PE>> The problem happens once a month with Bob with no seconds fixup.

RS> Yep, no argument that the seconds fixup is essential. The argument
RS> was tho about whether there is any practical difference between a
RS> 1 sec and a 2 sec increment. There isnt. And whether a real MSGID
RS> should be there TOO, it should.

PE> Right we'll stick with this, and I'll ignore all the crap about these
PE> imaginary systems which dupe-check on MSGID alone.

Are you seriously suggesting that there isnt anything else used world
wide except Squish 1.10 ?  Pull the other one.

PE> For practical difference, I don't know what the statistics are.

Its a complete wank IMO. And any design considerations have to take
account of whats used, even if its only on 2% of systems for something
like this.

PE> But basically, at the moment, if you post 38 messages, then the last
PE> 8 will have a "seconds" field exactly the same sas the first 8.

Yes, and its utterly unlikely that thats a problem because the CRUCIAL
thing is thats ONLY a problem if two of the messages in a pair are
edited in the SAME minute. Thats very very unlikely. And even when
that does indeed happen very occasionally, you STILL dont have a
problem UNLESS all the other fields like To From Subject match too.
Its fair to assume that the Date will match with daily mail runs.

PE> Then the problem of "did you edit them in the same minute" comes
PE> into play.

And is pretty damned rare. Even with my high reading and writing rate it
doesnt normally happen, tho you can sometimes get it with proof reading
the replys.

PE> I don't know what the chances are of you getting a dupe this way.

Sweet fuck all basically once you have a 2 second fixup.

PE> Certainly less than having 30 with the seconds fields all 0.

A hell of a lot less in fact. Its quite uncommon to have two messages
with the same time within a minute in one mail bundle.

PE> Anyway, what the new version does is set the seconds from 0 to 58
PE> for the first 30 messages, and then 1 to 59 for the second 30.

Yeah, I know that and my point all along has been that its really not
going to make any difference.

PE> This will guarantee safe passage past a Squish 1.10 system.
PE> And it will have the same problem as described in the previous
PE> paragraph getting past a Squish 1.01 system.  Nothing I can do about
PE> that without changing the minutes etc by looking at the others etc.

Yes, but in practice it doesnt actually achieve anything worthwhile
either and a real MSGID would be much better when the Squish 1.10
abortion is binned.

--- PQWK202