------------- Version 5.002 ------------- The main enhancement to the Perl core was the addition of prototypes. Many of the modules that come with Perl have been extensively upgraded. Other than that, nearly all the changes for 5.002 were bug fixes of one variety or another, so here's the bug list, along with the "resolution" for each of them. If you wish to correspond about any of them, please include the bug number (if any). Added APPLLIB_EXP for embedded perl library support. Files patched: perl.c Couldn't define autoloaded routine by assignment to typeglob. Files patched: pp_hot.c sv.c NETaa13525: Tiny patch to fix installman -n From: Larry Wall Files patched: installman NETaa13525: de-documented \v Files patched: pod/perlop.pod pod/perlre.pod NETaa13525: doc changes Files patched: pod/perlop.pod pod/perltrap.pod NETaa13525: perlxs update from Dean Roehrich Files patched: pod/perlxs.pod NETaa13525: rename powerunix to powerux Files patched: MANIFEST hints/powerux.sh NETaa13540: VMS uses CLK_TCK for HZ Files patched: pp_sys.c NETaa13721: pad_findlex core dumps on bad CvOUTSIDE() From: Carl Witty Files patched: op.c sv.c toke.c Each CV has a reference to the CV containing it lexically. Unfortunately, it didn't reference-count this reference, so when the outer CV was freed, we ended up with a pointer to memory that got reused later as some other kind of SV. NETaa13721: warning suppression Files patched: toke.c (same) NETaa13722: walk.c had inconsistent static declarations From: Tim Bunce Files patched: x2p/walk.c Consolidated the various declarations and made them consistent with the actual definitions. NETaa13724: -MPackage=args patch From: Tim Bunce Files patched: perl.c pod/perlrun.pod Added in the -MPackage=args patch too. NETaa13729: order-of-evaluation dependency in scope.c on leaving REGCONTEXT From: "Jason Shirk" Files patched: scope.c Did I32 delta = SSPOPINT; savestack_ix -= delta; /* regexp must have croaked */ instead. NETaa13731: couldn't assign external lexical array to itself From: oneill@cs.sfu.ca Files patched: op.c The pad_findmy routine was only checking previous statements for previous mention of external lexicals, so the fact that the current statement already mentioned @list was not noted. It therefore allocated another reference to the outside lexical, and this didn't compare equal when the assigment parsing code was trying to determine whether there was a common variable on either side of the equals. Since it didn't see the same variable, it thought it could avoid making copies of the values on the stack during list assignment. Unfortunately, before using those values, the list assignment has to zero out the target array, which destroys the values. The fix was to make pad_findmy search the current statement as well. This was actually a holdover from some old code that was trying to delay introduction of "my" variables until the next statement. This is now done with a different mechanism, so the fix should not adversely affect that. NETaa13733: s/// doesn't free old string when using copy mode From: Larry Wall Files patched: pp_ctl.c pp_hot.c When I removed the use of sv_replace(), I simply forgot to free the old char*. NETaa13736: closures leaked memory From: Carl Witty Files patched: op.c pp.c This is a specific example of a more general bug, fixed as NETaa13760, having to do with reference counts on comppads. NETaa13739: XSUB interface caches gimme in case XSUB clobbers it From: Dean Roehrich Files patched: pp_hot.c Applied suggest patch. Also deleted second gimme declaration as redundant. NETaa13760: comppad reference counts were inconsistent From: Larry Wall Files patched: op.c perl.c pp_ctl.c toke.c All official references to comppads are supposed to be through compcv now, but the transformation was not complete, resulting in memory leakage. NETaa13761: sv_2pv() wrongly preferred IV to NV when SV was readonly From: "Jack R. Lawler" Files patched: sv.c Okay, I understand how this one happened. This is a case where a beneficial fix uncovered a bug elsewhere. I changed the constant folder to prefer integer results over double if the numbers are the same. In this case, they aren't, but it leaves the integer value there anyway because the storage is already allocated for it, and it *might* be used in an integer context. And since it's producing a constant, it sets READONLY. Unfortunately, sv_2pv() bogusly preferred the integer value to the double when READONLY was set. This never showed up if you just said print 1.4142135623731; because in that case, there was already a string value. NETaa13772: shmwrite core dumps consistently From: Gabe Schaffer Files patched: opcode.h opcode.pl The shmwrite operator is a list operator but neglected to push a stack mark beforehand, because an 'm' was missing from opcode.pl. NETaa13773: $. was misdocumented as read-only. From: Inaba Hiroto Files patched: pod/perlvar.pod <1.array-element-read-only> % perl -le '$,=", "; $#w=5; for (@w) { $_=1; } print @w' Modification of a read-only value attempted at -e line 1. % perl4 -le '$,=", "; $#w=5; for (@w) { $_=1; } print @w' 1, 1, 1, 1, 1, 1 This one may stay the way it is for performance reasons. <2.begin-local-RS> % cat abc a b c % perl -e 'BEGIN { local $/ = ""; } print "$.:$_" while <>;' abc 1:a b c % perl -e '{ local $/ = ""; } print "$.:$_" while <>;' abc 1:a 2:b 3:c $/ wasn't initialized early enough, so local set it back to permanently undefined on exit from the block. <3.grep-x0-bug> % perl -le 'print grep(/^-/ ? ($x=$_) x 0 : 1, "a", "-b", "c");' a % perl4 -le 'print grep(/^-/ ? ($x=$_) x 0 : 1, "a", "-b", "c");' ac An extra mark was left on the stack if (('x') x $repeat) was used in a scalar context. <4.input-lineno-assign> # perl -w does not complain about assignment to $. (Is this just a feature?) # perlvar.pod says "This variable should be considered read-only." % cat abc a b c % perl -wnle '$. = 10 if $. == 2; print "$.:$_"' abc 1:a 10:b 11:c Fixed doc. <5.local-soft-ref.bug> % perl -e 'local ${"a"}=1;' zsh: 529 segmentation fault perl -e 'local ${"a"}=1;' Now says Can't localize a reference at -e line 1. <6.package-readline> % perl -e 'package foo; sub foo { 1; } package main; $_ = foo::foo(); print' 1 % perl -e ' package readline; sub foo { 1; } package main; $_ = readline::foo(); print' Undefined subroutine &main::foo called at -e line 1. % perl -e ' package readline; sub foo { 1; } package main; $_ = &readline::foo(); print' 1 Now treats foo::bar correctly even if foo is a keyword. <7.page-head-set-to-null-string> % cat page-head #From: russell@ccu1.auckland.ac.nz (Russell Fulton) #Newsgroups: comp.lang.perl #Subject: This script causes Perl 5.00 to sementation fault #Date: 15 Nov 1994 00:11:37 GMT #Message-ID: <3a8ubp$jrj@net.auckland.ac.nz> select((select(STDOUT), $^='')[0]); #this is the critical line $a = 'a'; write ; exit; format STDOUT = @<<<<<< $a . % perl page-head zsh: 1799 segmentation fault perl /tmp/page-head Now says Undefined top format "main::" called at ./try line 11. <8.sub-as-index> # parser bug? % perl -le 'sub foo {0}; $x[0]=0;$x[foo]<=0' Unterminated <> operator at -e line 1. % perl -le 'sub foo {0}; $x[0]=0;$x[foo()]<=0' A right square bracket now forces expectation of an operator. <9.unary-minus-to-regexp-var> % cat minus-reg #From: Michael Cook #Newsgroups: comp.lang.perl #Subject: bug: print -$1 #Date: 01 Feb 1995 15:31:25 GMT #Message-ID: $_ = "123"; /\d+/; print $&, "\n"; print -$&, "\n"; print 0-$&, "\n"; % perl minus-reg 123 123 -123 Apparently already fixed in my copy. <10.vec-segv> % cat vec-bug ## Offset values are changed for my machine. #From: augustin@gdstech.grumman.com (Conrad Augustin) #Subject: perl5 vec() bug? #Message-ID: <1994Nov22.193728.25762@gdstech.grumman.com> #Date: Tue, 22 Nov 1994 19:37:28 GMT #The following two statements each produce a segmentation fault in perl5: #vec($a, 21406, 32) = 1; # seg fault vec($a, 42813, 16) = 1; # seg fault #When the offset values are one less, all's well: #vec($a, 21405, 32) = 1; # ok #vec($a, 42812, 16) = 1; # ok #Interestingly, this is ok for all high values of N: #$N=1000000; vec($a, $N, 8) = 1; % perl vec-bug zsh: 1806 segmentation fault perl vec-bug Can't reproduce this one. NETaa13773: $/ not correctly localized in BEGIN Files patched: perl.c (same) NETaa13773: foo::bar was misparsed if foo was a reserved word Files patched: toke.c toke.c (same) NETaa13773: right square bracket didn't force expectation of operator Files patched: toke.c (same) NETaa13773: scalar ((x) x $repeat) left stack mark Files patched: op.c (same) NETaa13778: -w coredumps on <$> From: Hans Mulder Files patched: pp_hot.c toke.c Now produces suggested error message. Also installed guard in warning code that coredumped. NETaa13779: foreach didn't use savestack mechanism From: Hans Mulder Files patched: cop.h pp_ctl.c The foreach mechanism saved the old scalar value on the context stack rather than the savestack. It could consequently get out of sync if unexpectedly unwound. NETaa13785: GIMME sometimes used wrong context frame From: Greg Earle Files patched: embed.h global.sym op.h pp_ctl.c proto.h The expression inside the return was taking its context from the immediately surrounding block rather than the innermost surrounding subroutine call. NETaa13797: could modify sv_undef through auto-vivification From: Ilya Zakharevich Files patched: pp.c Inserted the missing check for readonly values on auto-vivification. NETaa13798: if (...) {print} treats print as quoted From: Larry Wall Files patched: toke.c The trailing paren of the condition was setting expectations to XOPERATOR rather than XBLOCK, so it was being treated like ${print}. NETaa13926: commonality was not detected in assignments using COND_EXPR From: Mark Hanson Files patched: opcode.h opcode.pl The assignment compiler didn't check the 2nd and 3rd args of a ?: for commonality. It still doesn't, but I made ?: into a "dangerous" operator so it is forced to treat it as common. NETaa13957: was marking the PUSHMARK as modifiable rather than the arg From: David Couture Files patched: op.c sv.c It was marking the PUSHMARK as modifiable rather than the arg. NETaa13962: documentation of behavior of scalar <*> was unclear From: Tom Christiansen Files patched: pod/perlop.pod Added the following to perlop: A glob only evaluates its (embedded) argument when it is starting a new list. All values must be read before it will start over. In a list context this isn't important, because you automatically get them all anyway. In a scalar context, however, the operator returns the next value each time it is called, or a FALSE value if you've just run out. Again, FALSE is returned only once. So if you're expecting a single value from a glob, it is much better to say ($file) = ; than $file = ; because the latter will alternate between returning a filename and returning FALSE. NETaa13986: split ignored /m pattern modifier From: Winfried Koenig Files patched: pp.c Fixed to work like m// and s///. NETaa13992: regexp comments not seen after + in non-extended regexp From: Mark Knutsen Files patched: regcomp.c The code to skip regexp comments was guarded by a conditional that only let it work when /x was in effect. NETaa14014: use subs should not count as definition, only as declaration From: Keith Thompson Files patched: sv.c On *foo = \&bar, doesn't set GVf_IMPORTED if foo and bar are in same package. NETaa14021: sv_inc and sv_dec "upgraded" magical SV to non-magical From: Paul A Sand Also: Andreas Koenig Files patched: sv.c The sv_inc() and sv_dec() routines "upgraded" null magical SVs to non-magical. NETaa14086: require should check tainting From: Karl Simon Berg Files patched: pp_ctl.c Since we shouldn't allow tainted requires anyway, it now says: Insecure dependency in require while running with -T switch at tst.pl line 1. NETaa14104: negation fails on magical variables like $1 From: tim Files patched: pp.c Negation was failing on magical values like $1. It was