I've not had time for a thorough look at this, but on a quick scan I noticed that CompareIndexInfo() missed checking if the Index AM matches the AM of the partitioned index.
Testing with:
create table p (a int not null) partition by range (a);
create table p1 partition of p for values from (1) to (10);
create table p2 partition of p for values from (10) to (20);
create index on p1 using btree (a);
create index on p2 using hash (a);
create index on p (a);
I see it ends up making use of the hash index on p2 to support the index that's stored as a btree on the partitioned table. I think these should match so that the operations we can perform on the index are all aligned.
