1PERL58DELTA(1) Perl Programmers Reference Guide PERL58DELTA(1)
2
6 perl58delta - what is new for perl v5.8.0
7
9 This document describes differences between the 5.6.0 release and the
10 5.8.0 release.
11 Many of the bug fixes in 5.8.0 were already seen in the 5.6.1 mainte‐
13 nance release since the two releases were kept closely coordinated
14 (while 5.8.0 was still called 5.7.something).
15 Changes that were integrated into the 5.6.1 release are marked "[561]".
17 Many of these changes have been further developed since 5.6.1 was
18 released, those are marked "[561+]".
19 You can see the list of changes in the 5.6.1 release (both from the
21 5.005_03 release and the 5.6.0 release) by reading perl561delta.
22
24 · Better Unicode support
25 · New IO Implementation
27 · New Thread Implementation
29 · Better Numeric Accuracy
31 · Safe Signals
33 · Many New Modules
35 · More Extensive Regression Testing
37
39 Binary Incompatibility
40 Perl 5.8 is not binary compatible with earlier releases of Perl.
42 You have to recompile your XS modules.
44 (Pure Perl modules should continue to work.)
46 The major reason for the discontinuity is the new IO architecture
48 called PerlIO. PerlIO is the default configuration because without it
49 many new features of Perl 5.8 cannot be used. In other words: you just
50 have to recompile your modules containing XS code, sorry about that.
51 In future releases of Perl, non-PerlIO aware XS modules may become com‐
53 pletely unsupported. This shouldn't be too difficult for module
54 authors, however: PerlIO has been designed as a drop-in replacement (at
55 the source code level) for the stdio interface.
56 Depending on your platform, there are also other reasons why we decided
58 to break binary compatibility, please read on.
59 64-bit platforms and malloc
61 If your pointers are 64 bits wide, the Perl malloc is no longer being
63 used because it does not work well with 8-byte pointers. Also, usually
64 the system mallocs on such platforms are much better optimized for such
65 large memory models than the Perl malloc. Some memory-hungry Perl
66 applications like the PDL don't work well with Perl's malloc. Finally,
67 other applications than Perl (such as mod_perl) tend to prefer the sys‐
68 tem malloc. Such platforms include Alpha and 64-bit HPPA, MIPS, PPC,
69 and Sparc.
70 AIX Dynaloading
72 The AIX dynaloading now uses in AIX releases 4.3 and newer the native
74 dlopen interface of AIX instead of the old emulated interface. This
75 change will probably break backward compatibility with compiled mod‐
76 ules. The change was made to make Perl more compliant with other
77 applications like mod_perl which are using the AIX native interface.
78 Attributes for "my" variables now handled at run-time
80 The "my EXPR : ATTRS" syntax now applies variable attributes at
82 run-time. (Subroutine and "our" variables still get attributes applied
83 at compile-time.) See attributes for additional details. In particu‐
84 lar, however, this allows variable attributes to be useful for "tie"
85 interfaces, which was a deficiency of earlier releases. Note that the
86 new semantics doesn't work with the Attribute::Handlers module (as of
87 version 0.76).
88 Socket Extension Dynamic in VMS
90 The Socket extension is now dynamically loaded instead of being stati‐
92 cally built in. This may or may not be a problem with ancient TCP/IP
93 stacks of VMS: we do not know since we weren't able to test Perl in
94 such configurations.
95 IEEE-format Floating Point Default on OpenVMS Alpha
97 Perl now uses IEEE format (T_FLOAT) as the default internal floating
99 point format on OpenVMS Alpha, potentially breaking binary compatibil‐
100 ity with external libraries or existing data. G_FLOAT is still avail‐
101 able as a configuration option. The default on VAX (D_FLOAT) has not
102 changed.
103 New Unicode Semantics (no more "use utf8", almost)
105 Previously in Perl 5.6 to use Unicode one would say "use utf8" and then
107 the operations (like string concatenation) were Unicode-aware in that
108 lexical scope.
109 This was found to be an inconvenient interface, and in Perl 5.8 the
111 Unicode model has completely changed: now the "Unicodeness" is bound to
112 the data itself, and for most of the time "use utf8" is not needed at
113 all. The only remaining use of "use utf8" is when the Perl script
114 itself has been written in the UTF-8 encoding of Unicode. (UTF-8 has
115 not been made the default since there are many Perl scripts out there
116 that are using various national eight-bit character sets, which would
117 be illegal in UTF-8.)
118 See perluniintro for the explanation of the current model, and utf8 for
120 the current use of the utf8 pragma.
121 New Unicode Properties
123 Unicode scripts are now supported. Scripts are similar to (and superior
125 to) Unicode blocks. The difference between scripts and blocks is that
126 scripts are the glyphs used by a language or a group of languages,
127 while the blocks are more artificial groupings of (mostly) 256 charac‐
128 ters based on the Unicode numbering.
129 In general, scripts are more inclusive, but not universally so. For
131 example, while the script "Latin" includes all the Latin characters and
132 their various diacritic-adorned versions, it does not include the vari‐
133 ous punctuation or digits (since they are not solely "Latin").
134 A number of other properties are now supported, including "\p{L&}",
136 "\p{Any}" "\p{Assigned}", "\p{Unassigned}", "\p{Blank}" [561] and
137 "\p{SpacePerl}" [561] (along with their "\P{...}" versions, of course).
138 See perlunicode for details, and more additions.
139 The "In" or "Is" prefix to names used with the "\p{...}" and "\P{...}"
141 are now almost always optional. The only exception is that a "In" pre‐
142 fix is required to signify a Unicode block when a block name conflicts
143 with a script name. For example, "\p{Tibetan}" refers to the script,
144 while "\p{InTibetan}" refers to the block. When there is no name con‐
145 flict, you can omit the "In" from the block name (e.g. "\p{BraillePat‐
146 terns}"), but to be safe, it's probably best to always use the "In").
147 REF(...) Instead Of SCALAR(...)
149 A reference to a reference now stringifies as "REF(0x81485ec)" instead
151 of "SCALAR(0x81485ec)" in order to be more consistent with the return
152 value of ref().
153 pack/unpack D/F recycled
155 The undocumented pack/unpack template letters D/F have been recycled
157 for better use: now they stand for long double (if supported by the
158 platform) and NV (Perl internal floating point type). (They used to be
159 aliases for d/f, but you never knew that.)
160 glob() now returns filenames in alphabetical order
162 The list of filenames from glob() (or <...>) is now by default sorted
164 alphabetically to be csh-compliant (which is what happened before in
165 most UNIX platforms). (bsd_glob() does still sort platform natively,
166 ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.) [561]
167 Deprecations
169 · The semantics of bless(REF, REF) were unclear and until someone
171 proves it to make some sense, it is forbidden.
172 · The obsolete chat2 library that should never have been allowed to
174 escape the laboratory has been decommissioned.
175 · Using chdir("") or chdir(undef) instead of explicit chdir() is
177 doubtful. A failure (think chdir(some_function()) can lead into
178 unintended chdir() to the home directory, therefore this behaviour
179 is deprecated.
180 · The builtin dump() function has probably outlived most of its use‐
182 fulness. The core-dumping functionality will remain in future
183 available as an explicit call to "CORE::dump()", but in future
184 releases the behaviour of an unqualified "dump()" call may change.
185 · The very dusty examples in the eg/ directory have been removed.
187 Suggestions for new shiny examples welcome but the main issue is
188 that the examples need to be documented, tested and (most impor‐
189 tantly) maintained.
190 · The (bogus) escape sequences \8 and \9 now give an optional warning
192 ("Unrecognized escape passed through"). There is no need to
193 \-escape any "\w" character.
194 · The *glob{FILEHANDLE} is deprecated, use *glob{IO} instead.
196 · The "package;" syntax ("package" without an argument) has been dep‐
198 recated. Its semantics were never that clear and its implementa‐
199 tion even less so. If you have used that feature to disallow all
200 but fully qualified variables, "use strict;" instead.
201 · The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are
203 still recognised but now cause fatal errors. The previous behav‐
204 iour of ignoring them by default and warning if requested was unac‐
205 ceptable since it, in a way, falsely promised that the features
206 could be used.
207 · In future releases, non-PerlIO aware XS modules may become com‐
209 pletely unsupported. Since PerlIO is a drop-in replacement for
210 stdio at the source code level, this shouldn't be that drastic a
211 change.
212 · Previous versions of perl and some readings of some sections of
214 Camel III implied that the ":raw" "discipline" was the inverse of
215 ":crlf". Turning off "clrfness" is no longer enough to make a
216 stream truly binary. So the PerlIO ":raw" layer (or "discipline",
217 to use the Camel book's older terminology) is now formally defined
218 as being equivalent to binmode(FH) - which is in turn defined as
219 doing whatever is necessary to pass each byte as-is without any
220 translation. In particular binmode(FH) - and hence ":raw" - will
221 now turn off both CRLF and UTF-8 translation and remove other lay‐
222 ers (e.g. :encoding()) which would modify byte stream.
223 · The current user-visible implementation of pseudo-hashes (the weird
225 use of the first array element) is deprecated starting from Perl
226 5.8.0 and will be removed in Perl 5.10.0, and the feature will be
227 implemented differently. Not only is the current interface rather
228 ugly, but the current implementation slows down normal array and
229 hash use quite noticeably. The "fields" pragma interface will
230 remain available. The restricted hashes interface is expected to
231 be the replacement interface (see Hash::Util). If your existing
232 programs depends on the underlying implementation, consider using
233 Class::PseudoHash from CPAN.
234 · The syntaxes "@a->[...]" and "%h->{...}" have now been deprecated.
236 · After years of trying, suidperl is considered to be too complex to
238 ever be considered truly secure. The suidperl functionality is
239 likely to be removed in a future release.
240 · The 5.005 threads model (module "Thread") is deprecated and
242 expected to be removed in Perl 5.10. Multithreaded code should be
243 migrated to the new ithreads model (see threads, threads::shared
244 and perlthrtut).
245 · The long deprecated uppercase aliases for the string comparison
247 operators (EQ, NE, LT, LE, GE, GT) have now been removed.
248 · The tr///C and tr///U features have been removed and will not
250 return; the interface was a mistake. Sorry about that. For simi‐
251 lar functionality, see pack('U0', ...) and pack('C0', ...). [561]
252 · Earlier Perls treated "sub foo (@bar)" as equivalent to "sub foo
254 (@)". The prototypes are now checked better at compile-time for
255 invalid syntax. An optional warning is generated ("Illegal charac‐
256 ter in prototype...") but this may be upgraded to a fatal error in
257 a future release.
258 · The "exec LIST" and "system LIST" operations now produce warnings
260 on tainted data and in some future release they will produce fatal
261 errors.
262 · The existing behaviour when localising tied arrays and hashes is
264 wrong, and will be changed in a future release, so do not rely on
265 the existing behaviour. See "Localising Tied Arrays and Hashes Is
266 Broken".
267
269 Unicode Overhaul
270 Unicode in general should be now much more usable than in Perl 5.6.0
272 (or even in 5.6.1). Unicode can be used in hash keys, Unicode in regu‐
273 lar expressions should work now, Unicode in tr/// should work now, Uni‐
274 code in I/O should work now. See perluniintro for introduction and
275 perlunicode for details.
276 · The Unicode Character Database coming with Perl has been upgraded
278 to Unicode 3.2.0. For more information, see http://www.uni‐
279 code.org/ . [561+] (5.6.1 has UCD 3.0.1.)
280 · For developers interested in enhancing Perl's Unicode capabilities:
282 almost all the UCD files are included with the Perl distribution in
283 the lib/unicore subdirectory. The most notable omission, for space
284 considerations, is the Unihan database.
285 · The properties \p{Blank} and \p{SpacePerl} have been added. "Blank"
287 is like C isblank(), that is, it contains only "horizontal white‐
288 space" (the space character is, the newline isn't), and the
289 "SpacePerl" is the Unicode equivalent of "\s" (\p{Space} isn't,
290 since that includes the vertical tabulator character, whereas "\s"
291 doesn't.)
292 See "New Unicode Properties" earlier in this document for addi‐
294 tional information on changes with Unicode properties.
295 PerlIO is Now The Default
297 · IO is now by default done via PerlIO rather than system's "stdio".
299 PerlIO allows "layers" to be "pushed" onto a file handle to alter
300 the handle's behaviour. Layers can be specified at open time via
301 3-arg form of open:
302 open($fh,'>:crlf :utf8', $path) ⎪⎪ ...
304 or on already opened handles via extended "binmode":
306 binmode($fh,':encoding(iso-8859-7)');
308 The built-in layers are: unix (low level read/write), stdio (as in
310 previous Perls), perlio (re-implementation of stdio buffering in a
311 portable manner), crlf (does CRLF <=> "\n" translation as on Win32,
312 but available on any platform). A mmap layer may be available if
313 platform supports it (mostly UNIXes).
314 Layers to be applied by default may be specified via the 'open'
316 pragma.
317 See "Installation and Configuration Improvements" for the effects
319 of PerlIO on your architecture name.
320 · If your platform supports fork(), you can use the list form of
322 "open" for pipes. For example:
323 open KID_PS, "-⎪", "ps", "aux" or die $!;
325 forks the ps(1) command (without spawning a shell, as there are
327 more than three arguments to open()), and reads its standard output
328 via the "KID_PS" filehandle. See perlipc.
329 · File handles can be marked as accepting Perl's internal encoding of
331 Unicode (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo
332 layer ":utf8" :
333 open($fh,">:utf8","Uni.txt");
335 Note for EBCDIC users: the pseudo layer ":utf8" is erroneously
337 named for you since it's not UTF-8 what you will be getting but
338 instead UTF-EBCDIC. See perlunicode, utf8, and http://www.uni‐
339 code.org/unicode/reports/tr16/ for more information. In future
340 releases this naming may change. See perluniintro for more infor‐
341 mation about UTF-8.
342 · If your environment variables (LC_ALL, LC_CTYPE, LANG) look like
344 you want to use UTF-8 (any of the variables match "/utf-?8/i"),
345 your STDIN, STDOUT, STDERR handles and the default open layer (see
346 open) are marked as UTF-8. (This feature, like other new features
347 that combine Unicode and I/O, work only if you are using PerlIO,
348 but that's the default.)
