I recently rediscovered this strange behaviour in Python’s Unicode handling.—Evan

►

-1; there’s no standard for UTF-8 BOMs—adding it to the codecs module was probably a mistake to begin with.—M.-A.

►

There is a standard for UTF-8 signatures, however.—Stephen

►

I would personally like to see an "utf-8-bom" codec (perhaps better named "utf-8-sig", which strips the BOM on reading (if present) and generates it on writing.—"Martin

+1. —M.-A.