Enabling pointer in config file for a custom type, doesn't work with `NOT NULL`

[adoublef]

Version

1.14.0

What happened?

Reading the docs here that its possible to override column type. For an in-built type like time.Time using the following configuration:

{
  "db_type": "timestamp",
  "go_type": {
    "import": "time",
    "type": "Time",
    "pointer": true
  },
  "nullable": true
}

Allows me to have a pointer as a struct field if I dont include NOT NULL in my schema.sql file:

create table example1 (
    /*...*/    
    createdAt timestamp
)

create table example2 (
    /*...*/    
    createdAt timestamp not null
)

produce these types:

type Example1 struct {
    /*...*/
    CreatedAt *time.Time
}

type Example2 struct {
    /*...*/
    CreatedAt time.Time
}

This does not work when overriding a column type with a package level type:

package mypkg

type Username string

{
    "column": "user.username",
    "go_type": {
        "import": "github.com/.../mypkg",
        "type": "Username",
        "pointer": true
    },
    "nullable": true
}

create table example3 (
    /*...*/    
    username varchar(255)
)

create table example4 (
    /*...*/    
    username varchar(255) not null
)

the generated types are both using pointers, even though, I would have expected, as with the previous example, it only be a pointer only if it could be nullable.

type Example3 struct {
    /*...*/
    Username *mypkg.Username
}

type Example4 struct {
    /*...*/
    Username *mypkg.Username
}

Is there a config missing or is this something that could be fixed as I am unsure via the documentation, if this is intended behaviour or not. This affects both mysql and psql engine.

Relevant log output

No response

Database schema

create table example4 (   
    username varchar(255) not null
)

SQL queries

No response

Configuration

{
    "column": "user.username",
    "go_type": {
        "import": "github.com/myrepo/mypkg",
        "type": "Username",
        "pointer": true
    },
    "nullable": true
}

Playground URL

No response

What operating system are you using?

Linux

What database engines are you using?

MySQL

What type of code are you generating?

Go

[kyleconroy]

Reproduced on the playground

[kyleconroy]

Reading over this again, this is working as intended. By using a column override, you're telling sqlc to always use the go_type specified. The nullable field is ignored for columns. Since nickname can never be null, you can remove the pointer field and get the code that you want.

https://play.sqlc.dev/p/01ee02b8f16afeb2788ae7d44c258b619fd348dcd921e06dd82947a9d89d7a86