Add support for constraint= attribute to override foreign reference constraint names#937
Conversation
…quired to support constraint= attribute for foreign key references.
…tation to Persistent-entity-syntax.md
parsonsmatt
left a comment
parsonsmatt
left a comment
There was a problem hiding this comment.
Code looks great! Version bump sounds reasonable.
Could this break anything? You're only adding a bit of functionality. Old versions of persistent will ignore the constraint=. It will generate a migration to rename a constraint if the constraint= is used, I think, but that's expected.
I don't think it can break anything on further thought. I'll merge this in later today. Thanks!
| AnotherVeryLongTableName | ||
| veryLongTableNameId VeryLongTableNameId constraint=short_foreign_key | ||
| ``` | ||
|
|
There was a problem hiding this comment.
Thanks for extending the docs!
| pure (DBName x, constraintName) | ||
| | Just x <- T.stripPrefix "constraint=" a = do | ||
| tableName <- fst <$> (ref c fe as) | ||
| pure (tableName, DBName x) |
There was a problem hiding this comment.
Oh, I see how this is implemented now! It doesn't require a breaking change addition to a datatype, which is great. It adds a new feature to the QuasiQuoter syntax, but doesn't change the persistent-template library at all.
So if you have a QuasiQuoter block with constraint=fk_name on an old version of persistent, then it'll get ignored. That's okay.
| _ -> [] | ||
| refAdd = case (ref == ref', ref) of | ||
| (False, Just (tname, _cname)) -> [(tname, addReference allDefs (refName tblName name) tname name)] | ||
| (False, Just (tname, cname)) -> [(tname, addReference allDefs cname tname name)] |
There was a problem hiding this comment.
This looks great to me. I think I would like to see it included in the other SQL backends, but that can come in a later PR.
3 participants
This change adds support for a
constraint=attribute that allows users to override the automatically generated foreign reference constraint names with their own. The motivation for this change is that we had two instances of very long table/column names that led to automatically generated constraint names exceeding 64 characters, which is not allowed in MySQL. Without a change like this we could not use theprintMigrationsoutput.Please let myself (@charukiewicz) or @belevy know if you have suggestions for a better solution or alternative approach.
Before submitting your PR, check that you've:
@sincedeclarations to the HaddockAfter submitting your PR: