Brendan O'Brien

need to look into this. Ideally we should bump up lots of our tests by passing `None` into places where `None` doesn't belong

From @dustmop : Other tests I think would be good to have: insert into the middle of an array: ``` [[1],[2],[4],[5]] [[1],[2],[3],[4],[5]] ``` row of a table completely changing: ```...

https://medium.com/qri-io/four-simple-steps-to-improve-u-s-census-data-csvs-b7358ee1b18?source=collection_detail----c8725b5a8089-----0-----------------------

We're _always_ performing a sync action when we start, but there's a chance the data we're trying to pull exists locally. We should do a quick check to the local...

We've recently made the decision to include the notion of a "remove request" within dsync as an optional protocol command. We should do the work of building out examples, tests,...

We need an interface abstraction for storing sessions. This would allow us to use different persistence & fetching mechanism in different contexts

We should arrange pulls into sessions the same way we do with push. That way we could potentially resume pulls, track progress, etc.

```go // Pull now accepts a meta param NewPull(cidStr string, lng ipld.NodeGetter, bapi coreiface.BlockAPI, rem DagSyncable, meta map[string]string) (pull *Pull, err error) // Push doesn't... func NewPush(lng ipld.NodeGetter, info *dag.Info,...

Test are are missing from recent additions: * [ ] test getDagInfo hook on http via a 3rd requesting node * [ ] test remove unpinning * [ ] test...

https://github.com/qri-io/desktop/issues/478 uncovered this with the following error message when running `qri save`: > validating data: error reading values: error writing row 129283: json: unsupported value: +Inf What's going on here:...