Commit | Line | Data |
cc0fca54 |
1 | =head1 NAME |
2 | |
f39f21d8 |
3 | perldelta - what is new for perl v5.8.0 |
cc0fca54 |
4 | |
5 | =head1 DESCRIPTION |
6 | |
f39f21d8 |
7 | This document describes differences between the 5.6.0 release and the |
8 | 5.8.0 release. |
9 | |
f39f21d8 |
10 | =head1 Incompatible Changes |
11 | |
12 | =over 4 |
13 | |
14 | =item * |
15 | |
f39f21d8 |
16 | The semantics of bless(REF, REF) were unclear and until someone proves |
17 | it to make some sense, it is forbidden. |
18 | |
19 | =item * |
20 | |
21 | A reference to a reference now stringify as "REF(0x81485ec)" instead |
22 | of "SCALAR(0x81485ec)" in order to be more consistent with the return |
23 | value of ref(). |
24 | |
25 | =item * |
26 | |
27 | The very dusty examples in the eg/ directory have been removed. |
28 | Suggestions for new shiny examples welcome but the main issue is that |
29 | the examples need to be documented, tested and (most importantly) |
30 | maintained. |
31 | |
32 | =item * |
33 | |
34 | The obsolete chat2 library that should never have been allowed |
35 | to escape the laboratory has been decommissioned. |
36 | |
37 | =item * |
38 | |
39 | The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still |
40 | recognised but now cause fatal errors. The previous behaviour of |
41 | ignoring them by default and warning if requested was unacceptable |
42 | since it, in a way, falsely promised that the features could be used. |
43 | |
44 | =item * |
45 | |
46 | The (bogus) escape sequences \8 and \9 now give an optional warning |
47 | ("Unrecognized escape passed through"). There is no need to \-escape |
48 | any C<\w> character. |
49 | |
50 | =item * |
51 | |
52 | lstat(FILEHANDLE) now gives a warning because the operation makes no sense. |
53 | In future releases this may become a fatal error. |
54 | |
55 | =item * |
56 | |
57 | The long deprecated uppercase aliases for the string comparison |
58 | operators (EQ, NE, LT, LE, GE, GT) have now been removed. |
59 | |
60 | =item * |
61 | |
62 | The regular expression captured submatches ($1, $2, ...) are now |
63 | more consistently unset if the match fails, instead of leaving false |
64 | data lying around in them. |
65 | |
66 | =item * |
67 | |
68 | The tr///C and tr///U features have been removed and will not return; |
69 | the interface was a mistake. Sorry about that. For similar |
70 | functionality, see pack('U0', ...) and pack('C0', ...). |
71 | |
77c8cf41 |
72 | =item * |
73 | |
74 | Although "you shouldn't do that", it was possible to write code that |
75 | depends on Perl's hashed key order (Data::Dumper does this). The new |
76 | algorithm "One-at-a-Time" produces a different hashed key order. |
77 | More details are in L</"Performance Enhancements">. |
78 | |
79 | =item * |
80 | |
81 | The list of filenames from glob() (or <...>) is now by default sorted |
82 | alphabetically to be csh-compliant. (bsd_glob() does still sort platform |
83 | natively, ASCII or EBCDIC, unless GLOB_ALPHASORT is specified.) |
84 | |
f39f21d8 |
85 | =back |
86 | |
77c8cf41 |
87 | =head2 64-bit platforms and malloc |
88 | |
89 | If your pointers are 64 bits wide, the Perl malloc is no more being |
90 | used because it simply does not work with 8-byte pointers. Also, |
91 | usually the system malloc on such platforms are much better optimized |
92 | for such large memory models than the Perl malloc. |
93 | |
94 | =head2 AIX Dynaloading |
95 | |
96 | The AIX dynaloading now uses in AIX releases 4.3 and newer the native |
97 | dlopen interface of AIX instead of the old emulated interface. This |
98 | change will probably break backward compatibility with compiled |
99 | modules. The change was made to make Perl more compliant with other |
100 | applications like modperl which are using the AIX native interface. |
101 | |
102 | =head2 Socket Extension Dynamic in VMS |
103 | |
104 | The Socket extension is now dynamically loaded instead of being |
105 | statically built in. This may or may not be a problem with ancient |
106 | TCP/IP stacks of VMS: we do not know since we weren't able to test |
107 | Perl in such configurations. |
108 | |
109 | =head2 Different Definition of the Unicode Character Classes \p{In...} |
110 | |
111 | As suggested by the Unicode consortium, the Unicode character classes |
112 | now prefer I<scripts> as opposed to I<blocks> (as defined by Unicode); |
113 | in Perl, when the C<\p{In....}> and the C<\p{In....}> regular expression |
114 | constructs are used. This has changed the definition of some of those |
115 | character classes. |
116 | |
117 | The difference between scripts and blocks is that scripts are the |
118 | glyphs used by a language or a group of languages, while the blocks |
119 | are more artificial groupings of 256 characters based on the Unicode |
120 | numbering. |
121 | |
122 | In general this change results in more inclusive Unicode character |
123 | classes, but changes to the other direction also do take place: |
124 | for example while the script C<Latin> includes all the Latin |
125 | characters and their various diacritic-adorned versions, it |
126 | does not include the various punctuation or digits (since they |
127 | are not solely C<Latin>). |
128 | |
129 | Changes in the character class semantics may have happened if a script |
130 | and a block happen to have the same name, for example C<Hebrew>. |
131 | In such cases the script wins and C<\p{InHebrew}> now means the script |
132 | definition of Hebrew. The block definition in still available, |
133 | though, by appending C<Block> to the name: C<\p{InHebrewBlock}> means |
134 | what C<\p{InHebrew}> meant in perl 5.6.0. For the full list |
135 | of affected character classes, see L<perlunicode/Blocks>. |
136 | |
137 | =head2 Deprecations |
138 | |
139 | The current user-visible implementation of pseudo-hashes (the weird |
140 | use of the first array element) is deprecated starting from Perl 5.8.0 |
141 | and will be removed in Perl 5.10.0, and the feature will be |
142 | implemented differently. Not only is the current interface rather |
143 | ugly, but the current implementation slows down normal array and hash |
144 | use quite noticeably. The C<fields> pragma interface will remain |
145 | available. |
146 | |
147 | The syntaxes C<@a->[...]> and C<@h->{...}> have now been deprecated. |
148 | |
149 | The suidperl is also considered to be too much a risk to continue |
150 | maintaining and the suidperl code is likely to be removed in a future |
151 | release. |
152 | |
153 | The C<package;> syntax (C<package> without an argument has been |
154 | deprecated. Its semantics were never that clear and its |
155 | implementation even less so. If you have used that feature to |
156 | disallow all but fully qualified variables, C<use strict;> instead. |
157 | |
158 | The chdir(undef) and chdir('') behaviors to match chdir() has been |
159 | deprecated. In future versions, chdir(undef) and chdir('') will |
160 | simply fail. |
161 | |
f39f21d8 |
162 | =head1 Core Enhancements |
163 | |
164 | =over 4 |
165 | |
166 | =item * |
167 | |
168 | C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass |
169 | in multiple arguments.) |
170 | |
171 | =item * |
172 | |
173 | my __PACKAGE__ $obj now works. |
174 | |
175 | =item * |
176 | |
177 | C<no Module;> now works even if there is no "sub unimport" in the Module. |
178 | |
179 | =item * |
180 | |
181 | The numerical comparison operators return C<undef> if either operand |
182 | is a NaN. Previously the behaviour was unspecified. |
183 | |
184 | =item * |
185 | |
186 | C<pack('U0a*', ...)> can now be used to force a string to UTF8. |
187 | |
188 | =item * |
189 | |
190 | prototype(\&) is now available. |
191 | |
192 | =item * |
193 | |
194 | There is now an UNTIE method. |
195 | |
196 | =back |
197 | |
77c8cf41 |
198 | =head2 AUTOLOAD Is Now Lvaluable |
f39f21d8 |
199 | |
77c8cf41 |
200 | AUTOLOAD is now lvaluable, meaning that you can add the :lvalue attribute |
201 | to AUTOLOAD subroutines and you can assign to the AUTOLOAD return value. |
202 | |
203 | =head2 PerlIO is Now The Default |
f39f21d8 |
204 | |
205 | =over 4 |
206 | |
207 | =item * |
208 | |
77c8cf41 |
209 | IO is now by default done via PerlIO rather than system's "stdio". |
210 | PerlIO allows "layers" to be "pushed" onto a file handle to alter the |
211 | handle's behaviour. Layers can be specified at open time via 3-arg |
212 | form of open: |
f39f21d8 |
213 | |
77c8cf41 |
214 | open($fh,'>:crlf :utf8', $path) || ... |
f39f21d8 |
215 | |
77c8cf41 |
216 | or on already opened handles via extended C<binmode>: |
f39f21d8 |
217 | |
77c8cf41 |
218 | binmode($fh,':encoding(iso-8859-7)'); |
f39f21d8 |
219 | |
77c8cf41 |
220 | The built-in layers are: unix (low level read/write), stdio (as in |
221 | previous Perls), perlio (re-implementation of stdio buffering in a |
222 | portable manner), crlf (does CRLF <=> "\n" translation as on Win32, |
223 | but available on any platform). A mmap layer may be available if |
224 | platform supports it (mostly UNIXes). |
f39f21d8 |
225 | |
77c8cf41 |
226 | Layers to be applied by default may be specified via the 'open' pragma. |
227 | |
228 | See L</"Installation and Configuration Improvements"> for the effects |
229 | of PerlIO on your architecture name. |
f39f21d8 |
230 | |
231 | =item * |
232 | |
77c8cf41 |
233 | File handles can be marked as accepting Perl's internal encoding of Unicode |
234 | (UTF-8 or UTF-EBCDIC depending on platform) by a pseudo layer ":utf8" : |
f39f21d8 |
235 | |
77c8cf41 |
236 | open($fh,">:utf8","Uni.txt"); |
f39f21d8 |
237 | |
77c8cf41 |
238 | Note for EBCDIC users: the pseudo layer ":utf8" is erroneously named |
239 | for you since it's not UTF-8 what you will be getting but instead |
240 | UTF-EBCDIC. See L<perlunicode>, L<utf8>, and |
241 | http://www.unicode.org/unicode/reports/tr16/ for more information. |
242 | In future releases this naming may change. |
f39f21d8 |
243 | |
244 | =item * |
245 | |
77c8cf41 |
246 | File handles can translate character encodings from/to Perl's internal |
247 | Unicode form on read/write via the ":encoding()" layer. |
f39f21d8 |
248 | |
249 | =item * |
250 | |
77c8cf41 |
251 | File handles can be opened to "in memory" files held in Perl scalars via: |
252 | |
253 | open($fh,'>', \$variable) || ... |
f39f21d8 |
254 | |
255 | =item * |
256 | |
77c8cf41 |
257 | Anonymous temporary files are available without need to |
258 | 'use FileHandle' or other module via |
f39f21d8 |
259 | |
77c8cf41 |
260 | open($fh,"+>", undef) || ... |
f39f21d8 |
261 | |
77c8cf41 |
262 | That is a literal undef, not an undefined value. |
f39f21d8 |
263 | |
264 | =item * |
265 | |
77c8cf41 |
266 | The list form of C<open> is now implemented for pipes (at least on UNIX): |
f39f21d8 |
267 | |
77c8cf41 |
268 | open($fh,"-|", 'cat', '/etc/motd') |
f39f21d8 |
269 | |
77c8cf41 |
270 | creates a pipe, and runs the equivalent of exec('cat', '/etc/motd') in |