testing the wrong bits and also failed to provide a final "else" if none of the bits matched. NETaa14107: deep sort return leaked contexts From: Quentin Fennessy Files patched: pp_ctl.c Needed to call dounwind() appropriately. NETaa14129: attempt to localize via a reference core dumps From: Michele Sardo Files patched: op.c pod/perldiag.pod Now produces an error "Can't localize a reference", with explanation in perldiag. NETaa14138: substr() and s/// can cause core dump From: Andrew Vignaux Files patched: pp_hot.c Forgot to call SvOOK_off() on the SV before freeing its string. NETaa14145: ${@INC}[0] dumped core in debugger From: Hans Mulder Files patched: sv.c Now croaks "Bizarre copy of ARRAY in block exit", which is better than a core dump. The fact that ${@INC}[0] means $INC[0] outside the debugger is a different bug. NETaa14147: bitwise assignment ops wipe out byte of target string From: Jim Richardson Files patched: doop.c The code was assuming that the target was not either of the two operands, which is false for an assignment operator. NETaa14153: lexing of lexicals in patterns fooled by character class From: Dave Bianchi Files patched: toke.c It never called the dwimmer, which is how it fooled it. NETaa14154: allowed autoloaded methods by recognizing sub method; declaration From: Larry Wall Files patched: gv.c Made sub method declaration sufficient for autoloader to stop searching on. NETaa14156: shouldn't optimize block scope on tainting From: Pete Peterson Files patched: op.c toke.c I totally disabled the block scope optimization when running tainted. NETaa14157: -T and -B only allowed 1/30 "odd" characters--changed to 1/3 From: Tor Lillqvist Files patched: pp_sys.c Applied suggested patch. NETaa14160: deref of null symbol should produce null list From: Jared Rhine Files patched: pp_hot.c It didn't check for list context before returning undef. NETaa14162: POSIX::gensym now returns a symbol reference From: Josh N. Pritikin Also: Tim Bunce Files patched: ext/POSIX/POSIX.pm Applied suggested patch. NETaa14164: POSIX autoloader now distinguishes non-constant "constants" From: Tim Bunce Files patched: ext/POSIX/POSIX.pm ext/POSIX/POSIX.xs The .xs file now distinguishes non-constant "constants" by setting EAGAIN. This will also let us use #ifdef within the .xs file to de-constantify any other macros that happen not to be constants even if they don't use an argument. NETaa14166: missing semicolon after "my" induces core dump From: Thomas Kofler Files patched: toke.c The parser was left thinking it was still processing a "my", and flubbed. I made it wipe out the "in_my" variable on a syntax error. NETaa14166: missing semicolon after "my" induces core dump" Files patched: toke.c (same) NETaa14206: can now use English and strict at the same time From: Andrew Wilcox Files patched: sv.c It now counts imported symbols as okay under "use strict". NETaa14206: can now use English and strict at the same time Files patched: gv.c pod/perldiag.pod (same) NETaa14265: elseif now produces severe warning From: Yutao Feng Files patched: pod/perldiag.pod toke.c Now complains explicitly about "elseif". NETaa14279: list assignment propagated taintedness to independent scalars From: Tim Freeman Files patched: pp_hot.c List assignment needed to be modified so that tainting didn't propagate between independent scalar values. NETaa14312: undef in @EXPORTS core dumps From: William Setzer Files patched: lib/Exporter.pm Now says: Unable to create sub named "t::" at lib/Exporter.pm line 159. Illegal null symbol in @t::EXPORT at -e line 1 BEGIN failed--compilation aborted at -e line 1. NETaa14312: undef in @EXPORTS core dumps Files patched: pod/perldiag.pod sv.c (same) NETaa14321: literal @array check shouldn't happen inside embedded expressions From: Mark H. Nodine Files patched: toke.c The general solution to this is to disable the literal @array check within any embedded expression. For instance, this also failed bogusly: print "$foo{@foo}"; The reason fixing this also fixes the s///e problem is that the lexer effectively puts the RHS into a do {} block, making the expression embedded within curlies, as far as the error message is concerned. NETaa14322: now localizes $! during POSIX::AUTOLOAD From: Larry Wall Files patched: ext/POSIX/POSIX.pm Added local $! = 0. NETaa14324: defined() causes spurious sub existence From: "Andreas Koenig" Files patched: op.c pp.c It called pp_rv2cv which wrongly assumed it could add any sub it referenced. NETaa14336: use Module () forces import of nothing From: Tim Bunce Files patched: op.c use Module () now refrains from calling import at all. NETaa14353: added special HE allocator From: Larry Wall Files patched: global.sym NETaa14353: added special HE allocator Files patched: hv.c perl.h NETaa14353: array extension now converts old memory to SV storage. Files patched: av.c av.h sv.c NETaa14353: hashes now convert old storage into SV arenas. Files patched: global.sym NETaa14353: hashes now convert old storage into SV arenas. Files patched: hv.c perl.h NETaa14353: upgraded SV arena allocation Files patched: proto.h NETaa14353: upgraded SV arena allocation Files patched: perl.c sv.c NETaa14422: added rudimentary prototypes From: Gisle Aas Files patched: Makefile.SH op.c op.c perly.c perly.c.diff perly.h perly.y proto.h sv.c toke.c Message-Id: <9509290018.AA21548@scalpel.netlabs.com> To: doughera@lafcol.lafayette.edu (Andy Dougherty) Cc: perl5-porters@africa.nicoh.com Subject: Re: Jumbo Configure patch vs. 1m. Date: Thu, 28 Sep 95 17:18:54 -0700 From: lwall@scalpel.netlabs.com (Larry Wall) : No. Larry's currently got the patch pumpkin for all such core perl topics. I dunno whether you should let me have the patch pumpkin or not. To fix a Sev 2 I just hacked in rudimentary prototypes. :-) We can now define true unary subroutines, as well as argumentless subroutines: sub baz () { 12; } # Must not have argument sub bar ($) { $_[0] * 7 } # Must have exactly one argument sub foo ($@) { print "@_\n" } # Must have at least one argument foo bar baz / 2 || "oops", "is the answer"; This prints "42 is the answer" on my machine. That is, it's the same as foo( bar( baz() / 2) || "oops", "is the answer"); Attempting to compile foo; results in Too few arguments for main::foo at ./try line 8, near "foo;" Compiling bar 1,2,3; results in Too many arguments for main::bar at ./try line 8, near "foo;" But @array = ('a','b','c'); foo @array, @array; prints "3 a b c" because the $ puts the first arg of foo into scalar context. The main win at this point is that we can say sub AAA () { 1; } sub BBB () { 2; } and the user can say AAA + BBB and get 3. I'm not quite sure how this interacts with autoloading though. I fear POSIX.pm will need to say sub E2BIG (); sub EACCES (); sub EAGAIN (); sub EBADF (); sub EBUSY (); ... sub _SC_STREAM_MAX (); sub _SC_TZNAME_MAX (); sub _SC_VERSION (); unless we can figure out how to efficiently declare a default prototype at import time. Meaning, not using eval. Currently *foo = \&bar; (the ordinary import mechanism) implicitly stubs &bar with no prototype if &bar is not yet declared. It's almost like you want an AUTOPROTO to go with your AUTOLOAD. Another thing to rub one's 5 o'clock shadow over is that there's no way to apply a prototype to a method call at compile time. And no, I don't want to have the sub howabout ($formal, @arguments) { ... } argument right now. Larry NETaa14422: couldn't take reference of a prototyped function Files patched: op.c (same) NETaa14423: use didn't allow expressions involving the scratch pad From: Graham Barr Files patched: op.c perly.c perly.c.diff perly.y proto.h vms/perly_c.vms Applied suggested patch. NETaa14444: lexical scalar didn't autovivify From: Gurusamy Sarathy Files patched: op.c pp_hot.c It didn't have code in pp_padsv to do the right thing. NETaa14448: caller could dump core when used within an eval or require From: Danny R. Faught Files patched: pp_ctl.c caller() was incorrectly assuming the context stack contained a subroutine context when it in fact contained an eval context. NETaa14451: improved error message on bad pipe filehandle From: Danny R. Faught Files patched: pp_sys.c Now says the slightly more informative Can't use an undefined value as filehandle reference at ./try line 3. NETaa14462: pp_dbstate had a scope leakage on recursion suppression From: Tim Bunce Files patched: pp_ctl.c Swapped the code in question around. NETaa14482: sv_unref freed ref prematurely at times From: Gurusamy Sarathy Files patched: sv.c Made sv_unref() mortalize rather than free the old reference. NETaa14484: appending string to array produced bizarre results From: Greg Ward Also: Malcolm Beattie Files patched: pp_hot.c Will now say, "Can't coerce ARRAY to string". NETaa14525: assignment to globs didn't reset them correctly From: Gurusamy Sarathy Files patched: sv.c Applied parts of patch not overridden by subsequent patch. NETaa14529: a partially matching subpattern could spoof infinity detector From: Wayne Berke Files patched: regexec.c A partial match on a subpattern could fool the infinite regress detector into thinking progress had been made. The previous workaround prevented another bug (NETaa14529) from being fixed, so I've backed it out. I'll need to think more about how to detect failure to progress. I'm still hopeful it's not equivalent to the halting problem. NETaa14535: patches from Gurusamy Sarathy From: Gurusamy Sarathy Files patched: op.c pp.c pp_hot.c regexec.c sv.c toke.c Applied most recent suggested patches. NETaa14537: select() can return too soon From: Matt Kimball Also: Andreas Gustafsson Files patched: pp_sys.c NETaa14538: method calls were treated like do {} under loop modifiers From: Ilya Zakharevich Files patched: perly.c perly.y Needed to take the OPf_SPECIAL flag off of entersubs from method reductions. (It was probably a cut-and-paste error from long ago.) NETaa14540: foreach (@array) no longer does extra stack copy From: darrinm@lmc.com Files patched: Todo op.c pp_ctl.c pp_hot.c Fixed by doing the foreach(@array) optimization, so it iterates directly through the array, and can detect the implicit shift from referencing <>. NETaa14541: new version of perlbug From: Kenneth Albanowski Files patched: README pod/perl.pod utils/perlbug.PL Brought it up to version 1.09. NETaa14541: perlbug 1.11 Files patched: utils/perlbug.PL (same) NETaa14548: magic sets didn't check private OK bits From: W. Bradley Rubenstein Files patched: mg.c The magic code was getting mixed up between private and public POK bits. NETaa14550: made ~ magic magical From: Tim Bunce Files patched: sv.c Applied suggested patch. NETaa14551: humongous header causes infinite loop in format From: Grace Lee Files patched: pp_sys.c Needed to check for page exhaustion after doing top-of-form. NETaa14558: attempt to call undefined top format core dumped From: Hallvard B Furuseth Files patched: pod/perldiag.pod pp_sys.c Now issues an error on attempts to call a non-existent top format. NETaa14561: Gurusamy Sarathy's G_KEEPERR patch From: Andreas Koenig Also: Gurusamy Sarathy Also: Tim Bunce Files patched: cop.h interp.sym perl.c perl.h pp_ctl.c pp_sys.c sv.c toke.c Applied latest patch. NETaa14581: shouldn't execute BEGIN when there are compilation errors From: Rickard Westman Files patched: op.c Perl should not try to execute BEGIN and END blocks if there's been a compilation error. NETaa14582: got SEGV sorting sparse array From: Rick Pluta Files patched: pp_ctl.c Now weeds out undefined values much like Perl 4 did. Now sorts undefined values to the front. NETaa14582: sort was letting unsortable values through to comparison routine Files patched: pp_ctl.c (same) NETaa14585: globs in pad space weren't properly cleaned up From: Gurusamy Sarathy Files patched: op.c pp.c pp_hot.c sv.c Applied suggested patch. NETaa14614: now does dbmopen with perl_eval_sv() From: The Man Files patched: perl.c pp_sys.c proto.h dbmopen now invokes perl_eval_sv(), which