[Question/Idea] Is it possible to use gh-page data.js content as cached value?

[RaulTrombin]

I would like to compare the PR's benchmark with the master one,

My ideia is: instead use a cache, to try something with:

      - name: Get previous benchmark data
        run: |
          echo "Fetching gh-pages branch"
          git fetch origin gh-pages

          echo "Checking out gh-pages branch"
          git checkout gh-pages

          if [ ! -d "./cache" ]; then
            echo "Cache folder does not exist, creating it"
            mkdir ./cache
          fi

          echo "Copying data file from gh-pages to cache"
          cp ./dev/bench/data.js ./cache/data.js || { echo "Failed to copy data file" ; exit 1; }

          echo "Checking out main branch"
          git checkout -

an then run the benchmark action with minimal setup, using this content as the json external file.

[ktrz]

Hi @RaulTrombin

I'm sorry for taking so long to get back to you.

I don't quite understand what you want to achieve. Could you please elaborate a bit more on what is your goal?

[RaulTrombin]

@ktrz Hey,
I can't remember exatcly what was the issue,
But I think it was the cache, it could be erased with long times without using the actions,
then I've used the gh-page to store my previous results.
So the idea proposed was to use gh-page to store cache.

This is how I'm using the actions for a Rust library running some drivers methods that read magnetometers/accelerometers/sensors...things.
https://github.com/bluerobotics/navigator-rs/blob/4425d95be16c786f3a562dbc2ed6043727804917/.github/workflows/action.yml#L60