Deployment Validation with coverage-formatters writes to nested coverage/ folder

[codefriar]

Summary

We use Github actions. One of our steps is to do a trial deployment to validate and do code coverage calculations. Specifically we use this command:

sf project deploy start --verbose --dry-run -x package/package.xml --target-org stage --ignore-conflicts --wait 45 --test-level RunSpecifiedTests --tests ${{ steps.related-tests.outputs.result }} --coverage-formatters json --coverage-formatters cobertura

This command executes normally. Here's the log output from that command:

... (skipping deployment and testing lines of 'In Progress'

Test Success [1]
✓ CaseRepoTest.unitTestConstructorPositive

Apex Code Coverage
 Name         % Covered Uncovered Lines 
 ──────────── ───────── ─────────────── 
 CaseRepo 100%                      

Test Results Summary
Passing: 1
Failing: 0
Total: 1
Time: 70
Code Coverage formats, [json, cobertura], written to coverage/
Dry-run complete.

This results in the cobertura file being written to coverage/coverage/cobertura.xml' not coverage/cobertura.xml` as I would expect, given the output line saying "Code Coverage formats, [json, cobertura], written to coverage/"

Steps To Reproduce

I'm unable to create a full repo, for security reasons. However, the operant bits of the workflow file are as follows:

      - name: Generate Deployment Delta
        run: |
          sf sgd source delta -t="HEAD" -f="HEAD~1" -o "." --loglevel=debug
      - name: Deploy to Stage and run Tests (Check Only)
        if: ${{ steps.related-tests.outputs.result != null }}
        run: |
          echo "Running related tests: ${{ steps.related-tests.outputs.result }}"
          sf project deploy start --verbose --dry-run -x package/package.xml --target-org stage --ignore-conflicts --wait 45 --test-level RunSpecifiedTests --tests ${{ steps.related-tests.outputs.result }} --coverage-formatters json --coverage-formatters cobertura
      - name: Deploy to Stage without running Tests (Check Only)
        if: ${{ steps.related-tests.outputs.result == null }}
        run: |
          echo "Running related tests: ${{ steps.related-tests.outputs.result }}"
          sf project deploy start --verbose --dry-run -x package/package.xml --target-org stage --ignore-conflicts --wait 45 --coverage-formatters json --coverage-formatters cobertura
      - name: "Check Cobertura file was created"
        id: does_cobetura_exist
        shell: bash
        run: |
          file=$(pwd)/coverage/coverage/cobertura.xml
          if [ -f $file ]; then
            echo "coverage/coverage/cobertura.xml exists"
            echo "FILE_EXISTS=true" >> $GITHUB_OUTPUT
          else
            echo "$(pwd)/coverage/coverage/cobertura.xml does not exist"
            echo "FILE_EXISTS=true" >> $GITHUB_OUTPUT
          fi

Expected result

cobertura.xml should be found in the WorkingDir's coverage/ folder.

Actual result

cobertura.xml (and it's sister Json file) are written to coverage/coverage/ folder.

System Information

Github Actions using the official SF cli docker image:

    container:
      image: salesforce/cli:latest-full

Additional information

[github-actions]

Hello @codefriar 👋 It looks like you didn't include the full Salesforce CLI version information in your issue.
Please provide the output of version --verbose --json for the CLI you're using (sf or sfdx).

A few more things to check:

• Make sure you've provided detailed steps to reproduce your issue.
  • A repository that clearly demonstrates the bug is ideal.
• Make sure you've installed the latest version of Salesforce CLI. (docs)
  • Better yet, try the rc or nightly versions. (docs)
• Try running the doctor command to diagnose common issues.
• Search GitHub for existing related issues.

Thank you!

[github-actions]

Thank you for filing this issue. We appreciate your feedback and will review the issue as soon as possible. Remember, however, that GitHub isn't a mechanism for receiving support under any agreement or SLA. If you require immediate assistance, contact Salesforce Customer Support.

