Bug #69180 for XML-Atom: XML::Atom chokes on doctypes--from Google Calendar, for example

Wed Jun 29 09:20:33 2011 len.budney [...] gmail.com - Ticket created

Subject:

XML::Atom chokes on doctypes--from Google Calendar, for example

Google Calendar returns atoms that begin with "<!doctype html>". I know that's utterly pointless, and it may be a violation of the Atom spec--I haven't checked that--but it results in a fatal error using XML::Atom to parse Google's feeds. XML::Atom needs to handle general SGML directives like doctype gracefully. If it can't, then it should at least accommodate some pretty important feeds by stripping the doctype statement before attempting to parse.

Wed Jun 29 09:45:48 2011 miyagawa [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #69180] XML::Atom chokes on doctypes--from Google Calendar, for example
Date:	Wed, 29 Jun 2011 09:45:19 -0400
To:	bug-XML-Atom [...] rt.cpan.org
From:	Tatsuhiko Miyagawa <miyagawa [...] gmail.com>

On Wed, Jun 29, 2011 at 9:20 AM, Leonard R Budney via RT <bug-XML-Atom@rt.cpan.org> wrote: Show quoted text

> Google Calendar returns atoms that begin with "<!doctype html>". I know that's utterly > pointless, and it may be a violation of the Atom spec--I haven't checked that--but it results in a > fatal error using XML::Atom to parse Google's feeds. > > XML::Atom needs to handle general SGML directives like doctype gracefully. If it can't, then it > should at least accommodate some pretty important feeds by stripping the doctype statement > before attempting to parse.

I disagree. Determining which feeds are "important" is totally arbitrary, and for generic purpose modules like XML::Atom it's dangerous thing to do and be even prune to security hole. That said, you should take a look at XML::Liberal first, and see if it can "sanitize" the broken XML like that before passing it to XML::Atom. -- Tatsuhiko Miyagawa

Wed Jun 29 09:45:49 2011 The RT System itself - Status changed from 'new' to 'open'

Wed Jun 29 09:48:27 2011 MIYAGAWA [...] cpan.org - Status changed from 'open' to 'rejected'

Wed Jun 29 10:14:34 2011 len.budney [...] gmail.com - Correspondence added

In this case, it's ALWAYS valid for XML to include a doctype; failing to handle that is a bug-- despite my agreement that doctypes are rather superfluous in XML and that spuriously adding them is a mistake. When a standard (like Atom) says "Foo is an XML document which...." then you're on the hook for supporting valid XML which...

Wed Jun 29 10:14:34 2011 The RT System itself - Status changed from 'rejected' to 'open'

Wed Jun 29 10:20:02 2011 miyagawa [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #69180] XML::Atom chokes on doctypes--from Google Calendar, for example
Date:	Wed, 29 Jun 2011 10:19:34 -0400
To:	bug-XML-Atom [...] rt.cpan.org
From:	Tatsuhiko Miyagawa <miyagawa [...] gmail.com>

You said you haven't checked if it validates on feed validator. Why not do it now. If it doesn't validate, there's no reason for XML::Atom to support them. On Wed, Jun 29, 2011 at 10:14 AM, Leonard R Budney via RT <bug-XML-Atom@rt.cpan.org> wrote: Show quoted text

> Queue: XML-Atom > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=69180 > > > In this case, it's ALWAYS valid for XML to include a doctype; failing to handle that is a bug-- > despite my agreement that doctypes are rather superfluous in XML and that spuriously adding > them is a mistake. > > When a standard (like Atom) says "Foo is an XML document which...." then you're on the hook for > supporting valid XML which... >

-- Tatsuhiko Miyagawa