should handle error conditions better. NETaa14618: exists doesn't work in GDBM_File From: Andrew Wilcox Files patched: ext/GDBM_File/GDBM_File.xs Applied suggested patch. NETaa14619: tied() From: Larry Wall Also: Paul Marquess Files patched: embed.h global.sym keywords.h keywords.pl opcode.h opcode.pl pp_sys.c toke.c Applied suggested patch. NETaa14636: Jumbo Dynaloader patch From: Tim Bunce Files patched: ext/DynaLoader/DynaLoader.pm 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 Applied suggested patches. NETaa14637: checkcomma routine was stupid about bareword sub calls From: Tim Bunce Files patched: toke.c The checkcomma routine was stupid about bareword sub calls. NETaa14639: (?i) didn't reset on runtime patterns From: Mark A. Scheel Files patched: op.h pp_ctl.c toke.c It didn't distinguish between permanent flags outside the pattern and temporary flags within the pattern. NETaa14649: selecting anonymous globs dumps core From: Chip Salzenberg Files patched: cop.h doio.c embed.h global.sym perl.c pp_sys.c proto.h Applied suggested patch, but reversed the increment and decrement to avoid decrementing and freeing what we're going to increment. NETaa14655: $? returned negative value on AIX From: Kim Frutiger Also: Stephen D. Lee Files patched: pp_sys.c Applied suggested patch. NETaa14668: {2,} could match once From: Hugo van der Sanden Files patched: regexec.c When an internal pattern failed a conjecture, it didn't back off on the number of times it thought it had matched. NETaa14673: open $undefined dumped core From: Samuli K{rkk{inen Files patched: pp_sys.c pp_open() didn't check its argument for globness. NETaa14683: stringifies were running pad out of space From: Robin Barker Files patched: op.h toke.c Increased PADOFFSET to a U32, and made lexer not put double-quoted strings inside OP_STRINGIFY unless they really needed it. NETaa14689: shouldn't have . in @INC when tainting From: William R. Somsky Files patched: perl.c Now does not put . into @INC when tainting. It may still be added with a use lib "."; or, to put it at the end, BEGIN { push(@INC, ".") } but this is not recommended unless a chdir to a known location has been done first. NETaa14690: values inside tainted SVs were ignored From: "James M. Stern" Files patched: pp.c pp_ctl.c It was assuming that a tainted value was a string. NETaa14692: format name required qualification under use strict From: Tom Christiansen Files patched: gv.c Now treats format names the same as subroutine names. NETaa14695: added simple regexp caching From: John Rowe Files patched: pp_ctl.c Applied suggested patch. NETaa14697: regexp comments were sometimes wrongly treated as literal text From: Tom Christiansen Files patched: regcomp.c The literal-character grabber didn't know about extended comments. N.B. '#' is treated as a comment character whenever the /x option is used now, so you can't include '#' as a simple literal in /x regexps. (By the way, Tom, the boxed form of quoting in the previous enclosure is exceeding antisocial when you want to extract the code from it.) NETaa14704: closure got wrong outer scope if outer sub was predeclared From: Marc Paquette Files patched: op.c The outer scope of the anonymous sub was set to the stub rather than to the actual subroutine. I kludged it by making the outer scope of the stub be the actual subroutine, if anything is depending on the stub. NETaa14705: $foo .= $foo did free memory read From: Gerd Knops Files patched: sv.c Now modifies address to copy if it was reallocated. NETaa14709: Chip's FileHandle stuff From: Larry Wall Also: Chip Salzenberg Files patched: MANIFEST ext/FileHandle/FileHandle.pm ext/FileHandle/FileHandle.xs ext/FileHandle/Makefile.PL ext/POSIX/POSIX.pm ext/POSIX/POSIX.pod ext/POSIX/POSIX.xs lib/FileCache.pm lib/Symbol.pm t/lib/filehand.t t/lib/posix.t Applied suggested patches. NETaa14711: added (&) and (*) prototypes for blocks and symbols From: Kenneth Albanowski Files patched: Makefile.SH op.c perly.c perly.h perly.y toke.c & now means that it must have an anonymous sub as that argument. If it's the first argument, the sub may be specified as a block in the indirect object slot, much like grep or sort, which have prototypes of (&@). Also added * so you can do things like sub myopen (*;$); myopen(FOO, $filename); NETaa14713: setuid FROM root now defaults to not do tainting From: Tony Camas Files patched: mg.c perl.c pp_hot.c Applied suggested patch. NETaa14714: duplicate magics could be added to an SV From: Yary Hluchan Files patched: sv.c sv.c The sv_magic() routine didn't properly check to see if it already had a magic of that type. Ordinarily it would have, but it was called during mg_get(), which forces the magic flags off temporarily. NETaa14721: sub defined during erroneous do-FILE caused core dump From: David Campbell Files patched: op.c Fixed the seg fault. I couldn't reproduce the return problem. NETaa14734: ref should never return undef From: Dale Amon Files patched: pp.c t/op/overload.t Now returns null string. NETaa14751: slice of undefs now returns null list From: Tim Bunce Files patched: pp.c pp_hot.c Null list clobberation is now done in lslice, not aassign. NETaa14789: select coredumped on Linux From: Ulrich Kunitz Files patched: pp_sys.c Applied suggested patches, more or less. NETaa14789: straightened out ins and out of duping Files patched: lib/IPC/Open3.pm (same) NETaa14791: implemented internal SUPER class From: Nick Ing-Simmons Also: Dean Roehrich Files patched: gv.c Applied suggested patch. NETaa14845: s/// didn't handle offset strings From: Ken MacLeod Files patched: pp_ctl.c Needed a call to SvOOK_off(targ) in pp_substcont(). NETaa14851: Use of << to mean <<"" is deprecated From: Larry Wall Files patched: toke.c NETaa14865: added HINT_BLOCK_SCOPE to "elsif" From: Jim Avera Files patched: perly.y Needed to set HINT_BLOCK_SCOPE on "elsif" to prevent the do block from being optimized away, which caused the statement transition in elsif to reset the stack too far back. NETaa14876: couldn't delete localized GV safely From: John Hughes Files patched: pp.c scope.c The reference count of the "borrowed" GV needed to be incremented while there was a reference to it in the savestack. NETaa14887: couldn't negate magical scalars From: ian Also: Gurusamy Sarathy Files patched: pp.c Applied suggested patch, more or less. (It's not necessary to test both SvNIOK and SvNIOKp, since the private bits are always set if the public bits are set.) NETaa14893: /m modifier was sticky From: Jim Avera Files patched: pp_ctl.c pp_match() and pp_subst() were using an improperly scoped SAVEINT to restore the value of the internal variable multiline. NETaa14893: /m modifier was sticky Files patched: cop.h pp_hot.c (same) NETaa14916: complete.pl retained old return value From: Martyn Pearce Files patched: lib/complete.pl Applied suggested patch. NETaa14928: non-const 3rd arg to split assigned to list could coredump From: Hans de Graaff Files patched: op.c The optimizer was assuming the OP was an OP_CONST. NETaa14942: substr as lvalue could disable magic From: Darrell Kindred Files patched: pp.c The substr was disabling the magic of $1. NETaa14990: "not" not parseable when expecting term From: "Randal L. Schwartz" Files patched: perly.c perly.c.diff perly.y vms/perly_c.vms The NOTOP production needed to be moved down into the terms. NETaa14993: Bizarre copy of formline From: Tom Christiansen Also: Charles Bailey Files patched: sv.c Applied suggested patch. NETaa14998: sv_add_arena() no longer leaks memory From: Andreas Koenig Files patched: av.c hv.c perl.h sv.c Now keeps one potential arena "on tap", but doesn't use it unless there's demand for SV headers. When an AV or HV is extended, its old memory becomes the next potential arena unless there already is one, in which case it is simply freed. This will have the desired property of not stranding medium-sized chunks of memory when extending a single array repeatedly, but will not degrade when there's no SV demand beyond keeping one chunk of memory on tap, which generally will be about 250 bytes big, since it prefers the earlier freed chunk over the later. See the nice_chunk variable. NETaa14999: $a and $b now protected from use strict and lexical declaration From: Tom Christiansen Files patched: gv.c pod/perldiag.pod toke.c Bare $a and $b are now allowed during "use strict". In addition, the following diag was added: =item Can't use "my %s" in sort comparison (F) The global variables $a and $b are reserved for sort comparisons. You mentioned $a or $b in the same line as the <=> or cmp operator, and the variable had earlier been declared as a lexical variable. Either qualify the sort variable with the package name, or rename the lexical variable. NETaa15034: use strict refs should allow calls to prototyped functions From: Roderick Schertler Files patched: perly.c perly.c.diff perly.y toke.c vms/perly_c.vms Applied patch suggested by Chip. NETaa15083: forced $AUTOLOAD to be untainted From: Tim Bunce Files patched: gv.c pp_hot.c Stripped any taintmagic from $AUTOLOAD after setting it. NETaa15084: patch for Term::Cap From: Mark Kaehny Also: Hugo van der Sanden Files patched: lib/Term/Cap.pm Applied suggested patch. NETaa15086: null pattern could cause coredump in s//_$1_/ From: "Paul E. Maisano" Files patched: cop.h pp_ctl.c If the replacement pattern was complicated enough to cause pp_substcont to be called, then it lost track of which REGEXP* it was supposed to be using. NETaa15087: t/io/pipe.t didn't work on AIX From: Andy Dougherty Files patched: t/io/pipe.t Applied suggested patch. NETaa15088: study was busted From: Hugo van der Sanden Files patched: opcode.h opcode.pl pp.c It was studying its scratch pad target rather than the argument supplied. NETaa15090: MSTATS patch From: Tim Bunce Files patched: global.sym malloc.c perl.c perl.h proto.h Applied suggested patch. NETaa15098: longjmp out of magic leaks memory From: Chip Salzenberg Files patched: mg.c sv.c Applied suggested patch. NETaa15102: getpgrp() is broken if getpgrp2() is available From: Roderick Schertler Files patched: perl.h pp_sys.c Applied suggested patch. NETaa15103: prototypes leaked opcodes From: Chip Salzenberg Files patched: op.c Applied suggested patch. NETaa15107: quotameta memory bug on all metacharacters From: Chip Salzenberg Files patched: pp.c Applied suggested patch. NETaa15108: Fix for incomplete string leak From: Chip Salzenberg Files patched: toke.c Applied suggested patch. NETaa15110: couldn't use $/ with 8th bit set on some architectures From: Chip Salzenberg Files 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 Applied suggested patches. NETaa15112: { a_1 => 2 } didn't parse as expected From: Stuart M. Weinstein Files patched: toke.c The little dwimmer was only skipping ALPHA rather than ALNUM chars. NETaa15123: bitwise ops produce spurious warnings From: Hugo van der Sanden Also: Chip Salzenberg Also: Andreas Gustafsson Files patched: sv.c Decided to suppress the warning in the conversion routines if merely converting a temporary, which can never be a user-supplied value anyway. NETaa15129: #if defined (foo) misparsed in h2ph From: Roderick Schertler Files patched: utils/h2ph.PL Applied suggested patch. NETaa15131: some POSIX functions assumed valid filehandles From: Chip Salzenberg Files patched: ext/POSIX/POSIX.xs Applied suggested patch. NETaa15151: don't optimize split on OPpASSIGN_COMMON From: Huw Rogers Files patched: op.c Had to swap the optimization down to after the assignment op is generated and COMMON is calculated, and then clean up the resultant tree differently. NETaa15154: MakeMaker-5.18 From: Andreas Koenig Files patched: MANIFEST lib/ExtUtils/Liblist.pm lib/ExtUtils/MM_VMS.