This is a live mirror of the Perl 5 development currently hosted at
Added warning about fixed /x regexps -- # now always starts a comment
[perl5.git] / Changes
748a9306 1-------------
2Version 5.002
5The main enhancement to the Perl core was the addition of prototypes.
6Many of the modules that come with Perl have been extensively upgraded.
8Other than that, nearly all the changes for 5.002 were bug fixes of one
9variety or another, so here's the bug list, along with the "resolution"
10for each of them. If you wish to correspond about any of them, please
11include the bug number (if any).
13Added APPLLIB_EXP for embedded perl library support.
14Files patched: perl.c
16Couldn't define autoloaded routine by assignment to typeglob.
17Files patched: pp_hot.c sv.c
a5f75d66 19NETaa13525: Tiny patch to fix installman -n
4633a7c4 20From: Larry Wall
21Files patched: installman
23NETaa13525: de-documented \v
24Files patched: pod/perlop.pod pod/perlre.pod
26NETaa13525: doc changes
27Files patched: pod/perlop.pod pod/perltrap.pod
29NETaa13525: perlxs update from Dean Roehrich
30Files patched: pod/perlxs.pod
4633a7c4 31
32NETaa13525: rename powerunix to powerux
33Files patched: MANIFEST hints/
35NETaa13540: VMS uses CLK_TCK for HZ
36Files patched: pp_sys.c
38NETaa13721: pad_findlex core dumps on bad CvOUTSIDE()
39From: Carl Witty
40Files patched: op.c sv.c toke.c
41 Each CV has a reference to the CV containing it lexically. Unfortunately,
42 it didn't reference-count this reference, so when the outer CV was freed,
43 we ended up with a pointer to memory that got reused later as some other kind
44 of SV.
46NETaa13721: warning suppression
47Files patched: toke.c
48 (same)
50NETaa13722: walk.c had inconsistent static declarations
51From: Tim Bunce
52Files patched: x2p/walk.c
53 Consolidated the various declarations and made them consistent with
54 the actual definitions.
56NETaa13724: -MPackage=args patch
57From: Tim Bunce
58Files patched: perl.c pod/perlrun.pod
59 Added in the -MPackage=args patch too.
61NETaa13729: order-of-evaluation dependency in scope.c on leaving REGCONTEXT
62From: "Jason Shirk"
63Files patched: scope.c
64 Did
66 I32 delta = SSPOPINT;
67 savestack_ix -= delta; /* regexp must have croaked */
69 instead.
71NETaa13731: couldn't assign external lexical array to itself
73Files patched: op.c
74 The pad_findmy routine was only checking previous statements for previous
75 mention of external lexicals, so the fact that the current statement
76 already mentioned @list was not noted. It therefore allocated another
77 reference to the outside lexical, and this didn't compare equal when
78 the assigment parsing code was trying to determine whether there was a
79 common variable on either side of the equals. Since it didn't see the
80 same variable, it thought it could avoid making copies of the values on
81 the stack during list assignment. Unfortunately, before using those
82 values, the list assignment has to zero out the target array, which
83 destroys the values.
85 The fix was to make pad_findmy search the current statement as well. This
86 was actually a holdover from some old code that was trying to delay
87 introduction of "my" variables until the next statement. This is now
88 done with a different mechanism, so the fix should not adversely affect
89 that.
91NETaa13733: s/// doesn't free old string when using copy mode
92From: Larry Wall
93Files patched: pp_ctl.c pp_hot.c
94 When I removed the use of sv_replace(), I simply forgot to free the old char*.
96NETaa13736: closures leaked memory
97From: Carl Witty
98Files patched: op.c pp.c
99 This is a specific example of a more general bug, fixed as NETaa13760, having
100 to do with reference counts on comppads.
102NETaa13739: XSUB interface caches gimme in case XSUB clobbers it
103From: Dean Roehrich
104Files patched: pp_hot.c
105 Applied suggest patch. Also deleted second gimme declaration as redundant.
107NETaa13760: comppad reference counts were inconsistent
108From: Larry Wall
109Files patched: op.c perl.c pp_ctl.c toke.c
110 All official references to comppads are supposed to be through compcv now,
111 but the transformation was not complete, resulting in memory leakage.
113NETaa13761: sv_2pv() wrongly preferred IV to NV when SV was readonly
114From: "Jack R. Lawler"
115Files patched: sv.c
116 Okay, I understand how this one happened. This is a case where a
117 beneficial fix uncovered a bug elsewhere. I changed the constant
118 folder to prefer integer results over double if the numbers are the
119 same. In this case, they aren't, but it leaves the integer value there
120 anyway because the storage is already allocated for it, and it *might*
121 be used in an integer context. And since it's producing a constant, it
122 sets READONLY. Unfortunately, sv_2pv() bogusly preferred the integer
123 value to the double when READONLY was set. This never showed up if you
124 just said
126 print 1.4142135623731;
128 because in that case, there was already a string value.
131NETaa13772: shmwrite core dumps consistently
132From: Gabe Schaffer
133Files patched: opcode.h
134 The shmwrite operator is a list operator but neglected to push a stack
135 mark beforehand, because an 'm' was missing from
137NETaa13773: $. was misdocumented as read-only.
138From: Inaba Hiroto
139Files patched: pod/perlvar.pod
140 <1.array-element-read-only>
141 % perl -le '$,=", "; $#w=5; for (@w) { $_=1; } print @w'
142 Modification of a read-only value attempted at -e line 1.
143 % perl4 -le '$,=", "; $#w=5; for (@w) { $_=1; } print @w'
144 1, 1, 1, 1, 1, 1
146 This one may stay the way it is for performance reasons.
148 <2.begin-local-RS>
149 % cat abc
150 a
151 b
152 c
153 % perl -e 'BEGIN { local $/ = ""; } print "$.:$_" while <>;' abc
154 1:a
155 b
156 c
157 % perl -e '{ local $/ = ""; } print "$.:$_" while <>;' abc
158 1:a
159 2:b
160 3:c
162 $/ wasn't initialized early enough, so local set it back to permanently
163 undefined on exit from the block.
165 <3.grep-x0-bug>
166 % perl -le 'print grep(/^-/ ? ($x=$_) x 0 : 1, "a", "-b", "c");'
167 a
169 % perl4 -le 'print grep(/^-/ ? ($x=$_) x 0 : 1, "a", "-b", "c");'
170 ac
172 An extra mark was left on the stack if (('x') x $repeat) was used in a scalar
173 context.
175 <4.input-lineno-assign>
176 # perl -w does not complain about assignment to $. (Is this just a feature?)
177 # perlvar.pod says "This variable should be considered read-only."
178 % cat abc
179 a
180 b
181 c
182 % perl -wnle '$. = 10 if $. == 2; print "$.:$_"' abc
183 1:a
184 10:b
185 11:c
187 Fixed doc.
189 <5.local-soft-ref.bug>
190 % perl -e 'local ${"a"}=1;'
191 zsh: 529 segmentation fault perl -e 'local ${"a"}=1;'
193 Now says
194 Can't localize a reference at -e line 1.
196 <6.package-readline>
197 % perl -e 'package foo; sub foo { 1; } package main; $_ = foo::foo(); print'
198 1
199 % perl -e '
200 package readline; sub foo { 1; } package main; $_ = readline::foo(); print'
201 Undefined subroutine &main::foo called at -e line 1.
202 % perl -e '
203 package readline; sub foo { 1; } package main; $_ = &readline::foo(); print'
204 1
206 Now treats foo::bar correctly even if foo is a keyword.
208 <>
209 % cat page-head
210 #From: (Russell Fulton)
211 #Newsgroups: comp.lang.perl
212 #Subject: This script causes Perl 5.00 to sementation fault
213 #Date: 15 Nov 1994 00:11:37 GMT
214 #Message-ID: <3a8ubp$>
216 select((select(STDOUT), $^='')[0]); #this is the critical line
217 $a = 'a';
218 write ;
219 exit;
221 format STDOUT =
222 @<<<<<<
223 $a
224 .
226 % perl page-head
227 zsh: 1799 segmentation fault perl /tmp/page-head
229 Now says
230 Undefined top format "main::" called at ./try line 11.
232 <8.sub-as-index>
233 # parser bug?
234 % perl -le 'sub foo {0}; $x[0]=0;$x[foo]<=0'
235 Unterminated <> operator at -e line 1.
236 % perl -le 'sub foo {0}; $x[0]=0;$x[foo()]<=0'
238 A right square bracket now forces expectation of an operator.
240 <9.unary-minus-to-regexp-var>
241 % cat minus-reg
242 #From: Michael Cook <>
243 #Newsgroups: comp.lang.perl
244 #Subject: bug: print -$1
245 #Date: 01 Feb 1995 15:31:25 GMT
246 #Message-ID: <>
248 $_ = "123";
249 /\d+/;
250 print $&, "\n";
251 print -$&, "\n";
252 print 0-$&, "\n";
254 % perl minus-reg
255 123
256 123
257 -123
259 Apparently already fixed in my copy.
261 <10.vec-segv>
262 % cat vec-bug
263 ## Offset values are changed for my machine.
265 #From: (Conrad Augustin)
266 #Subject: perl5 vec() bug?
267 #Message-ID: <>
268 #Date: Tue, 22 Nov 1994 19:37:28 GMT
270 #The following two statements each produce a segmentation fault in perl5:
272 #vec($a, 21406, 32) = 1; # seg fault
273 vec($a, 42813, 16) = 1; # seg fault
275 #When the offset values are one less, all's well:
276 #vec($a, 21405, 32) = 1; # ok
277 #vec($a, 42812, 16) = 1; # ok
279 #Interestingly, this is ok for all high values of N:
280 #$N=1000000; vec($a, $N, 8) = 1;
282 % perl vec-bug
283 zsh: 1806 segmentation fault perl vec-bug
285 Can't reproduce this one.
288NETaa13773: $/ not correctly localized in BEGIN
289Files patched: perl.c
290 (same)
292NETaa13773: foo::bar was misparsed if foo was a reserved word
293Files patched: toke.c toke.c
294 (same)
296NETaa13773: right square bracket didn't force expectation of operator
297Files patched: toke.c
298 (same)
300NETaa13773: scalar ((x) x $repeat) left stack mark
301Files patched: op.c
302 (same)
304NETaa13778: -w coredumps on <$>
305From: Hans Mulder
306Files patched: pp_hot.c toke.c
307 Now produces suggested error message. Also installed guard in warning code
308 that coredumped.
310NETaa13779: foreach didn't use savestack mechanism
311From: Hans Mulder
312Files patched: cop.h pp_ctl.c
313 The foreach mechanism saved the old scalar value on the context stack
314 rather than the savestack. It could consequently get out of sync if
315 unexpectedly unwound.
317NETaa13785: GIMME sometimes used wrong context frame
318From: Greg Earle
319Files patched: embed.h global.sym op.h pp_ctl.c proto.h
320 The expression inside the return was taking its context from the immediately
321 surrounding block rather than the innermost surrounding subroutine call.
323NETaa13797: could modify sv_undef through auto-vivification
324From: Ilya Zakharevich
325Files patched: pp.c
326 Inserted the missing check for readonly values on auto-vivification.
328NETaa13798: if (...) {print} treats print as quoted
329From: Larry Wall
330Files patched: toke.c
331 The trailing paren of the condition was setting expectations to XOPERATOR
332 rather than XBLOCK, so it was being treated like ${print}.
334NETaa13926: commonality was not detected in assignments using COND_EXPR
335From: Mark Hanson
336Files patched: opcode.h
337 The assignment compiler didn't check the 2nd and 3rd args of a ?:
338 for commonality. It still doesn't, but I made ?: into a "dangerous"
339 operator so it is forced to treat it as common.
341NETaa13957: was marking the PUSHMARK as modifiable rather than the arg
342From: David Couture
343Files patched: op.c sv.c
344 It was marking the PUSHMARK as modifiable rather than the arg.
346NETaa13962: documentation of behavior of scalar <*> was unclear
347From: Tom Christiansen
348Files patched: pod/perlop.pod
349 Added the following to perlop:
351 A glob only evaluates its (embedded) argument when it is starting a new
352 list. All values must be read before it will start over. In a list
353 context this isn't important, because you automatically get them all
354 anyway. In a scalar context, however, the operator returns the next value
355 each time it is called, or a FALSE value if you've just run out. Again,
356 FALSE is returned only once. So if you're expecting a single value from
357 a glob, it is much better to say
359 ($file) = <blurch*>;
361 than
363 $file = <blurch*>;
365 because the latter will alternate between returning a filename and
366 returning FALSE.
369NETaa13986: split ignored /m pattern modifier
370From: Winfried Koenig
371Files patched: pp.c
372 Fixed to work like m// and s///.
374NETaa13992: regexp comments not seen after + in non-extended regexp
375From: Mark Knutsen
376Files patched: regcomp.c
377 The code to skip regexp comments was guarded by a conditional that only
378 let it work when /x was in effect.
380NETaa14014: use subs should not count as definition, only as declaration
381From: Keith Thompson
382Files patched: sv.c
383 On *foo = \&bar, doesn't set GVf_IMPORTED if foo and bar are in same package.
385NETaa14021: sv_inc and sv_dec "upgraded" magical SV to non-magical
386From: Paul A Sand
387Also: Andreas Koenig
388Files patched: sv.c
389 The sv_inc() and sv_dec() routines "upgraded" null magical SVs to non-magical.
391NETaa14086: require should check tainting
392From: Karl Simon Berg
393Files patched: pp_ctl.c
394 Since we shouldn't allow tainted requires anyway, it now says:
396 Insecure dependency in require while running with -T switch at line 1.
398NETaa14104: negation fails on magical variables like $1
399From: tim
400Files patched: pp.c
401 Negation was failing on magical values like $1. It was testing the wrong
402 bits and also failed to provide a final "else" if none of the bits matched.
