Bug #37032 for XML-Writer: Numbered character references are botched

Mon Jun 23 16:22:29 2008 canipe_chris [...] yahoo.com - Ticket created

Subject:

Numbered character references are botched

The hex and decimal formats of "&#x...;" and "&#...;" become "&#x...;" and "&#...;" when written. I realize the raw method exists, but these are valid in XML and should not require the UNSAFE setting. The following code yields "<test>r&#xe9;sum&#233;</test>": use XML::Writer; my $string = 'résumé'; my $XML = XML::Writer->new(); $XML->dataElement('test', $string); $XML->end(); I'm running Perl 5.8.8 on SunOS 5.8. Thanks.

Mon Jun 23 17:05:26 2008 joe [...] kafsemo.org - Correspondence added

This is as intended - see examples/double-escaping-example.pl for a quick sample of why it has to behave in this way. Try: my $string = "r\x{e9}sum\x{e9}"; my $XML = XML::Writer->new(ENCODING => 'us-ascii'); for the output you want.

Mon Jun 23 17:05:28 2008 The RT System itself - Status changed from 'new' to 'open'

Mon Jun 23 17:05:29 2008 joe [...] kafsemo.org - Status changed from 'open' to 'rejected'

Tue Jun 24 18:39:24 2008 canipe_chris [...] yahoo.com - Correspondence added

Subject:	Re: [rt.cpan.org #37032] Numbered character references are botched
Date:	Tue, 24 Jun 2008 13:47:40 -0700 (PDT)
To:	bug-XML-Writer [...] rt.cpan.org
From:	chris canipe <canipe_chris [...] yahoo.com>

Joseph, There's no telling where data may be coming from these days. As a result, should the module not use a lookahead to determine which ampersands to escape? For example, s/&(?!#x?\d+;)/&/ would escape "Father & Son," but not "Father & Son" or "Father & Son." This could be expanded to check for amp, lt, gt, apos, and quot as well. Thanks, Chris --- On Mon, 6/23/08, Joseph Walton via RT <bug-XML-Writer@rt.cpan.org> wrote: From: Joseph Walton via RT <bug-XML-Writer@rt.cpan.org> Subject: [rt.cpan.org #37032] Numbered character references are botched To: canipe_chris@yahoo.com Date: Monday, June 23, 2008, 5:05 PM <URL: http://rt.cpan.org/Ticket/Display.html?id=37032 > This is as intended - see examples/double-escaping-example.pl for a quick sample of why it has to behave in this way. Try: my $string = "r\x{e9}sum\x{e9}"; my $XML = XML::Writer->new(ENCODING => 'us-ascii'); for the output you want.

Tue Jun 24 18:39:30 2008 The RT System itself - Status changed from 'rejected' to 'open'

Tue Jun 24 20:42:15 2008 joe [...] kafsemo.org - Correspondence added

Please see: http://rt.cpan.org/Ticket/Display.html?id=17862#txn-236936 I'm sure you can think of other cases where silently modifying data may cause problems - round-tripping '& & &' to '& & &', for example.

Tue Jun 24 20:42:32 2008 joe [...] kafsemo.org - Status changed from 'open' to 'rejected'