349 Note that after this Perl really does assume that everything is
351 UTF-8: for example if some input handle is not, Perl will probably
352 very soon complain about the input data like this "Malformed UTF-8
353 ..." since any old eight-bit data is not legal UTF-8.
354 Note for code authors: if you want to enable your users to use
356 UTF-8 as their default encoding but in your code still have eight-
357 bit I/O streams (such as images or zip files), you need to explic‐
358 itly open() or binmode() with ":bytes" (see "open" in perlfunc and
359 "binmode" in perlfunc), or you can just use "binmode(FH)" (nice for
360 pre-5.8.0 backward compatibility).
361 · File handles can translate character encodings from/to Perl's
363 internal Unicode form on read/write via the ":encoding()" layer.
364 · File handles can be opened to "in memory" files held in Perl
366 scalars via:
367 open($fh,'>', \$variable) ⎪⎪ ...
369 · Anonymous temporary files are available without need to 'use File‐
371 Handle' or other module via
372 open($fh,"+>", undef) ⎪⎪ ...
374 That is a literal undef, not an undefined value.
376 ithreads
378 The new interpreter threads ("ithreads" for short) implementation of
380 multithreading, by Arthur Bergman, replaces the old "5.005 threads"
381 implementation. In the ithreads model any data sharing between threads
382 must be explicit, as opposed to the model where data sharing was
383 implicit. See threads and threads::shared, and perlthrtut.
384 As a part of the ithreads implementation Perl will also use any neces‐
386 sary and detectable reentrant libc interfaces.
387 Restricted Hashes
389 A restricted hash is restricted to a certain set of keys, no keys out‐
391 side the set can be added. Also individual keys can be restricted so
392 that the key cannot be deleted and the value cannot be changed. No new
393 syntax is involved: the Hash::Util module is the interface.
394 Safe Signals
396 Perl used to be fragile in that signals arriving at inopportune moments
398 could corrupt Perl's internal state. Now Perl postpones handling of
399 signals until it's safe (between opcodes).
400 This change may have surprising side effects because signals no longer
402 interrupt Perl instantly. Perl will now first finish whatever it was
403 doing, like finishing an internal operation (like sort()) or an exter‐
404 nal operation (like an I/O operation), and only then look at any
405 arrived signals (and before starting the next operation). No more cor‐
406 rupt internal state since the current operation is always finished
407 first, but the signal may take more time to get heard. Note that
408 breaking out from potentially blocking operations should still work,
409 though.
410 Understanding of Numbers
412 In general a lot of fixing has happened in the area of Perl's under‐
414 standing of numbers, both integer and floating point. Since in many
415 systems the standard number parsing functions like "strtoul()" and
416 "atof()" seem to have bugs, Perl tries to work around their deficien‐
417 cies. This results hopefully in more accurate numbers.
418 Perl now tries internally to use integer values in numeric conversions
420 and basic arithmetics (+ - * /) if the arguments are integers, and
421 tries also to keep the results stored internally as integers. This
422 change leads to often slightly faster and always less lossy arith‐
423 metics. (Previously Perl always preferred floating point numbers in its
424 math.)
425 Arrays now always interpolate into double-quoted strings [561]
427 In double-quoted strings, arrays now interpolate, no matter what. The
429 behavior in earlier versions of perl 5 was that arrays would interpo‐
430 late into strings if the array had been mentioned before the string was
431 compiled, and otherwise Perl would raise a fatal compile-time error.
432 In versions 5.000 through 5.003, the error was
433 Literal @example now requires backslash
435 In versions 5.004_01 through 5.6.0, the error was
437 In string, @example now must be written as \@example
439 The idea here was to get people into the habit of writing "fred\@exam‐
441 ple.com" when they wanted a literal "@" sign, just as they have always
442 written "Give me back my \$5" when they wanted a literal "$" sign.
443 Starting with 5.6.1, when Perl now sees an "@" sign in a double-quoted
445 string, it always attempts to interpolate an array, regardless of
446 whether or not the array has been used or declared already. The fatal
447 error has been downgraded to an optional warning:
448 Possible unintended interpolation of @example in string
450 This warns you that "fred@example.com" is going to turn into "fred.com"
452 if you don't backslash the "@". See
453 http://www.plover.com/~mjd/perl/at-error.html for more details about
454 the history here.
455 Miscellaneous Changes
457 · AUTOLOAD is now lvaluable, meaning that you can add the :lvalue
459 attribute to AUTOLOAD subroutines and you can assign to the
460 AUTOLOAD return value.
461 · The $Config{byteorder} (and corresponding BYTEORDER in config.h)
463 was previously wrong in platforms if sizeof(long) was 4, but
464 sizeof(IV) was 8. The byteorder was only sizeof(long) bytes long
465 (1234 or 4321), but now it is correctly sizeof(IV) bytes long,
466 (12345678 or 87654321). (This problem didn't affect Windows plat‐
467 forms.)
468 Also, $Config{byteorder} is now computed dynamically--this is more
470 robust with "fat binaries" where an executable image contains bina‐
471 ries for more than one binary platform, and when cross-compiling.
472 · "perl -d:Module=arg,arg,arg" now works (previously one couldn't
474 pass in multiple arguments.)
475 · "do" followed by a bareword now ensures that this bareword isn't a
477 keyword (to avoid a bug where "do q(foo.pl)" tried to call a sub‐
478 routine called "q"). This means that for example instead of "do
479 format()" you must write "do &format()".
480 · The builtin dump() now gives an optional warning "dump() better
482 written as CORE::dump()", meaning that by default "dump(...)" is
483 resolved as the builtin dump() which dumps core and aborts, not as
484 (possibly) user-defined "sub dump". To call the latter, qualify
485 the call as "&dump(...)". (The whole dump() feature is to consid‐
486 ered deprecated, and possibly removed/changed in future releases.)
487 · chomp() and chop() are now overridable. Note, however, that their
489 prototype (as given by "prototype("CORE::chomp")" is undefined,
490 because it cannot be expressed and therefore one cannot really
491 write replacements to override these builtins.
492 · END blocks are now run even if you exit/die in a BEGIN block.
494 Internally, the execution of END blocks is now controlled by
495 PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new behav‐
496 iour for Perl embedders. This will default in 5.10. See perlembed.
497 · Formats now support zero-padded decimal fields.
499 · Although "you shouldn't do that", it was possible to write code
501 that depends on Perl's hashed key order (Data::Dumper does this).
502 The new algorithm "One-at-a-Time" produces a different hashed key
503 order. More details are in "Performance Enhancements".
504 · lstat(FILEHANDLE) now gives a warning because the operation makes
506 no sense. In future releases this may become a fatal error.
507 · Spurious syntax errors generated in certain situations, when glob()
509 caused File::Glob to be loaded for the first time, have been fixed.
510 [561]
511 · Lvalue subroutines can now return "undef" in list context. How‐
513 ever, the lvalue subroutine feature still remains experimental.
514 [561+]
515 · A lost warning "Can't declare ... dereference in my" has been
517 restored (Perl had it earlier but it became lost in later
518 releases.)
519 · A new special regular expression variable has been introduced: $^N,
521 which contains the most-recently closed group (submatch).
522 · "no Module;" does not produce an error even if Module does not have
524 an unimport() method. This parallels the behavior of "use" vis-a-
525 vis "import". [561]
526 · The numerical comparison operators return "undef" if either operand
528 is a NaN. Previously the behaviour was unspecified.
529 · "our" can now have an experimental optional attribute "unique" that
531 affects how global variables are shared among multiple inter‐
532 preters, see "our" in perlfunc.
533 · The following builtin functions are now overridable: each(),
535 keys(), pop(), push(), shift(), splice(), unshift(). [561]
536 · "pack() / unpack()" can now group template letters with "()" and
538 then apply repetition/count modifiers on the groups.
539 · "pack() / unpack()" can now process the Perl internal numeric
541 types: IVs, UVs, NVs-- and also long doubles, if supported by the
542 platform. The template letters are "j", "J", "F", and "D".
543 · "pack('U0a*', ...)" can now be used to force a string to UTF-8.
545 · my __PACKAGE__ $obj now works. [561]
547 · POSIX::sleep() now returns the number of unslept seconds (as the
549 POSIX standard says), as opposed to CORE::sleep() which returns the
550 number of slept seconds.
551 · printf() and sprintf() now support parameter reordering using the
553 "%\d+\$" and "*\d+\$" syntaxes. For example
554 printf "%2\$s %1\$s\n", "foo", "bar";
556 will print "bar foo\n". This feature helps in writing internation‐
558 alised software, and in general when the order of the parameters
559 can vary.
560 · The (\&) prototype now works properly. [561]
562 · prototype(\[$@%&]) is now available to implicitly create references
564 (useful for example if you want to emulate the tie() interface).
565 · A new command-line option, "-t" is available. It is the little
567 brother of "-T": instead of dying on taint violations, lexical
568 warnings are given. This is only meant as a temporary debugging
569 aid while securing the code of old legacy applications. This is
570 not a substitute for -T.
571 · In other taint news, the "exec LIST" and "system LIST" have now
573 been considered too risky (think "exec @ARGV": it can start any
574 program with any arguments), and now the said forms cause a warning
575 under lexical warnings. You should carefully launder the arguments
576 to guarantee their validity. In future releases of Perl the forms
577 will become fatal errors so consider starting laundering now.
578 · Tied hash interfaces are now required to have the EXISTS and DELETE
580 methods (either own or inherited).
581 · If tr/// is just counting characters, it doesn't attempt to modify
583 its target.
584 · untie() will now call an UNTIE() hook if it exists. See perltie
586 for details. [561]
587 · utime now supports "utime undef, undef, @files" to change the file
589 timestamps to the current time.
590 · The rules for allowing underscores (underbars) in numeric constants
592 have been relaxed and simplified: now you can have an underscore
593 simply between digits.
594 · Rather than relying on C's argv[0] (which may not contain a full
596 pathname) where possible $^X is now set by asking the operating
597 system. (eg by reading /proc/self/exe on Linux, /proc/curproc/file
598 on FreeBSD)
599 · A new variable, "${^TAINT}", indicates whether taint mode is
601 enabled.
602 · You can now override the readline() builtin, and this overrides
604 also the <FILEHANDLE> angle bracket operator.
605 · The command-line options -s and -F are now recognized on the she‐
607 bang (#!) line.
608 · Use of the "/c" match modifier without an accompanying "/g" modi‐
610 fier elicits a new warning: "Use of /c modifier is meaningless
611 without /g".
612 Use of "/c" in substitutions, even with "/g", elicits "Use of /c
614 modifier is meaningless in s///".
615 Use of "/g" with "split" elicits "Use of /g modifier is meaningless
617 in split".
618 · Support for the "CLONE" special subroutine had been added. With
620 ithreads, when a new thread is created, all Perl data is cloned,
621 however non-Perl data cannot be cloned automatically. In "CLONE"
622 you can do whatever you need to do, like for example handle the
623 cloning of non-Perl data, if necessary. "CLONE" will be executed
624 once for every package that has it defined or inherited. It will
625 be called in the context of the new thread, so all modifications
626 are made in the new area.
627 See perlmod
629
631 New Modules and Pragmata
632 · "Attribute::Handlers", originally by Damian Conway and now main‐
634 tained by Arthur Bergman, allows a class to define attribute han‐
635 dlers.
636 package MyPack;
638 use Attribute::Handlers;
639 sub Wolf :ATTR(SCALAR) { print "howl!\n" }
640 # later, in some package using or inheriting from MyPack...
642 my MyPack $Fluffy : Wolf; # the attribute handler Wolf will be called
644 Both variables and routines can have attribute handlers. Handlers
646 can be specific to type (SCALAR, ARRAY, HASH, or CODE), or specific
647 to the exact compilation phase (BEGIN, CHECK, INIT, or END). See
648 Attribute::Handlers.
649 · "B::Concise", by Stephen McCamant, is a new compiler backend for
651 walking the Perl syntax tree, printing concise info about ops. The
652 output is highly customisable. See B::Concise. [561+]
653 · The new bignum, bigint, and bigrat pragmas, by Tels, implement
655 transparent bignum support (using the Math::BigInt, Math::BigFloat,
656 and Math::BigRat backends).
657 · "Class::ISA", by Sean Burke, is a module for reporting the search
659 path for a class's ISA tree. See Class::ISA.
660 · "Cwd" now has a split personality: if possible, an XS extension is
662 used, (this will hopefully be faster, more secure, and more robust)
663 but if not possible, the familiar Perl implementation is used.
664 · "Devel::PPPort", originally by Kenneth Albanowski and now main‐
666 tained by Paul Marquess, has been added. It is primarily used by
667 "h2xs" to enhance portability of XS modules between different ver‐
668 sions of Perl. See Devel::PPPort.
669 · "Digest", frontend module for calculating digests (checksums), from
671 Gisle Aas, has been added. See Digest.
672 · "Digest::MD5" for calculating MD5 digests (checksums) as defined in
674 RFC 1321, from Gisle Aas, has been added. See Digest::MD5.
675 use Digest::MD5 'md5_hex';
677 $digest = md5_hex("Thirsty Camel");
679 print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1
681 NOTE: the "MD5" backward compatibility module is deliberately not
683 included since its further use is discouraged.
684 See also PerlIO::via::QuotedPrint.
686 · "Encode", originally by Nick Ing-Simmons and now maintained by Dan
688 Kogai, provides a mechanism to translate between different charac‐
689 ter encodings. Support for Unicode, ISO-8859-1, and ASCII are com‐
690 piled in to the module. Several other encodings (like the rest of
691 the ISO-8859, CP*/Win*, Mac, KOI8-R, three variants EBCDIC, Chi‐
692 nese, Japanese, and Korean encodings) are included and can be
693 loaded at runtime. (For space considerations, the largest Chinese
694 encodings have been separated into their own CPAN module,
695 Encode::HanExtra, which Encode will use if available). See Encode.
696 Any encoding supported by Encode module is also available to the
698 ":encoding()" layer if PerlIO is used.
699 · "Hash::Util" is the interface to the new restricted hashes feature.
701 (Implemented by Jeffrey Friedl, Nick Ing-Simmons, and Michael Schw‐
702 ern.) See Hash::Util.
