3 perldelta - what's new for perl5.004
7 This document describes differences between the 5.003 release (as
8 documented in I<Programming Perl>, second edition--the Camel Book) and
11 =head1 Supported Environments
13 Perl5.004 builds out of the box on Unix, Plan9, LynxOS, VMS, OS/2,
18 Most importantly, many bugs were fixed. See the F<Changes>
19 file in the distribution for details.
21 =head2 Compilation option: Binary compatibility with 5.003
23 There is a new Configure question that asks if you want to maintain
24 binary compatibility with Perl 5.003. If you choose binary
25 compatibility, you do not have to recompile your extensions, but you
26 might have symbol conflicts if you embed Perl in another application,
27 just as in the 5.003 release. By default, binary compatibility
28 is preserved at the expense of symbol table pollution.
30 =head2 $PERL5OPT environment variable
32 You may now put Perl options in the $PERL5OPT environment variable.
33 Unless Perl is running with taint checks, it will interpret this
34 variable as if its contents had appeared on a "#!perl" line at the
35 beginning of your script, except that hyphens are optional. PERL5OPT
36 may only be used to set the following switches: B<-[DIMUdmw]>.
38 =head2 Limitations on B<-M>, and C<-m>, and B<-T> options
40 The C<-M> and C<-m> options are no longer allowed on the C<#!> line of
41 a script. If a script needs a module, it should invoke it with the
44 The B<-T> option is also forbidden on the C<#!> line of a script,
45 unless it was present on the Perl command line. Due to the way C<#!>
46 works, this usually means that B<-T> must be in the first argument.
51 will probably work for an executable script invoked as C<scriptname>,
56 will probably fail under the same conditions. (Non-Unix systems will
57 probably not follow this rule.) But C<perl scriptname> is guaranteed
58 to fail, since then there is no chance of B<-T> being found on the
59 command line before it is found on the C<#!> line.
61 =head2 More precise warnings
63 If you removed the B<-w> option from your Perl 5.003 scripts because it
64 made Perl too verbose, we recommend that you try putting it back when
65 you upgrade to Perl 5.004. Each new perl version tends to remove some
66 undesirable warnings, while adding new warnings that may catch bugs in
69 =head2 Deprecated: Inherited C<AUTOLOAD> for non-methods
71 Before Perl 5.004, C<AUTOLOAD> functions were looked up as methods
72 (using the C<@ISA> hierarchy), even when the function to be autoloaded
73 was called as a plain function (e.g. C<Foo::bar()>), not a method
74 (e.g. C<Foo->bar()> or C<$obj->bar()>).
76 Perl 5.005 will use method lookup only for methods' C<AUTOLOAD>s.
77 However, there is a significant base of existing code that may be using
78 the old behavior. So, as an interim step, Perl 5.004 issues an optional
79 warning when a non-method uses an inherited C<AUTOLOAD>.
81 The simple rule is: Inheritance will not work when autoloading
82 non-methods. The simple fix for old code is: In any module that used to
83 depend on inheriting C<AUTOLOAD> for non-methods from a base class named
84 C<BaseClass>, execute C<*AUTOLOAD = \&BaseClass::AUTOLOAD> during startup.
86 =head2 Subroutine arguments created only when they're modified
88 In Perl 5.004, nonexistent array and hash elements used as subroutine
89 parameters are brought into existence only if they are actually
90 assigned to (via C<@_>).
92 Earlier versions of Perl vary in their handling of such arguments.
93 Perl versions 5.002 and 5.003 always brought them into existence.
94 Perl versions 5.000, 5.001, and 5.002 brought them into existence only
95 if they were not the first argument (which was almost certainly a
96 bug). Earlier versions of Perl never brought them into existence.
98 For example, given this code:
101 sub show { print $_[0] };
102 sub change { $_[0]++ };
106 After this code executes in Perl 5.004, $a{b} exists but $a[2] does
107 not. In Perl 5.002 and 5.003, both $a{b} and $a[2] would have existed
108 (but $a[2]'s value would have been undefined).
110 =head2 Group vector changeable with C<$)>
112 The C<$)> special variable has always (well, in Perl 5, at least)
113 reflected not only the current effective group, but also the group list
114 as returned by the C<getgroups()> C function (if there is one).
115 However, until this release, there has not been a way to call the
116 C<setgroups()> C function from Perl.
118 In Perl 5.004, assigning to C<$)> is exactly symmetrical with examining
119 it: The first number in its string value is used as the effective gid;
120 if there are any numbers after the first one, they are passed to the
121 C<setgroups()> C function (if there is one).
123 =head2 Fixed parsing of $$<digit>, &$<digit>, etc.
125 Perl versions before 5.004 misinterpreted any type marker followed by
126 "$" and a digit. For example, "$$0" was incorrectly taken to mean
127 "${$}0" instead of "${$0}". This bug is (mostly) fixed in Perl 5.004.
129 However, the developers of Perl 5.004 could not fix this bug completely,
130 because at least two widely-used modules depend on the old meaning of
131 "$$0" in a string. So Perl 5.004 still interprets "$$<digit>" in the
132 old (broken) way inside strings; but it generates this message as a
133 warning. And in Perl 5.005, this special treatment will cease.
