Matus Goljer
Matus Goljer
Yes it is and I was changing this yesterday. Const assignments in `let` will get promoted to the "type", `nil` will be infered to be `mixed` since we don't know....
but your example above should've at least updated it to `t` after the assignment. That might be a bug.
@DarwinAwardWinner Can you check on the new version if this is still the issue?
This will probably be fixed by the `elsa-save-scope` mechanism.
I think the first step would be to at least declare the places as existing in-scope variables. Later we can add rules to actually match the (possibly constant) stuff that...
I'm still considering if I want to cheat and do macroexpand. It would no longer be *static* analysis, because macros can do, well, whatever, so you can `rm -rf` your...
I guess something like a whitelist approach might work. I agree about the built-in stuff, that should be reasonably safe. I also trust dash because I know what's there. I...
`side-effect-free` set on `t` means the function has no side effect but can throw an error (e.g. `(+ "foo")`). If it is `error-free` it is side effect free and also...
TODO: what is the relationship between `pure` and `side-effect-free`? ``` (pure "If non-nil, the compiler can replace calls with their return value.\nThis may shift errors from run-time to compile-time.") (side-effect-free...
Yea, the `declare` syntax is the part which doesn't work :D (I commented in another issue). I'm in the process of updating the documentation. The problem is that I don't...