Changing to CopyTo aligns more with the idiomatic way of using spans. It also obviates the need for ISpanVisitor weirdness that made the logic flow for deserializing data from the data source confusing.
IDataSource now has a method CopyNodeBytesTo that takes in a hash and a buffer to populate. This solidifies the IDataSource as being concerned only with storing raw data, and separates the concern of dealing with "real" types to the serializer, which gets the raw bytes then deserializes them to a "real" type.
Changing to
CopyTo
aligns more with the idiomatic way of using spans. It also obviates the need forISpanVisitor
weirdness that made the logic flow for deserializing data from the data source confusing.IDataSource
now has a methodCopyNodeBytesTo
that takes in a hash and a buffer to populate. This solidifies theIDataSource
as being concerned only with storing raw data, and separates the concern of dealing with "real" types to the serializer, which gets the raw bytes then deserializes them to a "real" type.PANDO-29