ColumndsOfCurrentToken to struct tuple (dotnet#9219)

Author: cartermp

src/fsharp/service/ServiceLexing.fs

@@ -616,15 +616,15 @@ type FSharpLineTokenizer(lexbuf: UnicodeLexing.Lexbuf,
                     | Some mx when rightp.Line > leftp.Line -> mx
                     | _ -> rightp.Column
                 let rightc = rightc - 1
-                leftc, rightc
+                struct (leftc, rightc)
 
             // Get the token & position - either from a stack or from the lexer
             try
                 if (tokenStack.Count > 0) then true, tokenStack.Pop()
                 else
                   // Choose which lexer entry point to call and call it
                   let token = LexerStateEncoding.callLexCont lexcontInitial lexargs skip lexbuf
-                  let leftc, rightc = ColumnsOfCurrentToken()
+                  let struct (leftc, rightc) = ColumnsOfCurrentToken()
 
                   // Splits tokens like ">." into multiple tokens - this duplicates behavior from the 'lexfilter'
                   // which cannot be (easily) used from the language service. The rules here are not always valid,