jfecher

I was actually thinking the opposite, that we could get rid of `Strategy` altogether in favor of only having `recover_via` or other parsers implemented in terms of it. `skip_until` for...

I can use my own fork personally for recover_via, so I wouldn't need a quick merge or anything. If implementing it here would just end up being duplicate work then...

This PR implements #77 where the `recover_with` parser given there is defined here as `recover_via`

I've been thinking a bit on the correspondence between recovery strategies and parsers themselves. For starters, recovery strategies appear to be normal parsers that are called on otherwise failing input....

I don't have many specific thoughts on recovery that haven't been mentioned yet. Regarding discussion of a new release in #76, it would be out of scope to go through...

Definitely, I was even suggesting something simpler in the meantime like `.recover_with(optional(|span| default_value))` would work. Whichever solution you think would be a better fit though will work for me

Related PR: #107 which implemets the `recover_with` parser in this issue as `recover_via`

It would be possible to give better diagnostics if the arguments to a constrain have `const` types during type checking, letting users know it will either always succeed/fail. We can...

What is the unused constraint in the ssa pass? It was my impression that it only affected the interpreter. I still think it should be a warning since it is...

I don't see why the unused constraint would be indicative of a bug anymore than an used variable would be indicative of a bug in classical languages. It is true...