404NETaa14107: deep sort return leaked contexts
405From: Quentin Fennessy
406Files patched: pp_ctl.c
407 Needed to call dounwind() appropriately.
409NETaa14129: attempt to localize via a reference core dumps
410From: Michele Sardo
411Files patched: op.c pod/perldiag.pod
412 Now produces an error "Can't localize a reference", with explanation in
413 perldiag.
415NETaa14138: substr() and s/// can cause core dump
416From: Andrew Vignaux
417Files patched: pp_hot.c
418 Forgot to call SvOOK_off() on the SV before freeing its string.
420NETaa14145: ${@INC}[0] dumped core in debugger
421From: Hans Mulder
422Files patched: sv.c
423 Now croaks "Bizarre copy of ARRAY in block exit", which is better than
424 a core dump. The fact that ${@INC}[0] means $INC[0] outside the debugger
425 is a different bug.
427NETaa14147: bitwise assignment ops wipe out byte of target string
428From: Jim Richardson
429Files patched: doop.c
430 The code was assuming that the target was not either of the two operands,
431 which is false for an assignment operator.
433NETaa14153: lexing of lexicals in patterns fooled by character class
434From: Dave Bianchi
435Files patched: toke.c
436 It never called the dwimmer, which is how it fooled it.
438NETaa14154: allowed autoloaded methods by recognizing sub method; declaration
439From: Larry Wall
440Files patched: gv.c
441 Made sub method declaration sufficient for autoloader to stop searching on.
443NETaa14156: shouldn't optimize block scope on tainting
444From: Pete Peterson
445Files patched: op.c toke.c
446 I totally disabled the block scope optimization when running tainted.
448NETaa14157: -T and -B only allowed 1/30 "odd" characters--changed to 1/3
449From: Tor Lillqvist
450Files patched: pp_sys.c
451 Applied suggested patch.
453NETaa14160: deref of null symbol should produce null list
454From: Jared Rhine
455Files patched: pp_hot.c
456 It didn't check for list context before returning undef.
458NETaa14162: POSIX::gensym now returns a symbol reference
459From: Josh N. Pritikin
460Also: Tim Bunce
461Files patched: ext/POSIX/
462 Applied suggested patch.
464NETaa14164: POSIX autoloader now distinguishes non-constant "constants"
465From: Tim Bunce <>
466Files patched: ext/POSIX/ ext/POSIX/POSIX.xs
467 The .xs file now distinguishes non-constant "constants" by setting EAGAIN.
468 This will also let us use #ifdef within the .xs file to de-constantify
469 any other macros that happen not to be constants even if they don't use
470 an argument.
472NETaa14166: missing semicolon after "my" induces core dump
473From: Thomas Kofler
474Files patched: toke.c
475 The parser was left thinking it was still processing a "my", and flubbed.
476 I made it wipe out the "in_my" variable on a syntax error.
478NETaa14166: missing semicolon after "my" induces core dump"
479Files patched: toke.c
480 (same)
482NETaa14206: can now use English and strict at the same time
483From: Andrew Wilcox
484Files patched: sv.c
485 It now counts imported symbols as okay under "use strict".
487NETaa14206: can now use English and strict at the same time
488Files patched: gv.c pod/perldiag.pod
489 (same)
491NETaa14265: elseif now produces severe warning
492From: Yutao Feng
493Files patched: pod/perldiag.pod toke.c
494 Now complains explicitly about "elseif".
496NETaa14279: list assignment propagated taintedness to independent scalars
497From: Tim Freeman
498Files patched: pp_hot.c
499 List assignment needed to be modified so that tainting didn't propagate
500 between independent scalar values.
502NETaa14312: undef in @EXPORTS core dumps
503From: William Setzer
504Files patched: lib/
505 Now says:
507 Unable to create sub named "t::" at lib/ line 159.
508 Illegal null symbol in @t::EXPORT at -e line 1
509 BEGIN failed--compilation aborted at -e line 1.
512NETaa14312: undef in @EXPORTS core dumps
513Files patched: pod/perldiag.pod sv.c
514 (same)
516NETaa14321: literal @array check shouldn't happen inside embedded expressions
517From: Mark H. Nodine
518Files patched: toke.c
519 The general solution to this is to disable the literal @array check within
520 any embedded expression. For instance, this also failed bogusly:
522 print "$foo{@foo}";
524 The reason fixing this also fixes the s///e problem is that the lexer
525 effectively puts the RHS into a do {} block, making the expression
526 embedded within curlies, as far as the error message is concerned.
528NETaa14322: now localizes $! during POSIX::AUTOLOAD
529From: Larry Wall
530Files patched: ext/POSIX/
531 Added local $! = 0.
533NETaa14324: defined() causes spurious sub existence
534From: "Andreas Koenig"
535Files patched: op.c pp.c
536 It called pp_rv2cv which wrongly assumed it could add any sub it referenced.
538NETaa14336: use Module () forces import of nothing
539From: Tim Bunce
540Files patched: op.c
541 use Module () now refrains from calling import at all.
543NETaa14353: added special HE allocator
544From: Larry Wall
545Files patched: global.sym
547NETaa14353: added special HE allocator
548Files patched: hv.c perl.h
550NETaa14353: array extension now converts old memory to SV storage.
551Files patched: av.c av.h sv.c
553NETaa14353: hashes now convert old storage into SV arenas.
554Files patched: global.sym
556NETaa14353: hashes now convert old storage into SV arenas.
557Files patched: hv.c perl.h
559NETaa14353: upgraded SV arena allocation
560Files patched: proto.h
562NETaa14353: upgraded SV arena allocation
563Files patched: perl.c sv.c
565NETaa14422: added rudimentary prototypes
566From: Gisle Aas
567Files patched: Makefile.SH op.c op.c perly.c perly.c.diff perly.h perly.y proto.h sv.c toke.c
568 Message-Id: <>
569 To: (Andy Dougherty)
570 Cc:
571 Subject: Re: Jumbo Configure patch vs. 1m.
572 Date: Thu, 28 Sep 95 17:18:54 -0700
573 From: (Larry Wall)
575 : No. Larry's currently got the patch pumpkin for all such core perl topics.
577 I dunno whether you should let me have the patch pumpkin or not. To fix
578 a Sev 2 I just hacked in rudimentary prototypes. :-)
580 We can now define true unary subroutines, as well as argumentless
581 subroutines:
583 sub baz () { 12; } # Must not have argument
584 sub bar ($) { $_[0] * 7 } # Must have exactly one argument
585 sub foo ($@) { print "@_\n" } # Must have at least one argument
586 foo bar baz / 2 || "oops", "is the answer";
588 This prints "42 is the answer" on my machine. That is, it's the same as
590 foo( bar( baz() / 2) || "oops", "is the answer");
592 Attempting to compile
594 foo;
596 results in
598 Too few arguments for main::foo at ./try line 8, near "foo;"
600 Compiling
602 bar 1,2,3;
604 results in
606 Too many arguments for main::bar at ./try line 8, near "foo;"
608 But
610 @array = ('a','b','c');
611 foo @array, @array;
613 prints "3 a b c" because the $ puts the first arg of foo into scalar context.
615 The main win at this point is that we can say
617 sub AAA () { 1; }
618 sub BBB () { 2; }
620 and the user can say AAA + BBB and get 3.
622 I'm not quite sure how this interacts with autoloading though. I fear
623 will need to say
625 sub E2BIG ();
626 sub EACCES ();
627 sub EAGAIN ();
628 sub EBADF ();
629 sub EBUSY ();
630 ...
631 sub _SC_STREAM_MAX ();
632 sub _SC_TZNAME_MAX ();
633 sub _SC_VERSION ();
635 unless we can figure out how to efficiently declare a default prototype
636 at import time. Meaning, not using eval. Currently
638 *foo = \&bar;
640 (the ordinary import mechanism) implicitly stubs &bar with no prototype if
641 &bar is not yet declared. It's almost like you want an AUTOPROTO to
642 go with your AUTOLOAD.
644 Another thing to rub one's 5 o'clock shadow over is that there's no way
645 to apply a prototype to a method call at compile time.
647 And no, I don't want to have the
649 sub howabout ($formal, @arguments) { ... }
651 argument right now.
653 Larry
655NETaa14422: couldn't take reference of a prototyped function
656Files patched: op.c
657 (same)
659NETaa14423: use didn't allow expressions involving the scratch pad
660From: Graham Barr
661Files patched: op.c perly.c perly.c.diff perly.y proto.h vms/perly_c.vms
662 Applied suggested patch.
664NETaa14444: lexical scalar didn't autovivify
665From: Gurusamy Sarathy
666Files patched: op.c pp_hot.c
667 It didn't have code in pp_padsv to do the right thing.
669NETaa14448: caller could dump core when used within an eval or require
670From: Danny R. Faught
671Files patched: pp_ctl.c
672 caller() was incorrectly assuming the context stack contained a subroutine
673 context when it in fact contained an eval context.
675NETaa14451: improved error message on bad pipe filehandle
676From: Danny R. Faught
677Files patched: pp_sys.c
678 Now says the slightly more informative
680 Can't use an undefined value as filehandle reference at ./try line 3.
682NETaa14462: pp_dbstate had a scope leakage on recursion suppression
683From: Tim Bunce
684Files patched: pp_ctl.c
685 Swapped the code in question around.
687NETaa14482: sv_unref freed ref prematurely at times
688From: Gurusamy Sarathy
689Files patched: sv.c
690 Made sv_unref() mortalize rather than free the old reference.
692NETaa14484: appending string to array produced bizarre results
693From: Greg Ward
694Also: Malcolm Beattie
695Files patched: pp_hot.c
696 Will now say, "Can't coerce ARRAY to string".
698NETaa14525: assignment to globs didn't reset them correctly
699From: Gurusamy Sarathy
700Files patched: sv.c
701 Applied parts of patch not overridden by subsequent patch.
703NETaa14529: a partially matching subpattern could spoof infinity detector
704From: Wayne Berke
705Files patched: regexec.c
706 A partial match on a subpattern could fool the infinite regress detector
707 into thinking progress had been made.
708 The previous workaround prevented another bug (NETaa14529) from being fixed,
709 so I've backed it out. I'll need to think more about how to detect failure
710 to progress. I'm still hopeful it's not equivalent to the halting problem.
712NETaa14535: patches from Gurusamy Sarathy
713From: Gurusamy Sarathy
714Files patched: op.c pp.c pp_hot.c regexec.c sv.c toke.c
715 Applied most recent suggested patches.
717NETaa14537: select() can return too soon
718From: Matt Kimball
719Also: Andreas Gustafsson
720Files patched: pp_sys.c
722NETaa14538: method calls were treated like do {} under loop modifiers
723From: Ilya Zakharevich
724Files patched: perly.c perly.y
725 Needed to take the OPf_SPECIAL flag off of entersubs from method reductions.
726 (It was probably a cut-and-paste error from long ago.)
728NETaa14540: foreach (@array) no longer does extra stack copy
730Files patched: Todo op.c pp_ctl.c pp_hot.c
731 Fixed by doing the foreach(@array) optimization, so it iterates
732 directly through the array, and can detect the implicit shift from
733 referencing <>.
735NETaa14541: new version of perlbug
736From: Kenneth Albanowski
737Files patched: README pod/perl.pod utils/perlbug.PL
738 Brought it up to version 1.09.
740NETaa14541: perlbug 1.11
741Files patched: utils/perlbug.PL
742 (same)
744NETaa14548: magic sets didn't check private OK bits
745From: W. Bradley Rubenstein
746Files patched: mg.c
747 The magic code was getting mixed up between private and public POK bits.
749NETaa14550: made ~ magic magical
750From: Tim Bunce
751Files patched: sv.c
752 Applied suggested patch.
754NETaa14551: humongous header causes infinite loop in format
755From: Grace Lee
756Files patched: pp_sys.c
757 Needed to check for page exhaustion after doing top-of-form.
759NETaa14558: attempt to call undefined top format core dumped
760From: Hallvard B Furuseth
761Files patched: pod/perldiag.pod pp_sys.c
762 Now issues an error on attempts to call a non-existent top format.
764NETaa14561: Gurusamy Sarathy's G_KEEPERR patch
765From: Andreas Koenig
766Also: Gurusamy Sarathy
767Also: Tim Bunce
768Files patched: cop.h interp.sym perl.c perl.h pp_ctl.c pp_sys.c sv.c toke.c
769 Applied latest patch.
771NETaa14581: shouldn't execute BEGIN when there are compilation errors
772From: Rickard Westman
773Files patched: op.c
774 Perl should not try to execute BEGIN and END blocks if there's been a
775 compilation error.
777NETaa14582: got SEGV sorting sparse array
778From: Rick Pluta
779Files patched: pp_ctl.c
780 Now weeds out undefined values much like Perl 4 did.
781 Now sorts undefined values to the front.
783NETaa14582: sort was letting unsortable values through to comparison routine
784Files patched: pp_ctl.c
785 (same)
787NETaa14585: globs in pad space weren't properly cleaned up
788From: Gurusamy Sarathy
789Files patched: op.c pp.c pp_hot.c sv.c
790 Applied suggested patch.
792NETaa14614: now does dbmopen with perl_eval_sv()
793From: The Man
794Files patched: perl.c pp_sys.c proto.h
795 dbmopen now invokes perl_eval_sv(), which should handle error conditions
796 better.
798NETaa14618: exists doesn't work in GDBM_File
799From: Andrew Wilcox
800Files patched: ext/GDBM_File/GDBM_File.xs
801 Applied suggested patch.
803NETaa14619: tied()
804From: Larry Wall
805Also: Paul Marquess
806Files patched: embed.h global.sym keywords.h opcode.h pp_sys.c toke.c
807 Applied suggested patch.
809NETaa14636: Jumbo Dynaloader patch
810From: Tim Bunce
a5f75d66 811Files patched: ext/DynaLoader/ ext/DynaLoader/dl_dld.xs ext/DynaLoader/dl_dlopen.xs ext/DynaLoader/dl_hpux.xs ext/DynaLoader/dl_next.xs ext/DynaLoader/dl_vms.xs ext/DynaLoader/dlutils.c
812 Applied suggested patches.
814NETaa14637: checkcomma routine was stupid about bareword sub calls
815From: Tim Bunce <>
816Files patched: toke.c
817 The checkcomma routine was stupid about bareword sub calls.
819NETaa14639: (?i) didn't reset on runtime patterns
820From: Mark A. Scheel
821Files patched: op.h pp_ctl.c toke.c
822 It didn't distinguish between permanent flags outside the pattern and
823 temporary flags within the pattern.
