Request: Release Notes 20241224: add missing "Improve toString Performance"

[Simulant87]

Thank you for releasing a new version as Christmas present this year.

I noticed that in the release notes of version 20241224 my PR for Improving toString Performance is missing. It would be nice if you could add it to the release notes. Also this version is not yet marked as the "latest" version.

Speaking of it, I was a little bit confused that you rejected my other contribution to improve String parsing performance, as it is related to my first contribution mentioned above and makes the whole use case of the performance improvement complete improving the parsing and serialization performance. I also read from your one comment on the first PR that you first expected an improvement on the parsing, which I wanted to contribute with my second PR.

The comment from @anshchauhan shows that there is interest in performance improvements and I would be very happy if you would reconsider to accept my PR to improve the parsing performance.

Regards, and have a happy new year.

[stleary]

@rikkarth

my PR for #867 is missing.

Sorry, I missed this one. Release notes have been updated.

this [20241224] version is not yet marked as the "latest" version

It will be marked latest as soon as it shows up in Maven Central.

I was a little bit confused that you rejected my other contribution [#916]

PRs that only contain performance improvements are no longer being accepted. #867 probably should not have been allowed, either. I just updated the FAQ to make this a little more clear. In the cases of #867 and #916, I don't think it is a best practice to replace well-tested builtin classes for the purpose of improving performance.

[Simulant87]

Release notes have been updated.

Thank you for adding it.

PRs that only contain performance improvements are no longer being accepted. I just updated the FAQ to make this a little more clear.

I am sad to hear that, but thanks for stating this internal guideline more clearly in the FAQ. in my PR, the interface of the JSONTokener could also be broaden up to not only accept String, but the more general type of CharSequence, which is more commonly used in the Android Development. This would make it no longer a performance improvement only.

I don't think it is a best practice to replace well-tested builtin classes for the purpose of improving performance.

It might not be the best practice, considering that you want to maintain less code. On the other hand, I have also seen this approach in other libraries as well, e.g., in apache commons-io.

[stleary]

Reasonable people can disagree about best practices. I have updated the wiki with a better heuristic to use when deciding whether to submit a PR.

[stleary]

Closed due to release notes updated and release published.