RyanSkraba commented on PR #2529: URL: https://github.com/apache/avro/pull/2529#issuecomment-3587130391
Huh, I'm trying to revisit this PR to review the regression (it looks like the HUDI one is the same). I can't wrap my head around this PR, and I can't seem to figure out what is being fixed! Decimal `33.33000183105469f` is not really a number that exists as float, and isn't distinguishable from `33.33f` once it's represented as bits. `33.33000183105469d` is the closest double number to `33.33f` I suppose, but I don't think that it is correct to write it to JSON when encoding. The original behaviour of `"n": 33.33` is correct for a FLOAT type value in my opinion. The binary tests added in this PR pass, even in older versions of Avro without any other changes applied. Is there something else we're meant to be fixing here? Should we just revert this? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
