I write here because came another intervention. Sorry but I do not get what you guys mean. I tried to run on the internet a research about 'Ktor vs Springs', or 'limit of Ktor' and what i get is that is a microframework ktor, that means that lacks some advanced features. But if becomes popular will come more and more libraries right? And for what concern legacy code, can be also integrated in existing Spring project with a different stack that communicates with Spring right?

Since KotlinxSerializer setMapper is deprecated, what should be used instead?

When installing

Logging

(https://ktor.io/clients/http-client/features/logging.html) alongside other features that can change headers in a client, does the order matter ? Can I make sure to always get what's transmitted over the wire ?

But once you import //https://mvnrepository.com/artifact/org.webjars.npm/material-components-web It’s not exactly clear how to setup the static resource to use those css and js files.

Does Ktor CIO implement request pipelining? IE, reusing the same connection for multiple requests? If so, I think I may have another vulnerability in Ktor.

ktor version

1.3.0-rc2

in post call when call function

call.receiveParamters()

i get this error

java.lang.NoSuchMethodError: 'java.lang.Object io.ktor.request.ApplicationReceiveFunctionsKt.receive(io.ktor.application.ApplicationCall, kotlin.reflect.KType, kotlin.coroutines.Continuation)'

Any idea how fix it? it work fine if using ktor version

1.2.2

@hhariri

Hey all, sorry for spamming.. may be vague question for the larger audience. anyone give me why ktor is best bet when writing REST api in kotlin?

readUntilDelimiter should do the trick

I can't import class io.ktor.auth.AuthenticationProvider and other classes for ktor, even if I can run a ktor application. Why is it so? Do I need to reference some other dependency? The whole io.ktor.auth package is not available.

Hello, how would I go about measuring response times? Let's say I want to

println

something like "responded to /some/endpoint in 50ms" for every request. I have found I can access every request with

intercept(ApplicationCallPipeline.Monitoring){ /**/ }

but I'd like to access every response before it's sent and some information about it.

Hello, World! I have a client library using Ktor HttpClient, and a sample on the JVM works perfectly, while the exact same code on Native (MacOS) fails with:

Uncaught Kotlin exception: io.ktor.client.call.NoTransformationFoundException: No transformation found: class <http://io.ktor.utils.io|io.ktor.utils.io>.ByteChannelNative -> class org.jraf.klibqonto.internal.api.model.organizations.ApiOrganizationEnvelope

So I'm supposing the problem would have something to do with the Engine that's used (I use curl on MacOS). Any idea what the problem could be?

I am looking into Ktor interceptors (from a Ktor server perspective) to modify the original response before it is send to the client. A quick example of what I want to achieve: suppose I have a route which returns some response, for example:

route {
    get("sample") {
        call.respondText("Hello {Name}")
    }
}

And I want to use a global interceptor to replace the {Name} variable in the response with some other value. How do I access the contents of the response in an interceptor? (I know how I can send a custom response but not yet how I can access the original response, and modify it)

Hey! I’m trying to deploy my

Ktor

backend module from a

KMP

project to Heroku, but I’m getting

A problem occurred configuring project ':app'.  > kotlin.KotlinNullPointerException (no error message)

for

Task :buildSrc:build

. Anyone who had similar issues?

Hi guys! I am trying to implement simple crud in ktor and can not find any validations, something like javax.validation. Is it something existing i ktor? Or should I write my own?

We have a ktor back-end running in AWS ECS running on top of about 8 EC2 instances. It currently handles somewhere in the neighborhood of 1.5 million requests per day and has had some serious performance issues. This ktor server replaced a single PHP server that was handling the same load and while that server was larger, it didn’t have any of the autoscaling that our ktor servers have built in and handled the load just fine. The ktor server objectively has more computing resources to work with. This app doesn’t do any heavy-duty processing. It receives requests from clients, does some database queries, and sends http requests. If you were trying to optimize this code, where might you start looking? CC @Derek Seroky

Hi! Any good tutorial on how to make our own interceptors? I feel dumb trying to read other's and not understanding why it works.

Website says:

There is a performance penalty when using auto-reloading. So keep in mind that you should not use it in production or when doing benchmarks.

What exactly would cause a penalty here? Simply a non empty watchPaths, or watchPaths pointing to empty directories, or using reload instead of redeploy ?

override fun modify(item: User): CompletionStage<User> {
    return supplyAsync {
        transaction {
            Users.update({ Users.id eq item.id }) {
                it[name] = item.name
                it[gender] = item.gender
                it[first_name] = item.firstName;
            }
        }
    }.thenApply { return@thenApply item }
}

Is there a way to simplify an update transaction on object properties, i feel it is repetitive to keep on writing : it[name] = item.name it[gender] = item.gender ?

Hey All, Anyone used

aws paramstore

storing the config value and resolving at runtime at the application start?

