jiayuasu commented on PR #1188: URL: https://github.com/apache/sedona-db/pull/1188#issuecomment-5391795775
Thanks, that context is helpful. I agree the lazy-failure behavior has intentional value: it preserves wheels for debugging and lets extension/nightly work continue when a core wheel flakes. I will keep #1188 as a draft rather than treating removal of that behavior as the long-term fix. The main issue I traced is the pyogrio/GDAL concurrent-reader abort. #1187 is the proposed tactical fix: it serializes the full pyogrio reader lifecycle within each physical scan, and the original reproducer (which aborted at iteration 72) completed 2,002 affected test executions after the change. The native GDAL reader in #1092 remains the durable direction. I am not yet certain that the scan-local locking approach is directionally what we want, so I would appreciate your take on #1187. If it is, I will focus on getting #1187 in instead of moving #1188 forward. -- 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]
