BasPH commented on issue #5383: [AIRFLOW-4737] Increase and document celery queue name limit URL: https://github.com/apache/airflow/pull/5383#issuecomment-500216912 Few things to consider here: migrations are also part of the Airflow codebase, so for consistency sake we like to apply the same conventions. Every function performs some sort of logic, so it's convenient for other people and even your future self to explain _what_ the function does, even if it's a simple one-liner. Typically business logic code is very hard to debug and understand - because others don't have the context you had while writing the code. IMO this migration is a perfect example of this - the number of chars is raised. If it ever turns out to be an issue and somebody else would look at the code, they don't see any explanation in the code stating why the number was raised and why 256 was chosen. Only after reading your JIRA issue do I understand why these choices were made, which are perfectly fine. Unfortunately, not everybody explains the what and why in their Git commit/PR/JIRA issue. Reading the Airflow code base, there are a lot of such examples of undocumented logic, so now that we are in the process of adding Pylint I prefer not to lower the boundaries just because it's inconvenient. If you'd write a DAG at your work and you're the only person maintaining the thing, I wouldn't really be bothered, but Airflow is software used and developed by many people around the world so I strongly prefer to have consistently documented code. Sorry for the wall of text, hope you understand 😃
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
