Should "Maria Begona Lopez" under locales.ES keep Begona out of the family, as "María Begoña López" does? #1081

locales.ES and locales.ES_AR read two family names (rules/O6) with a list of given names keeping a listed given name out of the family. An entry written without accents matches the word with them (ines matches Inés), but the list's ñ is kept, so data that drops the tilde matches no entry:

Measured on master (25e698a plus #1079). ES's list holds 4 entries with ñ (begoña, beñat, iñaki, zuriñe), and ES_AR's holds 13 (iñigo among them).

Proposed rule shape (an Except for O6's given-name veto):

• When: a word standing where O6 would take a second family name is spelled like a listed given name with ñ written as n (Begona for begoña).
• Then: it is read as the listed given name, and the name reads as with the switch off for that word (middle Begona, family Lopez).
• Because: data from systems without ñ (ASCII exports, upper-case registries) writes n; an accent already folds this way, and the tilde is a writing of the same letter in that data.
• Boundary: a word the list does not hold under either spelling is taken as before: Ana Ordoñez Lopez and Ana Ordonez Lopez both read family Ordoñez Lopez / Ordonez Lopez under ES today, and keep it.

Open questions:

• Whether folding ñ can make a surname match a given-name entry: Peña/Pena and Muñoz/Munoz are surnames, but no entry spells either. A check against INE's surname table would settle it.
• Whether the fold belongs to the list's matching only (given_names), or to the vocabulary's _normalize generally, which would also move every other kind's ñ words.

Not in scope: the per-region list question #959 also raised; Argentina's is answered by locales.ES_AR (#1061).