[Copilot]

The deploy job in the Pages workflow failed before build/deploy due to an unresolved GitHub Action reference (Swatinem/rust-cache@main). This updates the workflow to use a resolvable, stable action version so the deploy pipeline can proceed.

• Root cause

  • pages.yml referenced Swatinem/rust-cache@main, which is not resolvable in GitHub Actions for this workflow run.

• Workflow fix

  • Updated the Pages deploy job cache step to pin Swatinem/rust-cache to @v2.
  • Kept existing workspace cache layout unchanged (. -> target, wasm -> wasm/target) to preserve current behavior.

• Example change

  - name: Cache cargo
    uses: Swatinem/rust-cache@v2
    with:
      workspaces: |
        . -> target
        wasm -> wasm/target

[wheevu]

Closing as superseded. The diagnosis was right: Swatinem/rust-cache@main does not resolve, and @v2 is what ci.yml already used.

The same fix reached main as c309fb3, together with two further fixes the deploy needed, both of which only appeared once the workflow ran for the first time:

• The artifact check read the WebAssembly magic with od -An -c | grep, which can never match. GNU od prints \0 where BSD od prints 0, and the pattern expected six bytes for a four-byte magic. It now reads hex.
• Staging copied all of site/ and deleted only the README afterwards, so it would have published the new browser tests, the npm manifest and the Playwright config to the public URL. It is now an allowlist of the thirteen page files.

main is green, Pages deployed from 112800e, and the page is live at https://wheevu.github.io/muninn/.