add result output

[thattommyhall]

I am using this to do chat-ops with peter-evans/{slash-command-dispatch|create-or-update-comment}

      - name: Apply
        id: apply
        uses: ./.github/actions/octodns-sync
        env:
          DNSIMPLE_TOKEN:  ${{ secrets.DNSIMPLE_TOKEN }}
        with:
          config_path: config.yaml

      - name: Add reaction
        uses: peter-evans/create-or-update-comment@v1
        with:
          token: ${{ secrets.PAT }}
          repository: ${{ github.event.client_payload.github.payload.repository.full_name }}
          issue-number: ${{ github.event.client_payload.github.payload.issue.number }}
          body: |
            DID IT!

            ```
            ${{ steps.apply.outputs.result }}
            ```

[solvaholic]

Here, rather than route the output through an environment variable will it work to store it in a file? For example:

octodns-sync --config-file="${_config_path}" --log-stream-stdout | tee /tmp/octodns-sync.log

Then you could use the Setting the comment body from a file approach to use the output in the next step of your job.

This way octodns-sync would still output directly for the Actions log and the unmodified output would remain available for other uses in addition to this specific case with peter-evans' Actions.

I'll setup a test for this later today or tomorrow. If you happen to check it in the meantime, @thattommyhall, I'll be curious to hear what you find.

[solvaholic]

🎉

@thattommyhall to illustrate that ☝️ I made an example in https://github.com/solvaholic/ideal-octo-pancake.

I pushed a new version of the solvaholic/octodns-sync container to Docker hub that writes the octodns-sync output to a file:

# Run octodns-sync.
echo "INFO: _config_path: ${_config_path}"
if [ "${_doit}" = "--doit" ]; then
  octodns-sync --config-file="${_config_path}" \
  --log-stream-stdout --doit | tee ${GITHUB_WORKSPACE}/octodns-sync.log
else
  octodns-sync --config-file="${_config_path}" \
  --log-stream-stdout | tee ${GITHUB_WORKSPACE}/octodns-sync.log
fi

And added a step to the workflow job to parse the output from the file:

      - id: get-comment-body
        run: |
          body=$(cat ${GITHUB_WORKSPACE}/octodns-sync.log)
          body="${body//'%'/'%25'}"
          body="${body//$'\n'/'%0A'}"
          body="${body//$'\r'/'%0D'}" 
          echo ::set-output name=body::$body

The automation added this: https://github.com/solvaholic/ideal-octo-pancake/pull/1#issuecomment-706599506

Would that approach meet your needs?

[solvaholic]

In #24 I switched this to use screen rather than tee, and preserve the exit code.

@thattommyhall I'll mark this pull request closed. Please do re-open if the log file won't meet your needs or we've more to discuss.

Thank you 🙇