ci(windows): retry submodule fetch on transient github.com 500s

[Fedr]

Summary

Windows CI hit transient github.com submodule-clone failures, e.g. run 24845654039:

error: RPC failed; HTTP 500 curl 22 The requested URL returned error: 500
fatal: expected 'packfile'
fatal: clone of 'https://github.com/AcademySoftwareFoundation/openvdb' into submodule path 'thirdparty/openvdb/v9/openvdb' failed

Six different third-party submodules failed this way within ~25 s (openvdb, parallel-hashmap, tinygltf, tinyxml2, zlib-ng, openvdb/v10) — all HTTP 500s from github.com itself, not an infra issue on our side.

actions/checkout@v6 has its own retry logic and the log actually shows Failed to clone 'thirdparty/openvdb/v9/openvdb'. Retry scheduled — but it only retries each submodule once and the second attempt hit the same 500 while github.com was still flaky.

Change

Replace submodules: true in actions/checkout with an explicit retry loop in the next step:

- name: Checkout
  uses: actions/checkout@v6
  with:
    submodules: false

- name: Checkout submodules (with retries)
  shell: bash
  run: |
    for i in 1 2 3; do
      if git -c protocol.version=2 submodule update --init --force --depth=1; then
        break
      fi
      if [ "$i" = "3" ]; then
        echo "::error::submodule update failed after 3 attempts"
        exit 1
      fi
      echo "::warning::submodule update attempt $i failed — retrying in 30s"
      sleep 30
    done

Three attempts with a 30 s wait — enough to ride out the ~25 s github.com 500 spike observed in the failing run. --force recovers from any partial clone left by an earlier attempt. --depth=1 and -c protocol.version=2 match the flags actions/checkout itself uses internally.

Why not Wandalen/wretry.action

An earlier attempt on this PR wrapped the whole actions/checkout step in Wandalen/wretry.action@v3.8.0. It triggered a startup_failure — the workflow parser refused to schedule any job. Root cause appears to be that wretry.action's outer composite-action layer dispatches to an inner _js_action that runs on node20, but actions/checkout@v6 uses node24; the handoff doesn't work and the whole workflow is rejected before any step runs. wretry.action is thinly maintained and has an open issue (#193) with no ETA for a fix, so a simple in-workflow retry is the right trade.

Scope

Only build-test-windows.yml's checkout area. Other platforms' workflows aren't touched; disable-build-* labels suppress non-windows CI in this PR.

Test plan

• Windows CI on this PR completes normally (attempt 1 succeeds; warnings silent).
• If a future CI run hits a github.com 500 during submodule fetch, retries kick in and the job still succeeds.