Git master can already avoid rewriting the table for column type changes like
varchar(10) -> varchar(20). However, if the column in question is on either
side of a FK relationship, we always revalidate the foreign key. Concretely,
I wanted these no-rewrite type changes to also assume FK validity:
- Any typmod-only change
- text <-> varchar
- domain <-> base type
To achieve that, this patch skips the revalidation when two conditions hold:
(a) Old and new pg_constraint.conpfeqop match exactly. This is actually
stronger than needed; we could loosen things by way of operator families.
However, no core type would benefit, and my head exploded when I tried to
define the more-generous test correctly.
(b) The functions, if any, implementing a cast from the foreign type to the
primary opcintype are the same. For this purpose, we can consider a binary
coercion equivalent to an exact type match. When the opcintype is
polymorphic, require that the old and new foreign types match exactly. (Since
ri_triggers.c does use the executor, the stronger check for polymorphic types
is no mere future-proofing. However, no core type exercises its necessity.)
These follow from the rules used to decide when to rebuild an index. I
further justify them in source comments.
To implement this, I have ATPostAlterTypeParse() stash the content of the old
pg_constraint.conpfeqop in the Constraint node. ATAddForeignKeyConstraint()
notices that and evaluates the above rules. If they both pass, it omits the
validation step just as though skip_validation had been given.
Thanks,
nm
