Commit | Line | Data |
d468ca04 |
1 | =head1 NAME |
2 | |
3 | perl570delta - what's new for perl v5.7.0 |
4 | |
5 | =head1 DESCRIPTION |
6 | |
7 | This document describes differences between the 5.6.0 release and |
8 | the 5.7.0 release. |
9 | |
10 | =head1 Security Vulnerability Closed |
11 | |
12 | A potential security vulnerability in the optional suidperl component |
13 | of Perl has been identified. suidperl is neither built nor installed |
14 | by default. As of September the 2nd, 2000, the only known vulnerable |
15 | platform is Linux, most likely all Linux distributions. CERT and |
16 | various vendors have been alerted about the vulnerability. |
17 | |
18 | The problem was caused by Perl trying to report a suspected security |
19 | exploit attempt using an external program, /bin/mail. On Linux |
20 | platforms the /bin/mail program had an undocumented feature which |
21 | when combined with suidperl gave access to a root shell, resulting in |
22 | a serious compromise instead of reporting the exploit attempt. If you |
23 | don't have /bin/mail, or if you have 'safe setuid scripts', or if |
24 | suidperl is not installed, you are safe. |
25 | |
26 | The exploit attempt reporting feature has been completely removed from |
27 | the Perl 5.7.0 release, so that particular vulnerability isn't there |
28 | anymore. However, further security vulnerabilities are, |
29 | unfortunately, always possible. The suidperl code is being reviewed |
30 | and if deemed too risky to continue to be supported, it may be |
31 | completely removed from future releases. In any case, suidperl should |
32 | only be used by security experts who know exactly what they are doing |
33 | and why they are using suidperl instead of some other solution such as |
1577cd80 |
34 | sudo ( see http://www.courtesan.com/sudo/ ). |
d468ca04 |
35 | |
36 | =head1 Incompatible Changes |
37 | |
38 | =over 4 |
39 | |
40 | =item * |
41 | |
42 | Arrays now always interpolate into double-quoted strings: |
43 | constructs like "foo@bar" now always assume C<@bar> is an array, |
44 | whether or not the compiler has seen use of C<@bar>. |
45 | |
46 | =item * |
47 | |
48 | The semantics of bless(REF, REF) were unclear and until someone proves |
49 | it to make some sense, it is forbidden. |
50 | |
51 | =item * |
52 | |
53 | A reference to a reference now stringify as "REF(0x81485ec)" instead |
54 | of "SCALAR(0x81485ec)" in order to be more consistent with the return |
55 | value of ref(). |
56 | |
57 | =item * |
58 | |
59 | The very dusty examples in the eg/ directory have been removed. |
60 | Suggestions for new shiny examples welcome but the main issue is that |
61 | the examples need to be documented, tested and (most importantly) |
62 | maintained. |
63 | |
64 | =item * |
65 | |
66 | The obsolete chat2 library that should never have been allowed |
67 | to escape the laboratory has been decommissioned. |
68 | |
69 | =item * |
70 | |
71 | The unimplemented POSIX regex features [[.cc.]] and [[=c=]] are still |
72 | recognised but now cause fatal errors. The previous behaviour of |
73 | ignoring them by default and warning if requested was unacceptable |
74 | since it, in a way, falsely promised that the features could be used. |
75 | |
76 | =item * |
77 | |
78 | The (bogus) escape sequences \8 and \9 now give an optional warning |
79 | ("Unrecognized escape passed through"). There is no need to \-escape |
80 | any C<\w> character. |
81 | |
82 | =item * |
83 | |
84 | lstat(FILEHANDLE) now gives a warning because the operation makes no sense. |