135 =head2 No resetting of $. on implicit close
137 The documentation for Perl 5.0 has always stated that C<$.> is I<not>
138 reset when an already-open file handle is reopened with no intervening
139 call to C<close>. Due to a bug, perl versions 5.000 through 5.003
140 I<did> reset C<$.> under that circumstance; Perl 5.004 does not.
142 =head2 C<wantarray> may return undef
144 The C<wantarray> operator returns true if a subroutine is expected to
145 return a list, and false otherwise. In Perl 5.004, C<wantarray> can
146 also return the undefined value if a subroutine's return value will
147 not be used at all, which allows subroutines to avoid a time-consuming
148 calculation of a return value if it isn't going to be used.
150 =head2 Changes to tainting checks
152 A bug in previous versions may have failed to detect some insecure
153 conditions when taint checks are turned on. (Taint checks are used
154 in setuid or setgid scripts, or when explicitly turned on with the
155 C<-T> invocation option.) Although it's unlikely, this may cause a
156 previously-working script to now fail -- which should be construed
157 as a blessing, since that indicates a potentially-serious security
158 hole was just plugged.
160 =head2 New Opcode module and revised Safe module
162 A new Opcode module supports the creation, manipulation and
163 application of opcode masks. The revised Safe module has a new API
164 and is implemented using the new Opcode module. Please read the new
165 Opcode and Safe documentation.
167 =head2 Embedding improvements
169 In older versions of Perl it was not possible to create more than one
170 Perl interpreter instance inside a single process without leaking like a
171 sieve and/or crashing. The bugs that caused this behavior have all been
172 fixed. However, you still must take care when embedding Perl in a C
173 program. See the updated perlembed manpage for tips on how to manage
176 =head2 Internal change: FileHandle class based on IO::* classes
178 File handles are now stored internally as type IO::Handle. The
179 FileHandle module is still supported for backwards compatibility, but
180 it is now merely a front end to the IO::* modules -- specifically,
181 IO::Handle, IO::Seekable, and IO::File. We suggest, but do not
182 require, that you use the IO::* modules in new code.
184 In harmony with this change, C<*GLOB{FILEHANDLE}> is now a
185 backward-compatible synonym for C<*STDOUT{IO}>.
187 =head2 Internal change: PerlIO abstraction interface
189 It is now possible to build Perl with AT&T's sfio IO package
190 instead of stdio. See L<perlapio> for more details, and
191 the F<INSTALL> file for how to use it.
193 =head2 New and changed builtin variables
199 Extended error message on some platforms. (Also known as
200 $EXTENDED_OS_ERROR if you C<use English>).
204 The current set of syntax checks enabled by C<use strict>. See the
205 documentation of C<strict> for more details. Not actually new, but
207 Because it is intended for internal use by Perl core components,
208 there is no C<use English> long name for this variable.
212 By default, running out of memory it is not trappable. However, if
213 compiled for this, Perl may use the contents of C<$^M> as an emergency
214 pool after die()ing with this message. Suppose that your Perl were
215 compiled with -DEMERGENCY_SBRK and used Perl's malloc. Then
219 would allocate a 64K buffer for use when in emergency.
220 See the F<INSTALL> file for information on how to enable this option.
221 As a disincentive to casual use of this advanced feature,
222 there is no C<use English> long name for this variable.
226 =head2 New and changed builtin functions
230 =item delete on slices
232 This now works. (e.g. C<delete @ENV{'PATH', 'MANPATH'}>)
236 is now supported on more platforms, prefers fcntl to lockf when
237 emulating, and always flushes before (un)locking.
239 =item printf and sprintf
241 now support "%i" as a synonym for "%d", and the "h" modifier.
242 So "%hi" means "short integer in decimal", and "%ho" means
243 "unsigned short integer as octal".
245 =item keys as an lvalue
247 As an lvalue, C<keys> allows you to increase the number of hash buckets
248 allocated for the given hash. This can gain you a measure of efficiency if
249 you know the hash is going to get big. (This is similar to pre-extending
250 an array by assigning a larger number to $#array.) If you say
254 then C<%hash> will have at least 200 buckets allocated for it. These
255 buckets will be retained even if you do C<%hash = ()>; use C<undef
256 %hash> if you want to free the storage while C<%hash> is still in scope.
257 You can't shrink the number of buckets allocated for the hash using
258 C<keys> in this way (but you needn't worry about doing this by accident,
259 as trying has no effect).
261 =item my() in Control Structures
263 You can now use my() (with or without the parentheses) in the control
264 expressions of control structures such as:
266 while (defined(my $line = <>)) {
272 if ((my $answer = <STDIN>) =~ /^y(es)?$/i) {
274 } elsif ($answer =~ /^n(o)?$/i) {
278 die "`$answer' is neither `yes' nor `no'";
281 Also, you can declare a foreach loop control variable as lexical by
282 preceding it with the word "my". For example, in:
284 foreach my $i (1, 2, 3) {
288 $i is a lexical variable, and the scope of $i extends to the end of
289 the loop, but not beyond it.
291 Note that you still cannot use my() on global punctuation variables
292 such as $_ and the like.
294 =item unpack() and pack()
296 A new format 'w' represents a BER compressed integer (as defined in
297 ASN.1). Its format is a sequence of one or more bytes, each of which
298 provides seven bits of the total value, with the most significant
299 first. Bit eight of each byte is set, except for the last byte, in
300 which bit eight is clear.
