OSF/1)

    The operating system version letter now recorded in $Config{osvers}. Allow compiling with gcc (previously explicitly forbidden). Compiling with gcc still not recommended because buggy code results, even with gcc 2.95.2.

  • Unicos

    Fixed various alignment problems that lead into core dumps either during build or later; no longer dies on math errors at runtime; now using full quad integers (64 bits), previously was using only 46 bit integers for speed.

  • VMS

    chdir() now works better despite a CRT bug; now works with MULTIPLICITY (see INSTALL); now works with Perl's malloc.

    The tainting of %ENV elements via keys or values was previously unimplemented. It now works as documented.

    The waitpid emulation has been improved. The worst bug (now fixed) was that a pid of -1 would cause a wildcard search of all processes on the system. The most significant enhancement is that we can now usually get the completion status of a terminated process.

    POSIX-style signals are now emulated much better on VMS versions prior to 7.0.

    The system function and backticks operator have improved functionality and better error handling.

    File access tests now use current process privileges rather than the user's default privileges, which could sometimes result in a mismatch between reported access and actual access.

  • Windows

    • accept() no longer leaks memory.

    • Borland C++ v5.5 is now a supported compiler that can build Perl. However, the generated binaries continue to be incompatible with those generated by the other supported compilers (GCC and Visual C++).

    • Better chdir() return value for a non-existent directory.

    • Duping socket handles with open(F, ">&MYSOCK") now works under Windows 9x.

    • New %ENV entries now propagate to subprocesses.

    • Current directory entries in %ENV are now correctly propagated to child processes.

    • $ENV{LIB} now used to search for libs under Visual C.

    • fork() emulation has been improved in various ways, but still continues to be experimental. See perlfork for known bugs and caveats.

    • A failed (pseudo)fork now returns undef and sets errno to EAGAIN.

    • Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root. Other bugs in chdir() and Cwd::cwd() have also been fixed.

    • HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html

    • The makefiles now provide a single switch to bulk-enable all the features enabled in ActiveState ActivePerl (a popular Win32 binary distribution).

    • Allow REG_EXPAND_SZ keys in the registry.

    • Can now send() from all threads, not just the first one.

    • Fake signal handling reenabled, bugs and all.

    • %SIG has been enabled under USE_ITHREADS, but its use is completely unsupported under all configurations.

    • Less stack reserved per thread so that more threads can run concurrently. (Still 16M per thread.)

    • File::Spec-&gt;tmpdir() now prefers C:/temp over /tmp (works better when perl is running as service).

    • Better UNC path handling under ithreads.

    • wait(), waitpid() and backticks now return the correct exit status under Windows 9x.

    • winsock handle leak fixed.

New or Changed Diagnostics

  • The lexical warnings category "deprecated" is no longer a sub-category of the "syntax" category. It is now a top-level category in its own right.

  • All regular expression compilation error messages are now hopefully easier to understand both because the error message now comes before the failed regex and because the point of failure is now clearly marked by a <-- HERE marker.

  • The various "opened only for", "on closed", "never opened" warnings drop the main:: prefix for filehandles in the main package, for example STDIN instead of main::STDIN.

  • The "Unrecognized escape" warning has been extended to include \8, \9, and \_. There is no need to escape any of the \w characters.

  • Two new debugging options have been added: if you have compiled your Perl with debugging, you can use the -DT and -DR options to trace tokenising and to add reference counts to displaying variables, respectively.

  • perl5db.pl has been modified to present a more consistent commands interface, via (CommandSet=580). perl5db.t was also added to test the changes, and as a placeholder for further tests.

    See perldebug

  • If an attempt to use a (non-blessed) reference as an array index is made, a warning is given.

  • push @a; and unshift @a; (with no values to push or unshift) now give a warning. This may be a problem for generated and evaled code.

  • If you try to "pack" in perlfunc a number less than 0 or larger than 255 using the "C" format you will get an optional warning. Similarly for the "c" format and a number less than -128 or more than 127.

  • Certain regex modifiers such as (?o) make sense only if applied to the entire regex. You will an optional warning if you try to do otherwise.

  • Using arrays or hashes as references (e.g. %foo-&gt;{bar} has been deprecated for a while. Now you will get an optional warning.

