From @cpansprout

I was a little buffled when I added eval{} to debug a fatal statement and the eval failed to catch the error.

It turns out that localisation that fails (due to a missing STORE method) will already have half-localised the variable, such that the unwinding of the eval scope will trigger another STORE, which will die and exit the next outer eval scope.

So eval{eval{ local $tied }} catches the error.

Localisation that fails obviously(?) shouldn’t half-happen, but should scope unwinding in general suppress errors that occur during STORE?

I don’t know which parts of perl need fixing. Maybe this is a won’t-fix. Or a useful feature.

package Acme​::Multidie;

@​EXPORT = multidie=>;
@​ISA = Exporter; require Exporter;

sub TIESCALAR{ shift; bless \@​_ }
sub FETCH {}
sub STORE { $_[0][0] --> 2 and return; die $_[0][1] }

sub multidie {
  my ($levels, $msg) = @​_;
  !defined $msg || $msg eq '' and $msg = (eval{die},$@​);
  die $msg if $levels < 2;
  local *bomb;
  tie $bomb, __PACKAGE__, ++$levels, $msg;
  eval 'local $bomb;' x $levels;
}

package main;

import Acme'Multidie;

eval {
eval {
  eval {
  eval {
  eval {
  eval {
  eval {
  multidie(4) # This takes us ---
  }; # `
  warn 1 # |
  }; # |
  warn 2 # /
  }; # /
  warn 3 # /
  }; # /
  warn 4 # <--- here -----------'
  };
  warn 5
};
warn 6
};
warn 7;

__END__

Here’s a test case​:

diff --git a/t/op/tie.t b/t/op/tie.t
index bbd789c..22fb260 100644
--- a/t/op/tie.t
+++ b/t/op/tie.t
@@ -1209,3 +1209,13 @@ $tyre = \tie $tied, "";
 print "ok\n" if \tied $tied == $tyre;
 EXPECT
 ok
+########
+
+# eval+local+tie
+sub TIESCALAR{bless[]}
+sub FETCH { }
+sub STORE { die }
+tie $x, "";
+eval { local $x; 1 } or warn "oenoeuhtn";
+EXPECT
+oenoeuhtn at - line 7.
---

Site configuration information for perl 5.15.5​:

Configured by sprout at Sat Nov 26 11​:40​:22 PST 2011.

Summary of my perl5 (revision 5 version 15 subversion 5) configuration​:
  Snapshot of​: c071f8d
  Platform​:
  osname=darwin, osvers=10.5.0, archname=darwin-thread-multi-2level
  uname='darwin pint.local 10.5.0 darwin kernel version 10.5.0​: fri nov 5 23​:20​:39 pdt 2010; root​:xnu-1504.9.17~1release_i386 i386 '
  config_args='-de -Dusedevel -Duseithreads -Dmad'
  hint=recommended, useposix=true, d_sigaction=define
  useithreads=define, usemultiplicity=define
  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 -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include',
  optimize='-O3',
  cppflags='-fno-common -DPERL_DARWIN -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=, 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 5.15.5​:
  /usr/local/lib/perl5/site_perl/5.15.5/darwin-thread-multi-2level
  /usr/local/lib/perl5/site_perl/5.15.5
  /usr/local/lib/perl5/5.15.5/darwin-thread-multi-2level
  /usr/local/lib/perl5/5.15.5
  /usr/local/lib/perl5/site_perl
  .

Environment for perl 5.15.5​:
  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

From @cpansprout

On Sun Dec 11 13​:13​:28 2011, sprout wrote​:

I was a little buffled when I added eval{} to debug a fatal statement
and the eval failed to catch the error.

It turns out that localisation that fails (due to a missing STORE
method) will already have half-localised the variable, such that
the unwinding of the eval scope will trigger another STORE, which
will die and exit the next outer eval scope.

So eval{eval{ local $tied }} catches the error.

Localisation that fails obviously(?) shouldn’t half-happen, but should
scope unwinding in general suppress errors that occur during STORE?

I don’t know which parts of perl need fixing. Maybe this is a won’t-
fix. Or a useful feature.

I suspect the correct fix is to use setjmp in leave_scope. But setjmp
is one of those scary things I stay away from.

--

Father Chrysostomos

From [Unknown Contact. See original ticket]

On Sun Dec 11 13​:13​:28 2011, sprout wrote​:

I was a little buffled when I added eval{} to debug a fatal statement
and the eval failed to catch the error.

It turns out that localisation that fails (due to a missing STORE
method) will already have half-localised the variable, such that
the unwinding of the eval scope will trigger another STORE, which
will die and exit the next outer eval scope.

So eval{eval{ local $tied }} catches the error.

Localisation that fails obviously(?) shouldn’t half-happen, but should
scope unwinding in general suppress errors that occur during STORE?

I don’t know which parts of perl need fixing. Maybe this is a won’t-
fix. Or a useful feature.

I suspect the correct fix is to use setjmp in leave_scope. But setjmp
is one of those scary things I stay away from.

--

Father Chrysostomos

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