Multiline footnote definitions were being reduced to empty/near-empty entries in the rendered footnote list (showing only [n] ←), with continuation paragraphs dropped after blank lines. This PR updates the footnote extraction/render path so full multiline bodies (including inline markdown) are preserved.

• Footnote parser: continuation block handling

  • Added explicit continuation-line detection and indent normalization for footnote definitions.
  • Continued parsing across blank lines when followed by indented continuation content, so multi-paragraph bodies stay attached to the same [^id]: definition.

• Footnote renderer: robust paragraph/body handling

  • Normalized paragraph splitting for mixed newline styles (LF/CRLF).
  • Avoided emitting empty footnote entries by storing/rendering only non-empty extracted definition content.
  • Preserved inline back-reference placement at the end of rendered footnote content.

• Codebase consistency

  • Applied identical logic in both mirrored frontend scripts:
    • script.js
    • desktop-app/resources/js/script.js

function isFootnoteContinuationLine(line, baseIndent) {
  if (!line.startsWith(baseIndent)) return false;
  const lineAfterBase = line.slice(baseIndent.length);
  return /^(?: {2,}|\t)/.test(lineAfterBase);
}

function stripFootnoteContinuationIndent(line, baseIndent) {
  const lineAfterBase = line.slice(baseIndent.length);
  if (lineAfterBase.startsWith("\t")) return lineAfterBase.slice(1);
  return lineAfterBase.replace(/^ {2}/, "");
}