Maybe a dumb question. But im working on a library (regretfully 😂) and we need to make network calls. would love to use okhttp + kotlinx.serialization, but we don't want to go all in and pull in retrofit. Is there an "easy" way to deserialization with okhttp without brining in retrofit?

Another fun one... This is weird right? We have a slim network layer (on top of okhttp) to just make a few things a bit easier. and we have a

fun sendSync(req: Req): Res
return runBlocking {
  sendAsync()
}

suspend fun sendAsync() {...}

With okhttp 5.1 my minified android build fails with

> Task :foo:app:minifyTstStableReleaseWithR8 FAILED
ERROR: Missing classes detected while running R8. Please add the missing classes or apply additional keep rules that are generated in /Users/bar/AndroidStudioProjects/foo/app/build/outputs/mapping/tstStableRelease/missing_rules.txt.
ERROR: R8: Missing class okhttp3.internal.Util (referenced from: void okhttp3.internal.sse.RealEventSource.processResponse(okhttp3.Response) and 1 other context)

Anyone else experiencing this?

Does anyone know if the workflow kt test library supports pure kotlin or multiplatform? I have an app built in multiplatform using workflow kt but can’t test it since last I checked the test library was only supported for the jvm target

Trying to understand which parts of okhttp are "value classes" Struggling with whether or not https://square.github.io/okhttp/3.x/okhttp/okhttp3/Request.html is a value class because of it's description

An HTTP request. Instances of this class are immutable if their body is null or itself immutable.

Like if this was written in kotlin would it just be a data class?

If

turbine

gurus could chime in

private val _state = MutableStateFlow<ChatState>(ChatState.Idle)
override val state: Flow<ChatState> get() = _state

scope.launch {
    combine(
        conversationDao.conversation.doOnNext { println("-- CONVERSATION=$it") },
        connector.connectionState.doOnNext { println("-- CONN_STATE=$it") },
        ::stateMapper
    )
        .collect {
            println("--- PRODUCED=$it")
            _state.value = it
        }
}

I have

combine

which produces as a result and pipes it into stateflow

_state

. That

state

flow is then asserted at test time with turbine (ommited) Issue is that for some reason this is not deterministic. Default value of

connectionState

is

Disconnected

and then obviously

Connecting etc

What I'm seeing is that sometimes I see

Connecting

getting emitted before the product of combining with

Disconnected

gets set to the

_state

Othertimes it's not and I see 2 `--- PRODUCED`as expected. Is this a

StateFlow

issue? A

combine

issue? A Turbine issue? A

TestDispatcher

issue? Should I maybe not assert all the intermediary emits and just the last one using?

Question for SqlDelight maintainers. Would you be willing to take a PR to either: 1. Expose

sqliter.Cursor

from SqliterSqlCursor,

ResultSet

from JdbcCursor, etc... or 2. Expose some kind of

CursorMetadata

API? I'm trying to see if I can integrate my libraries Terpal/ExoQuery with SQLDelight and the missing piece seems to be DB result-set metadata.

Hi there, I am using SQLDelight and I am trying to setup a dependency between two database (defined with the sqldelight gradle plugin). I followed the documentation and it shows clearly how to do this on a gradle subproject. There are two ways to setup these kind of dependencies

fun dependency(dependencyProject: Project)
fun dependency(delegatedProject: DelegatingProjectDependency)

The issue that I have here is that, the dependencies are defined inside a composite build (not just a subproject). And I am having a hard time getting an instance of that subproject. If anyone can help me how to achieve this, I will greatly appreciate

Hi everyone, how do we create a persistent DB using SQLDelight on JS? Web Workers driver by default only creates an in-memory DB.

Hello once again, I am building a hybrid app (with both online and offline support) and I am trying an unorthodox approach of sharing some tables on the client side and the server side The issue is that I need to write an SQL statement that is portable between SQLite (on the client) and MySQL (on the backend) Take the following statement as an example

CREATE TABLE IF NOT EXISTS ow_core_people(
    uid INTEGER PRIMARY KEY NOT NULL AUTO_INCREMENT,
    name TEXT NOT NULL,
    photo TEXT
);

Because AUTO_INCREMENT doesn't work on SQLite, it just works on MySQL but fails on SQLite. Is there a way for SQLDelight to support these multiple syntax and transform them in accordance to the dialect selected?

did the compose dialogs library from keith ever ship or is that still coming up?

When writing

sqldelight

kotlin migrations, I need to have atleast a placeholder

.sqm

file. This is error prone, as nothing seems to enforce the

.sqm

file. Has anyone figured out a way to blow up such state? (

.kt

migration present but

.sqm

not)

This message was deleted.

If anyone has a project using SqlDelight that could test it with sqldelight-androidx-driver it would be very much appreciated 😄 https://kotlinlang.slack.com/archives/C0BJ0GTE2/p1757703433402129

Can we re-open this by chance? Hitting this again with detekt 😭 https://github.com/sqldelight/sqldelight/issues/4157

HI, I'm an jvm

anvil

user. Now I want to experiment with KMP. I want to keep

@Inject

and

@ContributesBinding..

in common - and expect-actual them to typealiases of the jvm libs (and therefore have anvil/dagger take effect only on jvm)

@Inject

is easy enough

@Target(
    FUNCTION,
    PROPERTY_GETTER,
    PROPERTY_SETTER,
    CONSTRUCTOR,
    FIELD
)
@Retention(RUNTIME)
@MustBeDocumented
expect annotation class Inject()

actual typealias Inject = javax.inject.Inject

but I'm having trouble with

@ContributesBinding

because of all the params. See the full issue in thread

lazyweb: using mockwebserver in an existing project (it already had the dep added). writing a new test with it... now going to import and i see two options

com.squareup.okhttp3:mockwebserver3:5.1.0

or

com.squareup.okhttp3:mockwebserver:5.1.0

Are my depenedencies setup incorrectly or should i indeed be seeing both options (if so... which to pick)?

I want to encrypt my access tokens at rest, but the decryption being somehow expensive I'm thinking of caching the plain text of access token in memory. Then I remembered to use

ByteArray

instead of a `String`so I can deterministically clear the value from memory, because GC blah blah. But, when using

okhttp

interceptor to attach a token as

val tokenString = String(tokenBytes, Charsets.UTF_8) <-----
val newRequest = request.newBuilder()
    .header("Authorization", "Bearer $tokenString")
    .build()

then it's a moot point right? Or -- the question -- is there a smarter what of doing this in

okhttp

? Or do I just need to accept the transient token in memory (and therefore holding it as

ByteArray

is mostly pointless)?

I've got what I think is a pretty simple setup. I have an androidlibrary module... and a small app module to exercise said library. okhttp is only added to the androidlibrary module.

./gradlew androidlibrary:assemble

is successful, but

./gradlew app:assemble

fails due to

* What went wrong:
Execution failed for task ':app:checkDebugDuplicateClasses'.
> A failure occurred while executing com.android.build.gradle.internal.tasks.CheckDuplicatesRunnable
   > Duplicate class okhttp3.Address found in modules okhttp-jvm-5.1.0.jar -> okhttp-jvm-5.1.0 (com.squareup.okhttp3:okhttp-jvm:5.1.0) and okhttp-release.aar -> okhttp-release-runtime (com.squareup.okhttp3:okhttp-android:5.1.0)

seems like something regarding okhttp-jvm and okhttp-android? according to docs, this shouldn't be an issue since im using gradle and not maven?

⚠️ Heads up: OkHttp 5.2.0 is crashing on API levels 21 through 23. We’re cutting OkHttp 5.2.1 to fix. Details here.

Question about okhttps bom. It seems like all of okhttps artifacts are versioned together in lock step. why the bom? from what i can tell, bom seems to be more helpful in situations where artifact versions are all scattered?

I'm looking into making my SqlDriver use

QueryResult.AsyncValue

, not because the driver itself is async, but because I want to use coroutines in my connection pool. Is that "abusing" the system, or a valid use case?

As part of the above, I'm trying to bake in CoroutineDispatcher support directly to the SqlDriver so users don't have to manage that. I'm running into issues with

transaction

though, because it doesn't like the transaction escaping the thread it is started on, which is what happens with the dispatcher support (newTransaction, execute, and executeQuery use withContext to switch into the proper dispatcher). Is there any way to hook into the transacter to support this, and if not, is there any chance support for this could get added to SqlDelight?

Hi everyone, I'm running into a persistent

no such table: user

error with SQLDelight on a Kotlin Multiplatform project. I've tried the common solutions like manually calling

Schema.create(driver)

, deleting the old database file, and ensuring I'm using a singleton instance for the database, but the problem persists. It seems the schema creation is either not happening or is happening after my first query is executed. Full stacktrace and code in thread 👇🏽

Getting ready to distribute my first android-library, so I'm doing a final check. If I'm using okhttp in my android library (as an implementation dependency) do I need to include keep rules in consumer-rules.pro and proguard-rules.pro? Or since it's an android lib... will all of those proguard rules just work? I keep chasing my tail on this because I'm teetering on whether or not we will proguard our library so its a bit obfuscated (in the finance space so PM/TL really want obfuscation)