304 If the first argument to C<use> is a number, it is treated as a version
305 number instead of a module name. If the version of the Perl interpreter
306 is less than VERSION, then an error message is printed and Perl exits
307 immediately. Because C<use> occurs at compile time, this check happens
308 immediately during the compilation process, unlike C<require VERSION>,
309 which waits until runtime for the check. This is often useful if you
310 need to check the current Perl version before C<use>ing library modules
311 which have changed in incompatible ways from older versions of Perl.
312 (We try not to do this more than we have to.)
314 =item use Module VERSION LIST
316 If the VERSION argument is present between Module and LIST, then the
317 C<use> will call the VERSION method in class Module with the given
318 version as an argument. The default VERSION method, inherited from
319 the UNIVERSAL class, croaks if the given version is larger than the
320 value of the variable $Module::VERSION. (Note that there is not a
321 comma after VERSION!)
323 This version-checking mechanism is similar to the one currently used
324 in the Exporter module, but it is faster and can be used with modules
325 that don't use the Exporter. It is the recommended method for new
328 =item prototype(FUNCTION)
330 Returns the prototype of a function as a string (or C<undef> if the
331 function has no prototype). FUNCTION is a reference to or the name of the
332 function whose prototype you want to retrieve.
333 (Not actually new; just never documented before.)
337 The default seed for C<srand>, which used to be C<time>, has been changed.
338 Now it's a heady mix of difficult-to-predict system-dependent values,
339 which should be sufficient for most everyday purposes.
341 Previous to version 5.004, calling C<rand> without first calling C<srand>
342 would yield the same sequence of random numbers on most or all machines.
343 Now, when perl sees that you're calling C<rand> and haven't yet called
344 C<srand>, it calls C<srand> with the default seed. You should still call
345 C<srand> manually if your code might ever be run on a pre-5.004 system,
346 of course, or if you want a seed other than the default.
350 Functions documented in the Camel to default to $_ now in
351 fact do, and all those that do are so documented in L<perlfunc>.
353 =item C<m//g> does not trigger a pos() reset on failure
355 The C<m//g> match iteration construct used to reset the iteration
356 when it failed to match (so that the next C<m//g> match would start at
357 the beginning of the string). You now have to explicitly do a
358 C<pos $str = 0;> to reset the "last match" position, or modify the
359 string in some way. This change makes it practical to chain C<m//g>
360 matches together in conjunction with ordinary matches using the C<\G>
361 zero-width assertion. See L<perlop> and L<perlre>.
363 =item C<m//x> ignores whitespace before ?*+{}
365 The C<m//x> construct has always been intended to ignore all unescaped
366 whitespace. However, before Perl 5.004, whitespace had the effect of
367 escaping repeat modifiers like "*" or "?"; for example, C</a *b/x> was
368 (mis)interpreted as C</a\*b/x>. This bug has been fixed in 5.004.
370 =item nested C<sub{}> closures work now
372 Prior to the 5.004 release, nested anonymous functions didn't work
375 =item formats work right on changing lexicals
377 Just like anonymous functions that contain lexical variables
378 that change (like a lexical index variable for a C<foreach> loop),
379 formats now work properly. For example, this silently failed
380 before, and is fine now:
383 foreach $i ( 1 .. 10 ) {
393 =head2 New builtin methods
395 The C<UNIVERSAL> package automatically contains the following methods that
396 are inherited by all other classes:
402 C<isa> returns I<true> if its object is blessed into a subclass of C<CLASS>
404 C<isa> is also exportable and can be called as a sub with two arguments. This
405 allows the ability to check what a reference points to. Example:
407 use UNIVERSAL qw(isa);
409 if(isa($ref, 'ARRAY')) {
415 C<can> checks to see if its object has a method called C<METHOD>,
416 if it does then a reference to the sub is returned; if it does not then
417 I<undef> is returned.
419 =item VERSION( [NEED] )
421 C<VERSION> returns the version number of the class (package). If the
422 NEED argument is given then it will check that the current version (as
423 defined by the $VERSION variable in the given package) not less than
424 NEED; it will die if this is not the case. This method is normally
425 called as a class method. This method is called automatically by the
426 C<VERSION> form of C<use>.
428 use A 1.2 qw(some imported subs);
434 B<NOTE:> C<can> directly uses Perl's internal code for method lookup, and
435 C<isa> uses a very similar method and caching strategy. This may cause
436 strange effects if the Perl code dynamically changes @ISA in any package.
438 You may add other methods to the UNIVERSAL class via Perl or XS code.
439 You do not need to C<use UNIVERSAL> in order to make these methods
440 available to your program. This is necessary only if you wish to
441 have C<isa> available as a plain subroutine in the current package.
443 =head2 TIEHANDLE now supported
445 See L<perltie> for other kinds of tie()s.
449 =item TIEHANDLE classname, LIST
451 This is the constructor for the class. That means it is expected to
452 return an object of some sort. The reference can be used to
453 hold some internal information.
458 return bless \$i, shift;
461 =item PRINT this, LIST
463 This method will be triggered every time the tied handle is printed to.
464 Beyond its self reference it also expects the list that was passed to
470 return print join( $, => map {uc} @_), $\;
475 This method will be called when the handle is read from via the C<read>
476 or C<sysread> functions.
480 my($buf,$len,$offset) = @_;
481 print "READ called, \$buf=$buf, \$len=$len, \$offset=$offset";
486 This method will be called when the handle is read from. The method
487 should return undef when there is no more data.
