Commit | Line | Data |
---|---|---|
920dae64 AT |
1 | =head1 NAME |
2 | ||
3 | perl581delta - what is new for perl v5.8.1 | |
4 | ||
5 | =head1 DESCRIPTION | |
6 | ||
7 | This document describes differences between the 5.8.0 release and | |
8 | the 5.8.1 release. | |
9 | ||
10 | If you are upgrading from an earlier release such as 5.6.1, first read | |
11 | the L<perl58delta>, which describes differences between 5.6.0 and | |
12 | 5.8.0. | |
13 | ||
14 | In case you are wondering about 5.6.1, it was bug-fix-wise rather | |
15 | identical to the development release 5.7.1. Confused? This timeline | |
16 | hopefully helps a bit: it lists the new major releases, their maintenance | |
17 | releases, and the development releases. | |
18 | ||
19 | New Maintenance Development | |
20 | ||
21 | 5.6.0 2000-Mar-22 | |
22 | 5.7.0 2000-Sep-02 | |
23 | 5.6.1 2001-Apr-08 | |
24 | 5.7.1 2001-Apr-09 | |
25 | 5.7.2 2001-Jul-13 | |
26 | 5.7.3 2002-Mar-05 | |
27 | 5.8.0 2002-Jul-18 | |
28 | 5.8.1 2003-Sep-25 | |
29 | ||
30 | =head1 Incompatible Changes | |
31 | ||
32 | =head2 Hash Randomisation | |
33 | ||
34 | Mainly due to security reasons, the "random ordering" of hashes | |
35 | has been made even more random. Previously while the order of hash | |
36 | elements from keys(), values(), and each() was essentially random, | |
37 | it was still repeatable. Now, however, the order varies between | |
38 | different runs of Perl. | |
39 | ||
40 | B<Perl has never guaranteed any ordering of the hash keys>, and the | |
41 | ordering has already changed several times during the lifetime of | |
42 | Perl 5. Also, the ordering of hash keys has always been, and | |
43 | continues to be, affected by the insertion order. | |
44 | ||
45 | The added randomness may affect applications. | |
46 | ||
47 | One possible scenario is when output of an application has included | |
48 | hash data. For example, if you have used the Data::Dumper module to | |
49 | dump data into different files, and then compared the files to see | |
50 | whether the data has changed, now you will have false positives since | |
51 | the order in which hashes are dumped will vary. In general the cure | |
52 | is to sort the keys (or the values); in particular for Data::Dumper to | |
53 | use the C<Sortkeys> option. If some particular order is really | |
54 | important, use tied hashes: for example the Tie::IxHash module | |
55 | which by default preserves the order in which the hash elements | |
56 | were added. | |
57 | ||
58 | More subtle problem is reliance on the order of "global destruction". | |
59 | That is what happens at the end of execution: Perl destroys all data | |
60 | structures, including user data. If your destructors (the DESTROY | |
61 | subroutines) have assumed any particular ordering to the global | |
62 | destruction, there might be problems ahead. For example, in a | |
63 | destructor of one object you cannot assume that objects of any other | |
64 | class are still available, unless you hold a reference to them. | |
65 | If the environment variable PERL_DESTRUCT_LEVEL is set to a non-zero | |
66 | value, or if Perl is exiting a spawned thread, it will also destruct | |
67 | the ordinary references and the symbol tables that are no longer in use. | |
68 | You can't call a class method or an ordinary function on a class that | |
69 | has been collected that way. | |
70 | ||
71 | The hash randomisation is certain to reveal hidden assumptions about | |
72 | some particular ordering of hash elements, and outright bugs: it | |
73 | revealed a few bugs in the Perl core and core modules. | |
74 | ||
75 | To disable the hash randomisation in runtime, set the environment | |
76 | variable PERL_HASH_SEED to 0 (zero) before running Perl (for more | |
77 | information see L<perlrun/PERL_HASH_SEED>), or to disable the feature | |
78 | completely in compile time, compile with C<-DNO_HASH_SEED> (see F<INSTALL>). | |
79 | ||
80 | See L<perlsec/"Algorithmic Complexity Attacks"> for the original | |
81 | rationale behind this change. | |
82 | ||
83 | =head2 UTF-8 On Filehandles No Longer Activated By Locale | |
84 | ||
85 | In Perl 5.8.0 all filehandles, including the standard filehandles, | |
86 | were implicitly set to be in Unicode UTF-8 if the locale settings | |
87 | indicated the use of UTF-8. This feature caused too many problems, | |
88 | so the feature was turned off and redesigned: see L</"Core Enhancements">. | |
89 | ||
90 | =head2 Single-number v-strings are no longer v-strings before "=>" | |
91 | ||
92 | The version strings or v-strings (see L<perldata/"Version Strings">) | |
93 | feature introduced in Perl 5.6.0 has been a source of some confusion-- | |
94 | especially when the user did not want to use it, but Perl thought it | |
95 | knew better. Especially troublesome has been the feature that before | |
96 | a "=>" a version string (a "v" followed by digits) has been interpreted | |
97 | as a v-string instead of a string literal. In other words: | |
98 | ||
99 | %h = ( v65 => 42 ); | |
100 | ||
101 | has meant since Perl 5.6.0 | |
102 | ||
103 | %h = ( 'A' => 42 ); | |
104 | ||
105 | (at least in platforms of ASCII progeny) Perl 5.8.1 restores the | |
106 | more natural interpretation | |
107 | ||
108 | %h = ( 'v65' => 42 ); | |
109 | ||
110 | The multi-number v-strings like v65.66 and 65.66.67 still continue to | |
111 | be v-strings in Perl 5.8. | |
112 | ||
113 | =head2 (Win32) The -C Switch Has Been Repurposed | |
114 | ||
115 | The -C switch has changed in an incompatible way. The old semantics | |
116 | of this switch only made sense in Win32 and only in the "use utf8" | |
117 | universe in 5.6.x releases, and do not make sense for the Unicode | |
118 | implementation in 5.8.0. Since this switch could not have been used | |
119 | by anyone, it has been repurposed. The behavior that this switch | |
120 | enabled in 5.6.x releases may be supported in a transparent, | |
121 | data-dependent fashion in a future release. | |
