It is confusing to think of ARv1 and ARv2 as separate databases. There are not. In fact, when looking at a database entry, there's no way to determine which CRC was computed using ARv1 and which using ARv2. You can only guess that it's a V2 CRC when it matches your local CRC that you know you computed using V2. And all offset-detection CRCs are computed using ARv1, but this doesn't affect anything. It doesn't matter how offsets are detected.
So the first part of this log compares your V1 CRCs against a database entry. And the second part of this log compares your V2 CRCs against the same database entry. Those comparisons are currently done independently. This log is a mirror image of the log you quoted earlier, where the first part said 'No match but offset' and the second part said 'Accurately ripped'.
Would it be more intuitive if CUETools would have merged those two parts of log into somethings like follows?
Track [ CRC , CRCv2 ] Status
05 [071cfb94,87f9de6b] (3,2/5) Accurately ripped
06 [cf6b6325,c373159d] (3,2/5) Accurately ripped
07 [b4f969a5,86ceab88] (4,0/4) Accurately ripped
08 [2c45475e,577e6a52] (4,0/4) Accurately ripped
And your previous log from a CD that was submitted only using ARv2 would look like this:
01 [d48a93b6,eeae9727] (0,1/1) Accurately ripped
02 [58465981,eddfa49f] (0,1/1) Accurately ripped