Should `Type.AssemblyQualifiedName` be cached?

[Youssef1313]

On an MSTest project with huge parameterized test, I see:

Test class looks like this:

[TestClass]
public class TestClass
{
    [TestMethod]
    [DynamicData(nameof(Data))]
    public void Test3(int a, int b)
    {
        if (b == 10000000)
        {

        }
    }

    public static IEnumerable<(int A, int B)> Data
    {
        get
        {
            for (int i = 0; i < 10000000; i++)
            {
                yield return (i, i + 1);
            }
        }
    }
}

[AaronRobinsonMSFT]

@Youssef1313 I'm a bit suspicious of that callstack. It implies that get_AssemblyQualifiedName is calling back into the test platform, which seems rather odd.

Additionally, we already cache this information at the following location:

runtime/src/coreclr/System.Private.CoreLib/src/System/RuntimeType.CoreCLR.cs

Lines 3337 to 3352 in dfb2850

	public override string? FullName => GetCachedName(TypeNameKind.FullName);
	public override string? AssemblyQualifiedName
	{
	get
	{
	string? fullname = FullName;
	// FullName is null if this type contains generic parameters but is not a generic type definition
	// or if it is a function pointer.
	if (fullname == null)
	return null;
	return Assembly.CreateQualifiedName(Assembly.FullName, fullname);
	}
	}

[Youssef1313]

get_AssemblyQualifiedName is called by MSTest, not vice versa. I'm not sure why you think it's the other direction?

From the linked snippet, FullName is cached, but not AssemblyQualifiedName

[AaronRobinsonMSFT]

get_AssemblyQualifiedName is called by MSTest, not vice versa. I'm not sure why you think it's the other direction?

Oh, I see. This is the flipped call tree. Normally the caller is at the top, this is the callee view.

From the linked snippet, FullName is cached, but not AssemblyQualifiedName

Alright, let's see if we can cache this as well.