Skip to content

Latest commit

 

History

History
154 lines (120 loc) · 10.8 KB

outdated_mode.md

File metadata and controls

154 lines (120 loc) · 10.8 KB

Outdated Mode

Issue #253 contains a great discussion about the outdated mode and why it was added as a feature to this project.

The outdated_mode input option gives you the ability to determine how the branch-deploy Action should handle outdated branches. This input option is closely related to the update_branch input option as well.

Using the outdated_mode input option

The outdated_mode input option has three possible values:

  • strict (default): The branch-deploy Action will consider your branch (that you are deploying) to be out-of-date if it is behind the default|stable branch (i.e main) or if it is behind the branch that your pull request is merging into.
  • pr_base: The branch-deploy Action will consider your branch (that you are deploying) to be out-of-date if it is behind the branch that your pull request is merging into.
  • default_branch: The branch-deploy Action will consider your branch (that you are deploying) to be out-of-date if it is behind the default|stable branch (i.e main). The default_branch is determined by the value you have set for the stable_branch input option. By default, the value of the stable_branch input option is main.

How the outdated_mode input option works with the update_branch input option

If your branch is determined to be out-of-date, the branch-deploy Action will not deploy your branch. What happens next is determined by the value of the update_branch input option. Here are the three possible values for the update_branch input option and how they will operate:

  • warn: The branch-deploy Action will leave a comment on your pull request indicating that your branch is out-of-date and that you should update it before it can be deployed.
  • force: The branch-deploy Action will attempt to update your branch by merging the latest commits into your branch for you. You will need to re-run the deployment command after the branch-deploy Action has attempted to update your branch.
  • disabled: The branch-deploy Action will allow you to deploy your branch even if it is out-of-date. (dangerous)

It should be noted that if you are using the force value for the update_branch input option, the branch-deploy Action will only attempt to update your branch with the target base branch of the pull request. This means that if your target base branch is something other than the default|stable branch, the branch-deploy Action will only attempt to update your branch with the latest commits from the target base branch. You will still be warned about your branch being out-of-date with the default|stable branch but it will not try to update your branch with the latest commits from the default|stable branch due to the complexity and potential risk involved with such a merge. You will need to manually update your branch with the latest commits from the default|stable branch if you want to deploy your branch and that is the case.

The final bit of important information to note is that rollback deployments (i.e .deploy main) will always work even if the default|stable branch is considered out-of-date. This is for safety reasons so that users of this Action can always rollback their changes in the event of an issue.

About the outdated_mode input option

The outdated_mode input option is useful when you want to ensure that the branch-deploy Action only deploys branches that are up-to-date. This is especially useful when you have a large team of developers working on a project and you want to ensure that commits are not rolled back accidentally when you are deploying a branch. For the sake of availability and deployment confidence, this input option is set to strict by default. This will ensure that your branch is up-to-date with the latest commits from both the default|stable branch and the branch that your pull request is merging into.

Example Scenarios

This section will go into a few example scenarios to help you understand how the outdated_mode input option works in practice.

Big thanks to @jessew-albert for providing these examples!

For all scenarios, we assume that we are operating on a pull request that is at branch featureB (commit featB2) in the diagram. The target base branch of the pull request is featureA and the default|stable branch is main.

Scenario 1

This example diagram shows a scenario where a branch is created from the default|stable branch (main in this case). The name of this branch is featureA and a following branch is created from featureA called featureB. A few commits are made to the featureB branch and sometime later, another developer merges two commits into main. The featureB branch is now out-of-date with main even though it is up-to-date with the featureA branch. In this example, the target base branch of featureB is featureA.

gitGraph
   commit id: "A"
   commit id: "B"
   branch featureA
   checkout featureA
   commit id: "featA1"
   commit id: "featA2"
   branch featureB
   checkout featureB
   commit id: "featB1"
   commit id: "featB2"
   checkout main
   commit id: "C"
   commit id: "D"

Let's pretend that we are trying to run the .deploy command on the pull request associated with featureB (which is trying to merge into featureA).

