[youdie006]

Follow-up to #258. 333d0d7 made graphql_sync pass assume_not_async_iterable when check_sync is off, so objects that are both iterable and async-iterable (such as a Django QuerySet) are iterated synchronously. In #258 the plan was to set it "in graphql_sync and execute_sync", but execute_sync only got the is_awaitable part. It still uses the default async-iterable check, takes the async list path, and puts the coroutine into data:

before: execute_sync (dual): ExecutionResult(data={'listField': <coroutine object Executor.complete_async_iterator_value ...>}, errors=None)
        RuntimeWarning: coroutine 'Executor.complete_async_iterator_value' was never awaited
after:  execute_sync (dual): ExecutionResult(data={'listField': ['sync']}, errors=None)

A pure async generator now gives the same "Expected Iterable" field error as in graphql_sync, instead of a coroutine in data. With check_sync=True nothing changes. The helper is duplicated in execute.py, as assume_not_awaitable already is, because execute.py can't import from graphql.py. The tests in test_sync.py cover the dual iterable, a pure async iterable, and the check_sync=True error.

The full test suite passes with the 100% coverage gate on Python 3.14, and ruff check, ruff format --check and mypy src tests pass too. I pinned pytest below 9.2, because pytest-benchmark 5.3 errors on 9.2's benchmark fixtures.

Written with AI assistance (Claude); I have reviewed the change.