mhonarc-users

Re: one message instead of multiple - why?

1999-11-05 17:34:48
Andrzej Kasperowicz said:

I (Simeon Nevel) said:

The latest versions of Pegasus *can* store message "folders" in mbox
format if you tell it to.

How...?

I can see from your message headers:

  X-mailer: Pegasus Mail for Win32 (v3.12a)

that you're using the same version of Pegasus that I am.

Do File -> Mail Folders...

Click the "New" button of the Folders dialog box.

Fill in a name for the new folder and select "Unix Mailbox Format" from 
the drop-down list next to the "Message Folder" radio button (make sure 
that "Message Folder" not "Tray" is selected).  

Voila! you've got an mbox format Pegasus folder you can save messages to. 
You should be able to feed that file directly into MHonArc without any 
problems. 

I'm just not entirely sure if Pegasus keeps the "From " line when you do 
this.  

If not, never mind or play with the msgsep resource as needed. I don't 
know how Pegasus delimits messages it stores in mbox files (although I 
would *assume* if follows the relevant RFCs).  You can simply try it and
take a look at the resulting file.  

You can also "add" an external mbox file to your folder list by using the 
"Add mailbox to list" option of the Folders menu that is visible when your 
folders are being displayed  

This assumes, of course, that the mbox file you want to "add" is actually 
accessible from the machine where pegasus is running.   

FYI, messages in your "New mail folder" are actually individual files 
(ending in .cnm) that can almost certainly be fed into mhonarc one at a 
time.  


Apparently every message but the first is preceded by:

"-- End --"

If you could insert that string at the beginning of the file (to
properly delimit the first message), you *should* be able to use
the MSGSEP resource:

  ^-- End --$

I tried before ^-- End --
and Mhonarc didn't like the space after -- 

Which "--", the first one or the last one? What makes your think MHonArc 
"did't like the space"?  What was the symptom?  

What is that $ for? Does it help with the spaces?

The trailing "$" anchors the regexp to the end of the line (actully the 
newline char at the end of the line).  To handle any number (including 
zero) of trailing spaces, use:  

  ^-- End --\s*$

Unfortunatelly, things are more complicated than that, because the next day
when some new messages arrvive and are appended (by means of the command
added to subscribers list:

|/bin/cat >> /home/andyk/public_html/chemfan/archive/chemfan.mails.inbox 

to this file with pegasus stored past mails:

http://www.termisoc.org/~andyk/chemfan/archive/chemfan.mails.inbox

Mhonarc (run by cron without msgsep, since it seems not necessary to do it
for new incoming messages) creates new threads.html, and mailist.html
http://www.termisoc.org/~andyk/chemfan/archive/threads.html without all
those mails created yesterday with use of -msgsep. 

The same happens when I delete content of the file chemfan.mails.inbox (I
would prefer to delete it), after mhonarc run, so it seems that content of
this file cannot be deleted. What to do about those things?

Is the cron job invoking MHonArc with the "-add" parameter?  If not, 
you're simply rebuilding the archive each time from the current input.

see the documentation:

  http://www.oac.uci.edu/indiv/ehood/MHonArc/doc/quickstart.html#adding

PS. I also would prefer if mhonarc could run immediately after new message
using this command in subscribers file |/bin/cat >>
/home/andyk/public_html/chemfan/archive/chemfan.mails.inbox arrives. How
could I do that?

Rather than cat'ing incoming messages to an mbox file, you need to pipe
them into MHonArc.

Again, see the docs:

  http://www.oac.uci.edu/indiv/ehood/MHonArc/doc/faq/usage.html#stdin and
   http://www.oac.uci.edu/indiv/ehood/MHonArc/doc/faq/archives.html#dynamic

This may (or may not) actually be a "good" thing to do...

When I first started using MHonArc, I was archiving a very active mailing 
list (100+ messages a day) and invoking MHonArc for each incoming message
via procmail.

The sysadmin at my ISP made me stop 'cause:

1) As nice a program as MHonArc is, it *is* a bit of a resource hog.
   I was bogging down the mail server.

2) As pointed out in the FAQs, as your archive gets bigger, processing gets
   slower.  MHonArc *does* handle concurrent invocations with some sort of
   locking mechanism, but eventually, subsequent MHonArc processes will
   time out waiting for previous invocations to finish. 

3) An active mailling list can spawn *many* MHonArc processes, all
   serialized, exacerbating  1).

I simply use procmail to save all the list mail into an mbox file and then
periodically run a little shell script to feed the resulting file into
MHonArc (I can't do cron at my ISP).

It sounds to me like you've got a Pegasus folder full of old mail which
you want to html-ize and then you want to start adding the new messages.

You want to handle those groups of messages *separately*

You want to:

1) Rebuild the archive with the saved Pegasus messages (which I assume
   you've boosted up to your Unix host) using MHonArc *without* the -add
   parameter. You do this only once.

2) Then you want to feed new messages to MHonArc *with* the -add
   parameter. 

There are *many* ways to accomplish 2

You can continue to use the "|/bin/cat >>" method to append the incoming 
messages to an mbox file that you periodically feed to MHonArc with either
a cron job or a shell script invoked at the command line. Don't forget the
-add parm.

Or you can add messages as they arrive (subject to the caveats mentioned 
earlier) by invoking MHonArc with the -add parm from a sendmail alias or
from a .forward file or from a procmail recipe.  Which method you chose
depends on how much control you can exert on your system.  All of them
should work.

One final thought... 

Once you've added the old pegasus messages, you *might* have to respecify 
the msgsep resource the first time your run MHonArc again.  

Otherwise the msgsep resource will have been stored in the MHonArc db file 
and will be re-used when you start adding new messages that *don't* have 
the msgsep you defined imbedded in the file.  

If you decide to add single messages as they arrive, this may not matter. 

Earl?  

I hope this helps...  If you explain more about exactly what you're doing 
and what you're trying to accomplish, the list and I can try to expand on 
the answers.  

Simeon

P.S. Please respond either directly to me *or* to the list.  Getting 2
     copies of your response just confuses me.
--
snevel(_at_)sonic(_dot_)net      =-=-=-=-=   A rabbit is just an angel with big 
ears!
Send mail to snevel(_at_)sonic(_dot_)net with Subject: send index to get a 
listing of
files available from my e-mail robot.  <*>