Jaap Frolich

> Didn't it used to do a full decoder for all the data coming in? We don't do full checking of the JSON because the number of checks we need...

Um I think I didn't realize the impact. If we are parsing partial data (whether it's due to an errorPolicy or with the partialResults option), we should have the types...

Perhaps we can have a `parsePartial` function that is an alias of `parse` when `@ppxPartial` is used as a directive to instruct the `ppx` that all data can be partial.

so if you don't have the annotation, it will fail to compile because `parsePartial` is not available.

BTW the partialData story is interesting. I saw that Apollo is going to have a more precise partial rendering technique in 4.0 that is more similar to Relay with `useFragment`...

> I don't have a ton of clarity at this point how we make things typesafe for `errorPolicy` when set globally, or whether we can simply get around the issue...

No we have to wait for the `ppx` addition. But just saying that with the current approach it crashes mostly outside of the parse function, because most of the time...

Yeah I think shipping a runtime module is the best option to keep compatibility with < `4.08`

We can backfit this on the native ppx now. But the plan is to make the Bucklescript ppx universal so it should definitely be part of that PR.

Hey, good point. It should make the first letter lowercase in this situation. If that is not yet the case, PR is welcome. If a PR is not possible an...