122 | ||
123 | For the new life of this switch, see L<"UTF-8 no longer default under | |
124 | UTF-8 locales">, and L<perlrun/-C>. | |
125 | ||
126 | =head2 (Win32) The /d Switch Of cmd.exe | |
127 | ||
128 | Perl 5.8.1 uses the /d switch when running the cmd.exe shell | |
129 | internally for system(), backticks, and when opening pipes to external | |
130 | programs. The extra switch disables the execution of AutoRun commands | |
131 | from the registry, which is generally considered undesirable when | |
132 | running external programs. If you wish to retain compatibility with | |
133 | the older behavior, set PERL5SHELL in your environment to C<cmd /x/c>. | |
134 | ||
135 | =head1 Core Enhancements | |
136 | ||
137 | =head2 UTF-8 no longer default under UTF-8 locales | |
138 | ||
139 | In Perl 5.8.0 many Unicode features were introduced. One of them | |
140 | was found to be of more nuisance than benefit: the automagic | |
141 | (and silent) "UTF-8-ification" of filehandles, including the | |
142 | standard filehandles, if the user's locale settings indicated | |
143 | use of UTF-8. | |
144 | ||
145 | For example, if you had C<en_US.UTF-8> as your locale, your STDIN and | |
146 | STDOUT were automatically "UTF-8", in other words an implicit | |
147 | binmode(..., ":utf8") was made. This meant that trying to print, say, | |
148 | chr(0xff), ended up printing the bytes 0xc3 0xbf. Hardly what | |
149 | you had in mind unless you were aware of this feature of Perl 5.8.0. | |
150 | The problem is that the vast majority of people weren't: for example | |
151 | in RedHat releases 8 and 9 the B<default> locale setting is UTF-8, so | |
152 | all RedHat users got UTF-8 filehandles, whether they wanted it or not. | |
153 | The pain was intensified by the Unicode implementation of Perl 5.8.0 | |
154 | (still) having nasty bugs, especially related to the use of s/// and | |
155 | tr///. (Bugs that have been fixed in 5.8.1) | |
156 | ||
157 | Therefore a decision was made to backtrack the feature and change it | |
158 | from implicit silent default to explicit conscious option. The new | |
159 | Perl command line option C<-C> and its counterpart environment | |
160 | variable PERL_UNICODE can now be used to control how Perl and Unicode | |
161 | interact at interfaces like I/O and for example the command line | |
162 | arguments. See L<perlrun/-C> and L<perlrun/PERL_UNICODE> for more | |
163 | information. | |
164 | ||
165 | =head2 Unsafe signals again available | |
166 | ||
167 | In Perl 5.8.0 the so-called "safe signals" were introduced. This | |
168 | means that Perl no longer handles signals immediately but instead | |
169 | "between opcodes", when it is safe to do so. The earlier immediate | |
170 | handling easily could corrupt the internal state of Perl, resulting | |
171 | in mysterious crashes. | |
172 | ||
173 | However, the new safer model has its problems too. Because now an | |
174 | opcode, a basic unit of Perl execution, is never interrupted but | |
175 | instead let to run to completion, certain operations that can take a | |
176 | long time now really do take a long time. For example, certain | |
177 | network operations have their own blocking and timeout mechanisms, and | |
178 | being able to interrupt them immediately would be nice. | |
179 | ||
180 | Therefore perl 5.8.1 introduces a "backdoor" to restore the pre-5.8.0 | |
181 | (pre-5.7.3, really) signal behaviour. Just set the environment variable | |
182 | PERL_SIGNALS to C<unsafe>, and the old immediate (and unsafe) | |
183 | signal handling behaviour returns. See L<perlrun/PERL_SIGNALS> | |
184 | and L<perlipc/"Deferred Signals (Safe Signals)">. | |
185 | ||
186 | In completely unrelated news, you can now use safe signals with | |
187 | POSIX::SigAction. See L<POSIX/POSIX::SigAction>. | |
188 | ||
189 | =head2 Tied Arrays with Negative Array Indices | |
190 | ||
191 | Formerly, the indices passed to C<FETCH>, C<STORE>, C<EXISTS>, and | |
192 | C<DELETE> methods in tied array class were always non-negative. If | |
193 | the actual argument was negative, Perl would call FETCHSIZE implicitly | |
194 | and add the result to the index before passing the result to the tied | |
195 | array method. This behaviour is now optional. If the tied array class | |
196 | contains a package variable named C<$NEGATIVE_INDICES> which is set to | |
197 | a true value, negative values will be passed to C<FETCH>, C<STORE>, | |
198 | C<EXISTS>, and C<DELETE> unchanged. | |
199 | ||
200 | =head2 local ${$x} | |
201 | ||
202 | The syntaxes | |
203 | ||
204 | local ${$x} | |
205 | local @{$x} | |
206 | local %{$x} | |
207 | ||
208 | now do localise variables, given that the $x is a valid variable name. | |
209 | ||
210 | =head2 Unicode Character Database 4.0.0 | |
211 | ||
212 | The copy of the Unicode Character Database included in Perl 5.8 has | |
213 | been updated to 4.0.0 from 3.2.0. This means for example that the | |
214 | Unicode character properties are as in Unicode 4.0.0. | |
215 | ||
216 | =head2 Deprecation Warnings | |
217 | ||
218 | There is one new feature deprecation. Perl 5.8.0 forgot to add | |
219 | some deprecation warnings, these warnings have now been added. | |
220 | Finally, a reminder of an impending feature removal. | |
221 | ||
222 | =head3 (Reminder) Pseudo-hashes are deprecated (really) | |
223 | ||
224 | Pseudo-hashes were deprecated in Perl 5.8.0 and will be removed in | |
225 | Perl 5.10.0, see L<perl58delta> for details. Each attempt to access | |
226 | pseudo-hashes will trigger the warning C<Pseudo-hashes are deprecated>. | |
227 | If you really want to continue using pseudo-hashes but not to see the | |
228 | deprecation warnings, use: | |
229 | ||
230 | no warnings 'deprecated'; | |
231 | ||
232 | Or you can continue to use the L<fields> pragma, but please don't | |
233 | expect the data structures to be pseudohashes any more. | |
234 | ||
235 | =head3 (Reminder) 5.005-style threads are deprecated (really) | |
236 | ||
237 | 5.005-style threads (activated by C<use Thread;>) were deprecated in | |
238 | Perl 5.8.0 and will be removed after Perl 5.8, see L<perl58delta> for | |
239 | details. Each 5.005-style thread creation will trigger the warning | |
240 | C<5.005 threads are deprecated>. If you really want to continue | |
241 | using the 5.005 threads but not to see the deprecation warnings, use: | |
242 | ||
243 | no warnings 'deprecated'; | |
244 | ||
245 | =head3 (Reminder) The $* variable is deprecated (really) | |
246 | ||
247 | The C<$*> variable controlling multi-line matching has been deprecated | |
248 | and will be removed after 5.8. The variable has been deprecated for a | |
249 | long time, and a deprecation warning C<Use of $* is deprecated> is given, | |
250 | now the variable will just finally be removed. The functionality has | |
251 | been supplanted by the C</s> and C</m> modifiers on pattern matching. | |
252 | If you really want to continue using the C<$*>-variable but not to see | |
253 | the deprecation warnings, use: | |
254 | ||
255 | no warnings 'deprecated'; | |
256 | ||
257 | =head2 Miscellaneous Enhancements | |
258 | ||
259 | C<map> in void context is no longer expensive. C<map> is now context | |
260 | aware, and will not construct a list if called in void context. | |
261 | ||
262 | If a socket gets closed by the server while printing to it, the client | |
263 | now gets a SIGPIPE. While this new feature was not planned, it fell | |
264 | naturally out of PerlIO changes, and is to be considered an accidental | |
265 | feature. | |
266 | ||
267 | PerlIO::get_layers(FH) returns the names of the PerlIO layers | |
268 | active on a filehandle. | |
269 | ||
270 | PerlIO::via layers can now have an optional UTF8 method to | |
271 | indicate whether the layer wants to "auto-:utf8" the stream. | |
272 | ||
273 | utf8::is_utf8() has been added as a quick way to test whether | |
274 | a scalar is encoded internally in UTF-8 (Unicode). | |
275 | ||
276 | =head1 Modules and Pragmata | |
277 | ||
278 | =head2 Updated Modules And Pragmata | |
279 | ||
280 | The following modules and pragmata have been updated since Perl 5.8.0: | |
281 | ||
282 | =over 4 | |
283 | ||
284 | =item base | |
285 | ||
286 | =item B::Bytecode | |
287 | ||
288 | In much better shape than it used to be. Still far from perfect, but | |
289 | maybe worth a try. | |
290 | ||
291 | =item B::Concise | |
292 | ||
293 | =item B::Deparse | |
294 | ||
295 | =item Benchmark | |
296 | ||
297 | An optional feature, C<:hireswallclock>, now allows for high | |
298 | resolution wall clock times (uses Time::HiRes). | |
299 | ||
300 | =item ByteLoader | |
301 | ||
302 | See B::Bytecode. | |
303 | ||
304 | =item bytes | |
305 | ||
306 | Now has bytes::substr. | |
307 | ||
308 | =item CGI | |
309 | ||
310 | =item charnames | |
311 | ||
312 | One can now have custom character name aliases. | |
313 | ||
314 | =item CPAN | |
315 | ||
316 | There is now a simple command line frontend to the CPAN.pm | |
317 | module called F<cpan>. | |
318 | ||
319 | =item Data::Dumper | |
320 | ||
321 | A new option, Pair, allows choosing the separator between hash keys | |
322 | and values. | |
323 | ||
324 | =item DB_File | |
325 | ||
326 | =item Devel::PPPort | |
327 | ||
328 | =item Digest::MD5 | |
329 | ||
330 | =item Encode | |
331 | ||
332 | Significant updates on the encoding pragma functionality | |
333 | (tr/// and the DATA filehandle, formats). | |
334 | ||
335 | If a filehandle has been marked as to have an encoding, unmappable | |
336 | characters are detected already during input, not later (when the | |
337 | corrupted data is being used). | |
338 | ||
339 | The ISO 8859-6 conversion table has been corrected (the 0x30..0x39 | |
340 | erroneously mapped to U+0660..U+0669, instead of U+0030..U+0039). The | |
341 | GSM 03.38 conversion did not handle escape sequences correctly. The | |
342 | UTF-7 encoding has been added (making Encode feature-complete with | |
343 | Unicode::String). | |
344 | ||
345 | =item fields | |
346 | ||
347 | =item libnet | |
348 | ||
349 | =item Math::BigInt | |
350 | ||
351 | A lot of bugs have been fixed since v1.60, the version included in Perl | |
352 | v5.8.0. Especially noteworthy are the bug in Calc that caused div and mod to | |
353 | fail for some large values, and the fixes to the handling of bad inputs. | |
354 | ||
355 | Some new features were added, e.g. the broot() method, you can now pass | |
356 | parameters to config() to change some settings at runtime, and it is now | |
357 | possible to trap the creation of NaN and infinity. | |
358 | ||
359 | As usual, some optimizations took place and made the math overall a tad | |
360 | faster. In some cases, quite a lot faster, actually. Especially alternative | |
361 | libraries like Math::BigInt::GMP benefit from this. In addition, a lot of the | |
362 | quite clunky routines like fsqrt() and flog() are now much much faster. | |
363 | ||
364 | =item MIME::Base64 | |
365 | ||
366 | =item NEXT | |
367 | ||
368 | Diamond inheritance now works. | |
369 | ||
370 | =item Net::Ping | |
371 | ||
372 | =item PerlIO::scalar | |
373 | ||
374 | Reading from non-string scalars (like the special variables, see | |
375 | L<perlvar>) now works. | |
376 | ||
377 | =item podlators | |
378 | ||
379 | =item Pod::LaTeX | |
380 | ||
381 | =item PodParsers | |
382 | ||
383 | =item Pod::Perldoc | |
384 | ||
385 | Complete rewrite. As a side-effect, no longer refuses to startup when | |
386 | run by root. | |
387 | ||
388 | =item Scalar::Util | |
389 | ||
390 | New utilities: refaddr, isvstring, looks_like_number, set_prototype. | |
391 | ||
392 | =item Storable | |
393 | ||
394 | Can now store code references (via B::Deparse, so not foolproof). | |
395 | ||
396 | =item strict | |
397 | ||
398 | Earlier versions of the strict pragma did not check the parameters | |
399 | implicitly passed to its "import" (use) and "unimport" (no) routine. | |
400 | This caused the false idiom such as: | |
401 | ||
402 | use strict qw(@ISA); | |
403 | @ISA = qw(Foo); | |
404 | ||
405 | This however (probably) raised the false expectation that the strict | |
406 | refs, vars and subs were being enforced (and that @ISA was somehow | |
407 | "declared"). But the strict refs, vars, and subs are B<not> enforced | |
408 | when using this false idiom. | |
409 | ||
410 | Starting from Perl 5.8.1, the above B<will> cause an error to be | |
411 | raised. This may cause programs which used to execute seemingly | |
412 | correctly without warnings and errors to fail when run under 5.8.1. | |
413 | This happens because | |
414 | ||
415 | use strict qw(@ISA); | |
416 | ||
417 | will now fail with the error: | |
418 | ||
419 | Unknown 'strict' tag(s) '@ISA' | |
420 | ||
421 | The remedy to this problem is to replace this code with the correct idiom: | |
422 | ||
423 | use strict; | |
424 | use vars qw(@ISA); | |
425 | @ISA = qw(Foo); | |
426 | ||
427 | =item Term::ANSIcolor | |
428 | ||
429 | =item Test::Harness | |
430 | ||
431 | Now much more picky about extra or missing output from test scripts. | |
432 | ||
433 | =item Test::More | |
434 | ||
435 | =item Test::Simple | |
436 | ||
437 | =item Text::Balanced | |
438 | ||
439 | =item Time::HiRes | |
440 | ||
441 | Use of nanosleep(), if available, allows mixing subsecond sleeps with | |
442 | alarms. | |
443 | ||
444 | =item threads | |
445 | ||
446 | Several fixes, for example for join() problems and memory | |
447 | leaks. In some platforms (like Linux) that use glibc the minimum memory | |
448 | footprint of one ithread has been reduced by several hundred kilobytes. | |
449 | ||
450 | =item threads::shared | |
451 | ||
452 | Many memory leaks have been fixed. | |
453 | ||
454 | =item Unicode::Collate | |
455 | ||
456 | =item Unicode::Normalize | |
457 | ||
458 | =item Win32::GetFolderPath | |
459 | ||
460 | =item Win32::GetOSVersion | |
461 | ||
462 | Now returns extra information. | |
463 | ||
464 | =back | |
465 | ||
466 | =head1 Utility Changes | |
467 | ||
468 | The C<h2xs> utility now produces a more modern layout: | |
469 | F<Foo-Bar/lib/Foo/Bar.pm> instead of F<Foo/Bar/Bar.pm>. | |
470 | Also, the boilerplate test is now called F<t/Foo-Bar.t> | |
471 | instead of F<t/1.t>. | |
472 | ||
473 | The Perl debugger (F<lib/perl5db.pl>) has now been extensively | |
474 | documented and bugs found while documenting have been fixed. | |
475 | ||
476 | C<perldoc> has been rewritten from scratch to be more robust and | |
477 | featureful. | |
478 | ||
479 | C<perlcc -B> works now at least somewhat better, while C<perlcc -c> | |
480 | is rather more broken. (The Perl compiler suite as a whole continues | |
481 | to be experimental.) | |
482 | ||
483 | =head1 New Documentation | |
484 | ||
485 | perl573delta has been added to list the differences between the | |
486 | (now quite obsolete) development releases 5.7.2 and 5.7.3. | |
487 | ||
488 | perl58delta has been added: it is the perldelta of 5.8.0, detailing | |
489 | the differences between 5.6.0 and 5.8.0. | |
490 | ||
491 | perlartistic has been added: it is the Artistic License in pod format, | |
492 | making it easier for modules to refer to it. | |
493 | ||
494 | perlcheat has been added: it is a Perl cheat sheet. | |
495 | ||
496 | perlgpl has been added: it is the GNU General Public License in pod | |
497 | format, making it easier for modules to refer to it. | |
498 | ||
499 | perlmacosx has been added to tell about the installation and use | |
500 | of Perl in Mac OS X. | |
501 | ||
502 | perlos400 has been added to tell about the installation and use | |
503 | of Perl in OS/400 PASE. | |
504 | ||
505 | perlreref has been added: it is a regular expressions quick reference. | |
506 | ||
507 | =head1 Installation and Configuration Improvements | |
508 | ||
509 | The UNIX standard Perl location, F</usr/bin/perl>, is no longer | |
510 | overwritten by default if it exists. This change was very prudent | |
511 | because so many UNIX vendors already provide a F</usr/bin/perl>, | |
512 | but simultaneously many system utilities may depend on that | |
513 | exact version of Perl, so better not to overwrite it. | |
514 | ||
515 | One can now specify installation directories for site and vendor man | |
516 | and HTML pages, and site and vendor scripts. See F<INSTALL>. | |
517 | ||
518 | One can now specify a destination directory for Perl installation | |
519 | by specifying the DESTDIR variable for C<make install>. (This feature | |
520 | is slightly different from the previous C<Configure -Dinstallprefix=...>.) | |
521 | See F<INSTALL>. | |
522 | ||
523 | gcc versions 3.x introduced a new warning that caused a lot of noise | |
524 | during Perl compilation: C<gcc -Ialreadyknowndirectory (warning: | |
525 | changing search order)>. This warning has now been avoided by | |
526 | Configure weeding out such directories before the compilation. | |
527 | ||
528 | One can now build subsets of Perl core modules by using the | |
529 | Configure flags C<-Dnoextensions=...> and C<-Donlyextensions=...>, | |
530 | see F<INSTALL>. | |
531 | ||
532 | =head2 Platform-specific enhancements | |
533 | ||
534 | In Cygwin Perl can now be built with threads (C<Configure -Duseithreads>). | |
535 | This works with both Cygwin 1.3.22 and Cygwin 1.5.3. | |
536 | ||
537 | In newer FreeBSD releases Perl 5.8.0 compilation failed because of | |
538 | trying to use F<malloc.h>, which in FreeBSD is just a dummy file, and | |
539 | a fatal error to even try to use. Now F<malloc.h> is not used. | |
540 | ||
541 | Perl is now known to build also in Hitachi HI-UXMPP. | |
542 | ||
543 | Perl is now known to build again in LynxOS. | |
544 | ||
545 | Mac OS X now installs with Perl version number embedded in | |
546 | installation directory names for easier upgrading of user-compiled | |
547 | Perl, and the installation directories in general are more standard. | |