825NETaa14649: selecting anonymous globs dumps core
826From: Chip Salzenberg
827Files patched: cop.h doio.c embed.h global.sym perl.c pp_sys.c proto.h
828 Applied suggested patch, but reversed the increment and decrement to avoid
829 decrementing and freeing what we're going to increment.
831NETaa14655: $? returned negative value on AIX
832From: Kim Frutiger
833Also: Stephen D. Lee
834Files patched: pp_sys.c
835 Applied suggested patch.
837NETaa14668: {2,} could match once
838From: Hugo van der Sanden
839Files patched: regexec.c
840 When an internal pattern failed a conjecture, it didn't back off on the
841 number of times it thought it had matched.
843NETaa14673: open $undefined dumped core
844From: Samuli K{rkk{inen
845Files patched: pp_sys.c
846 pp_open() didn't check its argument for globness.
848NETaa14683: stringifies were running pad out of space
849From: Robin Barker
850Files patched: op.h toke.c
851 Increased PADOFFSET to a U32, and made lexer not put double-quoted strings
852 inside OP_STRINGIFY unless they really needed it.
854NETaa14689: shouldn't have . in @INC when tainting
855From: William R. Somsky
856Files patched: perl.c
857 Now does not put . into @INC when tainting. It may still be added with a
859 use lib ".";
861 or, to put it at the end,
863 BEGIN { push(@INC, ".") }
865 but this is not recommended unless a chdir to a known location has been done
866 first.
868NETaa14690: values inside tainted SVs were ignored
869From: "James M. Stern"
870Files patched: pp.c pp_ctl.c
871 It was assuming that a tainted value was a string.
873NETaa14692: format name required qualification under use strict
874From: Tom Christiansen
875Files patched: gv.c
876 Now treats format names the same as subroutine names.
878NETaa14695: added simple regexp caching
879From: John Rowe
880Files patched: pp_ctl.c
881 Applied suggested patch.
883NETaa14697: regexp comments were sometimes wrongly treated as literal text
884From: Tom Christiansen
885Files patched: regcomp.c
886 The literal-character grabber didn't know about extended comments.
cd48280b 887 N.B. '#' is treated as a comment character whenever the /x option is
888 used now, so you can't include '#' as a simple literal in /x regexps.
890 (By the way, Tom, the boxed form of quoting in the previous enclosure is
891 exceeding antisocial when you want to extract the code from it.)
893NETaa14704: closure got wrong outer scope if outer sub was predeclared
894From: Marc Paquette
895Files patched: op.c
896 The outer scope of the anonymous sub was set to the stub rather than to
897 the actual subroutine. I kludged it by making the outer scope of the
898 stub be the actual subroutine, if anything is depending on the stub.
900NETaa14705: $foo .= $foo did free memory read
901From: Gerd Knops
902Files patched: sv.c
903 Now modifies address to copy if it was reallocated.
905NETaa14709: Chip's FileHandle stuff
906From: Larry Wall
907Also: Chip Salzenberg
908Files patched: MANIFEST ext/FileHandle/ ext/FileHandle/FileHandle.xs ext/FileHandle/Makefile.PL ext/POSIX/ ext/POSIX/POSIX.pod ext/POSIX/POSIX.xs lib/ lib/ t/lib/filehand.t t/lib/posix.t
909 Applied suggested patches.
911NETaa14711: added (&) and (*) prototypes for blocks and symbols
912From: Kenneth Albanowski
913Files patched: Makefile.SH op.c perly.c perly.h perly.y toke.c
914 & now means that it must have an anonymous sub as that argument. If
915 it's the first argument, the sub may be specified as a block in the
916 indirect object slot, much like grep or sort, which have prototypes of (&@).
918 Also added * so you can do things like
920 sub myopen (*;$);
922 myopen(FOO, $filename);
924NETaa14713: setuid FROM root now defaults to not do tainting
925From: Tony Camas
926Files patched: mg.c perl.c pp_hot.c
927 Applied suggested patch.
929NETaa14714: duplicate magics could be added to an SV
930From: Yary Hluchan
931Files patched: sv.c sv.c
932 The sv_magic() routine didn't properly check to see if it already had a
933 magic of that type. Ordinarily it would have, but it was called during
934 mg_get(), which forces the magic flags off temporarily.
936NETaa14721: sub defined during erroneous do-FILE caused core dump
937From: David Campbell
938Files patched: op.c
939 Fixed the seg fault. I couldn't reproduce the return problem.
941NETaa14734: ref should never return undef
942From: Dale Amon
943Files patched: pp.c t/op/overload.t
944 Now returns null string.
946NETaa14751: slice of undefs now returns null list
947From: Tim Bunce
948Files patched: pp.c pp_hot.c
949 Null list clobberation is now done in lslice, not aassign.
951NETaa14789: select coredumped on Linux
952From: Ulrich Kunitz
953Files patched: pp_sys.c
954 Applied suggested patches, more or less.
956NETaa14789: straightened out ins and out of duping
957Files patched: lib/IPC/
958 (same)
960NETaa14791: implemented internal SUPER class
961From: Nick Ing-Simmons
962Also: Dean Roehrich
963Files patched: gv.c
964 Applied suggested patch.
966NETaa14845: s/// didn't handle offset strings
967From: Ken MacLeod
968Files patched: pp_ctl.c
969 Needed a call to SvOOK_off(targ) in pp_substcont().
971NETaa14851: Use of << to mean <<"" is deprecated
972From: Larry Wall
973Files patched: toke.c
975NETaa14865: added HINT_BLOCK_SCOPE to "elsif"
976From: Jim Avera
977Files patched: perly.y
978 Needed to set HINT_BLOCK_SCOPE on "elsif" to prevent the do block from
979 being optimized away, which caused the statement transition in elsif
980 to reset the stack too far back.
982NETaa14876: couldn't delete localized GV safely
983From: John Hughes
984Files patched: pp.c scope.c
985 The reference count of the "borrowed" GV needed to be incremented while
986 there was a reference to it in the savestack.
988NETaa14887: couldn't negate magical scalars
989From: ian
990Also: Gurusamy Sarathy
991Files patched: pp.c
992 Applied suggested patch, more or less. (It's not necessary to test both
993 SvNIOK and SvNIOKp, since the private bits are always set if the public
994 bits are set.)
996NETaa14893: /m modifier was sticky
997From: Jim Avera
998Files patched: pp_ctl.c
999 pp_match() and pp_subst() were using an improperly scoped SAVEINT to restore
1000 the value of the internal variable multiline.
1002NETaa14893: /m modifier was sticky
1003Files patched: cop.h pp_hot.c
1004 (same)
1006NETaa14916: retained old return value
1007From: Martyn Pearce
1008Files patched: lib/
1009 Applied suggested patch.
1011NETaa14928: non-const 3rd arg to split assigned to list could coredump
1012From: Hans de Graaff
1013Files patched: op.c
1014 The optimizer was assuming the OP was an OP_CONST.
1016NETaa14942: substr as lvalue could disable magic
1017From: Darrell Kindred <>
1018Files patched: pp.c
1019 The substr was disabling the magic of $1.
1021NETaa14990: "not" not parseable when expecting term
1022From: "Randal L. Schwartz"
1023Files patched: perly.c perly.c.diff perly.y vms/perly_c.vms
1024 The NOTOP production needed to be moved down into the terms.
1026NETaa14993: Bizarre copy of formline
1027From: Tom Christiansen
1028Also: Charles Bailey
1029Files patched: sv.c
1030 Applied suggested patch.
1032NETaa14998: sv_add_arena() no longer leaks memory
1033From: Andreas Koenig
1034Files patched: av.c hv.c perl.h sv.c
1035 Now keeps one potential arena "on tap", but doesn't use it unless there's
1036 demand for SV headers. When an AV or HV is extended, its old memory
1037 becomes the next potential arena unless there already is one, in which
1038 case it is simply freed. This will have the desired property of not
1039 stranding medium-sized chunks of memory when extending a single array
1040 repeatedly, but will not degrade when there's no SV demand beyond keeping
1041 one chunk of memory on tap, which generally will be about 250 bytes big,
1042 since it prefers the earlier freed chunk over the later. See the nice_chunk
1043 variable.
1045NETaa14999: $a and $b now protected from use strict and lexical declaration
1046From: Tom Christiansen
1047Files patched: gv.c pod/perldiag.pod toke.c
1048 Bare $a and $b are now allowed during "use strict". In addition,
1049 the following diag was added:
1051 =item Can't use "my %s" in sort comparison
1053 (F) The global variables $a and $b are reserved for sort comparisons.
1054 You mentioned $a or $b in the same line as the <=> or cmp operator,
1055 and the variable had earlier been declared as a lexical variable.
1056 Either qualify the sort variable with the package name, or rename the
1057 lexical variable.
1060NETaa15034: use strict refs should allow calls to prototyped functions
1061From: Roderick Schertler
1062Files patched: perly.c perly.c.diff perly.y toke.c vms/perly_c.vms
1063 Applied patch suggested by Chip.
1065NETaa15083: forced $AUTOLOAD to be untainted
1066From: Tim Bunce
1067Files patched: gv.c pp_hot.c
1068 Stripped any taintmagic from $AUTOLOAD after setting it.
1070NETaa15084: patch for Term::Cap
1071From: Mark Kaehny
1072Also: Hugo van der Sanden
1073Files patched: lib/Term/
1074 Applied suggested patch.
1076NETaa15086: null pattern could cause coredump in s//_$1_/
1077From: "Paul E. Maisano"
1078Files patched: cop.h pp_ctl.c
1079 If the replacement pattern was complicated enough to cause pp_substcont
1080 to be called, then it lost track of which REGEXP* it was supposed to
1081 be using.
1083NETaa15087: t/io/pipe.t didn't work on AIX
1084From: Andy Dougherty
1085Files patched: t/io/pipe.t
1086 Applied suggested patch.
1088NETaa15088: study was busted
1089From: Hugo van der Sanden
1090Files patched: opcode.h pp.c
1091 It was studying its scratch pad target rather than the argument supplied.
1093NETaa15090: MSTATS patch
1094From: Tim Bunce
1095Files patched: global.sym malloc.c perl.c perl.h proto.h
1096 Applied suggested patch.
1098NETaa15098: longjmp out of magic leaks memory
1099From: Chip Salzenberg
1100Files patched: mg.c sv.c
1101 Applied suggested patch.
1103NETaa15102: getpgrp() is broken if getpgrp2() is available
1104From: Roderick Schertler
1105Files patched: perl.h pp_sys.c
1106 Applied suggested patch.
1108NETaa15103: prototypes leaked opcodes
1109From: Chip Salzenberg
1110Files patched: op.c
1111 Applied suggested patch.
1113NETaa15107: quotameta memory bug on all metacharacters
1114From: Chip Salzenberg
1115Files patched: pp.c
1116 Applied suggested patch.
1118NETaa15108: Fix for incomplete string leak
1119From: Chip Salzenberg
1120Files patched: toke.c
1121 Applied suggested patch.
1123NETaa15110: couldn't use $/ with 8th bit set on some architectures
1124From: Chip Salzenberg
1125Files patched: doop.c interp.sym mg.c op.c perl.c perl.h pp_ctl.c pp_hot.c pp_sys.c sv.c toke.c util.c
1126 Applied suggested patches.
1128NETaa15112: { a_1 => 2 } didn't parse as expected
1129From: Stuart M. Weinstein
1130Files patched: toke.c
1131 The little dwimmer was only skipping ALPHA rather than ALNUM chars.
1133NETaa15123: bitwise ops produce spurious warnings
1134From: Hugo van der Sanden
1135Also: Chip Salzenberg
1136Also: Andreas Gustafsson
1137Files patched: sv.c
1138 Decided to suppress the warning in the conversion routines if merely converting
1139 a temporary, which can never be a user-supplied value anyway.
1141NETaa15129: #if defined (foo) misparsed in h2ph
1142From: Roderick Schertler <>
1143Files patched: utils/h2ph.PL
1144 Applied suggested patch.
1146NETaa15131: some POSIX functions assumed valid filehandles
1147From: Chip Salzenberg
1148Files patched: ext/POSIX/POSIX.xs
1149 Applied suggested patch.
1151NETaa15151: don't optimize split on OPpASSIGN_COMMON
1152From: Huw Rogers
1153Files patched: op.c
1154 Had to swap the optimization down to after the assignment op is generated
1155 and COMMON is calculated, and then clean up the resultant tree differently.
1157NETaa15154: MakeMaker-5.18
1158From: Andreas Koenig
1159Files patched: MANIFEST lib/ExtUtils/ lib/ExtUtils/ lib/ExtUtils/ lib/ExtUtils/
1160 Brought it up to 5.18.
1162NETaa15156: some Exporter tweaks
1163From: Roderick Schertler
1164Also: Tim Bunce
1165Files patched: lib/
1166 Also did Tim's Tiny Trivial patch.
1168NETaa15157: new version of Test::Harness
1169From: Andreas Koenig
1170Files patched: lib/Test/
1171 Applied suggested patch.
1173NETaa15175: overloaded nomethod has garbage 4th op
1174From: Ilya Zakharevich
1175Files patched: gv.c
1176 Applied suggested patch.
1178NETaa15179: SvPOK_only shouldn't back off on offset pointer
1180Files patched: sv.h
1181 SvPOK_only() was calling SvOOK_off(), which adjusted the string pointer
1182 after tr/// has already acquired it. It shouldn't really be necessary
1183 for SvPOK_only() to undo an offset string pointer, since there's no
1184 conflict with a possible integer value where the offset is stored.
