Hisham Muhammad
Hisham Muhammad
I think the addition of interfaces will require us to introduce a `self` type. I believe that should address this issue (by making any `function` whose first argument is of...
(for backwards-compatibility, we could even extend this behavior to functions whose first argument is the record itself and the argument _name_ is `self`... `self` is already a special name in...
> Is there a timeline for the introduction of interfaces? No timeline, sorry, but it's high in my priority list. > From my understanding, this could be implemented before interfaces...
> I think Teal does not complain about the third line due to the assumption that it MAY be correct, AND there is currently no way to mark function parameters...
> Can Teal distinguish between record objects and record instances? Internally, yes. A record instance is of type `record`, a record declaration is of type `typetype`, whose `typetype.def` is a...
For future reference: #640 says "closed" above but I merged it manually!
I see what you mean. From a quick glance at the code, I think one way to keep it single-pass is to trickle down the type name as an extra...
@JR-Mitchell To deal with generics, within `parse_record_body` you have `def.typeargs` available. If your function's self type has `typ.typevals`, you can check one against the other.
@LaylBongers Thanks for the report! I think one solution for this would be to move the `normalize` function from the `tl` script into the `tl.tl` module, and then use it...
It's not something I intend to implement for records in the foreseeable future.