Request to contribute again – small indexing perf fix (DocumentParser)

Hi Elastic team,

Last August I opened a burst of unsolicited build/tooling PRs (#131427–#133044) that weren't wanted. Sorry about that; I understand why they were closed.

I've since prepared one small, focused fix and would like to contribute it if possible. It looks like my GitHub account (Pankraz76) can no longer open PRs in elastic/elasticsearch.

The fix: elastic/elasticsearch#154503 removed a canAddIgnoredField() short-circuit, so DocumentParser#parseArrayElements and FallbackPostMapper#parseField now do pre-capture work for every array and field even when _ignored_source can't be written (e.g. stored _source). Restoring the guard is behaviour-neutral. In a JMH benchmark with 50 two-element arrays it cuts 17,710 ± 1,079 → 15,765 ± 635 ns/op and 12,776 → 9,976 B/op per document.

Branch: GitHub - Pankraz76/elasticsearch_frk at perf/skip-precapture-without-ignored-source · GitHub

Would a maintainer be willing to review it, or pick it up directly if contributions from my account aren't possible? I've signed (or will sign) the CLA.

Thanks!

Hey thanks that sounds like an interesting idea. I've sent a request to the Github admins to remove whatever block was imposed. It might take a few days for them to react, bear with us.

Hi David,

Thanks again for passing this on to the GitHub admins. I really appreciate it.

Just a quick heads-up: I gave it another try today and still can’t open a PR against elastic/elasticsearch from my account (Pankraz76). No rush at all, I know these things take time. I just wanted to let you know in case the request got lost along the way.

The branch is ready whenever it’s possible, and if it’s easier on your side, I’m also happy for someone to pick up the change directly:

Thanks for your help!

Vincent

Still chasing, sorry it's being so slow.

any chance you have a timeline on when the block might be lifted? would help plan next steps

Would you try again now @Pankraz76 ? I believe the block has been lifted.