The proposal is to make the flag in TypeCollectionTimeseriesFields to work like:

• If the collection may have mixed-schema buckets (mustConsiderMixedSchemaBucketsInReads == true), then set timeseriesBucketsMayHaveMixedSchemaData to true.
• If the collection does not have mixed-schema buckets (mustConsiderMixedSchemaBucketsInReads == false), then do not set the timeseriesBucketsMayHaveMixedSchemaData, i.e. leave it as none.

If we eventually also use TypeCollectionTimeseriesFields inside CollectionOptions in the shard catalog, then having none mean "no mixed-schema buckets" is convenient because e.g. it naturally won't "pollute" the output of listCollections for the 99.9% of customers without mixed-schema data