703 · "I18N::Langinfo" can be used to query locale information. See
705 I18N::Langinfo.
706 · "I18N::LangTags", by Sean Burke, has functions for dealing with
708 RFC3066-style language tags. See I18N::LangTags.
709 · "ExtUtils::Constant", by Nicholas Clark, is a new tool for exten‐
711 sion writers for generating XS code to import C header constants.
712 See ExtUtils::Constant.
713 · "Filter::Simple", by Damian Conway, is an easy-to-use frontend to
715 Filter::Util::Call. See Filter::Simple.
716 # in MyFilter.pm:
718 package MyFilter;
720 use Filter::Simple sub {
722 while (my ($from, $to) = splice @_, 0, 2) {
723 s/$from/$to/g;
724 }
725 };
726 1;
728 # in user's code:
730 use MyFilter qr/red/ => 'green';
732 print "red\n"; # this code is filtered, will print "green\n"
734 print "bored\n"; # this code is filtered, will print "bogreen\n"
735 no MyFilter;
737 print "red\n"; # this code is not filtered, will print "red\n"
739 · "File::Temp", by Tim Jenness, allows one to create temporary files
741 and directories in an easy, portable, and secure way. See
742 File::Temp. [561+]
743 · "Filter::Util::Call", by Paul Marquess, provides you with the
745 framework to write source filters in Perl. For most uses, the
746 frontend Filter::Simple is to be preferred. See Fil‐
747 ter::Util::Call.
748 · "if", by Ilya Zakharevich, is a new pragma for conditional inclu‐
750 sion of modules.
751 · libnet, by Graham Barr, is a collection of perl5 modules related to
753 network programming. See Net::FTP, Net::NNTP, Net::Ping (not part
754 of libnet, but related), Net::POP3, Net::SMTP, and Net::Time.
755 Perl installation leaves libnet unconfigured; use libnetcfg to con‐
757 figure it.
758 · "List::Util", by Graham Barr, is a selection of general-utility
760 list subroutines, such as sum(), min(), first(), and shuffle().
761 See List::Util.
762 · "Locale::Constants", "Locale::Country", "Locale::Currency"
764 "Locale::Language", and Locale::Script, by Neil Bowers, have been
765 added. They provide the codes for various locale standards, such
766 as "fr" for France, "usd" for US Dollar, and "ja" for Japanese.
767 use Locale::Country;
769 $country = code2country('jp'); # $country gets 'Japan'
771 $code = country2code('Norway'); # $code gets 'no'
772 See Locale::Constants, Locale::Country, Locale::Currency, and
774 Locale::Language.
775 · "Locale::Maketext", by Sean Burke, is a localization framework.
777 See Locale::Maketext, and Locale::Maketext::TPJ13. The latter is
778 an article about software localization, originally published in The
779 Perl Journal #13, and republished here with kind permission.
780 · "Math::BigRat" for big rational numbers, to accompany Math::BigInt
782 and Math::BigFloat, from Tels. See Math::BigRat.
783 · "Memoize" can make your functions faster by trading space for time,
785 from Mark-Jason Dominus. See Memoize.
786 · "MIME::Base64", by Gisle Aas, allows you to encode data in base64,
788 as defined in RFC 2045 - MIME (Multipurpose Internet Mail Exten‐
789 sions).
790 use MIME::Base64;
792 $encoded = encode_base64('Aladdin:open sesame');
794 $decoded = decode_base64($encoded);
795 print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ=="
797 See MIME::Base64.
799 · "MIME::QuotedPrint", by Gisle Aas, allows you to encode data in
801 quoted-printable encoding, as defined in RFC 2045 - MIME (Multipur‐
802 pose Internet Mail Extensions).
803 use MIME::QuotedPrint;
805 $encoded = encode_qp("\xDE\xAD\xBE\xEF");
807 $decoded = decode_qp($encoded);
808 print $encoded, "\n"; # "=DE=AD=BE=EF\n"
810 print $decoded, "\n"; # "\xDE\xAD\xBE\xEF\n"
811 See also PerlIO::via::QuotedPrint.
813 · "NEXT", by Damian Conway, is a pseudo-class for method redispatch.
815 See NEXT.
816 · "open" is a new pragma for setting the default I/O layers for
818 open().
819 · "PerlIO::scalar", by Nick Ing-Simmons, provides the implementation
821 of IO to "in memory" Perl scalars as discussed above. It also
822 serves as an example of a loadable PerlIO layer. Other future pos‐
823 sibilities include PerlIO::Array and PerlIO::Code. See Per‐
824 lIO::scalar.
825 · "PerlIO::via", by Nick Ing-Simmons, acts as a PerlIO layer and
827 wraps PerlIO layer functionality provided by a class (typically
828 implemented in Perl code).
829 · "PerlIO::via::QuotedPrint", by Elizabeth Mattijsen, is an example
831 of a "PerlIO::via" class:
832 use PerlIO::via::QuotedPrint;
834 open($fh,">:via(QuotedPrint)",$path);
835 This will automatically convert everything output to $fh to
837 Quoted-Printable. See PerlIO::via and PerlIO::via::QuotedPrint.
838 · "Pod::ParseLink", by Russ Allbery, has been added, to parse L<>
840 links in pods as described in the new perlpodspec.
841 · "Pod::Text::Overstrike", by Joe Smith, has been added. It converts
843 POD data to formatted overstrike text. See Pod::Text::Overstrike.
844 [561+]
845 · "Scalar::Util" is a selection of general-utility scalar subrou‐
847 tines, such as blessed(), reftype(), and tainted(). See
848 Scalar::Util.
849 · "sort" is a new pragma for controlling the behaviour of sort().
851 · "Storable" gives persistence to Perl data structures by allowing
853 the storage and retrieval of Perl data to and from files in a fast
854 and compact binary format. Because in effect Storable does serial‐
855 isation of Perl data structures, with it you can also clone deep,
856 hierarchical datastructures. Storable was originally created by
857 Raphael Manfredi, but it is now maintained by Abhijit Menon-Sen.
858 Storable has been enhanced to understand the two new hash features,
859 Unicode keys and restricted hashes. See Storable.
860 · "Switch", by Damian Conway, has been added. Just by saying
862 use Switch;
864 you have "switch" and "case" available in Perl.
866 use Switch;
868 switch ($val) {
870 case 1 { print "number 1" }
872 case "a" { print "string a" }
873 case [1..10,42] { print "number in list" }
874 case (@array) { print "number in list" }
875 case /\w+/ { print "pattern" }
876 case qr/\w+/ { print "pattern" }
877 case (%hash) { print "entry in hash" }
878 case (\%hash) { print "entry in hash" }
879 case (\&sub) { print "arg to subroutine" }
880 else { print "previous case not true" }
881 }
882 See Switch.
884 · "Test::More", by Michael Schwern, is yet another framework for
886 writing test scripts, more extensive than Test::Simple. See
887 Test::More.
888 · "Test::Simple", by Michael Schwern, has basic utilities for writing
890 tests. See Test::Simple.
891 · "Text::Balanced", by Damian Conway, has been added, for extracting
893 delimited text sequences from strings.
894 use Text::Balanced 'extract_delimited';
896 ($a, $b) = extract_delimited("'never say never', he never said", "'", '');
898 $a will be "'never say never'", $b will be ', he never said'.
900 In addition to extract_delimited(), there are also extract_brack‐
902 eted(), extract_quotelike(), extract_codeblock(), extract_vari‐
903 able(), extract_tagged(), extract_multiple(), gen_delimited_pat(),
904 and gen_extract_tagged(). With these, you can implement rather
905 advanced parsing algorithms. See Text::Balanced.
906 · "threads", by Arthur Bergman, is an interface to interpreter
908 threads. Interpreter threads (ithreads) is the new thread model
909 introduced in Perl 5.6 but only available as an internal interface
910 for extension writers (and for Win32 Perl for "fork()" emulation).
911 See threads, threads::shared, and perlthrtut.
912 · "threads::shared", by Arthur Bergman, allows data sharing for
914 interpreter threads. See threads::shared.
915 · "Tie::File", by Mark-Jason Dominus, associates a Perl array with
917 the lines of a file. See Tie::File.
918 · "Tie::Memoize", by Ilya Zakharevich, provides on-demand loaded
920 hashes. See Tie::Memoize.
921 · "Tie::RefHash::Nestable", by Edward Avis, allows storing hash ref‐
923 erences (unlike the standard Tie::RefHash) The module is contained
924 within Tie::RefHash. See Tie::RefHash.
925 · "Time::HiRes", by Douglas E. Wegscheid, provides high resolution
927 timing (ualarm, usleep, and gettimeofday). See Time::HiRes.
928 · "Unicode::UCD" offers a querying interface to the Unicode Character
930 Database. See Unicode::UCD.
931 · "Unicode::Collate", by SADAHIRO Tomoyuki, implements the UCA (Uni‐
933 code Collation Algorithm) for sorting Unicode strings. See Uni‐
934 code::Collate.
935 · "Unicode::Normalize", by SADAHIRO Tomoyuki, implements the various
937 Unicode normalization forms. See Unicode::Normalize.
938 · "XS::APItest", by Tim Jenness, is a test extension that exercises
940 XS APIs. Currently only "printf()" is tested: how to output vari‐
941 ous basic data types from XS.
942 · "XS::Typemap", by Tim Jenness, is a test extension that exercises
944 XS typemaps. Nothing gets installed, but the code is worth study‐
945 ing for extension writers.
946 Updated And Improved Modules and Pragmata
948 · The following independently supported modules have been updated to
950 the newest versions from CPAN: CGI, CPAN, DB_File, File::Spec,
951 File::Temp, Getopt::Long, Math::BigFloat, Math::BigInt, the podla‐
952 tors bundle (Pod::Man, Pod::Text), Pod::LaTeX [561+], Pod::Parser,
953 Storable, Term::ANSIColor, Test, Text-Tabs+Wrap.
954 · attributes::reftype() now works on tied arguments.
956 · AutoLoader can now be disabled with "no AutoLoader;".
958 · B::Deparse has been significantly enhanced by Robin Houston. It
960 can now deparse almost all of the standard test suite (so that the
961 tests still succeed). There is a make target "test.deparse" for
962 trying this out.
963 · Carp now has better interface documentation, and the @CARP_NOT
965 interface has been added to get optional control over where errors
966 are reported independently of @ISA, by Ben Tilly.
967 · Class::Struct can now define the classes in compile time.
969 · Class::Struct now assigns the array/hash element if the accessor is
971 called with an array/hash element as the sole argument.
972 · The return value of Cwd::fastcwd() is now tainted.
974 · Data::Dumper now has an option to sort hashes.
976 · Data::Dumper now has an option to dump code references using
978 B::Deparse.
979 · DB_File now supports newer Berkeley DB versions, among other
981 improvements.
982 · Devel::Peek now has an interface for the Perl memory statistics
984 (this works only if you are using perl's malloc, and if you have
985 compiled with debugging).
986 · The English module can now be used without the infamous performance
988 hit by saying
989 use English '-no_match_vars';
991 (Assuming, of course, that you don't need the troublesome variables
993 $`, $&, or $'.) Also, introduced @LAST_MATCH_START and
994 @LAST_MATCH_END English aliases for "@-" and "@+".
995 · ExtUtils::MakeMaker has been significantly cleaned up and fixed.
997 The enhanced version has also been backported to earlier releases
998 of Perl and submitted to CPAN so that the earlier releases can
999 enjoy the fixes.
1000 · The arguments of WriteMakefile() in Makefile.PL are now checked for
1002 sanity much more carefully than before. This may cause new warn‐
1003 ings when modules are being installed. See ExtUtils::MakeMaker for
1004 more details.
1005 · ExtUtils::MakeMaker now uses File::Spec internally, which hopefully
1007 leads to better portability.
1008 · Fcntl, Socket, and Sys::Syslog have been rewritten by Nicholas
1010 Clark to use the new-style constant dispatch section (see ExtU‐
1011 tils::Constant). This means that they will be more robust and
1012 hopefully faster.
1013 · File::Find now chdir()s correctly when chasing symbolic links.
1015 [561]
1016 · File::Find now has pre- and post-processing callbacks. It also
1018 correctly changes directories when chasing symbolic links. Call‐
1019 backs (naughtily) exiting with "next;" instead of "return;" now
1020 work.
1021 · File::Find is now (again) reentrant. It also has been made more
1023 portable.
1024 · The warnings issued by File::Find now belong to their own category.
1026 You can enable/disable them with "use/no warnings 'File::Find';".
1027 · File::Glob::glob() has been renamed to File::Glob::bsd_glob()
1029 because the name clashes with the builtin glob(). The older name
1030 is still available for compatibility, but is deprecated. [561]
1031 · File::Glob now supports "GLOB_LIMIT" constant to limit the size of
1033 the returned list of filenames.
1034 · IPC::Open3 now allows the use of numeric file descriptors.
1036 · IO::Socket now has an atmark() method, which returns true if the
1038 socket is positioned at the out-of-band mark. The method is also
1039 exportable as a sockatmark() function.
1040 · IO::Socket::INET failed to open the specified port if the service
1042 name was not known. It now correctly uses the supplied port number
1043 as is. [561]
1044 · IO::Socket::INET has support for the ReusePort option (if your
1046 platform supports it). The Reuse option now has an alias, Reuse‐
1047 Addr. For clarity, you may want to prefer ReuseAddr.
1048 · IO::Socket::INET now supports a value of zero for "LocalPort" (usu‐
1050 ally meaning that the operating system will make one up.)
1051 · 'use lib' now works identically to @INC. Removing directories with
1053 'no lib' now works.
1054 · Math::BigFloat and Math::BigInt have undergone a full rewrite by
1056 Tels. They are now magnitudes faster, and they support various
1057 bignum libraries such as GMP and PARI as their backends.
1058 · Math::Complex handles inf, NaN etc., better.
1060 · Net::Ping has been considerably enhanced by Rob Brown: multihoming
1062 is now supported, Win32 functionality is better, there is now time
1063 measuring functionality (optionally high-resolution using
1064 Time::HiRes), and there is now "external" protocol which uses
1065 Net::Ping::External module which runs your external ping utility
1066 and parses the output. A version of Net::Ping::External is avail‐
1067 able in CPAN.
