mckay@ECN.PURDUE.EDU (Dwight D. McKay) (11/17/90)
I'm trying to figure out how one gets mail fed into messages. I understand that one method is to run "eatmail". How? I've yet to find documentation describing how you do this. Also, is there a collected set of hints and how-to's for andrew and in particular messages? I'd like to do things like switch off the multi-media stuff for selected recipients, automatically removed really old mail from folders, etc, etc. Some hints and examples would really help! --Dwight D. McKay, ECN Workstation Software Support --Phone: 43561, Office: MSEE 104f --Office hours: Mon, Tues, Wed, Fri 8-11am; Thu 8-10am
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (11/20/90)
In general, mail gets automatically taken from /usr/spool/mail/xxx, and burst into separate files for each message. The separate files are stored in ~/Mailbox. From there, they are incorporated into your mail folders -- by default, they go into your "mali" folder, but a FLAMES program (.AMS.flames) can direct them elsewhere or respond to them automatically.) The "eatmail" program can be used to burst aport /usr/spool/mail format files into separate files as in a Mailbox directory. This is not needed for normal mail delivery, but the separate eatmail program is useful in certain circumstances (e.g. when your spool file is on another machine, and you just want to rsh a small program). In general, however, there are lots of ways to get mail fed into messages; you'll have to ask a more specific question to get more information, I'm afraid. Much of what you're looking for is probably buried in the help files. For example, to switch off multimedia for selected recipients, you can put certain lines in your .AMS_aliases file, e.g.: $forceformat Bill Cattey <wdc@athena.mit.edu> $forceformat wdc@athena.mit.edu $forcetrust @andrew.cmu.edu $forcetrust @andrew.cmu.edu> Type "help ms-aliases" for more information. For purging old messages, check out the CUI "epoch" command, in the CUI documentation. For other specific questions, if you can't find it in the documentation, feel free to ask... -- Nathaniel
howard@THUMPER.BELLCORE.COM (Howard Bussey) (11/20/90)
I run Nathaniel's eatmail program in a particular context: My mail is delivered to a general-purpose (and oftimes overloaded) machine, but I read mail from a variety of workstations. To do this, I use eatmail to grab mail from the un*x mailbox and move it to ~/Mailbox. eatmail is run from the per-use crontab available (at least) on SunOS 4.0.x. My crontab includes the following entry (but you can use at jobs or any other method to run eatmail periodically): 1,6,11,16,21,26,31,36,41,46,51,56 * * * * eatmail -l/u/howard/.EatMailLog The -l/path/to/file parameter is to make eatmail log something when it grabs some mail. (If it runs but my mailbox (/usr/spool/mail/howard for me) is empty, eatmail doesn't change my .EatMailLog file). I make xbiff look at my mail log via the following line in my .xinitrc: xbiff -file /u/howard/.EatMailLog -geometry -1+0 & Finally, I arrange to have mail collection empty .EatMailLog via the following line in my preferences file: *.PersonalMailCollectionCommand: cat /dev/null >/u/howard/.EatMailLog There may be a race condition between eatmail trying to append a line to .EatMailLog and the cat command run by messages/cui/vui trying to empty .EatMailLog, but as this is just a log file, an error either way is self-correcting and results in either (1) I get mail late; or (2) xbiff claims I got mail when I have already examined all new mail. I haven't run into either situation, so I don't think the potential race condition is terribly important. -------------------------------------------------------------- Howard Bussey <howard@thumper.bellcore.com> Bellcore Room 2P-288/445 South Street/Morristown NJ 07962-1910 voice: +201 829 44 79; fax: +201 984 22 83
Craig_Everhart@TRANSARC.COM (11/20/90)
Excerpts from internet.info-andrew: 16-Nov-90 messages, eatmail and other.. Dwight D. McKay@ecn.purd (606) > I'm trying to figure out how one gets mail fed into messages. I understand > that one method is to run "eatmail". How? I've yet to find documentation > describing how you do this. ``eatmail'' is only one method. Messages should find your mail in /usr/spool/mail/mckay just fine by itself. If it's not doing that, you probably need an AndrewSetup file telling it how to do so. The DESTDIR/help/setup.help file should explain this. > Also, is there a collected set of hints and how-to's for andrew and in > particular messages? I'd like to do things like switch off the multi-media > stuff for selected recipients, automatically removed really old mail from > folders, etc, etc. Some hints and examples would really help! Messages (acting alone) can strip off formatting on a per-message granularity; either a message is sent formatted or unformatted, but the same way to all recipients. Your ~/.AMS_aliases file can contain indications of what kinds of formatting a given recipient needs; I think that the help file for this is DESTDIR/help/msalses.help. As for removing old mail from folders (presumably public folders), you can use the ``epoch'' command in CUI for this. Check the CUI help files. Craig
datri@convex.com (Anthony A. Datri) (11/21/90)
>I run Nathaniel's eatmail program in a particular context: My mail is >delivered to a general-purpose (and oftimes overloaded) machine, Eatmail is indeed a fine program, but it does require ATK on the spool machine, which isn't available in my case. So, at first I wrote an eatmail-inspired program that did the same thing, learning all kinds of disturbing things about mkstemp in the process. This I ran by setting the AMSMailCollectionCommand preference to rsh it, but I found this to be slow (because of the rsh). I got a better idea (for me, at least) from someone's post on another group -- I wrote a [really simple] program that gets run as a pipe out of my .forward. It generates a filename in ~/Mailbox by concatenating its process-id and the time -- in my case, this should be enough to ensure unqiueness, since we have site-wide aliases with an entry for each user pointing to his/her home machine. -- datri@convex.com
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (11/26/90)
Why does eatmail require ATK on the spool machine? I think that's totally incorrect, actually. It requires a few libraries (libmail, liberrors, and libutil) but these can all be built without building ATK at all. -- Nathaniel
datri@convex.com (Anthony A. Datri) (11/28/90)
>Why does eatmail require ATK on the spool machine? I think that's >totally incorrect, actually. It requires a few libraries (libmail, >liberrors, and libutil) but these can all be built without building ATK >at all. -- Nathaniel I was considering the above to be part of ATK; forgive me if that wasn't correct. Can the above be built without class or any dynamic-linking code? --
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (11/29/90)
Excerpts from internet.info-andrew: 27-Nov-90 Re: messages, eatmail and o.. Anthony A. Datri@ucsd.ed (388) > >Why does eatmail require ATK on the spool machine? I think that's > >totally incorrect, actually. It requires a few libraries (libmail, > >liberrors, and libutil) but these can all be built without building ATK > >at all. -- Nathaniel > I was considering the above to be part of ATK; forgive me if that wasn't > correct. Can the above be built without class or any dynamic-linking code? Yes indeed. I believe all you need to build are the overhead/{errors,mail,util} trees. You may need to build config, first, actually. But certainly no dynamic-linking stuff should be necessary.
Craig_Everhart@TRANSARC.COM (11/29/90)
It's pretty simple to make the basic support from Andrew, if you want to omit ATK. (That's what we did to get AMDS running at Transarc at first.) You can omit these directories under your andrew distribution: overhead/index overhead/class overhead/conv overhead/fonts overhead/xicons overhead/cmenu atk atkams contrib (oda, odatrans) xmkfontd helpindex helpaliases There could, in principle, be an ATK_ENV option in the top-level system.h that controlled these directories. We built the rest of Andrew (Andrew minus ATK) here at Transarc on all platforms. I built ATK also for at least some of the platforms; for a long time the pmaxes were problematic, but are getting better. Craig