Incorrect casing for “ID”

[toastal]

All implicit generated IDs such as the ID for Foo is generated as FooId. While much of the documentation uses “ID”, the code representation is incorrect casing where “ID” is the standard abbreviation form for identifiers (IDs), not “Id”. Changing this would align other correctly-cased variables in the project like DB (along with uses of IO, UUID, & FromJSON from other libs), but still leaves incorrect cases like Sql instead of SQL for the initialism ‘structured query language’ (where documentation is largely using “SQL”).

What I would expect is the implicit ID for Foo becomes FooID… all fooId columns become fooID.

Moving from ‘ID’ to ‘Id’ is losing valuable casing information & if it weren’t important, we wouldn’t be bothering with upper case letters at all in English (& other languages with bicameral scripts).

[toastal]

I attempted to use something like

def :: ImplicitIdDef
def =
    autoIncrementingInteger
        { iidFieldType =
            \entName ->
                FTTypeCon Nothing $ unEntityNameHS entName `mappend` "ID"
        }

to no avail to fix the situation.

[toastal]

What I have not tried: sed -i "s/Id/ID/g" **/*.hs & seen if everything works.

[parsonsmatt]

Yeah, that assumption is baked into a lot of the TH code. I don't anticipate changing that to be easy or particularly flexible.

[toastal]

Is there a combinator route to go, or is Template Haskell the only/preferred way?

[parsonsmatt]

While you can define the PersistEntity stuff yourself, it's not really easy or well-supported.