548 | In other words, the default installation no longer breaks the | |
549 | Apple-provided Perl. On the other hand, with C<Configure -Dprefix=/usr> | |
550 | you can now really replace the Apple-supplied Perl (B<please be careful>). | |
551 | ||
552 | Mac OS X now builds Perl statically by default. This change was done | |
553 | mainly for faster startup times. The Apple-provided Perl is still | |
554 | dynamically linked and shared, and you can enable the sharedness for | |
555 | your own Perl builds by C<Configure -Duseshrplib>. | |
556 | ||
557 | Perl has been ported to IBM's OS/400 PASE environment. The best way | |
558 | to build a Perl for PASE is to use an AIX host as a cross-compilation | |
559 | environment. See README.os400. | |
560 | ||
561 | Yet another cross-compilation option has been added: now Perl builds | |
562 | on OpenZaurus, an Linux distribution based on Mandrake + Embedix for | |
563 | the Sharp Zaurus PDA. See the Cross/README file. | |
564 | ||
565 | Tru64 when using gcc 3 drops the optimisation for F<toke.c> to C<-O2> | |
566 | because of gigantic memory use with the default C<-O3>. | |
567 | ||
568 | Tru64 can now build Perl with the newer Berkeley DBs. | |
569 | ||
570 | Building Perl on WinCE has been much enhanced, see F<README.ce> | |
571 | and F<README.perlce>. | |
572 | ||
573 | =head1 Selected Bug Fixes | |
574 | ||
575 | =head2 Closures, eval and lexicals | |
576 | ||
577 | There have been many fixes in the area of anonymous subs, lexicals and | |
578 | closures. Although this means that Perl is now more "correct", it is | |
579 | possible that some existing code will break that happens to rely on | |
580 | the faulty behaviour. In practice this is unlikely unless your code | |
581 | contains a very complex nesting of anonymous subs, evals and lexicals. | |
582 | ||
583 | =head2 Generic fixes | |
584 | ||
585 | If an input filehandle is marked C<:utf8> and Perl sees illegal UTF-8 | |
586 | coming in when doing C<< <FH> >>, if warnings are enabled a warning is | |
587 | immediately given - instead of being silent about it and Perl being | |
588 | unhappy about the broken data later. (The C<:encoding(utf8)> layer | |
589 | also works the same way.) | |
590 | ||
591 | binmode(SOCKET, ":utf8") only worked on the input side, not on the | |
592 | output side of the socket. Now it works both ways. | |
593 | ||
594 | For threaded Perls certain system database functions like getpwent() | |
595 | and getgrent() now grow their result buffer dynamically, instead of | |
596 | failing. This means that at sites with lots of users and groups the | |
597 | functions no longer fail by returning only partial results. | |
598 | ||
599 | Perl 5.8.0 had accidentally broken the capability for users | |
600 | to define their own uppercase<->lowercase Unicode mappings | |
601 | (as advertised by the Camel). This feature has been fixed and | |
602 | is also documented better. | |
603 | ||
604 | In 5.8.0 this | |
605 | ||
606 | $some_unicode .= <FH>; | |
607 | ||
608 | didn't work correctly but instead corrupted the data. This has now | |
609 | been fixed. | |
610 | ||
611 | Tied methods like FETCH etc. may now safely access tied values, i.e. | |
612 | resulting in a recursive call to FETCH etc. Remember to break the | |
613 | recursion, though. | |
614 | ||
615 | At startup Perl blocks the SIGFPE signal away since there isn't much | |
616 | Perl can do about it. Previously this blocking was in effect also for | |
617 | programs executed from within Perl. Now Perl restores the original | |
618 | SIGFPE handling routine, whatever it was, before running external | |
619 | programs. | |
620 | ||
621 | Linenumbers in Perl scripts may now be greater than 65536, or 2**16. | |
622 | (Perl scripts have always been able to be larger than that, it's just | |
623 | that the linenumber for reported errors and warnings have "wrapped | |
624 | around".) While scripts that large usually indicate a need to rethink | |
625 | your code a bit, such Perl scripts do exist, for example as results | |
626 | from generated code. Now linenumbers can go all the way to | |
627 | 4294967296, or 2**32. | |
628 | ||
629 | =head2 Platform-specific fixes | |
630 | ||
631 | Linux | |
632 | ||
633 | =over 4 | |
634 | ||
635 | =item * | |
636 | ||
637 | Setting $0 works again (with certain limitations that | |
638 | Perl cannot do much about: see L<perlvar/$0>) | |
639 | ||
640 | =back | |
641 | ||
642 | HP-UX | |
643 | ||
644 | =over 4 | |
645 | ||
646 | =item * | |
647 | ||
648 | Setting $0 now works. | |
649 | ||
650 | =back | |
651 | ||
652 | VMS | |
653 | ||
654 | =over 4 | |
655 | ||
656 | =item * | |
657 | ||
658 | Configuration now tests for the presence of C<poll()>, and IO::Poll | |
659 | now uses the vendor-supplied function if detected. | |
660 | ||
661 | =item * | |
662 | ||
663 | A rare access violation at Perl start-up could occur if the Perl image was | |
664 | installed with privileges or if there was an identifier with the | |
665 | subsystem attribute set in the process's rightslist. Either of these | |
666 | circumstances triggered tainting code that contained a pointer bug. | |
667 | The faulty pointer arithmetic has been fixed. | |
668 | ||
669 | =item * | |
670 | ||
671 | The length limit on values (not keys) in the %ENV hash has been raised | |
672 | from 255 bytes to 32640 bytes (except when the PERL_ENV_TABLES setting | |
673 | overrides the default use of logical names for %ENV). If it is | |
674 | necessary to access these long values from outside Perl, be aware that | |
675 | they are implemented using search list logical names that store the | |
676 | value in pieces, each 255-byte piece (up to 128 of them) being an | |
677 | element in the search list. When doing a lookup in %ENV from within | |
678 | Perl, the elements are combined into a single value. The existing | |
679 | VMS-specific ability to access individual elements of a search list | |
