I see in the Okio docs that file system is available for all platforms, but I couldn’t see how to load a file from a directory in commonMain (I’m targeting JS, JVM, Native). Have I missed something in the docs? Is it possible to use

FileSystem.SYSTEM.read(path)

in commonMain?

It is not. The system filesystem is unavailable in JS and filesystem is generally not useful in JS because it's a fully synchronous interface.

You can really only use with in-memory things for JS, or for NodeJS there is a blocking implementation but you wouldn't want to use that in any kind of high-throughput server where one would normally choose NodeJS.

Hello, I want to use cash multi paging library (https://github.com/cashapp/multiplatform-paging) in my project which has support for following architectures:

iosX64() iosArm64() iosSimulatorArm64()

but this library gives error which used in my KMM project. Error stack 
    
     ```

Task widgetcompileKotlinIosSimulatorArm64 FAILED

e: Could not find "org.jetbrains.kotlin.native.platform.CoreFoundationBase" in [/Users/Documents/felix, /Users/.konan/klib, /Users/.konan/kotlin-native-prebuilt-macos-aarch64-1.7.21/klib/common, /Users/.konan/kotlin-native-prebuilt-macos-aarch64-1.7.21/klib/platform/ios_simulator_arm64]

FAILURE: Build failed with an exception. Any help is much appreciated!

I am trying to write a SqlDelight migration that adds a column to a table. I want to set a default value during migration but not on the fresh tables. I’ve tried this so far:

-- someTable.sq 0.sq
CREATE TABLE someTable(
	id TEXT NOT NULL PRIMARY KEY
)

-- someTable.sq 1.sq
CREATE TABLE someTable(
	id TEXT NOT NULL PRIMARY KEY,
	value INT NOT NULL
)

-- 1.sqm
ALTER TABLE someTable ADD COLUMN value INT NOT NULL DEFAULT 42;

This doesn’t pass SqlDelight tests because

someTable.sq

schema does not define default value, which I don’t want…Is it possible to make

ADD COLUMN

work here without forcing original schema to define default value, or is my best bet here to do manual table re-creation (create new table with new schema, move values there, delete old table, rename new table)?

This message was deleted.

Hello,I found such usage of

json_each()

is not valid in

.sq

file, will it be supported in the future?

How can I save my database to a user file when using SQLDelight? I plan to have the working database saved inside AppData folder (for crash recovery etc.) and when the user requests it, copy the database to the save path to implement save operation.

Using Turbine, I’ve got a fake impl of something like this (simplified)

class FakeFoo : Foo {
  val responseTurbine = Turbine<Bar>()
  override suspend fun invoke(): Bar {
    return responseTurbine.awaitItem()
  }
}

Then in a test, I am testing the result of a StateFlow, which comes from a result of combining two things, something like this

val data: StateFlow<State> = combine(someOtherFlow, flow { emit(FakeFoo().invoke()) }.onStart{ emit(null) }) { someOtherFlowValue, bar -> State(someOtherFlowValue, bar) }

But in one particular test I am not interested in the result of FakeFoo, so I want to rely on just the initial value emitted from the

onStart

(Is a StateFlow with a default value in my case) to only test the other half of the

combine

, whatever comes from

someOtherFlow

. However, since this

Turbine<Bar>

doesn’t ever get a value, but I am awaiting on it, it crashes my test with “no value received in X seconds”. Doing

cancelAndIgnoreRemainingEvents()

isn’t an option either, since now instead of infinitely waiting on

awaitItem()

it will instead crash with “Expected item but found Error”. Is there some other approach I can take instead? Maybe I shouldn’t be using a Turbine for my fake impl? I’ve taken inspiration to go with this approach from and it has worked flawlessly in all the other cases I’ve tried.

Is it possible to write suspend functions for an okhttp execute/enqueue call but have it use okhttps thread pool?

i don't follow. if you use enqueue it does use the okhttp dispatcher thread pool. and then when the response is done it resumes your coroutine on whatever dispatcher you called from

Hm. This probably just comes from me not being all that well versed in coroutines. But it sounds like if I wanted to create a suspend fun to encapsulate a network call, I would just use the enqueue method. Let me give that a shot

so looks like i have to do something like this? https://stackoverflow.com/questions/48552925/existing-3-function-callback-to-kotlin-coroutines/48562175#48562175

adapt to taste

two keys: ensure the response gets closed, and read the entire body before resuming the continuation because otherwise it's streaming and you'll wind up doing I/O on the wrong thread

Does okhttp provide any way to take an existing HttpUrl and change the path specificially (in cases where you want everything else to stay the same?)

if you don't care about query parameters or fragments you can call

resolve()

with a relative or absolute path. if you do care about query parameters and/or fragments you would need to call

newBuilder()

and then manipulate the path using its API

HI, I am trying to make a database backed sample on android with Workflow kotlin In the tutorial it says to install the file templates using this script I ran the scripts on Windows OS and it ran fine. (the script did not throw any error) Even after restarting Android Studio, I cannot see file templates installed by script in Android Studio I am using Android Studio Canary Giraffe Beta

I run into a strange behavior difference when using molecule+turbine, so I opened https://github.com/cashapp/molecule/issues/249 though I can't understand if it's a bug or working as intended nor who's the library responsible for this behavior (molecule, turbine or compose?). Any input is appreciated so I can move the issue to the appropriate repo, do more debugging or close it. The gist of it is that mutating a

MutableState

twice usually causes 2 recompositions (one per mutation), but doing the same inside a

launch{}

lambda will only cause 1 recomposition. Repro test case is included in the issue.

In what way is Postgres support in SqlDeligt "experimental"? Meaning that it might actually never get full support / be dropped from main SqlDelight? Or Postgres integration has very rough edges to be dealt with but eventually SqlDelight will for sure support it?

Postgres support is graduating to stable with SQLDelight 2.0

I've got a question about kotlinpoet: Why is

AnnotationSpec.get(annotation: Annotation, includeDefaultValues: Boolean)

annotated with

@DelicateKotlinPoetApi

? The message says "Java reflection APIs don't give complete information on Kotlin types. Consider using the kotlinpoet-metadata APIs instead.". However types in annotation are already pretty restricted in Kotlin due to JVM restrictions (no `null`s, only primitives, strings, enums, arrays, classes and other annotations). So I wasn't able to think of a situation that Java reflection wouldn't be able to extract all information needed for that function. Is there a situation that wouldn't work with Java reflection? And if not, is there another reason for the annotation?

I'm trying to update my KMM project to use Sqldelight 2.x and I'm getting this error. This appears to be the same code as what is in the docs and it worked with 1.x. Can somebody explain what's going on with "asFlow"? I have imported

import app.cash.sqldelight.coroutines.asFlow
import app.cash.sqldelight.coroutines.mapToList

idk why its not importing but that has a parameter now

you have to pass it a dispatcher

or coroutinecontext

whatever it is

Can I use zipline for using Kotlin/Js library that uses an npm library inside Kotlin/JVM ?

Unfortunately no. We decided not to support that case for now. If you need this, comment on https://github.com/cashapp/zipline/issues/725 in favor