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
Currently the status computation (of both PR and staging) depend on repository.status_ids.* (essentially).
This means when adding, removing, or modifying a status configuration every PR, batch, and staging for the repository gets retrieved and updated even though status configurations are only relevant for:
pending stagings
non-merged non-closed PRs
active non-merged batches
This causes a huge amount of recomputation and commonly a MemoryError for a somewhat innocuous change.
Now the easy option here would be to just remove those dependencies and not recompute anything, only new PRs and stagings would take the change in account and TBF for most situations that's probably what we want anyway?
Alternatively, create a search-only m2m for the relationships, such that only the "active" records get selected and recomputed.
The text was updated successfully, but these errors were encountered:
Currently the status computation (of both PR and staging) depend on
repository.status_ids.*
(essentially).This means when adding, removing, or modifying a status configuration every PR, batch, and staging for the repository gets retrieved and updated even though status configurations are only relevant for:
This causes a huge amount of recomputation and commonly a MemoryError for a somewhat innocuous change.
Now the easy option here would be to just remove those dependencies and not recompute anything, only new PRs and stagings would take the change in account and TBF for most situations that's probably what we want anyway?
Alternatively, create a search-only m2m for the relationships, such that only the "active" records get selected and recomputed.
The text was updated successfully, but these errors were encountered: