Fork me on GitHub
#graphql
<
2017-08-04
>
hmaurer15:08:17

@hlship do you think something like this would be doable with Lacinia? https://github.com/graphql/graphql-js/issues/407

hlship15:08:34

It's already there.

hmaurer15:08:35

In summary, modifying the context (passed to resolvers) during resolution (in a resolver)

hmaurer15:08:37

for all the subtree

hmaurer15:08:41

oh really?!

hmaurer15:08:02

I must have missed it in the doc; could you tell me the API please?

hmaurer15:08:57

Ah, silly me. Thanks, that’s great

theikkila16:08:56

@malcolmsparks @hlship thanks for input, I wrote my comment back there. I suggest we’re supporting more clients even with invalid headers

theikkila16:08:08

it’s easier to fix it there than from all of the clients

malcolmsparks16:08:34

Yes, Postel's Law applies here. Though iirc the Content-Type is strictly invalid if there are invalid parameters. And standards are important. But in this case it seems better to ignore the parameter. However, this is a bug in the client that should be communicated. The internet is only the internet when standards are upheld. :)

hlship16:08:44

I'm thinking that we could introduce a Pedestal interceptor that touches up the content-type for this case, to make it obvious that this is temporary.