Is there a way to specify a mapper for a query that selects a single column? This is mostly for aesthetics so patterns match regardless of the number of items returned -- I can obviously

map

the list after I've executed the query

I'm using MySQL database, and have this code for migration:

val dataSource = HikariDataSource(datasourceConfig)
val driver = dataSource.asJdbcDriver()
driver.migrate(Database.Schema)
//...
private class SqlDriverTransacter(driver: SqlDriver) : TransacterImpl(driver)

fun SqlDriver.migrate(schema: SqlDriver.Schema) {
    val sqlDriverTransacter = SqlDriverTransacter(this)
    // ...
    sqlDriverTransacter.transaction {
    	// ...
    }
}

I don't remember exactly where from this code came from, it looks kind of like this one. The problem is that in case of a crash inside

transaction

, the database is not getting rolled back. When in my migration I have

create table

and then some action, if the action fails, next time I'm getting a crash

Table already exists

. I've tried catching the error inside

transaction

and running

rollback

manually, but this had no effect.

Is there a trick to using types like BIGINT in a .sq file for use on Postgres? https://cashapp.github.io/sqldelight/jvm_postgresql/types/ Documentation suggests it should work, but code generation fails:

schema.sq line 17:12 - <type name real> expected, got 'BIGINT'

I'm using 1.5.3

there is no Molecule compiler plugin

Our presenters do not retain state. You're free to hook up the regular save instance state stuff like Compose UI does though.

Is there a way to dynamically switch between tables that have the exact same schema but different names in SQL Delight? I have a database with one table per language I'm wondering whether it would be feasible to switch between tables when the user selects a language (I'm not sure how to go about doing that since the schema classes are automatically generated and are named the same as the table, so it doesn't seem like a generic class name can be used).

no. you would need to copy it from your apk into the database directory of your private apk files before opening.

You are using Compose for presentation logic, but aren’t using Compose for UI?

by "you" are you referring to Square Block? or the library?

Short answer is no, because for Anvil a “scope” only needs to be a class reference. It doesn’t need to be a Dagger

@Scope

. For example our scope annotation looks like that:

@Scope
annotation class SingleIn(val value: KClass<*>)

The usage side looks similar to this:

@SingleIn(AppScope::class)
@ContributesTo(AppScope::class)
@MergeComponent(AppScope::class)

Anvil wouldn’t be able to make the connection that

@SingleIn

is the actual Dagger scope.

If I needed to cache something for "offline" mode for an api that doesn't send back any caching headers. Would you: 1. Write an interceptor that adds some sort of basic cache header so that even without a network connection the http client still returns data I can display on the screen 2. Just save my json response as a file and read from that if I'm "offline" Note: I know "offline" is not really a true state, but in my case my PM just wants data to show when the user is actually in airplane mode.

SQLDelight for JVM is ether not seeing my .sq file, or the gradle tasks for not generating my database is failing and saying it is succesfull or maybe i have no clue whats going on

What are the recommendations on testing flows? With runTest I always run into a TimeoutCancellationException. For example a very simple reproducer:

@Test
  fun myTest() = runTest {
    flow {
      delay(10.milliseconds)
      emit(Unit)
    }.test {
      awaitItem() shouldBe Unit
    }
  }

I've got a requirement to make my deserialization crash during runtime on "internal" builds, but not crash on external builds when I get data over the wire that is null when it's not supposed to be null. So with moshi, unless I declare something as explicitly null, I will not be made to handle nullability in that type in my code (makes sense) but "moshi" will still give me a null value if it's sent explicitly over the wire. So now I basically have a fork in the road in deciding on how to move forward. 2 options from what I see: 1. For every model... create two models, put one in /external and one in /internal. The ones in /internal would not be nullable, while the ones in /internal would be nullable which would make me handle nullability, but I would crash on deserialization in internal builds. 2. I can do a "check" on my data when it comes down and just make everything nullable, but in my checking layer I would just verify that my data is in a good state. I could have sworn I read a stackoverflow answer from Jesse Wilson on this topic a while back and he basically said something along the lines of "instead of depending on your deserializer to do this sort of checking for you, you can just do it yourself and the checking would then not be "hidden" in your deserializer." Anyone got any opinions on how to move forward with this? I think I'm leaning on option 2. Make everything nullable (because even though my backend says "hey we won't send you null" sometimes it does get through. And then there are some other invariants I could assert. i.e. Hypothetical, but person age will never be 0. If you get zero something is terribly wrong. In that case, I could crash in internal builds, but I could set some kind of default (or remove the feature from the UI) in external builds. I would love to just always crash in both internal/external, but I was overruled on that. 😢