1186NETaa15193: & now always bypasses prototype checking
1187From: Larry Wall
1188Files patched: dump.c op.c op.h perly.c perly.c.diff perly.y pod/perlsub.pod pp_hot.c proto.h toke.c vms/perly_c.vms vms/perly_h.vms
1189 Turned out to be a big hairy deal because the lexer turns foo() into &foo().
1190 But it works consistently now. Also fixed pod.
1192NETaa15197: 5.002b2 is 'appending' to $@
1193From: Gurusamy Sarathy
1194Files patched: pp_ctl.c
1195 Applied suggested patch.
1197NETaa15201: working around Linux DBL_DIG problems
1198From: Kenneth Albanowski
1199Files patched: hints/ sv.c
1200 Applied suggested patch.
1202NETaa15208: SelectSaver
1203From: Chip Salzenberg
1204Files patched: MANIFEST lib/
1205 Applied suggested patch.
1207NETaa15209: DirHandle
1208From: Chip Salzenberg
1209Files patched: MANIFEST lib/ t/lib/dirhand.t
1211NETaa15210: sysopen()
1212From: Chip Salzenberg
1213Files patched: doio.c lib/ExtUtils/typemap pod/perlfunc.pod pp_hot.c pp_sys.c proto.h toke.c
1214 Applied suggested patch. Hope it works...
1216NETaa15211: use mnemonic names in Safe setup
1217From: Chip Salzenberg
1218Files patched: ext/Safe/
1219 Applied suggested patch, more or less.
1221NETaa15214: prototype()
1222From: Chip Salzenberg
1223Files patched: ext/Safe/ global.sym pp.c toke.c
1224 Applied suggested patch.
1226NETaa15217: -w problem with -d:foo
1227From: Tim Bunce
1228Files patched: perl.c
1229 Applied suggested patch.
1231NETaa15218: *GLOB{ELEMENT}
1232From: Larry Wall
1233Files patched: Makefile.SH embed.h ext/Safe/ keywords.h opcode.h opcode.h perly.c perly.c.diff perly.y pp_hot.c t/lib/safe.t vms/perly_c.vms
1235NETaa15219: Make *x=\*y do like *x=*y
1236From: Chip Salzenberg
1237Files patched: sv.c
1238 Applied suggested patch.
1240NETaa15221: Indigestion with Carp::longmess and big eval '...'s
1241From: Tim Bunce
1242Files patched: lib/
1243 Applied suggested patch.
1245NETaa15222: VERSION patch for standard extensions
1246From: Paul Marquess
1247Files patched: ext/DB_File/Makefile.PL ext/DynaLoader/ ext/DynaLoader/Makefile.PL ext/Fcntl/ ext/Fcntl/Makefile.PL ext/GDBM_File/ ext/GDBM_File/Makefile.PL ext/NDBM_File/Makefile.PL ext/NDBM_File/ ext/ODBM_File/Makefile.PL ext/ODBM_File/ ext/POSIX/Makefile.PL ext/POSIX/ ext/SDBM_File/Makefile.PL ext/SDBM_File/ ext/Safe/Makefile.PL ext/Safe/ ext/Socket/Makefile.PL
1248 Applied suggested patch.
1250NETaa15222: VERSION patch for standard extensions (reprise)
1251Files patched: ext/DB_File/ ext/DynaLoader/ ext/Fcntl/ ext/GDBM_File/ ext/NDBM_File/ ext/ODBM_File/ ext/POSIX/ ext/SDBM_File/ ext/Safe/ ext/Socket/
1252 (same)
1254NETaa15227: $i < 10000 should optimize to integer op
1255From: Larry Wall
1256Files patched: op.c op.c
1257 The program
1259 for ($i = 0; $i < 100000; $i++) {
1260 push @foo, $i;
1261 }
1263 takes about one quarter the memory if the optimizer decides that it can
1264 use an integer < comparison rather than floating point. It now does so
1265 if one side is an integer constant and the other side a simple variable.
1266 This should really help some of our benchmarks. You can still force a
1267 floating point comparison by using 100000.0 instead.
1269NETaa15228: CPerl-mode patch
1270From: Ilya Zakharevich
1271Files patched: emacs/cperl-mode.el
1272 Applied suggested patch.
1274NETaa15231: Symbol::qualify()
1275From: Chip Salzenberg
1276Files patched: ext/FileHandle/ gv.c lib/ lib/ pp_hot.c
1277 Applied suggested patch.
1279NETaa15236: select select broke under use strict
1280From: Chip Salzenberg
1281Files patched: op.c
1282 Instead of inventing a new bit, I just turned off the HINT_STRICT_REFS bit.
1283 I don't think it's worthwhile distinguishing between qualified or unqualified
1284 names to select.
1286NETaa15237: use vars
1287From: Larry Wall
1288Files patched: MANIFEST gv.c lib/ lib/ sv.c
1290NETaa15240: keep op names _and_ descriptions
1291From: Chip Salzenberg
1292Files patched: doio.c embed.h ext/Safe/ ext/Safe/Safe.xs global.sym op.c opcode.h scope.c sv.c
1293 Applied suggested patch.
1295NETaa15259: study doesn't unset on string modification
1296From: Larry Wall
1297Files patched: mg.c pp.c
1298 Piggybacked on m//g unset magic to unset the study too.
1300NETaa15276: pick a better initial cxstack_max
1301From: Chip Salzenberg
1302Files patched: perl.c
1303 Added fudge in, and made it calculate how many it could fit into (most of) 8K,
1304 to avoid getting 16K of Kingsley malloc.
1306NETaa15287: numeric comparison optimization adjustments
1307From: Clark Cooper
1308Files patched: op.c
1309 Applied patch suggested by Chip, with liberalization to >= and <=.
1311NETaa15299: couldn't eval string containing pod or __DATA__
1312From: Andreas Koenig
1313Also: Gisle Aas
1314Files patched: toke.c
1315 Basically, eval didn't know how to bypass pods correctly.
1317NETaa15300: sv_backoff problems
1318From: Paul Marquess
1319Also: mtr
1320Also: Chip Salzenberg
1321Files patched: op.c sv.c sv.h
1322 Applied suggested patch.
1324NETaa15312: Avoid fclose(NULL)
1325From: Chip Salzenberg
1326Files patched: toke.c
1327 Applied suggested patch.
1329NETaa15318: didn't set up perl_init_i18nl14n for export
1330From: Ilya Zakharevich
1331Files patched: perl_exp.SH
1332 Applied suggested patch.
1334NETaa15331: File::Path::rmtree followed symlinks
1335From: Andreas Koenig
1336Files patched: lib/File/
1337 Added suggested patch, except I did
1339 if (not -l $root and -d _) {
1341 for efficiency, since if -d is true, the -l already called lstat on it.
1343NETaa15339: sv_gets() didn't reset count
1345Files patched: sv.c
1346 Applied suggested patch.
1348NETaa15341: differentiated importation of different types
1349From: Chip Salzenberg
1350Files patched: gv.c gv.h op.c perl.c pp.c pp_ctl.c sv.c sv.h toke.c
1351 Applied suggested patch.
1353NETaa15342: Consistent handling of e_{fp,tmpname}
1354From: Chip Salzenberg
1355Files patched: perl.c pp_ctl.c util.c
1356 Applied suggested patch.
1358NETaa15344: Safe gets confused about malloc on AIX
1359From: Tim Bunce
1360Files patched: ext/Safe/Safe.xs
1361 Applied suggested patch.
1363NETaa15348: -M upgrade
1364From: Tim Bunce
1365Files patched: perl.c pod/perlrun.pod
1366 Applied suggested patch.
1368NETaa15369: change in split optimization broke scalar context
1369From: Ulrich Pfeifer
1370Files patched: op.c
1371 The earlier patch to make the split optimization pay attention to
1372 OPpASSIGN_COMMON rearranged how the syntax tree is constructed, but kept
1373 the wrong context flags. This causes pp_split() do do the wrong thing.
1375NETaa15423: can't do subversion numbering because of %5.3f assumptions
1376From: Andy Dougherty
1377Files patched: configpm patchlevel.h perl.c perl.h pp_ctl.c
1378 Removed the %5.3f assumptions where appropriate. patchlevel.h now
1379 defines SUBVERSION, which if greater than 0 indicates a development version.
1381NETaa15424: Sigsetjmp patch
1382From: Kenneth Albanowski
1383Files patched: Configure config_h.SH op.c perl.c perl.h pp_ctl.c util.c
1384 Applied suggested patch.
1386Needed to make install paths absolute.
1387Files patched: installperl
1389h2xs 1.14
1390Files patched: utils/h2xs.PL
1392makedir() looped on a symlink to a directory.
1393Files patched: installperl
1395xsubpp 1.932
1396Files patched: lib/ExtUtils/xsubpp
1399Version 5.001
1402Nearly all the changes for 5.001 were bug fixes of one variety or another,
1403so here's the bug list, along with the "resolution" for each of them. If
1404you wish to correspond about any of them, please include the bug number.
1406There were a few that can be construed as enhancements:
1407 NETaa13059: now warns of use of \1 where $1 is necessary.
1408 NETaa13512: added $SIG{__WARN__} and $SIG{__DIE__} hooks
1409 NETaa13520: added closures
1410 NETaa13530: scalar keys now resets hash iterator
1411 NETaa13641: added Tim's fancy new import whizbangers
1412 NETaa13710: cryptswitch needed to be more "useable"
1413 NETaa13716: Carp now allows multiple packages to be skipped out of
1414 NETaa13716: now counts imported routines as "defined" for redef warnings
1415 (and, of course, much of the stuff from the perl5-porters)
1417NETaa12974: README incorrectly said it was a pre-release.
1418Files patched: README
1420NETaa13033: goto pushed a bogus scope on the context stack.
1421From: Steve Vinoski
1422Files patched: pp_ctl.c
1423 The goto operator pushed an extra bogus scope onto the context stack. (This
1424 often didn't matter, since many things pop extra unrecognized scopes off.)
1426NETaa13034: tried to get valid pointer from undef.
1427From: Castor Fu
1428Also: Achille Hui, the Day Dreamer
1429Also: Eric Arnold
1430Files patched: pp_sys.c
1431 Now treats undef specially, and calls SvPV_force on any non-numeric scalar
1432 value to get a real pointer to somewhere.
1434NETaa13035: included package info with filehandles.
1435From: Jack Shirazi - BIU
1436Files patched: pp_hot.c pp_sys.c
1437 Now passes a glob to filehandle methods to keep the package info intact.
1439NETaa13048: didn't give strict vars message on every occurrence.
1440From: Doug Campbell
1441Files patched: gv.c
1442 It now complains about every occurrence. (The bug resulted from an
1443 ill-conceived attempt to suppress a duplicate error message in a
1444 suboptimal fashion.)
1446NETaa13052: test for numeric sort sub return value fooled by taint magic.
1447From: Peter Jaspers-Fayer
1448Files patched: pp_ctl.c sv.h
1449 The test to see if the sort sub return value was numeric looked at the
1450 public flags rather than the private flags of the SV, so taint magic
1451 hid that info from the sort.
1453NETaa13053: forced a2p to use byacc
1454From: Andy Dougherty
1455Files patched: MANIFEST x2p/Makefile.SH x2p/a2p.c
1456 a2p.c is now pre-byacced and shipped with the kit.
1458NETaa13055: misnamed constant in previous patch.
1459From: Conrad Augustin
1460Files patched: op.c op.h toke.c
1461 The tokener translates $[ to a constant, but with a special marking in case
1462 the constant gets assigned to or localized. Unfortunately, the marking
1463 was done with a combination of OPf_SPECIAL and OPf_MOD that was easily
1464 spoofed. There is now a private OPpCONST_ARYLEN flag for this purpose.
1466NETaa13055: use of OPf_SPECIAL for $[ lvaluehood was too fragile.
1467Files patched: op.c op.h toke.c
1468 (same)
1470NETaa13056: convert needs to throw away any number info on its list.
1471From: Jack Shirazi - BIU
1472Files patched: op.c
1473 The listiness of the argument list leaked out to the subroutine call because
1474 of how prepend_elem and append_elem reuse an existing list. The convert()
1475 routine just needs to discard any listiness it finds on its argument.
1477NETaa13058: AUTOLOAD shouldn't assume size of @_ is meaningful.
1478From: Florent Guillaume
1479Files patched: ext/DB_File/ ext/Fcntl/ ext/GDBM_File/ ext/Socket/ h2xs.SH
1480 I just deleted the optimization, which is silly anyway since the eventual
1481 subroutine definition is cached.
1483NETaa13059: now warns of use of \1 where $1 is necessary.
1484From: Gustaf Neumann
1485Files patched: toke.c
1486 Now says
1488 Can't use \1 to mean $1 in expression at foo line 2
1490 along with an explanation in perldiag.
1492NETaa13060: no longer warns on attempt to read <> operator's transition state.
1493From: Chaim Frenkel
1494Files patched: pp_hot.c
1495 No longer warns on <> operator's transitional state.
1497NETaa13140: warning said $ when @ would be more appropriate.
1498From: David J. MacKenzie
1499Files patched: op.c pod/perldiag.pod
1500 Now says
1502 (Did you mean $ or @ instead of %?)
1504 and added more explanation to perldiag.
1506NETaa13149: was reading freed memory to make incorrect error message.
1507Files patched: pp_ctl.c
1508 It was reading freed memory to make an error message that would be
1509 incorrect in any event because it had the inner filename rather than
1510 the outer.
1512NETaa13149: confess was sometimes less informative than croak
1513From: Jack Shirazi
1514Files patched: lib/
1515 (same)
1517NETaa13150: stderr needs to be STDERR in package
1518From: Jack Shirazi
1519Files patched: lib/File/
1520 Also fixed pl2pm to translate the filehandles to uppercase.
1522NETaa13150: uppercases stdin, stdout and stderr
1523Files patched: pl2pm
1524 (same)
1526NETaa13154: array assignment didn't notice package magic.