1068 Note that some of the Net::Ping tests are disabled when running
1070 under the Perl distribution since one cannot assume one or more of
1071 the following: enabled echo port at localhost, full Internet con‐
1072 nectivity, or sympathetic firewalls. You can set the environment
1073 variable PERL_TEST_Net_Ping to "1" (one) before running the Perl
1074 test suite to enable all the Net::Ping tests.
1075 · POSIX::sigaction() is now much more flexible and robust. You can
1077 now install coderef handlers, 'DEFAULT', and 'IGNORE' handlers,
1078 installing new handlers was not atomic.
1079 · In Safe, %INC is now localised in a Safe compartment so that
1081 use/require work.
1082 · In SDBM_File on dosish platforms, some keys went missing because of
1084 lack of support for files with "holes". A workaround for the prob‐
1085 lem has been added.
1086 · In Search::Dict one can now have a pre-processing hook for the
1088 lines being searched.
1089 · The Shell module now has an OO interface.
1091 · In Sys::Syslog there is now a failover mechanism that will go
1093 through alternative connection mechanisms until the message is suc‐
1094 cessfully logged.
1095 · The Test module has been significantly enhanced.
1097 · Time::Local::timelocal() does not handle fractional seconds any‐
1099 more. The rationale is that neither does localtime(), and timelo‐
1100 cal() and localtime() are supposed to be inverses of each other.
1101 · The vars pragma now supports declaring fully qualified variables.
1103 (Something that "our()" does not and will not support.)
1104 · The "utf8::" name space (as in the pragma) provides various Perl-
1106 callable functions to provide low level access to Perl's internal
1107 Unicode representation. At the moment only length() has been
1108 implemented.
1109
1111 · Emacs perl mode (emacs/cperl-mode.el) has been updated to version
1112 4.31.
1113 · emacs/e2ctags.pl is now much faster.
1115 · "enc2xs" is a tool for people adding their own encodings to the
1117 Encode module.
1118 · "h2ph" now supports C trigraphs.
1120 · "h2xs" now produces a template README.
1122 · "h2xs" now uses "Devel::PPPort" for better portability between dif‐
1124 ferent versions of Perl.
1125 · "h2xs" uses the new ExtUtils::Constant module which will affect
1127 newly created extensions that define constants. Since the new code
1128 is more correct (if you have two constants where the first one is a
1129 prefix of the second one, the first constant never got defined),
1130 less lossy (it uses integers for integer constant, as opposed to
1131 the old code that used floating point numbers even for integer con‐
1132 stants), and slightly faster, you might want to consider regenerat‐
1133 ing your extension code (the new scheme makes regenerating easy).
1134 h2xs now also supports C trigraphs.
1135 · "libnetcfg" has been added to configure libnet.
1137 · "perlbug" is now much more robust. It also sends the bug report to
1139 perl.org, not perl.com.
1140 · "perlcc" has been rewritten and its user interface (that is, com‐
1142 mand line) is much more like that of the UNIX C compiler, cc. (The
1143 perlbc tools has been removed. Use "perlcc -B" instead.) Note
1144 that perlcc is still considered very experimental and unsupported.
1145 [561]
1146 · "perlivp" is a new Installation Verification Procedure utility for
1148 running any time after installing Perl.
1149 · "piconv" is an implementation of the character conversion utility
1151 "iconv", demonstrating the new Encode module.
1152 · "pod2html" now allows specifying a cache directory.
1154 · "pod2html" now produces XHTML 1.0.
1156 · "pod2html" now understands POD written using different line endings
1158 (PC-like CRLF versus UNIX-like LF versus MacClassic-like CR).
1159 · "s2p" has been completely rewritten in Perl. (It is in fact a full
1161 implementation of sed in Perl: you can use the sed functionality by
1162 using the "psed" utility.)
1163 · "xsubpp" now understands POD documentation embedded in the *.xs
1165 files. [561]
1166 · "xsubpp" now supports the OUT keyword.
1168
1170 · perl56delta details the changes between the 5.005 release and the
1171 5.6.0 release.
1172 · perlclib documents the internal replacements for standard C library
1174 functions. (Interesting only for extension writers and Perl core
1175 hackers.) [561+]
1176 · perldebtut is a Perl debugging tutorial. [561+]
1178 · perlebcdic contains considerations for running Perl on EBCDIC plat‐
1180 forms. [561+]
1181 · perlintro is a gentle introduction to Perl.
1183 · perliol documents the internals of PerlIO with layers.
1185 · perlmodstyle is a style guide for writing modules.
1187 · perlnewmod tells about writing and submitting a new module. [561+]
1189 · perlpacktut is a pack() tutorial.
1191 · perlpod has been rewritten to be clearer and to record the best
1193 practices gathered over the years.
1194 · perlpodspec is a more formal specification of the pod format,
1196 mainly of interest for writers of pod applications, not to people
1197 writing in pod.
1198 · perlretut is a regular expression tutorial. [561+]
1200 · perlrequick is a regular expressions quick-start guide. Yes, much
1202 quicker than perlretut. [561]
1203 · perltodo has been updated.
1205 · perltootc has been renamed as perltooc (to not to conflict with
1207 perltoot in filesystems restricted to "8.3" names).
1208 · perluniintro is an introduction to using Unicode in Perl. (perlu‐
1210 nicode is more of a detailed reference and background information)
1211 · perlutil explains the command line utilities packaged with the Perl
1213 distribution. [561+]
1214 The following platform-specific documents are available before the
1216 installation as README.platform, and after the installation as
1217 perlplatform:
1218 perlaix perlamiga perlapollo perlbeos perlbs2000
1220 perlce perlcygwin perldgux perldos perlepoc perlfreebsd perlhpux
1221 perlhurd perlirix perlmachten perlmacos perlmint perlmpeix
1222 perlnetware perlos2 perlos390 perlplan9 perlqnx perlsolaris
1223 perltru64 perluts perlvmesa perlvms perlvos perlwin32
1224 These documents usually detail one or more of the following subjects:
1226 configuring, building, testing, installing, and sometimes also using
1227 Perl on the said platform.
1228 Eastern Asian Perl users are now welcomed in their own languages:
1230 README.jp (Japanese), README.ko (Korean), README.cn (simplified Chi‐
1231 nese) and README.tw (traditional Chinese), which are written in normal
1232 pod but encoded in EUC-JP, EUC-KR, EUC-CN and Big5. These will get
1233 installed as
1234 perljp perlko perlcn perltw
1236 · The documentation for the POSIX-BC platform is called "BS2000", to
1238 avoid confusion with the Perl POSIX module.
1239 · The documentation for the WinCE platform is called perlce
1241 (README.ce in the source code kit), to avoid confusion with the
1242 perlwin32 documentation on 8.3-restricted filesystems.
1243
1245 · map() could get pathologically slow when the result list it gener‐
1246 ates is larger than the source list. The performance has been
1247 improved for common scenarios. [561]
1248 · sort() is also fully reentrant, in the sense that the sort function
1250 can itself call sort(). This did not work reliably in previous
1251 releases. [561]
1252 · sort() has been changed to use primarily mergesort internally as
1254 opposed to the earlier quicksort. For very small lists this may
1255 result in slightly slower sorting times, but in general the speedup
1256 should be at least 20%. Additional bonuses are that the worst case
1257 behaviour of sort() is now better (in computer science terms it now
1258 runs in time O(N log N), as opposed to quicksort's Theta(N**2)
1259 worst-case run time behaviour), and that sort() is now stable
1260 (meaning that elements with identical keys will stay ordered as
1261 they were before the sort). See the "sort" pragma for information.
1262 The story in more detail: suppose you want to serve yourself a lit‐
1264 tle slice of Pi.
1265 @digits = ( 3,1,4,1,5,9 );
1267 A numerical sort of the digits will yield (1,1,3,4,5,9), as
1269 expected. Which 1 comes first is hard to know, since one 1 looks
1270 pretty much like any other. You can regard this as totally triv‐
1271 ial, or somewhat profound. However, if you just want to sort the
1272 even digits ahead of the odd ones, then what will
1273 sort { ($a % 2) <=> ($b % 2) } @digits;
1275 yield? The only even digit, 4, will come first. But how about the
1277 odd numbers, which all compare equal? With the quicksort algorithm
1278 used to implement Perl 5.6 and earlier, the order of ties is left
1279 up to the sort. So, as you add more and more digits of Pi, the
1280 order in which the sorted even and odd digits appear will change.
1281 and, for sufficiently large slices of Pi, the quicksort algorithm
1282 in Perl 5.8 won't return the same results even if reinvoked with
1283 the same input. The justification for this rests with quicksort's
1284 worst case behavior. If you run
1285 sort { $a <=> $b } ( 1 .. $N , 1 .. $N );
1287 (something you might approximate if you wanted to merge two sorted
1289 arrays using sort), doubling $N doesn't just double the quicksort
1290 time, it quadruples it. Quicksort has a worst case run time that
1291 can grow like N**2, so-called quadratic behaviour, and it can hap‐
1292 pen on patterns that may well arise in normal use. You won't
1293 notice this for small arrays, but you will notice it with larger
1294 arrays, and you may not live long enough for the sort to complete
1295 on arrays of a million elements. So the 5.8 quicksort scrambles
1296 large arrays before sorting them, as a statistical defence against
1297 quadratic behaviour. But that means if you sort the same large
1298 array twice, ties may be broken in different ways.
1299 Because of the unpredictability of tie-breaking order, and the qua‐
1301 dratic worst-case behaviour, quicksort was almost replaced com‐
1302 pletely with a stable mergesort. Stable means that ties are broken
1303 to preserve the original order of appearance in the input array.
1304 So
1305 sort { ($a % 2) <=> ($b % 2) } (3,1,4,1,5,9);
1307 will yield (4,3,1,1,5,9), guaranteed. The even and odd numbers
1309 appear in the output in the same order they appeared in the input.
1310 Mergesort has worst case O(N log N) behaviour, the best value
1311 attainable. And, ironically, this mergesort does particularly well
1312 where quicksort goes quadratic: mergesort sorts (1..$N, 1..$N) in
1313 O(N) time. But quicksort was rescued at the last moment because it
1314 is faster than mergesort on certain inputs and platforms. For
1315 example, if you really don't care about the order of even and odd
1316 digits, quicksort will run in O(N) time; it's very good at sorting
1317 many repetitions of a small number of distinct elements. The
1318 quicksort divide and conquer strategy works well on platforms with
1319 relatively small, very fast, caches. Eventually, the problem gets
1320 whittled down to one that fits in the cache, from which point it
1321 benefits from the increased memory speed.
1322 Quicksort was rescued by implementing a sort pragma to control
1324 aspects of the sort. The stable subpragma forces stable behaviour,
1325 regardless of algorithm. The _quicksort and _mergesort subpragmas
1326 are heavy-handed ways to select the underlying implementation. The
1327 leading "_" is a reminder that these subpragmas may not survive
1328 beyond 5.8. More appropriate mechanisms for selecting the imple‐
1329 mentation exist, but they wouldn't have arrived in time to save
1330 quicksort.
1331 · Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm (
1333 http://burtleburtle.net/bob/hash/doobs.html ). This algorithm is
1334 reasonably fast while producing a much better spread of values than
1335 the old hashing algorithm (originally by Chris Torek, later tweaked
1336 by Ilya Zakharevich). Hash values output from the algorithm on a
1337 hash of all 3-char printable ASCII keys comes much closer to pass‐
1338 ing the DIEHARD random number generation tests. According to perl‐
1339 bench, this change has not affected the overall speed of Perl.
1340 · unshift() should now be noticeably faster.
1342
1344 Generic Improvements
1345 · INSTALL now explains how you can configure Perl to use 64-bit inte‐
1347 gers even on non-64-bit platforms.
1348 · Policy.sh policy change: if you are reusing a Policy.sh file (see
1350 INSTALL) and you use Configure -Dprefix=/foo/bar and in the old
1351 Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of
1352 them will now be changed to the new prefix, /foo/bar. (Previously
1353 only $prefix changed.) If you do not like this new behaviour,
1354 specify prefix, siteprefix, and vendorprefix explicitly.
1355 · A new optional location for Perl libraries, otherlibdirs, is avail‐
1357 able. It can be used for example for vendor add-ons without dis‐
1358 turbing Perl's own library directories.
1359 · In many platforms, the vendor-supplied 'cc' is too stripped-down to
1361 build Perl (basically, 'cc' doesn't do ANSI C). If this seems to
1362 be the case and 'cc' does not seem to be the GNU C compiler 'gcc',
1363 an automatic attempt is made to find and use 'gcc' instead.
1364 · gcc needs to closely track the operating system release to avoid
1366 build problems. If Configure finds that gcc was built for a differ‐
1367 ent operating system release than is running, it now gives a
1368 clearly visible warning that there may be trouble ahead.
1369 · Since Perl 5.8 is not binary-compatible with previous releases of
1371 Perl, Configure no longer suggests including the 5.005 modules in
1372 @INC.
1373 · Configure "-S" can now run non-interactively. [561]
1375 · Configure support for pdp11-style memory models has been removed
1377 due to obsolescence. [561]
1378 · configure.gnu now works with options with whitespace in them.
1380 · installperl now outputs everything to STDERR.
1382 · Because PerlIO is now the default on most platforms, "-perlio"
1384 doesn't get appended to the $Config{archname} (also known as $^O)
1385 anymore. Instead, if you explicitly choose not to use perlio (Con‐
1386 figure command line option -Uuseperlio), you will get "-stdio"
1387 appended.
1388 · Another change related to the architecture name is that "-64all"
1390 (-Duse64bitall, or "maximally 64-bit") is appended only if your
1391 pointers are 64 bits wide. (To be exact, the use64bitall is
1392 ignored.)
1393 · In AFS installations, one can configure the root of the AFS to be
1395 somewhere else than the default /afs by using the Configure parame‐
1396 ter "-Dafsroot=/some/where/else".
1397 · APPLLIB_EXP, a lesser-known configuration-time definition, has been
1399 documented. It can be used to prepend site-specific directories to
1400 Perl's default search path (@INC); see INSTALL for information.
1401 · The version of Berkeley DB used when the Perl (and, presumably, the
1403 DB_File extension) was built is now available as @Config{qw(db_ver‐
1404 sion_major db_version_minor db_version_patch)} from Perl and as
1405 "DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG DB_VERSION_PATCH_CFG"
1406 from C.
