Refactor doGetBeanNamesForType for improved readability and performance

[polyglot-k]

PR Description:

This pull request refactors the logic for collecting bean names by type in DefaultListableBeanFactory to improve readability and maintainability. The main change is splitting the logic into two helper methods: one for processing bean definitions and another for processing manually registered singletons. This makes the code easier to follow and maintain, without altering its functionality.

Refactoring and code organization:

• Extracted the core logic for collecting bean names into two new private methods: processBeanDefinitions (handles bean definitions) and processManualSingletons (handles manually registered singletons), improving separation of concerns and code readability.
• Simplified and clarified the logic for skipping aliases, handling abstract and lazy beans, and checking type matches within the new helper methods.

Error handling improvements:

• Improved exception handling by consolidating catch blocks and adding more descriptive logging when bean class loading or metadata resolution fails.

Note

This helps with hot path configuration at the jit computer level.

Thank you for reviewing this PR! Your feedback is highly appreciated.