Using XML::LibXML in Perl to alter XML file: newline chars are Unix, not Windows

Question

Dear Perl and XML gurus

I have a task to update values inside XML file using XPath.
I use XML::LibXML library in Perl to read, alter and save XML file:

# Read XML file
my $parser = XML::LibXML->new();
my $doc = $parser->load_xml(location => $config_file);
my $root = $doc->documentElement();

# Alter nodes\attributes
foreach (keys %environment_values) {
    my @nodes = $root->findnodes($_);
    if (scalar @nodes < 1) {
        print "ERROR: element not found in $config_file by XPath: $_\n";
        die;
    } elsif (scalar @nodes > 1) {
        print "ERROR: more than 1 element (" . scalar @nodes . ") is found in $config_file by XPath: $_\n";
        die;
    }
    my $node = $nodes[0];
    if  ($node->nodeType == XML_ELEMENT_NODE) {
        $node->removeChildNodes();
        $node->appendText($environment_values{$_});
    } elsif ($node->nodeType == XML_ATTRIBUTE_NODE) {
        $node->setValue($environment_values{$_});
    } else {
        print "ERROR: unknown node type: " . $node->nodeType . "\n";
        die;
    }
}

# Save the resulting XML file
open (my $fh, '>:raw', $config_file) or die $!;
print $fh $doc->toString();
close $fh;

Although it generates a file very similar to the original one, there are still a couple of nuisances:

1. Newline characters (line endings) are Unix-style, although original file has them in Windows-style.
2. Space before the closing /> gets removed, e.g. <node /> becomes <node/>.

Any chance to fix these? I'm hoping to get exactly the same XML file as the original one, with only differences being attribute values I'm amending...

P.S. I really like how simple is <xmlpoke> in NAnt. But have to use Perl for this work.

Answer 1

In general you're not going to get exactly what you seem to be asking for - e.g. the distinction between single and double quotes round attribute values will be lost as well as spaces inside tags.

Best might be to read in with Perl once and write out without changes, and then run your script, and compare those two files.

Answer 2

I think the answer to newline may be in the mode you are using.

As per http://perldoc.perl.org/PerlIO.html#Defaults-and-how-to-override-them :

If the platform is MS-DOS like and normally does CRLF to "\n" translation for text files then the default layers are :

   unix crlf

(The low level "unix" layer may be replaced by a platform specific low level layer.)