Removing the other (valid) points to focus on the part I'm interesting in...

Abel Rionda wrote:
> -I do no see a way to deactivate these features in our CSS validation
> tool (CSS Validator makes a complete grammar validation) and making an
> ad-hoc solution would tale a considerable development effort.
> Particularly, STYLE_SHEETS_USE test (currently these subtests are not
> made) is going to be hard to implement basing on the checker design and
> our limited time to dedicate to this.

Either I don't understand the problem, which is totally possible - I 
wonder if was even born at the time of the first Last Call - either the 
CSS validator we use already is correctly configured. But there is a bug 
in the checker I would say.

I created a simple test:

This is what the embedded CSS defines, and e embedded CSS defines:
  body { color: black; } -> good
  p { font-size: whatever; } -> unknown value, checker should raise a 
warning under STYLE_SHEET_USE
  ul { word-spacing: 42things; } -> unknown unit, checker should raise a 
warning under STYLE_SHEET_USE
  li { yeepee: 20px; } -> unknown property, checker should ignore the 
property and the value

The whole thing should not trigger any FAIL, but raise two WARN, if I 
understand things correctly.
It does trigger a FAIL currently on "yeepee: 20px" because of the use of 
an absolute unit since "yeepee" is an unknown property, but that's a bug.

In particular, the checker does not fail on the unknown properties and 
values because of the underlying CSS validator.
Did I miss something?
Would you have a simple example in that case, so that I understand what 
we're talking about?


Received on Friday, 4 July 2008 16:04:18 UTC