I recently received this (not the first time) in Mail.app on MacOS 11.4:

On Jun 24, 2021, at 10:02, web host <[email protected]> wrote:
"Head Geek (伟思ç��)" <[email protected]> wrote:

This is curious. I have everything on this end set to use UTF-8 output and to assume UTF-8 if the encoding header is missing. (Unless Apple changed it during an update as they sometimes do).

Why would the first two hanzi be rendered correctly and the third use ISOLatin1?  Is it a bug I should report, and to which (Apple or web host)? It looks correct in the Sent folder. It also looks correct in the To: line of the headers on the same message that has the error in the body.

It has not happened in recent replies for other places, so it's unlikely to be an Apple bug. Yet if the encoding header is missing, odd that the headers are correctly rendered differently.

  • What is the UTF-8 code for the incorrect glyph? The question marks look like the font does not have the relevant character.
    – mmmmmm
    Commented Jun 24, 2021 at 18:56
  • The question marks show that the font is missing two characters, but actually those are two of the three UTF-8 bytes for the character that the font does contain (as evidenced by it rendering correctly in other messages). For what it’s worth, the correct name is 伟思礼 but I typed it from a pinyin keyboard, so I don’t know the U+number.
    – WGroleau
    Commented Jun 24, 2021 at 19:30
  • The unicode number is #x793c of which the first byte is #xe7 whilst ç is confusing me but in UTF-16 is 0x00E7 fileformat.info/info/unicode/char/00e7/index.htm So somthing in the chain of sending and receiving mail has just knocked a leading 0 off. Seeing where it is wrong I think the software that the web host used to edit the text has the issue which is not the software that processes the message
    – mmmmmm
    Commented Jun 24, 2021 at 20:27
  • @mmmmmm, In a follow-up message, they said that is exactly what happened.
    – WGroleau
    Commented Jun 25, 2021 at 1:07

1 Answer 1


Partial answer from the hosting company, and my response:

Please accept our apologies for this, which is a known bug in our third-party ticket management system software. There's a bug in its UTF8 support that seems to affect lines that switch character sets more than once in a given line (as that one does, switching from Latin to UTF8 to Latin).

Since ASCII is a subset of UTF8, but ISOLatin1 isn’t, for it to assume ISOLatin1 and switch to UTF-8 is definitely a bug!

UNLESS Apple failed to provide an encoding header, in which case it is a bug from Apple. In that case, assuming ISOLatin1 is reasonable and attempting to detect and switch to UTF8 is clever, even if it fails.

  • You can see what was received in Mail. View->Message->Raw Source
    – mmmmmm
    Commented Jun 25, 2021 at 6:37
  • Of course. But that doesn’t tell me what they received.
    – WGroleau
    Commented Jun 25, 2021 at 7:01

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .