On Thu, 2 Jul 2026 at 06:15, Kevin Liu <[email protected]> wrote:
> We've discussed this issue during the parquet sync today. > > To summarize: > - We agreed that this is a bug and should be fixed. I'll put up a PR for > it. > - I proposed to backport the fix and create patch releases so older > parquet-java versions can be fixed without being forced to upgrade to newer > versions. Having a java 8 version would be good As well as Russel's work on .sh release scripts, ASF relaese automator is designed to do a lot of this, though it seems to leave out maven artifact promotion, which is a manual stage. You get to set up and suffer with the still-stabilising asf mfa setup here. This is all a WiP, expect discussion at the community over code event in october if you go for the whole week https://release-test.apache.org can I flag that my variant test data PR now includes a variant(2) declaration Every parquet lib that actually claims to parse variant(1) MUST reject new declaration as unsupported, at least if my reading is correct. https://github.com/apache/parquet-testing/pull/113 Is this right? *What SHALL a variant-aware parquet lib do if it encounters a version of variants it doesn't know about? It' can't parse the metadata after all, can it?* > >
