Dominic Raferd wrote in
 <[email protected]>:
 |On 05/07/2021 15:27, Steffen Nurpmeso wrote:
 |> Dominic Raferd wrote in
 |>   <[email protected]>:
 ...
 |>|2. Send same with s-nail v14.9.15, 2019-08-17 (built for Linux):
 |>
 |> Yes, pretty old and unfortunately distant from unbuggy.
 |This is the version in Ubuntu 20.04. Is there a safe/easy way to install 
 |latest'n'greatest s-nail in place of it?

Oh sorry, i have no idea of Ubuntu.  I know Paride does a good job
and even supports backporting on Debian (thought the last update
he did not backport i think).
If you have make and compiler it is just "make tangerine".
Google says

  Enabling Backports Manually

  Make sure the following line is in your /etc/apt/sources.list
  (substituting your release for trusty):

  deb http://archive.ubuntu.com/ubuntu trusty-backports main restricted 
universe multiverse

There is also something graphical, but i really do not know.

  ...
 |>|Am I doing something wrong? I realise that CR+LF line endings are not
 |>|'normal' in Linux environment but I think that s-nail should handle them
 |>
 |> You are feeding in a data file, POSIX says
 |>
 |>    3.403 Text File
 ...
 |>|My workaround is to remove CRs from header lines and the succeeding
 |>|blank line before passing source file to s-nail, but this is ugly.
 |>
 |> Not creating those sequences at first to turn valid text files
 |> into Windows 8-bit aka DOS ones at first glance may be a better
 |> option.  Earlier there were dos2unix and unix2dos programs iirc.
 |>
 |> Having said that, and since i am totally opposed to mutilating
 |> user data, we now at least keep the line ending type intact for -a
 |> ttachments, for example, after the v15 rewrite we have a different
 |> code flow and might be able to not only preserve newline type when
 |> saving edited files.  But i will not hack it into the current
 |> state of affairs.  Sorry.
 |
 |OK understood. Just to explain: the reason I have CR+LF line endings in 
 |real life (unlike the reproducible example I gave in my message) is that 
 |such text files come from Windows (or even DOS) machines.

It sounded mysterious with headers in Windows format and Body in
Linux format.  Otherwise i would claim the program backing your
sendmail(1) is broken, since it silently converts a Windows text
file into a Unix one.  I really had to look into the standard,
i wonder wether the truly right thing would be to convert to MIME
and use a content-transfer-encoding in order to be able to restore
to DOS format or not.  ..Thinking about it, likely not.  It may be
ok to simply treat this as a mail message and therefore just
ensure lines remain lines.  Well.  I mean for -t and only -t we
could do this without too many hurts, adding one more flag to this
already-too-big input consuming code would not be that dramatical.
Should maybe not be more than a twenty line patch, all in all.

 |I will continue with my workaround.

Well if you use this MUA until at least v14.10 you really have to,
then.

Ciao.

--steffen
|
|Der Kragenbaer,                The moon bear,
|der holt sich munter           he cheerfully and one by one
|einen nach dem anderen runter  wa.ks himself off
|(By Robert Gernhardt)

Reply via email to