On Tue, Feb 27, 2018 at 01:56:28 -0000, Mutt wrote: > As it is hard (and makes less to no sense) to support IDNA2003 and > IDNA2008 in parallel, I would recommend to replace IDNA2003 by IDNA2008 > support, which means that the requirement in the end switches over from > libidn to libidn2 simply.
This is probably still very welcome, e.g. for mutt-1.10 (wow, are we there already, it seems like only yesterday we were stuck at 1.4.xx forever ;->). The patch, which you integrated into the Fedora package, as I noticed, it not completely correct syntactically, as it continues to use the legacy return code checks (IDNA_SUCCESS) while libidn2 provides its own (e.g. IDN2_OK). Pre-2017 versions of libidn2 didn't provide the legacy enums. Moritz
