Re: improve performance of pg_dump with many sequences

I ran Euler's tests again on the v6 patch set.

    for i in `seq 1 10000`; do psql postgres -c "CREATE SEQUENCE s$i;"; done
    time pg_dump -f - -s -d postgres > /dev/null

    HEAD:        0.607s
    0001 + 0002: 0.094s
    all patches: 0.094s

Barring additional feedback, I am planning to commit these patches early
next week.

-- 
nathan