Changed Internals

  • perlapi.pod (a companion to perlguts) now attempts to document the internal API.

  • You can now build a really minimal perl called microperl. Building microperl does not require even running Configure; make -f Makefile.micro should be enough. Beware: microperl makes many assumptions, some of which may be too bold; the resulting executable may crash or otherwise misbehave in wondrous ways. For careful hackers only.

  • Added rsignal(), whichsig(), do_join(), op_clear, op_null, ptr_table_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8 interfaces to the publicised API. For the full list of the available APIs see perlapi.

  • Made possible to propagate customised exceptions via croak()ing.

  • Now xsubs can have attributes just like subs. (Well, at least the built-in attributes.)

  • dTHR and djSP have been obsoleted; the former removed (because it's a no-op) and the latter replaced with dSP.

  • PERL_OBJECT has been completely removed.

  • The MAGIC constants (e.g. 'P') have been macrofied (e.g. PERL_MAGIC_TIED) for better source code readability and maintainability.

  • The regex compiler now maintains a structure that identifies nodes in the compiled bytecode with the corresponding syntactic features of the original regex expression. The information is attached to the new offsets member of the struct regexp. See perldebguts for more complete information.

  • The C code has been made much more gcc -Wall clean. Some warning messages still remain in some platforms, so if you are compiling with gcc you may see some warnings about dubious practices. The warnings are being worked on.

  • perly.c, sv.c, and sv.h have now been extensively commented.

  • Documentation on how to use the Perl source repository has been added to Porting/repository.pod.

  • There are now several profiling make targets.

Security Vulnerability Closed

(This change was already made in 5.7.0 but bears repeating here.)

A potential security vulnerability in the optional suidperl component of Perl was identified in August 2000. suidperl is neither built nor installed by default. As of November 2001 the only known vulnerable platform is Linux, most likely all Linux distributions. CERT and various vendors and distributors have been alerted about the vulnerability. See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt for more information.

The problem was caused by Perl trying to report a suspected security exploit attempt using an external program, /bin/mail. On Linux platforms the /bin/mail program had an undocumented feature which when combined with suidperl gave access to a root shell, resulting in a serious compromise instead of reporting the exploit attempt. If you don't have /bin/mail, or if you have 'safe setuid scripts', or if suidperl is not installed, you are safe.

The exploit attempt reporting feature has been completely removed from Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also from all the Perl 5.7 releases), so that particular vulnerability isn't there anymore. However, further security vulnerabilities are, unfortunately, always possible. The suidperl functionality is most probably going to be removed in Perl 5.10. In any case, suidperl should only be used by security experts who know exactly what they are doing and why they are using suidperl instead of some other solution such as sudo (see http://www.courtesan.com/sudo/).

New Tests

Several new tests have been added, especially for the lib subsection. There are now about 34 000 individual tests (spread over about 530 test scripts), in the regression suite (5.6.1 has about 11700 tests, in 258 test scripts) Many of the new tests are introduced by the new modules, but still in general Perl is now more thoroughly tested.

Because of the large number of tests, running the regression suite will take considerably longer time than it used to: expect the suite to take up to 4-5 times longer to run than in perl 5.6. In a really fast machine you can hope to finish the suite in about 5 minutes (wallclock time).

The tests are now reported in a different order than in earlier Perls. (This happens because the test scripts from under t/lib have been moved to be closer to the library/extension they are testing.)

Known Problems

AIX

  • In AIX 4.2 Perl extensions that use C++ functions that use statics may have problems in that the statics are not getting initialized. In newer AIX releases this has been solved by linking Perl with the libC_r library, but unfortunately in AIX 4.2 the said library has an obscure bug where the various functions related to time (such as time() and gettimeofday()) return broken values, and therefore in AIX 4.2 Perl is not linked against the libC_r.

  • vac 5.0.0.0 May Produce Buggy Code For Perl

    The AIX C compiler vac version 5.0.0.0 may produce buggy code, resulting in few random tests failing, but when the failing tests are run by hand, they succeed. We suggest upgrading to at least vac version 5.0.1.0, that has been known to compile Perl correctly. "lslpp -L|grep vac.C" will tell you the vac version.

