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
At the moment, taskgraph hardcodes the OptimizationSchema with no way to override it. This means that even though you can register new optimization strategies, the task.py transform will not accept them.
I think rather than copy Gecko here, I'd rather the schema be provided as an argument to the @register_strategy decorator. That way the schema definition lives next to the thing that requires it and we don't need to call extra functions in the registration.
I think rather than copy Gecko here, I'd rather the schema be provided as an argument to the @register_strategy decorator. That way the schema definition lives next to the thing that requires it and we don't need to call extra functions in the registration.
Activity
ahal commentedon Nov 29, 2023
I think rather than copy Gecko here, I'd rather the schema be provided as an argument to the
@register_strategydecorator. That way the schema definition lives next to the thing that requires it and we don't need to call extra functions in the registration.ahal commentedon Nov 29, 2023
See the
group_bydecorator for an example of this:taskgraph/src/taskgraph/util/dependencies.py
Line 38 in 931ab6f
bhearsum commentedon Nov 29, 2023
Sounds very reasonable to me!
Register the optimization schema with `register_strategy`
Register the optimization schema with `register_strategy`