Corrige comportamento não-determinístico do atualizador de firmware (3rd/update_firmware.py)

[bmorcelli]

Motivation

• O script estava removendo todas as entradas de um par github+author em cada JSON e somente as re-inseria quando encontrava novas releases, causando desaparecimento/aparecimento aleatório de firmwares entre execuções.

Description

• Limitei a remoção/reconstrução para os fids esperados do JSON atual calculados com generate_fid, armazenados em expected_fids.
• Passei a filtrar existing_entries apenas por fid presente em expected_fids além de github e author, evitando apagar entradas de outros firmwares no mesmo arquivo.
• Removo apenas entradas cujo fid esteja em expected_fids e reescrevo-as combinando versões existentes com new_versions, preservando histórico quando não há novas releases.
• Ajustei logs para diferenciar "Nenhuma versão disponível" de "Nenhuma versão nova" e mantive o corte para as últimas 10 versões.

Testing

• Verificação de sintaxe com python -m py_compile 3rd/update_firmware.py executada com sucesso.

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 90eb068372

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Avoid duplicate name entries in t-deck firmware list

Adding another entry with name: "T-Deck" causes the downstream generator to silently overwrite other T-Deck firmwares because 3rd/create_json.py builds new_map as {item["name"].strip(): item} (keyed only by name). In this commit that behavior already changed 3rd/r/t-deck.json from Bruce's T-Deck to Ratdeck, so users consuming the /r feeds get a different firmware than before even though both should coexist; this is a data regression tied to duplicate names in the source list.

Useful? React with 👍 / 👎.