[I18n-sig] JapaneseCodecs 1.1 with ISO-2022-JP codec

[Martin v. Loewis]

> In addition, the ReadStream classes for all encodings are
> improved so that read(), readline() and readlines() do not
> return an empty string unless an EOF is reached.

Does readline() correctly deal with line breaks under all
circumstances? In some encodings, detecting line breaking is more
difficult than looking for a CR or LF byte; e.g. in UCS-2, you need to
consume two bytes for a line break, and a CR byte may be part of a
two-byte code that doesn't represent a line break.

Perhaps your encodings "play nicely" here?

Regards,
Martin