85 | In future releases this may become a fatal error. |
86 | |
87 | =item * |
88 | |
89 | The long deprecated uppercase aliases for the string comparison |
90 | operators (EQ, NE, LT, LE, GE, GT) have now been removed. |
91 | |
92 | =item * |
93 | |
94 | The regular expression captured submatches ($1, $2, ...) are now |
95 | more consistently unset if the match fails, instead of leaving false |
96 | data lying around in them. |
97 | |
98 | =item * |
99 | |
100 | The tr///C and tr///U features have been removed and will not return; |
101 | the interface was a mistake. Sorry about that. For similar |
102 | functionality, see pack('U0', ...) and pack('C0', ...). |
103 | |
104 | =back |
105 | |
106 | =head1 Core Enhancements |
107 | |
108 | =over 4 |
109 | |
110 | =item * |
111 | |
112 | C<perl -d:Module=arg,arg,arg> now works (previously one couldn't pass |
113 | in multiple arguments.) |
114 | |
115 | =item * |
116 | |
758e67a8 |
117 | my __PACKAGE__ $obj now works. |
d468ca04 |
118 | |
119 | =item * |
120 | |
121 | C<no Module;> now works even if there is no "sub unimport" in the Module. |
122 | |
123 | =item * |
124 | |
125 | The numerical comparison operators return C<undef> if either operand |
126 | is a NaN. Previously the behaviour was unspecified. |
127 | |
128 | =item * |
129 | |
1e54db1a |
130 | C<pack('U0a*', ...)> can now be used to force a string to UTF-8. |
d468ca04 |
131 | |
132 | =item * |
133 | |
134 | prototype(\&) is now available. |
135 | |
136 | =item * |
137 | |
138 | There is now an UNTIE method. |
139 | |
140 | =back |
141 | |
142 | =head1 Modules and Pragmata |
143 | |
144 | =head2 New Modules |
145 | |
146 | =over 4 |
147 | |
148 | =item * |
149 | |
150 | File::Temp allows one to create temporary files and directories in an |
151 | easy, portable, and secure way. |
152 | |
153 | =item * |
154 | |
155 | Storable gives persistence to Perl data structures by allowing the |
156 | storage and retrieval of Perl data to and from files in a fast and |
157 | compact binary format. |
158 | |
159 | =back |
160 | |
161 | =head2 Updated And Improved Modules and Pragmata |
162 | |
163 | =over 4 |
164 | |
165 | =item * |
166 | |
167 | The following independently supported modules have been updated to |
168 | newer versions from CPAN: CGI, CPAN, DB_File, File::Spec, Getopt::Long, |
169 | the podlators bundle, Pod::LaTeX, Pod::Parser, Term::ANSIColor, Test. |
170 | |
171 | =item * |
172 | |
173 | Bug fixes and minor enhancements have been applied to B::Deparse, |
174 | Data::Dumper, IO::Poll, IO::Socket::INET, Math::BigFloat, |
175 | Math::Complex, Math::Trig, Net::protoent, the re pragma, SelfLoader, |
176 | Sys::SysLog, Test::Harness, Text::Wrap, UNIVERSAL, and the warnings |
177 | pragma. |
178 | |
179 | =item * |
180 | |
181 | The attributes::reftype() now works on tied arguments. |
182 | |
183 | =item * |
184 | |
185 | AutoLoader can now be disabled with C<no AutoLoader;>, |
186 | |
187 | =item * |
188 | |
189 | The English module can now be used without the infamous performance |
190 | hit by saying |
191 | |
192 | use English '-no_performance_hit'; |
193 | |
194 | (Assuming, of course, that one doesn't need the troublesome variables |
195 | C<$`>, C<$&>, or C<$'>.) Also, introduced C<@LAST_MATCH_START> and |
196 | C<@LAST_MATCH_END> English aliases for C<@-> and C<@+>. |
197 | |
198 | =item * |
199 | |
200 | File::Find now has pre- and post-processing callbacks. It also |
201 | correctly changes directories when chasing symbolic links. Callbacks |
