It seems that pragma definitions in source are not only case insensitive, but can accept UK or US spelling variants.

I am going to need to capture SourceText values in the tokens, and bring them through into the AST.

Alan