Ayush0612005 commented on issue #14536: URL: https://github.com/apache/gravitino/issues/14536#issuecomment-6577586288
I'd like to work on this issue; could you assign it to me if it is still available? I checked the current `CatalogWrapperForREST` scan path: the non-incremental branch applies `useSnapshot(...)` and then the common request options, so I'd start with tests that distinguish selecting files from an older snapshot from selecting the schema used to bind the request. My first regression cases would cover a renamed column with an older snapshot under both `use-snapshot-schema` values, including projection/filter binding and the REST response. Separately, I'd reproduce the retired partition-spec case with actual data files, evolve the default spec and drop its former source column, then check both scan-task serialization and the HTTP error-model/content-type contract. I'd also include an empty filtered result because the response currently includes the table's full spec map, not only specs referenced by returned files. I haven't run these reproductions yet and won't assume one change solves both paths. Is there ongoing work in the Iceberg REST planning area I should coordinate with, or a preference for splitting the schema-binding and historical-spec serialization fixes into separate PRs? AI assistance: Codex helped inspect the issue/source and draft this plan. I'll provide reproduction evidence and regression coverage before proposing the implementation. -- 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]
