Going forward, I envision two issues with the current state of affairs:
the current "transpiler" (aka. generator) is not performing any checks for sanity of the passed yaml files. This results in the potential to ruin the data model. The transpiler code as it is now requires some relevant refactoring before proper checks are put in place.
Soon we'll face the problem of upgrading files compliant with an old file format, and port them to the new file format. The current design makes it extremely hard to do. I don't have a valid solution for this, because it would require a major investment in a migration strategy, and the generated code is able to handle the migration options and take them into account when it's time to read a file that is older than what the data model can read.
Going forward, I envision two issues with the current state of affairs:
the current "transpiler" (aka. generator) is not performing any checks for sanity of the passed yaml files. This results in the potential to ruin the data model. The transpiler code as it is now requires some relevant refactoring before proper checks are put in place.
Soon we'll face the problem of upgrading files compliant with an old file format, and port them to the new file format. The current design makes it extremely hard to do. I don't have a valid solution for this, because it would require a major investment in a migration strategy, and the generated code is able to handle the migration options and take them into account when it's time to read a file that is older than what the data model can read.
Just food for thoughts.