potiuk commented on issue #21496:
URL: https://github.com/apache/airflow/issues/21496#issuecomment-1198566403

   I think it makes very little sense to do it now. Our survey 
https://airflow.apache.org/blog/airflow-survey-2022/ shows that Airflow 2 
adoption is high and I think it makes very litle sense to make such upgrade now 
-I have not seen a single person having problem with it.
   
   > 1. preventing people who pip install apache-airflow-upgrade-check (without 
airflow installed) from being unable to use airflow upgrade_check
   
   Why would anyone do that? 
   
   > 2. preventing those who (dangerously) run pip install --upgrade 
apache-airflow-upgrade-check from breaking their python environment and ending 
up with whatever the latest airflow version is
   
   There are many other ways to break airflow by running random `pip install` 
command. I do not see why this particular command (which is neither endorsed or 
promoted by us) should be special?
   
   I think the real problem is with (dangerous like crasy and we spoke about it 
many times and warned against doing it) is promoting the installation of 
packages in docker via env variables as "production ready". This is exactly the 
kind of problems we foresaw and expected people to have. It's not the problem 
of the upgrade check, really. 
   
   I think a better way to handle it (if you still want to continue to support 
dynamic installation of packages) is to detect dangerous commands and prevent 
them. I think this is the only reasonable approach.


-- 
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]

Reply via email to