Hi team, hope you don’t mind a Cashapp/Turbine question here. This is a cross-post from #coroutines b/c I’m not quite sure if Turbine is influencing the outcome here.. I’m having trouble with a unit test. I’m trying to validate the behaviour of a StateFlow. But, the StateFlow is only in this particular state until a certain suspend function completes. During testing, my test double uses a delay to simulate a long running task. But, due to the nature of 

runBlockingTest()

 , this delay is skipped, so this function completes immediately, and the StateFlow essentially bypasses the value I’m trying to test.

I have a library which is going to be transformed to KMP and for testing I have a few json files inside resources directory. If the library has no JS target, I could use Okio, but if it has JS, I should transform those json files to strings or I could use them via Okio?

I'm doing some testing using moshi and trying to see what happens when I have a field defined as

val title: String = "default"

, but then my server sends a number by accident. It seems like moshi will convert the json number into a String type. Is that right?

Hello 👋 , I’m trying to intercept the request body with OkHttp when I’m doing a call with wire-grpc. How can I decode the

RequestBody

after doing

val buffer = Buffer()
requestBody?.writeTo(buffer)
buffer.readByteArray()

?

Need something similar to what https://github.com/mattprecious/protogram does. This is doable with Wire APIs

Using KotlinPoet, how would I create the ParameterSpec of class B in the following example?

abstract class A(
    open var field1: String
)

abstract class B(
    override var field1 : String
) : A(field1)

I tried to also create a corresponding PropertySpec for my class, but as a result I get the following:

public abstract class B(
  override field1: String
) : A() {
  public override var field1: String
}

what dialect are you on

If I was writing a feature/library that would intercept all network traffic and show it on the device (essentially building an on device charles proxy), it seems like I would use a network interceptor over an application interceptor. Is there anything that would still be hidden, or is it safe to say "an okhttp network interceptor would be able to show all of the same information as a typical desktop proxy (like charles)"

I updated android gradle plugin and android studio from alpha/canary 3 to alpha/canary 7. No other changes. Now after doing all sorts of nuking/restarting I'm left with these errors messages. I'm assuming there's no reason why okhttp and moshi wouldn't work with agp alpha 7 (released today)? I'm doing some more research on the android bugtracker to see if this is worth reporting, but figured id chime in here to see if anyone else upgraded to agp alpha 7 today. This error is actually in a java/kotlin lib. nothing android related in it. Edit: Should have mentioned my code actually still builds fine. just the IDE with these error messages.

I’m working with SQLDelight and I have the following relation A

1 to N

B B

1 to N

C Sometimes I’ll need a

List<A>

, sometimes just a

List<B>

. Constructing the objects, like

A

, based on queries seems a bit confusing to me, so wondering if someone could point me a to a sample that implements something similar? Thanks 👍

I don't know about Postgres but for MySQL on servers at Cash App we never just call the database to apply schema changes. Instead we use gh-ost so the DB isn't locked and unavailable during the schema change

Hi all. Trying to use SqlDelight on a KMP library module, (which also targets

js(IR)

). I’m making use of the npm-publish plugin to export my KMP library locally to npm, so that i can

link

these kotlin-js artefacts from pure js projects. However, when I invoke

initSqlDriver()

from jsMain, i get this on the console

wasm streaming compile failed: TypeError: Failed to execute 'compile' on 'WebAssembly': Incorrect response MIME type. Expected 'application/wasm'.
failed to asynchronously prepare wasm: CompileError: WebAssembly.instantiate(): expected magic word 00 61 73 6d, found 3c 21 44 4f @+0

Has anyone encountered this before?

Do you have sql.js as a dependency on your pure JS project?

For handling a typical oauth flow from my server where I get a 401 and then I want to try to use my refresh token to get a new access token. Best to just handle in an interceptor (network or application) or just use the authenticator apis in okhttp?

but outside of sqlite i think most use

deriveSchemaFromMigrations

What you wanna do with the iif?