[tamnd]

Summary

This adds the match operator @@ of text search for tsvector and tsquery, as ts_match_vq and TS_execute of tsvector_op.c have it. The functions that need a text search configuration, such as to_tsvector, are not in this change.

Changes

• rupg-func has the new module tsmatch, with ts_match_vq and ts_match_qv. The operators @@ and @@@ use these functions in both argument orders.
• An operand matches a lexeme with one of the weights of the operand. A lexeme with no positions matches each weight, as in PostgreSQL.
• A prefix operand such as cat:* matches each lexeme that starts with the prefix. Under <N>, the positions of all these lexemes are merged.
• The operator <N> follows TS_phrase_execute. It has the rules of PostgreSQL for !, & and | under <N>, and the width of a match.
• A lexeme with no positions under <N> gives the result "maybe", as TS_MAYBE does. At the top of the query, <N> with this result does not match.
• A position keeps 16 bits when the distances are added, as a WordEntryPos does. Thus a large <N> gives the same result as PostgreSQL.
• An empty query matches no vector.

Tests

• The session test ts_match is the output of PostgreSQL 19 for the same script in a new database. It has weights, prefixes, !, & and | inside and outside <N>, positions near the limit, vectors with no positions, GROUP BY on a match, @@@ and a query with parameters.
• A unit test for the match of each type of operand and operator.
• fmt, clippy, the workspace tests and the checks of xtask pass on server1.