> I bet. UTF-8 support -- for writing -- was added more or less in    
   > the last minute, before Johan disappeared from fidonet and left us all   
   > alone with his babies.   
      
    SS> Is that to say, that JamNNTPd needs a maintainer?   
      
   i cannot speak to that but if someone is rooting around in the code and can   
   provide easy patches to those of us who compile our own, well ;)   
      
   speaking of which, my system has numerous background processes that access my   
   JAMbases at any time... someone needs to look in jamnntpd and see where it   
   emits these and fix the problem...   
      
      Could not read messagebase header of    
      
   the JAM specs say to lock a specific byte in the header file but i seem to   
   recall that at least one library (MKMSG written in PASCAL) doesn't lock that   
   byte but another... AIR, the thought with this library is that byte/region   
   level locking may not be available but you can't get a lock on a file if any   
   part of the file is locked...   
      
   [time passes]   
      
   after digging for a while, i finally found some original JAM sources from the   
   JAM gods... it seems that the first byte is the lock byte after all... unless   
   i'm counting incorrectly and it is the second byte...   
      
   in any case, this is a bug and there needs to be some sort of timing loop that   
   retries the lock when it is needed or at least retries access if there is a   
   lock preventing reading... i would say, without looking at the nntp RFC, that   
   this lock needs to be at least 30 to 60 seconds in length... i'm guessing on   
   those time lengths based on how long it make take for comms to be seen as   
   failed over the 'net... FTP, for example, has something like a 5 or 10 minute   
   timeout...   
      
   i am not using the SMAPI stuff but the original stuff... there were things in   
   SMAPI that i didn't like way back and i've never used it with my JAMbases...   
   somethings are just enough slightly different to cause problems from expected   
   actions :?   
      
   )\/(ark   
      
    * Origin:  (1:3634/12)