Alec Zorab

It's not obvious to me why the context manager doesn't call `stream.cancel()` (ie an explicit hangup, rather than just signalling that we're not going to send more data). Once we...

a flag on `open` to indicate that the context manager should cancel on exit could be really useful here

perhaps a better indicator of the differences: ``` (l.to_pandas()["b"] / r["b"]) 0 NaN 1 0.099257 2 0.126360 3 0.146000 4 0.162464 ... 196 0.789660 197 0.790833 198 0.791998 199 0.793156...

I haven't worked it through to satisfy myself that it's the problem, but the fact that the impl doesn't look at the `adjust` flag looks suspicious

as I discovered today, a moderately satisfying solution to this is to fetch a new datasource, using the resource as an id. As an example: ```scala val config : Config...

Sadly, if you generate a new data source for each fetch then multiple queries from the "same" source won't be aggregated. That's why I ended up fetching the source instead

Actually, I say that, but I didn't investigate whether a proper equality definition would fix it.

can confirm that if you have a proper equality definition for the datasource then you can do ```scala def getToken(): Fetch[Token] = ??? implicit def fixingSource(implicit t:Token): DataSource[Id, Data] =...