ColorScheme.SetAttribute is defunct

[tig]

Current code:

		bool preparingScheme = false;

		Attribute SetAttribute (Attribute attribute, [CallerMemberName] string callerMemberName = null)
		{
			if (!Application._initialized && !preparingScheme)
				return attribute;

			if (preparingScheme)
				return attribute;

			preparingScheme = true;

			switch (caller) {
			case "TopLevel":
				switch (callerMemberName) {
				case "Normal":
					HotNormal = Application.Driver.MakeAttribute (HotNormal.Foreground, attribute.Background);
					break;
				case "Focus":
					HotFocus = Application.Driver.MakeAttribute (HotFocus.Foreground, attribute.Background);
					break;
....
			}
			preparingScheme = false;
			return attribute;
		}

Shouldn't preparingCode be a lock?

While I'm here, the logic in the switch statement has always confused me. I do not understand the intent! It seems arbitrary to be making style decisions in code like this. Can anyone explain clearly?

[tig]

I think i'm figuring out. New docs:

/// <summary>
		/// 
		/// </summary>
		/// <remarks>
		/// <para>
		/// If <see cref="Application"/> is not initialized yet and <see cref="_preparingScheme"/> is set to <see langword="false"/> 
		/// then the attribute is returned as is (because it's not possible to modify a Scheme wihtout a valid Driver). 
		/// </para>
		/// <para>
		/// If <see cref="Application"/> has been initialized and <see cref="_preparingScheme"/> is set to <see langword="true"/> it means
		/// <see cref="SetAttribute(Attribute, string)"/> is being called recursively and the attribute is returned as is.
		/// </para>
		/// <para>
		/// Otherwise, the attribute returned is created by Driver.MakeAttribute, using the 
		/// foreground and background colors of the attribute.
		/// </para>
		/// </remarks>
		/// <param name="attribute">THe attribute to set.</param>
		/// <param name="callerMemberName">The name of the property (e.g. <see cref="Normal"/> or <see cref="Focus"/>
		/// that called the function. This is used to determine how to make the attribute.</param>
		/// <returns>The attribute that was set.</returns>
		Attribute SetAttribute (Attribute attribute, [CallerMemberName] string callerMemberName = null)

[tig]

And... I just set a breakpoint here:

And I ran all unit tests and UICatalog and it was never hit.

Is this all defunct code!??!??!

[BDisp]

This was made to create automatic attributes, but since they are been creating on the Init method this wasn't used anymore. I don't know who create the preparingScheme flag but for sure was to avoid running the bellow code. I think this private method can be deleted and set the attributes directly with the value.

[tig]

After diving in deeper, I agree!