Jarek Potiuk

> A tight coupling between task-sdk and any "server side" component is the opposite to one of the goals of AIP-72 (I'm not sure we ever explicitly said this, but...

And following what we already do with providers - (but the other way round) we should be running complete suite of tests of all previously released `task-sdk` versions that we...

> If all packages have pinned versions, you can't independently update client or server. "Dependency conflicts for administrators supporting data teams using different versions of providers, libraries, or python packages"...

And yes - I think we want to split packages for different reasons :) - that's where confusion comes from. you want to split "client vs. server" - but there...

> Agreed on Providers relying on Task SDK alone apart from the Executors one. Can make it independent too but just need to think through the BaseExecutor interface which is...

> Yup, hasn't been forgotten https://github.com/apache/airflow/issues/51552 :) Yeah - but you can't solve the knot "partially" if you do not know how you are going to solve #51552. Just to...

> It is part of the same parent GH issue, check the GH dependency. Sure. But I am telling that splitting dag-processor and triggerer out of the common is necessary...

BTW. Is there a reason we are not discussing that in devlist? I think this is a very important topic - how airflow will be split into distributions and something...

Thanks for notifying - I think our documentation generation process accidentally removed SBOM generation step in https://github.com/apache/airflow/pull/50464 I will bring it back and regenerate SBOMs.

The SBOM links should be fixed now - for example https://airflow.apache.org/docs/apache-airflow/3.0.2/sbom/ - the "stable" links will be refreshed in about 3 hours when all cache expires.