m-creations / cl-migrations Goto Github PK

Each migration script is numbered sequentially (prefix of the file name like 1-some-migration) when migration script is generated with CL-MIGRATIONS:GENERATE. This breaks feature branch development.

Example

Developer A works in feature branch f1 and generate a new migration: 3-add-email-to-user.lisp. Then she runs (cl-migrations:migrate) and the user table gets its email column.

Simultaneously, one of her fellow co-workers (Dev B) works on feature branch f2 and adds a new migration 3-create-coffee-table.lisp.

Now, if A checks out the feature branch f2 of B (for testing purposes) and runs (cl-migrations:migrate) nothing happens. Because A is already on version 3 (she ran the migrations of her own f1 branch).

Which is kind of sub-optimal. Further, I assume (not tested yet) that this has also implications once both branches has been merged into main line.