1407 · Building Berkeley DB3 for compatibility modes for DB, NDBM, and
1409 ODBM has been documented in INSTALL.
1410 · If you have CPAN access (either network or a local copy such as a
1412 CD-ROM) you can during specify extra modules to Configure to build
1413 and install with Perl using the -Dextras=... option. See INSTALL
1414 for more details.
1415 · In addition to config.over, a new override file, config.arch, is
1417 available. This file is supposed to be used by hints file writers
1418 for architecture-wide changes (as opposed to config.over which is
1419 for site-wide changes).
1420 · If your file system supports symbolic links, you can build Perl
1422 outside of the source directory by
1423 mkdir perl/build/directory
1425 cd perl/build/directory
1426 sh /path/to/perl/source/Configure -Dmksymlinks ...
1427 This will create in perl/build/directory a tree of symbolic links
1429 pointing to files in /path/to/perl/source. The original files are
1430 left unaffected. After Configure has finished, you can just say
1431 make all test
1433 and Perl will be built and tested, all in perl/build/directory.
1435 [561]
1436 · For Perl developers, several new make targets for profiling and
1438 debugging have been added; see perlhack.
1439 · Use of the gprof tool to profile Perl has been documented
1441 in perlhack. There is a make target called "perl.gprof"
1442 for generating a gprofiled Perl executable.
1443 · If you have GCC 3, there is a make target called
1445 "perl.gcov" for creating a gcoved Perl executable for cov‐
1446 erage analysis. See perlhack.
1447 · If you are on IRIX or Tru64 platforms, new profiling/debug‐
1449 ging options have been added; see perlhack for more infor‐
1450 mation about pixie and Third Degree.
1451 · Guidelines of how to construct minimal Perl installations have been
1453 added to INSTALL.
1454 · The Thread extension is now not built at all under ithreads ("Con‐
1456 figure -Duseithreads") because it wouldn't work anyway (the Thread
1457 extension requires being Configured with "-Duse5005threads").
1458 Note that the 5.005 threads are unsupported and deprecated: if you
1460 have code written for the old threads you should migrate it to the
1461 new ithreads model.
1462 · The Gconvert macro ($Config{d_Gconvert}) used by perl for stringi‐
1464 fying floating-point numbers is now more picky about using sprintf
1465 %.*g rules for the conversion. Some platforms that used to use
1466 gcvt may now resort to the slower sprintf.
1467 · The obsolete method of making a special (e.g., debugging) flavor of
1469 perl by saying
1470 make LIBPERL=libperld.a
1472 has been removed. Use -DDEBUGGING instead.
1474 New Or Improved Platforms
1476 For the list of platforms known to support Perl, see "Supported Plat‐
1478 forms" in perlport.
1479 · AIX dynamic loading should be now better supported.
1481 · AIX should now work better with gcc, threads, and 64-bitness. Also
1483 the long doubles support in AIX should be better now. See perlaix.
1484 · AtheOS ( http://www.atheos.cx/ ) is a new platform.
1486 · BeOS has been reclaimed.
1488 · The DG/UX platform now supports 5.005-style threads. See perldgux.
1490 · The DYNIX/ptx platform (also known as dynixptx) is supported at or
1492 near osvers 4.5.2.
1493 · EBCDIC platforms (z/OS (also known as OS/390), POSIX-BC, and
1495 VM/ESA) have been regained. Many test suite tests still fail and
1496 the co-existence of Unicode and EBCDIC isn't quite settled, but the
1497 situation is much better than with Perl 5.6. See perlos390,
1498 perlbs2000 (for POSIX-BC), and perlvmesa for more information.
1499 · Building perl with -Duseithreads or -Duse5005threads now works
1501 under HP-UX 10.20 (previously it only worked under 10.30 or later).
1502 You will need a thread library package installed. See README.hpux.
1503 [561]
1504 · Mac OS Classic is now supported in the mainstream source package
1506 (MacPerl has of course been available since perl 5.004 but now the
1507 source code bases of standard Perl and MacPerl have been synchro‐
1508 nised) [561]
1509 · Mac OS X (or Darwin) should now be able to build Perl even on HFS+
1511 filesystems. (The case-insensitivity used to confuse the Perl
1512 build process.)
1513 · NCR MP-RAS is now supported. [561]
1515 · All the NetBSD specific patches (except for the installation spe‐
1517 cific ones) have been merged back to the main distribution.
1518 · NetWare from Novell is now supported. See perlnetware.
1520 · NonStop-UX is now supported. [561]
1522 · NEC SUPER-UX is now supported.
1524 · All the OpenBSD specific patches (except for the installation spe‐
1526 cific ones) have been merged back to the main distribution.
1527 · Perl has been tested with the GNU pth userlevel thread package (
1529 http://www.gnu.org/software/pth/pth.html ). All thread tests of
1530 Perl now work, but not without adding some yield()s to the tests,
1531 so while pth (and other userlevel thread implementations) can be
1532 considered to be "working" with Perl ithreads, keep in mind the
1533 possible non-preemptability of the underlying thread implementa‐
1534 tion.
1535 · Stratus VOS is now supported using Perl's native build method (Con‐
1537 figure). This is the recommended method to build Perl on VOS. The
1538 older methods, which build miniperl, are still available. See per‐
1539 lvos. [561+]
1540 · The Amdahl UTS UNIX mainframe platform is now supported. [561]
1542 · WinCE is now supported. See perlce.
1544 · z/OS (formerly known as OS/390, formerly known as MVS OE) now has
1546 support for dynamic loading. This is not selected by default, how‐
1547 ever, you must specify -Dusedl in the arguments of Configure. [561]
1548
1550 Numerous memory leaks and uninitialized memory accesses have been
1551 hunted down. Most importantly, anonymous subs used to leak quite a
1552 bit. [561]
1553 · The autouse pragma didn't work for Multi::Part::Function::Names.
1555 · caller() could cause core dumps in certain situations. Carp was
1557 sometimes affected by this problem. In particular, caller() now
1558 returns a subroutine name of "(unknown)" for subroutines that have
1559 been removed from the symbol table.
1560 · chop(@list) in list context returned the characters chopped in
1562 reverse order. This has been reversed to be in the right order.
1563 [561]
1564 · Configure no longer includes the DBM libraries (dbm, gdbm, db,
1566 ndbm) when building the Perl binary. The only exception to this is
1567 SunOS 4.x, which needs them. [561]
1568 · The behaviour of non-decimal but numeric string constants such as
1570 "0x23" was platform-dependent: in some platforms that was seen as
1571 35, in some as 0, in some as a floating point number (don't ask).
1572 This was caused by Perl's using the operating system libraries in a
1573 situation where the result of the string to number conversion is
1574 undefined: now Perl consistently handles such strings as zero in
1575 numeric contexts.
1576 · Several debugger fixes: exit code now reflects the script exit
1578 code, condition "0" now treated correctly, the "d" command now
1579 checks line number, $. no longer gets corrupted, and all debugger
1580 output now goes correctly to the socket if RemotePort is set. [561]
1581 · The debugger (perl5db.pl) has been modified to present a more con‐
1583 sistent commands interface, via (CommandSet=580). perl5db.t was
1584 also added to test the changes, and as a placeholder for further
1585 tests.
1586 See perldebug.
1588 · The debugger has a new "dumpDepth" option to control the maximum
1590 depth to which nested structures are dumped. The "x" command has
1591 been extended so that "x N EXPR" dumps out the value of EXPR to a
1592 depth of at most N levels.
1593 · The debugger can now show lexical variables if you have the CPAN
1595 module PadWalker installed.
1596 · The order of DESTROYs has been made more predictable.
1598 · Perl 5.6.0 could emit spurious warnings about redefinition of
1600 dl_error() when statically building extensions into perl. This has
1601 been corrected. [561]
1602 · dprofpp -R didn't work.
1604 · *foo{FORMAT} now works.
1606 · Infinity is now recognized as a number.
1608 · UNIVERSAL::isa no longer caches methods incorrectly. (This broke
1610 the Tk extension with 5.6.0.) [561]
1611 · Lexicals I: lexicals outside an eval "" weren't resolved correctly
1613 inside a subroutine definition inside the eval "" if they were not
1614 already referenced in the top level of the eval""ed code.
1615 · Lexicals II: lexicals leaked at file scope into subroutines that
1617 were declared before the lexicals.
1618 · Lexical warnings now propagating correctly between scopes and into
1620 "eval "..."".
1621 · "use warnings qw(FATAL all)" did not work as intended. This has
1623 been corrected. [561]
1624 · warnings::enabled() now reports the state of $^W correctly if the
1626 caller isn't using lexical warnings. [561]
1627 · Line renumbering with eval and "#line" now works. [561]
1629 · Fixed numerous memory leaks, especially in eval "".
1631 · Localised tied variables no longer leak memory
1633 use Tie::Hash;
1635 tie my %tied_hash => 'Tie::StdHash';
1636 ...
1638 # Used to leak memory every time local() was called;
1640 # in a loop, this added up.
1641 local($tied_hash{Foo}) = 1;
1642 · Localised hash elements (and %ENV) are correctly unlocalised to not
1644 exist, if they didn't before they were localised.
1645 use Tie::Hash;
1647 tie my %tied_hash => 'Tie::StdHash';
1648 ...
1650 # Nothing has set the FOO element so far
1652 { local $tied_hash{FOO} = 'Bar' }
1654 # This used to print, but not now.
1656 print "exists!\n" if exists $tied_hash{FOO};
1657 As a side effect of this fix, tied hash interfaces must define the
1659 EXISTS and DELETE methods.
1660 · mkdir() now ignores trailing slashes in the directory name, as man‐
1662 dated by POSIX.
1663 · Some versions of glibc have a broken modfl(). This affects builds
1665 with "-Duselongdouble". This version of Perl detects this broken‐
1666 ness and has a workaround for it. The glibc release 2.2.2 is known
1667 to have fixed the modfl() bug.
1668 · Modulus of unsigned numbers now works (4063328477 % 65535 used to
1670 return 27406, instead of 27047). [561]
1671 · Some "not a number" warnings introduced in 5.6.0 eliminated to be
1673 more compatible with 5.005. Infinity is now recognised as a num‐
1674 ber. [561]
1675 · Numeric conversions did not recognize changes in the string value
1677 properly in certain circumstances. [561]
1678 · Attributes (such as :shared) didn't work with our().
1680 · our() variables will not cause bogus "Variable will not stay
1682 shared" warnings. [561]
1683 · "our" variables of the same name declared in two sibling blocks
1685 resulted in bogus warnings about "redeclaration" of the variables.
1686 The problem has been corrected. [561]
1687 · pack "Z" now correctly terminates the string with "\0".
1689 · Fix password routines which in some shadow password platforms (e.g.
1691 HP-UX) caused getpwent() to return every other entry.
1692 · The PERL5OPT environment variable (for passing command line argu‐
1694 ments to Perl) didn't work for more than a single group of options.
1695 [561]
1696 · PERL5OPT with embedded spaces didn't work.
1698 · printf() no longer resets the numeric locale to "C".
1700 · "qw(a\\b)" now parses correctly as 'a\\b': that is, as three char‐
1702 acters, not four. [561]
1703 · pos() did not return the correct value within s///ge in earlier
1705 versions. This is now handled correctly. [561]
1706 · Printing quads (64-bit integers) with printf/sprintf now works
1708 without the q L ll prefixes (assuming you are on a quad-capable
1709 platform).
1710 · Regular expressions on references and overloaded scalars now work.
1712 [561+]
1713 · Right-hand side magic (GMAGIC) could in many cases such as string
1715 concatenation be invoked too many times.
1716 · scalar() now forces scalar context even when used in void context.
1718 · SOCKS support is now much more robust.
1720 · sort() arguments are now compiled in the right wantarray context
1722 (they were accidentally using the context of the sort() itself).
1723 The comparison block is now run in scalar context, and the argu‐
1724 ments to be sorted are always provided list context. [561]
1725 · Changed the POSIX character class "[[:space:]]" to include the
1727 (very rarely used) vertical tab character. Added a new POSIX-ish
1728 character class "[[:blank:]]" which stands for horizontal white‐
1729 space (currently, the space and the tab).
1730 · The tainting behaviour of sprintf() has been rationalized. It does
1732 not taint the result of floating point formats anymore, making the
1733 behaviour consistent with that of string interpolation. [561]
1734 · Some cases of inconsistent taint propagation (such as within hash
1736 values) have been fixed.
1737 · The RE engine found in Perl 5.6.0 accidentally pessimised certain
1739 kinds of simple pattern matches. These are now handled better.
1740 [561]
1741 · Regular expression debug output (whether through "use re 'debug'"
1743 or via "-Dr") now looks better. [561]
1744 · Multi-line matches like ""a\nxb\n" =~ /(?!\A)x/m" were flawed. The
1746 bug has been fixed. [561]
1747 · Use of $& could trigger a core dump under some situations. This is
1749 now avoided. [561]
1750 · The regular expression captured submatches ($1, $2, ...) are now
1752 more consistently unset if the match fails, instead of leaving
1753 false data lying around in them. [561]
1754 · readline() on files opened in "slurp" mode could return an extra ""
1756 (blank line) at the end in certain situations. This has been cor‐
1757 rected. [561]
1758 · Autovivification of symbolic references of special variables
1760 described in perlvar (as in "${$num}") was accidentally disabled.
1761 This works again now. [561]
1762 · Sys::Syslog ignored the "LOG_AUTH" constant.
1764 · $AUTOLOAD, sort(), lock(), and spawning subprocesses in multiple
1766 threads simultaneously are now thread-safe.
1767 · Tie::Array's SPLICE method was broken.
1769 · Allow a read-only string on the left-hand side of a non-modifying
1771 tr///.
1772 · If "STDERR" is tied, warnings caused by "warn" and "die" now cor‐
1774 rectly pass to it.
1775 · Several Unicode fixes.
1777 · BOMs (byte order marks) at the beginning of Perl files
1779 (scripts, modules) should now be transparently skipped.
1780 UTF-16 and UCS-2 encoded Perl files should now be read cor‐
1781 rectly.
1782 · The character tables have been updated to Unicode 3.2.0.
1784 · Comparing with utf8 data does not magically upgrade
1786 non-utf8 data into utf8. (This was a problem for example
1787 if you were mixing data from I/O and Unicode data: your
1788 output might have got magically encoded as UTF-8.)
