Caner Demirer

As someone mentioned on [Reddit](https://www.reddit.com/r/Zig/comments/xam2fd/andrew_kelleys_livestream_improving_the_user/), fixups done automatically by the compiler should probably have indicators: > `_ = somevar; // [[autofix unused variable]]` This is important for 2 reasons: *...

I guess `GetTypeInfo::VARIANT_TYPE` can be used instead of dummy Variants.

What should be the desired behavior here? The underlying assert operation can print anything just fine (not necessarily strings), which can be seen by removing the check (see #47165) that...

In that case, constant non-strings would also be fine, right? And the warning message would now be about the risk of side effects that wouldn't be included in the export,...

~~Maybe the programmer is lazy and just wants the value printed, like `assert(condition, my_vector)` :stuck_out_tongue: (it actually prints properly)~~ Nevermind that only makes sense for non-constant non-strings. Though if it's...

> While left side, i.e. the actual expression can still be non-constant. That's right! Until GDScript has const function support, I guess there is no reason to consider side effects...

I'd say there are good reasons to assume the parser would never look for a statement at the EOF token because in all expectable cases there would be some non-EOF...

> Perhaps a related case, shadowing a variable in a for in uses the value from the outer scope: > But it doesn't fix the issue with for. I think,...

I think this issue might be receiving less attention than needed due to the title. I think it needs to be updated to emphasize that dictionary keys are neither stored...

> If dictionary keys are neither stored by value nor reference, what are they stored as then? I intended to mean "not by value only or by reference only" :facepalm:...