491 return "PRINT called $$r times\n"
496 This method will be called when the C<getc> function is called.
498 sub GETC { print "Don't GETC, Get Perl"; return "a"; }
502 As with the other types of ties, this method will be called when the
503 tied handle is about to be destroyed. This is useful for debugging and
504 possibly for cleaning up.
512 =head2 Malloc enhancements
514 Four new compilation flags are recognized by malloc.c. (They have no
515 effect if perl is compiled with system malloc().)
519 =item -DDEBUGGING_MSTATS
521 If perl is compiled with C<DEBUGGING_MSTATS> defined, you can print
522 memory statistics at runtime by running Perl thusly:
524 env PERL_DEBUG_MSTATS=2 perl your_script_here
526 The value of 2 means to print statistics after compilation and on
527 exit; with a value of 1, the statistics ares printed only on exit.
528 (If you want the statistics at an arbitrary time, you'll need to
529 install the optional module Devel::Peek.)
531 =item -DEMERGENCY_SBRK
533 If this macro is defined, running out of memory need not be a fatal
534 error: a memory pool can allocated by assigning to the special
535 variable C<$^M>. See L<"$^M">.
539 Perl memory allocation is by bucket with sizes close to powers of two.
540 Because of these malloc overhead may be big, especially for data of
541 size exactly a power of two. If C<PACK_MALLOC> is defined, perl uses
542 a slightly different algorithm for small allocations (up to 64 bytes
543 long), which makes it possible to have overhead down to 1 byte for
544 allocations which are powers of two (and appear quite often).
546 Expected memory savings (with 8-byte alignment in C<alignbytes>) is
547 about 20% for typical Perl usage. Expected slowdown due to additional
548 malloc overhead is in fractions of a percent (hard to measure, because
549 of the effect of saved memory on speed).
551 =item -DTWO_POT_OPTIMIZE
553 Similarly to C<PACK_MALLOC>, this macro improves allocations of data
554 with size close to a power of two; but this works for big allocations
555 (starting with 16K by default). Such allocations are typical for big
556 hashes and special-purpose scripts, especially image processing.
558 On recent systems, the fact that perl requires 2M from system for 1M
559 allocation will not affect speed of execution, since the tail of such
560 a chunk is not going to be touched (and thus will not require real
561 memory). However, it may result in a premature out-of-memory error.
562 So if you will be manipulating very large blocks with sizes close to
563 powers of two, it would be wise to define this macro.
565 Expected saving of memory is 0-100% (100% in applications which
566 require most memory in such 2**n chunks); expected slowdown is
571 =head2 Miscellaneous efficiency enhancements
573 Functions that have an empty prototype and that do nothing but return
574 a fixed value are now inlined (e.g. C<sub PI () { 3.14159 }>).
576 Each unique hash key is only allocated once, no matter how many hashes
577 have an entry with that key. So even if you have 100 copies of the
578 same hash, the hash keys never have to be reallocated.
582 Six new pragmatic modules exist:
586 =item use autouse MODULE => qw(sub1 sub2 sub3)
588 Defers C<require MODULE> until someone calls one of the specified
589 subroutines (which must be exported by MODULE). This pragma should be
590 used with caution, and only when necessary.
596 Looks for MakeMaker-like I<'blib'> directory structure starting in
597 I<dir> (or current directory) and working back up to five levels of
600 Intended for use on command line with B<-M> option as a way of testing
601 arbitrary scripts against an uninstalled version of a package.
603 =item use constant NAME => VALUE
605 Provides a convenient interface for creating compile-time constants,
606 See L<perlsub/"Constant Functions">.
610 Tells the compiler to enable (or disable) the use of POSIX locales for
613 When C<use locale> is in effect, the current LC_CTYPE locale is used
614 for regular expressions and case mapping; LC_COLLATE for string
615 ordering; and LC_NUMERIC for numeric formating in printf and sprintf
616 (but B<not> in print). LC_NUMERIC is always used in write, since
617 lexical scoping of formats is problematic at best.
619 Each C<use locale> or C<no locale> affects statements to the end of
620 the enclosing BLOCK or, if not inside a BLOCK, to the end of the
621 current file. Locales can be switched and queried with
624 See L<perllocale> for more information.
628 Disable unsafe opcodes, or any named opcodes, when compiling Perl code.
632 Enable VMS-specific language features. Currently, there are three
633 VMS-specific features available: 'status', which makes C<$?> and
634 C<system> return genuine VMS status values instead of emulating POSIX;
635 'exit', which makes C<exit> take a genuine VMS status value instead of
636 assuming that C<exit 1> is an error; and 'time', which makes all times
637 relative to the local time zone, in the VMS tradition.