271 | the child process. |
f39f21d8 |
272 | |
273 | =item * |
274 | |
77c8cf41 |
275 | The following builtin functions are now overridable: chop(), chomp(), |
276 | each(), keys(), pop(), push(), shift(), splice(), unshift(). |
f39f21d8 |
277 | |
278 | =item * |
279 | |
77c8cf41 |
280 | Formats now support zero-padded decimal fields. |
f39f21d8 |
281 | |
282 | =item * |
283 | |
77c8cf41 |
284 | Perl now tries internally to use integer values in numeric conversions |
285 | and basic arithmetics (+ - * /) if the arguments are integers, and |
286 | tries also to keep the results stored internally as integers. |
287 | This change leads into often slightly faster and always less lossy |
288 | arithmetics. (Previously Perl always preferred floating point numbers |
289 | in its math.) |
f39f21d8 |
290 | |
291 | =item * |
292 | |
77c8cf41 |
293 | The printf() and sprintf() now support parameter reordering using the |
294 | C<%\d+\$> and C<*\d+\$> syntaxes. For example |
f39f21d8 |
295 | |
77c8cf41 |
296 | print "%2\$s %1\$s\n", "foo", "bar"; |
f39f21d8 |
297 | |
77c8cf41 |
298 | will print "bar foo\n"; This feature helps in writing |
299 | internationalised software. |
f39f21d8 |
300 | |
301 | =item * |
302 | |
77c8cf41 |
303 | Unicode in general should be now much more usable. Unicode can be |
304 | used in hash keys, Unicode in regular expressions should work now, |
305 | Unicode in tr/// should work now (though tr/// seems to be a |
306 | particularly tricky to get right, so you have been warned) |
f39f21d8 |
307 | |
308 | =item * |
309 | |
77c8cf41 |
310 | The Unicode Character Database coming with Perl has been upgraded |
311 | to Unicode 3.1. For more information, see http://www.unicode.org/, |
312 | and http://www.unicode.org/unicode/reports/tr27/ |
f39f21d8 |
313 | |
77c8cf41 |
314 | For developers interested in enhancing Perl's Unicode capabilities: |
315 | almost all the UCD files are included with the Perl distribution in |
316 | the lib/unicode subdirectory. The most notable omission, for space |
317 | considerations, is the Unihan database. |
f39f21d8 |
318 | |
319 | =item * |
320 | |
77c8cf41 |
321 | The Unicode character classes \p{Blank} and \p{SpacePerl} have been |
322 | added. "Blank" is like C isblank(), that is, it contains only |
323 | "horizontal whitespace" (the space character is, the newline isn't), |
324 | and the "SpacePerl" is the Unicode equivalent of C<\s> (\p{Space} |
325 | isn't, since that includes the vertical tabulator character, whereas |
326 | C<\s> doesn't.) |
f39f21d8 |
327 | |
328 | =back |
329 | |
77c8cf41 |
330 | =head2 Signals Are Now Safe |
331 | |
332 | Perl used to be fragile in that signals arriving at inopportune moments |
333 | could corrupt Perl's internal state. |
334 | |
335 | =head2 Understanding of Numbers |
336 | |
337 | In general a lot of fixing has happened in the area of Perl's |
338 | understanding of numbers, both integer and floating point. Since in |
339 | many systems the standard number parsing functions like C<strtoul()> |
340 | and C<atof()> seem to have bugs, Perl tries to work around their |
341 | deficiencies. This results hopefully in more accurate numbers. |
f39f21d8 |
342 | |
343 | =over 4 |
344 | |
345 | =item * |
346 | |
77c8cf41 |
347 | The rules for allowing underscores (underbars) in numeric constants |
348 | have been relaxed and simplified: now you can have an underscore |
349 | B<between digits>. |
f39f21d8 |
350 | |
351 | =item * |
352 | |
77c8cf41 |
353 | GMAGIC (right-hand side magic) could in many cases such as string |
354 | concatenation be invoked too many times. |
f39f21d8 |
355 | |
356 | =item * |
357 | |
77c8cf41 |
358 | Lexicals I: lexicals outside an eval "" weren't resolved |
359 | correctly inside a subroutine definition inside the eval "" if they |
360 | were not already referenced in the top level of the eval""ed code. |
f39f21d8 |
361 | |
362 | =item * |
363 | |
77c8cf41 |
364 | Lexicals II: lexicals leaked at file scope into subroutines that |
365 | were declared before the lexicals. |
f39f21d8 |
366 | |
367 | =item * |
368 | |
77c8cf41 |
369 | Lvalue subroutines can now return C<undef> in list context. |
f39f21d8 |
370 | |
371 | =item * |
372 | |
77c8cf41 |
373 | The C<op_clear> and C<op_null> are now exported. |
f39f21d8 |
374 | |
375 | =item * |
376 | |
77c8cf41 |
377 | A new special regular expression variable has been introduced: |
378 | C<$^N>, which contains the most-recently closed group (submatch). |
f39f21d8 |
379 | |
380 | =item * |
381 | |
77c8cf41 |
382 | L<utime> now supports C<utime undef, undef, @files> to change the |
383 | file timestamps to the current time. |
f39f21d8 |
384 | |
385 | =item * |
386 | |
77c8cf41 |
387 | The Perl parser has been stress tested using both random input and |
388 | Markov chain input. |
f39f21d8 |
389 | |
390 | =item * |
391 | |
77c8cf41 |
392 | C<eval "v200"> now works. |
f39f21d8 |
393 | |
394 | =item * |
395 | |
77c8cf41 |
396 | VMS now works under PerlIO. |
f39f21d8 |
397 | |
398 | =item * |
399 | |
77c8cf41 |
400 | END blocks are now run even if you exit/die in a BEGIN block. |
401 | The execution of END blocks is now controlled by |
402 | PL_exit_flags & PERL_EXIT_DESTRUCT_END. This enables the new |
403 | behaviour for perl embedders. This will default in 5.10. See |
404 | L<perlembed>. |
f39f21d8 |
405 | |
406 | =back |
407 | |
77c8cf41 |
408 | =head1 Modules and Pragmata |
f39f21d8 |
409 | |
77c8cf41 |
410 | =head2 New Modules |
f39f21d8 |
411 | |
412 | =over 4 |
413 | |
414 | =item * |
415 | |
77c8cf41 |
416 | File::Temp allows one to create temporary files and directories in an |
417 | easy, portable, and secure way. |
f39f21d8 |
418 | |
419 | =item * |
420 | |
77c8cf41 |
421 | Storable gives persistence to Perl data structures by allowing the |
422 | storage and retrieval of Perl data to and from files in a fast and |
423 | compact binary format. |
f39f21d8 |
424 | |
425 | =item * |
426 | |
427 | B::Concise, by Stephen McCamant, is a new compiler backend for |
428 | walking the Perl syntax tree, printing concise info about ops. |
429 | The output is highly customisable. |
430 | |
431 | See L<B::Concise> for more information. |
432 | |
433 | =item * |
434 | |
435 | Class::ISA, by Sean Burke, for reporting the search path for a |
436 | class's ISA tree, has been added. |
437 | |
438 | See L<Class::ISA> for more information. |
439 | |
440 | =item * |
441 | |
442 | Cwd has now a split personality: if possible, an extension is used, |
443 | (this will hopefully be both faster and more secure and robust) but |
444 | if not possible, the familiar Perl library implementation is used. |
445 | |
446 | =item * |
447 | |
448 | Digest, a frontend module for calculating digests (checksums), |
449 | from Gisle Aas, has been added. |
450 | |
451 | See L<Digest> for more information. |
452 | |
453 | =item * |
454 | |
455 | Digest::MD5 for calculating MD5 digests (checksums), by Gisle Aas, |
456 | has been added. |
457 | |
458 | use Digest::MD5 'md5_hex'; |
459 | |
460 | $digest = md5_hex("Thirsty Camel"); |
461 | |
462 | print $digest, "\n"; # 01d19d9d2045e005c3f1b80e8b164de1 |
463 | |
464 | NOTE: the MD5 backward compatibility module is deliberately not |
465 | included since its use is discouraged. |
466 | |
467 | See L<Digest::MD5> for more information. |
468 | |
469 | =item * |
470 | |
471 | Encode, by Nick Ing-Simmons, provides a mechanism to translate |
472 | between different character encodings. Support for Unicode, |
473 | ISO-8859-*, ASCII, CP*, KOI8-R, and three variants of EBCDIC are |
474 | compiled in to the module. Several other encodings (like Japanese, |
475 | Chinese, and MacIntosh encodings) are included and will be loaded at |
476 | runtime. |
477 | |
478 | Any encoding supported by Encode module is also available to the |
479 | ":encoding()" layer if PerlIO is used. |
480 | |
481 | See L<Encode> for more information. |
482 | |
483 | =item * |
484 | |
485 | Filter::Simple is an easy-to-use frontend to Filter::Util::Call, |
486 | from Damian Conway. |
487 | |
488 | # in MyFilter.pm: |
489 | |
490 | package MyFilter; |
491 | |
492 | use Filter::Simple sub { |
493 | while (my ($from, $to) = splice @_, 0, 2) { |
494 | s/$from/$to/g; |
495 | } |
496 | }; |
497 | |
498 | 1; |
499 | |
500 | # in user's code: |
501 | |
502 | use MyFilter qr/red/ => 'green'; |
503 | |
504 | print "red\n"; # this code is filtered, will print "green\n" |
505 | print "bored\n"; # this code is filtered, will print "bogreen\n" |
506 | |
507 | no MyFilter; |
508 | |
509 | print "red\n"; # this code is not filtered, will print "red\n" |
510 | |
511 | See L<Filter::Simple> for more information. |
512 | |
513 | =item * |
514 | |
515 | Filter::Util::Call, by Paul Marquess, provides you with the |
516 | framework to write I<Source Filters> in Perl. For most uses |
517 | the frontend Filter::Simple is to be preferred. |
518 | See L<Filter::Util::Call> for more information. |
519 | |
520 | =item * |
521 | |
522 | Locale::Constants, Locale::Country, Locale::Currency, and Locale::Language, |
523 | from Neil Bowers, have been added. They provide the codes for various |
524 | locale standards, such as "fr" for France, "usd" for US Dollar, and |
525 | "jp" for Japanese. |
526 | |
527 | use Locale::Country; |
528 | |
529 | $country = code2country('jp'); # $country gets 'Japan' |
530 | $code = country2code('Norway'); # $code gets 'no' |
531 | |
532 | See L<Locale::Constants>, L<Locale::Country>, L<Locale::Currency>, |
533 | and L<Locale::Language> for more information. |
534 | |
535 | =item * |
536 | |
537 | MIME::Base64, by Gisle Aas, allows you to encode data in base64. |
538 | |
539 | use MIME::Base64; |
540 | |
541 | $encoded = encode_base64('Aladdin:open sesame'); |
542 | $decoded = decode_base64($encoded); |
543 | |
544 | print $encoded, "\n"; # "QWxhZGRpbjpvcGVuIHNlc2FtZQ==" |
545 | |
546 | See L<MIME::Base64> for more information. |
547 | |
548 | =item * |
549 | |
550 | MIME::QuotedPrint, by Gisle Aas, allows you to encode data in |
551 | quoted-printable encoding. |
552 | |
553 | use MIME::QuotedPrint; |
554 | |
555 | $encoded = encode_qp("Smiley in Unicode: \x{263a}"); |
556 | $decoded = decode_qp($encoded); |
557 | |
558 | print $encoded, "\n"; # "Smiley in Unicode: =263A" |
559 | |
560 | MIME::QuotedPrint has been enhanced to provide the basic methods |
561 | necessary to use it with PerlIO::Via as in : |
562 | |
563 | use MIME::QuotedPrint; |
564 | open($fh,">Via(MIME::QuotedPrint)",$path) |
565 | |
566 | See L<MIME::QuotedPrint> for more information. |
567 | |
568 | =item * |
569 | |
570 | PerlIO::Scalar, by Nick Ing-Simmons, provides the implementation of |
571 | IO to "in memory" Perl scalars as discussed above. It also serves as |
572 | an example of a loadable layer. Other future possibilities include |
573 | PerlIO::Array and PerlIO::Code. See L<PerlIO::Scalar> for more |
574 | information. |
575 | |
576 | =item * |
577 | |
