[CI] Improve CI speed by optimizing Docker cache in code-quality.yml

[denver79]

The Problem:
In the .github/workflows/code-quality.yml workflow, the clang-tidy job builds a Docker image in two stages: base and code_quality. While the base stage is correctly built and its layers are saved to the GitHub Actions cache (using cache-to), the subsequent code_quality stage only reads from the cache (cache-from) but does not write its own resulting layers back to the cache.

This results in the code_quality stage being rebuilt from scratch on every single run. This leads to significantly longer CI times and inefficient use of resources.

Proposed Solution:

By adding the cache-to parameter to the code_quality build step, we can ensure its layers are also cached. This will dramatically speed up subsequent runs of the workflow. The cache-from should also be updated to look for its own cache first.

Current Code:

• name: Build code_quality stage
  uses: docker/build-push-action@v5
  with:
  context: .
  file: docker/Dockerfile
  target: code_quality
  cache-from: type=gha,scope=base

Suggested Change:

• name: Build code_quality stage
  uses: docker/build-push-action@v5
  with:
  context: .
  file: docker/Dockerfile
  target: code_quality

  Attempt to read from both 'code_quality' and 'base' caches

  cache-from: |
  type=gha,scope=code_quality
  type=gha,scope=base

  Write the resulting layers to the 'code_quality' cache scope

  cache-to: type=gha,scope=code_quality,mode=max
  This simple change will make the code quality checks run much faster, improving the overall developer experience and efficiency.