643 =head2 Required Updates
645 Though Perl 5.004 is compatible with almost all modules that work
646 with Perl 5.003, there are a few exceptions:
648 Module Required Version for Perl 5.004
649 ------ -------------------------------
652 Tk Tk400.202 (-w makes noise)
654 =head2 Installation directories
656 The I<installperl> script now places the Perl source files for
657 extensions in the architecture-specific library directory, which is
658 where the shared libraries for extensions have always been. This
659 change is intended to allow administrators to keep the Perl 5.004
660 library directory unchanged from a previous version, without running
661 the risk of binary incompatibility between extensions' Perl source and
664 =head2 Module information summary
666 Brand new modules, arranged by topic rather than strictly
669 CPAN interface to Comprehensive Perl Archive Network
670 CPAN::FirstTime create a CPAN configuration file
671 CPAN::Nox run CPAN while avoiding compiled extensions
673 IO.pm Top-level interface to IO::* classes
674 IO/File.pm IO::File extension Perl module
675 IO/Handle.pm IO::Handle extension Perl module
676 IO/Pipe.pm IO::Pipe extension Perl module
677 IO/Seekable.pm IO::Seekable extension Perl module
678 IO/Select.pm IO::Select extension Perl module
679 IO/Socket.pm IO::Socket extension Perl module
681 Opcode.pm Disable named opcodes when compiling Perl code
683 ExtUtils/Embed.pm Utilities for embedding Perl in C programs
684 ExtUtils/testlib.pm Fixes up @INC to use just-built extension
686 FindBin.pm Find path of currently executing program
688 Class/Struct.pm Declare struct-like datatypes as Perl classes
689 File/stat.pm By-name interface to Perl's built-in stat
690 Net/hostent.pm By-name interface to Perl's built-in gethost*
691 Net/netent.pm By-name interface to Perl's built-in getnet*
692 Net/protoent.pm By-name interface to Perl's built-in getproto*
693 Net/servent.pm By-name interface to Perl's built-in getserv*
694 Time/gmtime.pm By-name interface to Perl's built-in gmtime
695 Time/localtime.pm By-name interface to Perl's built-in localtime
696 Time/tm.pm Internal object for Time::{gm,local}time
697 User/grent.pm By-name interface to Perl's built-in getgr*
698 User/pwent.pm By-name interface to Perl's built-in getpw*
700 Tie/RefHash.pm Base class for tied hashes with references as keys
702 UNIVERSAL.pm Base class for *ALL* classes
706 New constants in the existing Fcntl modules are now supported,
707 provided that your operating system happens to support them:
710 O_ASYNC O_DEFER O_DSYNC O_FSYNC O_SYNC
713 These constants are intended for use with the Perl operators sysopen()
714 and fcntl() and the basic database modules like SDBM_File. For the
715 exact meaning of these and other Fcntl constants please refer to your
716 operating system's documentation for fcntl() and open().
718 In addition, the Fcntl module now provides these constants for use
719 with the Perl operator flock():
721 LOCK_SH LOCK_EX LOCK_NB LOCK_UN
723 These constants are defined in all environments (because where there is
724 no flock() system call, Perl emulates it). However, for historical
725 reasons, these constants are not exported unless they are explicitly
726 requested with the ":flock" tag (e.g. C<use Fcntl ':flock'>).
730 The IO module provides a simple mechanism to load all of the IO modules at one
731 go. Currently this includes:
739 For more information on any of these modules, please see its
740 respective documentation.
744 The Math::Complex module has been totally rewritten, and now supports
745 more operations. These are overloaded:
747 + - * / ** <=> neg ~ abs sqrt exp log sin cos atan2 "" (stringify)
749 And these functions are now exported:
752 log10 logn ln cbrt root
765 This new module provides a simpler interface to parts of Math::Complex for
766 those who need trigonometric functions only for real numbers.
770 There have been quite a few changes made to DB_File. Here are a few of
777 Fixed a handful of bugs.
781 By public demand, added support for the standard hash function exists().
785 Made it compatible with Berkeley DB 1.86.
789 Made negative subscripts work with RECNO interface.
793 Changed the default flags from O_RDWR to O_CREAT|O_RDWR and the default
794 mode from 0640 to 0666.
798 Made DB_File automatically import the open() constants (O_RDWR,
799 O_CREAT etc.) from Fcntl, if available.
803 Updated documentation.
807 Refer to the HISTORY section in DB_File.pm for a complete list of
808 changes. Everything after DB_File 1.01 has been added since 5.003.
812 Major rewrite - support added for both udp echo and real icmp pings.
814 =head2 Object-oriented overrides for builtin operators
816 Many of the Perl builtins returning lists now have
817 object-oriented overrides. These are:
829 For example, you can now say
833 $his = (stat($filename)->st_uid == pwent($whoever)->pw_uid);
835 =head1 Utility Changes
841 =item C<void> XSUBs now default to returning nothing
843 Due to a documentation/implementation bug in previous versions of
844 Perl, XSUBs with a return type of C<void> have actually been
845 returning one value. Usually that value was the GV for the XSUB,
846 but sometimes it was some already freed or reused value, which would
847 sometimes lead to program failure.
849 In Perl 5.004, if an XSUB is declared as returning C<void>, it
850 actually returns no value, i.e. an empty list (though there is a
851 backward-compatibility exception; see below). If your XSUB really
852 does return an SV, you should give it a return type of C<SV *>.
854 For backward compatibility, I<xsubpp> tries to guess whether a
855 C<void> XSUB is really C<void> or if it wants to return an C<SV *>.
856 It does so by examining the text of the XSUB: if I<xsubpp> finds
857 what looks like an assignment to C<ST(0)>, it assumes that the
858 XSUB's return type is really C<SV *>.
862 =head1 C Language API Changes
866 =item C<gv_fetchmethod> and C<perl_call_sv>
868 The C<gv_fetchmethod> function finds a method for an object, just like
869 in Perl 5.003. The GV it returns may be a method cache entry.
870 However, in Perl 5.004, method cache entries are not visible to users;
871 therefore, they can no longer be passed directly to C<perl_call_sv>.