578 | PerlIO::Via, by Nick Ing-Simmons, acts as a PerlIO layer and wraps |
579 | PerlIO layer functionality provided by a class (typically implemented |
580 | in perl code). |
581 | |
582 | use MIME::QuotedPrint; |
583 | open($fh,">Via(MIME::QuotedPrint)",$path) |
584 | |
585 | This will automatically convert everything output to C<$fh> |
586 | to Quoted-Printable. See L<PerlIO::Via> for more information. |
587 | |
588 | =item * |
589 | |
590 | Pod::Text::Overstrike, by Joe Smith, has been added. |
591 | It converts POD data to formatted overstrike text. |
592 | See L<Pod::Text::Overstrike> for more information. |
593 | |
594 | =item * |
595 | |
596 | Switch from Damian Conway has been added. Just by saying |
597 | |
598 | use Switch; |
599 | |
600 | you have C<switch> and C<case> available in Perl. |
601 | |
602 | use Switch; |
603 | |
604 | switch ($val) { |
605 | |
606 | case 1 { print "number 1" } |
607 | case "a" { print "string a" } |
608 | case [1..10,42] { print "number in list" } |
609 | case (@array) { print "number in list" } |
610 | case /\w+/ { print "pattern" } |
611 | case qr/\w+/ { print "pattern" } |
612 | case (%hash) { print "entry in hash" } |
613 | case (\%hash) { print "entry in hash" } |
614 | case (\&sub) { print "arg to subroutine" } |
615 | else { print "previous case not true" } |
616 | } |
617 | |
77c8cf41 |
618 | See L<Switch> for more information. |
619 | |
620 | =item * |
621 | |
622 | Text::Balanced from Damian Conway has been added, for |
623 | extracting delimited text sequences from strings. |
624 | |
625 | use Text::Balanced 'extract_delimited'; |
626 | |
627 | ($a, $b) = extract_delimited("'never say never', he never said", "'", ''); |
628 | |
629 | $a will be "'never say never'", $b will be ', he never said'. |
630 | |
631 | In addition to extract_delimited() there are also extract_bracketed(), |
632 | extract_quotelike(), extract_codeblock(), extract_variable(), |
633 | extract_tagged(), extract_multiple(), gen_delimited_pat(), and |
634 | gen_extract_tagged(). With these you can implement rather advanced |
635 | parsing algorithms. See L<Text::Balanced> for more information. |
636 | |
637 | =item * |
638 | |
639 | Tie::RefHash::Nestable, by Edward Avis, allows storing hash references |
640 | (unlike the standard Tie::RefHash) The module is contained within |
641 | Tie::RefHash. |
642 | |
643 | =item * |
644 | |
645 | XS::Typemap, by Tim Jenness, is a test extension that exercises XS |
646 | typemaps. Nothing gets installed but for extension writers the code |
647 | is worth studying. |
648 | |
649 | =item * |
650 | |
651 | L<Attribute::Handlers> - Simpler definition of attribute handlers |
652 | |
653 | =item * |
654 | |
655 | L<ExtUtils::Constant> - generate XS code to import C header constants |
656 | |
657 | =item * |
658 | |
659 | L<I18N::Langinfo> - query locale information |
660 | |
661 | =item * |
662 | |
663 | L<I18N::LangTags> - functions for dealing with RFC3066-style language tags |
664 | |
665 | =item * |
666 | |
667 | L<libnet> - a collection of perl5 modules related to network programming |
668 | |
669 | Perl installation leaves libnet unconfigured, use F<libnetcfg> to configure. |
670 | |
671 | =item * |
672 | |
673 | L<List::Util> - selection of general-utility list subroutines |
674 | |
675 | =item * |
676 | |
677 | L<Locale::Maketext> - framework for localization |
678 | |
679 | =item * |
680 | |
681 | L<Memoize> - Make your functions faster by trading space for time |
682 | |
683 | =item * |
684 | |
685 | L<NEXT> - pseudo-class for method redispatch |
686 | |
687 | =item * |
688 | |
689 | L<Scalar::Util> - selection of general-utility scalar subroutines |
690 | |
691 | =item * |
692 | |
693 | L<Test::More> - yet another framework for writing test scripts |
694 | |
695 | =item * |
696 | |
697 | L<Test::Simple> - Basic utilities for writing tests |
698 | |
699 | =item * |
700 | |
701 | L<Time::HiRes> - high resolution ualarm, usleep, and gettimeofday |
702 | |
703 | =item * |
704 | |
705 | L<Time::Piece> - Object Oriented time objects |
706 | |
707 | (Previously known as L<Time::Object>.) |
708 | |
709 | =item * |
710 | |
711 | L<Time::Seconds> - a simple API to convert seconds to other date values |
712 | |
713 | =item * |
714 | |
715 | L<UnicodeCD> - Unicode Character Database |
716 | |
717 | =back |
718 | |
719 | =head2 Updated And Improved Modules and Pragmata |
720 | |
721 | =over 4 |
722 | |
723 | =item * |
724 | |
725 | The following independently supported modules have been updated to |
726 | newer versions from CPAN: CGI, CPAN, DB_File, File::Spec, Getopt::Long, |
727 | the podlators bundle, Pod::LaTeX, Pod::Parser, Term::ANSIColor, Test. |
728 | |
729 | =item * |
730 | |
731 | Bug fixes and minor enhancements have been applied to B::Deparse, |
732 | Data::Dumper, IO::Poll, IO::Socket::INET, Math::BigFloat, |
733 | Math::Complex, Math::Trig, Net::protoent, the re pragma, SelfLoader, |
734 | Sys::SysLog, Test::Harness, Text::Wrap, UNIVERSAL, and the warnings |
735 | pragma. |
736 | |
737 | =item * |
738 | |
739 | The attributes::reftype() now works on tied arguments. |
740 | |
741 | =item * |
742 | |
743 | AutoLoader can now be disabled with C<no AutoLoader;>, |
744 | |
745 | =item * |
746 | |
747 | The English module can now be used without the infamous performance |
748 | hit by saying |
749 | |
750 | use English '-no_performance_hit'; |
751 | |
752 | (Assuming, of course, that one doesn't need the troublesome variables |
753 | C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and |
754 | C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>. |
755 | |
756 | =item * |
757 | |
758 | File::Find now has pre- and post-processing callbacks. It also |
759 | correctly changes directories when chasing symbolic links. Callbacks |
760 | (naughtily) exiting with "next;" instead of "return;" now work. |
761 | |
762 | =item * |
763 | |
764 | File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid |
765 | prototype mismatch with CORE::glob(). |
766 | |
767 | =item * |
768 | |
769 | IPC::Open3 now allows the use of numeric file descriptors. |
770 | |
771 | =item * |
772 | |
773 | use lib now works identically to @INC. Removing directories |
774 | with 'no lib' now works. |
775 | |
776 | =item * |
777 | |
778 | C<%INC> now localised in a Safe compartment so that use/require work. |
779 | |
780 | =item * |
781 | |
782 | The Shell module now has an OO interface. |
783 | |
784 | =item * |
785 | |
786 | B::Deparse should be now more robust. It still far from providing a full |
787 | round trip for any random piece of Perl code, though, and is under active |
788 | development: expect more robustness in 5.7.2. |
789 | |
790 | =item * |
791 | |
792 | Class::Struct can now define the classes in compile time. |
793 | |
794 | =item * |
795 | |
796 | Math::BigFloat has undergone much fixing, and in addition the fmod() |
797 | function now supports modulus operations. |
798 | |
799 | (The fixed Math::BigFloat module is also available in CPAN for those |
800 | who can't upgrade their Perl: http://www.cpan.org/authors/id/J/JP/JPEACOCK/) |
801 | |
802 | =item * |
803 | |
804 | Devel::Peek now has an interface for the Perl memory statistics |
805 | (this works only if you are using perl's malloc, and if you have |
806 | compiled with debugging). |
807 | |
808 | =item * |
809 | |
810 | IO::Socket has now atmark() method, which returns true if the socket |
811 | is positioned at the out-of-band mark. The method is also exportable |
812 | as a sockatmark() function. |
813 | |
814 | =item * |
815 | |
816 | IO::Socket::INET has support for ReusePort option (if your platform |
817 | supports it). The Reuse option now has an alias, ReuseAddr. For clarity |
818 | you may want to prefer ReuseAddr. |
819 | |
820 | =item * |
821 | |
822 | Net::Ping has been enhanced. There is now "external" protocol which |
823 | uses Net::Ping::External module which runs external ping(1) and parses |
824 | the output. An alpha version of Net::Ping::External is available in |
825 | CPAN and in 5.7.2 the Net::Ping::External may be integrated to Perl. |
826 | |
827 | =item * |
828 | |
829 | The C<open> pragma allows layers other than ":raw" and ":crlf" when |
830 | using PerlIO. |
831 | |
832 | =item * |
833 | |
834 | POSIX::sigaction() is now much more flexible and robust. |
835 | You can now install coderef handlers, 'DEFAULT', and 'IGNORE' |
836 | handlers, installing new handlers was not atomic. |
837 | |
838 | =item * |
839 | |
840 | The Test module has been significantly enhanced. Its use is |
841 | greatly recommended for module writers. |
842 | |
843 | =item * |
844 | |
845 | The utf8:: name space (as in the pragma) provides various |
846 | Perl-callable functions to provide low level access to Perl's |
847 | internal Unicode representation. At the moment only length() |
848 | has been implemented. |
849 | |
850 | =back |
851 | |
852 | The following modules have been upgraded from the versions at CPAN: |
853 | CPAN, CGI, DB_File, File::Temp, Getopt::Long, Pod::Man, Pod::Text, |
854 | Storable, Text-Tabs+Wrap. |
855 | |
856 | =item * |
857 | |
858 | L<B::Deparse> module has been significantly enhanced. It now |
859 | can deparse almost all of the standard test suite (so that the |
860 | tests still succeed). There is a make target "test.deparse" |
861 | for trying this out. |
862 | |
863 | =item * |
864 | |
865 | L<Class::Struct> now assigns the array/hash element if the accessor |
866 | is called with an array/hash element as the B<sole> argument. |
867 | |
868 | =item * |
869 | |
870 | L<Cwd> extension is now (even) faster. |
871 | |
872 | =item * |
873 | |
874 | L<DB_File> extension has been updated to version 1.77. |
f39f21d8 |
875 | |
876 | =item * |
877 | |
77c8cf41 |
878 | L<Fcntl>, L<Socket>, and L<Sys::Syslog> have been rewritten to use the |
879 | new-style constant dispatch section (see L<ExtUtils::Constant>). |
f39f21d8 |
880 | |
77c8cf41 |
881 | =item * |
f39f21d8 |
882 | |
77c8cf41 |
883 | L<File::Find> is now (again) reentrant. It also has been made |
884 | more portable. |
f39f21d8 |
885 | |
77c8cf41 |
886 | =item * |
f39f21d8 |
887 | |
77c8cf41 |
888 | L<File::Glob> now supports C<GLOB_LIMIT> constant to limit the |
889 | size of the returned list of filenames. |
f39f21d8 |
890 | |
891 | =item * |
892 | |
77c8cf41 |
893 | L<IO::Socket::INET> now supports C<LocalPort> of zero (usually meaning |
894 | that the operating system will make one up.) |
f39f21d8 |
895 | |
896 | =item * |
897 | |
77c8cf41 |
898 | The L<vars> pragma now supports declaring fully qualified variables. |
899 | (Something that C<our()> does not and will not support.) |
f39f21d8 |
900 | |
901 | =back |
902 | |
77c8cf41 |
903 | =head1 Utility Changes |
f39f21d8 |
904 | |
905 | =over 4 |
906 | |
907 | =item * |
908 | |
77c8cf41 |
909 | The Emacs perl mode (emacs/cperl-mode.el) has been updated to version |
910 | 4.31. |
f39f21d8 |
911 | |
912 | =item * |
913 | |
77c8cf41 |