1789 · Generating illegal Unicode code points such as U+FFFE, or
1791 the UTF-16 surrogates, now also generates an optional warn‐
1792 ing.
1793 · "IsAlnum", "IsAlpha", and "IsWord" now match titlecase.
1795 · Concatenation with the "." operator or via variable inter‐
1797 polation, "eq", "substr", "reverse", "quotemeta", the "x"
1798 operator, substitution with "s///", single-quoted UTF-8,
1799 should now work.
1800 · The "tr///" operator now works. Note that the "tr///CU"
1802 functionality has been removed (but see pack('U0', ...)).
1803 · "eval "v200"" now works.
1805 · Perl 5.6.0 parsed m/\x{ab}/ incorrectly, leading to spuri‐
1807 ous warnings. This has been corrected. [561]
1808 · Zero entries were missing from the Unicode classes such as
1810 "IsDigit".
1811 · Large unsigned numbers (those above 2**31) could sometimes lose
1813 their unsignedness, causing bogus results in arithmetic operations.
1814 [561]
1815 · The Perl parser has been stress tested using both random input and
1817 Markov chain input and the few found crashes and lockups have been
1818 fixed.
1819 Platform Specific Changes and Fixes
1821 · BSDI 4.*
1823 Perl now works on post-4.0 BSD/OSes.
1825 · All BSDs
1827 Setting $0 now works (as much as possible; see perlvar for
1829 details).
1830 · Cygwin
1832 Numerous updates; currently synchronised with Cygwin 1.3.10.
1834 · Previously DYNIX/ptx had problems in its Configure probe for non-
1836 blocking I/O.
1837 · EPOC
1839 EPOC now better supported. See README.epoc. [561]
1841 · FreeBSD 3.*
1843 Perl now works on post-3.0 FreeBSDs.
1845 · HP-UX
1847 README.hpux updated; "Configure -Duse64bitall" now works; now uses
1849 HP-UX malloc instead of Perl malloc.
1850 · IRIX
1852 Numerous compilation flag and hint enhancements; accidental mixing
1854 of 32-bit and 64-bit libraries (a doomed attempt) made much harder.
1855 · Linux
1857 · Long doubles should now work (see INSTALL). [561]
1859 · Linux previously had problems related to sockaddrlen when
1861 using accept(), recvfrom() (in Perl: recv()), getpeer‐
1862 name(), and getsockname().
1863 · Mac OS Classic
1865 Compilation of the standard Perl distribution in Mac OS Classic
1867 should now work if you have the Metrowerks development environment
1868 and the missing Mac-specific toolkit bits. Contact the macperl
1869 mailing list for details.
1870 · MPE/iX
1872 MPE/iX update after Perl 5.6.0. See README.mpeix. [561]
1874 · NetBSD/threads: try installing the GNU pth (should be in the pack‐
1876 ages collection, or http://www.gnu.org/software/pth/), and Config‐
1877 ure with -Duseithreads.
1878 · NetBSD/sparc
1880 Perl now works on NetBSD/sparc.
1882 · OS/2
1884 Now works with usethreads (see INSTALL). [561]
1886 · Solaris
1888 64-bitness using the Sun Workshop compiler now works.
1890 · Stratus VOS
1892 The native build method requires at least VOS Release 14.5.0 and
1894 GNU C++/GNU Tools 2.0.1 or later. The Perl pack function now maps
1895 overflowed values to +infinity and underflowed values to -infinity.
1896 · Tru64 (aka Digital UNIX, aka DEC OSF/1)
1898 The operating system version letter now recorded in $Con‐
1900 fig{osvers}. Allow compiling with gcc (previously explicitly for‐
1901 bidden). Compiling with gcc still not recommended because buggy
1902 code results, even with gcc 2.95.2.
1903 · Unicos
1905 Fixed various alignment problems that lead into core dumps either
1907 during build or later; no longer dies on math errors at runtime;
1908 now using full quad integers (64 bits), previously was using only
1909 46 bit integers for speed.
1910 · VMS
1912 See "Socket Extension Dynamic in VMS" and "IEEE-format Floating
1914 Point Default on OpenVMS Alpha" for important changes not otherwise
1915 listed here.
1916 chdir() now works better despite a CRT bug; now works with MULTI‐
1918 PLICITY (see INSTALL); now works with Perl's malloc.
1919 The tainting of %ENV elements via "keys" or "values" was previously
1921 unimplemented. It now works as documented.
1922 The "waitpid" emulation has been improved. The worst bug (now
1924 fixed) was that a pid of -1 would cause a wildcard search of all
1925 processes on the system.
1926 POSIX-style signals are now emulated much better on VMS versions
1928 prior to 7.0.
1929 The "system" function and backticks operator have improved func‐
1931 tionality and better error handling. [561]
1932 File access tests now use current process privileges rather than
1934 the user's default privileges, which could sometimes result in a
1935 mismatch between reported access and actual access. This improve‐
1936 ment is only available on VMS v6.0 and later.
1937 There is a new "kill" implementation based on "sys$sigprc" that
1939 allows older VMS systems (pre-7.0) to use "kill" to send signals
1940 rather than simply force exit. This implementation also allows
1941 later systems to call "kill" from within a signal handler.
1942 Iterative logical name translations are now limited to 10 itera‐
1944 tions in imitation of SHOW LOGICAL and other OpenVMS facilities.
1945 · Windows
1947 · Signal handling now works better than it used to. It is
1949 now implemented using a Windows message loop, and is there‐
1950 fore less prone to random crashes.
1951 · fork() emulation is now more robust, but still continues to
1953 have a few esoteric bugs and caveats. See perlfork for
1954 details. [561+]
1955 · A failed (pseudo)fork now returns undef and sets errno to
1957 EAGAIN. [561]
1958 · The following modules now work on Windows:
1960 ExtUtils::Embed [561]
1962 IO::Pipe
1963 IO::Poll
1964 Net::Ping
1965 · IO::File::new_tmpfile() is no longer limited to 32767 invo‐
1967 cations per-process.
1968 · Better chdir() return value for a non-existent directory.
1970 · Compiling perl using the 64-bit Platform SDK tools is now
1972 supported.
1973 · The Win32::SetChildShowWindow() builtin can be used to con‐
1975 trol the visibility of windows created by child processes.
1976 See Win32 for details.
1977 · Non-blocking waits for child processes (or pseudo-pro‐
1979 cesses) are supported via "waitpid($pid, &POSIX::WNOHANG)".
1980 · The behavior of system() with multiple arguments has been
1982 rationalized. Each unquoted argument will be automatically
1983 quoted to protect whitespace, and any existing whitespace
1984 in the arguments will be preserved. This improves the
1985 portability of system(@args) by avoiding the need for Win‐
1986 dows "cmd" shell specific quoting in perl programs.
1987 Note that this means that some scripts that may have relied
1989 on earlier buggy behavior may no longer work correctly.
1990 For example, "system("nmake /nologo", @args)" will now
1991 attempt to run the file "nmake /nologo" and will fail when
1992 such a file isn't found. On the other hand, perl will now
1993 execute code such as "system("c:/Program
1994 Files/MyApp/foo.exe", @args)" correctly.
1995 · The perl header files no longer suppress common warnings
1997 from the Microsoft Visual C++ compiler. This means that
1998 additional warnings may now show up when compiling XS code.
1999 · Borland C++ v5.5 is now a supported compiler that can build
2001 Perl. However, the generated binaries continue to be
2002 incompatible with those generated by the other supported
2003 compilers (GCC and Visual C++). [561]
2004 · Duping socket handles with open(F, ">&MYSOCK") now works
2006 under Windows 9x. [561]
2007 · Current directory entries in %ENV are now correctly propa‐
2009 gated to child processes. [561]
2010 · New %ENV entries now propagate to subprocesses. [561]
2012 · Win32::GetCwd() correctly returns C:\ instead of C: when at
2014 the drive root. Other bugs in chdir() and Cwd::cwd() have
2015 also been fixed. [561]
2016 · The makefiles now default to the features enabled in
2018 ActiveState ActivePerl (a popular Win32 binary distribu‐
2019 tion). [561]
2020 · HTML files will now be installed in c:\perl\html instead of
2022 c:\perl\lib\pod\html
2023 · REG_EXPAND_SZ keys are now allowed in registry settings
2025 used by perl. [561]
2026 · Can now send() from all threads, not just the first one.
2028 [561]
2029 · ExtUtils::MakeMaker now uses $ENV{LIB} to search for
2031 libraries. [561]
2032 · Less stack reserved per thread so that more threads can run
2034 concurrently. (Still 16M per thread.) [561]
2035 · "File::Spec->tmpdir()" now prefers C:/temp over /tmp (works
2037 better when perl is running as service).
2038 · Better UNC path handling under ithreads. [561]
2040 · wait(), waitpid(), and backticks now return the correct
2042 exit status under Windows 9x. [561]
2043 · A socket handle leak in accept() has been fixed. [561]
2045
2047 Please see perldiag for more details.
2048 · Ambiguous range in the transliteration operator (like a-z-9) now
2050 gives a warning.
2051 · chdir("") and chdir(undef) now give a deprecation warning because
2053 they cause a possible unintentional chdir to the home directory.
2054 Say chdir() if you really mean that.
2055 · Two new debugging options have been added: if you have compiled
2057 your Perl with debugging, you can use the -DT [561] and -DR options
2058 to trace tokenising and to add reference counts to displaying vari‐
2059 ables, respectively.
2060 · The lexical warnings category "deprecated" is no longer a sub-cate‐
2062 gory of the "syntax" category. It is now a top-level category in
2063 its own right.
2064 · Unadorned dump() will now give a warning suggesting to use explicit
2066 CORE::dump() if that's what really is meant.
2067 · The "Unrecognized escape" warning has been extended to include
2069 "\8", "\9", and "\_". There is no need to escape any of the "\w"
2070 characters.
2071 · All regular expression compilation error messages are now hopefully
2073 easier to understand both because the error message now comes
2074 before the failed regex and because the point of failure is now
2075 clearly marked by a "<-- HERE" marker.
2076 · Various I/O (and socket) functions like binmode(), close(), and so
2078 forth now more consistently warn if they are used illogically
2079 either on a yet unopened or on an already closed filehandle (or
2080 socket).
2081 · Using lstat() on a filehandle now gives a warning. (It's a non-
2083 sensical thing to do.)
2084 · The "-M" and "-m" options now warn if you didn't supply the module
2086 name.
2087 · If you in "use" specify a required minimum version, modules match‐
2089 ing the name and but not defining a $VERSION will cause a fatal
2090 failure.
2091 · Using negative offset for vec() in lvalue context is now a warnable
2093 offense.
2094 · Odd number of arguments to overload::constant now elicits a warn‐
2096 ing.
2097 · Odd number of elements in anonymous hash now elicits a warning.
2099 · The various "opened only for", "on closed", "never opened" warnings
2101 drop the "main::" prefix for filehandles in the "main" package, for
2102 example "STDIN" instead of "main::STDIN".
2103 · Subroutine prototypes are now checked more carefully, you may get
2105 warnings for example if you have used non-prototype characters.
2106 · If an attempt to use a (non-blessed) reference as an array index is
2108 made, a warning is given.
2109 · "push @a;" and "unshift @a;" (with no values to push or unshift)
2111 now give a warning. This may be a problem for generated and evaled
2112 code.
2113 · If you try to "pack" in perlfunc a number less than 0 or larger
2115 than 255 using the "C" format you will get an optional warning.
2116 Similarly for the "c" format and a number less than -128 or more
2117 than 127.
2118 · pack "P" format now demands an explicit size.
2120 · unpack "w" now warns of unterminated compressed integers.
2122 · Warnings relating to the use of PerlIO have been added.
2124 · Certain regex modifiers such as "(?o)" make sense only if applied
2126 to the entire regex. You will get an optional warning if you try
2127 to do otherwise.
2128 · Variable length lookbehind has not yet been implemented, trying to
2130 use it will tell that.
2131 · Using arrays or hashes as references (e.g. "%foo->{bar}" has been
2133 deprecated for a while. Now you will get an optional warning.
2134 · Warnings relating to the use of the new restricted hashes feature
2136 have been added.
2137 · Self-ties of arrays and hashes are not supported and fatal errors
2139 will happen even at an attempt to do so.
2140 · Using "sort" in scalar context now issues an optional warning.
2142 This didn't do anything useful, as the sort was not performed.
2143 · Using the /g modifier in split() is meaningless and will cause a
2145 warning.
2146 · Using splice() past the end of an array now causes a warning.
2148 · Malformed Unicode encodings (UTF-8 and UTF-16) cause a lot of warn‐
2150 ings, as does trying to use UTF-16 surrogates (which are unimple‐
2151 mented).
2152 · Trying to use Unicode characters on an I/O stream without marking
2154 the stream's encoding (using open() or binmode()) will cause "Wide
2155 character" warnings.
2156 · Use of v-strings in use/require causes a (backward) portability
2158 warning.
2159 · Warnings relating to the use interpreter threads and their shared
2161 data have been added.
2162
2164 · PerlIO is now the default.
2165 · perlapi.pod (a companion to perlguts) now attempts to document the
2167 internal API.
2168 · You can now build a really minimal perl called microperl. Building
2170 microperl does not require even running Configure; "make -f Make‐
2171 file.micro" should be enough. Beware: microperl makes many assump‐
2172 tions, some of which may be too bold; the resulting executable may
2173 crash or otherwise misbehave in wondrous ways. For careful hackers
2174 only.
2175 · Added rsignal(), whichsig(), do_join(), op_clear, op_null, ptr_ta‐
2177 ble_clear(), ptr_table_free(), sv_setref_uv(), and several UTF-8
2178 interfaces to the publicised API. For the full list of the avail‐
2179 able APIs see perlapi.
2180 · Made possible to propagate customised exceptions via croak()ing.
2182 · Now xsubs can have attributes just like subs. (Well, at least the
2184 built-in attributes.)
2185 · dTHR and djSP have been obsoleted; the former removed (because it's
2187 a no-op) and the latter replaced with dSP.
2188 · PERL_OBJECT has been completely removed.
2190 · The MAGIC constants (e.g. 'P') have been macrofied (e.g.
2192 "PERL_MAGIC_TIED") for better source code readability and maintain‐
2193 ability.