872 Instead, you should use the C<GvCV> macro on the GV to extract its CV,
873 and pass the CV to C<perl_call_sv>.
875 The most likely symptom of passing the result of C<gv_fetchmethod> to
876 C<perl_call_sv> is Perl's producing an "Undefined subroutine called"
877 error on the I<second> call to a given method (since there is no cache
880 =item Extended API for manipulating hashes
882 Internal handling of hash keys has changed. The old hashtable API is
883 still fully supported, and will likely remain so. The additions to the
884 API allow passing keys as C<SV*>s, so that C<tied> hashes can be given
885 real scalars as keys rather than plain strings (nontied hashes still
886 can only use strings as keys). New extensions must use the new hash
887 access functions and macros if they wish to use C<SV*> keys. These
888 additions also make it feasible to manipulate C<HE*>s (hash entries),
889 which can be more efficient. See L<perlguts> for details.
893 =head1 Documentation Changes
895 Many of the base and library pods were updated. These
896 new pods are included in section 1:
906 Locale support (internationalization and localization).
910 Tutorial on Perl OO programming.
914 Perl internal IO abstraction interface.
918 Although not new, this has been massively updated.
922 Although not new, this has been massively updated.
926 =head1 New Diagnostics
928 Several new conditions will trigger warnings that were
929 silent before. Some only affect certain platforms.
930 The following new warnings and errors outline these.
931 These messages are classified as follows (listed in
932 increasing order of desperation):
934 (W) A warning (optional).
935 (D) A deprecation (optional).
936 (S) A severe warning (mandatory).
937 (F) A fatal error (trappable).
938 (P) An internal error you should never see (trappable).
939 (X) A very fatal error (nontrappable).
940 (A) An alien error message (not generated by Perl).
944 =item "my" variable %s masks earlier declaration in same scope
946 (S) A lexical variable has been redeclared in the same scope, effectively
947 eliminating all access to the previous instance. This is almost always
948 a typographical error. Note that the earlier variable will still exist
949 until the end of the scope or until all closure referents to it are
952 =item %s argument is not a HASH element or slice
954 (F) The argument to delete() must be either a hash element, such as
957 $ref->[12]->{"susie"}
959 or a hash slice, such as
961 @foo{$bar, $baz, $xyzzy}
962 @{$ref->[12]}{"susie", "queue"}
964 =item Allocation too large: %lx
966 (X) You can't allocate more than 64K on an MS-DOS machine.
968 =item Allocation too large
970 (F) You can't allocate more than 2^31+"small amount" bytes.
972 =item Applying %s to %s will act on scalar(%s)
974 (W) The pattern match (//), substitution (s///), and translation (tr///)
975 operators work on scalar values. If you apply one of them to an array
976 or a hash, it will convert the array or hash to a scalar value -- the
977 length of an array, or the population info of a hash -- and then work on
978 that scalar value. This is probably not what you meant to do. See
979 L<perlfunc/grep> and L<perlfunc/map> for alternatives.
981 =item Attempt to free nonexistent shared string
983 (P) Perl maintains a reference counted internal table of strings to
984 optimize the storage and access of hash keys and other strings. This
985 indicates someone tried to decrement the reference count of a string
986 that can no longer be found in the table.
988 =item Attempt to use reference as lvalue in substr
990 (W) You supplied a reference as the first argument to substr() used
991 as an lvalue, which is pretty strange. Perhaps you forgot to
992 dereference it first. See L<perlfunc/substr>.
994 =item Can't use bareword ("%s") as %s ref while "strict refs" in use
996 (F) Only hard references are allowed by "strict refs". Symbolic references
997 are disallowed. See L<perlref>.
999 =item Cannot resolve method `%s' overloading `%s' in package `%s'
1001 (P) Internal error trying to resolve overloading specified by a method
1002 name (as opposed to a subroutine reference).
1004 =item Constant subroutine %s redefined
1006 (S) You redefined a subroutine which had previously been eligible for
1007 inlining. See L<perlsub/"Constant Functions"> for commentary and
1010 =item Constant subroutine %s undefined
1012 (S) You undefined a subroutine which had previously been eligible for
1013 inlining. See L<perlsub/"Constant Functions"> for commentary and
1016 =item Copy method did not return a reference
1018 (F) The method which overloads "=" is buggy. See L<overload/Copy Constructor>.
1022 (F) You passed die() an empty string (the equivalent of C<die "">) or
1023 you called it with no args and both C<$@> and C<$_> were empty.
1025 =item Exiting pseudo-block via %s
1027 (W) You are exiting a rather special block construct (like a sort block or
1028 subroutine) by unconventional means, such as a goto, or a loop control
1029 statement. See L<perlfunc/sort>.
1031 =item Illegal character %s (carriage return)
1033 (F) A carriage return character was found in the input. This is an
1034 error, and not a warning, because carriage return characters can break
1035 multi-line strings, including here documents (e.g., C<print E<lt>E<lt>EOF;>).
1037 =item Illegal switch in PERL5OPT: %s
1039 (X) The PERL5OPT environment variable may only be used to set the
1040 following switches: B<-[DIMUdmw]>.
