It seems the "mailx" package in main is actually Berkeley Mail, with sources
downloaded from the Arch mirror.
This actually does not provide mailx, but only mail.
Further, its compatability with POSIX mailx is not complete (mail
supports -iInNv, -b -c -f -s -u but not POSIX -HFe)
The missing options are as follows:
-e: scripting (check for mail)
-F: save a copy of the mail in a file named for the first recipient
-H: write header summaries for mails and exit
Beyond these missing options, it does not support MIME (which means no
attachments and no UTF-8) or maildir archives.
I'm not going to claim that it's unmaintained upstream since OpenBSD is
maintaining it (although we aren't tracking OpenBSD).
But there are a few implementations of mailx that address some of these
deficiencies, besides adding features:
heirloom-mailx - based on Berkeley Mail 8.1, but extensively modified.
Formerly known as 'nail'.
This adds support for POP3/IMAP/SMTP, with or without SSL (openssl and nss
are supported), attachments, MIME, multiple accounts, and a few other
The latest source (12.4 + patch to CVS HEAD) compiles cleanly.
S-nail - fork of heirloom-mailx, with ambitious plans involving code
refactoring. Currently supports all the features of heirloom mailx and
some UI enhancements.
Has a lot of releases (two last month, at least two the month before that)
Upstream expects to temporarily remove the IMAP support during refactoring.
mailutils - supports -eEfFHimnNpqrstu but not -bcIv
This is the GNU implementation; it ships a "mail" binary expecting
mailx to be a symlink to that.
There are also numerous other mail utilities.
This is in unmaintained/, and I'm not interested in adopting it; it
also doesn't support attachments.
I'm inclined to think that in the short term, heirloom-mailx would be
a significant improvement over our current main/mailx.
What I'd like to do is add testing/heirloom-mailx providing mailx;
I have a preliminary aport without the provides part.
In the future, upgrading to s-nail seems like a logical path; but
I think that using s-nail now might be hasty.
Received on Sat Sep 13 2014 - 13:33:17 UTC