> >Now what the error really means is that one of the objects that > >had the same connected value has 'disappeared'. disappeared > >could be any of various causes -a monster that was killed, or > >some other destroyable objects. > OK, now that's a bit freaky. You are saying the actual > button objects, etc. are being destroyed? That is what causes those error messages to be printed (ie, the object it added in to the buttonlist when the map was loaded does not match the object now there, or the object now there is a defunct object) > Putting everything together now, I'm *supposing* that what's > happening, is that when the affected maps are reset, at which > point the objects on it must be removed from memory (or have I > got that wrong?), the bit that calls update_button doesn't > realise they've gone (or maybe it even thinks the buttons are on > a different map, I dunno the CF internals). Are you running CVS or 0.98.0? There was a bug in 0.98.0 that resulted in something similar to what youu desribe - when maps with in memory were reset, the freeing of the data resulted in the above situation. That is fixed in CVS. > [BTW- I prolly should have mentioned earlier, yes I'm sure I've > seen the "update_item for item we don't have" error in cfclient, > but I don't remember them crashing it at all. And not anywhere near > titans, either :P] the update_item is a sort of harmless error message. Sort of harmless in the sense the client will just ignore it. Not completely harmless in that it does take up some bandwidth to send data we will discard. The fix for the bug (levers/other object player is standing on not getting the face updated) is not completely satisfactory. It does work, with the cost that we send some false updates to the clients.