Invalid test: CSS 3 Selectors Test "negation pseudo-class accepts only simple selectors for argument" about nwmatcher HOT 5 CLOSED

Thomas, I agree that having extra features may be a good thing in some cases, therefore I prefer to have that enabled/disabled by developers with a configuration switch in NWMatcher. Forcing the feature enabled by default will break consistency between browser having querySelectorAll() and those not having it.

From the test you can see that this was in the intentions of the test author, see here:

http://disruptive-innovations.com/zoo/css3tests/selectorTest.html#target

I instead suggest you also have a configuration helper to be able to control that at will and by default follow specifications. Having consistent results cross-browser is an objective we can't miss.

from nwmatcher.

Simply put, that test is invalid.

It is going beyond the W3C Recommendation into the realm of personal opinion.

We should all be supporting the standards as they are defined without limiting new ideas and extra features.

If you won't remove that test, please at least add a note that mentions that it is a contested test. That way people will not get the impression that an implementation that 'fails' that test is inferior to an implementation that 'passes' it.

from nwmatcher.

As I said that test is the official Selector API test suite, not one I invented.

My car can run at more than 200 Km/hour, still there is a 110 Km/hour limit on the roads and the car must be built in a way the user can respect that limit. Sounds ?

from nwmatcher.

Oh...forgot... I will happily put a notice in the "css3-compat" test for that, as I did for other contested tests like the ":root" pseudo-class.

from nwmatcher.

Ok... I am closing this since everything is now configurable in NWMatcher:
NW.Dom.configure({ SIMPLENOT: false });
will let developers use complex selectors in :not() pseudo-class.

from nwmatcher.