Cannot access result from mutation hoc

[doniyor2109]

I have used mutation-hoc for RN Component. I can access mutation function however I cannot access the _result from mutation.

react-apollo/src/mutation-hoc.tsx

Line 61 in 8d06f4f

{(mutate, _result) => {

[doniyor2109]

Here is PR #1844

[fenok]

I need this badly. However, I don't think that #1844 does this right. What if there are multiple mutations in HOCs chain? I believe the right way is:
let childProps = { [name]: Object.assign(mutate, _result) };
It's 100% backward-compatible and allows accessing each mutation's state individually.

[doniyor2109]

Maybe temporary workaround is to use old graphql from react-apollo

[fenok]

Could you explain that a bit further? What do you mean by “old graphql”? Currently graphql uses mutation-hoc internally.

[doniyor2109]

Sorry for mixing, I mean currently graphql

[fenok]

I still don't get it, because

Currently graphql uses mutation-hoc internally.

Can you provide an example?

[doniyor2109]

It is possible to get loading and error states with the help of some additional hocs. There is some solution in #421

[doniyor2109]

I opened new PR which only adds new prop called mutationResult #1881

[fenok]

It is possible to get loading and error states with the help of some additional hocs. There is some solution in #421

I'm aware of it, but I want this functionality included in react-apollo itself.

I opened new PR which only adds new prop called mutationResult #1881

Again, this will fail if component is wrapped by several mutation-hocs. You could at least use [`${name}Result`]: _result. But I'm still convinced that it should be done like let childProps = { [name]: Object.assign(mutate, _result) };.

[doniyor2109]

Yes you are right. It overrides when there are multiple hocs. You should also create PR for your solution. That could be the right solution.

[brendanmoore]

I am a little bit baffled as to why _result is not passed to the child with the mutation-hoc I cannot find any documentation explaining the decision. If somebody "in the know" could share, that would be great.

[ekfn]

Is there any good workarounds? Or may be news on this one?

[rosskevin]

Explanation:
#1844 (comment)

jbaxleyiii commented on Mar 26
@doniyor2109 this is actually intentional to keep backwards compat 100%. I'm curious why you wouldn't use the Mutation component instead?

FYI - In-house, we don't use the Mutation component or hoc, we just mutate directly on the client.

Given this explanation from a core maintainer, I'm going to close this in an effort to keep the tracker focused on actionable issues. I suspect a hooks implementation in 2019 will make us think about our patterns of use and perhaps transition us to something simpler yet to use.