feat: support loading multiple markdown sources via sourceUrls option

[dpecos]

Summary

• Adds sourceUrls: string[] to SlideshowOptions
• Fetches all URLs in parallel, concatenates results in order (joined with \n---\n slide separators), then parses as a single deck
• Exposes loadFromUrls(urls, callback) as a public method alongside the existing loadFromUrl / loadFromString
• When both sourceUrls and sourceUrl are provided, sourceUrls takes precedence
• 5 new tests covering: slide count, content preservation, public API, and precedence

API

mdeck.createSlideshow({
  sourceUrls: ['intro.md', 'chapter1.md', 'chapter2.md', 'outro.md']
});

// or imperatively:
slideshow.loadFromUrls(['part1.md', 'part2.md'], (ss) => console.log(ss.getSlideCount()));

Implementation notes

• Each URL is fetched via the existing XMLHttpRequest abstraction (same as loadFromUrl), wrapped in a Promise
• Promise.all fetches all concurrently; order is preserved by the array join
• Files are joined with \n---\n so each file's content starts as a new slide, even if the file itself doesn't end with a separator

Closes #7