Proposal: Multiple Dangers on the same service should create multiple tables

[orta]

In Danger ruby, if you run two instances of Danger they create separate comments.

I wonder if instead it could make two tables on the same comment. e.g.

[orta]

	Warnings
⚠️	Please add a changelog entry for your changes.

	Warnings
⚠️	Detected a package.json change but no lock file change.

Generated by 🚫 dangerJS

[orta]

They could get named, so that you can differentiate.

[macklinu]

I believe danger/danger#247 is a related Danger Ruby PR for this.

[orta]

It is! However this supports multiple comments - rather than single comment, multiple tables

[orta]

This is also an interesting take: danger/danger#764

Both of those could be built with the same infrastructure.

[orta]

Been thinking a lot about this due to Peril - I think the route I want to take is to be able to say where each individual message comes from, and have those merged into one comment.

The example being: Org Dangerfile for assigning users to PRs, but a Repo still having a Dangerfile with it's own rules. Changing the assignee and having the org Dangerfile run shouldn't delete messages from the repo Dangerfile.

[steprescott]

The uses case we used multiple dangerfiles were for Pre-CI and Post-CI checks.

Pre-CI validated the PR metadata and the Post-CI posted the results of the job on the ticket to make visuality of issues easier.

Now with peril the Pre-CI checks can now be ran outside of the CI world.

[orta]

Ah yeah, we have this exact thing on artsy/eigen too

I can see the value in the double table with titles

[Startouf]

Thanks for the Pre/post CI stuff, We recently switched to circleci workflows and this somewhat didn't play too well, didn't realize we could use two dangerfiles...