Zig: is there a way to know the name of an identifier

[ghost]

I want to print the name of a variable or a function.

[daurnimator]

Can you provide an example of what you mean?

[data-man]

Zig has:

@errorName
@memberName
@tagName
@typeName

Example:

var a = 1;
warn("{} = {}", .{@varName(a), a});

So, something like @varName, @fnName and @moduleName would be useful.
Or universal @name.

[ghost]

@daurnimator something like @data-man example

[pixelherodev]

But... all that does is literally emit its argument as a string. Replace @varName(a) with "a" and this does exactly what you want it to. I don't see the use in a builtin for this.

[kristoff-it]

Replace @varname(a) with "a"

Strings are not checked by the compiler when refactoring. The builtin would be more helpful.

[Sobeston]

I'm afraid I don't see the point. Are there any good use cases of this?

[ghost]

@pixelherodev if zig has something like the example, we don't have to do it manually.

[metaleap]

For use-cases, judging by how the C# nameof built-in that was introduced a few years ago is mostly used, mostly revolves around diagnostics code-paths (log outputs, in-production OpenTracing spans, debug-time printfs etc) and certain code-gen scenarios (incl. not just C#-code gen but whatever freakish query-langs or custom templated / config artifacts etc). And "ArgumentException" as in runtime / test-time errors for bad args. Yes, varName would probably be much rarer "needed" than memberName / typeName usages overall. Stably "important" items usually add up to structs and thus memberNames fairly quickly. But the above things come to mind.

[andrewrk]

Zig already has this. Put double quotes around the variable name.

[kavika13]

Maybe a new ticket that is more specific should be opened? Some of the chat here and in Discord thought that closing this issue meant Zig wouldn't support introspect on function/parameter/struct/field names, etc, and I don't think that's the case.

That aside, I think there's some value to not just putting it in quotes, even in the code example given above.

V1:

var a = 1;
warn("{} = {}", .{@varName(a), a});

V2:

var a = 1;
warn("{} = {}", .{"a", a});

If you do an IDE refactor-rename on a, then the string will get out of sync with the variable name in V2, but not in V1.

I might buy the argument that people would probably not even bother to find out that @varName() existed if they were just doing throw-away debug prints. However it might still be useful for logging code that was written to be more maintainable and robust? Also, maybe there's other use cases I haven't considered?

Either way, I think there's at least tiny bit of value to this feature, even if it is low priority.

Edit: I think this is basically what @kristoff-it meant, at least how I read it, but maybe spelling it out in its entirety helps make it more clear :)

[kristoff-it]

Yes, I was just pointing out what the common use is of this type of functionality, or at least how I've seen it being used in C#.

To be honest I'm not a big fan of it, I just wanted to clarify the concept a little bit to make sure people don't miss the point.

Anyway, Andrew seems to be even less of a fan than I am.