> Might this be a newly introduced Mac keyboard input problem? Thank you. I will look into it. I do use Karabiner Elements...
> I suspect there may be a change in behavior with iconv between BigSur and > Sequoia. It would be helpful if I could have an example to work with. > Would you be willing to provide an mbox with the problematic email to me, or > to attach it to a GitLab ticket? Thank you. I try to make an example with this e-mail. I set encode_from=yes so that this e-mail is quoted-printable encoded. It seems I require the following line for this e-mail to be actually quoted-printable encoded. From The five lines below consist of 65 to 69 ASCII number characters followed by two double-byte Japanese characters for "A" and "B". 12345678901234567890123456789012345678901234567890123456789012345ＡＢ 123456789012345678901234567890123456789012345678901234567890123456ＡＢ 1234567890123456789012345678901234567890123456789012345678901234567ＡＢ 12345678901234567890123456789012345678901234567890123456789012345678ＡＢ 123456789012345678901234567890123456789012345678901234567890123456789ＡＢ When you see the raw file of this quoted-printable e-mail, you see at the beginning of Japanese characters the start mark =1B$B, followed by Japanese characters themselves (in our case, #A for "A" and #B for "B"), followed by the end mark =1B(B. Now, since the lines are long, quoted-printable breaks each line by inserting =. The last parts of the above five lines read: 12345=1B$B#A#B= (line break) =1B(B 123456=1B$B#A#B= (line break) =1B(B 1234567=1B$B#A#= (line break) B=1B(B 12345678=1B$B#A= (line break) #B=1B(B 123456789=1B$B#= (line break) A#B=1B(B When I see this e-mail using mutt in Sequoia, the last parts of the above five lines read: 12345ＡＢ 123456ＡＢ 1234567Ａ#B 12345678ＡＢ 123456789ＡＢ Note that at the third line, the raw #B appears which should be the double-byte Ｂ. (I don't understand why the fifth line is OK. It seems #A is split by the newline.) But if I see the same e-mail in BigSur, I see: 12345ＡＢ 123456ＡＢ 1234567ＡＢ 12345678ＡＢ 123456789ＡＢ without the problem occurred in Sequoia. -- Kenichi Asai
