stefanbras / ado_pipeline_helper Goto Github PK

I'll need to approve for the tests to run, since it has secrets in it

Since ruamel.yaml keeps yaml comments (I think?) we could allow settings in the pipeline yaml itself. So just something like

# organization=a
# project=b
# pipeline_id=3
stages:
   ...

And no, token should not be allowed to be set this way.

Some expressions are on the type ${{ if something}}. This should be handled somehow

remember that you can use expressions in them

Done-ish but it should also

• do it in the correct order - first resolve, the add to context
• parse expressions in variables (i think it does that atleast)

Docs says that it will render an empty string which seems very annoying.

We should probably throw a warning; I think it just crashes for now

• Update actual correct information
• Make section with resolver
• Make installation section (recommend pipx)
• Roadmap, goal, vision

A lot of duplicate code

Apparently you can use system variables in template reference paths (and probably even whole expressions) which is a problem.

I'll probably make some way to provide overrides for system variables (especially DefaultWorkingDirectory because we use that at work).

If that's not enough I'll have to break out the expression resolver again (#32)

The old parameter parsing is now just a special case of the generic expression evaluation

If you change any of the test pipelines, those changes needs to be pushed for tests to run.

This has casued issues for me roughly a billion times.

It should be easy enough for a fixture to just report "hey you have unpushed changes in the test pipelines directory"

CLI arg, maybe a separate file for both that and token?

Is probably a lot of work hmm

You know you'll forget to lint

Union ☑️
X | Y ❌

OR; put _ in front of everything that isn't public

Stage template including jobs template including steps template

• test if they are actually allowed in ${{ }}
• if allowed, load the group and insert (yikes)

I was setting up a whole new system to parse all template time expressions, but I see now that we just need to substitute parameter retrieving.

Going into ADO and creating the pipeline is tedious and it's why I have a small amount of pipelines each testing a thousand things.

The az cli client works well enough for this, but I should probably make a script so it's easier to use.