680 | logical name via the $ENV{'foo;N'} syntax (where N is the search list | |
681 | index) is unimpaired. | |
682 | ||
683 | =item * | |
684 | ||
685 | The piping implementation now uses local rather than global DCL | |
686 | symbols for inter-process communication. | |
687 | ||
688 | =item * | |
689 | ||
690 | File::Find could become confused when navigating to a relative | |
691 | directory whose name collided with a logical name. This problem has | |
692 | been corrected by adding directory syntax to relative path names, thus | |
693 | preventing logical name translation. | |
694 | ||
695 | =back | |
696 | ||
697 | Win32 | |
698 | ||
699 | =over 4 | |
700 | ||
701 | =item * | |
702 | ||
703 | A memory leak in the fork() emulation has been fixed. | |
704 | ||
705 | =item * | |
706 | ||
707 | The return value of the ioctl() built-in function was accidentally | |
708 | broken in 5.8.0. This has been corrected. | |
709 | ||
710 | =item * | |
711 | ||
712 | The internal message loop executed by perl during blocking operations | |
713 | sometimes interfered with messages that were external to Perl. | |
714 | This often resulted in blocking operations terminating prematurely or | |
715 | returning incorrect results, when Perl was executing under environments | |
716 | that could generate Windows messages. This has been corrected. | |
717 | ||
718 | =item * | |
719 | ||
720 | Pipes and sockets are now automatically in binary mode. | |
721 | ||
722 | =item * | |
723 | ||
724 | The four-argument form of select() did not preserve $! (errno) properly | |
725 | when there were errors in the underlying call. This is now fixed. | |
726 | ||
727 | =item * | |
728 | ||
729 | The "CR CR LF" problem of has been fixed, binmode(FH, ":crlf") | |
730 | is now effectively a no-op. | |
731 | ||
732 | =back | |
733 | ||
734 | =head1 New or Changed Diagnostics | |
735 | ||
736 | All the warnings related to pack() and unpack() were made more | |
737 | informative and consistent. | |
738 | ||
739 | =head2 Changed "A thread exited while %d threads were running" | |
740 | ||
741 | The old version | |
742 | ||
743 | A thread exited while %d other threads were still running | |
744 | ||
745 | was misleading because the "other" included also the thread giving | |
746 | the warning. | |
747 | ||
748 | =head2 Removed "Attempt to clear a restricted hash" | |
749 | ||
750 | It is not illegal to clear a restricted hash, so the warning | |
751 | was removed. | |
752 | ||
753 | =head2 New "Illegal declaration of anonymous subroutine" | |
754 | ||
755 | You must specify the block of code for C<sub>. | |
756 | ||
757 | =head2 Changed "Invalid range "%s" in transliteration operator" | |
758 | ||
759 | The old version | |
760 | ||
761 | Invalid [] range "%s" in transliteration operator | |
762 | ||
763 | was simply wrong because there are no "[] ranges" in tr///. | |
764 | ||
765 | =head2 New "Missing control char name in \c" | |
766 | ||
767 | Self-explanatory. | |
768 | ||
769 | =head2 New "Newline in left-justified string for %s" | |
770 | ||
771 | The padding spaces would appear after the newline, which is | |
772 | probably not what you had in mind. | |
773 | ||
774 | =head2 New "Possible precedence problem on bitwise %c operator" | |
775 | ||
776 | If you think this | |
777 | ||
778 | $x & $y == 0 | |
779 | ||
780 | tests whether the bitwise AND of $x and $y is zero, | |
781 | you will like this warning. | |
782 | ||
783 | =head2 New "Pseudo-hashes are deprecated" | |
784 | ||
785 | This warning should have been already in 5.8.0, since they are. | |
786 | ||
787 | =head2 New "read() on %s filehandle %s" | |
788 | ||
789 | You cannot read() (or sysread()) from a closed or unopened filehandle. | |
790 | ||
791 | =head2 New "5.005 threads are deprecated" | |
792 | ||
793 | This warning should have been already in 5.8.0, since they are. | |
794 | ||
795 | =head2 New "Tied variable freed while still in use" | |
796 | ||
797 | Something pulled the plug on a live tied variable, Perl plays | |
798 | safe by bailing out. | |
799 | ||
800 | =head2 New "To%s: illegal mapping '%s'" | |
801 | ||
802 | An illegal user-defined Unicode casemapping was specified. | |
803 | ||
804 | =head2 New "Use of freed value in iteration" | |
805 | ||
806 | Something modified the values being iterated over. This is not good. | |
807 | ||
808 | =head1 Changed Internals | |
809 | ||
810 | These news matter to you only if you either write XS code or like to | |
811 | know about or hack Perl internals (using Devel::Peek or any of the | |
812 | C<B::> modules counts), or like to run Perl with the C<-D> option. | |
813 | ||
814 | The embedding examples of L<perlembed> have been reviewed to be | |
815 | uptodate and consistent: for example, the correct use of | |
816 | PERL_SYS_INIT3() and PERL_SYS_TERM(). | |
817 | ||
818 | Extensive reworking of the pad code (the code responsible | |
819 | for lexical variables) has been conducted by Dave Mitchell. | |
820 | ||
821 | Extensive work on the v-strings by John Peacock. | |
822 | ||
823 | UTF-8 length and position cache: to speed up the handling of Unicode | |
824 | (UTF-8) scalars, a cache was introduced. Potential problems exist if | |
825 | an extension bypasses the official APIs and directly modifies the PV | |
826 | of an SV: the UTF-8 cache does not get cleared as it should. | |
827 | ||
828 | APIs obsoleted in Perl 5.8.0, like sv_2pv, sv_catpvn, sv_catsv, | |
829 | sv_setsv, are again available. | |
830 | ||
831 | Certain Perl core C APIs like cxinc and regatom are no longer | |
832 | available at all to code outside the Perl core of the Perl core | |
833 | extensions. This is intentional. They never should have been | |
834 | available with the shorter names, and if you application depends on | |
835 | them, you should (be ashamed and) contact perl5-porters to discuss | |
836 | what are the proper APIs. | |
837 | ||
838 | Certain Perl core C APIs like C<Perl_list> are no longer available | |
