Adding a new migration when there is only one present results in broken code

[darricheng]

Description

When there is only one migration in the /migration/src/lib.rs file, adding a new migration via CLI results in broken code.

Before:

#![allow(elided_lifetimes_in_paths)]
#![allow(clippy::wildcard_imports)]
pub use sea_orm_migration::prelude::*;

mod m20220101_000001_users;

pub struct Migrator;

#[async_trait::async_trait]
impl MigratorTrait for Migrator {
    fn migrations() -> Vec<Box<dyn MigrationTrait>> {
        vec![Box::new(m20220101_000001_users::Migration)]
    }
}

After:

#![allow(elided_lifetimes_in_paths)]
#![allow(clippy::wildcard_imports)]
pub use sea_orm_migration::prelude::*;

mod m20220101_000001_users;

mod m20241020_051916_tests;
pub struct Migrator;

#[async_trait::async_trait]
impl MigratorTrait for Migrator {
    fn migrations() -> Vec<Box<dyn MigrationTrait>> {
            Box::new(m20241020_051916_tests::Migration),
        vec![Box::new(m20220101_000001_users::Migration)]
    }
}

To Reproduce

1. Remove all but one migration in the /migration/src/lib.rs file as shown above.
2. Run a command that generates a migration, e.g.cargo loco generate model test name:string

Expected Behavior

The new migration should be added into the vec![] macro correctly.

Misc

I'm willing to contribute a fix, though I'll need some guidance to find the source of the issue because I'm still unfamiliar with the codebase.

[jondot]

Thanks! looks like something we should fix.
I'm wondering how did you get a single migration? (most of the starters have multiple)

[jondot]

Look here: 26f0069
This should give you a sense for how to fix generator related issues in the future :)
Thanks for catching this!

[darricheng]

I was following the file changes in #846 on what files and code to edit, so that I could remove the notes model and related code from the saas starter.