zanmato1984 commented on issue #39577: URL: https://github.com/apache/arrow/issues/39577#issuecomment-1889090279
The cause seems to be that, the default buffer alignment (64b) doesn’t guarantee tail bytes safety when doing by-word operation for long fixed size types. Did some debugging, I found for this particular case, an encoded row took 19b, and there were 165 rows. They took 19b * 165 = 3135b, so 3136b is the actual size aligned by 64b. The last row access started at byte 3116 for 3 words (24b), which eventually exceeded the size 3136 buffer boundary by 4b. I’m working on a fix. -- 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]