914 | Perlbug is now much more robust. It also sends the bug report to |
915 | perl.org, not perl.com. |
f39f21d8 |
916 | |
917 | =item * |
918 | |
77c8cf41 |
919 | The perlcc utility has been rewritten and its user interface (that is, |
920 | command line) is much more like that of the UNIX C compiler, cc. |
f39f21d8 |
921 | |
77c8cf41 |
922 | =item * |
923 | |
924 | The xsubpp utility for extension writers now understands POD |
925 | documentation embedded in the *.xs files. |
f39f21d8 |
926 | |
927 | =item * |
928 | |
77c8cf41 |
929 | h2xs now produces template README. |
f39f21d8 |
930 | |
931 | =item * |
932 | |
77c8cf41 |
933 | s2p has been completely rewritten in Perl. (It is in fact a full |
934 | implementation of sed in Perl.) |
f39f21d8 |
935 | |
936 | =item * |
937 | |
77c8cf41 |
938 | xsubpp now supports OUT keyword. |
f39f21d8 |
939 | |
940 | =item * |
941 | |
77c8cf41 |
942 | The F<emacs/e2ctags.pl> is now much faster. |
f39f21d8 |
943 | |
944 | =item * |
945 | |
77c8cf41 |
946 | L<h2ph> now supports C trigraphs. |
f39f21d8 |
947 | |
948 | =item * |
949 | |
77c8cf41 |
950 | L<h2xs> uses the new L<ExtUtils::Constant> module which will affect |
951 | newly created extensions that define constants. Since the new code is |
952 | more correct (if you have two constants where the first one is a |
953 | prefix of the second one, the first constant B<never> gets defined), |
954 | less lossy (it uses integers for integer constant, as opposed to the |
955 | old code that used floating point numbers even for integer constants), |
956 | and slightly faster, you might want to consider regenerating your |
957 | extension code (the new scheme makes regenerating easy). |
958 | L<h2xs> now also supports C trigraphs. |
f39f21d8 |
959 | |
960 | =item * |
961 | |
77c8cf41 |
962 | L<libnetcfg> has been added to configure the libnet. |
f39f21d8 |
963 | |
964 | =item * |
965 | |
77c8cf41 |
966 | The F<Pod::Html> (and thusly L<pod2html>) now allows specifying |
967 | a cache directory. |
f39f21d8 |
968 | |
969 | =back |
970 | |
77c8cf41 |
971 | =head1 New Documentation |
f39f21d8 |
972 | |
973 | =over 4 |
974 | |
975 | =item * |
976 | |
77c8cf41 |
977 | perl56delta details the changes between the 5.005 release and the |
978 | 5.6.0 release. |
f39f21d8 |
979 | |
980 | =item * |
981 | |
77c8cf41 |
982 | perldebtut is a Perl debugging tutorial. |
f39f21d8 |
983 | |
77c8cf41 |
984 | =item * |
f39f21d8 |
985 | |
77c8cf41 |
986 | perlebcdic contains considerations for running Perl on EBCDIC platforms. |
987 | Note that unfortunately EBCDIC platforms that used to supported back in |
988 | Perl 5.005 are still unsupported by Perl 5.7.0; the plan, however, is to |
989 | bring them back to the fold. |
f39f21d8 |
990 | |
77c8cf41 |
991 | =item * |
992 | |
993 | perlnewmod tells about writing and submitting a new module. |
f39f21d8 |
994 | |
995 | =item * |
996 | |
77c8cf41 |
997 | perlposix-bc explains using Perl on the POSIX-BC platform |
998 | (an EBCDIC mainframe platform). |
f39f21d8 |
999 | |
1000 | =item * |
1001 | |
77c8cf41 |
1002 | perlretut is a regular expression tutorial. |
f39f21d8 |
1003 | |
1004 | =item * |
1005 | |
77c8cf41 |
1006 | perlrequick is a regular expressions quick-start guide. |
1007 | Yes, much quicker than perlretut. |
f39f21d8 |
1008 | |
77c8cf41 |
1009 | =item * |
f39f21d8 |
1010 | |
77c8cf41 |
1011 | perlutil explains the command line utilities packaged with the Perl |
1012 | distribution. |
1013 | |
1014 | =back |
f39f21d8 |
1015 | |
1016 | =head2 perlclib |
1017 | |
1018 | Internal replacements for standard C library functions. |
1019 | (Interesting only for extension writers and Perl core hackers.) |
1020 | |
1021 | =head2 perliol |
1022 | |
1023 | Internals of PerlIO with layers. |
1024 | |
1025 | =head2 README.aix |
1026 | |
1027 | Documentation on compiling Perl on AIX has been added. AIX has |
1028 | several different C compilers and getting the right patch level |
1029 | is essential. On install README.aix will be installed as L<perlaix>. |
1030 | |
1031 | =head2 README.bs2000 |
1032 | |
1033 | Documentation on compiling Perl on the POSIX-BC platform (an EBCDIC |
1034 | mainframe environment) has been added. |
1035 | |
1036 | This was formerly known as README.posix-bc but the name was considered |
1037 | to be too confusing (it has nothing to do with the POSIX module or the |
1038 | POSIX standard). On install README.bs2000 will be installed as L<perlbs2000>. |
1039 | |
1040 | =head2 README.macos |
1041 | |
1042 | In perl 5.7.1 (and in the 5.6.1) the MacPerl sources have been |
1043 | synchronised with the standard Perl sources. To compile MacPerl |
1044 | some additional steps are required, and this file documents those |
1045 | steps. On install README.macos will be installed as L<perlmacos>. |
1046 | |
1047 | =head2 README.mpeix |
1048 | |
1049 | The README.mpeix has been podified, which means that this information |
1050 | about compiling and using Perl on the MPE/iX miniframe platform will |
1051 | be installed as L<perlmpeix>. |
1052 | |
1053 | =head2 README.solaris |
1054 | |
1055 | README.solaris has been created and Solaris wisdom from elsewhere |
1056 | in the Perl documentation has been collected there. On install |
1057 | README.solaris will be installed as L<perlsolaris>. |
1058 | |
1059 | =head2 README.vos |
1060 | |
1061 | The README.vos has been podified, which means that this information |
1062 | about compiling and using Perl on the Stratus VOS miniframe platform |
1063 | will be installed as L<perlvos>. |
1064 | |
77c8cf41 |
1065 | =head2 Porting/repository.pod |
1066 | |
1067 | Documentation on how to use the Perl source repository has been added. |
1068 | |
1069 | =over 4 |
1070 | |
1071 | =item * |
1072 | |
1073 | L<Locale::Maketext::TPJ13> is an article about software localization, |
1074 | originally published in The Perl Journal #13, republished here with |
1075 | kind permission. |
1076 | |
1077 | =item * |
1078 | |
1079 | More README.$PLATFORM files have been converted into pod, which also |
1080 | means that they also be installed as perl$PLATFORM documentation |
1081 | files. The new files are L<perlapollo>, L<perlbeos>, L<perldgux>, |
1082 | L<perlhurd>, L<perlmint>, L<perlnetware>, L<perlplan9>, L<perlqnx>, |
1083 | and L<perltru64>. |
1084 | |
1085 | =item * |
1086 | |
1087 | The F<Todo> and F<Todo-5.6> files have been merged into L<perltodo>. |
1088 | |
1089 | =item * |
1090 | |
1091 | Use of the F<gprof> tool to profile Perl has been documented in |
1092 | L<perlhack>. There is a make target "perl.gprof" for generating a |
1093 | gprofiled Perl executable. |
1094 | |
1095 | =back |
1096 | |
1097 | =head1 Performance Enhancements |
1098 | |
1099 | =over 4 |
1100 | |
1101 | =item * |
1102 | |
1103 | map() that changes the size of the list should now work faster. |
1104 | |
1105 | =item * |
1106 | |
1107 | sort() has been changed to use mergesort internally as opposed to the |
1108 | earlier quicksort. For very small lists this may result in slightly |
1109 | slower sorting times, but in general the speedup should be at least |
1110 | 20%. Additional bonuses are that the worst case behaviour of sort() |
1111 | is now better (in computer science terms it now runs in time O(N log N), |
1112 | as opposed to quicksort's Theta(N**2) worst-case run time behaviour), |
1113 | and that sort() is now stable (meaning that elements with identical |
1114 | keys will stay ordered as they were before the sort). |
1115 | |
1116 | =item * |
1117 | |
1118 | Hashes now use Bob Jenkins "One-at-a-Time" hashing key algorithm |
1119 | (http://burtleburtle.net/bob/hash/doobs.html). This algorithm is |
1120 | reasonably fast while producing a much better spread of values than |
1121 | the old hashing algorithm (originally by Chris Torek, later tweaked by |
1122 | Ilya Zakharevich). Hash values output from the algorithm on a hash of |
1123 | all 3-char printable ASCII keys comes much closer to passing the |
1124 | DIEHARD random number generation tests. According to perlbench, this |
1125 | change has not affected the overall speed of Perl. |
1126 | |
1127 | =item * |
1128 | |
1129 | unshift() should now be noticeably faster. |
1130 | |
1131 | =back |
1132 | |
1133 | =head1 Installation and Configuration Improvements |
1134 | |
1135 | =head2 Generic Improvements |
1136 | |
1137 | =over 4 |
1138 | |
1139 | =item * |
1140 | |
1141 | INSTALL now explains how you can configure Perl to use 64-bit |
1142 | integers even on non-64-bit platforms. |
1143 | |
1144 | =item * |
1145 | |
1146 | Policy.sh policy change: if you are reusing a Policy.sh file |
1147 | (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old |
1148 | Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of |
1149 | them will now be changed to the new prefix, /foo/bar. (Previously |
1150 | only $prefix changed.) If you do not like this new behaviour, |
1151 | specify prefix, siteprefix, and vendorprefix explicitly. |
1152 | |
1153 | =item * |
1154 | |
1155 | A new optional location for Perl libraries, otherlibdirs, is available. |
1156 | It can be used for example for vendor add-ons without disturbing Perl's |
1157 | own library directories. |
1158 | |
1159 | =item * |
1160 | |
1161 | In many platforms the vendor-supplied 'cc' is too stripped-down to |
1162 | build Perl (basically, 'cc' doesn't do ANSI C). If this seems |
1163 | to be the case and 'cc' does not seem to be the GNU C compiler |
1164 | 'gcc', an automatic attempt is made to find and use 'gcc' instead. |
1165 | |
1166 | =item * |
1167 | |
1168 | gcc needs to closely track the operating system release to avoid |
1169 | build problems. If Configure finds that gcc was built for a different |
1170 | operating system release than is running, it now gives a clearly visible |
1171 | warning that there may be trouble ahead. |
1172 | |
1173 | =item * |
1174 | |
1175 | If binary compatibility with the 5.005 release is not wanted, Configure |
1176 | no longer suggests including the 5.005 modules in @INC. |
1177 | |
1178 | =item * |
1179 | |
1180 | Configure C<-S> can now run non-interactively. |
1181 | |
1182 | =item * |
1183 | |
1184 | configure.gnu now works with options with whitespace in them. |
f39f21d8 |
1185 | |
77c8cf41 |
1186 | =item * |
f39f21d8 |
1187 | |
77c8cf41 |
1188 | installperl now outputs everything to STDERR. |
f39f21d8 |
1189 | |
77c8cf41 |
1190 | =item * |
1191 | |
1192 | $Config{byteorder} is now computed dynamically (this is more robust |
1193 | with "fat binaries" where an executable image contains binaries for |
1194 | more than one binary platform.) |
f39f21d8 |
1195 | |
1196 | =item * |
1197 | |
1198 | Because PerlIO is now the default on most platforms, "-perlio" doesn't |
1199 | get appended to the $Config{archname} (also known as $^O) anymore. |
1200 | Instead, if you explicitly choose not to use perlio (Configure command |
1201 | line option -Uuseperlio), you will get "-stdio" appended. |
1202 | |
1203 | =item * |
