[gw1500se]

I reported this to Apple as a bug quite some time ago. Although I have my doubts, they claim the reason they are not fixing it on Panther is because it is fixed on Tiger

As a side note (you may already have gotten this complaint) if the password expires after the user logs in and the screen saver is activated it will no longer unlock with the expired password (a bug fixed in Tiger?) and Applemail will no longer connect to the POP server (another bug fixed in Tiger?). This is driving our users crazy every 30 days. They have to then quit everything they are doing, log out, log back in, change their password, run keychain first aid and edit their Applemail password settings (because Kerberos authentication is broken on our POP server which is also supposedly fixed in Tiger). Effective single signon is only a pipe dream at this point (fixed in Tiger?).

[gw1500se]

Thanks for the suggestion. Unfortunately it didn’t help.

[gw1500se]

Thanks. This is in the user’s startup items list. It works for some users but not others. At one time it worked for all users but as time goes on it stops working for more and more users.

#!/bin/sh
#
# This script mounts the email archive volume for the user
#

user=`logname`
usertrunc=`echo $user|cut -c1-12`
if [ ! -d "/Volumes/$usertrunc on EMAIL-ARCHIVE" ] ; then
	mkdir "/Volumes/$usertrunc on EMAIL-ARCHIVE"
fi
if [ ! -f ~/Library/Mail/Mailboxes/Archive.mbox ] ; then
	ln -s "/Volumes/$usertrunc on EMAIL-ARCHIVE/Archive.mbox" ~/Library/Mail/Mailboxes/Archive.mbox
fi
mount_afp -o nobrowse "afp://$user:xxxxxxxx@email-archive/$usertrunc EMAIL-ARCHIVE" "/Volumes/$usertrunc on EMAIL-ARCHIVE"
if [ ! -d "/Volumes/$usertrunc on EMAIL-ARCHIVE/Archive.mbox" ] ; then
	mkdir "/Volumes/$usertrunc on EMAIL-ARCHIVE/Archive.mbox"
fi

[gw1500se]

Rats!!! You had my hopes up. I’m afraid it didn’t help.

[gw1500se]

Thanks. I know what you meant. I’m more a Unix guy then OS X anyway.

[gw1500se]

Thanks for the reply. If only it were that simple. There is no mounted volume with the same name (visible) that would cause the “-1” prefix to be added. There must be, either a hidden mount with that name that is auto-mounted even after a reboot (nearly impossible, IMO) or some peice of information kept somewhere (that can survive reboots) that is incorrect and/or corrupt.

[gw1500se]

I need to reserect this problem as now I have several users that have it. I do have a little more information that might be helpful.

First there is an error with the errno. The error is -5019. It would be a big help if someone can tell me where to look up these errors.

Second, I noticed that when I use the finder connect function to mount the same volume it is successful. However, the auto-created mount point has a ‘-1’ suffix on it. The only time I have seen that is when 2 volumes with the same name are used. The implication is that it thinks there is a volume with that name already there (that may be what the error -5019 is trying to tell me) and unlike the finder connect, I specify a specific volume in the afp_mount command.

Assuming that is true for now, then somewere there must be a file that tells afp what is mounted and that file must be corrupt or write protected. Does anyone know what/where that file is so I can fix it? TIA.

[gw1500se]

I finally got it working if anyone is interested. I had to install another mail reader because Apple mail would not let me see either the error or the message. The problem turned out to be the extremely fussy way headers need to be set up (to a fault, IMO). It also was compounded by the way PHP ‘mail()’ works.

The spacing of headers and ‘boundry’ lines is critial (i.e. there cannot be a blank line between the ‘boundry’ line and the ‘content-type’ line and there MUST be a blank line between the ‘content-type’ line and text of that part of the body.

The crux of the problem was that the headers need to be terminated with ‘\r\n’ which I already knew. However, when the ‘boundry’ and ‘content-type’ are part of the body, the ‘\r\n’ is apparently not recognized by the ‘mail()’ function and thus not properly translated. The solution was to put chr(13) and chr(10) at the end of those lines when in the body.

Sheesh!

[gw1500se]

Thanks, that worked. Unfortunately the man page says that option means use simple authentication instead of SASL. How that translates to NO authentication or anonymous is beyond me.

[gw1500se]

Thanks for the reply. The ‘dscl’ command was the key. I wish I knew how to find commands when I don’t know what to look for.

[gw1500se]

Thanks for the reply.

Yes, on the OD server.

[gw1500se]

Has anyone come up with a reliable way to implement this yet? Apple is still claiming there is no problem (with the instructions in the admin manual) but I sure can’t get anything to work and I’ve tried every methodology suggested. Apple disputes the claim that something is wrong with ‘getsslpassphrase’.

P.S. The last one from developer.apple.com doesn’t work either because:

1) The directory for for the Keychain does not exist.
2) Even if I create the directory, ‘certtool’ gets the error “Error getting keychain handle.”

[gw1500se]

Just to close this out, this is apparenlty a known Microsoft mail server bug and has nothing to do with Panther Server. In the event someone else runs into this here is the Microsoft suggested workaround one might want to send to the message originator:

http://support.microsoft.com/default.aspx?scid=http://support.microsoft.com:80/support/kb/articles/Q138/0/53.ASP&NoWebContent=1

[gw1500se]

Oops. I just realized you said client side. I don’t know what you mean now. It is definately NOT the client as I can see in my server log that the file is winmail.dat before the client even gets it.

[gw1500se]

Thanks. Any idea what I might say to their tech support people? That it singles out only certain users, not all, in the same domain is really odd.

[Author]

Posts