v2.2.0

🚀 Release v2.2.0

👍 Added

• max_threads input in action.yml to control concurrent translation calls.
• Call‑out block exclusion: lines beginning with > [!…] (e.g. > [!WARNING] …) are now automatically skipped (preserved verbatim) rather than run through the translator.
• Regex‐cleanup step in Processor.build_file to strip stray spaces inside **bold** and __italic__ markers after placeholder restoration.

🛠 Changed

• Action metadata (action.yml)

  • Switched fully to inputs/with syntax.
  • Added max_threads (default: 5) alongside existing files, langs, debug and max_line_length.

• Dockerfile

  • Pinned to python:3.9.16‑slim.
  • Runs under a non‑root appuser, copies the entire core/ tree, and invokes core/main.py.

• core/main.py

  • Now parses exactly five sys.argv parameters (files, langs, debug, max_threads, max_line_length).
  • Dynamically adjusts logging level when debug: 'true'.

• Concurrency model

  • Switched translation calls to use an asyncio.Semaphore(max_threads) instead of default fully‐parallel gathers, preventing API‑quota bursts.

• Placeholder handling (processor.py)

  • Always store “original” text as a string (not tuple) to avoid the TypeError in _restore_placeholders.
  • Introduced a new _MD_BLOCK_NOTE placeholder type for skipped call‑outs.

🐛 Fixed

• Markdown‐element spacing: fixed bug where **test** could become ** test ** after placeholder round‑trip.
• Placeholder restore: corrected replacement logic so every placeholder is properly re‑inserted as its original string.

🚫 Removed

• CPU‐process pool sketch from v2.2.0 pre‑releases—decided to preserve the existing escape/placeholder stability rather than introduce a max_processes fork.