1204 | |
1205 | Another change related to the architecture name is that "-64all" |
1206 | (-Duse64bitall, or "maximally 64-bit") is appended only if your |
1207 | pointers are 64 bits wide. (To be exact, the use64bitall is ignored.) |
1208 | |
1209 | =item * |
1210 | |
1211 | APPLLIB_EXP, a less-know configuration-time definition, has been |
1212 | documented. It can be used to prepend site-specific directories |
1213 | to Perl's default search path (@INC), see INSTALL for information. |
1214 | |
1215 | =item * |
1216 | |
1217 | Building Berkeley DB3 for compatibility modes for DB, NDBM, and ODBM |
1218 | has been documented in INSTALL. |
1219 | |
1220 | =item * |
1221 | |
1222 | If you are on IRIX or Tru64 platforms, new profiling/debugging options |
1223 | have been added, see L<perlhack> for more information about pixie and |
1224 | Third Degree. |
1225 | |
77c8cf41 |
1226 | =item * |
1227 | |
1228 | In AFS installations one can configure the root of the AFS to be |
1229 | somewhere else than the default F</afs> by using the Configure |
1230 | parameter C<-Dafsroot=/some/where/else>. |
1231 | |
1232 | =item * |
1233 | |
1234 | The version of Berkeley DB used when the Perl (and, presumably, the |
1235 | DB_File extension) was built is now available as |
1236 | C<@Config{qw(db_version_major db_version_minor db_version_patch)}> |
1237 | from Perl and as C<DB_VERSION_MAJOR_CFG DB_VERSION_MINOR_CFG |
1238 | DB_VERSION_PATCH_CFG> from C. |
1239 | |
1240 | =item * |
1241 | |
1242 | The Thread extension is now not built at all under ithreads |
1243 | (C<Configure -Duseithreads>) because it wouldn't work anyway (the |
1244 | Thread extension requires being Configured with C<-Duse5005threads>). |
1245 | |
1246 | =item * |
1247 | |
1248 | The C<B::Deparse> compiler backend has been so significantly improved |
1249 | that almost the whole Perl test suite passes after being deparsed. A |
1250 | make target has been added to help in further testing: C<make test.deparse>. |
1251 | |
f39f21d8 |
1252 | =back |
1253 | |
1254 | =head2 New Or Improved Platforms |
1255 | |
1256 | For the list of platforms known to support Perl, |
1257 | see L<perlport/"Supported Platforms">. |
1258 | |
1259 | =over 4 |
1260 | |
1261 | =item * |
1262 | |
1263 | AIX dynamic loading should be now better supported. |
1264 | |
1265 | =item * |
1266 | |
1267 | After a long pause, AmigaOS has been verified to be happy with Perl. |
1268 | |
1269 | =item * |
1270 | |
1271 | EBCDIC platforms (z/OS, also known as OS/390, POSIX-BC, and VM/ESA) |
1272 | have been regained. Many test suite tests still fail and the |
1273 | co-existence of Unicode and EBCDIC isn't quite settled, but the |
1274 | situation is much better than with Perl 5.6. See L<perlos390>, |
1275 | L<perlbs2000> (for POSIX-BC), and L<perlvmesa> for more information. |
1276 | |
1277 | =item * |
1278 | |
1279 | Building perl with -Duseithreads or -Duse5005threads now works under |
1280 | HP-UX 10.20 (previously it only worked under 10.30 or later). You will |
1281 | need a thread library package installed. See README.hpux. |
1282 | |
1283 | =item * |
1284 | |
1285 | MacOS Classic (MacPerl has of course been available since |
1286 | perl 5.004 but now the source code bases of standard Perl |
1287 | and MacPerl have been synchronised) |
1288 | |
1289 | =item * |
1290 | |
1291 | NCR MP-RAS is now supported. |
1292 | |
1293 | =item * |
1294 | |
1295 | NonStop-UX is now supported. |
1296 | |
1297 | =item * |
1298 | |
1299 | Amdahl UTS is now supported. |
1300 | |
1301 | =item * |
1302 | |
1303 | z/OS (formerly known as OS/390, formerly known as MVS OE) has now |
1304 | support for dynamic loading. This is not selected by default, |
1305 | however, you must specify -Dusedl in the arguments of Configure. |
1306 | |
f39f21d8 |
1307 | =item * |
1308 | |
77c8cf41 |
1309 | AIX should now work better with gcc, threads, and 64-bitness. Also the |
1310 | long doubles support in AIX should be better now. See L<perlaix>. |
f39f21d8 |
1311 | |
1312 | =item * |
1313 | |
77c8cf41 |
1314 | AtheOS (http://www.atheos.cx/) is a new platform. |
f39f21d8 |
1315 | |
77c8cf41 |
1316 | =item * |
f39f21d8 |
1317 | |
77c8cf41 |
1318 | DG/UX platform now supports the 5.005-style threads. See L<perldgux>. |
f39f21d8 |
1319 | |
1320 | =item * |
1321 | |
77c8cf41 |
1322 | DYNIX/ptx platform (a.k.a. dynixptx) is supported at or near osvers 4.5.2. |
f39f21d8 |
1323 | |
1324 | =item * |
1325 | |
77c8cf41 |
1326 | Several MacOS (Classic) portability patches have been applied. We |
1327 | hope to get a fully working port by 5.8.0. (The remaining problems |
1328 | relate to the changed IO model of Perl.) See L<perlmacos>. |
f39f21d8 |
1329 | |
1330 | =item * |
1331 | |
77c8cf41 |
1332 | MacOS X (or Darwin) should now be able to build Perl even on HFS+ |
1333 | filesystems. (The case-insensitivity confused the Perl build process.) |
f39f21d8 |
1334 | |
77c8cf41 |
1335 | =item * |
f39f21d8 |
1336 | |
77c8cf41 |
1337 | NetWare from Novell is now supported. See L<perlnetware>. |
f39f21d8 |
1338 | |
77c8cf41 |
1339 | =item * |
f39f21d8 |
1340 | |
77c8cf41 |
1341 | The Amdahl UTS UNIX mainframe platform is now supported. |
f39f21d8 |
1342 | |
1343 | =back |
1344 | |
1345 | =head1 Selected Bug Fixes |
1346 | |
f39f21d8 |
1347 | =over 4 |
1348 | |
1349 | =item * |
1350 | |
77c8cf41 |
1351 | Several debugger fixes: exit code now reflects the script exit code, |
1352 | condition C<"0"> now treated correctly, the C<d> command now checks |
1353 | line number, the C<$.> no longer gets corrupted, all debugger output now |
1354 | goes correctly to the socket if RemotePort is set. |
f39f21d8 |
1355 | |
1356 | =item * |
1357 | |
77c8cf41 |
1358 | C<*foo{FORMAT}> now works. |
f39f21d8 |
1359 | |
1360 | =item * |
1361 | |
77c8cf41 |
1362 | Lexical warnings now propagating correctly between scopes. |
f39f21d8 |
1363 | |
1364 | =item * |
1365 | |
77c8cf41 |
1366 | Line renumbering with eval and C<#line> now works. |
f39f21d8 |
1367 | |
1368 | =item * |
1369 | |
77c8cf41 |
1370 | Fixed numerous memory leaks, especially in eval "". |
f39f21d8 |
1371 | |
1372 | =item * |
1373 | |
77c8cf41 |
1374 | Modulus of unsigned numbers now works (4063328477 % 65535 used to |
1375 | return 27406, instead of 27047). |
f39f21d8 |
1376 | |
1377 | =item * |
1378 | |
77c8cf41 |
1379 | Some "not a number" warnings introduced in 5.6.0 eliminated to be |
1380 | more compatible with 5.005. Infinity is now recognised as a number. |
f39f21d8 |
1381 | |
1382 | =item * |
1383 | |
77c8cf41 |
1384 | our() variables will not cause "will not stay shared" warnings. |
f39f21d8 |
1385 | |
1386 | =item * |
1387 | |
77c8cf41 |
1388 | pack "Z" now correctly terminates the string with "\0". |
f39f21d8 |
1389 | |
1390 | =item * |
1391 | |
77c8cf41 |
1392 | Fix password routines which in some shadow password platforms |
1393 | (e.g. HP-UX) caused getpwent() to return every other entry. |
f39f21d8 |
1394 | |
1395 | =item * |
1396 | |
77c8cf41 |
1397 | printf() no longer resets the numeric locale to "C". |
f39f21d8 |
1398 | |
1399 | =item * |
1400 | |
77c8cf41 |
1401 | C<q(a\\b)> now parses correctly as C<'a\\b'>. |
f39f21d8 |
1402 | |
1403 | =item * |
1404 | |
77c8cf41 |
1405 | Printing quads (64-bit integers) with printf/sprintf now works |
1406 | without the q L ll prefixes (assuming you are on a quad-capable platform). |
f39f21d8 |
1407 | |
1408 | =item * |
1409 | |
77c8cf41 |
1410 | Regular expressions on references and overloaded scalars now work. |
f39f21d8 |
1411 | |
1412 | =item * |
1413 | |
77c8cf41 |
1414 | scalar() now forces scalar context even when used in void context. |
f39f21d8 |
1415 | |
1416 | =item * |
1417 | |
77c8cf41 |
1418 | sort() arguments are now compiled in the right wantarray context |
1419 | (they were accidentally using the context of the sort() itself). |
f39f21d8 |
1420 | |
1421 | =item * |
1422 | |
77c8cf41 |
1423 | Changed the POSIX character class C<[[:space:]]> to include the (very |
1424 | rare) vertical tab character. Added a new POSIX-ish character class |
1425 | C<[[:blank:]]> which stands for horizontal whitespace (currently, |
1426 | the space and the tab). |
f39f21d8 |
1427 | |
1428 | =item * |
1429 | |
77c8cf41 |
1430 | $AUTOLOAD, sort(), lock(), and spawning subprocesses |
1431 | in multiple threads simultaneously are now thread-safe. |
f39f21d8 |
1432 | |
1433 | =item * |
1434 | |
77c8cf41 |
1435 | Allow read-only string on left hand side of non-modifying tr///. |
f39f21d8 |
1436 | |
77c8cf41 |
1437 | =item * |
f39f21d8 |
1438 | |
77c8cf41 |
1439 | Several Unicode fixes (but still not perfect). |
f39f21d8 |
1440 | |
77c8cf41 |
1441 | =over 8 |
f39f21d8 |
1442 | |
1443 | =item * |
1444 | |
77c8cf41 |
1445 | BOMs (byte order marks) in the beginning of Perl files |
1446 | (scripts, modules) should now be transparently skipped. |
1447 | UTF-16 (UCS-2) encoded Perl files should now be read correctly. |
f39f21d8 |
1448 | |
1449 | =item * |
1450 | |
77c8cf41 |
1451 | The character tables have been updated to Unicode 3.0.1. |
f39f21d8 |
1452 | |
1453 | =item * |
1454 | |
77c8cf41 |
1455 | chr() for values greater than 127 now create utf8 when under use |
1456 | utf8. |
f39f21d8 |
1457 | |
77c8cf41 |
1458 | =item * |
f39f21d8 |
1459 | |
77c8cf41 |
1460 | Comparing with utf8 data does not magically upgrade non-utf8 data into |
1461 | utf8. |
f39f21d8 |
1462 | |
77c8cf41 |
1463 | =item * |
f39f21d8 |
1464 | |
77c8cf41 |
1465 | C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase. |
f39f21d8 |
1466 | |
77c8cf41 |
1467 | =item * |
f39f21d8 |
1468 | |
77c8cf41 |
1469 | Concatenation with the C<.> operator or via variable interpolation, |
1470 | C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator, |
1471 | substitution with C<s///>, single-quoted UTF8, should now work--in |
1472 | theory. |
f39f21d8 |
1473 | |
77c8cf41 |
1474 | =item * |
f39f21d8 |
1475 | |
77c8cf41 |
1476 | The C<tr///> operator now works I<slightly> better but is still rather |
1477 | broken. Note that the C<tr///CU> functionality has been removed (but |
1478 | see pack('U0', ...)). |
f39f21d8 |
1479 | |
77c8cf41 |
1480 | =item * |
f39f21d8 |
1481 | |
77c8cf41 |
1482 | vec() now refuses to deal with characters >255. |
f39f21d8 |
1483 | |
77c8cf41 |
1484 | =item * |
f39f21d8 |
1485 | |
77c8cf41 |
1486 | Zero entries were missing from the Unicode classes like C<IsDigit>. |
f39f21d8 |
1487 | |
77c8cf41 |
1488 | =back |
f39f21d8 |
1489 | |
77c8cf41 |
1490 | =item * |
f39f21d8 |
1491 | |
77c8cf41 |
1492 | UNIVERSAL::isa no longer caches methods incorrectly. (This broke |
1493 | the Tk extension with 5.6.0.) |
f39f21d8 |
1494 | |
77c8cf41 |
1495 | =item * |
f39f21d8 |
1496 | |
77c8cf41 |
1497 | Configure no longer includes the DBM libraries (dbm, gdbm, db, ndbm) |
1498 | when building the Perl binary. The only exception to this is SunOS 4.x, |
1499 | which needs them. |
f39f21d8 |
1500 | |
77c8cf41 |
1501 | =item * |
f39f21d8 |
1502 | |
77c8cf41 |
