Michael pointed out over at [1] that the new tiny.json is pretty
inscrutable given its size, and I have to agree. Attached is a patch
to pare it down 98% or so. I think people wanting to run the
performance comparisons will need to come up with their own gigantic
files.
Michael, with your "Jacob might be a nefarious cabal of
state-sponsored hackers" hat on, is this observable enough, or do we
need to get it smaller? I was thinking we may want to replace the URLs
with stuff that doesn't link randomly around the Internet. Delicious
in its original form is long gone.
Thanks,
--Jacob
[1] https://www.postgresql.org/message-id/ZhCTtpGoXbERacoN@paquier.xyz