839 | without their C<Perl_> prefix. If your XS module stops working | |
840 | because some functions cannot be found, in many cases a simple fix is | |
841 | to add the C<Perl_> prefix to the function and the thread context | |
842 | C<aTHX_> as the first argument of the function call. This is also how | |
843 | it should always have been done: letting the Perl_-less forms to leak | |
844 | from the core was an accident. For cleaner embedding you can also | |
845 | force this for all APIs by defining at compile time the cpp define | |
846 | PERL_NO_SHORT_NAMES. | |
847 | ||
848 | Perl_save_bool() has been added. | |
849 | ||
850 | Regexp objects (those created with C<qr>) now have S-magic rather than | |
851 | R-magic. This fixed regexps of the form /...(??{...;$x})/ to no | |
852 | longer ignore changes made to $x. The S-magic avoids dropping | |
853 | the caching optimization and making (??{...}) constructs obscenely | |
854 | slow (and consequently useless). See also L<perlguts/"Magic Variables">. | |
855 | Regexp::Copy was affected by this change. | |
856 | ||
857 | The Perl internal debugging macros DEBUG() and DEB() have been renamed | |
858 | to PERL_DEBUG() and PERL_DEB() to avoid namespace conflicts. | |
859 | ||
860 | C<-DL> removed (the leaktest had been broken and unsupported for years, | |
861 | use alternative debugging mallocs or tools like valgrind and Purify). | |
862 | ||
863 | Verbose modifier C<v> added for C<-DXv> and C<-Dsv>, see L<perlrun>. | |
864 | ||
865 | =head1 New Tests | |
866 | ||
867 | In Perl 5.8.0 there were about 69000 separate tests in about 700 test files, | |
868 | in Perl 5.8.1 there are about 77000 separate tests in about 780 test files. | |
869 | The exact numbers depend on the Perl configuration and on the operating | |
870 | system platform. | |
871 | ||
872 | =head1 Known Problems | |
873 | ||
874 | The hash randomisation mentioned in L</Incompatible Changes> is definitely | |
875 | problematic: it will wake dormant bugs and shake out bad assumptions. | |
876 | ||
877 | If you want to use mod_perl 2.x with Perl 5.8.1, you will need | |
878 | mod_perl-1.99_10 or higher. Earlier versions of mod_perl 2.x | |
879 | do not work with the randomised hashes. (mod_perl 1.x works fine.) | |
880 | You will also need Apache::Test 1.04 or higher. | |
881 | ||
882 | Many of the rarer platforms that worked 100% or pretty close to it | |
883 | with perl 5.8.0 have been left a little bit untended since their | |
884 | maintainers have been otherwise busy lately, and therefore there will | |
885 | be more failures on those platforms. Such platforms include Mac OS | |
886 | Classic, IBM z/OS (and other EBCDIC platforms), and NetWare. The most | |
887 | common Perl platforms (Unix and Unix-like, Microsoft platforms, and | |
888 | VMS) have large enough testing and expert population that they are | |
889 | doing well. | |
890 | ||
891 | =head2 Tied hashes in scalar context | |
892 | ||
893 | Tied hashes do not currently return anything useful in scalar context, | |
894 | for example when used as boolean tests: | |
895 | ||
896 | if (%tied_hash) { ... } | |
897 | ||
898 | The current nonsensical behaviour is always to return false, | |
899 | regardless of whether the hash is empty or has elements. | |
900 | ||
901 | The root cause is that there is no interface for the implementors of | |
902 | tied hashes to implement the behaviour of a hash in scalar context. | |
903 | ||
904 | =head2 Net::Ping 450_service and 510_ping_udp failures | |
905 | ||
906 | The subtests 9 and 18 of lib/Net/Ping/t/450_service.t, and the | |
907 | subtest 2 of lib/Net/Ping/t/510_ping_udp.t might fail if you have | |
908 | an unusual networking setup. For example in the latter case the | |
909 | test is trying to send a UDP ping to the IP address 127.0.0.1. | |
910 | ||
911 | =head2 B::C | |
912 | ||
913 | The C-generating compiler backend B::C (the frontend being | |
914 | C<perlcc -c>) is even more broken than it used to be because of | |
915 | the extensive lexical variable changes. (The good news is that | |
916 | B::Bytecode and ByteLoader are better than they used to be.) | |
917 | ||
918 | =head1 Platform Specific Problems | |
919 | ||
920 | =head2 EBCDIC Platforms | |
921 | ||
922 | IBM z/OS and other EBCDIC platforms continue to be problematic | |
923 | regarding Unicode support. Many Unicode tests are skipped when | |
924 | they really should be fixed. | |
925 | ||
926 | =head2 Cygwin 1.5 problems | |
927 | ||
928 | In Cygwin 1.5 the F<io/tell> and F<op/sysio> tests have failures for | |
929 | some yet unknown reason. In 1.5.5 the threads tests stress_cv, | |
930 | stress_re, and stress_string are failing unless the environment | |
931 | variable PERLIO is set to "perlio" (which makes also the io/tell | |
932 | failure go away). | |
933 | ||
934 | Perl 5.8.1 does build and work well with Cygwin 1.3: with (uname -a) | |
935 | C<CYGWIN_NT-5.0 ... 1.3.22(0.78/3/2) 2003-03-18 09:20 i686 ...> | |
936 | a 100% "make test" was achieved with C<Configure -des -Duseithreads>. | |
937 | ||
938 | =head2 HP-UX: HP cc warnings about sendfile and sendpath | |
939 | ||
940 | With certain HP C compiler releases (e.g. B.11.11.02) you will | |
941 | get many warnings like this (lines wrapped for easier reading): | |
942 | ||
943 | cc: "/usr/include/sys/socket.h", line 504: warning 562: | |
944 | Redeclaration of "sendfile" with a different storage class specifier: | |
945 | "sendfile" will have internal linkage. | |
946 | cc: "/usr/include/sys/socket.h", line 505: warning 562: | |
947 | Redeclaration of "sendpath" with a different storage class specifier: | |
948 | "sendpath" will have internal linkage. | |
949 | ||
950 | The warnings show up both during the build of Perl and during certain | |
951 | lib/ExtUtils tests that invoke the C compiler. The warning, however, | |
952 | is not serious and can be ignored. | |
953 | ||
954 | =head2 IRIX: t/uni/tr_7jis.t falsely failing | |