1042 =item Integer overflow in hex number
1044 (S) The literal hex number you have specified is too big for your
1045 architecture. On a 32-bit architecture the largest hex literal is
1048 =item Integer overflow in octal number
1050 (S) The literal octal number you have specified is too big for your
1051 architecture. On a 32-bit architecture the largest octal literal is
1054 =item internal error: glob failed
1056 (P) Something went wrong with the external program(s) used for C<glob>
1057 and C<E<lt>*.cE<gt>>. This may mean that your csh (C shell) is
1058 broken. If so, you should change all of the csh-related variables in
1059 config.sh: If you have tcsh, make the variables refer to it as if it
1060 were csh (e.g. C<full_csh='/usr/bin/tcsh'>); otherwise, make them all
1061 empty (except that C<d_csh> should be C<'undef'>) so that Perl will
1062 think csh is missing. In either case, after editing config.sh, run
1063 C<./Configure -S> and rebuild Perl.
1065 =item Name "%s::%s" used only once: possible typo
1067 (W) Typographical errors often show up as unique variable names.
1068 If you had a good reason for having a unique name, then just mention
1069 it again somehow to suppress the message (the C<use vars> pragma is
1070 provided for just this purpose).
1072 =item Null picture in formline
1074 (F) The first argument to formline must be a valid format picture
1075 specification. It was found to be empty, which probably means you
1076 supplied it an uninitialized value. See L<perlform>.
1078 =item Offset outside string
1080 (F) You tried to do a read/write/send/recv operation with an offset
1081 pointing outside the buffer. This is difficult to imagine.
1082 The sole exception to this is that C<sysread()>ing past the buffer
1083 will extend the buffer and zero pad the new area.
1085 =item Out of memory!
1087 (X|F) The malloc() function returned 0, indicating there was insufficient
1088 remaining memory (or virtual memory) to satisfy the request.
1090 The request was judged to be small, so the possibility to trap it
1091 depends on the way Perl was compiled. By default it is not trappable.
1092 However, if compiled for this, Perl may use the contents of C<$^M> as
1093 an emergency pool after die()ing with this message. In this case the
1094 error is trappable I<once>.
1096 =item Out of memory during request for %s
1098 (F) The malloc() function returned 0, indicating there was insufficient
1099 remaining memory (or virtual memory) to satisfy the request. However,
1100 the request was judged large enough (compile-time default is 64K), so
1101 a possibility to shut down by trapping this error is granted.
1103 =item Possible attempt to put comments in qw() list
1105 (W) qw() lists contain items separated by whitespace; as with literal
1106 strings, comment characters are not ignored, but are instead treated
1107 as literal data. (You may have used different delimiters than the
1108 exclamation marks parentheses shown here; braces are also frequently
1111 You probably wrote something like this:
1118 when you should have written this:
1125 If you really want comments, build your list the
1126 old-fashioned way, with quotes and commas:
1130 'b', # another comment
1133 =item Possible attempt to separate words with commas
1135 (W) qw() lists contain items separated by whitespace; therefore commas
1136 aren't needed to separate the items. (You may have used different
1137 delimiters than the parentheses shown here; braces are also frequently
1140 You probably wrote something like this:
1144 which puts literal commas into some of the list items. Write it without
1145 commas if you don't want them to appear in your data:
1149 =item Recursive substitution detected
1151 (F) The replacement string of a substitution caused the recursive
1152 execution of that very same substituion. Perl cannot keep track of
1153 special variables (C<$1>, etc.) under such circumstances.
1155 =item Scalar value @%s{%s} better written as $%s{%s}
1157 (W) You've used a hash slice (indicated by @) to select a single element of
1158 a hash. Generally it's better to ask for a scalar value (indicated by $).
1159 The difference is that C<$foo{&bar}> always behaves like a scalar, both when
1160 assigning to it and when evaluating its argument, while C<@foo{&bar}> behaves
1161 like a list when you assign to it, and provides a list context to its
1162 subscript, which can do weird things if you're expecting only one subscript.
1164 =item Stub found while resolving method `%s' overloading `%s' in package `%s'
1166 (P) Overloading resolution over @ISA tree may be broken by importing stubs.
1167 Stubs should never be implicitely created, but explicit calls to C<can>
1170 =item Too late for "B<-T>" option
1172 (X) The #! line (or local equivalent) in a Perl script contains the
1173 B<-T> option, but Perl was not invoked with B<-T> in its argument
1174 list. This is an error because, by the time Perl discovers a B<-T> in
1175 a script, it's too late to properly taint everything from the
1176 environment. So Perl gives up.
1178 =item untie attempted while %d inner references still exist
1180 (W) A copy of the object returned from C<tie> (or C<tied>) was still
1181 valid when C<untie> was called.
1183 =item Unrecognized character %s
1185 (F) The Perl parser has no idea what to do with the specified character
1186 in your Perl script (or eval). Perhaps you tried to run a compressed
1187 script, a binary program, or a directory as a Perl program.
1189 =item Unsupported function fork
1191 (F) Your version of executable does not support forking.
1193 Note that under some systems, like OS/2, there may be different flavors of
1194 Perl executables, some of which may support fork, some not. Try changing
1195 the name you call Perl by to C<perl_>, C<perl__>, and so on.
1197 =item Use of "$$<digit>" to mean "${$}<digit>" is deprecated
1199 (D) Perl versions before 5.004 misinterpreted any type marker followed
1200 by "$" and a digit. For example, "$$0" was incorrectly taken to mean
1201 "${$}0" instead of "${$0}". This bug is (mostly) fixed in Perl 5.004.
