代码之家 › 专栏 › 技术社区 › Ashar

如何在需要的工作中为不同的工作设置多个条件?

github-actions github

Ashar · 技术社区 · 2 年前

我有一个可重复使用的工作流,有三个作业:作业1、作业2和作业3。

我希望满足的条件是:

作业3需要始终在作业2成功完成后运行,因此我提到 needs: [job1,job2] 在job3中。
如果跳过作业2,则应运行作业3
job3应该始终依赖于/需要job1来运行。

以下是我的工作流程:

on:

  workflow_call:

    inputs:

        enable_scan:
          required: false
          type: boolean
          default: true 

        runner_name:
          required: false
          type: string
          default: 'myselfhostedrunner' 

jobs:

 

  job1:

    runs-on: '${{ inputs.runner_name }}'
    outputs:
      myrunner: '${{ inputs.runner_name }}'

    steps:

      - name: Runner details
        run: |
          echo "Job1 called"

  job2:

  if: inputs.enable_scan

    runs-on: '${{ inputs.runner_name }}'

    steps:

      - name: Runner details
        run: |
          echo "Job2 called"

  job3:

    runs-on: '${{ needs.job1.outputs.myrunner }}'

    needs: [job1, job2]
    if: ${{ ! success() && ! cancelled() }}

    steps:

      - name: Checkout folderX
        uses: actions/checkout@v4

如何在job3中满足job1和job2的两个不同要求?

1 回复 | 直到 2 年前

VonC 2 年前

你有

 job1 --[always]--> job3
 job2 --[success]--> job3

根据您提供的配置和您想要满足的条件,您可以考虑:

调整 if 语句,以检查作业2是否已跳过或成功。
使用GitHub Actions的条件 如果 语法来创建允许job3在所需条件下运行的逻辑。
通过保持 needs: [job1, job2] 在job3定义中。
使用 outputs 从…起 job2 以发出信号 job3 是否应该运行。

例如:

on:
  workflow_call:
    inputs:
      enable_scan:
        required: false
        type: boolean
        default: true 
      runner_name:
        required: false
        type: string
        default: 'myselfhostedrunner' 

jobs:
  job1:
    runs-on: '${{ inputs.runner_name }}'
    outputs:
      myrunner: '${{ inputs.runner_name }}'
    steps:
      - name: Runner details
        run: |
          echo "Job1 called"

  job2:
    if: inputs.enable_scan
    runs-on: '${{ inputs.runner_name }}'
    outputs:
      status: ${{ steps.check-status.outputs.status }}  # Set an output to signal the status of this job to job3
    steps:
      - name: Runner details
        run: |
          echo "Job2 called"
      - id: check-status
        run: echo "status=completed" >> $GITHUB_OUTPUT  # Update to use GITHUB_OUTPUT environment file

  job3:
    runs-on: '${{ needs.job1.outputs.myrunner }}'
    needs: [job1, job2]
    if: ${{ fromJson(needs.job2.outputs.status).status == 'completed' || needs.job2.result == 'SKIPPED' }}  # Check the status output from job2
    steps:
      - name: Checkout folderX
        uses: actions/checkout@v4

现在,如果job2被跳过或成功,job3将执行,并且它将始终依赖于job1。

 job1 --[always]--> job3
 job2 --[success or skipped]--> job3

一种可能的替代方案是使用工件,在 作业2 和 job3 :

on:
  workflow_call:
    inputs:
      enable_scan:
        required: false
        type: boolean
        default: true 
      runner_name:
        required: false
        type: string
        default: 'myselfhostedrunner' 

jobs:
  job1:
    runs-on: '${{ inputs.runner_name }}'
    steps:
      - name: Runner details
        run: |
          echo "Job1 called"

  job2:
    if: inputs.enable_scan
    runs-on: '${{ inputs.runner_name }}'
    steps:
      - name: Runner details
        run: |
          echo "Job2 called"
      - name: Set status
        id: set-status
        run: echo "completed" > status.txt
      - name: Upload status
        uses: actions/upload-artifact@v2
        with:
          name: status
          path: status.txt

  job3:
    runs-on: '${{ needs.job1.outputs.myrunner }}'
    needs: [job1, job2]
    steps:
      - name: Download status
        uses: actions/download-artifact@v2
        with:
          name: status
      - name: Check status
        id: check-status
        run: |
          if [[ -f status.txt ]]; then
            STATUS=$(cat status.txt)
          else
            STATUS="SKIPPED"
          fi
          echo "::set-output name=status::$STATUS"
      - name: Conditional step
        if: ${{ steps.check-status.outputs.status == 'completed' || needs.job2.result == 'SKIPPED' }}
        run: |
          echo "Job3 running..."