Re: Bug#203498: ITP: decss -- utility for stripping CSS tags from an HTML page.

[Sam Hocevar <sam@zoy.org>]

On Wed, Jul 30, 2003, Tollef Fog Heen wrote:

> |    And what is the point of confusing our users and cluttering the package/
> | executable namespace with a useless program that could be replaced with
> | a sed one-liner?
> 
> oh?  what sed one-liner would that be?

   That trivial one, for instance:

> sed -e 's%\(<link[^>]*rel="stylesheet"[^>]*>\|<style>.*</style>\|\(style\|class\|id\)="[^"]*"\)%%g'

   It can be made better, of course. But honestly, the original DeCSS
Perl version is an utter piece of crap, too. I now additionally object
to the ITP on the grounds of poor software quality.

   For instance it fails to remove this construct:

<link type="text/css"
      rel="stylesheet"
      href="/foo.css" />

   And it wrongly removes style="blah" in this one:

<p> Hello, this paragraph is about the famous style="blah" phrase! <p>

   Without a correct HTML parser, such a DeCSS program cannot be
reliable.

Cheers,
-- 
Sam.