`prisma pull` should add `///` comments based on postgres's `COMMENT`

Problem

On Postgres database one can have COMMENT ON different database objects (tables, columns, types).

When generating prisma file from such database comments are written in the generated file. Tho doing so would be great!

Suggested solution

When generating prisma file look up object descriptions and use those as comments in generated prisma file.

Example

If your database dump looks like this:

CREATE TYPE foo AS ENUM('fiz', 'foz');
COMMENT ON TYPE foo IS 'foo comment';

CREATE TABLE IF NOT EXISTS bar
(
  id bigserial unique primary key,
  foo foo not null
);

COMMENT ON TABLE bar IS 'bar comment';
COMMENT ON COLUMN bar.id IS 'bar.id comment';

in a Postgres database, and then you run prisma pull, you shold get:

/// bar comment
model bar {
  /// bar.id comment
  id BigInt @id @default(autoincrement())
  foo foo
}

/// foo comment
enum foo {
  fiz
  foz
}

---

Related:

• https://github.com/prisma/prisma/issues/5113
• https://github.com/prisma/prisma/issues/2546

[janpio]

Not right now, now: https://pris.ly/roadmap But that does not mean it could be implemented soon - we are rather conservative how much we plan.

[janpio]

It is just a statement of a current fact: Prisma currently does not support reading or modifying the comments of tables and columns in databases, but we do support // and /// in the Prisma Schema file and use that data in different situations and locations.