Just watched the talk by Alexander and this looks promising. Is there already a way to handle errors and reconnections? I skimmed though the doc and it doesn't seem to have anything about it.
a
Alexander af Trolle
05/24/2024, 11:28 AM
Hi, I can't find the link to the talk? do you have it in the history
s
Spoudel347
05/24/2024, 1:01 PM
@Alexander af Trolle here
https://youtu.be/G8pS1FzZlhk?t=2227▾
a
Alexander af Trolle
05/24/2024, 3:49 PM
Thank you
a
Alexander Sysoev
05/27/2024, 9:02 AM
Hi!
The current way errors are handled with kRPC protocol - user errors are propagated with the reconstructed stacktrace, RPC errors are thrown as is
Right now no customisations of this process are available
Reconnections are not supported, but are in the plans. Right now any unhandled failures cause service to cancel
👍 1
a
Alexander af Trolle
05/27/2024, 9:10 AM
But, It's fine if you create your own RPC connection & Service holder to handle reconnections yourself?
a
Alexander Sysoev
05/27/2024, 9:12 AM
Of course, all of the above is applicable to kRPC protocol and thus Ktor integration at the current stage
If you create our own RPCClient/RPCServer implementation - you can handle request in whatever way you may want