Validating XML

networker2149 has asked for the wisdom of the Perl Monks concerning the following question:

I got an error recently after modifying a script I wrote to parse XML from juniper router commands. The change was to run a different show command that produces more information. The problem is that routers are apparently bad at XML and this caused the mal-formatted XML I'm used to to turn into invalid XML. Will post more code if required but It worked pretty well before the invalid characters. I put in a regex to remove the invalid character, but I was curious if there was a way to do this automatically. Or at least another module that provides better error handling. Here's the error: reference to invalid character number at line 1, column 1373, byte 1373 at /usr/lib/perl5/site_perl/5.8.8/i386-linux-thread-multi/XML/Parser.pm line 187 Any advice would be helpful. Some clue as to how to zero in on line 1, column 1373, byte 1373 of a file would be helpful as well. Thanks

use XML::Simple;
my $xml = new XML::Simple (KeyAttr=>[]);
my $data = $xml->XMLin("cmd-out\/$hostname.xml");
[download]

Comment on Validating XML Download Code

Replies are listed 'Best First'.
Re: Validating XML by ikegami (Patriarch) on Dec 03, 2011 at 22:05 UTC
XML::LibXML has an option to `recover` from errors you might find useful.	[reply] [d/l]
Re: Validating XML by JavaFan (Canon) on Dec 03, 2011 at 23:05 UTC
Some clue as to how to zero in on line 1, column 1373, byte 1373 of a file would be helpful as well. Untested code: `use autodie; use Fcntl ':seek'; open my $fh, "<:raw", "your-file-name-goes-here"; seek $fh, 1373, SEEK_SET; read $fh, $_, 1; # No sysread! print $_, "\n";` [download]	[reply] [d/l]
Re: Validating XML by vinian (Beadle) on Dec 04, 2011 at 06:47 UTC
`use XML::Simple; my $xml = new XML::Simple (KeyAttr=>[]); my $data; eval { $data = $xml->XMLin("cmd-out\/$hostname.xml"); }; if ( $@ ){ die "$@"; }` [download] use `eval` catch error	[reply] [d/l] [select]
Re: Validating XML by grantm (Parson) on Dec 04, 2011 at 23:25 UTC
I use the following Perl 1-liner to help identify the offending part of a broken XML file: `perl -MXML::Parser -e "XML::Parser->new( ErrorContext => 3 )->parsefil +e(shift)" filename.xml` [download] The diagnostic output from XML::Parser includes an ASCII arrow pointing to the bad bit. Of course if the line in question is over 1300 bytes long then that might be a little hard to read :-(	[reply] [d/l]