955 | ||
956 | The test t/uni/tr_7jis.t is known to report failure under 'make test' | |
957 | or the test harness with certain releases of IRIX (at least IRIX 6.5 | |
958 | and MIPSpro Compilers Version 7.3.1.1m), but if run manually the test | |
959 | fully passes. | |
960 | ||
961 | =head2 Mac OS X: no usemymalloc | |
962 | ||
963 | The Perl malloc (C<-Dusemymalloc>) does not work at all in Mac OS X. | |
964 | This is not that serious, though, since the native malloc works just | |
965 | fine. | |
966 | ||
967 | =head2 Tru64: No threaded builds with GNU cc (gcc) | |
968 | ||
969 | In the latest Tru64 releases (e.g. v5.1B or later) gcc cannot be used | |
970 | to compile a threaded Perl (-Duseithreads) because the system | |
971 | C<< <pthread.h> >> file doesn't know about gcc. | |
972 | ||
973 | =head2 Win32: sysopen, sysread, syswrite | |
974 | ||
975 | As of the 5.8.0 release, sysopen()/sysread()/syswrite() do not behave | |
976 | like they used to in 5.6.1 and earlier with respect to "text" mode. | |
977 | These built-ins now always operate in "binary" mode (even if sysopen() | |
978 | was passed the O_TEXT flag, or if binmode() was used on the file | |
979 | handle). Note that this issue should only make a difference for disk | |
980 | files, as sockets and pipes have always been in "binary" mode in the | |
981 | Windows port. As this behavior is currently considered a bug, | |
982 | compatible behavior may be re-introduced in a future release. Until | |
983 | then, the use of sysopen(), sysread() and syswrite() is not supported | |
984 | for "text" mode operations. | |
985 | ||
986 | =head1 Future Directions | |
987 | ||
988 | The following things B<might> happen in future. The first publicly | |
989 | available releases having these characteristics will be the developer | |
990 | releases Perl 5.9.x, culminating in the Perl 5.10.0 release. These | |
991 | are our best guesses at the moment: we reserve the right to rethink. | |
992 | ||
993 | =over 4 | |
994 | ||
995 | =item * | |
996 | ||
997 | PerlIO will become The Default. Currently (in Perl 5.8.x) the stdio | |
998 | library is still used if Perl thinks it can use certain tricks to | |
999 | make stdio go B<really> fast. For future releases our goal is to | |
1000 | make PerlIO go even faster. | |
1001 | ||
1002 | =item * | |
1003 | ||
1004 | A new feature called I<assertions> will be available. This means that | |
1005 | one can have code called assertions sprinkled in the code: usually | |
1006 | they are optimised away, but they can be enabled with the C<-A> option. | |
1007 | ||
1008 | =item * | |
1009 | ||
1010 | A new operator C<//> (defined-or) will be available. This means that | |
1011 | one will be able to say | |
1012 | ||
1013 | $a // $b | |
1014 | ||
1015 | instead of | |
1016 | ||
1017 | defined $a ? $a : $b | |
1018 | ||
1019 | and | |
1020 | ||
1021 | $c //= $d; | |
1022 | ||
1023 | instead of | |
1024 | ||
1025 | $c = $d unless defined $c; | |
1026 | ||
1027 | The operator will have the same precedence and associativity as C<||>. | |
1028 | A source code patch against the Perl 5.8.1 sources will be available | |
1029 | in CPAN as F<authors/id/H/HM/HMBRAND/dor-5.8.1.diff>. | |
1030 | ||
1031 | =item * | |
1032 | ||
1033 | C<unpack()> will default to unpacking the C<$_>. | |
1034 | ||
1035 | =item * | |
1036 | ||
1037 | Various Copy-On-Write techniques will be investigated in hopes | |
1038 | of speeding up Perl. | |
1039 | ||
1040 | =item * | |
1041 | ||
1042 | CPANPLUS, Inline, and Module::Build will become core modules. | |
1043 | ||
1044 | =item * | |
1045 | ||
1046 | The ability to write true lexically scoped pragmas will be introduced. | |
1047 | ||
1048 | =item * | |
1049 | ||
1050 | Work will continue on the bytecompiler and byteloader. | |
1051 | ||
1052 | =item * | |
1053 | ||
1054 | v-strings as they currently exist are scheduled to be deprecated. The | |
1055 | v-less form (1.2.3) will become a "version object" when used with C<use>, | |
1056 | C<require>, and C<$VERSION>. $^V will also be a "version object" so the | |
1057 | printf("%vd",...) construct will no longer be needed. The v-ful version | |
1058 | (v1.2.3) will become obsolete. The equivalence of strings and v-strings (e.g. | |
1059 | that currently 5.8.0 is equal to "\5\8\0") will go away. B<There may be no | |
1060 | deprecation warning for v-strings>, though: it is quite hard to detect when | |
1061 | v-strings are being used safely, and when they are not. | |
1062 | ||
1063 | =item * | |
1064 | ||
1065 | 5.005 Threads Will Be Removed | |
1066 | ||
1067 | =item * | |
1068 | ||
1069 | The C<$*> Variable Will Be Removed | |
1070 | (it was deprecated a long time ago) | |
1071 | ||
1072 | =item * | |
1073 | ||
1074 | Pseudohashes Will Be Removed | |
1075 | ||
1076 | =back | |
1077 | ||
1078 | =head1 Reporting Bugs | |
1079 | ||
1080 | If you find what you think is a bug, you might check the articles | |
1081 | recently posted to the comp.lang.perl.misc newsgroup and the perl | |
1082 | bug database at http://bugs.perl.org/ . There may also be | |
1083 | information at http://www.perl.com/ , the Perl Home Page. | |
1084 | ||
1085 | If you believe you have an unreported bug, please run the B<perlbug> | |
1086 | program included with your release. Be sure to trim your bug down | |
1087 | to a tiny but sufficient test case. Your bug report, along with the | |
1088 | output of C<perl -V>, will be sent off to perlbug@perl.org to be | |
1089 | analysed by the Perl porting team. You can browse and search | |
1090 | the Perl 5 bugs at http://bugs.perl.org/ | |
1091 | ||
1092 | =head1 SEE ALSO | |
1093 | ||
1094 | The F<Changes> file for exhaustive details on what changed. | |
1095 | ||
1096 | The F<INSTALL> file for how to build Perl. | |
1097 | ||
1098 | The F<README> file for general stuff. | |
1099 | ||
1100 | The F<Artistic> and F<Copying> files for copyright information. | |
1101 | ||
1102 | =cut |