`-vet-unused-variables` produces unnecessary errors when variable is used as a index about odin HOT 8 OPEN

So this is kind of a chicken and egg problem. It's saying it is not used because it technically isn't. But it cannot know that because non_existing doesn't exist.

from odin.

It can be confusing, which is why I'm reporting it. But I understand that the solution might not be trivial. It actually happens frequently so maybe I'll just get used to it, or stop using the flag.

here is another example:

for val, i in non_existing {
	fmt.println(val, i)
}

odin-bug/main.odin(6:16) Error: Undeclared name: non_existing 
	for val, i in non_existing { 
	              ^~~~~~~~~~~^ 
 
odin-bug/main.odin(7:15) Error: Undeclared name: val 
	fmt.println(val, i) 
	            ^~^ 
 
odin-bug/main.odin(7:20) Error: Undeclared name: i 
	fmt.println(val, i) 
	                 ^ 

from odin.

A possible solution is only showing vet errors when there are 0 actual compiler errors. Is that a good solution? Maybe.

from odin.

There is a compiler error though, which is the whole premise of the issue.

from odin.

There is a compiler error though, which is the whole premise of the issue.

Which is why a solution is to not show vet errors when there are compiler errors.

In this case it would only show the compiler error about using an undeclared variable, and not the vet errors about an unused variable.

from odin.

the errors in my last example are not from vet flags actually, so that's not a full solution.
Also it wouldn't be fun to compile, see few errors, correct them, and compile again only to see more errors that got ignored on the first run.

from odin.

I don't think there's an (actionable) solution to this then

from odin.

^ I actually like laytan's proposal to only show vet errors only when there are no compiler errors

The more generic version of this problem is showing inter-dependent errors which I don't think can be solved by patching in one or two lines of code, it has to be done on a case-by-case basis.

from odin.