tied gets scalars and globs confused

[p5pRT]

Migrated from rt.perl.org#77496 (status was 'resolved')

Searchable as RT77496$

[p5pRT]

From @cpansprout

(I must be tired today. This is the second bug report I’ve muddled up. I sent it first without the perlbug output. It disappeared, but—who knows?—it may pop up again someday.)

In trying to write a test for an assignment bug, I came across this anomaly with tied()​:

sub TIESCALAR {bless[]}
sub FETCH { 1 }
sub STORE{}
tie $rin, "";
$rin = *rin;
print qw[good bad][!tied $rin];

tied returns undef, so this prints ‘bad’. tied checks first to see whether its argument is a glob. If it is, it returns undef if the IO slot is empty.

So the the tiedness, according to tied(), varies depending on what the last value happened to be that was returned or assigned.

---

Flags​:
  category=core
  severity=low

---

Site configuration information for perl 5.13.4​:

Configured by sprout at Fri Aug 20 23​:24​:53 PDT 2010.

Summary of my perl5 (revision 5 version 13 subversion 4 patch v5.13.4-16-g16c9153) configuration​:
  Snapshot of​: 9b47cdd
  Platform​:
  osname=darwin, osvers=10.4.0, archname=darwin-2level
  uname='darwin pint.local 10.4.0 darwin kernel version 10.4.0​: fri apr 23 18​:28​:53 pdt 2010; root​:xnu-1504.7.4~1release_i386 i386 '
  config_args='-de -Dusedevel'
  hint=recommended, useposix=true, d_sigaction=define
  useithreads=undef, usemultiplicity=undef
  useperlio=define, d_sfio=undef, uselargefiles=define, usesocks=undef
  use64bitint=undef, use64bitall=undef, uselongdouble=undef
  usemymalloc=n, bincompat5005=undef
  Compiler​:
  cc='cc', ccflags ='-fno-common -DPERL_DARWIN -no-cpp-precomp -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include',
  optimize='-O3',
  cppflags='-no-cpp-precomp -fno-common -DPERL_DARWIN -no-cpp-precomp -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include'
  ccversion='', gccversion='4.2.1 (Apple Inc. build 5664)', gccosandvers=''
  intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=1234
  d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
  ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=8
  alignbytes=8, prototype=define
  Linker and Libraries​:
  ld='env MACOSX_DEPLOYMENT_TARGET=10.3 cc', ldflags =' -fstack-protector -L/usr/local/lib'
  libpth=/usr/local/lib /usr/lib
  libs=-ldbm -ldl -lm -lutil -lc
  perllibs=-ldl -lm -lutil -lc
  libc=/usr/lib/libc.dylib, so=dylib, useshrplib=false, libperl=libperl.a
  gnulibc_version=''
  Dynamic Linking​:
  dlsrc=dl_dlopen.xs, dlext=bundle, d_dlsymun=undef, ccdlflags=' '
  cccdlflags=' ', lddlflags=' -bundle -undefined dynamic_lookup -L/usr/local/lib -fstack-protector'

Locally applied patches​:
 

---

@​INC for perl 5.13.4​:
  /usr/local/lib/perl5/site_perl/5.13.4/darwin-2level
  /usr/local/lib/perl5/site_perl/5.13.4
  /usr/local/lib/perl5/5.13.4/darwin-2level
  /usr/local/lib/perl5/5.13.4
  /usr/local/lib/perl5/site_perl
  .

---

Environment for perl 5.13.4​:
  DYLD_LIBRARY_PATH (unset)
  HOME=/Users/sprout
  LANG=en_US.UTF-8
  LANGUAGE (unset)
  LD_LIBRARY_PATH (unset)
  LOGDIR (unset)
  PATH=/usr/bin​:/bin​:/usr/sbin​:/sbin​:/usr/local/bin​:/usr/X11/bin​:/usr/local/bin
  PERL_BADLANG (unset)
  SHELL=/bin/bash

[p5pRT]

From @cpansprout

On Aug 29, 2010, at 1​:38 PM, Father Chrysostomos wrote​:

In trying to write a test for an assignment bug, I came across this anomaly with tied()​:

sub TIESCALAR {bless[]}
sub FETCH { 1 }
sub STORE{}
tie $rin, "";
$rin = *rin;
print qw[good bad][!tied $rin];

tied returns undef, so this prints ‘bad’. tied checks first to see whether its argument is a glob. If it is, it returns undef if the IO slot is empty.

So the the tiedness, according to tied(), varies depending on what the last value happened to be that was returned or assigned.

There should be a patch attached hereto. I put the explanation of OPf_SPECIAL above OP_HELEM to avoid a conflict with the patch for bug #20444.

[p5pRT]

From @cpansprout

From​: Father Chrysostomos <sprout@​cpan.org>

[perl #77496] tied gets scalars and globs confused

• Give a tied op a special flag if the RHS is an rv2gv.
• Make pp_tied use that to distinguish between tied scalars and
  tied handles.

This fixes the problem of tied($scalar) ignoring the tie if the last
thing returned or assigned happened to be a glob.

