You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Oct 2, 2022. It is now read-only.
It has a shortcoming in the situation I find myself in: We have several people working on their own local DBs, but not in a linear fashion. They are all working on their own branches and adding columns/tables as and when needed. With the current numerical approach to revisions it would be very difficult for them to not use the same revision number when adding a revision and also there is no way to keep track of "missed" revisions as it only works on numbers, so they can't just jump ahead a few numbers to make sure they don't collide.
My proposal would be to allow the following: Remove the numeric requirement for revisions. Users create their revisions with a unique identifier that is highly unlikely to collide with another users revision on a different branch, so for example "2018-03-22_11-12_AddTreeHutCountToApiKeys"
When it comes to revision application, the revision file in data/meta should contain the list of revisions that have already been applied locally, as opposed to just the latest revision number applied. That way revisions made by other users that get merged into a mainline branch could be applied later.
Your thoughts?
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Love this project.
It has a shortcoming in the situation I find myself in: We have several people working on their own local DBs, but not in a linear fashion. They are all working on their own branches and adding columns/tables as and when needed. With the current numerical approach to revisions it would be very difficult for them to not use the same revision number when adding a revision and also there is no way to keep track of "missed" revisions as it only works on numbers, so they can't just jump ahead a few numbers to make sure they don't collide.
My proposal would be to allow the following: Remove the numeric requirement for revisions. Users create their revisions with a unique identifier that is highly unlikely to collide with another users revision on a different branch, so for example "2018-03-22_11-12_AddTreeHutCountToApiKeys"
When it comes to revision application, the revision file in data/meta should contain the list of revisions that have already been applied locally, as opposed to just the latest revision number applied. That way revisions made by other users that get merged into a mainline branch could be applied later.
Your thoughts?
The text was updated successfully, but these errors were encountered: