I assume that a lot of get-data users will want to get the whole blob, and I assume this is done by passing start = 0, end = u64::MAX, so I assume that an end beyond the length of the blob is okay and we just return as much as exists.
But that feels like a lot of assuming and it would be good to see it specified.
start
If start is past the end of the blob, do the incoming-value functions return an empty list/stream, or is that an error? If an error, is it reported at get-data time, or when an incoming-value function is called, or (in the async case) when the input-stream is first read?
end
I assume that a lot of
get-datausers will want to get the whole blob, and I assume this is done by passing start = 0, end = u64::MAX, so I assume that anendbeyond the length of the blob is okay and we just return as much as exists.But that feels like a lot of assuming and it would be good to see it specified.
start
If
startis past the end of the blob, do theincoming-valuefunctions return an empty list/stream, or is that an error? If an error, is it reported atget-datatime, or when anincoming-valuefunction is called, or (in the async case) when theinput-streamis first read?