Re: [nail-devel] Announcing S-nail 14.3.1

[Martin N. <ne...@ga...>]

A mixed bag of replies the for the two/three recent commits:

(1) I have merge conficts with cc-test.sh, without any clue why this
    should be so.  (s-nail still builds, so it's not fatal.)

    I am on the "master" branch and didn't touch cc-test.sh in any way.
    The previous pull (just a few days ago) went flawlessly.

(2) While looking at cc-test.sh:

    Putting any kind of non-ASCII-characters into a source file is
    a bad idea.  These days, programs such as vim and less have become
    so electric that they try to guess and automatically "correct"/mutilate
    encodings on the fly, and often users aren't even aware of that.

    So, instead of using 8-bit characters in the source, I recommend
    to use circumlocutions by strict 7-bit ASCII chars.  For example:

	#!/bin/sh
	aeLATIN1=`printf '\344'`
	aeUTF8=`printf '\303\244'`
	... echo "gef${aeUTF8}hrlich" | s-nail ...

     That may appear as not so pretty but is a MUCH safer way for
     maintaining source.  And it IS much prettier for any developer who
     happens to use a character encoding different from you.

     Last not least:  national language comments such as
	"Die letzte Zeile war nur ein Punkt.\r\n"\
     are very nice for the Huns but what about the rest of
     us developers?

(3)  > - WANT_JUNK and WANT_SCORE have been disabled by default.

And, by the way, I missed them immediately ;-)

>   They don't seem to be too useful; i hope i can implement
>   a SpamAssassin hook for (downloaded) mail messages for v14.4.
>   If so, expect these two "modules" to become removed completely.

While I'm expressing my mere personal preference for keeping the nail
classifiers, I might as well mention my willingness to try my best
in the following areas:

(4) Trying to coax more diagnostics from the openssl layer.
    (For example, a simple "verified OK" is pretty meaningless as long
    as the certificate subject is not even mentioned.)

(5) Trying to support PGP-mails in addition to X.509-SMIME.

(6) Trying to re-wrap text/plain msgs marked as "format=flowed".

For tackling (6), it would be most helpful to have some README giving
bird's eye view how messages are processed (ie. decoded, filtered,
paged and whatnot else), and which knobs control which stage.
Such an overview would be very helpful for me for either the old
or your upcoming, revised system.

(In fact, it would be very help to have such an overwiew in the man
page.  It certainly took me quite some time to gather all required
switches whenever I needed to get something done.)

				A nice weekend everybody, Martin