Consider ordering of tag-value pairs; ignore unknown tags

[ryansch]

From the spec:

A DMARC policy record MUST comply with the formal specification found
in Section 5.3 in that the "v" and "p" tags MUST be present and MUST
appear in that order.  Unknown tags MUST be ignored.  Syntax errors
in the remainder of the record SHOULD be discarded in favor of
default values (if any) or ignored outright.

We are currently enforcing the order of all of the tags (not just v and p). We also blow up on unknown tags.

I've added a failing test with yahoo's dmarc record as an example of the pct tag being out of order.

[postmodern]

So we need to move back to the Array of Hashes model? For unknown tags, we could add a catch rule. Not sure how we'd safely recover from syntax errors within tags. Maybe we'd have to split by ';' first, and parse each tag individually?

[ryansch]

I'm not sure. I'm heads down on some other stuff for the rest of the week. I'll have another look early next week.

[ryansch]

I've reached out to the parslet community on the mailing list for ideas.

[ryansch]

Possible example:

rule(:a) { a_properly_formed_a | consume_anything_upto_the_next_tag_start } 
rule(:consume_anything_upto_the_next_tag_start) {  (start_tag_token.absent? >> any).repeat(1) }

[postmodern]

Ah clever. Instead of any how about match["^#{dmarc_sep}"].repeat(1)?

[postmodern]

Closing/Recreating this now that we have the Alexa Top 500 gauntlet tests.