1527From: Brian Reichert
1528Files patched: pp_hot.c
1529 The list assignment operator looked for only set magic, but set magic is
1530 only on the elements of a magical hash, not on the hash as a whole. I made
1531 the operator look for any magic at all on the target array or hash.
1533NETaa13155: &DB::DB left trash on the stack.
1534From: Thomas Koenig
1535Files patched: lib/ pp_ctl.c
1536 The call by pp_dbstate() to &DB::DB left trash on the stack. It now
1537 calls DB in list context, and DB returns ().
1539NETaa13156: lexical variables didn't show up in debugger evals.
1540From: Joergen Haegg
1541Files patched: op.c
1542 The code that searched back up the context stack for the lexical scope
1543 outside the eval only partially took into consideration that there
1544 might be extra debugger subroutine frames that shouldn't be used, and
1545 ended up comparing the wrong statement sequence number to the range of
1546 valid sequence numbers for the scope of the lexical variable. (There
1547 was also a bug fixed in passing that caused the scope of lexical to go
1548 clear to the end of the subroutine even if it was within an inner block.)
1550NETaa13157: any request for autoloaded DESTROY should create a null one.
1551From: Tom Christiansen
1552Files patched: lib/
1553 If is not located, it now creates sub DESTROY {} automatically.
1555NETaa13158: now preserves $@ around destructors while leaving eval.
1556From: Tim Bunce
1557Files patched: pp_ctl.c
1558 Applied supplied patch, except the whole second hunk can be replaced with
1560 sv_insert(errsv, 0, 0, message, strlen(message));
1562NETaa13160: clarified behavior of split without arguments
1563From: Harry Edmon
1564Files patched: pod/perlfunc.pod
1565 Clarified the behavior of split without arguments.
1567NETaa13162: eval {} lost list/scalar context
1568From: Dov Grobgeld
1569Files patched: op.c
1570 LEAVETRY didn't propagate number to ENTERTRY.
1572NETaa13163: clarified documentation of foreach using my variable
1573From: Tom Christiansen
1574Files patched: pod/perlsyn.pod
1575 Explained that foreach using a lexical is still localized.
1577NETaa13164: the dot detector for the end of formats was over-rambunctious.
1578From: John Stoffel
1579Files patched: toke.c
1580 The dot detector for the end of formats was over-rambunctious. It would
1581 pick up any dot that didn't have a space in front of it.
1583NETaa13165: do {} while 1 never linked outer block into next chain.
1584From: Gisle Aas
1585Files patched: op.c
1586 When the conditional of do {} while 1; was optimized away, it confused the
1587 postfix order construction so that the block that ordinarily sits around the
1588 whole loop was never executed. So when the loop tried to unstack between
1589 iterations, it got the wrong context, and blew away the lexical variables
1590 of the outer scope. Fixed it by introducing a NULL opcode that will be
1591 optimized away later.
1593NETaa13167: coercion was looking at public bits rather than private bits.
1594From: Randal L. Schwartz
1595Also: Thomas Riechmann
1596Also: Shane Castle
1597Files patched: sv.c
1598 There were some bad ifdefs around the various varieties of set*id(). In
1599 addition, tainting was interacting badly with assignment to $> because
1600 sv_2iv() was examining SvPOK rather than SvPOKp, and so couldn't coerce
1601 a string uid to an integer one.
1603NETaa13167: had some ifdefs wrong on set*id.
1604Files patched: mg.c pp_hot.c
1605 (same)
1607NETaa13168: relaxed test for comparison of new and old fds
1608From: Casper H.S. Dik
1609Files patched: t/lib/posix.t
1610 I relaxed the comparison to just check that the new fd is greater.
1612NETaa13169: autoincrement can corrupt scalar value state.
1613From: Gisle Aas
1614Also: Tom Christiansen
1615Files patched: sv.c
1616 It assumed a PV didn't need to be upgraded to become an NV.
1618NETaa13169: previous patch could leak a string pointer.
1619Files patched: sv.c
1620 (same)
1622NETaa13170: symbols missing from global.sym
1623From: Tim Bunce
1624Files patched: global.sym
1625 Applied suggested patch.
1627NETaa13171: \\ in <<'END' shouldn't reduce to \.
1628From: Randal L. Schwartz
1629Files patched: toke.c
1630 <<'END' needed to bypass ordinary single-quote processing.
1632NETaa13172: 'use integer' turned off magical autoincrement.
1633From: Erich Rickheit KSC
1634Files patched: pp.c pp_hot.c
1635 The integer versions of the increment and decrement operators were trying too
1636 hard to be efficient.
1638NETaa13172: deleted duplicate increment and decrement code
1639Files patched: opcode.h pp.c
1640 (same)
1642NETaa13173: install should make shared libraries executable.
1643From: Brian Grossman
1644Also: Dave Nadler
1645Also: Eero Pajarre
1646Files patched: installperl
1647 Now gives permission 555 to any file ending with extension specified by $dlext.
1649NETaa13176: ck_rvconst didn't free the const it used up.
1650From: Nick Duffek
1651Files patched: op.c
1652 I checked in many random memory leaks under this bug number, since it
1653 was an eval that brought many of them out.
1655NETaa13176: didn't delete XRV for temp ref of destructor.
1656Files patched: sv.c
1657 (same)
1659NETaa13176: didn't delete op_pmshort in matching operators.
1660Files patched: op.c
1661 (same)
1663NETaa13176: eval leaked the name of the eval.
1664Files patched: scope.c
1665 (same)
1667NETaa13176: gp_free didn't free the format.
1668Files patched: gv.c
1669 (same)
1671NETaa13176: minor leaks in loop exits and constant subscript optimization.
1672Files patched: op.c
1673 (same)
1675NETaa13176: plugged some duplicate struct allocation memory leaks.
1676Files patched: perl.c
1677 (same)
1679NETaa13176: sv_clear of an FM didn't clear anything.
1680Files patched: sv.c
1681 (same)
1683NETaa13176: tr/// didn't mortalize its return value.
1684Files patched: pp.c
1685 (same)
1687NETaa13177: SCOPE optimization hid line number info
1688From: David J. MacKenzie
1689Also: Hallvard B Furuseth
1690Files patched: op.c
1691 Every pass on the syntax tree has to keep track of the current statement.
1692 Unfortunately, the single-statement block was optimized into a single
1693 statement between the time the variable was parsed and the time the
1694 void code scan was done, so that pass didn't see the OP_NEXTSTATE
1695 operator, because it has been optimized to an OP_NULL.
1697 Fortunately, null operands remember what they were, so it was pretty easy
1698 to make it set the correct line number anyway.
1700NETaa13178: some linux doesn't handle nm well
1701From: Alan Modra
1702Files patched: hints/
1703 Applied supplied patch.
1705NETaa13180: localized slice now pre-extends array
1706From: Larry Schuler
1707Files patched: pp.c
1708 A localized slice now pre-extends its array to avoid reallocation during
1709 the scope of the local.
1711NETaa13181: m//g didn't keep track of whether previous match matched null.
1712From: "philippe.verdret"
1713Files patched: mg.h pp_hot.c
1714 A pattern isn't allowed to match a null string in the same place twice in
1715 a row. m//g wasn't keeping track of whether the previous match matched
1716 the null string.
1718NETaa13182: now includes whitespace as a regexp metacharacter.
1719From: Larry Wall
1720Files patched: toke.c
1721 scan_const() now counts " \t\n\r\f\v" as metacharacters when scanning a pattern.
1723NETaa13183: sv_setsv shouldn't try to clone an object.
1724From: Peter Gordon
1725Files patched: sv.c
1726 The sv_mortalcopy() done by the return in STORE called sv_setsv(),
1727 which cloned the object. sv_setsv() shouldn't be in the business of
1728 cloning objects.
1730NETaa13184: bogus warning on quoted signal handler name removed.
1731From: Dan Carson
1732Files patched: toke.c
1733 Now doesn't complain unless the first non-whitespace character after the =
1734 is an alphabetic character.
1736NETaa13186: now croaks on chop($')
1737From: Casper H.S. Dik
1738Files patched: doop.c
1739 Now croaks on chop($') and such.
1741NETaa13187: "${foo::bar}" now counts as mere delimitation, not as a bareword.
1742From: Jay Rogers
1743Files patched: toke.c
1744 "${foo::bar}" now counts as mere delimitation, not as a bareword inside a
1745 reference block.
1747NETaa13188: for backward compatibility, looks for "perl -" before "perl".
1748From: Russell Mosemann
1749Files patched: toke.c
1750 Now allows non-whitespace characters on the #! line between the "perl"
1751 and the "-".
1753NETaa13188: now allows non-whitespace after #!...perl before switches.
1754Files patched: toke.c
1755 (same)
1757NETaa13189: derivative files need to be removed before recreation
1758From: Simon Leinen
1759Also: Dick Middleton
1760Also: David J. MacKenzie
1761Files patched: x2p/Makefile.SH
1762 Fixed various little nits as suggested in several messages.
1764NETaa13190: certain assignments can spoof pod directive recognizer
1765From: Ilya Zakharevich
1766Files patched: toke.c
1767 The lexer now only recognizes pod directives where a statement is expected.
1769NETaa13194: now returns undef when there is no curpm.
1770From: lusol@Dillon.CC.Lehigh.EDU
1771Files patched: mg.c
1772 Since there was no regexp prior to the "use", it was returning whatever the
1773 last successful match was within the "use", because there was no current
1774 regexp, so it treated it as a normal variable. It now returns undef.
1776NETaa13195: semop had one S too many.
1777From: Joachim Huober
1778Files patched:
1779 The entry in had one too many S's.
1781NETaa13196: always assumes it's a Perl script if -c is used.
1782From: Dan Carson
1783Files patched: toke.c
1784 It now will assume it's a Perl script if the -c switch is used.
1786NETaa13197: changed implicit -> message to be more understandable.
1787From: Bruce Barnett
1788Files patched: op.c pod/perldiag.pod
1789 I changed the error message to be more understandable. It now says
1791 Can't use subscript on sort...
1794NETaa13201: added OPpCONST_ENTERED flag to properly enter filehandle symbols.
1795From: E. Jay Berkenbilt
1796Also: Tom Christiansen
1797Files patched: op.c op.h toke.c
1798 The grammatical reduction of a print statement didn't properly count
1799 the filehandle as a symbol reference because it couldn't distinguish
1800 between a symbol entered earlier in the program and a symbol entered
1801 for the first time down in the lexer.
1803NETaa13203: README shouldn't mention uperl.o any more.
1804From: Anno Siegel
1805Files patched: README
1807NETaa13204: .= shouldn't warn on uninitialized target.
1808From: Pete Peterson
1809Files patched: pp_hot.c
1810 No longer warns on uninitialized target of .= operator.
1812NETaa13206: handy macros in XSUB.h
1813From: Tim Bunce
1814Files patched: XSUB.h
1815 Added suggested macros.
1817NETaa13228: commonality checker didn't treat lexicals as variables.
1819Files patched: op.c
1820 The list assignment operator tries to avoid unnecessary copies by doing the
1821 assignment directly if there are no common variables on either side of the
1822 equals. Unfortunately, the code that decided that only recognized references
1823 to dynamic variables, not lexical variables.
1825NETaa13229: fixed sign stuff for complement, integer coercion.
1826From: Larry Wall
1827Files patched: perl.h pp.c sv.c
1828 Fixed ~0 and integer coercions.
1830NETaa13230: no longer tries to reuse scratchpad temps if tainting in effect.
1831From: Luca Fini
1832Files patched: op.c
1833 I haven't reproduced it, but I believe the problem is the reuse of scratchpad
1834 temporaries between statements. I've made it not try to reuse them if
1835 tainting is in effect.
1837NETaa13231: *foo = *bar now prevents typo warnings on "foo"
1838From: Robin Barker
1839Files patched: sv.c
1840 Aliasing of the form *foo = *bar is now protected from the typo warnings.
1841 Previously only the *foo = \$bar form was.
1843NETaa13235: require BAREWORD now introduces package name immediately.
1844From: Larry Wall
1845Files patched: toke.c
1846 require BAREWORD now introduces package name immediately. This lets the
1847 method intuit code work right even though the require hasn't actually run
1848 yet.
1850NETaa13289: didn't calculate correctly using arybase.
1851From: Jared Rhine
1852Files patched: pp.c pp_hot.c
1853 The runtime code didn't use curcop->cop_arybase correctly.
1855NETaa13301: store now throws exception on error
1856From: Barry Friedman
1857Files patched: ext/GDBM_File/GDBM_File.xs ext/NDBM_File/NDBM_File.xs ext/ODBM_File/ODBM_File.xs ext/SDBM_File/SDBM_File.xs
1858 Changed warn to croak in ext/*DBM_File/*.xs.
1860NETaa13302: ctime now takes Time_t rather than Time_t*.
1861From: Rodger Anderson
1862Files patched: ext/POSIX/POSIX.xs
1863 Now declares a Time_t and takes the address of that in CODE.
1865NETaa13302: shorter way to do this patch
1866Files patched: ext/POSIX/POSIX.xs
1867 (same)
1869NETaa13304: could feed too large $@ back into croak, whereupon it croaked.
1870From: Larry Wall
1871Files patched: perl.c
1872 callist() could feed $@ back into croak with more than a bare %s. (croak()
1873 handles long strings with a bare %s okay.)
1875NETaa13305: compiler misoptimized RHS to outside of s/a/print/e
1876From: Brian S. Cashman <>
1877Files patched: op.c
1878 The syntax tree was being misconstructed because the compiler felt that
1879 the RHS was invariant, so it did it outside the s///.
1881NETaa13314: assigning mortal to lexical leaks
1882From: Larry Wall
1883Files patched: sv.c
1884 In stealing strings, sv_setsv was checking SvPOK to see if it should free
1885 the destination string. It should have been checking SvPVX.