202 | (naughtily) exiting with "next;" instead of "return;" now work. |
203 | |
204 | =item * |
205 | |
206 | File::Glob::glob() renamed to File::Glob::bsd_glob() to avoid |
207 | prototype mismatch with CORE::glob(). |
208 | |
209 | =item * |
210 | |
211 | IPC::Open3 now allows the use of numeric file descriptors. |
212 | |
213 | =item * |
214 | |
215 | use lib now works identically to @INC. Removing directories |
216 | with 'no lib' now works. |
217 | |
218 | =item * |
219 | |
220 | C<%INC> now localised in a Safe compartment so that use/require work. |
221 | |
222 | =item * |
223 | |
224 | The Shell module now has an OO interface. |
225 | |
226 | =back |
227 | |
228 | =head1 Utility Changes |
229 | |
230 | =over 4 |
231 | |
232 | =item * |
233 | |
234 | The Emacs perl mode (emacs/cperl-mode.el) has been updated to version |
235 | 4.31. |
236 | |
237 | =item * |
238 | |
239 | Perlbug is now much more robust. It also sends the bug report to |
240 | perl.org, not perl.com. |
241 | |
242 | =item * |
243 | |
244 | The perlcc utility has been rewritten and its user interface (that is, |
245 | command line) is much more like that of the UNIX C compiler, cc. |
246 | |
247 | =item * |
248 | |
249 | The xsubpp utility for extension writers now understands POD |
250 | documentation embedded in the *.xs files. |
251 | |
252 | =back |
253 | |
254 | =head1 New Documentation |
255 | |
256 | =over 4 |
257 | |
258 | =item * |
259 | |
260 | perl56delta details the changes between the 5.005 release and the |
261 | 5.6.0 release. |
262 | |
263 | =item * |
264 | |
265 | perldebtut is a Perl debugging tutorial. |
266 | |
267 | =item * |
268 | |
269 | perlebcdic contains considerations for running Perl on EBCDIC platforms. |
270 | Note that unfortunately EBCDIC platforms that used to supported back in |
271 | Perl 5.005 are still unsupported by Perl 5.7.0; the plan, however, is to |
272 | bring them back to the fold. |
273 | |
274 | =item * |
275 | |
276 | perlnewmod tells about writing and submitting a new module. |
277 | |
278 | =item * |
279 | |
280 | perlposix-bc explains using Perl on the POSIX-BC platform |
281 | (an EBCDIC mainframe platform). |
282 | |
283 | =item * |
284 | |
285 | perlretut is a regular expression tutorial. |
286 | |
287 | =item * |
288 | |
289 | perlrequick is a regular expressions quick-start guide. |
290 | Yes, much quicker than perlretut. |
291 | |
292 | =item * |
293 | |
294 | perlutil explains the command line utilities packaged with the Perl |
295 | distribution. |
296 | |
297 | =back |
298 | |
299 | =head1 Performance Enhancements |
300 | |
301 | =over 4 |
302 | |
303 | =item * |
304 | |
305 | map() that changes the size of the list should now work faster. |
306 | |
307 | =item * |
308 | |
309 | sort() has been changed to use mergesort internally as opposed to the |
310 | earlier quicksort. For very small lists this may result in slightly |
311 | slower sorting times, but in general the speedup should be at least |
312 | 20%. Additional bonuses are that the worst case behaviour of sort() |
313 | is now better (in computer science terms it now runs in time O(N log N), |
314 | as opposed to quicksort's Theta(N**2) worst-case run time behaviour), |
315 | and that sort() is now stable (meaning that elements with identical |
316 | keys will stay ordered as they were before the sort). |
317 | |
318 | =back |
319 | |
320 | =head1 Installation and Configuration Improvements |
321 | |
322 | =head2 Generic Improvements |
323 | |
324 | =over 4 |
325 | |
326 | =item * |
327 | |
328 | INSTALL now explains how you can configure Perl to use 64-bit |
329 | integers even on non-64-bit platforms. |
330 | |
331 | =item * |
332 | |
333 | Policy.sh policy change: if you are reusing a Policy.sh file |
334 | (see INSTALL) and you use Configure -Dprefix=/foo/bar and in the old |
335 | Policy $prefix eq $siteprefix and $prefix eq $vendorprefix, all of |
336 | them will now be changed to the new prefix, /foo/bar. (Previously |
337 | only $prefix changed.) If you do not like this new behaviour, |
338 | specify prefix, siteprefix, and vendorprefix explicitly. |
339 | |
340 | =item * |
341 | |
342 | A new optional location for Perl libraries, otherlibdirs, is available. |
343 | It can be used for example for vendor add-ons without disturbing Perl's |
344 | own library directories. |
345 | |
346 | =item * |
347 | |
348 | In many platforms the vendor-supplied 'cc' is too stripped-down to |
349 | build Perl (basically, 'cc' doesn't do ANSI C). If this seems |
350 | to be the case and 'cc' does not seem to be the GNU C compiler |
351 | 'gcc', an automatic attempt is made to find and use 'gcc' instead. |
352 | |
353 | =item * |
354 | |
355 | gcc needs to closely track the operating system release to avoid |
356 | build problems. If Configure finds that gcc was built for a different |
357 | operating system release than is running, it now gives a clearly visible |
358 | warning that there may be trouble ahead. |
359 | |
360 | =item * |
361 | |
362 | If binary compatibility with the 5.005 release is not wanted, Configure |
363 | no longer suggests including the 5.005 modules in @INC. |
364 | |
365 | =item * |
366 | |
367 | Configure C<-S> can now run non-interactively. |
368 | |
369 | =item * |
370 | |
371 | configure.gnu now works with options with whitespace in them. |
372 | |
373 | =item * |
374 | |
375 | installperl now outputs everything to STDERR. |
376 | |
377 | =item * |
378 | |
379 | $Config{byteorder} is now computed dynamically (this is more robust |
380 | with "fat binaries" where an executable image contains binaries for |
381 | more than one binary platform.) |
382 | |
383 | =back |
384 | |
385 | =head1 Selected Bug Fixes |
386 | |
387 | =over 4 |
388 | |
389 | =item * |
390 | |
391 | Several debugger fixes: exit code now reflects the script exit code, |
392 | condition C<"0"> now treated correctly, the C<d> command now checks |
393 | line number, the C<$.> no longer gets corrupted, all debugger output now |
394 | goes correctly to the socket if RemotePort is set. |
395 | |
396 | =item * |
397 | |
398 | C<*foo{FORMAT}> now works. |
399 | |
400 | =item * |
401 | |
402 | Lexical warnings now propagating correctly between scopes. |
403 | |
404 | =item * |
405 | |
406 | Line renumbering with eval and C<#line> now works. |
407 | |
408 | =item * |
409 | |
410 | Fixed numerous memory leaks, especially in eval "". |
411 | |
412 | =item * |
413 | |
414 | Modulus of unsigned numbers now works (4063328477 % 65535 used to |
415 | return 27406, instead of 27047). |
416 | |
417 | =item * |
418 | |
419 | Some "not a number" warnings introduced in 5.6.0 eliminated to be |
420 | more compatible with 5.005. Infinity is now recognised as a number. |
421 | |
422 | =item * |
423 | |
424 | our() variables will not cause "will not stay shared" warnings. |
425 | |
426 | =item * |
427 | |
428 | pack "Z" now correctly terminates the string with "\0". |
429 | |
430 | =item * |
431 | |
432 | Fix password routines which in some shadow password platforms |
433 | (e.g. HP-UX) caused getpwent() to return every other entry. |
434 | |
435 | =item * |
436 | |
