Ruben Verborgh

Not everything is implemented in the underlying OIDC library, and as @dmitrizagidulin wrote above, not likely to happen soon.

By default, LDflex will indeed only query the one document corresponding to the URL of the entity. For example, `data[https://ruben.verborgh.org/profile/#me]` will only read data from `https://ruben.verborgh.org/profile/`. This is by design,...

Thanks for reporting; however, I don't fully understand the case yet. What is the relationship between `baseIterator` and `transformIt`?

@jeswr Some questions about the circular list: 1. Shall we extract `CircularList` to a separate file like `LinkedList`? 2. Is a `CircularList` strictly needed? I.e., do the performance gains really...

> both require the traversal of O(n) elements I agree; but it seems to be that it would be `O(10)` at most; unless we are looking at scenarios where the...

Okay; I'm happy to have a circular list. (Although, side-note, what we actually need is a list with O(1) item removal. Which can perfectly be a `LinkedList` with lazy deletion:...

> this would be _preferred_ over a circular linked list implementation Not necessarily; my remark was more that all we need is O(1) deletion. Circular lists as such doesn't offer...

> Yes, since the erroring behavior will change in some cases this will need to be a v4. Could you details these a bit? Because we should line up all...

> we do not subscribe to the error event of iterators until after we start reading from them Can't we change that behavior? Not just for backwards compatibility reasons, but...

> but `follow-redirects` (https://github.com/follow-redirects/follow-redirects) still patches `http` module It does not. Note that `nativeProtocols` and `exports` are local objects. > However, their copying logic fools the Proxy implementation in this...