Amiga Perl Invoking Mystery

One cannot call Perl using the volume: syntax, that is, perl -v works, but for example bin:perl -v doesn't. The exact reason isn't known but the current suspect is the ixemul library.

lib/ftmp-security tests warn 'system possibly insecure'

Don't panic. Read INSTALL 'make test' section instead.

Cygwin intermittent failures of lib/Memoize/t/expire_file 11 and 12

The subtests 11 and 12 sometimes fail and sometimes work.

FreeBSD 4.5 fails lib/File/Spec/t/rel2abs2rel.t

lib/File/Spec/t/rel2abs2rel.t tests that "`` works" by running a a perl 1 liner in backticks, using "$^X" as the path to perl. It is failing on FreeBSD 4.5, but only when run as part of make test. This seems to be a kernel problem rather than perl - reading the symlink /proc/curproc/file returns "unknown" rather than the path to perl, and a kernel debugger reveals that variable numfullpathfail2 in /usr/src/sys/kern/vfs_cache.c is being incremented whenever /proc/curproc/file fails to return the perl executable's path.

HP-UX lib/io_multihomed Fails When LP64-Configured

The lib/io_multihomed test may hang in HP-UX if Perl has been configured to be 64-bit. Because other 64-bit platforms do not hang in this test, HP-UX is suspect. All other tests pass in 64-bit HP-UX. The test attempts to create and connect to "multihomed" sockets (sockets which have multiple IP addresses).

HP-UX lib/posix Subtest 9 Fails When LP64-Configured

If perl is configured with -Duse64bitall, the successful result of the subtest 10 of lib/posix may arrive before the successful result of the subtest 9, which confuses the test harness so much that it thinks the subtest 9 failed.

Linux With Sfio Fails op/misc Test 48

No known fix.

Mac OS X

The following tests are known to fail:

 Failed Test                 Stat Wstat Total Fail  Failed  List of Failed
 -------------------------------------------------------------------------
 ../ext/DB_File/t/db-btree.t    0    11    ??   ??       %  ??
 ../ext/DB_File/t/db-recno.t              149    3   2.01%  61 63 65
 ../ext/POSIX/t/posix.t                    31    1   3.23%  10

OS/390

OS/390 has rather many test failures but the situation is actually better than it was in 5.6.0, it's just that so many new modules and tests have been added.

Failed 10/611 test scripts, 98.36% okay. 72/53809 subtests failed, 99.87% okay. Failed Test Stat Wstat Total Fail Failed List of Failed ------------------------------------------------------------------------------- ../ext/B/t/deparse.t 17 1 5.88% 14 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5 ../lib/utf8.t 94 13 13.83% 27 30-31 43 46 73 76 79 82 85 88 91 94 ../lib/Benchmark.t 1 256 159 1 0.63% 75 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9 ../lib/ExtUtils/t/ExtUtils.t 27 19 70.37% 5-23 op/pat.t 858 9 1.05% 242-243 665 776 785 832-834 845 op/sprintf.t 224 3 1.34% 98 100 136 op/tr.t 97 5 5.15% 63 71-74 uni/fold.t 767 8 1.04% 25-26 62 169 196 648 697-698 57 tests and 377 subtests skipped.

op/sprintf tests 129 and 130

The op/sprintf tests 129 and 130 are known to fail on some platforms. Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX. The failing platforms do not comply with the ANSI C Standard, line 19ff on page 134 of ANSI X3.159 1989 to be exact. (They produce something other than "1" and "-1" when formatting 0.6 and -0.6 using the printf format "%.0f", most often they produce "0" and "-0".)

Failure of Thread tests

Note that support for 5.005-style threading remains experimental and practically unsupported.

The following tests are known to fail due to fundamental problems in the 5.005 threading implementation. These are not new failures--Perl 5.005_0x has the same bugs, but didn't have these tests.

  ext/List/Util/t/first         2
  lib/autouse                   4
  ext/Thread/thr5005            19-20

These failures are unlikely to get fixed as the 5.005-style threads are considered fundamentally broken.