My application uses SQLDelight to save chat history and generated images. This functionality has suddenly stopped working (Everthing works before) . To ensure this wasn't a local issue, I completely deleted my local repository and cloned it again. The problem persists. The app fails with the following errors:

text
org.sqlite.SQLiteException: [SQLITE_ERROR] SQL error or missing database (no such table: chats)
org.sqlite.SQLiteException: [SQLITE_ERROR] SQL error or missing database (no such table: generated_images)

To reproduce the error: 1. Clone the repository: https://github.com/yassineAbou/LLMS.git 2. Run the composeApp desktop target 3. The errors appear immediately during the build

PSA for anyone using sqldelight-androidx-driver: I am strongly considering making the driver async (the underlying SQLite APIs are still blocking though). This will allow more efficient integration with the connection pool that I've added to it, as well as make it possible to handle all of the details about dispatching internally, so SQLDelight APIs can be used without worrying about what

CoroutineDispatcher

you are using. If you have any thoughts, questions, or concerns, please discuss here

I created new project (from android studio wizard). it defaults to using kotlin 2.0.20. i added okhttp latest as a dep and don't do anything with it, and then i get a somewhat obscure error message, but after bumping to kotlin 2.2.x. the error is resolved. still a noob on this topic. im assuming thats expected? that once any library I use moves to a newer version of kotlin that the consumer must also move to that version (bonus question... as a library author... is it in your best interest to just target 2.0.x instead?)

As mentioned in the Zipline WASM support discussion, replacing the QuickJS interpreter with Wasm3 or WasmEdge would be the only way to achieve a web-based loader, correct? The current loader module doesn't support the web; it's limited to Android, iOS, and desktop. While I have created a simple program to simulate the effect of dynamically loading a JS plugin, trying to somehow insert a JS file into the existing Zipline files would be quite messy... However, the work involved in replacing the interpreter, if possible, is massive. As for inserting a dynamic JS plugin into the current structure, perhaps a 'web' folder could be created in Zipline's root directory to store the JS, though this JS would likely need to consider encryption. https://github.com/cashapp/zipline/discussions/843 https://github.com/crowforkotlin/compose-multiplatform-web-dynamic-load-js-plugin

Hello I am getting an error with SQLDelight Browser WASM: WebWorkerException: {"message":"Uncaught SyntaxError: Unexpected token '<'","type":"error"}isTrusted,true overrideMethod@ The build.gradle.kts:

kotlin {
   @OptIn(ExperimentalWasmDsl::class)
   wasmJs("store") {
      browser {
         commonWebpackConfig {
            outputFileName = "store.js"
         }
      }
      binaries.executable()
   }
storeMain.dependencies {
   // DB Web worker
   implementation(libs.web.worker.driver)
   implementation(devNpm("copy-webpack-plugin", "9.1.0"))
   implementation(npm("@cashapp/sqldelight-sqljs-worker", "2.1.0"))
   implementation(npm("sql.js", "1.8.0"))
}
}
sqldelight {
   databases {
      create("Database") {
         packageName.set("ar.com.cancerbero")
         deriveSchemaFromMigrations.set(true)
         verifyMigrations.set(true)
         generateAsync.set(true)
      }
   }
}

the webpack.config.d/sqljs.js seems to be working...

config.resolve = {
   fallback: {
      fs: false,
      path: false,
      crypto: false,
   }
};

const CopyWebpackPlugin = require('copy-webpack-plugin');
config.plugins.push(
   new CopyWebpackPlugin({
      patterns: [
         '../../node_modules/sql.js/dist/sql-wasm.wasm'
      ]
   })
);

the provideDbDriver:

private val workerScriptUrl: String = js("""new URL("@cashapp/sqldelight-sqljs-worker/sqljs.worker.js", import.meta.url)""")

actual suspend fun provideDbDriver(
   schema: SqlSchema<QueryResult.AsyncValue<Unit>>
): SqlDriver {
   return WebWorkerDriver(
      Worker(workerScriptUrl)
   ).also { schema.create(it).await() }
}

The initialization logic:

object StoreApp {
   val scope = CoroutineScope(Dispatchers.Default)

   private var database: Database? = null

   init {
      scope.launch {
         val driver = provideDbDriver(Database.Schema)
         database = createDatabase(driver)
      }
   }
}

This error appears at first moment, with no use... So LATER when i try to do something like: it crashes:

db!!.key_valuesQueries.selectValue("selectedStoreId").asFlow().mapToOneOrNull(Dispatchers.Default),

because of null pointer exception... The Database,kt in commonMain:

expect suspend fun provideDbDriver(
   schema: SqlSchema<QueryResult.AsyncValue<Unit>>
): SqlDriver

fun createDatabase(driver: SqlDriver): Database {
   return Database(driver)

   // Do more work with the database (see below).
}