Henry Tung

@iamdanfox Don't disagree, but am moving a bunch of code into pgdev, so not looking into changing the format yet. Can revisit after that's settled.

@carterkozak Well, "cleanup" here is more of a generalization - might involve some diagnostic logging that might be useful to user but risky, for example. Basically may involve some things...

@cowtowncoder Given that `@JsonAnyGetter` and `@JsonAnySetter` exist, how correct/incorrect would an implementation be to have `@JsonUnwrapped` imply the other two annotations when used with `MapSerializer`?

Ping! Running into this issue right now; this commit seems straightforward enough, but concerned that it doesn't look like it's gotten attention for over a year. See https://github.com/substack/node-browserify/issues/1351. Would be...

Ping?

@zertosh Can you provide a repro or description of the case you ran into? May be able to take a look if I have a place to start from. Thanks!

I would second not the approach highlighted in the last comment (mutating existing definitions) but the original goal of being able to exclude definitions from the scope of `tsd link`....

@chrisalice https://github.com/palantir/gradle-processors/blob/master/src/main/groovy/org/inferred/gradle/ProcessorsPlugin.groovy#L103 `.minus += [compile/runtime]` ?

Although, to contribute some more semantics to the discussion: `processor` is effectively an _unnecessary-at-runtime_ compile-dependency. It compiles code against the library (as an annotation provider) but, by default, does not...

Would add https://square.github.io/okhttp/3.x/okhttp/okhttp3/OkHttpClient.html docs about immediate shutdown, in particular `ConnectionPool.evictAll()` (probably to gracefully shut down idle connections with other servers rather than just vanishing).