[WIP] Fix run-all.bat and Docker build issues in NPU-STACK

[Copilot]

Thanks for asking me to work on this. I will get started on it and keep this PR's description up to date as I form a plan and make progress.

Original prompt

User reports new breakages in chainchopper/NPU-STACK after running run-all.bat on Windows and Docker build.

Observed issues:

1. run-all.bat prints:
   Backend: http://localhost:8000
   'Frontend:' is not recognized as an internal or external command
   This indicates a batch syntax/quoting/line-continuation bug where literal label text is being interpreted as a command.

2. Docker backend build fails in Dockerfile llama-builder stage because it still builds llama.cpp with make:
   make -j$(nproc) llama-quantize llama-gguf-split llama-imatrix llama-cli llama-server llama-export-lora
   Upstream llama.cpp changed build system to CMake; Makefile now errors with:
   "Build system changed: The Makefile build has been replaced by CMake."

Please update the repository to fix both problems.

Required changes:

• Locate and fix run-all.bat so frontend startup message and command execution are valid on Windows cmd.

  • Ensure labels/messages are emitted with echo and not executed as commands.
  • Keep behavior of launching backend + frontend intact.

• Update Dockerfile(s) for backend llama tools build:

  • Replace deprecated make-based llama.cpp build with CMake-based build per upstream docs.
  • Build required binaries: llama-quantize, llama-gguf-split, llama-imatrix, llama-cli, llama-server, llama-export-lora (or current equivalent targets if names changed upstream).
  • Make build robust to upstream target renames where practical and fail with clear error when missing.

• Verify docker compose build/start path works after changes.

• Update docs/README quick-start notes if commands or build expectations changed.

Acceptance criteria:

• Running run-all.bat no longer emits "'Frontend:' is not recognized...".
• docker compose up --build no longer fails at llama.cpp make step; uses CMake successfully.
• Required llama tooling is available in resulting image (or documented equivalents).
• Changes are committed in a PR against main in chainchopper/NPU-STACK.

This pull request was created from Copilot chat.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[Copilot]

Copilot wasn't able to review any files in this pull request.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.