Remove BitmapSelector.Suffix property

[eerhardt]

With #22761 and dotnet/corefx#22833, BitmapSelector.Suffix will always be null. This means this feature is dead code, and users are unable to use it.

Removing this dead code because:

1. It doesn't do anything.
2. It is causing ILLink warnings, and it is easier to delete than to try to address the warnings.

I logged dotnet/winforms#8832 to follow up and either re-implement this feature, or obsolete the attributes that no longer have any effect on the app.

[ghost]

Tagging subscribers to this area: @safern, @tarekgh
See info in area-owners.md if you want to be subscribed.

Issue Details

---

With #22761 and dotnet/corefx#22833, BitmapSelector.Suffix will always be null. This means this feature is dead code, and users are unable to use it.

Removing this dead code because:

1. It doesn't do anything.
2. It is causing ILLink warnings, and it is easier to delete than to try to address the warnings.

I logged dotnet/winforms#8832 to follow up and either re-implement this feature, or obsolete the attributes that no longer have any effect on the app.

Author:	eerhardt
Assignees:	-
Labels:	area-System.Drawing
Milestone:	-

[tarekgh]

It is causing ILLink warnings, and it is easier to delete than to try to address the warnings.

I am not objecting the changes here as it makes sense. Just wondering, if we decided to implement this feature, do you think the linker can complain at that time again? or it depends on what we'll do during then?

[eerhardt]

if we decided to implement this feature, do you think the linker can complain at that time again? or it depends on what we'll do during then?

The linker was complaining about the following lines:

runtime/src/libraries/System.Drawing.Common/src/System/Drawing/BitmapSelector.cs

Lines 88 to 131 in 72a26f9

	[RequiresUnreferencedCode("Calls Assembly.GetType which may be trimmed")]
	private static bool DoesAssemblyHaveCustomAttribute(Assembly assembly, string typeName)
	{
	return DoesAssemblyHaveCustomAttribute(assembly, assembly.GetType(typeName));
	}
	private static bool DoesAssemblyHaveCustomAttribute(Assembly assembly, Type? attrType)
	{
	if (attrType != null)
	{
	var attr = assembly.GetCustomAttributes(attrType, false);
	if (attr.Length > 0)
	{
	return true;
	}
	}
	return false;
	}
	// internal for unit tests
	internal static bool SatelliteAssemblyOptIn(Assembly assembly)
	{
	// Try 4.5 public attribute type first
	if (DoesAssemblyHaveCustomAttribute(assembly, typeof(BitmapSuffixInSatelliteAssemblyAttribute)))
	{
	return true;
	}
	// Also load attribute type by name for dlls compiled against older frameworks
	return DoesAssemblyHaveCustomAttribute(assembly, "System.Drawing.BitmapSuffixInSatelliteAssemblyAttribute");
	}
	// internal for unit tests
	internal static bool SameAssemblyOptIn(Assembly assembly)
	{
	// Try 4.5 public attribute type first
	if (DoesAssemblyHaveCustomAttribute(assembly, typeof(BitmapSuffixInSameAssemblyAttribute)))
	{
	return true;
	}
	// Also load attribute type by name for dlls compiled against older frameworks
	return DoesAssemblyHaveCustomAttribute(assembly, "System.Drawing.BitmapSuffixInSameAssemblyAttribute");
	}

What this was doing was looking for these 2 attributes in the same assembly as the Type. If the public attribute wasn't found, it fell back to looking for the attribute by name (assembly.GetType(typeName)), which is what caused the linker warning. These attributes have existed since .NET 4.5, but this was supporting assemblies built against older frameworks.

If we decided to implement this feature, we could decide to not support looking these attributes up by name in the assembly and only look for our publicly declared attributes. This wouldn't cause any warnings.

Alternatively, the linker will only trim attributes it is explicitly told to trim, for example here. So another approach could be to just suppress the linker warning, and assuming no one is trying to trim these attributes.