Hi,

On 2016-10-31 13:26:38, mark lewis wrote to Wilfred van Velzen:
  about: "DUPES!":

 WV>> Dupes have by definition the same dupe-checksum as the original, so
 WV>> storing them is useless!

 ml> maybe to you but to another developer who may be doing other things than
 ml> /just/ dupes, it may be needed or desired... don't be so square that your
 ml> mind is closed to other possibilities and thought processes... perhaps
 ml> someone is keeping count of duplicate MSGIDs and needing also to store the
 ml> CRC/hash for each of them...

If the CRC/hash is the same, because it's a dupe, you don't have to store it
again.

If you want to keep count of duplicate MSGID's you store those, or just a
count of them. That's got little to do with dupe hash's...

 ml> example: yesterday i saw two messages with exactly the same MSGID,
 ml> header, time stamps, and what looks to be the exact same message
 ml> body... the seenbys and paths were different yet HPT with its
 ml> HASH+MSGID dupe checking missed seeing the second one as a dupe...
 ml> both were back-to-back in my message base so it was easy to flip back
 ml> and forth between them to try to see any differences... golded+'s
 ml> [I]nfo showed that they were virtually identical but there looked to
 ml> be one space character immediately after the MSGID that was in one
 ml> post and not the other... now, get this, both MSGIDs, even though they
 ml> are the same, are in the dupe database but with different hashes...

I fail to see what's your point here. They weren't dupes according to hpt, so
the different crc/hash's were stored.

Bye, Wilfred.


--- FMail-W32 1.73.1.23-B20161031
 * Origin: Native IPv6 connectable node (2:280/464)