2194 · The regex compiler now maintains a structure that identifies nodes
2196 in the compiled bytecode with the corresponding syntactic features
2197 of the original regex expression. The information is attached to
2198 the new "offsets" member of the "struct regexp". See perldebguts
2199 for more complete information.
2200 · The C code has been made much more "gcc -Wall" clean. Some warning
2202 messages still remain in some platforms, so if you are compiling
2203 with gcc you may see some warnings about dubious practices. The
2204 warnings are being worked on.
2205 · perly.c, sv.c, and sv.h have now been extensively commented.
2207 · Documentation on how to use the Perl source repository has been
2209 added to Porting/repository.pod.
2210 · There are now several profiling make targets.
2212
2214 (This change was already made in 5.7.0 but bears repeating here.)
2215 (5.7.0 came out before 5.6.1: the development branch 5.7 released ear‐
2216 lier than the maintenance branch 5.6)
2217 A potential security vulnerability in the optional suidperl component
2219 of Perl was identified in August 2000. suidperl is neither built nor
2220 installed by default. As of November 2001 the only known vulnerable
2221 platform is Linux, most likely all Linux distributions. CERT and vari‐
2222 ous vendors and distributors have been alerted about the vulnerability.
2223 See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt
2224 for more information.
2225 The problem was caused by Perl trying to report a suspected security
2227 exploit attempt using an external program, /bin/mail. On Linux plat‐
2228 forms the /bin/mail program had an undocumented feature which when com‐
2229 bined with suidperl gave access to a root shell, resulting in a serious
2230 compromise instead of reporting the exploit attempt. If you don't have
2231 /bin/mail, or if you have 'safe setuid scripts', or if suidperl is not
2232 installed, you are safe.
2233 The exploit attempt reporting feature has been completely removed from
2235 Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also
2236 from all the Perl 5.7 releases), so that particular vulnerability isn't
2237 there anymore. However, further security vulnerabilities are, unfortu‐
2238 nately, always possible. The suidperl functionality is most probably
2239 going to be removed in Perl 5.10. In any case, suidperl should only be
2240 used by security experts who know exactly what they are doing and why
2241 they are using suidperl instead of some other solution such as sudo (
2242 see http://www.courtesan.com/sudo/ ).
2243
2245 Several new tests have been added, especially for the lib and ext sub‐
2246 sections. There are now about 69 000 individual tests (spread over
2247 about 700 test scripts), in the regression suite (5.6.1 has about 11
2248 700 tests, in 258 test scripts) The exact numbers depend on the plat‐
2249 form and Perl configuration used. Many of the new tests are of course
2250 introduced by the new modules, but still in general Perl is now more
2251 thoroughly tested.
2252 Because of the large number of tests, running the regression suite will
2254 take considerably longer time than it used to: expect the suite to take
2255 up to 4-5 times longer to run than in perl 5.6. On a really fast
2256 machine you can hope to finish the suite in about 6-8 minutes (wall‐
2257 clock time).
2258 The tests are now reported in a different order than in earlier Perls.
2260 (This happens because the test scripts from under t/lib have been moved
2261 to be closer to the library/extension they are testing.)
2262
2264 The Compiler Suite Is Still Very Experimental
2265 The compiler suite is slowly getting better but it continues to be
2267 highly experimental. Use in production environments is discouraged.
2268 Localising Tied Arrays and Hashes Is Broken
2270 local %tied_array;
2272 doesn't work as one would expect: the old value is restored incor‐
2274 rectly. This will be changed in a future release, but we don't know
2275 yet what the new semantics will exactly be. In any case, the change
2276 will break existing code that relies on the current (ill-defined)
2277 semantics, so just avoid doing this in general.
2278 Building Extensions Can Fail Because Of Largefiles
2280 Some extensions like mod_perl are known to have issues with `large‐
2282 files', a change brought by Perl 5.6.0 in which file offsets default to
2283 64 bits wide, where supported. Modules may fail to compile at all, or
2284 they may compile and work incorrectly. Currently, there is no good
2285 solution for the problem, but Configure now provides appropriate non-
2286 largefile ccflags, ldflags, libswanted, and libs in the %Config hash
2287 (e.g., $Config{ccflags_nolargefiles}) so the extensions that are having
2288 problems can try configuring themselves without the largefileness.
2289 This is admittedly not a clean solution, and the solution may not even
2290 work at all. One potential failure is whether one can (or, if one can,
2291 whether it's a good idea to) link together at all binaries with differ‐
2292 ent ideas about file offsets; all this is platform-dependent.
2293 Modifying $_ Inside for(..)
2295 for (1..5) { $_++ }
2297 works without complaint. It shouldn't. (You should be able to modify
2299 only lvalue elements inside the loops.) You can see the correct behav‐
2300 iour by replacing the 1..5 with 1, 2, 3, 4, 5.
2301 mod_perl 1.26 Doesn't Build With Threaded Perl
2303 Use mod_perl 1.27 or higher.
2305 lib/ftmp-security tests warn 'system possibly insecure'
2307 Don't panic. Read the 'make test' section of INSTALL instead.
2309 libwww-perl (LWP) fails base/date #51
2311 Use libwww-perl 5.65 or later.
2313 PDL failing some tests
2315 Use PDL 2.3.4 or later.
2317 Perl_get_sv
2319 You may get errors like 'Undefined symbol "Perl_get_sv"' or "can't
2321 resolve symbol 'Perl_get_sv'", or the symbol may be "Perl_sv_2pv".
2322 This probably means that you are trying to use an older shared Perl
2323 library (or extensions linked with such) with Perl 5.8.0 executable.
2324 Perl used to have such a subroutine, but that is no more the case.
2325 Check your shared library path, and any shared Perl libraries in those
2326 directories.
2327 Sometimes this problem may also indicate a partial Perl 5.8.0 installa‐
2329 tion, see "Mac OS X dyld undefined symbols" for an example and how to
2330 deal with it.
2331 Self-tying Problems
2333 Self-tying of arrays and hashes is broken in rather deep and hard-to-
2335 fix ways. As a stop-gap measure to avoid people from getting frus‐
2336 trated at the mysterious results (core dumps, most often), it is for‐
2337 bidden for now (you will get a fatal error even from an attempt).
2338 A change to self-tying of globs has caused them to be recursively ref‐
2340 erenced (see: "Two-Phased Garbage Collection" in perlobj). You will
2341 now need an explicit untie to destroy a self-tied glob. This behaviour
2342 may be fixed at a later date.
2343 Self-tying of scalars and IO thingies works.
2345 ext/threads/t/libc
2347 If this test fails, it indicates that your libc (C library) is not
2349 threadsafe. This particular test stress tests the localtime() call to
2350 find out whether it is threadsafe. See perlthrtut for more informa‐
2351 tion.
2352 Failure of Thread (5.005-style) tests
2354 Note that support for 5.005-style threading is deprecated, experimental
2356 and practically unsupported. In 5.10, it is expected to be removed.
2357 You should migrate your code to ithreads.
2358 The following tests are known to fail due to fundamental problems in
2360 the 5.005 threading implementation. These are not new failures--Perl
2361 5.005_0x has the same bugs, but didn't have these tests.
2362 ../ext/B/t/xref.t 255 65280 14 12 85.71% 3-14
2364 ../ext/List/Util/t/first.t 255 65280 7 4 57.14% 2 5-7
2365 ../lib/English.t 2 512 54 2 3.70% 2-3
2366 ../lib/FileCache.t 5 1 20.00% 5
2367 ../lib/Filter/Simple/t/data.t 6 3 50.00% 1-3
2368 ../lib/Filter/Simple/t/filter_only. 9 3 33.33% 1-2 5
2369 ../lib/Math/BigInt/t/bare_mbf.t 1627 4 0.25% 8 11 1626-1627
2370 ../lib/Math/BigInt/t/bigfltpm.t 1629 4 0.25% 10 13 1628-
2371 1629
2372 ../lib/Math/BigInt/t/sub_mbf.t 1633 4 0.24% 8 11 1632-1633
2373 ../lib/Math/BigInt/t/with_sub.t 1628 4 0.25% 9 12 1627-1628
2374 ../lib/Tie/File/t/31_autodefer.t 255 65280 65 32 49.23% 34-65
2375 ../lib/autouse.t 10 1 10.00% 4
2376 op/flip.t 15 1 6.67% 15
2377 These failures are unlikely to get fixed as 5.005-style threads are
2379 considered fundamentally broken. (Basically what happens is that com‐
2380 peting threads can corrupt shared global state, one good example being
2381 regular expression engine's state.)
2382 Timing problems
2384 The following tests may fail intermittently because of timing problems,
2386 for example if the system is heavily loaded.
2387 t/op/alarm.t
2389 ext/Time/HiRes/HiRes.t
2390 lib/Benchmark.t
2391 lib/Memoize/t/expmod_t.t
2392 lib/Memoize/t/speed.t
2393 In case of failure please try running them manually, for example
2395 ./perl -Ilib ext/Time/HiRes/HiRes.t
2397 Tied/Magical Array/Hash Elements Do Not Autovivify
2399 For normal arrays "$foo = \$bar[1]" will assign "undef" to $bar[1]
2401 (assuming that it didn't exist before), but for tied/magical arrays and
2402 hashes such autovivification does not happen because there is currently
2403 no way to catch the reference creation. The same problem affects slic‐
2404 ing over non-existent indices/keys of a tied/magical array/hash.
2405 Unicode in package/class and subroutine names does not work
2407 One can have Unicode in identifier names, but not in package/class or
2409 subroutine names. While some limited functionality towards this does
2410 exist as of Perl 5.8.0, that is more accidental than designed; use of
2411 Unicode for the said purposes is unsupported.
2412 One reason of this unfinishedness is its (currently) inherent unporta‐
2414 bility: since both package names and subroutine names may need to be
2415 mapped to file and directory names, the Unicode capability of the
2416 filesystem becomes important-- and there unfortunately aren't portable
2417 answers.
2418
2420 AIX
2421 · If using the AIX native make command, instead of just "make" issue
2423 "make all". In some setups the former has been known to spuriously
2424 also try to run "make install". Alternatively, you may want to use
2425 GNU make.
2426 · In AIX 4.2, Perl extensions that use C++ functions that use statics
2428 may have problems in that the statics are not getting initialized.
2429 In newer AIX releases, this has been solved by linking Perl with
2430 the libC_r library, but unfortunately in AIX 4.2 the said library
2431 has an obscure bug where the various functions related to time
2432 (such as time() and gettimeofday()) return broken values, and
2433 therefore in AIX 4.2 Perl is not linked against libC_r.
2434 · vac 5.0.0.0 May Produce Buggy Code For Perl
2436 The AIX C compiler vac version 5.0.0.0 may produce buggy code,
2438 resulting in a few random tests failing when run as part of "make
2439 test", but when the failing tests are run by hand, they succeed.
2440 We suggest upgrading to at least vac version 5.0.1.0, that has been
2441 known to compile Perl correctly. "lslpp -L⎪grep vac.C" will tell
2442 you the vac version. See README.aix.
2443 · If building threaded Perl, you may get compilation warning from
2445 pp_sys.c:
2446 "pp_sys.c", line 4651.39: 1506-280 (W) Function argument assignment between types "unsigned char*" and "const void*" is not allowed.
2448 This is harmless; it is caused by the getnetbyaddr() and getnet‐
2450 byaddr_r() having slightly different types for their first argu‐
2451 ment.
2452 Alpha systems with old gccs fail several tests
2454 If you see op/pack, op/pat, op/regexp, or ext/Storable tests failing in
2456 a Linux/alpha or *BSD/Alpha, it's probably time to upgrade your gcc.
2457 gccs prior to 2.95.3 are definitely not good enough, and gcc 3.1 may be
2458 even better. (RedHat Linux/alpha with gcc 3.1 reported no problems, as
2459 did Linux 2.4.18 with gcc 2.95.4.) (In Tru64, it is preferable to use
2460 the bundled C compiler.)
2461 AmigaOS
2463 Perl 5.8.0 doesn't build in AmigaOS. It broke at some point during the
2465 ithreads work and we could not find Amiga experts to unbreak the prob‐
2466 lems. Perl 5.6.1 still works for AmigaOS (as does the 5.7.2 develop‐
2467 ment release).
2468 BeOS
2470 The following tests fail on 5.8.0 Perl in BeOS Personal 5.03:
2472 t/op/lfs............................FAILED at test 17
2474 t/op/magic..........................FAILED at test 24
2475 ext/Fcntl/t/syslfs..................FAILED at test 17
2476 ext/File/Glob/t/basic...............FAILED at test 3
2477 ext/POSIX/t/sigaction...............FAILED at test 13
2478 ext/POSIX/t/waitpid.................FAILED at test 1
2479 See perlbeos (README.beos) for more details.
2481 Cygwin "unable to remap"
2483 For example when building the Tk extension for Cygwin, you may get an
2485 error message saying "unable to remap". This is known problem with
2486 Cygwin, and a workaround is detailed in here: http://sources.red‐
2487 hat.com/ml/cygwin/2001-12/msg00894.html
2488 Cygwin ndbm tests fail on FAT
2490 One can build but not install (or test the build of) the NDBM_File on
2492 FAT filesystems. Installation (or build) on NTFS works fine. If one
2493 attempts the test on a FAT install (or build) the following failures
2494 are expected:
2495 ../ext/NDBM_File/ndbm.t 13 3328 71 59 83.10% 1-2 4 16-71
2497 ../ext/ODBM_File/odbm.t 255 65280 ?? ?? % ??
2498 ../lib/AnyDBM_File.t 2 512 12 2 16.67% 1 4
2499 ../lib/Memoize/t/errors.t 0 139 11 5 45.45% 7-11
2500 ../lib/Memoize/t/tie_ndbm.t 13 3328 4 4 100.00% 1-4
2501 run/fresh_perl.t 97 1 1.03% 91
2502 NDBM_File fails and ODBM_File just coredumps.
2504 If you intend to run only on FAT (or if using AnyDBM_File on FAT), run
2506 Configure with the -Ui_ndbm and -Ui_dbm options to prevent NDBM_File
2507 and ODBM_File being built.