437 | printf() no longer resets the numeric locale to "C". |
438 | |
439 | =item * |
440 | |
441 | C<q(a\\b)> now parses correctly as C<'a\\b'>. |
442 | |
443 | =item * |
444 | |
445 | Printing quads (64-bit integers) with printf/sprintf now works |
446 | without the q L ll prefixes (assuming you are on a quad-capable platform). |
447 | |
448 | =item * |
449 | |
450 | Regular expressions on references and overloaded scalars now work. |
451 | |
452 | =item * |
453 | |
454 | scalar() now forces scalar context even when used in void context. |
455 | |
456 | =item * |
457 | |
458 | sort() arguments are now compiled in the right wantarray context |
459 | (they were accidentally using the context of the sort() itself). |
460 | |
461 | =item * |
462 | |
463 | Changed the POSIX character class C<[[:space:]]> to include the (very |
464 | rare) vertical tab character. Added a new POSIX-ish character class |
465 | C<[[:blank:]]> which stands for horizontal whitespace (currently, |
466 | the space and the tab). |
467 | |
468 | =item * |
469 | |
470 | $AUTOLOAD, sort(), lock(), and spawning subprocesses |
471 | in multiple threads simultaneously are now thread-safe. |
472 | |
473 | =item * |
474 | |
475 | Allow read-only string on left hand side of non-modifying tr///. |
476 | |
477 | =item * |
478 | |
479 | Several Unicode fixes (but still not perfect). |
480 | |
481 | =over 8 |
482 | |
483 | =item * |
484 | |
485 | BOMs (byte order marks) in the beginning of Perl files |
486 | (scripts, modules) should now be transparently skipped. |
487 | UTF-16 (UCS-2) encoded Perl files should now be read correctly. |
488 | |
489 | =item * |
490 | |
491 | The character tables have been updated to Unicode 3.0.1. |
492 | |
493 | =item * |
494 | |
495 | chr() for values greater than 127 now create utf8 when under use |
496 | utf8. |
497 | |
498 | =item * |
499 | |
500 | Comparing with utf8 data does not magically upgrade non-utf8 data into |
501 | utf8. |
502 | |
503 | =item * |
504 | |
505 | C<IsAlnum>, C<IsAlpha>, and C<IsWord> now match titlecase. |
506 | |
507 | =item * |
508 | |
509 | Concatenation with the C<.> operator or via variable interpolation, |
510 | C<eq>, C<substr>, C<reverse>, C<quotemeta>, the C<x> operator, |
1e54db1a |
511 | substitution with C<s///>, single-quoted UTF-8, should now work--in |
d468ca04 |
512 | theory. |
513 | |
514 | =item * |
515 | |
516 | The C<tr///> operator now works I<slightly> better but is still rather |
517 | broken. Note that the C<tr///CU> functionality has been removed (but |
518 | see pack('U0', ...)). |
519 | |
520 | =item * |
521 | |
522 | vec() now refuses to deal with characters >255. |
523 | |
524 | =item * |
525 | |
526 | Zero entries were missing from the Unicode classes like C<IsDigit>. |
527 | |
528 | =back |
529 | |
530 | =item * |
531 | |
532 | UNIVERSAL::isa no longer caches methods incorrectly. (This broke |
533 | the Tk extension with 5.6.0.) |
534 | |
535 | =back |
536 | |
537 | =head2 Platform Specific Changes and Fixes |
538 | |
539 | =over 4 |
540 | |
541 | =item * |
542 | |
543 | BSDI 4.* |
544 | |
545 | Perl now works on post-4.0 BSD/OSes. |
546 | |
547 | =item * |
548 | |
549 | All BSDs |
550 | |
551 | Setting C<$0> now works (as much as possible; see perlvar for details). |
552 | |
553 | =item * |
554 | |
555 | Cygwin |
556 | |
557 | Numerous updates; currently synchronised with Cygwin 1.1.4. |
558 | |
559 | =item * |
560 | |
561 | EPOC |
562 | |
563 | EPOC update after Perl 5.6.0. See README.epoc. |
564 | |
565 | =item * |
566 | |
567 | FreeBSD 3.* |
568 | |
