Hi team I’ve added a new built-in function

URL_DECODE_RECURSIVE

to handle multi-level URL encoding cases that frequently occur in real-world log pipelines (e.g., double/triple encoded request URIs, redirect targets, tracking parameters, or upstream systems that encode multiple times). The function recursively decodes until the result stabilizes (with a safety upper bound), which helps when the encoding depth is unknown. This is particularly useful for log ETL, clickstream analysis, ad/tracking data cleaning, and security rule preprocessing (e.g., detecting encoded path traversal patterns). Would appreciate feedback on: 1. API design (separate function vs optional parameter), 2. Performance considerations in streaming jobs, 3. Potential regression or misuse risks. [FLINK-38822] Extend URL_DECODE with optional recursive decoding https://github.com/apache/flink/pull/27351 Thanks in advance for the review!