Open nmccready opened 4 years ago
Another argument for this is for use cases when a manual migration must be run. Thus, again losing traceability of what has exactly run.
For example, what if the name
was identical for the run but it was not run off the master branch but a feature.
So currently the
migrations
table hasid
,name
, andrun_on
as columns. I am proposing to either add asql
column to actually save off what has run for traceability.Granted this is pumped to stdout and could just be utilized in your CI architecture.
Even if this is not accepted would it be possible to make
createMigrationsTable
's options more flexible so that custom columns could be added more easily.Seems hardcoded here