When you have a Ktor server watching class files for hot-reloads, what’s the best way to re-compile class files in the IDE. If you use a gradle IDE configuration, that seems to be equivalent to using gradle from the terminal, which takes additional time because gradle has to perform startup. How do I leverage the already started daemon that the IDE owns? All means to do that seem to involve providing a main class.

oi... are you all aware that you can only use the ktor client from the main thread for native targets?

I'm not sure but... I think ktor client (Apache) is leaking TextContent objects for some reason? My app crashed and when looking at the memory dump there are 654 TextContent objects, in total they are using almost 4GBs (!) The engine is created with timeouts, so it isn't leaking because the connection never succeeds.

this.socketTimeout = 25_000
			this.connectTimeout = 25_000
			this.connectionRequestTimeout = 25_000

My unit test code would be a whole lot simpler if the

MockClientEngine

request handling could be updated after the object is created

Hello. Is there a way to make the HttpClient JsonFeature reuse the same serializer/mapper we have configured for the server?

let's say I am using HttpClient w/ serialization, and I want to handle the case of invalid JSON by getting the response text. is that possible, or do I need to receive the response as text and deserialize manually?

i seem to be experiencing a deadlock in

<http://HttpClient.post|HttpClient.post>

on iOS. the method never returns. it's called from within a

runBlocking

block from the main thread (from a unit test). has anyone experienced this?

📣 ktor 1.3.0 has been released! • ktor client proxy support • Introduced

HttpStatement

and deprecated potentially dangerous resource-leaking client API • Eliminated kotlinx.io dependency • Fixed server identity compression handling: keep original content length • Fixed handling GET requests with body (#1302) • Fixed curl request with empty body • Added iOS url session configuration • Fixed CIO engine no longer sends port in “Host” header (#1295) • Add INTRINSIC value to TLS signature algorithms • Introduced ability to send string in request body • Improved client and server typeOf support with kotlinx.serialization • Gradle 5.4.1+ with newer metadata (metadata 1.0) • Improved exceptions handling in client and server on Android • Added missing TLS parameters and relaxed TLS parsing to ignore unsupported features • Improved session diagnostics (#1368) • Fixed

hookRequests

in test engine (#1300) • Deprecate java.time related API and related cleanup (for future kotlin.time support) • Restricted CIO HTTP headers parser • Introduced header name and value validation • Fixed must-revalidate on the request side in ktor client (#1406) • Fixed OkHttp client resource cleanup on close • Added watchos/tvos native targets • Fixed content truncation at native and JS targets • Fixed server’s

If-Range

header parsing to avoid crash at date parsing (#1377) • Fixed server’s conditional headers processing • Reduced required JDK version for

DefaultHeaders

server feature • Fixed client hanging due to exception in response pipeline • Replaced HttpClientJvmEngine to HttpClientEngineBase that is now common for all platforms (affects only custom client engines) • Fixed hierarchy of execution and call contexts in clients that allows to properly handle request lifetime using execution context. • Optimize JS module import time (#1464) • Upgraded versions of Netty, Jetty and Tomcat implementations • Added Pebble template engine (#1374) • Introduced localPort route that is always tied to actual socket port (#1392) • Fixed cookie expiration date parsing (#1390) • Server authentication feature’s phases are now public (#1160) • Fixed auth header resending after redirect (#1467) • TCP half-close made optional for CIO client engine and disabled by default. • Apache client random timeouts fixed • Fixed locale-dependant code (#1491) • Fixed unclosed websocket channels if cancelled too early • TCP half-close made optional for CIO client engine and disabled by default. (#1456) • Improved ktor-client-mock engine to be thread safe (#1505) • Fixed client cookies logging (#1506) • Fixed multiple application stop events in test engine (#1498) • Fixed CIO ActorSelectorManager to not spin due to cancelled keys (affects both CIO client and server) • Made default auth validate functions fail to force users to implement them • Introduced test client instance in the test server • Fixed various server and client engines to return

null

for missing headers rather than empty list • Introduced support for json structures in client and server (#1519) • Introduced ktor-server-core binary compatibility tracking • kotlinx.coroutines 1.3.3 See migration guide: https://ktor.io/quickstart/migration/1.3.0.html Known issues: • Netty HTTP/2 push sometimes get broken • Running ios and native client engine on non-main worker is still not supported (there are only few working scenarios so not guaranteed to work)

I’m currently in the process of upgrading to ktor 1.3 and I have this issue with where previously

readRemaining()

did not need arguments, now it needs a limit and headerSizeHint. What would be appropriate values here?

This message was deleted.

An error occurred communicating with http://localhost:8080/validate: Failed to parse request body: request body length should be specified, chunked transfer encoding should be used or keep-alive should be disabled (connection: close) CIO client engine running on an Android emulator. Server side is Ktor Netty engine.