[codefriar]

{
  "architecture": "linux-x64",
  "cliVersion": "@salesforce/cli/2.35.6",
  "nodeVersion": "node-v18.20.1",
  "osVersion": "Linux 6.5.0-1017-azure",
  "rootPath": "/usr/local/lib/nodejs/lib/node_modules/@salesforce/cli",
  "shell": "bash",
  "pluginVersions": [
    "@oclif/plugin-autocomplete 3.0.[13](https://github.com/department-of-veterans-affairs/va-teams/actions/runs/8635526082/job/23673436239#step:14:14) (core)",
    "@oclif/plugin-commands 3.2.2 (core)",
    "@oclif/plugin-help 6.0.20 (core)",
    "@oclif/plugin-not-found 3.1.1 (core)",
    "@oclif/plugin-plugins 5.0.1 (core)",
    "@oclif/plugin-search 1.0.20 (core)",
    "@oclif/plugin-update 4.2.2 (core)",
    "@oclif/plugin-version 2.0.16 (core)",
    "@oclif/plugin-warn-if-update-available 3.0.[15](https://github.com/department-of-veterans-affairs/va-teams/actions/runs/8635526082/job/23673436239#step:14:16) (core)",
    "@oclif/plugin-which 3.1.7 (core)",
    "@salesforce/cli 2.35.6 (core)",
    "apex 3.1.0 (core)",
    "auth 3.5.0 (core)",
    "data 3.2.2 (core)",
    "deploy-retrieve 3.4.0 (core)",
    "info 3.1.0 (core)",
    "limits 3.2.0 (core)",
    "marketplace 1.1.0 (core)",
    "org 3.6.0 (core)",
    "packaging 2.2.0 (core)",
    "schema 3.2.0 (core)",
    "settings 2.1.0 (core)",
    "sobject 1.2.0 (core)",
    "source 3.2.0 (core)",
    "telemetry 3.1.[17](https://github.com/department-of-veterans-affairs/va-teams/actions/runs/8635526082/job/23673436239#step:14:18) (core)",
    "templates 56.1.0 (core)",
    "trust 3.4.0 (core)",
    "user 3.4.0 (core)",
    "sfdx-git-delta 5.34.0 (user)"
  ]
}

You should change your info-required-bot to accept a docker image tag as well.

[WillieRuemmele]

hey @codefriar - nice catch - we'll get this fixed

[codefriar]

@WillieRuemmele - at the risk of being told off here,

it'd be nice if i could specify the folder / filename ;)

[WillieRuemmele]

@codefriar more than just --output-dir?

update on the PR, but this seems to be edging towards a breaking change, or at least breaking an existing bug that everyone's got accustomed to... how about making the message correct, e.g.

Code Coverage formats, [cobertura], written to /coverage/coverage

[codefriar]

@WillieRuemmele Apologies. I didn't realize --output-dir was an option. It's not listed here: https://developer.salesforce.com/docs/atlas.en-us.sfdx_cli_reference.meta/sfdx_cli_reference/cli_reference_project_commands_unified.htm#cli_reference_project_deploy_start_unified

Nor is it listed in the command output:

Perhaps that's a docs bug? I'm currently testing adding --output-dir to my ci setup.

I understand this is likely a breaking change. is there any other solution than just updating the output of where it's stored?

[WillieRuemmele]

sorry, I mislead you, it's the --results-dir flag

[codefriar]

ah, then it's my bad for not having caught --results-dir I kept looking for an output option.

[mcarvin8]

@codefriar - I'm not affiliated with Salesforce, but this is consistent with what I've seen developing a plugin to convert the code coverage output for SonarQube and what I've listed in that plugin's readme (not to advertise it or suggest you use it - https://github.com/mcarvin8/apex-code-coverage-transformer).

The --results-dir flag is the one you use to specify a different directory besides the default one (which I believe defaults to the deploy ID). But this still creates an additional coverage sub-folder no matter what results-dir you specify.

I usually append this to my deploy command (my plugin converts the JSON output over the cobertura output but I think the directory logic works the same) - --coverage-formatters json --results-dir coverage.

This still creates the coverage file in ../coverage/coverage/coverage.json

I've tested json-summary and that gets added to the same folder path, just a different file name.

I think it would be nice if this extra sub-folder created by the CLI was removed since I've definitely had failed pipelines when passing the coverage file path to my plugin.