Hi Walter,
i did not look closely at the patch yet, and i did not dig for standards
documents, which one should almost certainly do before committing such
a patch unless one knows all the relevant standards by heart (which i
do not), so i'm not saying this must be done differently, but instead
i am merely asking questions.
1. Are you really sure that a header like
MIME-Version: 1.0
is not needed when you add Content-*: headers?
2. Are you really sure that a header like
Content-Disposition: inline
is not needed?
3. What is the reason for not simply hardcoding
Content-Transfer-Encoding: 8bit
when sending UTF-8 mail?
Are there really still MTAs that choke on that in 2023?
quoted-printable is definitely a PITA no matter the context,
so in my book, if it can be avoided, avoiding it would be a plus.
4. What's the motivation for the -y flag taking an argument
and not simply hardcoding "text/plain;charset=utf-8"?
OpenBSD does not support any other charset and does not plan to
change that in the future.
I hope your next patch isn't going to be support for text/html. =:-S
5. What's the motivation for supporting -y without -e
and for supporting -e without -y ?
In general, we want as few options as possible and as little
configurabity as possible. If there is a sane use case for something -
in this case, sending UTF-8 mail - *one* option is possibly warranted.
But adding more than one option would need a very robust justification,
and so would adding an option that takes an argument.
Note that mail(1) is not mail/swaks. Its purpose is reading and
sending mail in a *simple* way, not low-level testing or protocol
debugging.
I'll postpone code review and testing, maybe you can simplify this
first?
Yours,
Ingo