1203 However, the developers of Perl 5.004 could not fix this bug completely,
1204 because at least two widely-used modules depend on the old meaning of
1205 "$$0" in a string. So Perl 5.004 still interprets "$$<digit>" in the
1206 old (broken) way inside strings; but it generates this message as a
1207 warning. And in Perl 5.005, this special treatment will cease.
1209 =item Value of %s can be "0"; test with defined()
1211 (W) In a conditional expression, you used <HANDLE>, <*> (glob), C<each()>,
1212 or C<readdir()> as a boolean value. Each of these constructs can return a
1213 value of "0"; that would make the conditional expression false, which is
1214 probably not what you intended. When using these constructs in conditional
1215 expressions, test their values with the C<defined> operator.
1217 =item Variable "%s" may be unavailable
1219 (W) An inner (nested) I<anonymous> subroutine is inside a I<named>
1220 subroutine, and outside that is another subroutine; and the anonymous
1221 (innermost) subroutine is referencing a lexical variable defined in
1222 the outermost subroutine. For example:
1224 sub outermost { my $a; sub middle { sub { $a } } }
1226 If the anonymous subroutine is called or referenced (directly or
1227 indirectly) from the outermost subroutine, it will share the variable
1228 as you would expect. But if the anonymous subroutine is called or
1229 referenced when the outermost subroutine is not active, it will see
1230 the value of the shared variable as it was before and during the
1231 *first* call to the outermost subroutine, which is probably not what
1234 In these circumstances, it is usually best to make the middle
1235 subroutine anonymous, using the C<sub {}> syntax. Perl has specific
1236 support for shared variables in nested anonymous subroutines; a named
1237 subroutine in between interferes with this feature.
1239 =item Variable "%s" will not stay shared
1241 (W) An inner (nested) I<named> subroutine is referencing a lexical
1242 variable defined in an outer subroutine.
1244 When the inner subroutine is called, it will probably see the value of
1245 the outer subroutine's variable as it was before and during the
1246 *first* call to the outer subroutine; in this case, after the first
1247 call to the outer subroutine is complete, the inner and outer
1248 subroutines will no longer share a common value for the variable. In
1249 other words, the variable will no longer be shared.
1251 Furthermore, if the outer subroutine is anonymous and references a
1252 lexical variable outside itself, then the outer and inner subroutines
1253 will I<never> share the given variable.
1255 This problem can usually be solved by making the inner subroutine
1256 anonymous, using the C<sub {}> syntax. When inner anonymous subs that
1257 reference variables in outer subroutines are called or referenced,
1258 they are automatically rebound to the current values of such
1261 =item Warning: something's wrong
1263 (W) You passed warn() an empty string (the equivalent of C<warn "">) or
1264 you called it with no args and C<$_> was empty.
1266 =item Ill-formed logical name |%s| in prime_env_iter
1268 (W) A warning peculiar to VMS. A logical name was encountered when preparing
1269 to iterate over %ENV which violates the syntactic rules governing logical
1270 names. Since it cannot be translated normally, it is skipped, and will not
1271 appear in %ENV. This may be a benign occurrence, as some software packages
1272 might directly modify logical name tables and introduce nonstandard names,
1273 or it may indicate that a logical name table has been corrupted.
1275 =item Got an error from DosAllocMem
1277 (P) An error peculiar to OS/2. Most probably you're using an obsolete
1278 version of Perl, and this should not happen anyway.
1280 =item Malformed PERLLIB_PREFIX
1282 (F) An error peculiar to OS/2. PERLLIB_PREFIX should be of the form
1290 with nonempty prefix1 and prefix2. If C<prefix1> is indeed a prefix
1291 of a builtin library search path, prefix2 is substituted. The error
1292 may appear if components are not found, or are too long. See
1293 "PERLLIB_PREFIX" in F<README.os2>.
1295 =item PERL_SH_DIR too long
1297 (F) An error peculiar to OS/2. PERL_SH_DIR is the directory to find the
1298 C<sh>-shell in. See "PERL_SH_DIR" in F<README.os2>.
1300 =item Process terminated by SIG%s
1302 (W) This is a standard message issued by OS/2 applications, while *nix
1303 applications die in silence. It is considered a feature of the OS/2
1304 port. One can easily disable this by appropriate sighandlers, see
1305 L<perlipc/"Signals">. See also "Process terminated by SIGTERM/SIGINT"
1312 If you find what you think is a bug, you might check the headers of
1313 recently posted articles in the comp.lang.perl.misc newsgroup.
1314 There may also be information at http://www.perl.com/perl/, the Perl
1317 If you believe you have an unreported bug, please run the B<perlbug>
1318 program included with your release. Make sure you trim your bug down
1319 to a tiny but sufficient test case. Your bug report, along with the
1320 output of C<perl -V>, will be sent off to <F<perlbug@perl.com>> to be
1321 analysed by the Perl porting team.
1325 The F<Changes> file for exhaustive details on what changed.
1327 The F<INSTALL> file for how to build Perl. This file has been
1328 significantly updated for 5.004, so even veteran users should
1331 The F<README> file for general stuff.
1333 The F<Copying> file for copyright information.
1337 Constructed by Tom Christiansen, grabbing material with permission
1338 from innumerable contributors, with kibitzing by more than a few Perl
1341 Last update: Sat Mar 8 19:51:26 EST 1997