Re: mhbuild and long header fields

[Philipp]

[2023-08-25 23:46] Steffen Nurpmeso <steffen@sdaoden.eu>
> Steffen Nurpmeso wrote in
>  <20230825213130.xthnK%steffen@sdaoden.eu>:
>  |Philipp wrote in
>  | <93c190b84a23e620d07b9504f619b1f2.philipp@bureaucracy.de>:
>  ||[2023-08-24 21:07] Philipp Takacs <philipp@bureaucracy.de>
>  ||> [2023-08-23 14:29] Philipp <philipp+mmh@bureaucracy.de>
>  ||>> [2023-08-20 22:14] David Levine <levinedl@acm.org>
>  ||>>> Ken Hornstein wrote:
>  | ...
>  ||I have tested this and fixed some bugs. A commit with a test is attached.
>  |
>  |I have not looked at your code (for real), but .. i have a test
>  |for the MUA i maintain like
>  |
>  |  i=$(awk 'BEGIN{for(i=0; i<92; ++i) printf "0123456789_"}')
>  |
>  |and then i use "$i" as a subject.
>  |The problem with RFC 5322 is that artificial spaces may have to be
>  |introduced to break a line without whitespace.  (That is, you have
>  |to break somewhere, and have to start the next line with
>  |a whitespace, but that is included then, modifying the original
>  |data.)
>
> That is: you should switch to MIME RFC 2047 then for the purpose
> of uniting maybe even.  Or however nmh wants to deal with this.

Thanks for your feedback, but I don't think general encoding every field
with a line longer the 78 (998) bytes is the way to go. Because fields
which are to long are uncommen. Also fields which tend to be to long have
most of the time contain white space[0]. So folding is possible.

That said, my code detects lines without WSP (and are longer then 78
bytes). Some rfc 2047 encoding could be added there. But first the
folding should be implemented.

Philipp

[0] I speak of fields you would include raw in a draft file, not binary
data like a dkim signature. Such data is mostly direct rfc 2047 encoded.