Inline Patch

diff -Nup blead-77362-glob2lv8/op.c blead-77362-glob2lv9.tied/op.c
--- blead-77362-glob2lv8/op.c	2010-08-20 18:55:11.000000000 -0700
+++ blead-77362-glob2lv9.tied/op.c	2010-08-27 22:10:24.000000000 -0700
@@ -7155,6 +7155,9 @@ Perl_ck_fun(pTHX_ OP *o)
 	    return too_many_arguments(o,OP_DESC(o));
 #endif
 	listkids(o);
+
+	if (type == OP_TIED && cLISTOPo->op_first->op_type == OP_RV2GV)
+	    o->op_flags |= OPf_SPECIAL;
     }
     else if (PL_opargs[type] & OA_DEFGV) {
 #ifdef PERL_MAD
diff -Nup blead-77362-glob2lv8/op.h blead-77362-glob2lv9.tied/op.h
--- blead-77362-glob2lv8/op.h	2010-08-20 18:55:11.000000000 -0700
+++ blead-77362-glob2lv9.tied/op.h	2010-08-27 22:00:46.000000000 -0700
@@ -139,6 +139,7 @@ Deprecated.  Use C<GIMME_V> instead.
 				/*  On OP_ANONHASH and OP_ANONLIST, create a
 				    reference to the new anon hash or array */
 				/*  On OP_ENTER, store caller context */
+				/*  On OP_TIED, RHS is a handle */
 				/*  On OP_HELEM and OP_HSLICE, localization will be followed
 				    by assignment, so do not wipe the target if it is special
 				    (e.g. a glob or a magic SV) */
diff -Nup blead-77362-glob2lv8/pp_sys.c blead-77362-glob2lv9.tied/pp_sys.c
--- blead-77362-glob2lv8/pp_sys.c	2010-08-25 18:03:31.000000000 -0700
+++ blead-77362-glob2lv9.tied/pp_sys.c	2010-08-27 23:02:50.000000000 -0700
@@ -941,7 +941,8 @@ PP(pp_tied)
     const char how = (SvTYPE(sv) == SVt_PVHV || SvTYPE(sv) == SVt_PVAV)
 		? PERL_MAGIC_tied : PERL_MAGIC_tiedscalar;
 
-    if (isGV_with_GP(sv) && !(sv = MUTABLE_SV(GvIOp(sv))))
+    if (isGV_with_GP(sv) && (PL_op->op_flags & OPf_SPECIAL || !SvFAKE(sv))
+     && !(sv = MUTABLE_SV(GvIOp(sv))))
 	RETPUSHUNDEF;
 
     if ((mg = SvTIED_mg(sv, how))) {
diff -Nurp blead-77362-glob2lv8/t/op/tie.t blead-77362-glob2lv9.tied/t/op/tie.t
--- blead-77362-glob2lv8/t/op/tie.t	2010-08-11 04:53:10.000000000 -0700
+++ blead-77362-glob2lv9.tied/t/op/tie.t	2010-08-27 17:47:48.000000000 -0700
@@ -939,3 +939,17 @@ sub IO::File::TIEARRAY {
 fileno FOO; tie @a, "FOO"
 EXPECT
 Can't locate object method "TIEARRAY" via package "FOO" at - line 5.
+########
+
+# tied() should still work on tied scalars after glob assignment
+sub TIESCALAR {bless[]}
+sub FETCH {*foo}
+sub f::TIEHANDLE{bless[],f}
+tie *foo, "f";
+tie $rin, "";
+[$rin]; # call FETCH
+print ref tied $rin, "\n";
+print ref tied *$rin, "\n";
+EXPECT
+main
+f

[p5pRT]

From @cpansprout

On Sun Aug 29 13​:43​:32 2010, sprout wrote​:

On Aug 29, 2010, at 1​:38 PM, Father Chrysostomos wrote​:

In trying to write a test for an assignment bug, I came across this
anomaly with tied()​:

sub TIESCALAR {bless[]}
sub FETCH { 1 }
sub STORE{}
tie $rin, "";
$rin = *rin;
print qw[good bad][!tied $rin];

tied returns undef, so this prints ‘bad’. tied checks first to see
whether its argument is a glob. If it is, it returns undef if the IO
slot is empty.

So the the tiedness, according to tied(), varies depending on what
the last value happened to be that was returned or assigned.

There should be a patch attached hereto.

I have applied a simplified version as 8752206. (2acc331 allows pp_tied
to check just the FAKE flag on a glob, to determine whether it should be
considered one.)

[p5pRT]

From [Unknown Contact. See original ticket]

On Sun Aug 29 13​:43​:32 2010, sprout wrote​:

On Aug 29, 2010, at 1​:38 PM, Father Chrysostomos wrote​:

In trying to write a test for an assignment bug, I came across this
anomaly with tied()​:

sub TIESCALAR {bless[]}
sub FETCH { 1 }
sub STORE{}
tie $rin, "";
$rin = *rin;
print qw[good bad][!tied $rin];

tied returns undef, so this prints ‘bad’. tied checks first to see
whether its argument is a glob. If it is, it returns undef if the IO
slot is empty.

So the the tiedness, according to tied(), varies depending on what
the last value happened to be that was returned or assigned.

There should be a patch attached hereto.

I have applied a simplified version as 8752206. (2acc331 allows pp_tied
to check just the FAKE flag on a glob, to determine whether it should be
considered one.)

[p5pRT]

@cpansprout - Status changed from 'new' to 'resolved'

[p5pRT]

@cpansprout - Status changed from 'resolved' to 'open'

[p5pRT]

From @cpansprout

I’ve re-opened this, since the fix was reverted for 5.14. See #79528.

[p5pRT]

From @cpansprout

On Wed Apr 20 14​:32​:56 2011, sprout wrote​:

I’ve re-opened this, since the fix was reverted for 5.14. See #79528.

And now I’m closing it, as the revert was reverted as 4be76e1.

[p5pRT]

@cpansprout - Status changed from 'open' to 'resolved'