Stephen R. van den Berg (AKA BuGless) wrote,
[snipped; ">" is Stephen; more quote marks are others]
| If the digest generator is MIME compatible, it would contain
| Content-Length fields (e.g. a SmartList managed mailinglist).
-d Tell formail that the messages it is supposed to
mailbox formats). This disables recognition of the
Content-Length: field. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^
So if a digest article contains another piece of mail with unindented
headers, a Content-Length: header at the beginning of the article won't
help.
Indeed. Forgot about that. That's why there are man pages :-).
The only recourse would be to teach formail a bit more of
MIME (or use a real MIME-tool to split the digests).
Is "Content-Length:" a MIME header only? I'm just learning and
exploring the maze of RFC's, but haven't found the one defining that
one yet. Didn't seem to me that this is a MIME issue... none
of the posts (OK, rarely) to that digest are MIME. It's just a
straightforward "sucked up this article and included it here"
problem that I wrote about in the first place.
The digest I wrote about seems to me to be RFC 1153 compatible.
I somehow think there should be a digest RFC requirement that *if*
an article body contains "^From" then either it be escaped with ">"
or a "Content-Length" header added to the message. It wouldn't
make the current formail able to handle this, but it would make
it *possible* for a future one to do so. I'm not sure the current
standards actually permit an unambiguous splitting of digests.
Could get *really* interesting if one wanted support for a
"digest of digests" to be splittable into either separate digest
messages, or individual posters' messages. But if the standard
were written to handle *that*, we'd have a lot less trouble with
the mundane inclusions (IMHO).
On a sort of related issue...
A couple of months ago (and for the life of me, I can't now remember
why!) I came up with a case where I would have liked formail to
generate a Content-Length: header for me. An option to do that
might be nice, although you could also work it off the -a/-A switches
and fill in the value if empty. Just a suggestion.
Or would piping the mail body to "wc -c" suffice? I know blank lines
get added here and there, so the "wc" approach might be off.
Might make more sense if I could remember *why* I wanted to do that,
but the memory is the second thing to go. (What's first? I can't
remember.) (Apologies to McGyver.)
Cheers,
Stan.