1503 | Some new Configure symbols, useful for extension writers: |
f39f21d8 |
1504 | |
77c8cf41 |
1505 | =over 8 |
f39f21d8 |
1506 | |
77c8cf41 |
1507 | =item d_cmsghdr |
f39f21d8 |
1508 | |
77c8cf41 |
1509 | For struct cmsghdr. |
f39f21d8 |
1510 | |
77c8cf41 |
1511 | =item d_fcntl_can_lock |
f39f21d8 |
1512 | |
77c8cf41 |
1513 | Whether fcntl() can be used for file locking. |
f39f21d8 |
1514 | |
77c8cf41 |
1515 | =item d_fsync |
f39f21d8 |
1516 | |
77c8cf41 |
1517 | =item d_getitimer |
f39f21d8 |
1518 | |
77c8cf41 |
1519 | =item d_getpagsz |
f39f21d8 |
1520 | |
77c8cf41 |
1521 | For getpagesize(), though you should prefer POSIX::sysconf(_SC_PAGE_SIZE)) |
f39f21d8 |
1522 | |
77c8cf41 |
1523 | =item d_msghdr_s |
f39f21d8 |
1524 | |
77c8cf41 |
1525 | For struct msghdr. |
f39f21d8 |
1526 | |
77c8cf41 |
1527 | =item need_va_copy |
f39f21d8 |
1528 | |
77c8cf41 |
1529 | Whether one needs to use Perl_va_copy() to copy varargs. |
f39f21d8 |
1530 | |
77c8cf41 |
1531 | =item d_readv |
f39f21d8 |
1532 | |
77c8cf41 |
1533 | =item d_recvmsg |
f39f21d8 |
1534 | |
77c8cf41 |
1535 | =item d_sendmsg |
f39f21d8 |
1536 | |
77c8cf41 |
1537 | =item sig_size |
f39f21d8 |
1538 | |
77c8cf41 |
1539 | The number of elements in an array needed to hold all the available signals. |
f39f21d8 |
1540 | |
77c8cf41 |
1541 | =item d_sockatmark |
f39f21d8 |
1542 | |
77c8cf41 |
1543 | =item d_strtoq |
f39f21d8 |
1544 | |
77c8cf41 |
1545 | =item d_u32align |
f39f21d8 |
1546 | |
77c8cf41 |
1547 | Whether one needs to access character data aligned by U32 sized pointers. |
f39f21d8 |
1548 | |
77c8cf41 |
1549 | =item d_ualarm |
f39f21d8 |
1550 | |
77c8cf41 |
1551 | =item d_usleep |
f39f21d8 |
1552 | |
77c8cf41 |
1553 | =back |
f39f21d8 |
1554 | |
77c8cf41 |
1555 | =item * |
f39f21d8 |
1556 | |
77c8cf41 |
1557 | Removed Configure symbols: the PDP-11 memory model settings: huge, |
1558 | large, medium, models. |
f39f21d8 |
1559 | |
77c8cf41 |
1560 | =item * |
f39f21d8 |
1561 | |
77c8cf41 |
1562 | SOCKS support is now much more robust. |
f39f21d8 |
1563 | |
77c8cf41 |
1564 | =item * |
f39f21d8 |
1565 | |
77c8cf41 |
1566 | If your file system supports symbolic links you can build Perl outside |
1567 | of the source directory by |
f39f21d8 |
1568 | |
77c8cf41 |
1569 | mkdir /tmp/perl/build/directory |
1570 | cd /tmp/perl/build/directory |
1571 | sh /path/to/perl/source/Configure -Dmksymlinks ... |
f39f21d8 |
1572 | |
77c8cf41 |
1573 | This will create in /tmp/perl/build/directory a tree of symbolic links |
1574 | pointing to files in /path/to/perl/source. The original files are left |
1575 | unaffected. After Configure has finished you can just say |
f39f21d8 |
1576 | |
77c8cf41 |
1577 | make all test |
f39f21d8 |
1578 | |
77c8cf41 |
1579 | and Perl will be built and tested, all in /tmp/perl/build/directory. |
f39f21d8 |
1580 | |
77c8cf41 |
1581 | =back |
f39f21d8 |
1582 | |
77c8cf41 |
1583 | =head2 Platform Specific Changes and Fixes |
f39f21d8 |
1584 | |
1585 | =over 4 |
1586 | |
1587 | =item * |
1588 | |
77c8cf41 |
1589 | BSDI 4.* |
f39f21d8 |
1590 | |
77c8cf41 |
1591 | Perl now works on post-4.0 BSD/OSes. |
f39f21d8 |
1592 | |
1593 | =item * |
1594 | |
77c8cf41 |
1595 | All BSDs |
f39f21d8 |
1596 | |
77c8cf41 |
1597 | Setting C<$0> now works (as much as possible; see perlvar for details). |
f39f21d8 |
1598 | |
1599 | =item * |
1600 | |
77c8cf41 |
1601 | Cygwin |
f39f21d8 |
1602 | |
77c8cf41 |
1603 | Numerous updates; currently synchronised with Cygwin 1.1.4. |
f39f21d8 |
1604 | |
1605 | =item * |
1606 | |
77c8cf41 |
1607 | EPOC |
f39f21d8 |
1608 | |
77c8cf41 |
1609 | EPOC update after Perl 5.6.0. See README.epoc. |
f39f21d8 |
1610 | |
1611 | =item * |
1612 | |
77c8cf41 |
1613 | FreeBSD 3.* |
f39f21d8 |
1614 | |
77c8cf41 |
1615 | Perl now works on post-3.0 FreeBSDs. |
f39f21d8 |
1616 | |
1617 | =item * |
1618 | |
77c8cf41 |
1619 | HP-UX |
1620 | |
1621 | README.hpux updated; C<Configure -Duse64bitall> now almost works. |
f39f21d8 |
1622 | |
1623 | =item * |
1624 | |
77c8cf41 |
1625 | IRIX |
f39f21d8 |
1626 | |
77c8cf41 |
1627 | Numerous compilation flag and hint enhancements; accidental mixing |
1628 | of 32-bit and 64-bit libraries (a doomed attempt) made much harder. |
f39f21d8 |
1629 | |
77c8cf41 |
1630 | =item * |
f39f21d8 |
1631 | |
77c8cf41 |
1632 | Linux |
f39f21d8 |
1633 | |
77c8cf41 |
1634 | Long doubles should now work (see INSTALL). |
f39f21d8 |
1635 | |
1636 | =item * |
1637 | |
77c8cf41 |
1638 | MacOS Classic |
f39f21d8 |
1639 | |
77c8cf41 |
1640 | Compilation of the standard Perl distribution in MacOS Classic should |
1641 | now work if you have the Metrowerks development environment and |
1642 | the missing Mac-specific toolkit bits. Contact the macperl mailing |
1643 | list for details. |
f39f21d8 |
1644 | |
1645 | =item * |
1646 | |
77c8cf41 |
1647 | MPE/iX |
f39f21d8 |
1648 | |
77c8cf41 |
1649 | MPE/iX update after Perl 5.6.0. See README.mpeix. |
f39f21d8 |
1650 | |
1651 | =item * |
1652 | |
77c8cf41 |
1653 | NetBSD/sparc |
f39f21d8 |
1654 | |
77c8cf41 |
1655 | Perl now works on NetBSD/sparc. |
f39f21d8 |
1656 | |
1657 | =item * |
1658 | |
77c8cf41 |
1659 | OS/2 |
f39f21d8 |
1660 | |
77c8cf41 |
1661 | Now works with usethreads (see INSTALL). |
f39f21d8 |
1662 | |
1663 | =item * |
1664 | |
77c8cf41 |
1665 | Solaris |
f39f21d8 |
1666 | |
77c8cf41 |
1667 | 64-bitness using the Sun Workshop compiler now works. |
f39f21d8 |
1668 | |
1669 | =item * |
1670 | |
77c8cf41 |
1671 | Tru64 (aka Digital UNIX, aka DEC OSF/1) |
f39f21d8 |
1672 | |
77c8cf41 |
1673 | The operating system version letter now recorded in $Config{osvers}. |
1674 | Allow compiling with gcc (previously explicitly forbidden). Compiling |
1675 | with gcc still not recommended because buggy code results, even with |
1676 | gcc 2.95.2. |
f39f21d8 |
1677 | |
1678 | =item * |
1679 | |
77c8cf41 |
1680 | Unicos |
1681 | |
1682 | Fixed various alignment problems that lead into core dumps either |
1683 | during build or later; no longer dies on math errors at runtime; |
1684 | now using full quad integers (64 bits), previously was using |
1685 | only 46 bit integers for speed. |
f39f21d8 |
1686 | |
1687 | =item * |
1688 | |
77c8cf41 |
1689 | VMS |
1690 | |
1691 | chdir() now works better despite a CRT bug; now works with MULTIPLICITY |
1692 | (see INSTALL); now works with Perl's malloc. |
f39f21d8 |
1693 | |
1694 | =item * |
1695 | |
77c8cf41 |
1696 | Windows |
f39f21d8 |
1697 | |
77c8cf41 |
1698 | =over 8 |
f39f21d8 |
1699 | |
1700 | =item * |
1701 | |
77c8cf41 |
1702 | accept() no longer leaks memory. |
f39f21d8 |
1703 | |
1704 | =item * |
1705 | |
77c8cf41 |
1706 | Better chdir() return value for a non-existent directory. |
f39f21d8 |
1707 | |
77c8cf41 |
1708 | =item * |
f39f21d8 |
1709 | |
77c8cf41 |
1710 | New %ENV entries now propagate to subprocesses. |
f39f21d8 |
1711 | |
1712 | =item * |
1713 | |
77c8cf41 |
1714 | $ENV{LIB} now used to search for libs under Visual C. |
1715 | |
1716 | =item * |
1717 | |
1718 | A failed (pseudo)fork now returns undef and sets errno to EAGAIN. |
f39f21d8 |
1719 | |
1720 | =item * |
1721 | |
77c8cf41 |
1722 | Allow REG_EXPAND_SZ keys in the registry. |
f39f21d8 |
1723 | |
1724 | =item * |
1725 | |
77c8cf41 |
1726 | Can now send() from all threads, not just the first one. |
f39f21d8 |
1727 | |
1728 | =item * |
1729 | |
77c8cf41 |
1730 | Fake signal handling reenabled, bugs and all. |
f39f21d8 |
1731 | |
1732 | =item * |
1733 | |
77c8cf41 |
1734 | Less stack reserved per thread so that more threads can run |
1735 | concurrently. (Still 16M per thread.) |
f39f21d8 |
1736 | |
1737 | =item * |
1738 | |
77c8cf41 |
1739 | C<File::Spec->tmpdir()> now prefers C:/temp over /tmp |
1740 | (works better when perl is running as service). |
f39f21d8 |
1741 | |
1742 | =item * |
1743 | |
77c8cf41 |
1744 | Better UNC path handling under ithreads. |
f39f21d8 |
1745 | |
1746 | =item * |
1747 | |
77c8cf41 |
1748 | wait() and waitpid() now work much better. |
f39f21d8 |
1749 | |
1750 | =item * |
1751 | |
77c8cf41 |
1752 | winsock handle leak fixed. |
f39f21d8 |
1753 | |
1754 | =back |
1755 | |
77c8cf41 |
1756 | =back |
f39f21d8 |
1757 | |
77c8cf41 |
1758 | =head1 New or Changed Diagnostics |
f39f21d8 |
1759 | |
77c8cf41 |
1760 | All regular expression compilation error messages are now hopefully |
1761 | easier to understand both because the error message now comes before |
1762 | the failed regex and because the point of failure is now clearly |
1763 | marked. |
f39f21d8 |
1764 | |
77c8cf41 |
1765 | The various "opened only for", "on closed", "never opened" warnings |
1766 | drop the C<main::> prefix for filehandles in the C<main> package, |
1767 | for example C<STDIN> instead of <main::STDIN>. |
f39f21d8 |
1768 | |
77c8cf41 |
1769 | The "Unrecognized escape" warning has been extended to include C<\8>, |
1770 | C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters. |
f39f21d8 |
1771 | |
77c8cf41 |
1772 | =over 4 |
f39f21d8 |
1773 | |
77c8cf41 |
1774 | Two new debugging options have been added: if you have compiled your |
1775 | Perl with debugging, you can use the -DT and -DR options to trace |
1776 | tokenising and to add reference counts to displaying variables, |
1777 | respectively. |
f39f21d8 |
1778 | |
1779 | =item * |
1780 | |
77c8cf41 |
1781 | If an attempt to use a (non-blessed) reference as an array index |
1782 | is made, a warning is given. |
f39f21d8 |
1783 | |
1784 | =item * |
1785 | |
77c8cf41 |
1786 | C<push @a;> and C<unshift @a;> (with no values to push or unshift) |
1787 | now give a warning. This may be a problem for generated and evaled |
1788 | code. |
f39f21d8 |
1789 | |
1790 | =back |
1791 | |
77c8cf41 |
1792 | =head1 Changed Internals |
f39f21d8 |
1793 | |
1794 | =over 4 |
1795 | |
1796 | =item * |
1797 | |
77c8cf41 |
1798 | perlapi.pod (a companion to perlguts) now attempts to document the |
1799 | internal API. |
f39f21d8 |
1800 | |
1801 | =item * |
1802 | |
77c8cf41 |
1803 | You can now build a really minimal perl called microperl. |
1804 | Building microperl does not require even running Configure; |
1805 | C<make -f Makefile.micro> should be enough. Beware: microperl makes |
1806 | many assumptions, some of which may be too bold; the resulting |
1807 | executable may crash or otherwise misbehave in wondrous ways. |
1808 | For careful hackers only. |
f39f21d8 |
1809 | |
1810 | =item * |
1811 | |
77c8cf41 |
1812 | Added rsignal(), whichsig(), do_join() to the publicised API. |
f39f21d8 |
1813 | |
1814 | =item * |
1815 | |
77c8cf41 |
1816 | Made possible to propagate customised exceptions via croak()ing. |
f39f21d8 |
1817 | |
77c8cf41 |
1818 | =item * |
f39f21d8 |
1819 | |
77c8cf41 |
1820 | Added is_utf8_char(), is_utf8_string(), bytes_to_utf8(), and utf8_to_bytes(). |
f39f21d8 |
1821 | |
77c8cf41 |
1822 | =item * |
f39f21d8 |
1823 | |
77c8cf41 |
1824 | Now xsubs can have attributes just like subs. |
f39f21d8 |
1825 | |
1826 | =item * |
1827 | |
77c8cf41 |
1828 | Some new APIs: ptr_table_clear(), ptr_table_free(), sv_setref_uv(). |
1829 | For the full list of the available APIs see L<perlapi>. |
f39f21d8 |
1830 | |
1831 | =item * |
1832 | |
77c8cf41 |
1833 | dTHR and djSP have been obsoleted; the former removed (because it's |
1834 | a no-op) and the latter replaced with dSP. |
f39f21d8 |
1835 | |
1836 | =item * |
1837 | |
77c8cf41 |
1838 | Perl now uses system malloc instead of Perl malloc on all 64-bit |
1839 | platforms, and even in some not-always-64-bit platforms like AIX, |
1840 | IRIX, and Solaris. This change breaks backward compatibility but |
1841 | Perl's malloc has problems with large address spaces and also the |
1842 | speed of vendors' malloc is generally better in large address space |
1843 | machines (Perl's malloc is mostly tuned for space). |
f39f21d8 |
1844 | |
77c8cf41 |
1845 | =back |
f39f21d8 |
1846 | |
77c8cf41 |
1847 | =head1 Security Vulnerability Closed |
f39f21d8 |
1848 | |
77c8cf41 |
1849 | (This change was already made in 5.7.0 but bears repeating here.) |
f39f21d8 |
1850 | |
77c8cf41 |
1851 | A potential security vulnerability in the optional suidperl component |
1852 | of Perl was identified in August 2000. suidperl is neither built nor |
1853 | installed by default. As of November 2001 the only known vulnerable |
1854 | platform is Linux, most likely all Linux distributions. CERT and |
1855 | various vendors and distributors have been alerted about the vulnerability. |
1856 | See http://www.cpan.org/src/5.0/sperl-2000-08-05/sperl-2000-08-05.txt |
1857 | for more information. |
f39f21d8 |
1858 | |
77c8cf41 |
1859 | The problem was caused by Perl trying to report a suspected security |
1860 | exploit attempt using an external program, /bin/mail. On Linux |
1861 | platforms the /bin/mail program had an undocumented feature which |
1862 | when combined with suidperl gave access to a root shell, resulting in |
1863 | a serious compromise instead of reporting the exploit attempt. If you |
1864 | don't have /bin/mail, or if you have 'safe setuid scripts', or if |
1865 | suidperl is not installed, you are safe. |
f39f21d8 |
1866 | |
77c8cf41 |
1867 | The exploit attempt reporting feature has been completely removed from |
1868 | Perl 5.8.0 (and the maintenance release 5.6.1, and it was removed also |
1869 | from all the Perl 5.7 releases), so that particular vulnerability |
1870 | isn't there anymore. However, further security vulnerabilities are, |
1871 | unfortunately, always possible. The suidperl code is being reviewed |
1872 | and if deemed too risky to continue to be supported, it may be |
1873 | completely removed from future releases. In any case, suidperl should |
1874 | only be used by security experts who know exactly what they are doing |
1875 | and why they are using suidperl instead of some other solution such as |
1876 | sudo (see http://www.courtesan.com/sudo/). |
f39f21d8 |
1877 | |
77c8cf41 |
1878 | =head1 Selected Bug Fixes |
f39f21d8 |
1879 | |
77c8cf41 |
1880 | Numerous memory leaks and uninitialized memory accesses have been hunted down. |
1881 | Most importantly anonymous subs used to leak quite a bit. |
1882 | |
1883 | =over 4 |
f39f21d8 |
1884 | |
1885 | =item * |
1886 | |
77c8cf41 |
1887 | chop(@list) in list context returned the characters chopped in |
1888 | reverse order. This has been reversed to be in the right order. |
f39f21d8 |
1889 | |
77c8cf41 |
1890 | =item * |
f39f21d8 |
1891 | |
77c8cf41 |
1892 | The order of DESTROYs has been made more predictable. |
f39f21d8 |
1893 | |
77c8cf41 |
1894 | =item * |
1895 | |
1896 | mkdir() now ignores trailing slashes in the directory name, |
1897 | as mandated by POSIX. |
f39f21d8 |
1898 | |
1899 | =item * |
1900 | |
77c8cf41 |
1901 | Attributes (like :shared) didn't work with our(). |
f39f21d8 |
1902 | |
1903 | =item * |
1904 | |
77c8cf41 |
1905 | The PERL5OPT environment variable (for passing command line arguments |
1906 | to Perl) didn't work for more than a single group of options. |
f39f21d8 |
1907 | |
1908 | =item * |
1909 | |
77c8cf41 |
1910 | The tainting behaviour of sprintf() has been rationalized. It does |
1911 | not taint the result of floating point formats anymore, making the |
1912 | behaviour consistent with that of string interpolation. |
f39f21d8 |
1913 | |
1914 | =item * |
1915 | |
77c8cf41 |
1916 | All but the first argument of the IO syswrite() method are now optional. |
f39f21d8 |
1917 | |
77c8cf41 |
1918 | =item * |
f39f21d8 |
1919 | |
77c8cf41 |
1920 | Tie::ARRAY SPLICE method was broken. |
1921 | |
1922 | =item * |
f39f21d8 |
1923 | |
77c8cf41 |
1924 | vec() now tries to work with characters <= 255 when possible, but it leaves |
1925 | higher character values in place. In that case, if vec() was used to modify |
1926 | the string, it is no longer considered to be utf8-encoded. |
f39f21d8 |
1927 | |
1928 | =item * |
1929 | |
1930 | The autouse pragma didn't work for Multi::Part::Function::Names. |
1931 | |
1932 | =item * |
1933 | |
1934 | The behaviour of non-decimal but numeric string constants such as |
1935 | "0x23" was platform-dependent: in some platforms that was seen as 35, |
1936 | in some as 0, in some as a floating point number (don't ask). This |
1937 | was caused by Perl using the operating system libraries in a situation |
1938 | where the result of the string to number conversion is undefined: now |
1939 | Perl consistently handles such strings as zero in numeric contexts. |
1940 | |
1941 | =item * |
1942 | |
1943 | L<dprofpp> -R didn't work. |
1944 | |
1945 | =item * |
1946 | |
1947 | PERL5OPT with embedded spaces didn't work. |
1948 | |
1949 | =item * |
1950 | |
1951 | L<Sys::Syslog> ignored the C<LOG_AUTH> constant. |
1952 | |
1953 | =back |
1954 | |
1955 | =head2 Platform Specific Changes and Fixes |
1956 | |
1957 | =over 4 |
1958 | |
1959 | =item * |
1960 | |
1961 | Some versions of glibc have a broken modfl(). This affects builds |
1962 | with C<-Duselongdouble>. This version of Perl detects this brokenness |
1963 | and has a workaround for it. The glibc release 2.2.2 is known to have |
1964 | fixed the modfl() bug. |
1965 | |
1966 | =back |
1967 | |
77c8cf41 |
1968 | =head2 Platform Specific Changes and Fixes |
1969 | |
1970 | =over 4 |
1971 | |
1972 | =item * |
1973 | |
1974 | Linux previously had problems related to sockaddrlen when using |
1975 | accept(), revcfrom() (in Perl: recv()), getpeername(), and getsockname(). |
1976 | |
1977 | =item * |
1978 | |
1979 | Previously DYNIX/ptx had problems in its Configure probe for non-blocking I/O. |
1980 | |
1981 | =item * |
1982 | |
1983 | Windows |
1984 | |
1985 | =over 8 |
1986 | |
1987 | =item * |
1988 | |
1989 | Borland C++ v5.5 is now a supported compiler that can build Perl. |
1990 | However, the generated binaries continue to be incompatible with those |
1991 | generated by the other supported compilers (GCC and Visual C++). |
1992 | |
1993 | =item * |
1994 | |
1995 | Win32::GetCwd() correctly returns C:\ instead of C: when at the drive root. |
1996 | Other bugs in chdir() and Cwd::cwd() have also been fixed. |
1997 | |
1998 | =item * |
1999 | |
2000 | Duping socket handles with open(F, ">&MYSOCK") now works under Windows 9x. |
2001 | |
2002 | =item * |
2003 | |
2004 | HTML files will be installed in c:\perl\html instead of c:\perl\lib\pod\html |
2005 | |
2006 | =item * |
2007 | |
2008 | The makefiles now provide a single switch to bulk-enable all the features |
2009 | enabled in ActiveState ActivePerl (a popular binary distribution). |
2010 | |
2011 | =back |
2012 | |
2013 | =head1 New Tests |
2014 | |
2015 | Several new tests have been added, especially for the F<lib> subsection. |
2016 | |
2017 | The tests are now reported in a different order than in earlier Perls. |
2018 | (This happens because the test scripts from under t/lib have been moved |
2019 | to be closer to the library/extension they are testing.) |
2020 | |
f39f21d8 |
2021 | =head1 New or Changed Diagnostics |
2022 | |
2023 | =over 4 |
2024 | |
2025 | =item * |
2026 | |
2027 | In the regular expression diagnostics the C<E<lt>E<lt> HERE> marker |
2028 | introduced in 5.7.0 has been changed to be C<E<lt>-- HERE> since too |
2029 | many people found the C<E<lt>E<lt>> to be too similar to here-document |
2030 | starters. |
2031 | |
2032 | =item * |
2033 | |
2034 | If you try to L<perlfunc/pack> a number less than 0 or larger than 255 |
2035 | using the C<"C"> format you will get an optional warning. Similarly |
2036 | for the C<"c"> format and a number less than -128 or more than 127. |
2037 | |
2038 | =item * |
2039 | |
2040 | Certain regex modifiers such as C<(?o)> make sense only if applied to |
2041 | the entire regex. You will an optional warning if you try to do otherwise. |
2042 | |
2043 | =item * |
2044 | |
2045 | Using arrays or hashes as references (e.g. C<%foo->{bar}> has been |
2046 | deprecated for a while. Now you will get an optional warning. |
2047 | |
2048 | =back |
2049 | |
2050 | =head1 Source Code Enhancements |
2051 | |
2052 | =head2 MAGIC constants |
2053 | |
2054 | The MAGIC constants (e.g. C<'P'>) have been macrofied |
2055 | (e.g. C<PERL_MAGIC_TIED>) for better source code readability |
2056 | and maintainability. |
2057 | |
2058 | =head2 Better commented code |
2059 | |
2060 | F<perly.c>, F<sv.c>, and F<sv.h> have now been extensively commented. |
2061 | |
2062 | =head2 Regex pre-/post-compilation items matched up |
2063 | |
2064 | The regex compiler now maintains a structure that identifies nodes in |
2065 | the compiled bytecode with the corresponding syntactic features of the |
2066 | original regex expression. The information is attached to the new |
2067 | C<offsets> member of the C<struct regexp>. See L<perldebguts> for more |
2068 | complete information. |
2069 | |
2070 | =head2 gcc -Wall |
2071 | |
2072 | The C code has been made much more C<gcc -Wall> clean. Some warning |
2073 | messages still remain, though, so if you are compiling with gcc you |
2074 | will see some warnings about dubious practices. The warnings are |
2075 | being worked on. |
2076 | |
f39f21d8 |
2077 | =head1 Known Problems |
2078 | |
2079 | Note that unlike other sections in this document (which describe |
2080 | changes since 5.7.0) this section is cumulative containing known |
2081 | problems for all the 5.7 releases. |
2082 | |
2083 | =head2 AIX |
2084 | |
2085 | =over 4 |
2086 | |
2087 | =item * |
2088 | |
2089 | In AIX 4.2 Perl extensions that use C++ functions that use statics |
2090 | may have problems in that the statics are not getting initialized. |
2091 | In newer AIX releases this has been solved by linking Perl with |
2092 | the libC_r library, but unfortunately in AIX 4.2 the said library |
2093 | has an obscure bug where the various functions related to time |
2094 | (such as time() and gettimeofday()) return broken values, and |
2095 | therefore in AIX 4.2 Perl is not linked against the libC_r. |
2096 | |
2097 | =item * |
2098 | |
2099 | vac 5.0.0.0 May Produce Buggy Code For Perl |
