nglenn has asked for the wisdom of the Perl Monks concerning the following question:

I am using XML::Simple to parse some utf8 xml files. However, utf8 files with BOM cause it to crash and burn with the following error:

Wide character in subroutine entry at C:/Perl/lib/Encode.pm line 174.

The offending code is:

use XML::Simple;
local $XML::Simple::PREFERRED_PARSER = 'XML::Parser';
$xml->XMLin($file,ForceArray => ['map'], KeyAttr =>{},);

[download]

The error does NOT occur with utf8 files that have no BOM. Seems bazaar, and I have no idea what to do. Any suggestions?

Update:

It has to be the way I'm using it. Here's more context for the code I posted earlier:

In ETTX.pm:

package ETTX;
use strict;
use XML::Simple;
local $XML::Simple::PREFERRED_PARSER = 'XML::Parser';
use XML::SemanticDiff;

sub new(){#scalar file name
    my $class = shift;
    my $self = 
        {
            ettxFile => '',
            ettx => {},
        };
    bless $self, $class;
    load($self,shift) if @_ ==1;
    return $self;
};

sub load(){#scalar file name
    my ($self, $ettxFile) = @_;
    print "loading $ettxFile";
    open(my $fh, '<', $ettxFile);
    binmode($fh);
    my $xml = XML::Simple->new();
    $self->{ettx} = 
        $xml->XMLin($fh, 
            ForceArray => ['map'], KeyAttr => {}, 
        )
    ->{table};
    $self->{ettxFile} = $ettxFile;
    # print Dumper($self);
    1;
}

[download]

The code above was changed to to use the binmode() thing, but still fails just the same.

Called like so:

use lib 'C:\Texts\Programs';    #wherever ETTX.pm is
use ETTX;
my $ettxFile = 'someXMLfile.ettx';
my $ettx->load($ettxFile);

[download]
And here's a sample xml file:

<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<ettx ver="2">
  <table id="{4fa6cd7a-f7b6-416d-8f59-3acc0eab9bdb}" name="TestFile">
    <level type="V">
      <map sync="Title" src="someText"/>
      <map sync="Title Page" src="someText"/>
    </level>
  </table>
</ettx>

[download]

Replies are listed 'Best First'.
Re: Encode throws "Wide character in subroutine entry" when using XML::Simple
by Jim (Curate) on Dec 12, 2010 at 06:19 UTC

    This works for me.

    D:\>file Simple.xml
Simple.xml:     Text file, UTF-8 format

D:\>cat Simple.xml
<?xml version="1.0" encoding="UTF-8"?>
<root>
    <foo>
        bar
    </foo>
</root>

D:\>od -h Simple.xml | head -1 | cut -c 1-24
0000000000    EF  BB  BF

D:\>cat Simple.pl
#!perl

use strict;
use warnings;

use XML::Simple;

local $XML::Simple::PREFERRED_PARSER = 'XML::Parser';

my $file = shift @ARGV;
my $xml  = XMLin($file, ForceArray => ['map'], KeyAttr => {});

D:\>perl Simple.pl Simple.xml

D:\>perl -v | head -8 | fmt -w 68

This is perl 5, version 12, subversion 2 (v5.12.2) built for
MSWin32-x86-multi-thread (with 8 registered patches, see perl -V for
more detail)

Copyright 1987-2010, Larry Wall

Binary build 1202 [293621] provided by ActiveState
http://www.ActiveState.com Built Sep  6 2010 23:36:03

D:\>

    [download]

    Simple.xml is a simple XML document in the UTF-8 character encoding scheme of the Unicode coded character set. It has a UTF-8 byte order mark in it. XML::Simple handles the XML document properly.

[reply]
[d/l]
      Not directly related but I had the same error and google made me land here. You get the same error when you try to decode and specify a wrong encoding: I tried to decode to iso 8859-1 (latin-1) when it was utf-8.
[reply]
Re: Encode throws "Wide character in subroutine entry" when using XML::Simple
by Jim (Curate) on Dec 12, 2010 at 00:10 UTC

    Try this (untested):

    use strict;
use warnings;
use autodie;

use XML::Simple;

local $XML::Simple::PREFERRED_PARSER = 'XML::Parser';

open my $fh, '<:encoding(UTF-8)', $file;

$xml->XMLin(
    $fh,
    ForceArray => ['map'],
    KeyAttr    => {},
);

    [download]

    Just a thought... But see 850187.

[reply]
[d/l]
      Backwards. That would cause the problem. 
      open(my $fh, '<:encoding(UTF-8)', $file);

      [download]
      should be 
      open(my $fh, '<', $file);
binmode($fh);

      [download]
[reply]
[d/l]
[select]

        That is as counterintuitive as a thing can be. You have a file that is encoded in UTF-8 and has a UTF-8 byte order mark in it, yet to solve a problem with it not being interpreted properly as UTF-8 text by a module, you have to use binmode, not the proper UTF-8 encoding layer :encoding(UTF-8). It just doesn't make sense. Who would intuit that? Obviously, not I. :-(

[reply]
[d/l]
[select]

        Nope. This still throws the same errors:

            open(my $fh, '<', $file);
    binmode($fh);
    my $xml = XML::Simple->new();
    $self->{ettx} = 
        $xml->XMLin($fh, 
            ForceArray => ['map'], KeyAttr => {}, 
        )
    ->{table};

        [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom