Hello Pierre-Elliot El 04/05/18 a las 13:52, Pierre-Elliott Bécue escribió: > So, I wonder, maybe this error is triggered because the *source* actually > still contains this wrongfully encoded character? But theoretically I'm not > supposed to alter upstream's source. > Yes, that seems to be the case: I think our gen-material scripts don't handle well the application of patches for packages with format 3.0 (quilt). This is reported in bug #748716.
I'm excluding django-allauth from the l10n stats generation until #748716 is solved, or upstream fixes the bad-encoded string. Should I close this bug? Thanks for your help -- Laura Arjona Reina https://wiki.debian.org/LauraArjona