1887NETaa13316: wait4pid now recalled when errno == EINTR
1888From: Robert J. Pankratz
1889Files patched: pp_sys.c util.c
1890 system() and the close() of a piped open now recall wait4pid if it returned
1891 prematurely with errno == EINTR.
1893NETaa13329: needed to localize taint magic
1894From: Brian Katzung
1895Files patched: sv.c doio.c mg.c pp_hot.c pp_sys.c scope.c taint.c
1896 Taint magic is now localized better, though I had to resort to a kludge
1897 to allow a value to be both tainted and untainted simultaneously during
1898 the assignment of
1900 local $foo = $_[0];
1902 when $_[0] is a reference to the variable $foo already.
1904NETaa13341: clarified interaction of AnyDBM_File::ISA and "use"
1905From: Ian Phillipps
1906Files patched: pod/modpods/AnyDBMFile.pod
1907 The doc was misleading.
1909NETaa13342: grep and map with block would enter block but never leave it.
1910From: Ian Phillipps
1911Files patched: op.c
1912 The compiler use some sort-checking code to handle the arguments of
1913 grep and map. Unfortunately, this wiped out the block exit opcode while
1914 leaving the block entry opcode. This doesn't matter to sort, but did
1915 matter to grep and map. It now leave the block entry intact.
1917 The reason it worked without the my is because the block entry and exit
1918 were optimized away to an OP_SCOPE, which it doesn't matter if it's there
1919 or not.
1921NETaa13343: goto needed to longjmp when in a signal handler.
1922From: Robert Partington
1923Files patched: pp_ctl.c
1924 goto needed to longjmp() when in a signal handler to get back into the
1925 right run() context.
1928NETaa13344: strict vars shouldn't apply to globs or filehandles.
1929From: Andrew Wilcox
1930Files patched: gv.c
1931 Filehandles and globs will be excepted from "strict vars", so that you can
1932 do the standard Perl 4 trick of
1934 use strict;
1935 sub foo {
1936 local(*IN);
1937 open(IN,"file");
1938 }
1941NETaa13345: didn't use package DB
1942From: Hans Mulder
1943Files patched: lib/
1944 Now it does.
1946NETaa13348: av_undef didn't free scalar representing $#foo.
1947From: David Filo
1948Files patched: av.c
1949 av_undef didn't free scalar representing $#foo.
1951NETaa13349: sort sub accumulated save stack entries
1952From: David Filo
1953Files patched: pp_ctl.c
1954 COMMON only gets set if assigning to @_, which is reasonable. Most of the
1955 problem was a memory leak.
1957NETaa13351: didn't treat indirect filehandles as references.
1958From: Andy Dougherty
1959Files patched: op.c
1960 Now produces
1962 Can't use an undefined value as a symbol reference at ./foo line 3.
1965NETaa13352: OP_SCOPE allocated as UNOP rather than LISTOP.
1966From: Andy Dougherty
1967Files patched: op.c
1969NETaa13353: scope() didn't release filegv on OP_SCOPE optimization.
1970From: Larry Wall
1971Files patched: op.c
1972 When scope() nulled out a NEXTSTATE, it didn't release its filegv reference.
1974NETaa13355: hv_delete now avoids useless mortalcopy
1975From: Larry Wall
1976Files patched: hv.c op.c pp.c pp_ctl.c proto.h scope.c util.c
1977 hv_delete now avoids useless mortalcopy.
1980NETaa13359: comma operator section missing its heading
1981From: Larry Wall
1982Files patched: pod/perlop.pod
1984NETaa13359: random typo
1985Files patched: pod/perldiag.pod
1987NETaa13360: code to handle partial vec values was bogus.
1988From: Conrad Augustin
1989Files patched: pp.c
1990 The code that Mark J. added a long time ago to handle values that were partially
1991 off the end of the string was incorrect.
1993NETaa13361: made it not interpolate inside regexp comments
1994From: Martin Jost
1995Files patched: toke.c
1996 To avoid surprising people, it no longer interpolates inside regexp
1999NETaa13362: ${q[1]} should be interpreted like it used to
2000From: Hans Mulder
2001Files patched: toke.c
2002 Now resolves ${keyword[1]} to $keyword[1] and warns if -w. Likewise for {}.
2004NETaa13363: meaning of repeated search chars undocumented in tr///
2005From: Stephen P. Potter
2006Files patched: pod/perlop.pod
2007 Documented that repeated characters use the first translation given.
2009NETaa13365: if closedir fails, don't try it again.
2010From: Frank Crawford
2011Files patched: pp_sys.c
2012 Now does not attempt to closedir a second time.
2014NETaa13366: can't do block scope optimization on $1 et al when tainting.
2015From: Andrew Vignaux
2016Files patched: toke.c
2017 The tainting mechanism assumes that every statement starts out
2018 untainted. Unfortunately, the scope removal optimization for very
2019 short blocks removed the statementhood of statements that were
2020 attempting to read $1 as an untainted value, with the effect that $1
2021 appeared to be tainted anyway. The optimization is now disabled when
2022 tainting and the block contains $1 (or equivalent).
2024NETaa13366: fixed this a better way in toke.c.
2025Files patched: op.c
2026 (same)
2028NETaa13366: need to disable scope optimization when tainting.
2029Files patched: op.c
2030 (same)
2032NETaa13367: Did a SvCUR_set without nulling out final char.
2033From: "Rob Henderson" <>
2034Files patched: doop.c pp.c pp_sys.c
2035 When do_vop set the length on its result string it neglected to null-terminate
2036 it.
2038NETaa13368: bigrat::norm sometimes chucked sign
2039From: Greg Kuperberg
2040Files patched: lib/
2041 The normalization routine was assuming that the gcd of two numbers was
2042 never negative, and based on that assumption managed to move the sign
2043 to the denominator, where it was deleted on the assumption that the
2044 denominator is always positive.
2046NETaa13368: botched previous patch
2047Files patched: lib/
2048 (same)
2050NETaa13369: # is now a comment character, and \# should be left for regcomp.
2051From: Simon Parsons
2052Files patched: toke.c
2053 It was not skipping the comment when it skipped the white space, and constructed
2054 an opcode that tried to match a null string. Unfortunately, the previous
2055 star tried to use the first character of the null string to optimize where
2056 to recurse, so it never matched.
2058NETaa13369: comment after regexp quantifier induced non-match.
2059Files patched: regcomp.c
2060 (same)
2062NETaa13370: some code assumed SvCUR was of type int.
2063From: Spider Boardman
2064Files patched: pp_sys.c
2065 Did something similar to the proposed patch. I also fixed the problem that
2066 it assumed the type of SvCUR was int. And fixed get{peer,sock}name the
2067 same way.
2069NETaa13375: sometimes dontbother wasn't added back into strend.
2070From: Jamshid Afshar
2071Files patched: regexec.c
2072 When the /g modifier was used, the regular expression code would calculate
2073 the end of $' too short by the minimum number of characters the pattern could
2074 match.
2076NETaa13375: sv_setpvn now disallows negative length.
2077Files patched: sv.c
2078 (same)
2080NETaa13376: suspected indirect objecthood prevented recognition of lexical.
2082Files patched: toke.c
2083 When $data[0] is used in a spot that might be an indirect object, the lexer
2084 was getting confused over the rule that says the $data in $$data[0] isn't
2085 an array element. (The lexer uses XREF state for both indirect objects
2086 and for variables used as names.)
2088NETaa13377: -I processesing ate remainder of #! line.
2089From: Darrell Schiebel
2090Files patched: perl.c
2091 I made the -I processing in moreswitches look for the end of the string,
2092 delimited by whitespace.
2094NETaa13379: ${foo} now treated the same outside quotes as inside
2095From: Hans Mulder
2096Files patched: toke.c
2097 ${bareword} is now treated the same outside quotes as inside.
2099NETaa13379: previous fix for this bug was botched
2100Files patched: toke.c
2101 (same)
2103NETaa13381: TEST should check for perl link
2104From: Andy Dougherty
2105Files patched: t/TEST
2106 die "You need to run \"make test\" first to set things up.\n" unless -e 'perl';
2109NETaa13384: fixed version 0.000 botch.
2110From: Larry Wall
2111Files patched: installperl
2113NETaa13385: return 0 from required file loses message
2114From: Malcolm Beattie
2115Files patched: pp_ctl.c
2116 Works right now.
2118NETaa13387: added pod2latex
2119From: Taro KAWAGISHI
2120Files patched: MANIFEST pod/pod2latex
2121 Added most recent copy to pod directory.
2123NETaa13388: constant folding now prefers integer results over double
2124From: Ilya Zakharevich
2125Files patched: op.c
2126 Constant folding now prefers integer results over double.
2128NETaa13389: now treats . and exec as shell metathingies
2129From: Hans Mulder
2130Files patched: doio.c
2131 Now treats . and exec as shell metathingies.
2133NETaa13395: eval didn't check taintedness.
2134From: Larry Wall
2135Files patched: pp_ctl.c
2137NETaa13396: $^ coredumps at end of string
2138From: Paul Rogers
2139Files patched: toke.c
2140 The scan_ident() didn't check for a null following $^.
2142NETaa13397: improved error messages when operator expected
2143From: Larry Wall
2144Files patched: toke.c
2145 Added message (Do you need to predeclare BAR?). Also fixed the missing
2146 semicolon message.
2148NETaa13399: cleanup by Andy
2149From: Larry Wall
2150Files patched: Changes Configure Makefile.SH README cflags.SH config.H config_h.SH deb.c doop.c dump.c ext/DB_File/ ext/DB_File/DB_File.xs ext/DynaLoader/ ext/Fcntl/ ext/GDBM_File/ ext/POSIX/ ext/SDBM_File/sdbm/sdbm.h ext/Socket/ ext/util/make_ext h2xs.SH hints/ hints/ hints/ hints/ hints/ hints/ hints/ hints/ hints/ hints/ installperl lib/ lib/ lib/ExtUtils/ lib/ExtUtils/xsubpp lib/Term/ mg.c miniperlmain.c perl.c perl.h perl_exp.SH pod/Makefile pod/perldiag.pod pod/pod2html pp.c pp_ctl.c pp_hot.c pp_sys.c proto.h sv.h t/re_tests util.c x2p/Makefile.SH x2p/a2p.h x2p/a2py.c x2p/handy.h x2p/hash.c x2p/hash.h x2p/str.c x2p/str.h x2p/util.c x2p/util.h x2p/walk.c
2152NETaa13399: cleanup from Andy
2153Files patched: MANIFEST
2155NETaa13399: configuration cleanup
2156Files patched: Configure Configure MANIFEST MANIFEST Makefile.SH Makefile.SH README config.H config.H config_h.SH config_h.SH configpm ext/DynaLoader/ ext/DynaLoader/dl_hpux.xs ext/NDBM_File/Makefile.PL ext/ODBM_File/Makefile.PL ext/util/make_ext handy.h hints/ hints/ hints/ hints/ hints/ hints/ hints/ hints/ hints/ installperl installperl lib/ lib/ExtUtils/ lib/ExtUtils/ lib/ExtUtils/xsubpp lib/Getopt/ lib/Text/ makedepend.SH makedepend.SH mg.c op.c perl.h perl_exp.SH pod/perl.pod pod/perldiag.pod pod/perlsyn.pod pod/pod2man pp_sys.c proto.h proto.h unixish.h util.c util.c vms/config.vms writemain.SH x2p/a2p.h x2p/a2p.h x2p/a2py.c x2p/a2py.c x2p/handy.h x2p/util.c x2p/walk.c x2p/walk.c
2158NETaa13399: new files from Andy
2159Files patched: ext/DB_File/Makefile.PL ext/DynaLoader/Makefile.PL ext/Fcntl/Makefile.PL ext/GDBM_File/Makefile.PL ext/NDBM_File/Makefile.PL ext/ODBM_File/Makefile.PL ext/POSIX/Makefile.PL ext/SDBM_File/Makefile.PL ext/SDBM_File/sdbm/Makefile.PL ext/Socket/Makefile.PL globals.c hints/ hints/
2161NETaa13399: patch0l from Andy
2162Files patched: Configure MANIFEST Makefile.SH config.H config_h.SH ext/DB_File/Makefile.PL ext/GDBM_File/Makefile.PL ext/NDBM_File/Makefile.PL ext/POSIX/POSIX.xs ext/SDBM_File/sdbm/Makefile.PL ext/util/make_ext h2xs.SH hints/ hints/ hints/ installperl lib/ lib/ExtUtils/ makeaperl.SH vms/config.vms x2p/util.c x2p/util.h
2164NETaa13399: stuff from Andy
2165Files patched: Configure MANIFEST Makefile.SH configpm hints/ hints/ hints/ lib/ExtUtils/ util.c
2167NETaa13399: Patch 0k from Andy
2168Files patched: Configure MANIFEST Makefile.SH config.H config_h.SH hints/ hints/ hints/ hints/ installperl lib/ExtUtils/ lib/File/ makeaperl.SH minimod.PL perl.c proto.h vms/config.vms vms/ext/ x2p/a2p.h
2170NETaa13399: Patch 0m from Andy
2171Files patched: Configure MANIFEST Makefile.SH README config.H config_h.SH ext/DynaLoader/README ext/POSIX/POSIX.xs ext/SDBM_File/sdbm/sdbm.h ext/util/extliblist hints/ hints/ hints/ lib/ExtUtils/ malloc.c perl.h pp_sys.c util.c
2173NETaa13400: pod2html update from Bill Middleton
2174From: Larry Wall
2175Files patched: pod/pod2html
2177NETaa13401: Boyer-Moore code attempts to compile string longer than 255.
2178From: Kyriakos Georgiou
2179Files patched: util.c
2180 The Boyer-Moore table uses unsigned char offsets, but the BM compiler wasn't
2181 rejecting strings longer than 255 chars, and was miscompiling them.