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/Mksymlists.pm Brought it up to 5.18. NETaa15156: some Exporter tweaks From: Roderick Schertler Also: Tim Bunce Files patched: lib/Exporter.pm Also did Tim's Tiny Trivial patch. NETaa15157: new version of Test::Harness From: Andreas Koenig Files patched: lib/Test/Harness.pm Applied suggested patch. NETaa15175: overloaded nomethod has garbage 4th op From: Ilya Zakharevich Files patched: gv.c Applied suggested patch. NETaa15179: SvPOK_only shouldn't back off on offset pointer From: Gutorm.Hogasen@oslo.teamco.telenor.no Files patched: sv.h SvPOK_only() was calling SvOOK_off(), which adjusted the string pointer after tr/// has already acquired it. It shouldn't really be necessary for SvPOK_only() to undo an offset string pointer, since there's no conflict with a possible integer value where the offset is stored. NETaa15193: & now always bypasses prototype checking From: Larry Wall Files 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 Turned out to be a big hairy deal because the lexer turns foo() into &foo(). But it works consistently now. Also fixed pod. NETaa15197: 5.002b2 is 'appending' to $@ From: Gurusamy Sarathy Files patched: pp_ctl.c Applied suggested patch. NETaa15201: working around Linux DBL_DIG problems From: Kenneth Albanowski Files patched: hints/linux.sh sv.c Applied suggested patch. NETaa15208: SelectSaver From: Chip Salzenberg Files patched: MANIFEST lib/SelectSaver.pm Applied suggested patch. NETaa15209: DirHandle From: Chip Salzenberg Files patched: MANIFEST lib/DirHandle.pm t/lib/dirhand.t NETaa15210: sysopen() From: Chip Salzenberg Files patched: doio.c keywords.pl lib/ExtUtils/typemap opcode.pl pod/perlfunc.pod pp_hot.c pp_sys.c proto.h toke.c Applied suggested patch. Hope it works... NETaa15211: use mnemonic names in Safe setup From: Chip Salzenberg Files patched: ext/Safe/Safe.pm Applied suggested patch, more or less. NETaa15214: prototype() From: Chip Salzenberg Files patched: ext/Safe/Safe.pm global.sym keywords.pl opcode.pl pp.c toke.c Applied suggested patch. NETaa15217: -w problem with -d:foo From: Tim Bunce Files patched: perl.c Applied suggested patch. NETaa15218: *GLOB{ELEMENT} From: Larry Wall Files patched: Makefile.SH embed.h ext/Safe/Safe.pm keywords.h opcode.h opcode.h opcode.pl perly.c perly.c.diff perly.y pp_hot.c t/lib/safe.t vms/perly_c.vms NETaa15219: Make *x=\*y do like *x=*y From: Chip Salzenberg Files patched: sv.c Applied suggested patch. NETaa15221: Indigestion with Carp::longmess and big eval '...'s From: Tim Bunce Files patched: lib/Carp.pm Applied suggested patch. NETaa15222: VERSION patch for standard extensions From: Paul Marquess Files patched: ext/DB_File/Makefile.PL ext/DynaLoader/DynaLoader.pm ext/DynaLoader/Makefile.PL ext/Fcntl/Fcntl.pm ext/Fcntl/Makefile.PL ext/GDBM_File/GDBM_File.pm ext/GDBM_File/Makefile.PL ext/NDBM_File/Makefile.PL ext/NDBM_File/NDBM_File.pm ext/ODBM_File/Makefile.PL ext/ODBM_File/ODBM_File.pm ext/POSIX/Makefile.PL ext/POSIX/POSIX.pm ext/SDBM_File/Makefile.PL ext/SDBM_File/SDBM_File.pm ext/Safe/Makefile.PL ext/Safe/Safe.pm ext/Socket/Makefile.PL Applied suggested patch. NETaa15222: VERSION patch for standard extensions (reprise) Files patched: ext/DB_File/DB_File.pm ext/DynaLoader/DynaLoader.pm ext/Fcntl/Fcntl.pm ext/GDBM_File/GDBM_File.pm ext/NDBM_File/NDBM_File.pm ext/ODBM_File/ODBM_File.pm ext/POSIX/POSIX.pm ext/SDBM_File/SDBM_File.pm ext/Safe/Safe.pm ext/Socket/Socket.pm (same) NETaa15227: $i < 10000 should optimize to integer op From: Larry Wall Files patched: op.c op.c The program for ($i = 0; $i < 100000; $i++) { push @foo, $i; } takes about one quarter the memory if the optimizer decides that it can use an integer < comparison rather than floating point. It now does so if one side is an integer constant and the other side a simple variable. This should really help some of our benchmarks. You can still force a floating point comparison by using 100000.0 instead. NETaa15228: CPerl-mode patch From: Ilya Zakharevich Files patched: emacs/cperl-mode.el Applied suggested patch. NETaa15231: Symbol::qualify() From: Chip Salzenberg Files patched: ext/FileHandle/FileHandle.pm gv.c lib/SelectSaver.pm lib/Symbol.pm pp_hot.c Applied suggested patch. NETaa15236: select select broke under use strict From: Chip Salzenberg Files patched: op.c Instead of inventing a new bit, I just turned off the HINT_STRICT_REFS bit. I don't think it's worthwhile distinguishing between qualified or unqualified names to select. NETaa15237: use vars From: Larry Wall Files patched: MANIFEST gv.c lib/subs.pm lib/vars.pm sv.c NETaa15240: keep op names _and_ descriptions From: Chip Salzenberg Files patched: doio.c embed.h ext/Safe/Safe.pm ext/Safe/Safe.xs global.sym op.c opcode.h opcode.pl scope.c sv.c Applied suggested patch. NETaa15259: study doesn't unset on string modification From: Larry Wall Files patched: mg.c pp.c Piggybacked on m//g unset magic to unset the study too. NETaa15276: pick a better initial cxstack_max From: Chip Salzenberg Files patched: perl.c Added fudge in, and made it calculate how many it could fit into (most of) 8K, to avoid getting 16K of Kingsley malloc. NETaa15287: numeric comparison optimization adjustments From: Clark Cooper Files patched: op.c Applied patch suggested by Chip, with liberalization to >= and <=. NETaa15299: couldn't eval string containing pod or __DATA__ From: Andreas Koenig Also: Gisle Aas Files patched: toke.c Basically, eval didn't know how to bypass pods correctly. NETaa15300: sv_backoff problems From: Paul Marquess Also: mtr Also: Chip Salzenberg Files patched: op.c sv.c sv.h Applied suggested patch. NETaa15312: Avoid fclose(NULL) From: Chip Salzenberg Files patched: toke.c Applied suggested patch. NETaa15318: didn't set up perl_init_i18nl14n for export From: Ilya Zakharevich Files patched: perl_exp.SH Applied suggested patch. NETaa15331: File::Path::rmtree followed symlinks From: Andreas Koenig Files patched: lib/File/Path.pm Added suggested patch, except I did if (not -l $root and -d _) { for efficiency, since if -d is true, the -l already called lstat on it. NETaa15339: sv_gets() didn't reset count From: alanburlison@unn.unisys.com Files patched: sv.c Applied suggested patch. NETaa15341: differentiated importation of different types From: Chip Salzenberg Files patched: gv.c gv.h op.c perl.c pp.c pp_ctl.c sv.c sv.h toke.c Applied suggested patch. NETaa15342: Consistent handling of e_{fp,tmpname} From: Chip Salzenberg Files patched: perl.c pp_ctl.c util.c Applied suggested patch. NETaa15344: Safe gets confused about malloc on AIX From: Tim Bunce Files patched: ext/Safe/Safe.xs Applied suggested patch. NETaa15348: -M upgrade From: Tim Bunce Files patched: perl.c pod/perlrun.pod Applied suggested patch. NETaa15369: change in split optimization broke scalar context From: Ulrich Pfeifer Files patched: op.c The earlier patch to make the split optimization pay attention to OPpASSIGN_COMMON rearranged how the syntax tree is constructed, but kept the wrong context flags. This causes pp_split() do do the wrong thing. NETaa15423: can't do subversion numbering because of %5.3f assumptions From: Andy Dougherty Files patched: configpm patchlevel.h perl.c perl.h pp_ctl.c Removed the %5.3f assumptions where appropriate. patchlevel.h now defines SUBVERSION, which if greater than 0 indicates a development version. NETaa15424: Sigsetjmp patch From: Kenneth Albanowski Files patched: Configure config_h.SH op.c perl.c perl.h pp_ctl.c util.c Applied suggested patch. Needed to make install paths absolute. Files patched: installperl h2xs 1.14 Files patched: utils/h2xs.PL makedir() looped on a symlink to a directory. Files patched: installperl xsubpp 1.932 Files patched: lib/ExtUtils/xsubpp ------------- Version 5.001 ------------- Nearly all the changes for 5.001 were bug fixes of one variety or another, so here's the bug list, along with the "resolution" for each of them. If you wish to correspond about any of them, please include the bug number. There were a few that can be construed as enhancements: NETaa13059: now warns of use of \1 where $1 is necessary. NETaa13512: added $SIG{__WARN__} and $SIG{__DIE__} hooks NETaa13520: added closures NETaa13530: scalar keys now resets hash iterator NETaa13641: added Tim's fancy new import whizbangers NETaa13710: cryptswitch needed to be more "useable" NETaa13716: Carp now allows multiple packages to be skipped out of NETaa13716: now counts imported routines as "defined" for redef warnings (and, of course, much of the stuff from the perl5-porters) NETaa12974: README incorrectly said it was a pre-release. Files patched: README NETaa13033: goto pushed a bogus scope on the context stack. From: Steve Vinoski Files patched: pp_ctl.c The goto operator pushed an extra bogus scope onto the context stack. (This often didn't matter, since many things pop extra unrecognized scopes off.) NETaa13034: tried to get valid pointer from undef. From: Castor Fu Also: Achille Hui, the Day Dreamer Also: Eric Arnold Files patched: pp_sys.c Now treats undef specially, and calls SvPV_force on any non-numeric scalar value to get a real pointer to somewhere. NETaa13035: included package info with filehandles. From: Jack Shirazi - BIU Files patched: pp_hot.c pp_sys.c Now passes a glob to filehandle methods to keep the package info intact. NETaa13048: didn't give strict vars message on every occurrence. From: Doug Campbell Files patched: gv.c It now complains about every occurrence. (The bug resulted from an ill-conceived attempt to suppress a duplicate error message in a suboptimal fashion.) NETaa13052: test for numeric sort sub return value fooled by taint magic. From: Peter Jaspers-Fayer Files patched: pp_ctl.c sv.h The test to see if the sort sub return value was numeric looked at the public flags rather than the private flags of the SV, so taint magic hid that info from the sort. NETaa13053: forced a2p to use byacc From: Andy Dougherty Files patched: MANIFEST x2p/Makefile.SH x2p/a2p.c a2p.c is now pre-byacced and shipped with the kit. NETaa13055: misnamed constant in previous patch. From: Conrad Augustin Files patched: op.c op.h toke.c The tokener translates $[ to a constant, but with a special marking in case the constant gets assigned to or localized. Unfortunately, the marking was done with a combination of OPf_SPECIAL and OPf_MOD that was easily spoofed. There is now a private OPpCONST_ARYLEN flag for this purpose. NETaa13055: use of OPf_SPECIAL for $[ lvaluehood was too fragile. Files patched: op.c op.h toke.c (same) NETaa13056: convert needs to throw away any number info on its list. From: Jack Shirazi - BIU Files patched: op.c The listiness of the argument list leaked out to the subroutine call because of how prepend_elem and append_elem reuse an existing list. The convert() routine just needs to discard any listiness it finds on its argument. NETaa13058: AUTOLOAD shouldn't assume size of @_ is meaningful. From: Florent Guillaume Files patched: ext/DB_File/DB_File.pm ext/Fcntl/Fcntl.pm ext/GDBM_File/GDBM_File.pm ext/Socket/Socket.pm h2xs.SH I just deleted the optimization, which is silly anyway since the eventual subroutine definition is cached. NETaa13059: now warns of use of \1 where $1 is necessary. From: Gustaf Neumann Files patched: toke.c Now says Can't use \1 to mean $1 in expression at foo line 2 along with an explanation in perldiag. NETaa13060: no longer warns on attempt to read <> operator's transition state. From: Chaim Frenkel Files patched: pp_hot.c No longer warns on <> operator's transitional state. NETaa13140: warning said $ when @ would be more appropriate. From: David J. MacKenzie Files patched: op.c pod/perldiag.pod Now says (Did you mean $ or @ instead of %?) and added more explanation to perldiag. NETaa13149: was reading freed memory to make incorrect error message. Files patched: pp_ctl.c It was reading freed memory to make an error message that would be incorrect in any event because it had the inner filename rather than the outer. NETaa13149: confess was sometimes less informative than croak From: Jack Shirazi Files patched: lib/Carp.pm (same) NETaa13150: stderr needs to be STDERR in package From: Jack Shirazi Files patched: lib/File/CheckTree.pm Also fixed pl2pm to translate the filehandles to uppercase. NETaa13150: uppercases stdin, stdout and stderr Files patched: pl2pm (same) NETaa13154: array assignment didn't notice package magic. From: Brian Reichert Files patched: pp_hot.c The list assignment operator looked for only set magic, but set magic is only on the elements of a magical hash, not on the hash as a whole. I made the operator look for any magic at all on the target array or hash. NETaa13155: &DB::DB left trash on the stack. From: Thomas Koenig Files patched: lib/perl5db.pl pp_ctl.c The call by pp_dbstate() to &DB::DB left trash on the stack. It now calls DB in list context, and DB returns (). NETaa13156: lexical variables didn't show up in debugger evals. From: Joergen Haegg Files patched: op.c The code that searched back up the context stack for the lexical scope outside the eval only partially took into consideration that there might be extra debugger subroutine frames that shouldn't be used, and ended up comparing the wrong statement sequence number to the range of valid sequence numbers for the scope of the lexical variable. (There was also a bug fixed in passing that caused the scope of lexical to go clear to the end of the subroutine even if it was within an inner block.) NETaa13157: any request for autoloaded DESTROY should create a null one. From: Tom Christiansen Files patched: lib/AutoLoader.pm If DESTROY.al is not located, it now creates sub DESTROY {} automatically. NETaa13158: now preserves $@ around destructors while leaving eval. From: Tim Bunce Files patched: pp_ctl.c Applied supplied patch, except the whole second hunk can be replaced with sv_insert(errsv, 0, 0, message, strlen(message)); NETaa13160: clarified behavior of split without arguments From: Harry Edmon Files patched: pod/perlfunc.pod Clarified the behavior of split without arguments. NETaa13162: eval {} lost list/scalar context From: Dov Grobgeld Files patched: op.c LEAVETRY didn't propagate number to ENTERTRY. NETaa13163: clarified documentation of foreach using my variable From: Tom Christiansen Files patched: pod/perlsyn.pod Explained that foreach using a lexical is still localized. NETaa13164: the dot detector for the end of formats was over-rambunctious. From: John Stoffel Files patched: toke.c The dot detector for the end of formats was over-rambunctious. It would pick up any dot that didn't have a space in front of it. NETaa13165: do {} while 1 never linked outer block into next chain. From: Gisle Aas Files patched: op.c When the conditional of do {} while 1; was optimized away, it confused the postfix order construction so that the block that ordinarily sits around the whole loop was never executed. So when the loop tried to unstack between iterations, it got the wrong context, and blew away the lexical variables of the outer scope. Fixed it by introducing a NULL opcode that will be optimized away later. NETaa13167: coercion was looking at public bits rather than private bits. From: Randal L. Schwartz Also: Thomas Riechmann Also: Shane Castle Files patched: sv.c There were some bad ifdefs around the various varieties of set*id(). In addition, tainting was interacting badly with assignment to $> because sv_2iv() was examining SvPOK rather than SvPOKp, and so couldn't coerce a string uid to an integer one. NETaa13167: had some ifdefs wrong on set*id. Files patched: mg.c pp_hot.c (same) NETaa13168: relaxed test for comparison of new and old fds From: Casper H.S. Dik Files patched: t/lib/posix.t I relaxed the comparison to just check that the new fd is greater. NETaa13169: autoincrement can corrupt scalar value state. From: Gisle Aas Also: Tom Christiansen Files patched: sv.c It assumed a PV didn't need to be upgraded to become an NV. NETaa13169: previous patch could leak a string pointer. Files patched: sv.c (same) NETaa13170: symbols missing from global.sym From: Tim Bunce Files patched: global.sym Applied suggested patch. NETaa13171: \\ in <<'END' shouldn't reduce to \. From: Randal L. Schwartz Files patched: toke.c <<'END' needed to bypass ordinary single-quote processing. NETaa13172: 'use integer' turned off magical autoincrement. From: Erich Rickheit KSC Files patched: pp.c pp_hot.c The integer versions of the increment and decrement operators were trying too hard to be efficient. NETaa13172: deleted duplicate increment and decrement code Files patched: opcode.h opcode.pl pp.c (same) NETaa13173: install should make shared libraries executable. From: Brian Grossman Also: Dave Nadler Also: Eero Pajarre Files patched: installperl Now gives permission 555 to any file ending with extension specified by $dlext. NETaa13176: ck_rvconst didn't free the const it used up. From: Nick Duffek Files patched: op.c I checked in many random memory leaks under this bug number, since it was an eval that brought many of them out. NETaa13176: didn't delete XRV for temp ref of destructor. Files patched: sv.c (same) NETaa13176: didn't delete op_pmshort in matching operators. Files patched: op.c (same) NETaa13176: eval leaked the name of the eval. Files patched: scope.c (same) NETaa13176: gp_free didn't free the format. Files patched: gv.c (same) NETaa13176: minor leaks in loop exits and constant subscript optimization. Files patched: op.c (same) NETaa13176: plugged some duplicate struct allocation memory leaks. Files patched: perl.c (same) NETaa13176: sv_clear of an FM didn't clear anything. Files patched: sv.c (same) NETaa13176: tr/// didn't mortalize its return value. Files patched: pp.c (same) NETaa13177: SCOPE optimization hid line number info From: David J. MacKenzie Also: Hallvard B Furuseth Files patched: op.c Every pass on the syntax tree has to keep track of the current statement. Unfortunately, the single-statement block was optimized into a single statement between the time the variable was parsed and the time the void code scan was done, so that pass didn't see the OP_NEXTSTATE operator, because it has been optimized to an OP_NULL. Fortunately, null operands remember what they were, so it was pretty easy to make it set the correct line number anyway. NETaa13178: some linux doesn't handle nm well From: Alan Modra Files patched: hints/linux.sh Applied supplied patch. NETaa13180: localized slice now pre-extends array From: Larry Schuler Files patched: pp.c A localized slice now pre-extends its array to avoid reallocation during the scope of the local. NETaa13181: m//g didn't keep track of whether previous match matched null. From: "philippe.verdret" Files patched: mg.h pp_hot.c A pattern isn't allowed to match a null string in the same place twice in a row. m//g wasn't keeping track of whether the previous match matched the null string. NETaa13182: now includes whitespace as a regexp metacharacter. From: Larry Wall Files patched: toke.c scan_const() now counts " \t\n\r\f\v" as metacharacters when scanning a pattern. NETaa13183: sv_setsv shouldn't try to clone an object. From: Peter Gordon Files patched: sv.c The sv_mortalcopy() done by the return in STORE called sv_setsv(), which cloned the object. sv_setsv() shouldn't be in the business of cloning objects. NETaa13184: bogus warning on quoted signal handler name removed. From: Dan Carson Files patched: toke.c Now doesn't complain unless the first non-whitespace character after the = is an alphabetic character. NETaa13186: now croaks on chop($') From: Casper H.S. Dik Files patched: doop.c Now croaks on chop($') and such. NETaa13187: "${foo::bar}" now counts as mere delimitation, not as a bareword. From: Jay Rogers Files patched: toke.c "${foo::bar}" now counts as mere delimitation, not as a bareword inside a reference block. NETaa13188: for backward compatibility, looks for "perl -" before "perl". From: Russell Mosemann Files patched: toke.c Now allows non-whitespace characters on the #! line between the "perl" and the "-". NETaa13188: now allows non-whitespace after #!...perl before switches. Files patched: toke.c (same) NETaa13189: derivative files need to be removed before recreation From: Simon Leinen Also: Dick Middleton Also: David J. MacKenzie Files patched: embed_h.sh x2p/Makefile.SH Fixed various little nits as suggested in several messages. NETaa13190: certain assignments can spoof pod directive recognizer From: Ilya Zakharevich Files patched: toke.c The lexer now only recognizes pod directives where a statement is expected. NETaa13194: now returns undef when there is no curpm. From: lusol@Dillon.CC.Lehigh.EDU Files patched: mg.c Since there was no regexp prior to the "use", it was returning whatever the last successful match was within the "use", because there was no current regexp, so it treated it as a normal variable. It now returns undef. NETaa13195: semop had one S too many. From: Joachim Huober Files patched: opcode.pl The entry in opcode.pl had one too many S's. NETaa13196: always assumes it's a Perl script if -c is used. From: Dan Carson Files patched: toke.c It now will assume it's a Perl script if the -c switch is used. NETaa13197: changed implicit -> message to be more understandable. From: Bruce Barnett Files patched: op.c pod/perldiag.pod I changed the error message to be more understandable. It now says Can't use subscript on sort... NETaa13201: added OPpCONST_ENTERED flag to properly enter filehandle symbols. From: E. Jay Berkenbilt Also: Tom Christiansen Files patched: op.c op.h toke.c The grammatical reduction of a print statement didn't properly count the filehandle as a symbol reference because it couldn't distinguish between a symbol entered earlier in the program and a symbol entered for the first time down in the lexer. NETaa13203: README shouldn't mention uperl.o any more. From: Anno Siegel Files patched: README NETaa13204: .= shouldn't warn on uninitialized target. From: Pete Peterson Files patched: pp_hot.c No longer warns on uninitialized target of .= operator. NETaa13206: handy macros in XSUB.h From: Tim Bunce Files patched: XSUB.h Added suggested macros. NETaa13228: commonality checker didn't treat lexicals as variables. From: mcook@cognex.com Files patched: op.c opcode.pl The list assignment operator tries to avoid unnecessary copies by doing the assignment directly if there are no common variables on either side of the equals. Unfortunately, the code that decided that only recognized references to dynamic variables, not lexical variables. NETaa13229: fixed sign stuff for complement, integer coercion. From: Larry Wall Files patched: perl.h pp.c sv.c Fixed ~0 and integer coercions. NETaa13230: no longer tries to reuse scratchpad temps if tainting in effect. From: Luca Fini Files patched: op.c I haven't reproduced it, but I believe the problem is the reuse of scratchpad temporaries between statements. I've made it not try to reuse them if tainting is in effect. NETaa13231: *foo = *bar now prevents typo warnings on "foo" From: Robin Barker Files patched: sv.c Aliasing of the form *foo = *bar is now protected from the typo warnings. Previously only the *foo = \$bar form was. NETaa13235: require BAREWORD now introduces package name immediately. From: Larry Wall Files patched: toke.c require BAREWORD now introduces package name immediately. This lets the method intuit code work right even though the require hasn't actually run yet. NETaa13289: didn't calculate correctly using arybase. From: Jared Rhine Files patched: pp.c pp_hot.c The runtime code didn't use curcop->cop_arybase correctly. NETaa13301: store now throws exception on error From: Barry Friedman Files 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 Changed warn to croak in ext/*DBM_File/*.xs. NETaa13302: ctime now takes Time_t rather than Time_t*. From: Rodger Anderson Files patched: ext/POSIX/POSIX.xs Now declares a Time_t and takes the address of that in CODE. NETaa13302: shorter way to do this patch Files patched: ext/POSIX/POSIX.xs (same) NETaa13304: could feed too large $@ back into croak, whereupon it croaked. From: Larry Wall Files patched: perl.c callist() could