You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@airflow.apache.org by GitBox <gi...@apache.org> on 2020/02/19 20:59:28 UTC

[GitHub] [airflow] potiuk edited a comment on issue #7191: [AIRFLOW-4030] second attempt to add singularity to airflow

potiuk edited a comment on issue #7191: [AIRFLOW-4030] second attempt to add singularity to airflow
URL: https://github.com/apache/airflow/pull/7191#issuecomment-588461323
 
 
   > One quick note about the docs for pylint - the files are now changed to have pylint_main.sh and pylint_tests.sh. Also, this unexpectedly started that huge build process that failed before. Given a .pylintrc file in some root and running pylint, it would be great to see a much simpler (not wrapped / abstracted) run of pylint on a single file.
   
   I'd really recommend (again) pre-commits. They will run pylint on all modified files only. This is far better than having to run it manually - the will simply run pylint (or whatever of the other tests) automatically on all the files you modified when committing it. You don't even have to remember about it. And the reason for the build process is to make sure that you run it in the same environment as Travis CI and other people. The problems with tools like pylint is that the results depend often on installed libraries, pylint version etc. For example you need to have installed stubs for python libraries in order to do full static analysis. So there is huge variability depending on your local environment and you can simply get different results locally than on Travis. That's why we run pylint tests inside the docker container which mirrors this container that runs on Travis. This is the only way to make sure everyone runs in the same environment.
   
   It happened many times that people reported "I see different errors than those on Travis" - and it turned out that they have different pylint or python version. This is the real benefit of the "huge build". you start to appreciate it when your team of contributors is totally distributed, maintain their own environment and consists of 600 individuals - each with their own setup.
   
   Simply - it looks differently from the point of view of single person and differently from the point of view of team of few hundred people. Sometimes efficiency of team requires some sacrifice of efficiency of single individual. 

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services