Re: [Lynx-dev] A patch for lynx.

[David Woolley]

Walter Ian Kaye wrote:

> At 07:25 a +0100 04/20/2007, David Woolley didst inscribe:
> > As a first thing, one should do a much more careful bugwise 
> > compatibility reverse engineering job to see whether, for
> > example, a missing </script> is handled correctly even when
> > <script src=.... /> is not used.
>
> Exactly. :-)

And in fact there is a public attempt to create such a set of parsing
rules, and all the mainstream browser developers, except, possibly,
Microsoft, are involved.  WHATWG is creating a specification to
maximise bugwise compatibility between browsers, HTML5.  The current
assumption seems to be that web authors are incapable of following
standards, so the standards should specify explicit behaviour for
every possible error case.  The result is a large specification,
full of ad hoc rules.

If Lynx really wants to maintain compatibility with real life HTML,
it should be implementing that specification.  That almost certainly
means a complete rewrite of the core of the program, and is going to be
a major job!

The current working specification is at:

http://www.whatwg.org/specs/web-apps/current-work/multipage/section-parsing.html#parsing

Note that WHATWG considers / at the end of an opening tag for both
script and a to be a parse error:

http://www.whatwg.org/specs/web-apps/current-work/multipage/section-tokenisation.html#permitted

and ignores it under all circumstances.  So this change will make Lynx
incompatible with the mainstream browsers.  src does change the content
model for script in HTML5, although I haven't followed through the rules
to make sure that this then results in content other than </noscript>
generating a parse error and recovery.  Seems likely though.