2183NETaa13403: missing a $ on variable name
2184From: Wayne Scott
2185Files patched: installperl
2186 Yup, it was missing.
2188NETaa13406: didn't wipe out dead match when proceeding to next BRANCH
2189From: Michael P. Clemens
2190Files patched: regexec.c
2191 The code to check alternatives didn't invalidate backreferences matched by the
2192 failed branch.
2194NETaa13407: overload upgrade
2196Also: Ilya Zakharevich
2197Files patched: MANIFEST gv.c lib/Math/ perl.h pod/perlovl.pod pp.c pp.h pp_hot.c sv.c t/lib/bigintpm.t t/op/overload.t
2198 Applied supplied patch, and fixed bug induced by use of sv_setsv to do
2199 a deep copy, since sv_setsv no longer copies objecthood.
2201NETaa13409: sv_gets tries to grow string at EOF
2202From: Harold O Morris
2203Files patched: sv.c
2204 Applied suggested patch, only two statements earlier, since the end code
2205 also does SvCUR_set.
2207NETaa13410: delaymagic did =~ instead of &= ~
2208From: Andreas Schwab
2209Files patched: pp_hot.c
2210 Applied supplied patch.
2212NETaa13411: POSIX didn't compile under -DLEAKTEST
2213From: Frederic Chauveau
2214Files patched: ext/POSIX/POSIX.xs
2215 Used NEWSV instead of newSV.
2217NETaa13412: new version from Tony Sanders
2218From: Tony Sanders
2219Files patched: lib/Term/
2220 Installed as
2222NETaa13413: regmust extractor needed to restart loop on BRANCH for (?:) to work
2224Files patched: regcomp.c
2225 The BRANCH skipper should have restarted the loop from the top.
2227NETaa13414: the check for accidental list context was done after pm_short check
2228From: Michael H. Coen
2229Files patched: pp_hot.c
2230 Moved check for accidental list context to before the pm_short optimization.
2232NETaa13418: perlre.pod babbled nonsense about | in character classes
2233From: Philip Hazel
2234Files patched: pod/perlre.pod
2235 Removed bogus brackets. Now reads:
2236 Note however that "|" is interpreted as a literal with square brackets,
2237 so if you write C<[fee|fie|foe]> you're really only matching C<[feio|]>.
2239NETaa13419: need to document introduction of lexical variables
2240From: "Heading, Anthony"
2241Files patched: pod/perlfunc.pod
2242 Now mentions that lexicals aren't introduced till after the current statement.
2244NETaa13420: formats that overflowed a page caused endless top of forms
2245From: Hildo@CONSUL.NL
2246Files patched: pp_sys.c
2247 If a record is too large to fit on a page, it now prints whatever will
2248 fit and then calls top of form again on the remainder.
2250NETaa13423: the code to do negative list subscript in scalar context was missing
2251From: Steve McDougall
2252Files patched: pp.c
2253 The negative subscript code worked right in list context but not in scalar
2254 context. In fact, there wasn't code to do it in the scalar context.
2256NETaa13424: existing but undefined CV blocked inheritance
2257From: Spider Boardman
2258Files patched: gv.c
2259 Applied supplied patch.
2261NETaa13425: removed extra argument to croak
2262From: "R. Bernstein"
2263Files patched: regcomp.c
2264 Removed extra argument.
2266NETaa13427: added return types
2267From: "R. Bernstein"
2268Files patched: x2p/a2py.c
2269 Applied suggested patch.
2271NETaa13427: added static declarations
2272Files patched: x2p/walk.c
2273 (same)
2275NETaa13428: split was assuming that all backreferences were defined
2276From: Dave Schweisguth
2277Files patched: pp.c
2278 split was assuming that all backreferences were defined.
2280NETaa13430: hoistmust wasn't hoisting anchored shortcircuit's length
2281From: Tom Christiansen
2282Also: Rob Hooft
2283Files patched: toke.c
2285NETaa13432: couldn't call code ref under debugger
2286From: Mike Fletcher
2287Files patched: op.c pp_hot.c sv.h
2288 The debugging code assumed it could remember a name to represent a subroutine,
2289 but anonymous subroutines don't have a name. It now remembers a CV reference
2290 in that case.
2292NETaa13435: 1' dumped core
2293From: Larry Wall
2294Files patched: toke.c
2295 Didn't check a pointer for nullness.
2297NETaa13436: print foo(123) didn't treat foo as subroutine
2299Files patched: toke.c
2300 Now treats it as a subroutine rather than a filehandle.
2302NETaa13437: &$::foo didn't think $::foo was a variable name
2304Files patched: toke.c
2305 Now treats $::foo as a global variable.
2307NETaa13439: referred to old package name
2308From: Tom Christiansen
2309Files patched: lib/Sys/
2310 Wasn't a strict refs problem after all. It was simply referring to package
2311 syslog, which had been renamed to Sys::Syslog.
2313NETaa13440: stat operations didn't know what to do with glob or ref to glob
2315Files patched: doio.c pp_sys.c
2316 Now knows about the kinds of filehandles returned by FileHandle constructors
2317 and such.
2319NETaa13442: couldn't find name of copy of deleted symbol table entry
2320From: Spider Boardman
2321Files patched: gv.c gv.h
2322 I did a much simpler fix. When gp_free notices that it's freeing the
2323 master GV, it nulls out gp_egv. The GvENAME and GvESTASH macros know
2324 to revert to gv if egv is null.
2326 This has the advantage of not creating a reference loop.
2328NETaa13443: couldn't override an XSUB
2329From: William Setzer
2330Files patched: op.c
2331 When the newSUB and newXS routines checked for whether the old sub was
2332 defined, they only looked at CvROOT(cv), not CvXSUB(cv).
2334NETaa13443: needed to do same thing in newXS
2335Files patched: op.c
2336 (same)
2338NETaa13444: -foo now doesn't warn unless sub foo is defined
2339From: Larry Wall
2340Files patched: toke.c
2341 Made it not warn on -foo, unless there is a sub foo defined.
2343NETaa13451: in scalar context, pp_entersub now guarantees one item from XSUB
2344From: Nick Gianniotis
2345Files patched: pp_hot.c
2346 The pp_entersub routine now guarantees that an XSUB in scalar context
2347 returns one and only one value. If there are fewer, it pushes undef,
2348 and if there are more, it returns the last one.
2350NETaa13457: now explicitly disallows printf format with 'n' or '*'.
2352Files patched: doop.c
2353 Now says
2355 Use of n in printf format not supported at ./foo line 3.
2358NETaa13458: needed to call SvPOK_only() in pp_substr
2359From: Wayne Scott
2360Files patched: pp.c
2361 Needed to call SvPOK_only() in pp_substr.
2363NETaa13459: umask and chmod now warn about missing initial 0 even with paren
2364From: Andreas Koenig
2365Files patched: toke.c
2366 Now skips parens as well as whitespace looking for argument.
2368NETaa13460: backtracking didn't work on .*? because reginput got clobbered
2369From: Andreas Koenig
2370Files patched: regexec.c
2371 When .*? did a probe of the rest of the string, it clobbered reginput,
2372 so the next call to match a . tried to match the newline and failed.
2374NETaa13475: \(@ary) now treats array as list of scalars
2375From: Tim Bunce
2376Files patched: op.c
2377 The mod() routine now refrains from marking @ary as an lvalue if it's in parens
2378 and is the subject of an OP_REFGEN.
2380NETaa13481: accept buffer wasn't aligned good enough
2381From: Holger Bechtold
2382Also: Christian Murphy
2383Files patched: pp_sys.c
2384 Applied suggested patch.
2386NETaa13486: while (<>) now means while (defined($_ = <>))
2387From: Jim Balter
2388Files patched: op.c pod/perlop.pod
2389 while (<HANDLE>) now means while (defined($_ = <HANDLE>)).
2391NETaa13500: needed DESTROY in FileHandle
2392From: Tim Bunce
2393Files patched: ext/POSIX/
2394 Added DESTROY method. Also fixed ungensym to use POSIX:: instead of _POSIX.
2395 Removed ungensym from close method, since DESTROY should do that now.
2397NETaa13502: now complains if you use local on a lexical variable
2398From: Larry Wall
2399Files patched: op.c
2400 Now says something like
2402 Can't localize lexical variable $var at ./try line 6.
2404NETaa13512: added $SIG{__WARN__} and $SIG{__DIE__} hooks
2405From: Larry Wall
2406Files patched: embed.h gv.c interp.sym mg.c perl.h pod/perlvar.pod pp_ctl.c util.c Todo pod/perldiag.pod
2408NETaa13514: statements before intro of lex var could see lex var
2409From: William Setzer
2410Files patched: op.c
2411 When a lexical variable is declared, introduction is delayed until
2412 the start of the next statement, so that any initialization code runs
2413 outside the scope of the new variable. Thus,
2415 my $y = 3;
2416 my $y = $y;
2417 print $y;
2419 should print 3. Unfortunately, the declaration was marked with the
2420 beginning location at the time that "my $y" was processed instead of
2421 when the variable was introduced, so any embedded statements within
2422 an anonymous subroutine picked up the wrong "my". The declaration
2423 is now labelled correctly when the variable is actually introduced.
2425NETaa13520: added closures
2426From: Larry Wall
2427Files patched: Todo cv.h embed.h global.sym gv.c interp.sym op.c perl.c perl.h pod/perlform.pod pp.c pp_ctl.c pp_hot.c sv.c sv.h toke.c
2429NETaa13520: test to see if lexical works in a format now
2430Files patched: t/op/write.t
2432NETaa13522: substitution couldn't be used on a substr()
2433From: Hans Mulder
2434Files patched: pp_ctl.c pp_hot.c
2435 Changed pp_subst not to use sv_replace() anymore, which didn't handle lvalues
2436 and was overkill anyway. Should be slightly faster this way too.
2438NETaa13525: G_EVAL mode in perl_call_sv didn't return values right.
2439Files patched: perl.c
2441NETaa13525: consolidated error message
2442From: Larry Wall
2443Files patched: perl.h toke.c
2445NETaa13525: derived it
2446Files patched: perly.h
2448NETaa13525: missing some values from embed.h
2449Files patched: embed.h
2451NETaa13525: random cleanup
2452Files patched: MANIFEST Todo cop.h lib/ lib/ opcode.h patchlevel.h pod/perldata.pod pod/perlsub.pod t/op/ref.t toke.c
2454NETaa13525: random cleanup
2455Files patched: pp_ctl.c util.c
2457NETaa13527: File::Find needed to export $name and $dir
2458From: Chaim Frenkel
2459Files patched: lib/File/
2460 They are now exported.
2462NETaa13528: cv_undef left unaccounted-for GV pointer in CV
2463From: Tye McQueen
2464Also: Spider Boardman
2465Files patched: op.c
2467NETaa13530: scalar keys now resets hash iterator
2468From: Tim Bunce
2469Files patched: doop.c
2470 scalar keys() now resets the hash iterator.
2472NETaa13531: h2ph doesn't check defined right
2473From: Casper H.S. Dik
2474Files patched: h2ph.SH
2476NETaa13540: VMS update
2477From: Larry Wall
2478Files patched: MANIFEST README.vms doio.c embed.h ext/DynaLoader/dl_vms.xs interp.sym lib/ lib/ExtUtils/xsubpp lib/File/ lib/File/ lib/File/ mg.c miniperlmain.c perl.c perl.h perly.c perly.c.diff pod/perldiag.pod pp_ctl.c pp_hot.c pp_sys.c proto.h util.c vms/Makefile vms/config.vms vms/descrip.mms vms/ext/ vms/ext/ vms/ext/VMS/stdio/Makefile.PL vms/ext/VMS/stdio/ vms/ext/VMS/stdio/stdio.xs vms/ vms/perlvms.pod vms/sockadapt.c vms/sockadapt.h vms/vms.c vms/vmsish.h vms/
2480NETaa13540: got some duplicate code
2481Files patched: lib/File/
2483NETaa13540: stuff from Charles
2484Files patched: MANIFEST README.vms lib/ExtUtils/ lib/ExtUtils/ lib/ExtUtils/xsubpp lib/File/ lib/File/ perl.c perl.h pod/perldiag.pod pod/perldiag.pod vms/Makefile vms/Makefile vms/config.vms vms/config.vms vms/descrip.mms vms/descrip.mms vms/ext/ vms/ext/ vms/ext/ vms/ext/ vms/ext/VMS/stdio/ vms/ext/VMS/stdio/stdio.xs vms/ vms/ vms/ vms/ vms/ vms/perlvms.pod vms/sockadapt.h vms/ vms/vms.c vms/vms.c vms/vmsish.h vms/vmsish.h vms/
2486NETaa13540: tweak from Charles
2487Files patched: lib/File/
2489NETaa13552: scalar unpack("P4",...) ignored the 4
2490From: Eric Arnold
2491Files patched: pp.c
2492 The optimization that tried to do only one item in a scalar context didn't
2493 realize that the argument to P was not a repeat count.
2495NETaa13553: now warns about 8 or 9 in octal escapes
2496From: Mike Rogers
2497Files patched: util.c
2498 Now warns if it finds 8 or 9 before the end of the octal escape sequence.
2499 So \039 produces a warning, but \0339 does not.
2501NETaa13554: now allows foreach ${"name"}
2502From: Johan Holtman
2503Files patched: op.c
2504 Instead of trying to remove OP_RV2SV, the compiler now just transmutes it into an
2505 OP_RV2GV, which is a no-op for ordinary variables and does the right
2506 thing for ${"name"}.
2508NETaa13559: substitution now always checks for readonly
2509From: Rodger Anderson
2510Files patched: pp_hot.c
2511 Substitution now always checks for readonly.
2513NETaa13561: added explanations of closures and curly-quotes
2514From: Larry Wall
2515Files patched: pod/perlref.pod
2517NETaa13562: null components in path cause indigestion
2518From: Ambrose Kofi Laing
2519Files patched: lib/ lib/
2521NETaa13575: documented semantics of negative substr length
2522From: Jeff Bouis
2523Files patched: pod/perlfunc.pod
2524 Documented the fact that negative length now leaves characters off the end,
2525 and while I was at it, made it work right even if offset wasn't 0.
