fix(event_handler): preserve comma-separated values in scalar parameters and model fields - #8514

Issue number: closes #8508

Summary

Changes

In the OpenAPI validation middleware:

• Updated _process_scalar_param to check if a parameter's value was split into a list of strings (by proxies such as API Gateway HTTP API v2 or Lambda Function URLs representing multi-value query or header parameters) and rejoin them with ",".join(value) when expecting a scalar string. Previously, _process_scalar_param only unwrapped single-element lists (len(value) == 1), leaving multi-item lists intact and causing Pydantic to reject valid comma-containing queries or headers with a 422 string_type validation error.
• Updated _normalize_field_value for Pydantic model fields to rejoin string lists with ",".join(value) when the target field annotation is not a sequence, instead of returning value[0] which caused strings containing commas to be silently truncated.
• Added comprehensive unit tests in tests/functional/event_handler/_pydantic/test_openapi_validation_middleware.py verifying comma preservation for scalar query parameters, headers (e.g. X-Forwarded-For), and Pydantic models.

User experience

Before:

1. A scalar query parameter containing commas (e.g. GET /search?search=hello,world where search: Annotated[str, Query()]) was rejected with HTTP 422:
   {"statusCode": 422, "body": "[{\"type\":\"string_type\",\"loc\":[\"query\",\"search\"],\"msg\":\"Input should be a valid string\",\"input\":[\"hello\",\"world\"]}]"}
2. A header containing commas (e.g. X-Forwarded-For: 203.0.113.9, 10.0.0.1 where x_forwarded_for: Annotated[str, Header()]) was rejected with HTTP 422 string_type.
3. A query parameter mapped to a Pydantic model with a string field (e.g. search: str) was truncated to only the first segment before the comma ("hello" instead of "hello,world").

After:
All query parameters and headers containing commas are properly validated and preserved as full strings ("hello,world", "203.0.113.9, 10.0.0.1") across scalar parameters and Pydantic models, while sequence fields continue to be parsed as list items.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.