kaxil commented on issue #67515: URL: https://github.com/apache/airflow/issues/67515#issuecomment-4553627252
@dwreeves Apologies if you felt any disrespect, none was intended. re: What's being commented Anything that comes from a my handle, you can be assured that while it might have been assisted by AI is coming from me. I am using it word it in some way. You can also be rest assured that I have read every single line of what you have wrote, which is why I went so far to close all the PRs as well. > You, or rather the AI that is writing these messages, isn't even saying anything that contradicted what I said. If you feel so, I think you should read the comments again! "The scheduler also does not process raw Dag files, that's just factually incorrect in Airflow 3 land!" is literally part of https://github.com/apache/airflow/issues/67515#issuecomment-4548272652 which is in response to the following line yo had in the Issue description: > On the margin, lazy loading improves performance by reducing the memory footprint and speeding up start-ups during the scheduler’s DAG file processing and during task execution. Even in Airflow 2 land, Scheduler as a component does not do file processing and you can have deploy separate dag file processor. > It's a little disrespectful that I am talking to AI and not a human, but it's also a little disrespectful that you are assuming my problem either doesn't exist or it's because I'm some silly guy who's loading a 10 GB Pandas dataframe into a 2 GB worker. See, we as committers, need to think of the entire project and all users not one. I do not want to have Spaghetti code lying around because we are the ones who will need to deal with it -- even after years. The details are also not just for you, but future myself, future contributors who might have similar issues and "proper" solution rather than a temporary solution that doesn't make it significantly better. Which is why I had include AWS & Django's example. > "we'll wait for PEP 810 because that's easier." No, not because it is easier, but because that is the right answer! It was designed specifically for this sort of use-case. So to dismiss it and say that the message reads like AI-generated is then disrespectful! As I said above, if it is coming my user handle, anything posted from it is my responsibility including accuracy, including tone. Even the [last comment](https://github.com/apache/airflow/issues/67515#issuecomment-4550501544) was because you asked about the common use-case and the impact on it. Airflow does support multiple executors so the profile looks really different for each! If you would have rather preferred discussion about the problem in hand, that would have been a different story. But not when you come with a solution in mind!. https://lucumr.pocoo.org/2026/5/24/pi-oss/ is worth reading if you don't mind the suggestion -- on why Committers are frustrated as well when people come in issue with a solution that might not be the best long-term solution. If you were to dismiss my comments as AI-generated when they clearly are coming from my profile -- that is what I'd call disrespectful . I have not just spend a couple of years to this project, more than 8 years. So yes, I am protective of what goes into the project. -- 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]
