two listeners when starting sample locally about ktor-server-lambda HOT 6 CLOSED

actually, we do not need to start the application locally, right? with this in mind, i removed the whole handler stuff from sample, so what's left to do for a user is to specify the ktor application and reference the LambdaAdapter::handle method in the serverless.yml.
should be very easy to use.

from ktor-server-lambda.

Looks great so far. I wonder if we can keep (parts of) the Handler.kt file just for local development purposes? I would really love to be able to start the API on my machine 🤗

from ktor-server-lambda.

Yeah, I would like that too :) what do you think about some configuration option when starting the app? E.g. declaring a constant in LambdaAdapter, that can be passed to the LambdaEngine from the main function, which in turn enables delegation to the CIO Engine? Something like this 😅

from ktor-server-lambda.

Im not quite sure about this 🤔I would like to keep it as simple as possible for us and the users. Correct me if Im wrong but if we keep Handler.kt until this line and rename it to something like Launcher.kt it should work without modifications of the LambdaAdapter/Engine, shouldnt it? Especially after your last changes to the LambdaAdapter.

from ktor-server-lambda.

the problem with the current LambdaEngine is, that the main thread won't be kept alive. If you run the main function, the program will be stopped after a short period of time.
If you have a deeper look into CIOEngine, you'll see that there are coroutines being started that stop the application from being shut down.

I think the question is: what do you want to test locally?
if it's only the ktor application itself, I would say it's enough to re-use the CIOEngine.
if you want to simulate the lambda environment, then we'll need to block in the LambdaEngine ourselves.

from ktor-server-lambda.

after talking to @otbe my confusion is now resolved and i simply re-added the original main function

from ktor-server-lambda.