569 | Perl now works on post-3.0 FreeBSDs. |
570 | |
571 | =item * |
572 | |
573 | HP-UX |
574 | |
575 | README.hpux updated; C<Configure -Duse64bitall> now almost works. |
576 | |
577 | =item * |
578 | |
579 | IRIX |
580 | |
581 | Numerous compilation flag and hint enhancements; accidental mixing |
582 | of 32-bit and 64-bit libraries (a doomed attempt) made much harder. |
583 | |
584 | =item * |
585 | |
586 | Linux |
587 | |
588 | Long doubles should now work (see INSTALL). |
589 | |
590 | =item * |
591 | |
8939ba94 |
592 | Mac OS Classic |
d468ca04 |
593 | |
8939ba94 |
594 | Compilation of the standard Perl distribution in Mac OS Classic should |
595 | now work if you have the Metrowerks development environment and the |
596 | missing Mac-specific toolkit bits. Contact the macperl mailing list |
597 | for details. |
d468ca04 |
598 | |
599 | =item * |
600 | |
601 | MPE/iX |
602 | |
603 | MPE/iX update after Perl 5.6.0. See README.mpeix. |
604 | |
605 | =item * |
606 | |
607 | NetBSD/sparc |
608 | |
609 | Perl now works on NetBSD/sparc. |
610 | |
611 | =item * |
612 | |
613 | OS/2 |
614 | |
615 | Now works with usethreads (see INSTALL). |
616 | |
617 | =item * |
618 | |
619 | Solaris |
620 | |
621 | 64-bitness using the Sun Workshop compiler now works. |
622 | |
623 | =item * |
624 | |
625 | Tru64 (aka Digital UNIX, aka DEC OSF/1) |
626 | |
627 | The operating system version letter now recorded in $Config{osvers}. |
628 | Allow compiling with gcc (previously explicitly forbidden). Compiling |
629 | with gcc still not recommended because buggy code results, even with |
630 | gcc 2.95.2. |
631 | |
632 | =item * |
633 | |
634 | Unicos |
635 | |
636 | Fixed various alignment problems that lead into core dumps either |
637 | during build or later; no longer dies on math errors at runtime; |
638 | now using full quad integers (64 bits), previously was using |
639 | only 46 bit integers for speed. |
640 | |
641 | =item * |
642 | |
643 | VMS |
644 | |
645 | chdir() now works better despite a CRT bug; now works with MULTIPLICITY |
646 | (see INSTALL); now works with Perl's malloc. |
647 | |
648 | =item * |
649 | |
650 | Windows |
651 | |
652 | =over 8 |
653 | |
654 | =item * |
655 | |
656 | accept() no longer leaks memory. |
657 | |
658 | =item * |
659 | |
660 | Better chdir() return value for a non-existent directory. |
661 | |
662 | =item * |
663 | |
664 | New %ENV entries now propagate to subprocesses. |
665 | |
666 | =item * |
667 | |
668 | $ENV{LIB} now used to search for libs under Visual C. |
669 | |
670 | =item * |
671 | |
672 | A failed (pseudo)fork now returns undef and sets errno to EAGAIN. |
673 | |
674 | =item * |
675 | |
676 | Allow REG_EXPAND_SZ keys in the registry. |
677 | |
678 | =item * |
679 | |
680 | Can now send() from all threads, not just the first one. |
681 | |
682 | =item * |
683 | |
684 | Fake signal handling reenabled, bugs and all. |
685 | |
686 | =item * |
687 | |
688 | Less stack reserved per thread so that more threads can run |
689 | concurrently. (Still 16M per thread.) |
690 | |
691 | =item * |
692 | |
ad0f383a |
693 | C<< File::Spec->tmpdir() >> now prefers C:/temp over /tmp |
d468ca04 |
694 | (works better when perl is running as service). |
695 | |
696 | =item * |
697 | |
698 | Better UNC path handling under ithreads. |
699 | |
700 | =item * |
701 | |
702 | wait() and waitpid() now work much better. |
703 | |
704 | =item * |
705 | |
706 | winsock handle leak fixed. |
707 | |
708 | =back |
709 | |
710 | =back |
711 | |
712 | =head1 New or Changed Diagnostics |
