On  5 June 2019 at 19:29 CEST, Eli Zaretskii <eliz@gnu.org> wrote:
>
>> From: Christophe TROESTLER <Christophe.TROESTLER@umons.ac.be>
>> CC: "36098@debbugs.gnu.org" <36098@debbugs.gnu.org>
>> Date: Wed, 5 Jun 2019 17:05:45 +0000
>> 
>> > What I mean was only to let that function return a unibyte string, and
>> > let the caller perform coding conversions.
>> 
>> I can do that.  To avoid repetitions, what about introducing epg--decode-percent-escape-as-utf-8?
>
> I think that would be fine, yes.

Updated patch doing that.