On Tue, 2 Sep 2025 14:00:52 -0400
Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Jim Jones <jim.jones@uni-muenster.de> writes:
> > LGTM; I'll mark the CF entry as Ready for Committer.
>
> Pushed with some trivial cosmetic adjustments, including the
> perhaps-not-so-trivial fix of removing the comment you falsified.
Thank you for committing the patch and for fixing it.
> I was concerned about Rahila's upthread worry about the performance
> of this approach, but in some quick testing it seemed to add only
> barely-noticeable overhead even at 1000 enum labels. At 10000
> labels it's slightly annoying: my machine goes from ~80ms to ~250ms.
> But that seems well beyond what anybody would be likely to use,
> so I judge it not worth trying to be smarter.
>
> The obvious solution if we did wish to avoid the O(N^2) behavior would
> be to qsort the labels and then compare only adjacent ones. That'd
> require a temporary array though, and I'd bet it's actually slower
> than this way for normal-sized enums. Another possibility perhaps is
> to apply the check only when there are fewer than say 1000 labels,
> reasoning that anything bigger is probably machine-generated anyhow.
I also thought the O(N^2) behavior was acceptable, since it seemed unlikely
that users would try to create such a large number of enum labels.
If any complaints arise in the future, we can address them then.
Regards,
Yugo Nagata
--
Yugo Nagata <nagata@sraoss.co.jp>
