mirror of
https://github.com/carbon-language/carbon-lang.git
synced 2026-11-17 19:30:30 +00:00
We plumb through the ID so that we can later see if a requirement was in the IdentifiedFacetType, and potentially grab a final witness from the witness source if it's a FacetValue. Instead of plumbing through the IdentifiedFacetTypeId, collect all the requirements up front, and the final witnesses if they exist. We want to do this because in future work we will need to subst/evaluate the outputs of IdentifiedFacetType to use it since it will contain IdentifiedWitnesses, and we'd like to do that once up front instead of multiple places. That is more efficient and ensures less time/code with access to the intermediate state of the IdentifiedFacetTypeId.