Jake Wharton

https://github.com/square/okhttp/blob/master/okhttp/src/main/java/okhttp3/internal/http/HttpEngine.java#L756-L803 `cacheBody` would be `Okio.buffer(Okio.source(targetFile))` and everything else should be straightforward to modify.

By the way you can use `ResponseBody.charStream()` to get a `Reader` to pass to `Gson` that will use the appropriate charset defined in the `Content-Type` header automatically.

Would be nice if these were inline extension functions on reading a signed value and then "wrapping" it. That makes them zero overhead instead of actually requiring additional methods on...

Related, and also making the rounds: https://itnext.io/modern-storage-is-plenty-fast-it-is-the-apis-that-are-bad-6a68319fbc1a On Thu, Nov 26, 2020, at 11:47 PM, Kevin Cianfarini wrote: > > I found this article interesting on the topic of non-blocking...

This is a pain in the ass because in order to add an Android variant we need to remove `withJava()` from the JVM variant. It's only needed for tests, so...

What is the code you are trying to emit? As far as I can tell, this is working as intended. You either want just `"null"` for a null literal in...

Yeah I'm not super enthusiastic about that behavior, and I'm tempted to even classify the existing `CodeBlock` behavior a bug. Having a null literal produced from `%S` seems quite odd.

I think we can do a few niceties like #39 and then maybe `.apply` and `.let` will get us 90% of what a DSL would without the overhead trying to...

In that case you might want to run ktfmt or ktlint's formatter on the output as part of the generation process. BTW you can replace `endControlFlow().beginControlFlow()` calls with `nextControlFlow()`

How about TypeVariableModifier