The GitHub Actions job "flink-connector-aws: build on pull request" on flink-connector-aws.git/new-combined-implementation has failed. Run started by GitHub user fmorillo7694 (triggered by fmorillo7694).
Head commit for run: 5e057b2760f0f21061b9654f61c9bc7683567f2b / Francisco <[email protected]> [FLINK-30481][Connectors/AWS] Address review: sync client option precedence, schema metadata robustness Sync HTTP client: the Apache-specific options (http-client.apache.max-connections, http-client.socket-timeout-ms) were merged as the lower-precedence side of AttributeMap.merge, so a generic aws.http-client.max-concurrency or read timeout silently won when both were set. The specific option now wins; documented, with both-set and generic-only tests. Affects Glue, Kinesis source and DynamoDB streams. Schema fidelity metadata: a physical column with a declared-type override that another engine dropped from Glue was resurrected from the recorded override; it is now omitted and warned like any dropped column. Column name lists (declared order, NOT NULL, primary key) escape the list separator so backtick-quoted identifiers containing a comma or backslash round-trip; values written before escaping decode unchanged. Dead code: ConnectorRegistry entries whose option can never hold a URI (kinesis, kafka, dynamodb, hbase, hive) are removed, and the registry test now proves every registered key actually yields a Glue location. The unwired CLIENT_TYPE_URLCONNECTION constant is removed. Docs state the storage location rule and the option precedence. Report URL: https://github.com/apache/flink-connector-aws/actions/runs/37109938695 With regards, GitHub Actions via GitBox
