Question: how should analysis-result be used?

[pauldreik]

Thanks for writing this action! I got it running, but now I want to store the results as an artifact.
I don't understand how the analysis-result is to be interpreted. Should it be set to a filename where to write the results? How do I get the results into a file I can store as a github action artifact?

On a side note. Is it possible to setup some limit, like failing the job if the binary size increases by more than x bytes (or percent)?

[djarek]

analysis-result contains the raw string of the result. You should just be able to useit, just like any workflow variable, with ${{ analysis-result }}. I planned on using the output in boost.json's CI, but I ended up not using it. If it doesn't work let me know, I'll try to fix it.

As for setting a hard limit - it might make sense for a constrained embedded system (you obviously have a hard cap on flash size), but I'd avoid doing it in other projects. I think it's hard to make a heuristic that determines whether the size increase is warranted.

Oh, btw, Bloaty has mutliple output formats so if you want to parse the results with something else, I can add an option to use that flag. I didn't include it, because I didn't have a need for it so far. CSV might be easier to parse if you end up deciding to include a size-limiting job.

[pauldreik]

Ok, how do I get the contents of that variable into a file? I tried

      - name: "store results"
        run: |
          echo -E ${{ analysis-result }} > bloaty-sizes.txt

But it gives me this error:

The workflow is not valid. .github/workflows/bloaty.yml (Line: 29, Col: 14): Unrecognized named-value: 'analysis-result'. Located at position 1 within expression: analysis-result,.github/workflows/bloaty.yml (Line: 50, Col: 14): Unrecognized named-value: 'analysis-result'. Located at position 1 within expression: analysis-result

I suspect that variable "ends it's scope" when the bloaty action ends?

What is the best way for me to get the result into bloaty-sizes.txt so I can store it as an artifact?

[djarek]

https://docs.github.com/en/free-pro-team@latest/actions/reference/workflow-syntax-for-github-actions#jobsjob_idoutputs

This suggests that workflow variables are scoped, so in order to refer to the variable, you need to use the id: ${{ my-step-id.analysis-result }}, where the id is defined in the bloaty step, e.g.:

    - name: Bloat analysis
        id: my-step-id
        uses: djarek/bloaty-analyze@v1
        with:
            file: build/mylib.a

[pauldreik]

Thanks for the reference! I made progress. However, I think the action only assigns the first row to the variable. I get

    FILE SIZE        VM SIZE    

from the analysis-results variable. If I look in the log output during the action, the body of the table seems to be written to the terminal rather than captured by the variable.

I do not know if you can access it, but here is a link to the output: https://github.com/simdjson/simdjson/runs/1240273039?check_suite_focus=true and here is the workflow: https://github.com/pauldreik/simdjson/blob/pauldreik/bloaty/.github/workflows/bloaty.yml

[djarek]

Hmm, I didn't expect this behavior. I'll consider it to be a bug and I'll fix it.

[pauldreik]

Thanks. While we are at it, may I wish for an optional parameter specifying a file which the results are written to?

[pauldreik]

@djarek friendly reminder, do you intend to fix this in the near future?