Ruben De Smet

> I've been using `divan` instead of `criterion` lately, the API is friendlier. Ack on `divan`! Haven't used it yet, but I have heard about it enough :-)

> From what I see most of the time is spent parsing/allocating regex. Looks like it indeed! What do you propose going forward? I'm still quite sure we can find...

> Ah yes @ds-cbo notified me about the merge (we work in the same company 😆) Small world :') > But in general there is a lot we can do...

As a side note: I recall a discussion on Discord where there was a parameter dump for a DO curve that embeds Curve25519, but I don't have the link at...

Pushed a few fixups, will discuss the rest with @Tarinn when we both have time at the office! Thanks for the review so far.

@Pratyush, @samvandevelde noticed that the current implementation was tailored to Jq255s, where A=-1. This has meanwhile been resolved (766b91a). He'll also take care of your further comments. Thanks for the...

> [..] should we propagate the unparsable `id()` call too? ```diff diff --git a/src/error.rs b/src/error.rs index c7799df..0c4fa1a 100644 --- a/src/error.rs +++ b/src/error.rs @@ -60,7 +60,7 @@ pub enum Parse {...

If someone wants to tackle this: `#[deny(clippy:unwrap_used,clippy::expect_used)]` is a thing, but yields quite a lot of errors.

> One on each wrist??  In principle, sure, but people were getting worried about the size of my left arm.

> I love my PTS. Getting a bit anxious about what I'll do if it breaks - it's seven years old now and still going strong. Ebay still has them...