How to handle SpanExtensions methods to prevent shadowing MemoryExtensions methods

[Octanoz]

Currently when you take a row span from the Span2D you're working with by calling mySpan2D.GetRowSpan(rowNumber) and, say, you wish to know the index of 'S', the compiler will tell you you need to use a variable because it is passed to a ref readonly parameter. The using statement using CommunityToolkit.HighPerformance at the top of the file will make the compiler ignore all the methods in MemoryExtensions if they have the same name. So I ended up scratching my head, "Didn't I use this before with the same parameter and it worked just fine?").

So you could wrap everything in spans beforehand:

var (start, exit) = (Coord.Invalid, Coord.Invalid);
Span<char> markers = ['S', 'E'];
Span<char> startMarker = [markers[0]];
Span<char> exitMarker = [markers[1]];

int row = 0;
while (start == Coord.Invalid || exit == Coord.Invalid)
{
     var rowSpan = grid.GetRowSpan(row);
     if (rowSpan.ContainsAny(markers))
     {
         switch ((start == Coord.Invalid, exit == Coord.Invalid))
         {
             case (true, true):
                 int sIndex = rowSpan.IndexOf(startMarker);
                 int eIndex = rowSpan.IndexOf(exitMarker);
                 //etc...

You could write the MemoryExtensions extension method in full.

int sIndex = System.MemoryExtensions.IndexOf(rowSpan, 'S');

You could use a static using statement to the base library

using static System.MemoryExtensions;
//...
int sIndex = rowSpan.IndexOf('S');

Or wrap the value within the call itself

if (rowSpan.ContainsAny(['S', 'E']))
//...
int sIndex = rowSpan.IndexOf(['S']);

The last approach works for everything that's shadowing its MemoryExtensions cousin, I think. But it took me a while to figure out, perhaps because I haven't been programming for very long and not that used to working with Span yet (that was the first ref readonly parameter I ever saw).

The most elegant approach is definitely the static using statement but I only figured that out while writing this post. I had assumed it would make the calls ambiguous but that's not the case so perhaps that could be mentioned in the documentation or in the tooltip?

An overload to the MemoryExtensions method would work as well but I suppose it's a bit awkward that the high performance method calls the regular method...