Artyom Shalkhakov

Looks like a type provider can be run via `fsc` on command line. Maybe it's all that I need, will check it out this evening.

What is needed to convert a generative TP to an erasing one? I don't have any experience implementing TPs. But I would like to use Rezoom.SQL on very, very old...

I also hit this, but on Linux, and with MySQL 8. Since this changes the way MySQL works by default, it might break other applications using the same database.

Regarding proposal 3: I think I've implemented a [drop-in replacement](https://gist.github.com/ashalkhakov/b3e4a9c2593600c930d1) for Runtime.Data.NativeRow that will store the elements in unboxed form, and will allow getting/setting their values using type-parameterized methods.

Attempted proposal 1. Issue: NilException is really pervasive! There are many places where nil is a perfectly fine result of evaluation, and there are way too many places where Evaluate...

> Having said that, I like the suggestion that an operator be able to inform the compiler about it's nil propagation behavior so that the nilable characteristic could be inferred...

Please find an [example](https://github.com/ashalkhakov/Dataphor/commit/50c01a49683f88ef8bc75814808f3472f1fab92c) of proposal 2. Obviously it needs more polish (I don't follow coding conventions..), but it shows concretely what I am proposing to do. I would like...

Stumbled on a generic nil. In an expression like `nil+5`, nil is assigned the type Generic. How should it be treated? Its type is unknown at compile-time. How is generic...

Thanks for the note on generic. I've implemented some basic imperative node compilation. Currently using the interpreter stack (so every assignment and variable reference involve boxing/unboxing), but I'd like to...

Note that I've been working on this some more, and the progress so far is as follows. 1. added more operators and more types 2. ensured that existing Dataphor tests...