Eliaaazzz opened a new pull request, #58515:
URL: https://github.com/apache/spark/pull/58515

   ### What changes were proposed in this pull request?
   
   `_extract_jvm_stacktrace` in `pyspark/errors/exceptions/connect.py` follows 
the `cause_idx` chain from `FetchErrorDetailsResponse` recursively, without 
validating the index. This PR walks the chain iteratively instead and stops 
when the next index is out of range or has already been visited, so each error 
is formatted at most once.
   
   The Scala client has a similar unguarded recursion in 
`GrpcExceptionConverter.errorsToThrowable`; this PR only covers the Python 
client.
   
   ### Why are the changes needed?
   
   If the error details returned by `FetchErrorDetails` contain a malformed 
cause chain, the client crashes inside error conversion and the original server 
error is lost:
   
   - a cyclic `cause_idx` (two errors pointing at each other, or an error 
listed as its own cause) raises `RecursionError: maximum recursion depth 
exceeded`
   - an out-of-range `cause_idx` raises `IndexError: list index out of range`
   
   Reproducible by calling `convert_exception` with a 
`FetchErrorDetailsResponse` where `errors[0].cause_idx = 1` and 
`errors[1].cause_idx = 0`.
   
   ### Does this PR introduce _any_ user-facing change?
   
   Only for malformed error details: the client now raises the converted server 
exception instead of `RecursionError`/`IndexError`. Well-formed cause chains 
are formatted exactly as before.
   
   ### How was this patch tested?
   
   Added `test_convert_exception_with_malformed_cause_chain` to 
`pyspark.errors.tests.test_connect_errors_conversion` covering the mutual 
cycle, the self-cycle, and the out-of-range case. All tests in the module pass:
   
   ```
   python -m pytest 
python/pyspark/errors/tests/test_connect_errors_conversion.py
   13 passed
   ```
   
   ### Was this patch authored or co-authored using generative AI tooling?
   
   Generated-by: Claude Code
   


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to