713 | |
714 | All regular expression compilation error messages are now hopefully |
715 | easier to understand both because the error message now comes before |
716 | the failed regex and because the point of failure is now clearly |
717 | marked. |
718 | |
719 | The various "opened only for", "on closed", "never opened" warnings |
720 | drop the C<main::> prefix for filehandles in the C<main> package, |
721 | for example C<STDIN> instead of <main::STDIN>. |
722 | |
723 | The "Unrecognized escape" warning has been extended to include C<\8>, |
724 | C<\9>, and C<\_>. There is no need to escape any of the C<\w> characters. |
725 | |
726 | =head1 Changed Internals |
727 | |
728 | =over 4 |
729 | |
730 | =item * |
731 | |
732 | perlapi.pod (a companion to perlguts) now attempts to document the |
733 | internal API. |
734 | |
735 | =item * |
736 | |
737 | You can now build a really minimal perl called microperl. |
738 | Building microperl does not require even running Configure; |
739 | C<make -f Makefile.micro> should be enough. Beware: microperl makes |
740 | many assumptions, some of which may be too bold; the resulting |
741 | executable may crash or otherwise misbehave in wondrous ways. |
742 | For careful hackers only. |
743 | |
744 | =item * |
745 | |
746 | Added rsignal(), whichsig(), do_join() to the publicised API. |
747 | |
748 | =item * |
749 | |
750 | Made possible to propagate customised exceptions via croak()ing. |
751 | |
752 | =item * |
753 | |
754 | Added is_utf8_char(), is_utf8_string(), bytes_to_utf8(), and utf8_to_bytes(). |
755 | |
756 | =item * |
757 | |
758 | Now xsubs can have attributes just like subs. |
759 | |
760 | =back |
761 | |
762 | =head1 Known Problems |
763 | |
764 | =head2 Unicode Support Still Far From Perfect |
765 | |
766 | We're working on it. Stay tuned. |
767 | |
768 | =head2 EBCDIC Still A Lost Platform |
769 | |
770 | The plan is to bring them back. |
771 | |
772 | =head2 Building Extensions Can Fail Because Of Largefiles |
773 | |
774 | Certain extensions like mod_perl and BSD::Resource are known to have |
775 | issues with `largefiles', a change brought by Perl 5.6.0 in which file |
776 | offsets default to 64 bits wide, where supported. Modules may fail to |
777 | compile at all or compile and work incorrectly. Currently there is no |
778 | good solution for the problem, but Configure now provides appropriate |
779 | non-largefile ccflags, ldflags, libswanted, and libs in the %Config |
780 | hash (e.g., $Config{ccflags_nolargefiles}) so the extensions that are |
781 | having problems can try configuring themselves without the |
782 | largefileness. This is admittedly not a clean solution, and the |
783 | solution may not even work at all. One potential failure is whether |
784 | one can (or, if one can, whether it's a good idea) link together at |
785 | all binaries with different ideas about file offsets, all this is |
786 | platform-dependent. |
787 | |
788 | =head2 ftmp-security tests warn 'system possibly insecure' |
789 | |
790 | Don't panic. Read INSTALL 'make test' section instead. |
791 | |
792 | =head2 Test lib/posix Subtest 9 Fails In LP64-Configured HP-UX |
793 | |
794 | If perl is configured with -Duse64bitall, the successful result of the |
795 | subtest 10 of lib/posix may arrive before the successful result of the |
796 | subtest 9, which confuses the test harness so much that it thinks the |
797 | subtest 9 failed. |
798 | |
799 | =head2 Long Doubles Still Don't Work In Solaris |
800 | |
801 | The experimental long double support is still very much so in Solaris. |
