Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 21 additions & 0 deletions .github/workflows/copilot-setup-steps.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
name: "Copilot Setup Steps"

# Allow testing of the setup steps from your repository's "Actions" tab.
on: workflow_dispatch

jobs:
# The job MUST be called `copilot-setup-steps` or it will not be picked up by Copilot.
# See https://docs.github.com/en/copilot/customizing-copilot/customizing-the-development-environment-for-copilot-coding-agent
copilot-setup-steps:
runs-on: ubuntu-latest

permissions:
contents: read

# You can define any steps you want, and they will run before the agent starts.
# If you do not check out your code, Copilot will do this for you.
steps:
- uses: actions/checkout@v4.2.2

- name: Restore solution
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thinking aloud here - would it also help to set some things in the environment so that it would use the dotnet that it pulled down on the path? I'm not sure what sort of resources the agent will be using, but we should try to set it up for local-development as close as possible. How do we see / measure how good these steps are working for the agent? Can we see it try them and then see how successful it is at doing things?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

it seems to successfully know to go to the root of the repo and run 'build.sh' or whatever the onboarding instructions say.

But, if you add a .github\copilot-instructions.md you can be explicit about how it should do incremental subtree builds and how to run tests too.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, I think putting dotnet on the path and including some instructions for building individual projects, testing individual projects might be a good idea since I expect the agent is actually doing these things so giving it the quicker inner-loops like real devs use should make it more efficient.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree it's a good idea to add some specific instructions. I'll try it out in a separate PR.

run: ./build.sh --restore --excludecibinarylog --warnaserror false /p:BuildAllConfigurations=true /p:DotNetBuildAllRuntimePacks=true