UNICOS

  ../ext/Socket/socketpair.t    1   256    45    1   2.22%  12
  ../lib/Math/Trig.t                       26    1   3.85%  25
  ../lib/warnings.t                       460    1   0.22%  425
  io/fs.t                                  36    1   2.78%  31
  op/numconvert.t                        1440   13   0.90%  208 509-510
  657-658 665-666 829-830 989-990 1149-1150

UNICOS and UNICOS/mk

The io/fs test #31 is failing because in UNICOS and UNICOS/mk truncate() cannot be used to grow the size of filehandles, only to reduce the size. The workaround is to truncate files instead of filehandles.

UTS

There are a few known test failures, see perluts.

VMS

There should be no reported test failures with a default configuration, though there are a number of tests marked TODO that point to areas needing further debugging and/or porting work.

Win32

In multi-CPU boxes there are some problems with the I/O buffering: some output may appear twice. The Win32 following failures are known as of 5.7.3:

  ..\ext/Encode/t/JP.t      4  1024    22    4  18.18%  9 14 18 21
  ..\ext/threads/t/end.t                6    4  66.67%  3-6
  ..\lib/blib.t             3   768     7    3  42.86%  1 4-5

Localising a Tied Variable Leaks Memory

    use Tie::Hash;
    tie my %tie_hash => 'Tie::StdHash';

    ...

    local($tie_hash{Foo}) = 1; # leaks

Code like the above is known to leak memory every time the local() is executed.

Localising Tied Arrays and Hashes Is Broken

    local %tied_array;

doesn't work as one would expect: the old value is restored incorrectly.

Self-tying of Arrays and Hashes Is Forbidden

Self-tying of arrays and hashes is broken in rather deep and hard-to-fix ways. As a stop-gap measure to avoid people from getting frustrated at the mysterious results (core dumps, most often) it is for now forbidden (you will get a fatal error even from an attempt).

Building Extensions Can Fail Because Of Largefiles

Some extensions like mod_perl are known to have issues with `largefiles', a change brought by Perl 5.6.0 in which file offsets default to 64 bits wide, where supported. Modules may fail to compile at all or compile and work incorrectly. Currently there is no good solution for the problem, but Configure now provides appropriate non-largefile ccflags, ldflags, libswanted, and libs in the %Config hash (e.g., $Config{ccflags_nolargefiles}) so the extensions that are having problems can try configuring themselves without the largefileness. This is admittedly not a clean solution, and the solution may not even work at all. One potential failure is whether one can (or, if one can, whether it's a good idea) link together at all binaries with different ideas about file offsets, all this is platform-dependent.

Unicode Support on EBCDIC Still Spotty

Though mostly working, Unicode support still has problem spots on EBCDIC platforms. One such known spot are the \p{} and \P{} regular expression constructs for code points less than 256: the pP are testing for Unicode code points, not knowing about EBCDIC.

The Compiler Suite Is Still Experimental

The compiler suite is slowly getting better but it continues to be highly experimental. Use in production environments is discouraged.

The Long Double Support Is Still Experimental

The ability to configure Perl's numbers to use "long doubles", floating point numbers of hopefully better accuracy, is still experimental. The implementations of long doubles are not yet widespread and the existing implementations are not quite mature or standardised, therefore trying to support them is a rare and moving target. The gain of more precision may also be offset by slowdown in computations (more bits to move around, and the operations are more likely to be executed by less optimised libraries).

Seen In Perl 5.7 But Gone Now

Time::Piece (previously known as Time::Object) was removed because it was felt that it didn't have enough value in it to be a core module. It is still a useful module, though, and is available from the CPAN.

Reporting Bugs

If you find what you think is a bug, you might check the articles recently posted to the comp.lang.perl.misc newsgroup and the perl bug database at http://bugs.perl.org. There may also be information at http://www.perl.com/, the Perl Home Page.

If you believe you have an unreported bug, please run the perlbug program included with your release. Be sure to trim your bug down to a tiny but sufficient test case. Your bug report, along with the output of perl -V, will be sent off to perlbug@perl.org to be analysed by the Perl porting team.

SEE ALSO

The Changes file for exhaustive details on what changed.

The INSTALL file for how to build Perl.

The README file for general stuff.

The Artistic and Copying files for copyright information.

HISTORY

Written by Jarkko Hietaniemi <jhi@iki.fi>.