On 4/12/06, Professional Software Engineering
<PSE-L(_at_)mail(_dot_)professional(_dot_)org> wrote:
(w/ bcc to Timmothy, since you're probably not getting your procmail list
traffic right now...)
Actually the only part which was working was the first recipe which
forwards a copy of all my incoming mail to my GMail account... which
is actually how I realized that there was a problem in the first place
:-)
At 11:03 2006-04-12 -0400, TjL wrote:
My webhosting company (Dreamhost) seems to have absolutely no clue.
Are their heads in a cloud, or some less fluffy place?
Well I think at the very least that they aren't used to dealing with
the nuances of the Unix commandline.
For example I reported the error in 'formail' and part of the reply
back (from someone who was very helpful) asked how I was executing
'formmail' in procmail, so I explained the difference between formail
(which he wasn't familiar with) and formmail (which he was)
I reported this as a problem and removed my .forward-postfix
Can you post the content of that file please?
"|/usr/bin/procmail -t"
so mail would get delivered until they solved the problem... Some 8 hours
later they "tested" the problem by sending me an email, and when it
was received they said "OH, well the problem must be solved, you're
getting email now."
DUH! Of course it was received, but procmail/formail still isn't working!
Check the Received: headers on the test message amd those on prior archived
mail. I'm guessing they "upgraded" the mail software.
That's what I thought too, but it turns out to be something different:
The "cannot execute binary file" is not a message embedded within
procmail. It is possible it is the result of a system call to return an
error message associated with an error code (though I don't know if
procmail actually makes any such calls).
A greater possibility lies in security config changes with the
MTA. Postfix probably has a feature similar to Sendmail's smrsh (sendmail
restricted shell). Some guy probably started reading up on something and
went "oh, we should turn this on" without considering compatability issues.
Apparently there was an NFS issue which prevented the home directories
from being mounted, which is why part of the error message said
'cannot chdir to $HOME'
What is the easiest way to test for whether or not a directory exists,
and if it doesn't, send a message?
I'm thinking about something like
:0
* ? test -d $MAILDIR
| (echo "Cannot find $MAILDIR" | Mail -s No-Maildir me(_at_)gmail(_dot_)com
but I'm wondering if there's a lighter-weight way of doing it w/o
invoking 'test'
TjL
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail