Print an RSS pubDate as a date, not its first ten characters

src/distill.js

@@ -724,7 +724,7 @@ export function feedToHTML(text) {
       || rssLink(entry)
       || (guidIsPermalink ? field(entry, 'guid') : '');
     const author = field(entry, 'author name') || field(entry, 'author');
-    const date = (field(entry, 'updated') || field(entry, 'published') || field(entry, 'pubdate')).slice(0, 10);
+    const date = feedDate(field(entry, 'updated') || field(entry, 'published') || field(entry, 'pubdate'));
     const byline = [author && `by ${author}`, date].filter(Boolean).join(', ');
     // Atom escapes the entry body, so textContent of content/summary is the
     // HTML itself, ready to be embedded and parsed like any page.
@@ -752,6 +752,27 @@ export function feedToHTML(text) {
   return `<html><head><title>${esc(feedTitle)}</title></head><body>\n${parts.join('\n')}\n</body></html>`;
 }
 
+const MONTHS = ['jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug', 'sep', 'oct', 'nov', 'dec'];
+const RFC822_DATE = /^(?:[a-z]{3},\s*)?(\d{1,2})\s+([a-z]{3})\s+(\d{4})\b/i;
+
+/**
+ * An entry's date as YYYY-MM-DD. Atom writes ISO 8601, so the day is its
+ * first ten characters. RSS 2.0 writes RFC 822 ('Mon, 17 Aug 2026 00:00:00
+ * GMT'), and ten characters of that are 'Mon, 17 Au'. Its day, month, and
+ * year are read as written rather than through Date, which would move a late
+ * evening post to the next day in UTC; the Atom slice keeps the feed's own
+ * day too. Anything else is printed as the feed wrote it, not cut.
+ * @param {string} s
+ * @returns {string}
+ */
+function feedDate(s) {
+  if (/^\d{4}-\d\d-\d\d/.test(s)) return s.slice(0, 10);
+  const m = s.match(RFC822_DATE);
+  const month = m ? MONTHS.indexOf(m[2].toLowerCase()) + 1 : 0;
+  if (!month) return s;
+  return `${m[3]}-${String(month).padStart(2, '0')}-${m[1].padStart(2, '0')}`;
+}
+
 const escHTML = (s) => String(s ?? '').replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/"/g, '&quot;');
 
 /**

tests/distill.test.js

@@ -267,6 +267,27 @@ test('an RSS 2.0 entry is opened by its link, not a non-permalink guid', () => {
   assert.ok(text.includes('Second lede.'), 'item without content:encoded lost its description');
 });
 
+test('an RSS 2.0 pubDate prints as a date, not as its first ten characters', () => {
+  // pubDate is RFC 822, the day name first. Cutting it where an Atom
+  // timestamp is cut printed "Mon, 17 Au" under every WordPress post.
+  const rss = readFileSync(new URL('./pages/rss20.xml', import.meta.url), 'utf8');
+  const lines = distill(rss, 'https://blog.example.test/feed.xml').blocks.map((b) => b.text);
+  assert.ok(lines.includes('2026-08-17'), `pubDate not read as a date: ${JSON.stringify(lines)}`);
+  assert.ok(lines.includes('2026-08-18'), 'second item lost its date');
+  assert.ok(!lines.some((l) => /\bAu$/.test(l)), 'a date was cut mid-month');
+  // The day is the one the feed wrote, in its own zone, the way the slice of
+  // an Atom timestamp already reads, not the UTC day it converts to.
+  const late = `<?xml version="1.0"?><rss version="2.0"><channel><title>Late</title>
+    <item><title>Late post</title><link>https://example.test/late</link>
+    <pubDate>Sat, 5 Sep 2026 23:30:00 -0700</pubDate><description>Body text here.</description></item>
+    </channel></rss>`;
+  assert.ok(distill(late, 'https://example.test/rss').blocks.some((b) => b.text === '2026-09-05'),
+    'a single-digit day or a zone offset threw the date off');
+  // A date in neither form is printed as the feed wrote it, rather than cut.
+  const odd = late.replace('Sat, 5 Sep 2026 23:30:00 -0700', 'yesterday');
+  assert.ok(distill(odd, 'https://example.test/rss').blocks.some((b) => b.text === 'yesterday'));
+});
+
 test('a youtube watch page renders as title, byline, description, and transcript links', () => {
   const watch = readFileSync(new URL('./pages/youtube_watch.html', import.meta.url), 'utf8');
   const p = distill(watch, 'https://www.youtube.com/watch?v=fixture123');