feed $@ back into croak with more than a bare %s. (croak() handles long strings with a bare %s okay.) NETaa13305: compiler misoptimized RHS to outside of s/a/print/e From: Brian S. Cashman Files patched: op.c The syntax tree was being misconstructed because the compiler felt that the RHS was invariant, so it did it outside the s///. NETaa13314: assigning mortal to lexical leaks From: Larry Wall Files patched: sv.c In stealing strings, sv_setsv was checking SvPOK to see if it should free the destination string. It should have been checking SvPVX. NETaa13316: wait4pid now recalled when errno == EINTR From: Robert J. Pankratz Files patched: pp_sys.c util.c system() and the close() of a piped open now recall wait4pid if it returned prematurely with errno == EINTR. NETaa13329: needed to localize taint magic From: Brian Katzung Files patched: sv.c doio.c mg.c pp_hot.c pp_sys.c scope.c taint.c Taint magic is now localized better, though I had to resort to a kludge to allow a value to be both tainted and untainted simultaneously during the assignment of local $foo = $_[0]; when $_[0] is a reference to the variable $foo already. NETaa13341: clarified interaction of AnyDBM_File::ISA and "use" From: Ian Phillipps Files patched: pod/modpods/AnyDBMFile.pod The doc was misleading. NETaa13342: grep and map with block would enter block but never leave it. From: Ian Phillipps Files patched: op.c The compiler use some sort-checking code to handle the arguments of grep and map. Unfortunately, this wiped out the block exit opcode while leaving the block entry opcode. This doesn't matter to sort, but did matter to grep and map. It now leave the block entry intact. The reason it worked without the my is because the block entry and exit were optimized away to an OP_SCOPE, which it doesn't matter if it's there or not. NETaa13343: goto needed to longjmp when in a signal handler. From: Robert Partington Files patched: pp_ctl.c goto needed to longjmp() when in a signal handler to get back into the right run() context. NETaa13344: strict vars shouldn't apply to globs or filehandles. From: Andrew Wilcox Files patched: gv.c Filehandles and globs will be excepted from "strict vars", so that you can do the standard Perl 4 trick of use strict; sub foo { local(*IN); open(IN,"file"); } NETaa13345: assert.pl didn't use package DB From: Hans Mulder Files patched: lib/assert.pl Now it does. NETaa13348: av_undef didn't free scalar representing $#foo. From: David Filo Files patched: av.c av_undef didn't free scalar representing $#foo. NETaa13349: sort sub accumulated save stack entries From: David Filo Files patched: pp_ctl.c COMMON only gets set if assigning to @_, which is reasonable. Most of the problem was a memory leak. NETaa13351: didn't treat indirect filehandles as references. From: Andy Dougherty Files patched: op.c Now produces Can't use an undefined value as a symbol reference at ./foo line 3. NETaa13352: OP_SCOPE allocated as UNOP rather than LISTOP. From: Andy Dougherty Files patched: op.c NETaa13353: scope() didn't release filegv on OP_SCOPE optimization. From: Larry Wall Files patched: op.c When scope() nulled out a NEXTSTATE, it didn't release its filegv reference. NETaa13355: hv_delete now avoids useless mortalcopy From: Larry Wall Files patched: hv.c op.c pp.c pp_ctl.c proto.h scope.c util.c hv_delete now avoids useless mortalcopy. NETaa13359: comma operator section missing its heading From: Larry Wall Files patched: pod/perlop.pod NETaa13359: random typo Files patched: pod/perldiag.pod NETaa13360: code to handle partial vec values was bogus. From: Conrad Augustin Files patched: pp.c The code that Mark J. added a long time ago to handle values that were partially off the end of the string was incorrect. NETaa13361: made it not interpolate inside regexp comments From: Martin Jost Files patched: toke.c To avoid surprising people, it no longer interpolates inside regexp comments. NETaa13362: ${q[1]} should be interpreted like it used to From: Hans Mulder Files patched: toke.c Now resolves ${keyword[1]} to $keyword[1] and warns if -w. Likewise for {}. NETaa13363: meaning of repeated search chars undocumented in tr/// From: Stephen P. Potter Files patched: pod/perlop.pod Documented that repeated characters use the first translation given. NETaa13365: if closedir fails, don't try it again. From: Frank Crawford Files patched: pp_sys.c Now does not attempt to closedir a second time. NETaa13366: can't do block scope optimization on $1 et al when tainting. From: Andrew Vignaux Files patched: toke.c The tainting mechanism assumes that every statement starts out untainted. Unfortunately, the scope removal optimization for very short blocks removed the statementhood of statements that were attempting to read $1 as an untainted value, with the effect that $1 appeared to be tainted anyway. The optimization is now disabled when tainting and the block contains $1 (or equivalent). NETaa13366: fixed this a better way in toke.c. Files patched: op.c (same) NETaa13366: need to disable scope optimization when tainting. Files patched: op.c (same) NETaa13367: Did a SvCUR_set without nulling out final char. From: "Rob Henderson" Files patched: doop.c pp.c pp_sys.c When do_vop set the length on its result string it neglected to null-terminate it. NETaa13368: bigrat::norm sometimes chucked sign From: Greg Kuperberg Files patched: lib/bigrat.pl The normalization routine was assuming that the gcd of two numbers was never negative, and based on that assumption managed to move the sign to the denominator, where it was deleted on the assumption that the denominator is always positive. NETaa13368: botched previous patch Files patched: lib/bigrat.pl (same) NETaa13369: # is now a comment character, and \# should be left for regcomp. From: Simon Parsons Files patched: toke.c It was not skipping the comment when it skipped the white space, and constructed an opcode that tried to match a null string. Unfortunately, the previous star tried to use the first character of the null string to optimize where to recurse, so it never matched. NETaa13369: comment after regexp quantifier induced non-match. Files patched: regcomp.c (same) NETaa13370: some code assumed SvCUR was of type int. From: Spider Boardman Files patched: pp_sys.c Did something similar to the proposed patch. I also fixed the problem that it assumed the type of SvCUR was int. And fixed get{peer,sock}name the same way. NETaa13375: sometimes dontbother wasn't added back into strend. From: Jamshid Afshar Files patched: regexec.c When the /g modifier was used, the regular expression code would calculate the end of $' too short by the minimum number of characters the pattern could match. NETaa13375: sv_setpvn now disallows negative length. Files patched: sv.c (same) NETaa13376: suspected indirect objecthood prevented recognition of lexical. From: Gisle.Aas@nr.no Files patched: toke.c When $data[0] is used in a spot that might be an indirect object, the lexer was getting confused over the rule that says the $data in $$data[0] isn't an array element. (The lexer uses XREF state for both indirect objects and for variables used as names.) NETaa13377: -I processesing ate remainder of #! line. From: Darrell Schiebel Files patched: perl.c I made the -I processing in moreswitches look for the end of the string, delimited by whitespace. NETaa13379: ${foo} now treated the same outside quotes as inside From: Hans Mulder Files patched: toke.c ${bareword} is now treated the same outside quotes as inside. NETaa13379: previous fix for this bug was botched Files patched: toke.c (same) NETaa13381: TEST should check for perl link From: Andy Dougherty Files patched: t/TEST die "You need to run \"make test\" first to set things up.\n" unless -e 'perl'; NETaa13384: fixed version 0.000 botch. From: Larry Wall Files patched: installperl NETaa13385: return 0 from required file loses message From: Malcolm Beattie Files patched: pp_ctl.c Works right now. NETaa13387: added pod2latex From: Taro KAWAGISHI Files patched: MANIFEST pod/pod2latex Added most recent copy to pod directory. NETaa13388: constant folding now prefers integer results over double From: Ilya Zakharevich Files patched: op.c Constant folding now prefers integer results over double. NETaa13389: now treats . and exec as shell metathingies From: Hans Mulder Files patched: doio.c Now treats . and exec as shell metathingies. NETaa13395: eval didn't check taintedness. From: Larry Wall Files patched: pp_ctl.c NETaa13396: $^ coredumps at end of string From: Paul Rogers Files patched: toke.c The scan_ident() didn't check for a null following $^. NETaa13397: improved error messages when operator expected From: Larry Wall Files patched: toke.c Added message (Do you need to predeclare BAR?). Also fixed the missing semicolon message. NETaa13399: cleanup by Andy From: Larry Wall Files patched: Changes Configure Makefile.SH README cflags.SH config.H config_h.SH deb.c doop.c dump.c ext/DB_File/DB_File.pm ext/DB_File/DB_File.xs ext/DynaLoader/DynaLoader.pm ext/Fcntl/Fcntl.pm ext/GDBM_File/GDBM_File.pm ext/POSIX/POSIX.pm ext/SDBM_File/sdbm/sdbm.h ext/Socket/Socket.pm ext/util/make_ext h2xs.SH hints/aix.sh hints/bsd386.sh hints/dec_osf.sh hints/esix4.sh hints/freebsd.sh hints/irix_5.sh hints/next_3_2.sh hints/sunos_4_1.sh hints/svr4.sh hints/ultrix_4.sh installperl lib/AutoSplit.pm lib/Cwd.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/xsubpp lib/Term/Cap.pm 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 NETaa13399: cleanup from Andy Files patched: MANIFEST NETaa13399: configuration cleanup Files patched: Configure Configure MANIFEST MANIFEST Makefile.SH Makefile.SH README config.H config.H config_h.SH config_h.SH configpm ext/DynaLoader/DynaLoader.pm ext/DynaLoader/dl_hpux.xs ext/NDBM_File/Makefile.PL ext/ODBM_File/Makefile.PL ext/util/make_ext handy.h hints/aix.sh hints/hpux_9.sh hints/hpux_9.sh hints/irix_4.sh hints/linux.sh hints/mpeix.sh hints/next_3_2.sh hints/solaris_2.sh hints/svr4.sh installperl installperl lib/AutoSplit.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/xsubpp lib/Getopt/Long.pm lib/Text/Tabs.pm 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 NETaa13399: new files from Andy Files 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/convexos.sh hints/irix_6.sh NETaa13399: patch0l from Andy Files 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/next_3_2.sh hints/solaris_2.sh hints/unicos.sh installperl lib/Cwd.pm lib/ExtUtils/MakeMaker.pm makeaperl.SH vms/config.vms x2p/util.c x2p/util.h NETaa13399: stuff from Andy Files patched: Configure MANIFEST Makefile.SH configpm hints/dec_osf.sh hints/linux.sh hints/machten.sh lib/ExtUtils/MakeMaker.pm util.c NETaa13399: Patch 0k from Andy Files patched: Configure MANIFEST Makefile.SH config.H config_h.SH hints/dec_osf.sh hints/mpeix.sh hints/next_3_0.sh hints/ultrix_4.sh installperl lib/ExtUtils/MakeMaker.pm lib/File/Path.pm makeaperl.SH minimod.PL perl.c proto.h vms/config.vms vms/ext/MM_VMS.pm x2p/a2p.h NETaa13399: Patch 0m from Andy Files 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/cxux.sh hints/linux.sh hints/powerunix.sh lib/ExtUtils/MakeMaker.pm malloc.c perl.h pp_sys.c util.c NETaa13400: pod2html update from Bill Middleton From: Larry Wall Files patched: pod/pod2html NETaa13401: Boyer-Moore code attempts to compile string longer than 255. From: Kyriakos Georgiou Files patched: util.c The Boyer-Moore table uses unsigned char offsets, but the BM compiler wasn't rejecting strings longer than 255 chars, and was miscompiling them. NETaa13403: missing a $ on variable name From: Wayne Scott Files patched: installperl Yup, it was missing. NETaa13406: didn't wipe out dead match when proceeding to next BRANCH From: Michael P. Clemens Files patched: regexec.c The code to check alternatives didn't invalidate backreferences matched by the failed branch. NETaa13407: overload upgrade From: owner-perl5-porters@nicoh.com Also: Ilya Zakharevich Files patched: MANIFEST