Hisham Muhammad

``` local record f g: function(integer) g: function({string: any}) end f.g({true, hello="world"}) ``` It seems to get confused with tables that mix array and map parts (which doesn't fully surprise...

Ok, so here's what's happening: * on **non-polymorphic functions**, we use bidirectional inference (that is, we use the knowledge that the table literal is inside a function argument, sending typing...

Hi @pigpigyyy, thank you for taking the time to make a PR! The code itself looks good, but I'm uncertain about adding these features to the language. These are the...

> I think it is best to treat as as how Rust treats Transmute, it is there if you really need/want it but... YOU need to make sure you don't...

> As for nominal records, I just changed the way to compare records with embeds in order to make subtypes to be not equal. Thanks @pigpigyyy, I'll give it another...

@muratg Thanks for the feedback! Regarding this feature proposal, I think the direction Teal should go is to adopt real intersection types. Various features in the language already hint in...

Hi there, thank you for all the feedback! Here's my feedback on the items above > Ability to pretty print the AST as the original tl source instead of just...

Update: I just saw #203... it does add a lot of code, and that's even without the formatting preservation. I'm not sure if I want to take the printer function...

No plans to do that at this point. This kind of feature interacts with the entire type system in super subtle ways, and as these things are added, they compound...

Thank you for the writeup! It would be really awesome to have something like this indeed. I guess we could start with the syntax for doc comments: the usual standard...