TOut of curiosity: are you using react18?
TInteresting. Is there maybe something nextjs is already doing with suspense on the server 
...
...T
TIf you can, I would appreciate to know where it comes from. We can rule out trpc if you can reproduce it with next + react-query alone
TI think this is worth an issue in react-query then. can you file it? it seems that throwing a promise in react18 and next 12.2 triggers this behaviour?
Tokay this is extra weird because the component is rendered on the client only - there is no gSSP or something... also, I forked it and played around a bit and ... now it works?
https://codesandbox.io/s/suspicious-kapitsa-pmg9jp?file=/hooks/usePosts.js
https://codesandbox.io/s/suspicious-kapitsa-pmg9jp?file=/hooks/usePosts.js
Tokay nevermind, I think the sandbox didn't save 
. a new for still shows the error
however, if I go back to react17, I get the error:
. a new for still shows the errorhowever, if I go back to react17, I get the error:
Thas that always been that way with next?
Tokay let me downgrade to v3, too
Tah okay, yeah that makes sense that it works there
Tcan you try:
enabled: typeof window !== "undefined"Tthat should turn off fetching on the server; I am getting a hydration mismatch though!
TSebastian Markbåge (@sebmarkbage)
Twitter
•
@TkDodo @nextjs @reactjs The recommendation is that libraries throw an error on the server if they implement experimental suspense: true when there's no way to guarantee that the same request on the client will line up during hydration. It's up to the library to ensure that it implements this properly.
9/23/22, 9:16 PM
TI think suspense for data fetching on the server will work better with the new nextjs router (see the layouts rfc). For now I wouldn't use it
Athat' doesn't feel too good -- i really want us to be able to use proper RSC on the server soon