Rework pipeline data structure

[qwerty287]

supersedes #3421
closes #2971
closes #4677
closes #4195

Currently, our pipeline model has fields like the message, title etc., but every field is used differently, dependending on the event.

This PR changes this and splits these reused fields to have a more clear pipeline model.

The model now has fields like:

• commit (containing sha, author, forge url and message)
• pull request (index, title, from fork, labels)
• release title

and similar.

Advantages:

1. cleaner model and less differently used fields
2. fix the problem that there is a difference between commit and pipeline author (pipeline author = pusher) Environment Variable CI_COMMIT_AUTHOR_EMAIL is wrong #2971
3. store and show the commit message for every pipeline
4. store and show the context of a pipeline (PR title, deployment target, release title etc.) and show it

Especially 3 and 4 and are the enhancement parts here, as we can now store both context and commit message and not only one.

[pat-s]

@qwerty287 Why is that breaking? Because of the env var change?

[qwerty287]

Yes, env var change and it also changes the json for the pipeline model

[qwerty287]

@woodpecker-ci/maintainers could I ask again for reviews here? This is open for almost a year now. And it simplifies the pipeline structure a lot. (Also feel free to directly modify the branch if there's something to fix)

[anbraten]

Thanks for updating to main again. I will try to have a look at it and will do some testing. Could you try to test with a copy of our ci.woodpecker-ci.org db as well. Last time it had some hickups with the migrations.