Parsing XML

The Challenge of RSS Parsing

RSS parsing sounds easy. It’s just XML. In practice, it’s surprisingly tricky:

• Multiple formats: RSS 0.91, 1.0, 2.0 and Atom 1.0 are all in use
• Non-standard extensions: Content namespaces, Dublin Core, Media RSS, etc.
• Malformed XML: Many publishers have invalid HTML inside <description> tags
• Encoding issues: Some feeds incorrectly declare their encoding

SparkFeed handles all of these cases.

The Parser

SparkFeed uses a custom parser built on top of Node.js’s fast-xml-parser library:

// src/server/lib/rss-parser.ts
import { XMLParser } from 'fast-xml-parser';

const parser = new XMLParser({
  ignoreAttributes: false,
  attributeNamePrefix: '@_',
  parseAttributeValue: false,
  trimValues: true,
  parseTagValue: false,
});

RSS 2.0 Parsing

function parseRSS2(xml: unknown): ParsedFeed {
  const channel = xml?.rss?.channel;

  return {
    title: extractText(channel?.title),
    url: extractText(channel?.link),
    items: (channel?.item ?? []).map((item: unknown) => ({
      guid:        extractText(item?.guid) ?? extractText(item?.link),
      title:       extractText(item?.title),
      url:         extractText(item?.link),
      content:     extractText(item?.['content:encoded']) ?? extractText(item?.description),
      summary:     stripHtml(extractText(item?.description) ?? ''),
      author:      extractText(item?.author) ?? extractText(item?.['dc:creator']),
      publishedAt: parseDate(extractText(item?.pubDate)),
    })),
  };
}

Atom 1.0 Parsing

function parseAtom(xml: unknown): ParsedFeed {
  const feed = xml?.feed;

  return {
    title: extractText(feed?.title),
    url:   extractText(feed?.link?.['@_href']) ?? extractText(feed?.id),
    items: (feed?.entry ?? []).map((entry: unknown) => ({
      guid:        extractText(entry?.id),
      title:       extractText(entry?.title),
      url:         extractLink(entry?.link),
      content:     extractText(entry?.content) ?? extractText(entry?.summary),
      summary:     stripHtml(extractText(entry?.summary) ?? ''),
      author:      extractText(entry?.author?.name),
      publishedAt: parseDate(extractText(entry?.published) ?? extractText(entry?.updated)),
    })),
  };
}

HTML Sanitization

Feed content often contains HTML. SparkFeed sanitizes it before storing to prevent XSS:

import DOMPurify from 'isomorphic-dompurify';

const ALLOWED_TAGS = ['p', 'h1', 'h2', 'h3', 'h4', 'ul', 'ol', 'li',
                      'blockquote', 'code', 'pre', 'a', 'strong', 'em',
                      'img', 'figure', 'figcaption', 'hr', 'br'];

export function sanitizeHtml(dirty: string): string {
  return DOMPurify.sanitize(dirty, {
    ALLOWED_TAGS,
    ALLOWED_ATTR: ['href', 'src', 'alt', 'title', 'class'],
    FORCE_BODY: true,
  });
}

This strips scripts, iframes, and other dangerous elements while preserving the article’s visual structure.

Auto-Detecting Feed Type

Before parsing, SparkFeed detects whether a feed is RSS or Atom:

export function detectFeedType(xml: unknown): 'rss' | 'atom' | 'unknown' {
  if (xml?.rss) return 'rss';
  if (xml?.feed?.['@_xmlns'] === 'http://www.w3.org/2005/Atom') return 'atom';
  if (xml?.feed) return 'atom';
  return 'unknown';
}