otbe / ktor-server-lambda Goto Github PK

Investigate how to remove EngineAPI annotations from userland (example) code without getting a warning from intellij

Interesting for authorization context provided by APIG if your using Cognito or Custom Authorizers

...

If yes we must encode the buffer as a base64 encoded string and set the proper flag.

There should be some description about what this library is and how to use it.

Check current impl for correctness

Add tests for APIG integration

https://phauer.com/2018/best-practices-unit-testing-kotlin/

• LambdaApplicationRequest (@alxgrk using kotlin-test, MockK & AssertK)
• LambdaApplicationResponse (@otbe )
• LambdaApplicationCall (attributes) (@otbe using same 😊)
• LambdaEngine (request mapping HTTP -> APIG -> HTTP)

Unfortunately, the formatting rules seem to differ between developers.
Find a way to unify formatting and maybe also introduce ktlint to this project.

Maybe something like com.mercateo.ktor.server.lambda?

Check current impl for correctness (especially the headers part)

We should provide a basic example. Maybe a small serverless + REST example?

Should it be possible to deploy the ktor function to any other cloud provider than AWS? Could be helpful to at least structure the project by creating a separate ktor-server-lambda-aws module and map from AWS specific request model to a more generic one...

Clenaup and verify correctness

• header generation
• output (ByteChannel)

Create a library ready build setup + Travis (or circle?) setup

...

we encountered that the application is started twice, when running locally.

--- [main] INFO  Application - No ktor.deployment.watch patterns specified, automatic reload is not active
--- [main] INFO  Application - Responding at http://0.0.0.0:8080
--- [DefaultDispatcher-worker-1] INFO  Application - No ktor.deployment.watch patterns specified, automatic reload is not active
--- [DefaultDispatcher-worker-1] INFO  Application - Responding at http://0.0.0.0:8080

this seems to happen because the we probably don't use the lambda engine correctly: we specify the main function like fun main(args: Array<String>): Unit = io.ktor.server.cio.EngineMain.main(args) but shortly afterwards we also start the LambdaEngine using the LambdaAdapter.
This also shows, that it's probably a good idea to follow the pattern used in EngineMain and let LambdaEngine be a singleton, that is started by the library user.