My general plan here is a little weird, but it's to help keep this piece-meal. I've updated most of the code to support multiple-schemas. Nevertheless, I'm leaving it gated in this PR because I haven't actually added any tests.
In a follow-up PR, I will actually add acceptance tests and ungate the behavior. This might find a few more bugs in multi-schema support, but I don't suspect there will be many.
I could do one larger PR, but I think it's acceptable to do it this way, since our current acceptance test suite would identify any regressions introduced.
Description
My general plan here is a little weird, but it's to help keep this piece-meal. I've updated most of the code to support multiple-schemas. Nevertheless, I'm leaving it gated in this PR because I haven't actually added any tests.
In a follow-up PR, I will actually add acceptance tests and ungate the behavior. This might find a few more bugs in multi-schema support, but I don't suspect there will be many.
I could do one larger PR, but I think it's acceptable to do it this way, since our current acceptance test suite would identify any regressions introduced.
Motivation
https://github.com/stripe/pg-schema-diff/issues/94
Testing
Acceptance tests