Juraj Kirchheim

There are 3 issues I see with this: 1. I'm not sure this is the right place for this to go. It feels more like this belongs on framework level,...

> I think redirect is as common as ofString or ofBytes. If they are included (and they are), probably redirect should be too. I understand your argument, and FWIW I...

I'm not saying I'm against this either. I just want the library to have a really clear scope and we are in the process of shaping that. My concern is...

There is some duplication between this and some changes I made in the meantime. I will decide how to go about this after the WWX.

Yeah, sorry. So, uhm, as I said I implemented Set-Cookie too, which is here: https://github.com/haxetink/tink_http/blob/master/src/tink/http/Header.hx#L114 I can't say I'm overly happy with it (except the fact that httpOnly is the...

Yeah. I've seen that, although it's kind of silly to add 15KLOC to get that. There's also DateTime, which is smaller, but still has around 4KLOC. I prefer to err...

Whoops, wrong button ^^ Anyway, a similar option to what Steve suggests would be to fallback to a default value in a getter, as that will deal with other null...

That's how modules work. Type `A` is defined in module `Test`, so outside of that module, you have to either: 1. refer to it by fully qualified name, i.e. `Test.A`....

We want `EnumValueMap` to satisfy `map[Some(42)] == map[Some(42)]`, which is why `ObjectMap` won't do the trick. And so we're left with using a tree, to not have everything run at...

> On the JavaScript target, enabling full DCE actually causes getProperty('myField') to fail, presumably because the getter function is getting culled. Getter and setter functions should definitely never get culled...