Closed rmg closed 6 years ago
Now that I've fixed the coverage reporting I see that my approach causes a bunch of bucket checks in other places throughout the repo to no longer be tested.
@timshannon would you rather I change my approach to do the bucket check closer to the Find()
entrypoint so that it impacts less code paths?
On the one hand that seems the better approach since it would dramatically reduce the likelihood of bugs.. but on the other hand, the reduction in coverage is a pretty good indicator that the current approach actually does remove a bunch of code paths and that is normally a good thing.
I think the check makes more sense in the runQuery func. It's used in updates, and deleted as well. I'm still reviewing this, but first glance this looks great. I'll look and see if we want to actually be removing the unused code paths further down the line or not.
Thanks
Fixes #30 by immediately returning from queries against missing/empty buckets.