A part past the second comma holds no name word (rules.md#C2), so since #636 group makes no particle chain there. P5's bound given-name join still runs there after a suffix comma. Group passes BoundJoin.STRICT to every segment of a structure that isn't a family comma, so in a tail part abdul joins the word after it into one piece. When that word is a title, the joined piece is no title word for C2 to read:
| Input |
Today |
John Smith, Jr., Berg Dr. Smith |
title 'Dr.', suffix 'Jr., Berg Smith' |
John Smith, Jr., abdul Dr. Smith |
suffix 'Jr., abdul Dr. Smith' |
Smith, John, abdul Dr. Smith (family comma; the join is disabled in its tail) |
title 'Dr.', suffix 'abdul Smith' |
1.4.0 reads both suffix-comma forms wholly as suffixes. #636's review found this: on #636's branch, John Smith, Jr., abdul St Prof. reads title 'St Prof.' while John Smith, Jr., abdul St Secretary of State reads suffix 'abdul St'.
Proposal: disable the join in a tail part, as #636 did the chain, on the same argument: a bound given-name pair is a name unit, and nothing in such a part is read as a name. It's garbage input either way, so the question is whether consistency is worth the change.
A part past the second comma holds no name word (rules.md#C2), so since #636 group makes no particle chain there. P5's bound given-name join still runs there after a suffix comma. Group passes
BoundJoin.STRICTto every segment of a structure that isn't a family comma, so in a tail partabduljoins the word after it into one piece. When that word is a title, the joined piece is no title word for C2 to read:John Smith, Jr., Berg Dr. SmithJohn Smith, Jr., abdul Dr. SmithSmith, John, abdul Dr. Smith(family comma; the join is disabled in its tail)1.4.0 reads both suffix-comma forms wholly as suffixes. #636's review found this: on #636's branch,
John Smith, Jr., abdul St Prof.reads title 'St Prof.' whileJohn Smith, Jr., abdul St Secretary of Statereads suffix 'abdul St'.Proposal: disable the join in a tail part, as #636 did the chain, on the same argument: a bound given-name pair is a name unit, and nothing in such a part is read as a name. It's garbage input either way, so the question is whether consistency is worth the change.