One tradeoff I have to do using run is that I can't autoformat scripts, make linters run properly, etc. It would be pretty bad just to put all commands in separate script files and just declare them in Runfile as this way, parameters, and code will be in two different files.
I suggest adding a way to include/import scripts by using wildcards or directories. Ideally, the folder hierarchy could be used to prefix the commands, i.e., scripts in the terraform folder will automatically prefix with contained commands with terraform-, so ./terraform/lock.sh would be terraform-lock. The manifest in comments could be searched for after the optional shebang.
This way, not having a grammar for Runfile files won't be such a big deal as the vast majority of the code will be simply shell scripts.
One tradeoff I have to do using
runis that I can't autoformat scripts, make linters run properly, etc. It would be pretty bad just to put all commands in separate script files and just declare them inRunfileas this way, parameters, and code will be in two different files.I suggest adding a way to include/import scripts by using wildcards or directories. Ideally, the folder hierarchy could be used to prefix the commands, i.e., scripts in the
terraformfolder will automatically prefix with contained commands withterraform-, so./terraform/lock.shwould beterraform-lock. The manifest in comments could be searched for after the optional shebang.This way, not having a grammar for
Runfilefiles won't be such a big deal as the vast majority of the code will be simply shell scripts.