2527NETaa13575: negative length to substr didn't work when offset non-zero
2528Files patched: pp.c
2529 (same)
2531NETaa13575: random cleanup
2532Files patched: pod/perlfunc.pod
2533 (same)
2535NETaa13580: couldn't localize $ACCUMULATOR
2536From: Larry Wall
2537Files patched: gv.c lib/ mg.c perl.c sv.c
2538 Needed to make $^A a real magical variable. Also lib/ wasn't
2539 exporting good.
2541NETaa13583: doc mods from Tom
2542From: Larry Wall
2543Files patched: pod/modpods/AnyDBMFile.pod pod/modpods/Basename.pod pod/modpods/Benchmark.pod pod/modpods/Cwd.pod pod/modpods/Dynaloader.pod pod/modpods/Exporter.pod pod/modpods/Find.pod pod/modpods/Finddepth.pod pod/modpods/Getopt.pod pod/modpods/MakeMaker.pod pod/modpods/Open2.pod pod/modpods/POSIX.pod pod/modpods/Ping.pod pod/modpods/less.pod pod/modpods/strict.pod pod/perlapi.pod pod/perlbook.pod pod/perldata.pod pod/perlform.pod pod/perlfunc.pod pod/perlipc.pod pod/perlmod.pod pod/perlobj.pod pod/perlref.pod pod/perlrun.pod pod/perlsec.pod pod/perlsub.pod pod/perltrap.pod pod/perlvar.pod
2545NETaa13589: return was enforcing list context on its arguments
2546From: Tim Freeman
2547Files patched:
2548 A return was being treated like a normal list operator, in that it was
2549 setting list context on its arguments. This was bogus.
2551NETaa13591: POSIX::creat used wrong argument
2552From: Paul Marquess
2553Files patched: ext/POSIX/
2554 Applied suggested patch.
2556NETaa13605: use strict refs error message now displays bad ref
2557From: Peter Gordon
2558Files patched: perl.h pod/perldiag.pod pp.c pp_hot.c
2559 Now says
2561 Can't use string ("2") as a HASH ref while "strict refs" in use at ./foo line 12.
2563NETaa13630: eof docs were unclear
2564From: Hallvard B Furuseth
2565Files patched: pod/perlfunc.pod
2566 Applied suggested patch.
2568NETaa13636: $< and $> weren't refetched on undump restart
2569From: Steve Pearlmutter
2570Files patched: perl.c
2571 The code in main() bypassed perl_construct on an undump restart, which bypassed
2572 the code that set $< and $>.
2574NETaa13641: added Tim's fancy new import whizbangers
2575From: Tim Bunce
2576Files patched: lib/
2577 Applied suggested patch.
2579NETaa13649: couldn't AUTOLOAD a symbol reference
2580From: Larry Wall
2581Files patched: pp_hot.c
2582 pp_entersub needed to guarantee a CV so it would get to the AUTOLOAD code.
2584NETaa13651: renamed file had wrong package name
2585From: Andreas Koenig
2586Files patched: lib/File/
2587 Applied suggested patch.
2589NETaa13660: now that we're testing distribution we can diagnose RANDBITS errors
2590From: Karl Glazebrook
2591Files patched: t/op/rand.t
2592 Changed to suggested algorithm. Also duplicated it to test rand(100) too.
2594NETaa13660: rand.t didn't test for proper distribution within range
2595Files patched: t/op/rand.t
2596 (same)
2598NETaa13671: array slice misbehaved in a scalar context
2599From: Tye McQueen
2600Files patched: pp.c
2601 A spurious else prevented the scalar-context-handling code from running.
2603NETaa13672: filehandle constructors in POSIX don't return failure successfully
2604From: Ian Phillipps
2605Files patched: ext/POSIX/
2606 Applied suggested patch.
2609NETaa13678: forced $1 to always be untainted
2610From: Ka-Ping Yee
2611Files patched: mg.c
2612 I believe the bug that triggered this was fixed elsewhere, but just in case,
2613 I put in explicit code to force $1 et al not to be tainted regardless.
2615NETaa13682: formline doc need to discuss ~ and ~~ policy
2616From: Peter Gordon
2617Files patched: pod/perlfunc.pod
2619NETaa13686: POSIX::open and POSIX::mkfifo didn't check tainting
2620From: Larry Wall
2621Files patched: ext/POSIX/POSIX.xs
2622 open() and mkfifo() now check tainting.
2624NETaa13687: new
2625From: Tim Bunce
2626Files patched: lib/
2627 Added suggested changes, except for @EXPORTABLE, because it looks too much
2628 like @EXPORTTABLE. Decided to stick with @EXPORT_OK because it looks more
2629 like an adjunct. Also added an export_tags routine. The keys in the
2630 %EXPORT_TAGS hash no longer use colons, to make the initializers prettier.
2632NETaa13687: new
2633Files patched: ext/POSIX/
2634 (same)
2636NETaa13694: add sockaddr_in to
2637From: Tim Bunce
2638Files patched: ext/Socket/
2639 Applied suggested patch.
2641NETaa13695: library routines should use qw() as good example
2642From: Dean Roehrich
2643Files patched: ext/DB_File/ ext/DynaLoader/ ext/Fcntl/ ext/GDBM_File/ ext/POSIX/ ext/Socket/
2644 Applied suggested patch.
2646NETaa13696: myconfig should be a routine in
2647From: Kenneth Albanowski
2648Files patched: configpm
2649 Applied suggested patch.
2651NETaa13704: fdopen closed fd on failure
2652From: Hallvard B Furuseth
2653Files patched: doio.c
2654 Applied suggested patch.
2656NETaa13706: Term::Cap doesn't work
2657From: Dean Roehrich
2658Files patched: lib/Term/
2659 Applied suggested patch.
2661NETaa13710: cryptswitch needed to be more "useable"
2662From: Tim Bunce
2663Files patched: embed.h global.sym perl.h toke.c
2664 The cryptswitch_fp function now can operate in two modes. It can
2665 modify the global rsfp to redirect input as before, or it can modify
2666 linestr and return true, indicating that it is not necessary for yylex
2667 to read another line since cryptswitch_fp has just done it.
2669NETaa13712: new_tmpfile() can't be called as constructor
2670From: Hans Mulder
2671Files patched: ext/POSIX/POSIX.xs
2672 Now allows new_tmpfile() to be called as a constructor.
2674NETaa13714: variable method call not documented
2675From: "Randal L. Schwartz"
2676Files patched: pod/perlobj.pod
2677 Now indicates that OBJECT->$method() works.
2679NETaa13715: PACK->$method produces spurious warning
2680From: Larry Wall
2681Files patched: toke.c
2682 The -> operator was telling the lexer to expect an operator when the
2683 next thing was a variable.
2685NETaa13716: Carp now allows multiple packages to be skipped out of
2686From: Larry Wall
2687Files patched: lib/
2688 The subroutine redefinition warnings now warn on import collisions.
2690NETaa13716: Exporter catches warnings and gives a better line number
2691Files patched: lib/
2692 (same)
2694NETaa13716: now counts imported routines as "defined" for redef warnings
2695Files patched: op.c sv.c
2696 (same)
2699Version 5.000
2702New things
2704 The -w switch is much more informative.
2706 References. See t/op/ref.t for examples. All entities in Perl 5 are
2707 reference counted so that it knows when each item should be destroyed.
2709 Objects. See t/op/ref.t for examples.
2711 => is now a synonym for comma. This is useful as documentation for
2712 arguments that come in pairs, such as initializers for associative arrays,
2713 or named arguments to a subroutine.
2715 All functions have been turned into list operators or unary operators,
2716 meaning the parens are optional. Even subroutines may be called as
2717 list operators if they've already been declared.
3edbfbe5 2719 More embeddible. See main.c and Multiple interpreters
2720 in the same process are supported (though not with interleaved
2721 execution yet).
2723 The interpreter is now flattened out. Compare Perl 4's eval.c with
2724 the perl 5's pp.c. Compare Perl 4's 900 line interpreter loop in cmd.c
2725 with Perl 5's 1 line interpreter loop in run.c. Eventually we'll make
2726 everything non-blocking so we can interface nicely with a scheduler.
2728 eval is now treated more like a subroutine call. Among other things,
2729 this means you can return from it.
2731 Format value lists may be spread over multiple lines by enclosing in
85e6fe83 2732 a do {} block.
2734 You may now define BEGIN and END subroutines for each package. The BEGIN
2735 subroutine executes the moment it's parsed. The END subroutine executes
2736 just before exiting.
2738 Flags on the #! line are interpreted even if the script wasn't
2739 executed directly. (And even if the script was located by "perl -x"!)
2741 The ?: operator is now legal as an lvalue.
2743 List context now propagates to the right side of && and ||, as well
2744 as the 2nd and 3rd arguments to ?:.
2746 The "defined" function can now take a general expression.
2748 Lexical scoping available via "my". eval can see the current lexical
2749 variables.
2751 The preferred package delimiter is now :: rather than '.
2753 tie/untie are now preferred to dbmopen/dbmclose. Multiple DBM
2754 implementations are allowed in the same executable, so you can
2755 write scripts to interchange data among different formats.
2757 New "and" and "or" operators work just like && and || but with
2758 a precedence lower than comma, so they work better with list operators.
2760 New functions include: abs(), chr(), uc(), ucfirst(), lc(), lcfirst(),
2761 chomp(), glob()
2763 require with a number checks to see that the version of Perl that is
2764 currently running is at least that number.
2766 Dynamic loading of external modules is now supported.
2768 There is a new quote form qw//, which is equivalent to split(' ', q//).
2770 Assignment of a reference to a glob value now just replaces the
2771 single element of the glob corresponding to the reference type:
2772 *foo = \$bar, *foo = \&bletch;
2774 Filehandle methods are now supported:
2775 output_autoflush STDOUT 1;
2777 There is now an "English" module that provides human readable translations
2778 for cryptic variable names.
2780 Autoload stubs can now call the replacement subroutine with goto &realsub.
2782 Subroutines can be defined lazily in any package by declaring an AUTOLOAD
2783 routine, which will be called if a non-existent subroutine is called in
2784 that package.
2786 Several previously added features have been subsumed under the new
2787 keywords "use" and "no". Saying "use Module LIST" is short for
2788 BEGIN { require Module; import Module LIST; }
2789 The "no" keyword is identical except that it calls "unimport" instead.
2790 The earlier pragma mechanism now uses this mechanism, and two new
2791 modules have been added to the library to implement "use integer"
2792 and variations of "use strict vars, refs, subs".
2794 Variables may now be interpolated literally into a pattern by prefixing
2795 them with \Q, which works just like \U, but backwhacks non-alphanumerics
2796 instead. There is also a corresponding quotemeta function.
2798 Any quantifier in a regular expression may now be followed by a ? to
2799 indicate that the pattern is supposed to match as little as possible.
2801 Pattern matches may now be followed by an m or s modifier to explicitly
2802 request multiline or singleline semantics. An s modifier makes . match
2803 newline.
2805 Patterns may now contain \A to match only at the beginning of the string,
2806 and \Z to match only at the end. These differ from ^ and $ in that
2807 they ignore multiline semantics. In addition, \G matches where the
2808 last interation of m//g or s///g left off.
2810 Non-backreference-producing parens of various sorts may now be
2811 indicated by placing a ? directly after the opening parenthesis,
2812 followed by a character that indicates the purpose of the parens.
2813 An :, for instance, indicates simple grouping. (?:a|b|c) will
2814 match any of a, b or c without producing a backreference. It does
2815 "eat" the input. There are also assertions which do not eat the
2816 input but do lookahead for you. (?=stuff) indicates that the next
2817 thing must be "stuff". (?!nonsense) indicates that the next thing
2818 must not be "nonsense".
2820 The negation operator now treats non-numeric strings specially.
2821 A -"text" is turned into "-text", so that -bareword is the same
2822 as "-bareword". If the string already begins with a + or -, it
2823 is flipped to the other sign.
85e6fe83 2824
2827 @ now always interpolates an array in double-quotish strings. Some programs
2828 may now need to use backslash to protect any @ that shouldn't interpolate.
2830 Ordinary variables starting with underscore are no longer forced into
2831 package main.
2833 s'$lhs'$rhs' now does no interpolation on either side. It used to
2834 interplolate $lhs but not $rhs.
2836 The second and third arguments of splice are now evaluated in scalar
2837 context (like the book says) rather than list context.
2839 Saying "shift @foo + 20" is now a semantic error because of precedence.
2841 "open FOO || die" is now incorrect. You need parens around the filehandle.
2843 The elements of argument lists for formats are now evaluated in list
2844 context. This means you can interpolate list values now.
2846 You can't do a goto into a block that is optimized away. Darn.
2848 It is no longer syntactically legal to use whitespace as the name
a0d0e21e 2849 of a variable, or as a delimiter for any kind of quote construct.
2851 Some error messages will be different.
2853 The caller function now returns a false value in a scalar context if there
2854 is no caller. This lets library files determine if they're being required.
2856 m//g now attaches its state to the searched string rather than the
2857 regular expression.
2859 "reverse" is no longer allowed as the name of a sort subroutine.
2861 taintperl is no longer a separate executable. There is now a -T
2862 switch to turn on tainting when it isn't turned on automatically.
2864 Symbols starting with _ are no longer forced into package main, except
2865 for $_ itself (and @_, etc.).
2867 Double-quoted strings may no longer end with an unescaped $ or @.
2869 Negative array subscripts now count from the end of the array.
2871 The comma operator in a scalar context is now guaranteed to give a
2872 scalar context to its arguments.
2874 The ** operator now binds more tightly than unary minus.
2876 Setting $#array lower now discards array elements so that destructors
2877 work reasonably.
2879 delete is not guaranteed to return the old value for tied arrays,
2880 since this capability may be onerous for some modules to implement.
2882 Attempts to set $1 through $9 now result in a run-time error.