I've been working on trying to get the new generics-based stdlib to actually work.
Most of this work is then related to improving the diagnostics around overload resolution, and then filling in gaps in how the generic argument inference system works.
These changes leave the codebase in a state where some things are starting to work, but there needs to be a bit of re-architecting of the generic inference/constraint code (to resolve differences between the "unify" and "join" cases) before we can resolve the remaining issues.
I've been working on trying to get the new generics-based stdlib to actually work.
Most of this work is then related to improving the diagnostics around overload resolution, and then filling in gaps in how the generic argument inference system works.
These changes leave the codebase in a state where some things are starting to work, but there needs to be a bit of re-architecting of the generic inference/constraint code (to resolve differences between the "unify" and "join" cases) before we can resolve the remaining issues.