Support setting Cron in Pipeline

[JakobDev]

Clear and concise description of the problem

If you want to setup a Cronjob, you have to do it in the UI and add a cron event filter to the Pipelne. It would be great, if Cron can be used directly in the Pipeline.

Suggested solution

Something like this:

when:
    event: cron
    cron: some-name
    schedule: @daily

Alternative

No response

Additional context

This would make it easier to copy and paste Pipelines in other Projects. I also think, that as much as possible should be configured with the Pipeline files and not with the GUI.

Validations

• Read the Contributing Guidelines.
• Read the docs.
• Check that there isn't already an issue that request the same feature to avoid creating a duplicate.

[6543]

we would somehow get that entry into DB ... and maintain it acordingly

To Decide: how to determine the reference to run on and can there ocure config conflicts between branches ..

yes you could have multible crons with same name specifyed and so it would not be unique 💣

[Janik-Haag]

Then we would still need to figure out how to do migrations, because just wiping the existing cronjobs with the declarative ones seems like it could make a few users quite unhappy but that's probably a implementation detail to worry about later.

[qwerty287]

I think updating the yaml should delete all crons that were previously defined but removed then. What would you suggest otherwise?

If you mean crons defined via ui: I'd still support both ways, at least for now. You could add a new boolean db column which is used to check whether it's yaml or ui defined.

Or do you mean something completely different?

[Janik-Haag]

Im fine with just deleting, but I thought about maybe require the user to maybe have to acknowledge that the current crons will be deleted.

[Janik-Haag]

I should probably add a file like pipeline/frontend/yaml/types/cron.go and just add a struct with stuff like branch, timer, etc..
How does the data flow from .woodpecker file/folder in the git repo to getting parsed to writing stuff to the database work (I read the architecture diagram but couldn't make much sense out of it)?
While looking through the code I also found:

	// Cron is the JSON data of a cron job
	Cron struct {
		ID        int64  `json:"id"`
		Name      string `json:"name"`
		RepoID    int64  `json:"repo_id"`
		CreatorID int64  `json:"creator_id"`
		NextExec  int64  `json:"next_exec"`
		Schedule  string `json:"schedule"`
		Created   int64  `json:"created_at"`
		Branch    string `json:"branch"`
	}

Name, Schedule, and Branch are obvious. we probably also can get the repo_id from somewhere:tm: and use the git commit time as created timestap, we are probably able to leave next_exec untouched, but I don't know what to do about id and creatore_id. we could probably make creator_id a optional value because git blame would tell you who modified the cron job.

[qwerty287]

This struct isn't the db model I think(it's here: https://github.com/woodpecker-ci/woodpecker/blob/main/server/model/cron.go). Yes, I think you can skip the creator. The ID field is automatically set when you insert it into the db, you don't need to care about this.