Matthias Benkort

To give some context, this comment mostly stemmed from empirical measurement and benchmarks using profiling on this project: https://github.com/CardanoSolutions/ogmios -- which has heavy needs on JSON-serialization (hundred of megabytes of...

Just to be clear from my end: I am not particularly interested in this conversation / outcome. I've got other matters to tackle and I was merely providing a bit...

Possibly, but it may be difficult since the API of `fast-bech32` is actually quite different from this one... Simplified, but also arguably less typesafe. So, porting part of the (optimized)...

Related to: https://github.com/aiken-lang/aiken/pull/809

Any update about this :) ? Edit : Forked the repo v6.0.0.beta, and applied changes in the source code to take in account what I called a "foreign-type". Accessible and...

> @zliu41 Do scripts care about the order of ScriptPurposes? Why would they care? Scripts don't generally; but compiler makers and library builders do 🤓! > @michaelpj: I don't think...

> @michaelpj: I was very confused before, but it seems like you are mostly talking about the Map constructor of Data, and not about how the ScriptContext type is represented...

> @zliu41: doesn't Aiken (or any language targeting UPLC) need to provide a map type that enforces uniqueness, since users may want to use maps in the scripts? Yes, but...

@zliu41 that is _already_ what we do; there's actually no extra cost in constructing a map from the script context. but to circle back to the root of the problem...

> @zliu41: As far as I understand this shouldn't lead to higher script evaluation cost, since you can use the specialize map whenever applicable, just like how you'd use IntMap...