802 | (Other platforms like Linux and Tru64 are beginning to solidify in |
803 | this area.) |
804 | |
805 | =head2 Linux With Sfio Fails op/misc Test 48 |
806 | |
807 | No known fix. |
808 | |
809 | =head2 Storable tests fail in some platforms |
810 | |
811 | If any Storable tests fail the use of Storable is not advisable. |
812 | |
813 | =over 4 |
814 | |
815 | =item * |
816 | |
817 | Many Storable tests fail on AIX configured with 64 bit integers. |
818 | |
819 | So far unidentified problems break Storable in AIX if Perl is |
820 | configured to use 64 bit integers. AIX in 32-bit mode works and |
821 | other 64-bit platforms work with Storable. |
822 | |
823 | =item * |
824 | |
825 | DOS DJGPP may hang when testing Storable. |
826 | |
827 | =item * |
828 | |
829 | st-06compat fails in UNICOS and UNICOS/mk. |
830 | |
831 | This means that you cannot read old (pre-Storable-0.7) Storable images |
832 | made in other platforms. |
833 | |
834 | =item * |
835 | |
836 | st-store.t and st-retrieve may fail with Compaq C 6.2 on OpenVMS Alpha 7.2. |
837 | |
838 | =back |
839 | |
840 | =head2 Threads Are Still Experimental |
841 | |
842 | Multithreading is still an experimental feature. Some platforms |
843 | emit the following message for lib/thr5005 |
844 | |
845 | # |
846 | # This is a KNOWN FAILURE, and one of the reasons why threading |
847 | # is still an experimental feature. It is here to stop people |
848 | # from deploying threads in production. ;-) |
849 | # |
850 | |
851 | and another known thread-related warning is |
852 | |
853 | pragma/overload......Unbalanced saves: 3 more saves than restores |
854 | panic: magic_mutexfree during global destruction. |
855 | ok |
856 | lib/selfloader.......Unbalanced saves: 3 more saves than restores |
857 | panic: magic_mutexfree during global destruction. |
858 | ok |
859 | lib/st-dclone........Unbalanced saves: 3 more saves than restores |
860 | panic: magic_mutexfree during global destruction. |
861 | ok |
862 | |
863 | =head2 The Compiler Suite Is Still Experimental |
864 | |
865 | The compiler suite is slowly getting better but is nowhere near |
866 | working order yet. The backend part that has seen perhaps the most |
867 | progress is the bytecode compiler. |
868 | |
869 | =head1 Reporting Bugs |
870 | |
871 | If you find what you think is a bug, you might check the articles |
872 | recently posted to the comp.lang.perl.misc newsgroup and the perl |
f224927c |
873 | bug database at http://bugs.perl.org/ There may also be |
874 | information at http://www.perl.com/perl/ , the Perl Home Page. |
d468ca04 |
875 | |
876 | If you believe you have an unreported bug, please run the B<perlbug> |
877 | program included with your release. Be sure to trim your bug down |
878 | to a tiny but sufficient test case. Your bug report, along with the |
879 | output of C<perl -V>, will be sent off to perlbug@perl.org to be |
880 | analysed by the Perl porting team. |
881 | |
882 | =head1 SEE ALSO |
883 | |
884 | The F<Changes> file for exhaustive details on what changed. |
885 | |
886 | The F<INSTALL> file for how to build Perl. |
887 | |
888 | The F<README> file for general stuff. |
889 | |
890 | The F<Artistic> and F<Copying> files for copyright information. |
891 | |
892 | =head1 HISTORY |
893 | |
894 | Written by Jarkko Hietaniemi <F<jhi@iki.fi>>, with many contributions |
895 | from The Perl Porters and Perl Users submitting feedback and patches. |
896 | |
897 | Send omissions or corrections to <F<perlbug@perl.org>>. |
898 | |
899 | =cut |