RPF moves away from a regex-based parsing model, to a LR parser where the input string is first lexed for space-delimited tokens, and then parsed as a newline-delimited string. This fixes all kinds of problems, including queries that confused the regex:
homer (a panda) versus home (a tag)
credit 218_yu (I think the number at the beginning of this query was confusing the regex parser)
Right now this is a one-for-one replacement of the old query parser, but it's expansible enough that I can finally support things like "multiple-tag" search, set intersections, and family relationship queries!
RPF moves away from a regex-based parsing model, to a LR parser where the input string is first lexed for space-delimited tokens, and then parsed as a newline-delimited string. This fixes all kinds of problems, including queries that confused the regex:
homer(a panda) versushome(a tag)credit 218_yu(I think the number at the beginning of this query was confusing the regex parser)Right now this is a one-for-one replacement of the old query parser, but it's expansible enough that I can finally support things like "multiple-tag" search, set intersections, and family relationship queries!