gv.c lib/Math/BigInt.pm perl.h pod/perlovl.pod pp.c pp.h pp_hot.c sv.c t/lib/bigintpm.t t/op/overload.t Applied supplied patch, and fixed bug induced by use of sv_setsv to do a deep copy, since sv_setsv no longer copies objecthood. NETaa13409: sv_gets tries to grow string at EOF From: Harold O Morris Files patched: sv.c Applied suggested patch, only two statements earlier, since the end code also does SvCUR_set. NETaa13410: delaymagic did =~ instead of &= ~ From: Andreas Schwab Files patched: pp_hot.c Applied supplied patch. NETaa13411: POSIX didn't compile under -DLEAKTEST From: Frederic Chauveau Files patched: ext/POSIX/POSIX.xs Used NEWSV instead of newSV. NETaa13412: new version from Tony Sanders From: Tony Sanders Files patched: lib/Term/Cap.pm Installed as Term::Cap.pm NETaa13413: regmust extractor needed to restart loop on BRANCH for (?:) to work From: DESARMENIEN Files patched: regcomp.c The BRANCH skipper should have restarted the loop from the top. NETaa13414: the check for accidental list context was done after pm_short check From: Michael H. Coen Files patched: pp_hot.c Moved check for accidental list context to before the pm_short optimization. NETaa13418: perlre.pod babbled nonsense about | in character classes From: Philip Hazel Files patched: pod/perlre.pod Removed bogus brackets. Now reads: Note however that "|" is interpreted as a literal with square brackets, so if you write C<[fee|fie|foe]> you're really only matching C<[feio|]>. NETaa13419: need to document introduction of lexical variables From: "Heading, Anthony" Files patched: pod/perlfunc.pod Now mentions that lexicals aren't introduced till after the current statement. NETaa13420: formats that overflowed a page caused endless top of forms From: Hildo@CONSUL.NL Files patched: pp_sys.c If a record is too large to fit on a page, it now prints whatever will fit and then calls top of form again on the remainder. NETaa13423: the code to do negative list subscript in scalar context was missing From: Steve McDougall Files patched: pp.c The negative subscript code worked right in list context but not in scalar context. In fact, there wasn't code to do it in the scalar context. NETaa13424: existing but undefined CV blocked inheritance From: Spider Boardman Files patched: gv.c Applied supplied patch. NETaa13425: removed extra argument to croak From: "R. Bernstein" Files patched: regcomp.c Removed extra argument. NETaa13427: added return types From: "R. Bernstein" Files patched: x2p/a2py.c Applied suggested patch. NETaa13427: added static declarations Files patched: x2p/walk.c (same) NETaa13428: split was assuming that all backreferences were defined From: Dave Schweisguth Files patched: pp.c split was assuming that all backreferences were defined. NETaa13430: hoistmust wasn't hoisting anchored shortcircuit's length From: Tom Christiansen Also: Rob Hooft Files patched: toke.c NETaa13432: couldn't call code ref under debugger From: Mike Fletcher Files patched: op.c pp_hot.c sv.h The debugging code assumed it could remember a name to represent a subroutine, but anonymous subroutines don't have a name. It now remembers a CV reference in that case. NETaa13435: 1' dumped core From: Larry Wall Files patched: toke.c Didn't check a pointer for nullness. NETaa13436: print foo(123) didn't treat foo as subroutine From: mcook@cognex.com Files patched: toke.c Now treats it as a subroutine rather than a filehandle. NETaa13437: &$::foo didn't think $::foo was a variable name From: mcook@cognex.com Files patched: toke.c Now treats $::foo as a global variable. NETaa13439: referred to old package name From: Tom Christiansen Files patched: lib/Sys/Syslog.pm Wasn't a strict refs problem after all. It was simply referring to package syslog, which had been renamed to Sys::Syslog. NETaa13440: stat operations didn't know what to do with glob or ref to glob From: mcook@cognex.com Files patched: doio.c pp_sys.c Now knows about the kinds of filehandles returned by FileHandle constructors and such. NETaa13442: couldn't find name of copy of deleted symbol table entry From: Spider Boardman Files patched: gv.c gv.h I did a much simpler fix. When gp_free notices that it's freeing the master GV, it nulls out gp_egv. The GvENAME and GvESTASH macros know to revert to gv if egv is null. This has the advantage of not creating a reference loop. NETaa13443: couldn't override an XSUB From: William Setzer Files patched: op.c When the newSUB and newXS routines checked for whether the old sub was defined, they only looked at CvROOT(cv), not CvXSUB(cv). NETaa13443: needed to do same thing in newXS Files patched: op.c (same) NETaa13444: -foo now doesn't warn unless sub foo is defined From: Larry Wall Files patched: toke.c Made it not warn on -foo, unless there is a sub foo defined. NETaa13451: in scalar context, pp_entersub now guarantees one item from XSUB From: Nick Gianniotis Files patched: pp_hot.c The pp_entersub routine now guarantees that an XSUB in scalar context returns one and only one value. If there are fewer, it pushes undef, and if there are more, it returns the last one. NETaa13457: now explicitly disallows printf format with 'n' or '*'. From: lees@cps.msu.edu Files patched: doop.c Now says Use of n in printf format not supported at ./foo line 3. NETaa13458: needed to call SvPOK_only() in pp_substr From: Wayne Scott Files patched: pp.c Needed to call SvPOK_only() in pp_substr. NETaa13459: umask and chmod now warn about missing initial 0 even with paren From: Andreas Koenig Files patched: toke.c Now skips parens as well as whitespace looking for argument. NETaa13460: backtracking didn't work on .*? because reginput got clobbered From: Andreas Koenig Files patched: regexec.c When .*? did a probe of the rest of the string, it clobbered reginput, so the next call to match a . tried to match the newline and failed. NETaa13475: \(@ary) now treats array as list of scalars From: Tim Bunce Files patched: op.c The mod() routine now refrains from marking @ary as an lvalue if it's in parens and is the subject of an OP_REFGEN. NETaa13481: accept buffer wasn't aligned good enough From: Holger Bechtold Also: Christian Murphy Files patched: pp_sys.c Applied suggested patch. NETaa13486: while (<>) now means while (defined($_ = <>)) From: Jim Balter Files patched: op.c pod/perlop.pod while () now means while (defined($_ = )). NETaa13500: needed DESTROY in FileHandle From: Tim Bunce Files patched: ext/POSIX/POSIX.pm Added DESTROY method. Also fixed ungensym to use POSIX:: instead of _POSIX. Removed ungensym from close method, since DESTROY should do that now. NETaa13502: now complains if you use local on a lexical variable From: Larry Wall Files patched: op.c Now says something like Can't localize lexical variable $var at ./try line 6. NETaa13512: added $SIG{__WARN__} and $SIG{__DIE__} hooks From: Larry Wall Files patched: embed.h gv.c interp.sym mg.c perl.h pod/perlvar.pod pp_ctl.c util.c Todo pod/perldiag.pod NETaa13514: statements before intro of lex var could see lex var From: William Setzer Files patched: op.c When a lexical variable is declared, introduction is delayed until the start of the next statement, so that any initialization code runs outside the scope of the new variable. Thus, my $y = 3; my $y = $y; print $y; should print 3. Unfortunately, the declaration was marked with the beginning location at the time that "my $y" was processed instead of when the variable was introduced, so any embedded statements within an anonymous subroutine picked up the wrong "my". The declaration is now labelled correctly when the variable is actually introduced. NETaa13520: added closures From: Larry Wall Files 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 NETaa13520: test to see if lexical works in a format now Files patched: t/op/write.t NETaa13522: substitution couldn't be used on a substr() From: Hans Mulder Files patched: pp_ctl.c pp_hot.c Changed pp_subst not to use sv_replace() anymore, which didn't handle lvalues and was overkill anyway. Should be slightly faster this way too. NETaa13525: G_EVAL mode in perl_call_sv didn't return values right. Files patched: perl.c NETaa13525: consolidated error message From: Larry Wall Files patched: perl.h toke.c NETaa13525: derived it Files patched: perly.h NETaa13525: missing some values from embed.h Files patched: embed.h NETaa13525: random cleanup Files patched: MANIFEST Todo cop.h lib/TieHash.pm lib/perl5db.pl opcode.h patchlevel.h pod/perldata.pod pod/perlsub.pod t/op/ref.t toke.c NETaa13525: random cleanup Files patched: pp_ctl.c util.c NETaa13527: File::Find needed to export $name and $dir From: Chaim Frenkel Files patched: lib/File/Find.pm They are now exported. NETaa13528: cv_undef left unaccounted-for GV pointer in CV From: Tye McQueen Also: Spider Boardman Files patched: op.c NETaa13530: scalar keys now resets hash iterator From: Tim Bunce Files patched: doop.c scalar keys() now resets the hash iterator. NETaa13531: h2ph doesn't check defined right From: Casper H.S. Dik Files patched: h2ph.SH NETaa13540: VMS update From: Larry Wall Files patched: MANIFEST README.vms doio.c embed.h ext/DynaLoader/dl_vms.xs interp.sym lib/Cwd.pm lib/ExtUtils/xsubpp lib/File/Basename.pm lib/File/Find.pm lib/File/Path.pm 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/Filespec.pm vms/ext/MM_VMS.pm vms/ext/VMS/stdio/Makefile.PL vms/ext/VMS/stdio/stdio.pm vms/ext/VMS/stdio/stdio.xs vms/genconfig.pl vms/perlvms.pod vms/sockadapt.c vms/sockadapt.h vms/vms.c vms/vmsish.h vms/writemain.pl NETaa13540: got some duplicate code Files patched: lib/File/Path.pm NETaa13540: stuff from Charles Files patched: MANIFEST README.vms lib/ExtUtils/MakeMaker.pm lib/ExtUtils/MakeMaker.pm lib/ExtUtils/xsubpp lib/File/Basename.pm lib/File/Path.pm 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/Filespec.pm vms/ext/Filespec.pm vms/ext/MM_VMS.pm vms/ext/MM_VMS.pm vms/ext/VMS/stdio/stdio.pm vms/ext/VMS/stdio/stdio.xs vms/gen_shrfls.pl vms/gen_shrfls.pl vms/genconfig.pl vms/genconfig.pl vms/mms2make.pl vms/perlvms.pod vms/sockadapt.h vms/test.com vms/vms.c vms/vms.c vms/vmsish.h vms/vmsish.h vms/writemain.pl NETaa13540: tweak from Charles Files patched: lib/File/Path.pm NETaa13552: scalar unpack("P4",...) ignored the 4 From: Eric Arnold Files patched: pp.c The optimization that tried to do only one item in a scalar context didn't realize that the argument to P was not a repeat count. NETaa13553: now warns about 8 or 9 in octal escapes From: Mike Rogers Files patched: util.c Now warns if it finds 8 or 9 before the end of the octal escape sequence. So \039 produces a warning, but \0339 does not. NETaa13554: now allows foreach ${"name"} From: Johan Holtman Files patched: op.c Instead of trying to remove OP_RV2SV, the compiler now just transmutes it into an OP_RV2GV, which is a no-op for ordinary variables and does the right thing for ${"name"}. NETaa13559: substitution now always checks for readonly From: Rodger Anderson Files patched: pp_hot.c Substitution now always checks for readonly. NETaa13561: added explanations of closures and curly-quotes From: Larry Wall Files patched: pod/perlref.pod NETaa13562: null components in path cause indigestion From: Ambrose Kofi Laing Files patched: lib/Cwd.pm lib/pwd.pl NETaa13575: documented semantics of negative substr length From: Jeff Bouis Files patched: pod/perlfunc.pod Documented the fact that negative length now leaves characters off the end, and while I was at it, made it work right even if offset wasn't 0. NETaa13575: negative length to substr didn't work when offset non-zero Files patched: pp.c (same) NETaa13575: random cleanup Files patched: pod/perlfunc.pod (same) NETaa13580: couldn't localize $ACCUMULATOR From: Larry Wall Files patched: gv.c lib/English.pm mg.c perl.c sv.c Needed to make $^A a real magical variable. Also lib/English.pm wasn't exporting good. NETaa13583: doc mods from Tom From: Larry Wall Files 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 NETaa13589: return was