My bug? I think using .nui or .utl as file extensions is a bug..
Kidding aside. There is no 100% reliable method for detecting the file encoding on files that do not have a BOM. Whatever notepad++ might be doing can't be anything other than guessing.
When it gets slow is when it tries to interpret an OEM or ANSI file, that contains extended characters, as UTF-8. It can't just perform a binary search because most searches are performed case insensitive.
It currently assumes that everything is UTF-8 unless the file extension is a known DataFlex source file and doesn't have a BOM, for those it assumes OEM.
Now I am sure there are smarter ways to do this, and we might investigate how other tools do searches across multiple files. But for DataFlex 20 we will likely leave it as it is now.