On Mon, 13 May 2013, RJ Clay wrote to mark lewis:

 RC>>>    I don't recall if I've brought this up again but are there 
 RC>>> echomail import limits that may not be explictly configurable? I 
 RC>>> ask because I noticed again (on my v0.92.0 system at 1:120/544) 
 RC>>> that although a rescan of an echo to my uplink resulted in a 
 RC>>> message noting that 400 messages were scanned but my system only 
 RC>>> ended up importing 47.  Need to figure out where that is...

 AL>> Is it possible the others were discarded as duplicates?

 RC>>     Shouldn't have been; I'd just added the echo to the 
 RC>> configuration...

 ml> that depends on the tosser's method of dupe detection...

 RC>    True but in this case none of the messages would be in any other
 RC> echo... 

understandable... but that doesn't mean that the MSGID hasn't been seen
before and it is possible that there may be a CRC16 or CRC32 collision if
the tosser uses such methods and then that also depends on what data in the
message they use for those calculations... some take only the header
information... some take the header information plus 40 bytes or so to try
to include some of the hidden control lines... they are hoping to catch the
MSGID line in there but if the line is too far away that that won't always
work...

duplicate detection is definitely an art and generating messages to pass
the various duplicate detection schemes can be even more of an art...
however, in some cases, the duplicate detection is very brain dead (eg:
header only or message body only) and no amount of crafting will get a
valid new message by them... sbbsecho is one that insists that a message
with an identicle body is a duplicate even thought the header plainly shows
that it is a new posting... DB is another one with a similar method of
duplicate detection...

)\/(ark

--- FMail/Win32 1.60