Some part of bug from 2.3.4 still in 2.3.5

[Hajimu UMEMOTO]

Hi,

>>>>> On Thu, 25 May 2006 18:00:00 +1000
>>>>> "Bron Gondwana" <brong at fastmail.fm> said:

brong> ... it appears reconstruct just blindly re-applied the value of
brong> modseq from the already existing entry without checking if it
brong> was valid.

brong> The attached patch contains my previous patch to fix the
brong> append_copy code path along with an additional check in
brong> reconstruct which will set modseq to 1 if it doesn't 
brong> already have a value.  Since modseq always starts at 1 and
brong> is increased each operation, I think that's OK for people
brong> running in the CONDSTORE universe as well.  It's certainly
brong> fine for those of us where modseq will always be 1.

I met this problem, too, and your patch solved the problem.  Thank
you!

Sincerely,

--
Hajimu UMEMOTO @ Internet Mutual Aid Society Yokohama, Japan
ume at mahoroba.org  ume@{,jp.}FreeBSD.org
http://www.imasy.org/~ume/