On 08 Jan 17  09:16:24, Mark Lewis said the following to Nick Andre:   
      
   ML> exactly sure on the details but i know that they greatly influenced me to    
   ML> my software place the MSGID as close to the beginning of the control lines    
   ML> possible so that db would not detect messages posted within one second as   
   ML> dupes... this was especially important when testing at 100+ posts per   
   ML> second... are you willing to share information about how db does its dupe   
   ML> detection so others can understand more? please?   
      
   Its not that hard to understand. A CRC is computed from the header and date of   
   the tossed message. I would have to dig into the code and I'm not sure how    
   many bytes are being included from the start of the message.   
      
   Each Echomail area has a cache database file. In the case of *.MSG, this is   
   called DBRIDGE.DUP and resides in each area and for Hudson/QBBS there is    
   one database segmented slightly different. The CRC's are kept in there. I    
   believe the code sets the cache database size at 1,024 entries.   
      
   Interestingly it appears that there is a "reputation" method for the cache    
   database. It appears as it is loaded into RAM during a toss, any time a CRC   
   match is encountered, that CRC is pushed up the cache table, while CRC's of   
   legitimate messages end up being pushed down. The CRC table is saved into that   
   cache file every time the Echomail area changes in the toss cycle; or there    
   are no more packets to toss.   
      
   Nick   
      
   --- Renegade vY2Ka2   
    * Origin: Joey, do you like movies about gladiators? (1:229/426)