Downgrade CSS Validator failures to warnings.

[mikewest]

In light of #321, it seems reasonable to ignore CSS Validator failures until such time as the validator validates all valid CSS. The alternative is for every bikeshed user to apply for exemptions individually, which seems like a lot of overhead.

[tripu]

(The exception can be defined for all specs, as Denis pointed out in #321 (comment)).

[plehegar]

Given that #321 got closed and resolved, can we close this one?

[tripu]

In favour of closing, too. Waiting for @mikewest's comments.

[mikewest]

This bug was predicated upon the validator being out of date, so if y'all are confident that the validator will be doing the right thing going forward, great!

[tripu]

The CSS validator may misbehave again in the future. When that happens, and as @deniak pointed out, we can ignore that false positive for all specs at once. It is not ideal, but IMHO it's better than ignoring all false positives for all specs pre-emptively, as this issue seems to suggest.

[dontcallmedom]

I would like to revisit the idea of making CSS validation optional in pubrules, since there is yet another case of mis-firing errors in #500 .

While I think it's great to use valid CSS, I don't think it's a particularly important criteria for being able to publish a new draft of a spec; and given the recurrent false positives, I would suggest we leave CSS validation as a post-CR requirement (if at all).

[plehegar]

@fantasai , got an opinion?

[fantasai]

I'm OK with dom's proposal to reduce publishing friction. We should require validation for CR+ if possible, though, and provide exemptions or validator fixes as necessary. That will help to catch errors (in both the specs and the validator).

[tripu]

I don't have a strong opinion about this.

I wonder if these “false positives”, when they happen, have at least the benefit of pushing people into updating the checkers from time to time. That has happened a few times already.

[plehegar]

We shouldn't use pubrules as a way to force the checkers imho. +1 to downgrade css validator for <= CR