(original) (raw)

\[Tim\]
\>> I'm keen to get feedback on this before merging the PR, because this
\>> case is so very much larger than anything I've ever tried that I'm
\>> wary that there may be more than one "surprise" lurking here. ...

\[Inada Naoki <songofacandy@gmail.com>\]
\> I started r5a.4xlarge EC2 instance and started arena.py.
\> I will post the result in next 12 hours.

Thank you! Wrapping this up, Inada attached the results to the bug
report. For the OP's original big case, the time to reclaim the tree
storage dropped from nearly 5 hours to about 70 seconds. The time to
build the tree didn't materially change (it was a bit faster after the
patch, but offhand didn't look significantly faster to me).

Since I called this a "design flaw" rather than "a bug", I'm not
inclined to backport it. If someone else thinks it should be, that's
up to them. I'm \_assuming\_ Github won't decide to backport it on its
own - but maybe I'm wrong (Github workflow is still pretty magical to
me).