  • If the outdated_mode input option is set to strict, the branch-deploy Action will consider the featureB branch to be out-of-date and prevent the deployment from happening. This is because the featureB branch is behind the main branch. If this branch were to be deployed, it would roll back the commits made to main and that could cause issues. Even though the featureB branch is up-to-date with the featureA branch, it is still considered out-of-date because it is behind the main branch when the outdated_mode input option is set to strict.
  • If the outdated_mode input option is set to pr_base, the branch-deploy Action will consider the featureB branch to be up-to-date and allow the deployment to happen. This is because the featureB branch is up-to-date with the featureA branch. The featureB branch is behind the main branch but that is not relevant in this case because the target base branch of the pull request is featureA and not main.
  • If the outdated_mode input option is set to default_branch, the branch-deploy Action will consider the featureB branch to be out-of-date and prevent the deployment from happening. This is because the featureB branch is behind the main branch. If this branch were to be deployed, it would roll back the commits made to main and that could cause issues. Even though the featureB branch is up-to-date with the featureA branch, it is still considered out-of-date because it is behind the main branch when the outdated_mode input option is set to default_branch.

Scenario 2

This scenario shows the "ideal" situation for a successful deployment. In this example, the featureB branch is up-to-date with both the featureA branch and the main branch. When this is the case, it does not matter which value you have set for the outdated_mode input option. Whether you have it set to strict, pr_base, or default_branch, all options in this case are up-to-date and the branch-deploy Action will allow the deployment to happen. Hooray!

This diagram expands upon Scenario 1 and shows the featureB branch being updated by having main merged into it:

gitGraph
   commit id: "A"
   commit id: "B"
   branch featureA
   checkout featureA
   commit id: "featA1"
   commit id: "featA2"
   branch featureB
   checkout featureB
   commit id: "featB1"
   commit id: "featB2"
   checkout main
   commit id: "C"
   commit id: "D"
   checkout featureB
   merge main

Alternatively, you could rebase the stack with featureA and featureB on top of main:

gitGraph
   commit id: "A"
   commit id: "B"
   commit id: "C"
   commit id: "D"
   branch featureA
   checkout featureA
   commit id: "featA1"
   commit id: "featA2"
   branch featureB
   checkout featureB
   commit id: "featB1"
   commit id: "featB2"

Scenario 3

In this scenario, the branch featureB is now out-of-date in all cases. This is because the featureB branch is behind both the featureA branch and the main branch. This is the worst-case scenario and the branch-deploy Action will prevent the deployment from happening in all cases. It does not matter what value you have set for the outdated_mode input option. Whether you have it set to strict, pr_base, or default_branch, all options in this case are out-of-date and the branch-deploy Action will prevent the deployment from happening.

gitGraph
   commit id: "A"
   commit id: "B"
   branch featureA
   checkout featureA
   commit id: "featA1"
   commit id: "featA2"
   branch featureB
   checkout featureB
   commit id: "featB1"
   commit id: "featB2"
   checkout featureA
   commit id: "featA3"
   checkout main
   commit id: "C"
   commit id: "D"

Scenario 4

In this example, the featureB branch is behind featureA but up-to-date with main. In this case, the outdated_mode input option will determine whether the branch-deploy Action will allow the deployment to happen or not.

  • If the outdated_mode input option is set to strict (default), the branch-deploy Action will consider the featureB branch to be out-of-date and prevent the deployment from happening. This is because the featureB branch is behind the featureA branch. Even though the featureB branch is up-to-date with the main branch, it is still considered out-of-date because it is behind the featureA branch when the outdated_mode input option is set to strict so both the target base branch (of the pull request) and the default|stable branch are considered.
  • If the outdated_mode input option is set to pr_base, the branch-deploy Action will consider the featureB branch to be out-of-date and prevent the deployment from happening. This is because the featureB branch is behind the featureA branch. In this case, the default|stable branch is not considered, only the target base branch (of the pull request) is considered.
  • If the outdated_mode input option is set to default_branch, the branch-deploy Action will consider the featureB branch to be up-to-date and allow the deployment to happen. This is because the featureB branch is up-to-date with the main branch. Even though the featureB branch is behind the featureA branch, that is not relevant in this case because default_branch option only evaluates the default|stable branch.
gitGraph
   commit id: "A"
   commit id: "B"
   branch featureA
   checkout featureA
   commit id: "featA1"
   commit id: "featA2"
   branch featureB
   checkout featureB
   commit id: "featB1"
   commit id: "featB2"
   checkout featureA
   commit id: "featA3"