enforcing list context on its arguments From: Tim Freeman Files patched: opcode.pl A return was being treated like a normal list operator, in that it was setting list context on its arguments. This was bogus. NETaa13591: POSIX::creat used wrong argument From: Paul Marquess Files patched: ext/POSIX/POSIX.pm Applied suggested patch. NETaa13605: use strict refs error message now displays bad ref From: Peter Gordon Files patched: perl.h pod/perldiag.pod pp.c pp_hot.c Now says Can't use string ("2") as a HASH ref while "strict refs" in use at ./foo line 12. NETaa13630: eof docs were unclear From: Hallvard B Furuseth Files patched: pod/perlfunc.pod Applied suggested patch. NETaa13636: $< and $> weren't refetched on undump restart From: Steve Pearlmutter Files patched: perl.c The code in main() bypassed perl_construct on an undump restart, which bypassed the code that set $< and $>. NETaa13641: added Tim's fancy new import whizbangers From: Tim Bunce Files patched: lib/Exporter.pm Applied suggested patch. NETaa13649: couldn't AUTOLOAD a symbol reference From: Larry Wall Files patched: pp_hot.c pp_entersub needed to guarantee a CV so it would get to the AUTOLOAD code. NETaa13651: renamed file had wrong package name From: Andreas Koenig Files patched: lib/File/Path.pm Applied suggested patch. NETaa13660: now that we're testing distribution we can diagnose RANDBITS errors From: Karl Glazebrook Files patched: t/op/rand.t Changed to suggested algorithm. Also duplicated it to test rand(100) too. NETaa13660: rand.t didn't test for proper distribution within range Files patched: t/op/rand.t (same) NETaa13671: array slice misbehaved in a scalar context From: Tye McQueen Files patched: pp.c A spurious else prevented the scalar-context-handling code from running. NETaa13672: filehandle constructors in POSIX don't return failure successfully From: Ian Phillipps Files patched: ext/POSIX/POSIX.pm Applied suggested patch. NETaa13678: forced $1 to always be untainted From: Ka-Ping Yee Files patched: mg.c I believe the bug that triggered this was fixed elsewhere, but just in case, I put in explicit code to force $1 et al not to be tainted regardless. NETaa13682: formline doc need to discuss ~ and ~~ policy From: Peter Gordon Files patched: pod/perlfunc.pod NETaa13686: POSIX::open and POSIX::mkfifo didn't check tainting From: Larry Wall Files patched: ext/POSIX/POSIX.xs open() and mkfifo() now check tainting. NETaa13687: new Exporter.pm From: Tim Bunce Files patched: lib/Exporter.pm Added suggested changes, except for @EXPORTABLE, because it looks too much like @EXPORTTABLE. Decided to stick with @EXPORT_OK because it looks more like an adjunct. Also added an export_tags routine. The keys in the %EXPORT_TAGS hash no longer use colons, to make the initializers prettier. NETaa13687: new Exporter.pm Files patched: ext/POSIX/POSIX.pm (same) NETaa13694: add sockaddr_in to Socket.pm From: Tim Bunce Files patched: ext/Socket/Socket.pm Applied suggested patch. NETaa13695: library routines should use qw() as good example From: Dean Roehrich Files patched: ext/DB_File/DB_File.pm ext/DynaLoader/DynaLoader.pm ext/Fcntl/Fcntl.pm ext/GDBM_File/GDBM_File.pm ext/POSIX/POSIX.pm ext/Socket/Socket.pm Applied suggested patch. NETaa13696: myconfig should be a routine in Config.pm From: Kenneth Albanowski Files patched: configpm Applied suggested patch. NETaa13704: fdopen closed fd on failure From: Hallvard B Furuseth Files patched: doio.c Applied suggested patch. NETaa13706: Term::Cap doesn't work From: Dean Roehrich Files patched: lib/Term/Cap.pm Applied suggested patch. NETaa13710: cryptswitch needed to be more "useable" From: Tim Bunce Files patched: embed.h global.sym perl.h toke.c The cryptswitch_fp function now can operate in two modes. It can modify the global rsfp to redirect input as before, or it can modify linestr and return true, indicating that it is not necessary for yylex to read another line since cryptswitch_fp has just done it. NETaa13712: new_tmpfile() can't be called as constructor From: Hans Mulder Files patched: ext/POSIX/POSIX.xs Now allows new_tmpfile() to be called as a constructor. NETaa13714: variable method call not documented From: "Randal L. Schwartz" Files patched: pod/perlobj.pod Now indicates that OBJECT->$method() works. NETaa13715: PACK->$method produces spurious warning From: Larry Wall Files patched: toke.c The -> operator was telling the lexer to expect an operator when the next thing was a variable. NETaa13716: Carp now allows multiple packages to be skipped out of From: Larry Wall Files patched: lib/Carp.pm The subroutine redefinition warnings now warn on import collisions. NETaa13716: Exporter catches warnings and gives a better line number Files patched: lib/Exporter.pm (same) NETaa13716: now counts imported routines as "defined" for redef warnings Files patched: op.c sv.c (same) ------------- Version 5.000 ------------- New things ---------- The -w switch is much more informative. References. See t/op/ref.t for examples. All entities in Perl 5 are reference counted so that it knows when each item should be destroyed. Objects. See t/op/ref.t for examples. => is now a synonym for comma. This is useful as documentation for arguments that come in pairs, such as initializers for associative arrays, or named arguments to a subroutine. All functions have been turned into list operators or unary operators, meaning the parens are optional. Even subroutines may be called as list operators if they've already been declared. More embeddible. See main.c and embed_h.sh. Multiple interpreters in the same process are supported (though not with interleaved execution yet). The interpreter is now flattened out. Compare Perl 4's eval.c with the perl 5's pp.c. Compare Perl 4's 900 line interpreter loop in cmd.c with Perl 5's 1 line interpreter loop in run.c. Eventually we'll make everything non-blocking so we can interface nicely with a scheduler. eval is now treated more like a subroutine call. Among other things, this means you can return from it. Format value lists may be spread over multiple lines by enclosing in a do {} block. You may now define BEGIN and END subroutines for each package. The BEGIN subroutine executes the moment it's parsed. The END subroutine executes just before exiting. Flags on the #! line are interpreted even if the script wasn't executed directly. (And even if the script was located by "perl -x"!) The ?: operator is now legal as an lvalue. List context now propagates to the right side of && and ||, as well as the 2nd and 3rd arguments to ?:. The "defined" function can now take a general expression. Lexical scoping available via "my". eval can see the current lexical variables. The preferred package delimiter is now :: rather than '. tie/untie are now preferred to dbmopen/dbmclose. Multiple DBM implementations are allowed in the same executable, so you can write scripts to interchange data among different formats. New "and" and "or" operators work just like && and || but with a precedence lower than comma, so they work better with list operators. New functions include: abs(), chr(), uc(), ucfirst(), lc(), lcfirst(), chomp(), glob() require with a number checks to see that the version of Perl that is currently running is at least that number. Dynamic loading of external modules is now supported. There is a new quote form qw//, which is equivalent to split(' ', q//). Assignment of a reference to a glob value now just replaces the single element of the glob corresponding to the reference type: *foo = \$bar, *foo = \&bletch; Filehandle methods are now supported: output_autoflush STDOUT 1; There is now an "English" module that provides human readable translations for cryptic variable names. Autoload stubs can now call the replacement subroutine with goto &realsub. Subroutines can be defined lazily in any package by declaring an AUTOLOAD routine, which will be called if a non-existent subroutine is called in that package. Several previously added features have been subsumed under the new keywords "use" and "no". Saying "use Module LIST" is short for BEGIN { require Module; import Module LIST; } The "no" keyword is identical except that it calls "unimport" instead. The earlier pragma mechanism now uses this mechanism, and two new modules have been added to the library to implement "use integer" and variations of "use strict vars, refs, subs". Variables may now be interpolated literally into a pattern by prefixing them with \Q, which works just like \U, but backwhacks non-alphanumerics instead. There is also a corresponding quotemeta function. Any quantifier in a regular expression may now be followed by a ? to indicate that the pattern is supposed to match as little as possible. Pattern matches may now be followed by an m or s modifier to explicitly request multiline or singleline semantics. An s modifier makes . match newline. Patterns may now contain \A to match only at the beginning of the string, and \Z to match only at the end. These differ from ^ and $ in that they ignore multiline semantics. In addition, \G matches where the last interation of m//g or s///g left off. Non-backreference-producing parens of various sorts may now be indicated by placing a ? directly after the opening parenthesis, followed by a character that indicates the purpose of the parens. An :, for instance, indicates simple grouping. (?:a|b|c) will match any of a, b or c without producing a backreference. It does "eat" the input. There are also assertions which do not eat the input but do lookahead for you. (?=stuff) indicates that the next thing must be "stuff". (?!nonsense) indicates that the next thing must not be "nonsense". The negation operator now treats non-numeric strings specially. A -"text" is turned into "-text", so that -bareword is the same as "-bareword". If the string already begins with a + or -, it is flipped to the other sign. Incompatibilities ----------------- @ now always interpolates an array in double-quotish strings. Some programs may now need to use backslash to protect any @ that shouldn't interpolate. Ordinary variables starting with underscore are no longer forced into package main. s'$lhs'$rhs' now does no interpolation on either side. It used to interplolate $lhs but not $rhs. The second and third arguments of splice are now evaluated in scalar context (like the book says) rather than list context. Saying "shift @foo + 20" is now a semantic error because of precedence. "open FOO || die" is now incorrect. You need parens around the filehandle. The elements of argument lists for formats are now evaluated in list context. This means you can interpolate list values now. You can't do a goto into a block that is optimized away. Darn. It is no longer syntactically legal to use whitespace as the name of a variable, or as a delimiter for any kind of quote construct. Some error messages will be different. The caller function now returns a false value in a scalar context if there is no caller. This lets library files determine if they're being required. m//g now attaches its state to the searched string rather than the regular expression. "reverse" is no longer allowed as the name of a sort subroutine. taintperl is no longer a separate executable. There is now a -T switch to turn on tainting when it isn't turned on automatically. Symbols starting with _ are no longer forced into package main, except for $_ itself (and @_, etc.). Double-quoted strings may no longer end with an unescaped $ or @. Negative array subscripts now count from the end of the array. The comma operator in a scalar context is now guaranteed to give a scalar context to its arguments. The ** operator now binds more tightly than unary minus. Setting $#array lower now discards array elements so that destructors work reasonably. delete is not guaranteed to return the old value for tied arrays, since this capability may be onerous for some modules to implement. Attempts to set $1 through $9 now result in a run-time error.