asterisk users - Sep 2004 - UK Caller ID - todays CVS update knocks out a channel

I've updated to the latest CVS as of today (and rebuilt RedHat 9).

My setup is as follows:

Wildcard X101P - channel 1
TDM400P - channels 2-3 - fxs cards with fxo signalling,  channels 4-5 - fxo
cards with fxs signalling

I got CallerID to work on channel 4 with an old CVS, despite the usual
"Didn't finish CallerID spill" message.

However, as soon as I insert the following commands with my new build channel 5
"recognises" incoming calls but does nothing:

cidsignalling=v23
cidstart=polarity

If I comment them out, no problem. It works, but not CID.

I get the following log entries:

*CLI>   == Starting post polarity CID detection on channel 5
    -- Starting simple switch on 'Zap/5-1'
Sep 30 19:04:29 NOTICE[1111234112]: chan_zap.c:5161 ss_thread: Got event 2
(Ring/Answered)...
Sep 30 19:04:32 NOTICE[1111234112]: chan_zap.c:5161 ss_thread: Got event 2
(Ring/Answered)...
Sep 30 19:04:35 NOTICE[1111234112]: chan_zap.c:5161 ss_thread: Got event 2
(Ring/Answered)...
Sep 30 19:04:38 NOTICE[1111234112]: chan_zap.c:5161 ss_thread: Got event 2
(Ring/Answered)...
Sep 30 19:04:39 NOTICE[1111234112]: chan_zap.c:5161 ss_thread: Got event 17
(Polarity Reversal)...
Sep 30 19:07:52 NOTICE[1111234112]: chan_zap.c:5161 ss_thread: Got event 2
(Ring/Answered)...


Eventually it answers and you get one high pitched screech and the line goes
dead.

On channel 4 it works fine with cidsignalling=v23 and cidstart=polarity enabled.

Is this an oddity to do with my hardware layout?  I seem to recall a previous
posting which suggested that the TDM400P card doesn't like an fxo
daughterboard in the first physical slot.  Did I recall that correctly?

Any suggestions would be most welcome.

Many thanks,
George