2100 | |
2101 | The AIX C compiler vac version 5.0.0.0 may produce buggy code, |
2102 | resulting in few random tests failing, but when the failing tests |
2103 | are run by hand, they succeed. We suggest upgrading to at least |
2104 | vac version 5.0.1.0, that has been known to compile Perl correctly. |
2105 | "lslpp -L|grep vac.C" will tell you the vac version. |
2106 | |
2107 | =back |
2108 | |
2109 | =head2 Amiga Perl Invoking Mystery |
2110 | |
2111 | One cannot call Perl using the C<volume:> syntax, that is, C<perl -v> |
2112 | works, but for example C<bin:perl -v> doesn't. The exact reason is |
2113 | known but the current suspect is the F<ixemul> library. |
2114 | |
2115 | =head2 lib/ftmp-security tests warn 'system possibly insecure' |
2116 | |
2117 | Don't panic. Read INSTALL 'make test' section instead. |
2118 | |
2119 | =head2 Cygwin intermittent failures of lib/Memoize/t/expire_file 11 and 12 |
2120 | |
2121 | The subtests 11 and 12 sometimes fail and sometimes work. |
2122 | |
2123 | =head2 HP-UX lib/io_multihomed Fails When LP64-Configured |
2124 | |
2125 | The lib/io_multihomed test may hang in HP-UX if Perl has been |
2126 | configured to be 64-bit. Because other 64-bit platforms do not hang in |
2127 | this test, HP-UX is suspect. All other tests pass in 64-bit HP-UX. The |
2128 | test attempts to create and connect to "multihomed" sockets (sockets |
2129 | which have multiple IP addresses). |
2130 | |
2131 | =head2 HP-UX lib/posix Subtest 9 Fails When LP64-Configured |
2132 | |
2133 | If perl is configured with -Duse64bitall, the successful result of the |
2134 | subtest 10 of lib/posix may arrive before the successful result of the |
2135 | subtest 9, which confuses the test harness so much that it thinks the |
2136 | subtest 9 failed. |
2137 | |
2138 | =head2 Linux With Sfio Fails op/misc Test 48 |
2139 | |
2140 | No known fix. |
2141 | |
2142 | =head2 OS/390 |
2143 | |
2144 | OS/390 has rather many test failures but the situation is actually |
2145 | better than it was in 5.6.0, it's just that so many new modules and |
2146 | tests have been added. |
2147 | |
2148 | Failed Test Stat Wstat Total Fail Failed List of Failed |
2149 | ----------------------------------------------------------------------------- |
2150 | ../ext/B/Deparse.t 14 1 7.14% 14 |
2151 | ../ext/B/Showlex.t 1 1 100.00% 1 |
2152 | ../ext/Encode/Encode/Tcl.t 610 13 2.13% 592 594 596 598 |
2153 | 600 602 604-610 |
2154 | ../ext/IO/lib/IO/t/io_unix.t 113 28928 5 3 60.00% 3-5 |
2155 | ../ext/POSIX/POSIX.t 29 1 3.45% 14 |
2156 | ../ext/Storable/t/lock.t 255 65280 5 3 60.00% 3-5 |
2157 | ../lib/locale.t 129 33024 117 19 16.24% 99-117 |
2158 | ../lib/warnings.t 434 1 0.23% 75 |
2159 | ../lib/ExtUtils.t 27 1 3.70% 25 |
2160 | ../lib/Math/BigInt/t/bigintpm.t 1190 1 0.08% 1145 |
2161 | ../lib/Unicode/UCD.t 81 48 59.26% 1-16 49-64 66-81 |
2162 | ../lib/User/pwent.t 9 1 11.11% 4 |
2163 | op/pat.t 660 6 0.91% 242-243 424-425 |
2164 | 626-627 |
2165 | op/split.t 0 9 ?? ?? % ?? |
2166 | op/taint.t 174 3 1.72% 156 162 168 |
2167 | op/tr.t 70 3 4.29% 50 58-59 |
2168 | Failed 16/422 test scripts, 96.21% okay. 105/23251 subtests failed, 99.55% okay. |
2169 | |
2170 | =head2 op/sprintf tests 129 and 130 |
2171 | |
2172 | The op/sprintf tests 129 and 130 are known to fail on some platforms. |
2173 | Examples include any platform using sfio, and Compaq/Tandem's NonStop-UX. |
2174 | The failing platforms do not comply with the ANSI C Standard, line |
2175 | 19ff on page 134 of ANSI X3.159 1989 to be exact. (They produce |
2176 | something other than "1" and "-1" when formatting 0.6 and -0.6 using |
2177 | the printf format "%.0f", most often they produce "0" and "-0".) |
2178 | |
2179 | =head2 Failure of Thread tests |
2180 | |
2181 | B<Note that support for 5.005-style threading remains experimental.> |
2182 | |
2183 | The following tests are known to fail due to fundamental problems in |
2184 | the 5.005 threading implementation. These are not new failures--Perl |
2185 | 5.005_0x has the same bugs, but didn't have these tests. |
2186 | |
2187 | lib/autouse.t 4 |
2188 | t/lib/thr5005.t 19-20 |
2189 | |
2190 | =head2 UNICOS |
2191 | |
2192 | =over 4 |
2193 | |
2194 | =item * |
2195 | |
2196 | ext/POSIX/sigaction subtests 6 and 13 may fail. |
2197 | |
2198 | =item * |
2199 | |
2200 | lib/ExtUtils may spuriously claim that subtest 28 failed, |
2201 | which is interesting since the test only has 27 tests. |
2202 | |
2203 | =item * |
2204 | |
2205 | Numerous numerical test failures |
2206 | |
2207 | op/numconvert 209,210,217,218 |
2208 | op/override 7 |
2209 | ext/Time/HiRes/HiRes 9 |
2210 | lib/Math/BigInt/t/bigintpm 1145 |
2211 | lib/Math/Trig 25 |
2212 | |
2213 | These tests fail because of yet unresolved floating point inaccuracies. |
2214 | |
2215 | =back |
2216 | |
2217 | =head2 UTS |
2218 | |
2219 | There are a few known test failures, see L<perluts>. |
2220 | |
2221 | =head2 VMS |
2222 | |
2223 | Rather many tests are failing in VMS but that actually more tests |
2224 | succeed in VMS than they used to, it's just that there are many, |
2225 | many more tests than there used to be. |
2226 | |
2227 | Here are the known failures from some compiler/platform combinations. |
2228 | |
2229 | DEC C V5.3-006 on OpenVMS VAX V6.2 |
2230 | |
2231 | [-.ext.list.util.t]tainted..............FAILED on test 3 |
2232 | [-.ext.posix]sigaction..................FAILED on test 7 |
2233 | [-.ext.time.hires]hires.................FAILED on test 14 |
2234 | [-.lib.file.find]taint..................FAILED on test 17 |
2235 | [-.lib.math.bigint.t]bigintpm...........FAILED on test 1183 |
2236 | [-.lib.test.simple.t]exit...............FAILED on test 1 |
2237 | [.lib]vmsish............................FAILED on test 13 |
2238 | [.op]sprintf............................FAILED on test 12 |
2239 | Failed 8/399 tests, 91.23% okay. |
2240 | |
2241 | DEC C V6.0-001 on OpenVMS Alpha V7.2-1 and |
2242 | Compaq C V6.2-008 on OpenVMS Alpha V7.1 |
2243 | |
2244 | [-.ext.list.util.t]tainted..............FAILED on test 3 |
2245 | [-.lib.file.find]taint..................FAILED on test 17 |
2246 | [-.lib.test.simple.t]exit...............FAILED on test 1 |
2247 | [.lib]vmsish............................FAILED on test 13 |
2248 | Failed 4/399 tests, 92.48% okay. |
2249 | |
2250 | Compaq C V6.4-005 on OpenVMS Alpha 7.2.1 |
2251 | |
2252 | [-.ext.b]showlex........................FAILED on test 1 |
2253 | [-.ext.list.util.t]tainted..............FAILED on test 3 |
2254 | [-.lib.file.find]taint..................FAILED on test 17 |
2255 | [-.lib.test.simple.t]exit...............FAILED on test 1 |
2256 | [.lib]vmsish............................FAILED on test 13 |
2257 | [.op]misc...............................FAILED on test 49 |
2258 | Failed 6/401 tests, 92.77% okay. |
2259 | |
2260 | =head2 Win32 |
2261 | |
2262 | In multi-CPU boxes there are some problems with the I/O buffering: |
2263 | some output may appear twice. |
2264 | |
2265 | =head2 Localising a Tied Variable Leaks Memory |
2266 | |
2267 | use Tie::Hash; |
2268 | tie my %tie_hash => 'Tie::StdHash'; |
2269 | |
2270 | ... |
2271 | |
2272 | local($tie_hash{Foo}) = 1; # leaks |
2273 | |
2274 | Code like the above is known to leak memory every time the local() |
2275 | is executed. |
2276 | |
2277 | =head2 Self-tying of Arrays and Hashes Is Forbidden |
2278 | |
2279 | Self-tying of arrays and hashes is broken in rather deep and |
2280 | hard-to-fix ways. As a stop-gap measure to avoid people from getting |
2281 | frustrated at the mysterious results (core dumps, most often) it is |
2282 | for now forbidden (you will get a fatal error even from an attempt). |
2283 | |
2284 | =head2 Variable Attributes are not Currently Usable for Tieing |
2285 | |
2286 | This limitation will hopefully be fixed in future. (Subroutine |
2287 | attributes work fine for tieing, see L<Attribute::Handlers>). |
2288 | |
2289 | =head2 Building Extensions Can Fail Because Of Largefiles |
2290 | |
2291 | Some extensions like mod_perl are known to have issues with |
2292 | `largefiles', a change brought by Perl 5.6.0 in which file offsets |
2293 | default to 64 bits wide, where supported. Modules may fail to compile |
2294 | at all or compile and work incorrectly. Currently there is no good |
2295 | solution for the problem, but Configure now provides appropriate |
2296 | non-largefile ccflags, ldflags, libswanted, and libs in the %Config |
2297 | hash (e.g., $Config{ccflags_nolargefiles}) so the extensions that are |
2298 | having problems can try configuring themselves without the |
2299 | largefileness. This is admittedly not a clean solution, and the |
2300 | solution may not even work at all. One potential failure is whether |
2301 | one can (or, if one can, whether it's a good idea) link together at |
2302 | all binaries with different ideas about file offsets, all this is |
2303 | platform-dependent. |
2304 | |
2305 | =head2 The Compiler Suite Is Still Experimental |
2306 | |
2307 | The compiler suite is slowly getting better but is nowhere near |
2308 | working order yet. |
2309 | |
2310 | =head2 The Long Double Support is Still Experimental |
2311 | |
2312 | The ability to configure Perl's numbers to use "long doubles", |
2313 | floating point numbers of hopefully better accuracy, is still |
2314 | experimental. The implementations of long doubles are not yet |
2315 | widespread and the existing implementations are not quite mature |
2316 | or standardised, therefore trying to support them is a rare |
2317 | and moving target. The gain of more precision may also be offset |
2318 | by slowdown in computations (more bits to move around, and the |
2319 | operations are more likely to be executed by less optimised |
2320 | libraries). |
33a87e58 |
2321 | |
cc0fca54 |
2322 | =head1 Reporting Bugs |
2323 | |
d4ad863d |
2324 | If you find what you think is a bug, you might check the articles |
2325 | recently posted to the comp.lang.perl.misc newsgroup and the perl |
2326 | bug database at http://bugs.perl.org. There may also be |
2327 | information at http://www.perl.com/perl/, the Perl Home Page. |
cc0fca54 |
2328 | |
2329 | If you believe you have an unreported bug, please run the B<perlbug> |
2330 | program included with your release. Be sure to trim your bug down |
2331 | to a tiny but sufficient test case. Your bug report, along with the |
d4ad863d |
2332 | output of C<perl -V>, will be sent off to perlbug@perl.org to be |
cc0fca54 |
2333 | analysed by the Perl porting team. |
2334 | |
2335 | =head1 SEE ALSO |
2336 | |
2337 | The F<Changes> file for exhaustive details on what changed. |
2338 | |
2339 | The F<INSTALL> file for how to build Perl. |
2340 | |
2341 | The F<README> file for general stuff. |
2342 | |
2343 | The F<Artistic> and F<Copying> files for copyright information. |
2344 | |
2345 | =head1 HISTORY |
2346 | |
d468ca04 |
2347 | Written by Jarkko Hietaniemi <F<jhi@iki.fi>>. |
cc0fca54 |
2348 | |
2349 | =cut |