Use filelib:wildcard rather than filelib:fold_files. It is 44 times faster.