Separating `schema_migration` from the search_path

Description

I have a security situation where I can not allow the schema_migration table to be located in the same schema that my tables and such are located in.

I've tried a couple of different ways to do this ( have the schema_migration in its own schema )

My first method was to specify the DBMATE_MIGRATIONS_TABLE by prepending the schema name.

file: .env

DATABASE_URL="postgres://postgresUser:postgresPW@localhost:5432/test_database?search_path=test_schema&sslmode=disable"
DBMATE_MIGRATIONS_TABLE=dbmate.schema_migration

In this case I expected the schema_migration table be in a new dbmate schema AND it is. I expected my tables to be created in the test_schema BUT my new tables are in the dbmate schema

Trying a second method, i see in the documentation that.

If multiple comma-separated schemas are passed, the first will be used for the schema_migrations table.

I assumed then if the first schema is for the schema_migrations table then the second would be used for my objects like tables etc.

My .env I tried the following

DATABASE_URL="postgres://postgresUser:postgresPW@localhost:5432/test_database?search_path=dbmate,test_schema&sslmode=disable"

But again everything is created in the dbmate schema.

I think dbmate is working as designed and the issue is that I'm trying to something that is not supported. A schema_migrations schema name variable would be nice to be able to explicitly put and use the schema_migrations table in another schema and not effect the search_path.

• Version: 2.2.0
• Database: psql (14.7 (Homebrew), server 15.2 (Debian 15.2-1.pgdg110+1))
• Operating System: MacOS Ventura 13.2 (22D49)

Steps To Reproduce

Expected Behavior