Ariel Davis

re: indicating it’s on: i was thinking a config option in millet.toml. as for implementing this feature, it looks like you can translate down the “lazy” language into the regular...

smaller reproducer follows. the problem is not with the `&` but rather with type-checking the commented out line marked BAD. ```sml structure Product = struct datatype ('a, 'b) product =...

the culprit appears to be the usage of 'fru' aka functional record update. according to the [fru sig](https://github.com/MLton/mltonlib/blob/04477d37e709c3bcff9c26a9b00494ee85469f34/com/ssh/extended-basis/unstable/public/fold/fru.sig#L10), we should check http://mlton.org/FunctionalRecordUpdate for more info. this page give a different...

possible optimization opportunities: - structural type sharing to reduce allocations - https://okmij.org/ftp/ML/generalization.html

did the first one (structural sharing) but it's still slow

changing this to be an enhancement request to implement https://okmij.org/ftp/ML/generalization.html, which should make millet as a whole generally faster

This is a bit non trivial especially with larger projects. For instance, if i have a project with a mlb file ``` a.sml b.sml c.sml ``` and my cursor is...

yeah, this is an old todo: https://github.com/azdavis/millet/blob/afb85566b1c1aa6f296bf25044eff105f8f06049/crates/sml-statics/src/top_dec/enrich.rs#L3 it's definitely possible in common cases like this, it would just require us to be better about knowing what `Idx` to emit errors...

I believe this affects more things than just exception copy. Basically whenever in the grammar we have a name or path that is a usage site that is not directly...

some issues i see with 2: 1. it requires the user to have a local sml installation. probably many users will anyway, but if they don't their experience will be...