Introduce an easy way to block components by amp-consent

This is part of the amp-consent-v2 work. We need a feature to block all amp-ad components on the page in an easy way. So that CMP can manage all ads on the page.

Approach 1: Block all <amp-ad> components on the page if there’s <amp-consent type='cmp_name'> on the page.

• Pro: Requires minimal work for the publisher if they use CMP to manage consent
• Con: Introduces extra delay to every <amp-ad> (even if the ad decides not to be blocked by CMP) because the <amp-ad> will need to wait for the #buildCallback of <amp-consent> element.

Approach 2: Introduce a new meta tag that’s defined by the publisher. Block corresponding components on the page if the meta tag is found.

• Pro: Better performance because all information is in the header. Cleaner code because everything is synchronous.
• Con: Need the publisher’s effort to add meta tag to the header. If the publisher fail to add the meta tag, it will lead to cmp not blocking anything. (It could also be a pro if the CMP does not exist, because the publisher don’t need to individually tag every element they want to block)

After a discussion with @lannka, we decided Approach is more preferable. Please let us know if there’s any early feedback. I’ll post the detail design of the meta tag to the issue later.

@jasti @lannka @torch2424

[jasti]

Based on the AMP principle: the user experience > developer experience > work we have to do. This seems to fall under the user experience bucket because we want to load the ads as quickly as possible without waiting for build callback. Option 2 SGTM.

[zhouyx]

Great! Thank you @honeybadgerdontcare. I’ll try that.