Fix CI failure: escape inline <pre data-test="exec"> in linguistics-asia READMEs

[Copilot]

The main / test CI job was crashing with ERROR: expected string or bytes-like object, got 'NoneType' when processing the new examples/linguistics-asia/ READMEs.

Root cause: The test framework (test/test.py) parses raw Markdown as HTML via BeautifulSoup. All 8 linguistics-asia READMEs contain `<pre data-test="exec">` in their prose description of the testing convention. BeautifulSoup interprets this as a real HTML tag, creating a phantom <pre data-test="exec"> element that wraps the subsequent text and actual test blocks. This element has multiple children → tag.string is None → re.sub(pattern, value, None) throws TypeError.

Changes

• 8 linguistics-asia/0*/README.md files: Escape the inline prose reference from <pre data-test="exec"> to <pre data-test="exec"> so BeautifulSoup doesn't create a phantom element.

• test/test.py: Add a None guard in parse_cmds() to fail gracefully rather than crash if tag.string is ever None:

  def parse_cmds(pre, attrs):
      if pre is None:
          return []
      ...