> Has anyone gotten the POSIXness mail command from the Dachstein
> pre-releases or release candidate to work?  I'm experiencing the
> problem Jeff Clark asked about in
>
>   http://www.geocrawler.com/archives/3/7325/2001/9/200/6589580/
>
> where the mail process never ends, doesn't send the mail, and
> displays the message "{pid}: User defined signal 1".
>
> I've verified that I can use netcat from my Dachstein box to
> manually send mail to the mail server I have listed in
> /etc/POSIXness.conf and I've also verified that the $fdata
> file is properly formed and that the right hostname and port
> are used in the call to the nc command.

I've verified your problems with the POSIXness mail script.

In addition to the mods you posted regarding reading the config file, I've
noticed the following:

- For some as yet unknown reason, the 'jobs' command in ash is not providing
information about the background processes, which causes the mainline script
to instantly exit, thinking the child processes are finished, when they are
acutally just getting started.  A quick hack to fix this is to replace the
test of the jobs command output with  a test for the childs PID.  Since the
ps command doesn't suport querying by pid, I just look in the /proc
filesystem for a $PID folder:

  while [ -e "/proc/$pidctrl" -o -e "/proc/$pidnc" ] ; do

In addition, there is apparently a race condition between launching the
control process and signaling it with USR1...apparently, the signal can
appear before the child has registered it's handler.  The ironic thing is
this signal is sent to prevent a different race condition...the control
process running too fast and reading from the recieve pipe too early.  The
solution is to go to a full two-way handshake between the main process and
the control process, which can be done using the existing USR1 signal sent
from ctrl to the main process to reset $count to 0.

The quick & dirty fix is to put a "sleep 1" line before the main thred
executes "kill -10 $pidctrl"

In addition, I noticed the attachment function seems to be broken.  This is
now fixed as well.

Fixes will be applied to the next release...

Charles Steinkuehler
http://lrp.steinkuehler.net
http://c0wz.steinkuehler.net (lrp.c0wz.com mirror)



_______________________________________________
Leaf-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/leaf-user

Reply via email to