Support cloud usage patterns about dud HOT 1 CLOSED

Thanks for your thorough feature request! I certainly see the benefits to this idea: Less commands for the developer to run, and a simpler path to automation and cloud compute from local dev pipelines. This idea is important to me, and I agree that we need better tooling in this space.

You also rightly point out the reasons not to implement this in Dud: Dud strives to do one thing well and exist orthogonally to other engineering tools such as source control systems and cloud provider APIs. Because Dud strives to be a good UNIX citizen, I think the best way to accomplish your goal is to wrap Dud in shell scripts. For example, here's a pseudocode shell script derived from your steps above:

git fetch
git checkout $COMMIT_SHA
dud pull [stage(s) in question]
dud run [stage(s) in question]
dud commit [...]
git commit -am 'automated commit ...'
dud push [...]
git push

This script template could also be run in most CI/CD systems, thus abstracting the orchestration for developers.

I agree that this approach isn't perfect. As you point out above, Dud still requires artifacts to be staged locally, so developers have to think carefully about local caching and network speeds and costs. To address these drawbacks, my current opinion is that a new "cloud native" tool should be created, and Dud is not this tool. I'm still stewing on this idea, but I'd be happy to keep chatting about it.

For now, please consider the shell script approach as the recommended way to "extend" Dud's interoperability with source control and other systems. I'm going to close this issue, but please feel free to continue using it as a discussion thread. Thanks again for the feedback; I really appreciate it!

from dud.