2508 DJGPP Failures
2510 t/op/stat............................FAILED at test 29
2512 lib/File/Find/t/find.................FAILED at test 1
2513 lib/File/Find/t/taint................FAILED at test 1
2514 lib/h2xs.............................FAILED at test 15
2515 lib/Pod/t/eol........................FAILED at test 1
2516 lib/Test/Harness/t/strap-analyze.....FAILED at test 8
2517 lib/Test/Harness/t/test-harness......FAILED at test 23
2518 lib/Test/Simple/t/exit...............FAILED at test 1
2519 The above failures are known as of 5.8.0 with native builds with long
2521 filenames, but there are a few more if running under dosemu because of
2522 limitations (and maybe bugs) of dosemu:
2523 t/comp/cpp...........................FAILED at test 3
2525 t/op/inccode.........................(crash)
2526 and a few lib/ExtUtils tests, and several hundred Encode/t/Aliases.t
2528 failures that work fine with long filenames. So you really might pre‐
2529 fer native builds and long filenames.
2530 FreeBSD built with ithreads coredumps reading large directories
2532 This is a known bug in FreeBSD 4.5's readdir_r(), it has been fixed in
2534 FreeBSD 4.6 (see perlfreebsd (README.freebsd)).
2535 FreeBSD Failing locale Test 117 For ISO 8859-15 Locales
2537 The ISO 8859-15 locales may fail the locale test 117 in FreeBSD. This
2539 is caused by the characters \xFF (y with diaeresis) and \xBE (Y with
2540 diaeresis) not behaving correctly when being matched case-insensi‐
2541 tively. Apparently this problem has been fixed in the latest FreeBSD
2542 releases. ( http://www.freebsd.org/cgi/query-pr.cgi?pr=34308 )
2543 IRIX fails ext/List/Util/t/shuffle.t or Digest::MD5
2545 IRIX with MIPSpro 7.3.1.2m or 7.3.1.3m compiler may fail the List::Util
2547 test ext/List/Util/t/shuffle.t by dumping core. This seems to be a
2548 compiler error since if compiled with gcc no core dump ensues, and no
2549 failures have been seen on the said test on any other platform.
2550 Similarly, building the Digest::MD5 extension has been known to fail
2552 with "*** Termination code 139 (bu21)".
2553 The cure is to drop optimization level (Configure -Doptimize=-O2).
2555 HP-UX lib/posix Subtest 9 Fails When LP64-Configured
2557 If perl is configured with -Duse64bitall, the successful result of the
2559 subtest 10 of lib/posix may arrive before the successful result of the
2560 subtest 9, which confuses the test harness so much that it thinks the
2561 subtest 9 failed.
2562 Linux with glibc 2.2.5 fails t/op/int subtest #6 with -Duse64bitint
2564 This is a known bug in the glibc 2.2.5 with long long integers. (
2566 http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=65612 )
2567 Linux With Sfio Fails op/misc Test 48
2569 No known fix.
2571 Mac OS X
2573 Please remember to set your environment variable LC_ALL to "C" (setenv
2575 LC_ALL C) before running "make test" to avoid a lot of warnings about
2576 the broken locales of Mac OS X.
2577 The following tests are known to fail in Mac OS X 10.1.5 because of
2579 buggy (old) implementations of Berkeley DB included in Mac OS X:
2580 Failed Test Stat Wstat Total Fail Failed List of Failed
2582 -------------------------------------------------------------------------
2583 ../ext/DB_File/t/db-btree.t 0 11 ?? ?? % ??
2584 ../ext/DB_File/t/db-recno.t 149 3 2.01% 61 63 65
2585 If you are building on a UFS partition, you will also probably see
2587 t/op/stat.t subtest #9 fail. This is caused by Darwin's UFS not sup‐
2588 porting inode change time.
2589 Also the ext/POSIX/t/posix.t subtest #10 fails but it is skipped for
2591 now because the failure is Apple's fault, not Perl's (blocked signals
2592 are lost).
2593 If you Configure with ithreads, ext/threads/t/libc.t will fail. Again,
2595 this is not Perl's fault-- the libc of Mac OS X is not threadsafe (in
2596 this particular test, the localtime() call is found to be threadun‐
2597 safe.)
2598 Mac OS X dyld undefined symbols
2600 If after installing Perl 5.8.0 you are getting warnings about missing
2602 symbols, for example
2603 dyld: perl Undefined symbols
2605 _perl_sv_2pv
2606 _perl_get_sv
2607 you probably have an old pre-Perl-5.8.0 installation (or parts of one)
2609 in /Library/Perl (the undefined symbols used to exist in pre-5.8.0
2610 Perls). It seems that for some reason "make install" doesn't always
2611 completely overwrite the files in /Library/Perl. You can move the old
2612 Perl shared library out of the way like this:
2613 cd /Library/Perl/darwin/CORE
2615 mv libperl.dylib libperlold.dylib
2616 and then reissue "make install". Note that the above of course is
2618 extremely disruptive for anything using the /usr/local/bin/perl. If
2619 that doesn't help, you may have to try removing all the .bundle files
2620 from beneath /Library/Perl, and again "make install"-ing.
2621 OS/2 Test Failures
2623 The following tests are known to fail on OS/2 (for clarity only the
2625 failures are shown, not the full error messages):
2626 ../lib/ExtUtils/t/Mkbootstrap.t 1 256 18 1 5.56% 8
2628 ../lib/ExtUtils/t/Packlist.t 1 256 34 1 2.94% 17
2629 ../lib/ExtUtils/t/basic.t 1 256 17 1 5.88% 14
2630 lib/os2_process.t 2 512 227 2 0.88% 174 209
2631 lib/os2_process_kid.t 227 2 0.88% 174 209
2632 lib/rx_cmprt.t 255 65280 18 3 16.67% 16-18
2633 op/sprintf tests 91, 129, and 130
2635 The op/sprintf tests 91, 129, and 130 are known to fail on some plat‐
2637 forms. Examples include any platform using sfio, and Compaq/Tandem's
2638 NonStop-UX.
2639 Test 91 is known to fail on QNX6 (nto), because "sprintf '%e',0" incor‐
2641 rectly produces 0.000000e+0 instead of 0.000000e+00.
2642 For tests 129 and 130, the failing platforms do not comply with the
2644 ANSI C Standard: lines 19ff on page 134 of ANSI X3.159 1989, to be
2645 exact. (They produce something other than "1" and "-1" when formatting
2646 0.6 and -0.6 using the printf format "%.0f"; most often, they produce
2647 "0" and "-0".)
2648 SCO
2650 The socketpair tests are known to be unhappy in SCO 3.2v5.0.4:
2652 ext/Socket/socketpair.t...............FAILED tests 15-45
2654 Solaris 2.5
2656 In case you are still using Solaris 2.5 (aka SunOS 5.5), you may expe‐
2658 rience failures (the test core dumping) in lib/locale.t. The suggested
2659 cure is to upgrade your Solaris.
2660 Solaris x86 Fails Tests With -Duse64bitint
2662 The following tests are known to fail in Solaris x86 with Perl config‐
2664 ured to use 64 bit integers:
2665 ext/Data/Dumper/t/dumper.............FAILED at test 268
2667 ext/Devel/Peek/Peek..................FAILED at test 7
2668 SUPER-UX (NEC SX)
2670 The following tests are known to fail on SUPER-UX:
2672 op/64bitint...........................FAILED tests 29-30, 32-33, 35-36
2674 op/arith..............................FAILED tests 128-130
2675 op/pack...............................FAILED tests 25-5625
2676 op/pow................................
2677 op/taint..............................# msgsnd failed
2678 ../ext/IO/lib/IO/t/io_poll............FAILED tests 3-4
2679 ../ext/IPC/SysV/ipcsysv...............FAILED tests 2, 5-6
2680 ../ext/IPC/SysV/t/msg.................FAILED tests 2, 4-6
2681 ../ext/Socket/socketpair..............FAILED tests 12
2682 ../lib/IPC/SysV.......................FAILED tests 2, 5-6
2683 ../lib/warnings.......................FAILED tests 115-116, 118-119
2684 The op/pack failure ("Cannot compress negative numbers at op/pack.t
2686 line 126") is serious but as of yet unsolved. It points at some prob‐
2687 lems with the signedness handling of the C compiler, as do the
2688 64bitint, arith, and pow failures. Most of the rest point at problems
2689 with SysV IPC.
2690 Term::ReadKey not working on Win32
2692 Use Term::ReadKey 2.20 or later.
2694 UNICOS/mk
2696 · During Configure, the test
2698 Guessing which symbols your C compiler and preprocessor define...
2700 will probably fail with error messages like
2702 CC-20 cc: ERROR File = try.c, Line = 3
2704 The identifier "bad" is undefined.
2705 bad switch yylook 79bad switch yylook 79bad switch yylook 79bad switch yylook 79#ifdef A29K
2707 ^
2708 CC-65 cc: ERROR File = try.c, Line = 3
2710 A semicolon is expected at this point.
2711 This is caused by a bug in the awk utility of UNICOS/mk. You can
2713 ignore the error, but it does cause a slight problem: you cannot
2714 fully benefit from the h2ph utility (see h2ph) that can be used to
2715 convert C headers to Perl libraries, mainly used to be able to
2716 access from Perl the constants defined using C preprocessor, cpp.
2717 Because of the above error, parts of the converted headers will be
2718 invisible. Luckily, these days the need for h2ph is rare.
2719 · If building Perl with interpreter threads (ithreads), the get‐
2721 grent(), getgrnam(), and getgrgid() functions cannot return the
2722 list of the group members due to a bug in the multithreaded support
2723 of UNICOS/mk. What this means is that in list context the func‐
2724 tions will return only three values, not four.
2725 UTS
2727 There are a few known test failures, see perluts (README.uts).
2729 VOS (Stratus)
2731 When Perl is built using the native build process on VOS Release 14.5.0
2733 and GNU C++/GNU Tools 2.0.1, all attempted tests either pass or result
2734 in TODO (ignored) failures.
2735 VMS
2737 There should be no reported test failures with a default configuration,
2739 though there are a number of tests marked TODO that point to areas
2740 needing further debugging and/or porting work.
2741 Win32
2743 In multi-CPU boxes, there are some problems with the I/O buffering:
2745 some output may appear twice.
2746 XML::Parser not working
2748 Use XML::Parser 2.31 or later.
2750 z/OS (OS/390)
2752 z/OS has rather many test failures but the situation is actually much
2754 better than it was in 5.6.0; it's just that so many new modules and
2755 tests have been added.
2756 Failed Test Stat Wstat Total Fail Failed List of Failed
2758 ---------------------------------------------------------------------------
2759 ../ext/Data/Dumper/t/dumper.t 357 8 2.24% 311 314 325 327
2760 331 333 337 339
2761 ../ext/IO/lib/IO/t/io_unix.t 5 4 80.00% 2-5
2762 ../ext/Storable/t/downgrade.t 12 3072 169 12 7.10% 14-15 46-47 78-79
2763 110-111 150 161
2764 ../lib/ExtUtils/t/Constant.t 121 30976 48 48 100.00% 1-48
2765 ../lib/ExtUtils/t/Embed.t 9 9 100.00% 1-9
2766 op/pat.t 922 7 0.76% 665 776 785 832-
2767 834 845
2768 op/sprintf.t 224 3 1.34% 98 100 136
2769 op/tr.t 97 5 5.15% 63 71-74
2770 uni/fold.t 780 6 0.77% 61 169 196 661
2771 710-711
2772 The failures in dumper.t and downgrade.t are problems in the tests,
2774 those in io_unix and sprintf are problems in the USS (UDP sockets and
2775 printf formats). The pat, tr, and fold failures are genuine Perl prob‐
2776 lems caused by EBCDIC (and in the pat and fold cases, combining that
2777 with Unicode). The Constant and Embed are probably problems in the
2778 tests (since they test Perl's ability to build extensions, and that
2779 seems to be working reasonably well.)
2780 Unicode Support on EBCDIC Still Spotty
2782 Though mostly working, Unicode support still has problem spots on
2784 EBCDIC platforms. One such known spot are the "\p{}" and "\P{}" regu‐
2785 lar expression constructs for code points less than 256: the "pP" are
2786 testing for Unicode code points, not knowing about EBCDIC.
2787 Seen In Perl 5.7 But Gone Now
2789 "Time::Piece" (previously known as "Time::Object") was removed because
2791 it was felt that it didn't have enough value in it to be a core module.
2792 It is still a useful module, though, and is available from the CPAN.
2793 Perl 5.8 unfortunately does not build anymore on AmigaOS; this broke
2795 accidentally at some point. Since there are not that many Amiga devel‐
2796 opers available, we could not get this fixed and tested in time for
2797 5.8.0. Perl 5.6.1 still works for AmigaOS (as does the 5.7.2 develop‐
2798 ment release).
2799 The "PerlIO::Scalar" and "PerlIO::Via" (capitalised) were renamed as
2801 "PerlIO::scalar" and "PerlIO::via" (all lowercase) just before 5.8.0.
2802 The main rationale was to have all core PerlIO layers to have all low‐
2803 ercase names. The "plugins" are named as usual, for example "Per‐
2804 lIO::via::QuotedPrint".
2805 The "threads::shared::queue" and "threads::shared::semaphore" were
2807 renamed as "Thread::Queue" and "Thread::Semaphore" just before 5.8.0.
2808 The main rationale was to have thread modules to obey normal naming,
2809 "Thread::" (the "threads" and "threads::shared" themselves are more
2810 pragma-like, they affect compile-time, so they stay lowercase).
2811
2813 If you find what you think is a bug, you might check the articles
2814 recently posted to the comp.lang.perl.misc newsgroup and the perl bug
2815 database at http://bugs.perl.org/ . There may also be information at
2816 http://www.perl.com/ , the Perl Home Page.
2817 If you believe you have an unreported bug, please run the perlbug pro‐
2819 gram included with your release. Be sure to trim your bug down to a
2820 tiny but sufficient test case. Your bug report, along with the output
2821 of "perl -V", will be sent off to perlbug@perl.org to be analysed by
2822 the Perl porting team.
2823
2825 The Changes file for exhaustive details on what changed.
2826 The INSTALL file for how to build Perl.
2828 The README file for general stuff.
2830 The Artistic and Copying files for copyright information.
2832
2834 Written by Jarkko Hietaniemi <jhi@iki.fi>.
2835perl v5.8.8 2006-01-07 PERL58DELTA(1)