Merge pull request #3 from pupperpowell/rss-feed-fixes

fix(rss): stop future-verse leaks, correct timezone dates, fix feed discovery

src/app.html

@@ -4,6 +4,7 @@
 		<meta charset="utf-8" />
 		<meta name="viewport" content="width=device-width, initial-scale=1, viewport-fit=cover" />
 			<link rel="icon" href="/favicon.png" type="image/png" />
+		<link rel="alternate" type="application/rss+xml" title="Bibdle RSS Feed" href="/feed.xml" />
 		%sveltekit.head%
 	</head>
 	<body data-sveltekit-preload-data="hover">

src/lib/server/rss.ts

@@ -0,0 +1,165 @@
+import type { DailyVerse } from '$lib/server/db/schema';
+
+/**
+ * The feed presents calendar days, so every timestamp is anchored to midnight
+ * in this reference timezone. Using a fixed zone (rather than the server's
+ * local timezone) keeps output stable across UTC dev machines and NY hosts.
+ */
+export const FEED_TIMEZONE = 'America/New_York';
+
+/** Number of most-recent verses exposed in the feed. */
+export const FEED_ITEM_LIMIT = 30;
+
+/** Escape the five XML predefined entities. */
+export function escapeXml(text: string): string {
+	return text
+		.replace(/&/g, '&')
+		.replace(/</g, '&lt;')
+		.replace(/>/g, '&gt;')
+		.replace(/"/g, '&quot;')
+		.replace(/'/g, '&apos;');
+}
+
+/**
+ * The current calendar date (`YYYY-MM-DD`) in the feed's reference timezone.
+ * Used to cap the feed so future/scheduled verses are never published early.
+ */
+export function todayInTimeZone(timeZone: string = FEED_TIMEZONE, now: Date = new Date()): string {
+	const parts = new Intl.DateTimeFormat('en-US', {
+		timeZone,
+		year: 'numeric',
+		month: '2-digit',
+		day: '2-digit'
+	}).formatToParts(now);
+	const get = (type: string) => parts.find((part) => part.type === type)?.value ?? '';
+	return `${get('year')}-${get('month')}-${get('day')}`;
+}
+
+/**
+ * The UTC offset in effect on a given `YYYY-MM-DD` for a timezone, formatted
+ * as `±HH:MM` (e.g. `-04:00` during EDT, `-05:00` during EST).
+ */
+export function timeZoneOffset(dateStr: string, timeZone: string = FEED_TIMEZONE): string {
+	const [year, month, day] = dateStr.split('-').map(Number);
+	// Noon UTC is safely inside the calendar day in any US offset, so DST is
+	// resolved correctly regardless of the time of day.
+	const reference = new Date(Date.UTC(year, month - 1, day, 12));
+	const name = new Intl.DateTimeFormat('en-US', {
+		timeZone,
+		timeZoneName: 'longOffset'
+	})
+		.formatToParts(reference)
+		.find((part) => part.type === 'timeZoneName')?.value;
+
+	const offset = (name ?? 'GMT+00:00').replace('GMT', '');
+	return offset === '' ? '+00:00' : offset;
+}
+
+/**
+ * Format a `YYYY-MM-DD` day as an RFC 822 / RFC 2822 timestamp at midnight in
+ * the feed's reference timezone. Always emits a valid numeric/GMT timestamp
+ * (never a bare, ambiguous abbreviation like `EST`).
+ */
+export function formatRFC822(dateStr: string, timeZone: string = FEED_TIMEZONE): string {
+	const offset = timeZoneOffset(dateStr, timeZone);
+	return new Date(`${dateStr}T00:00:00${offset}`).toUTCString();
+}
+
+/** Human-readable date that does not shift with the server's local timezone. */
+export function formatReadableDate(dateStr: string): string {
+	const [year, month, day] = dateStr.split('-').map(Number);
+	return new Date(Date.UTC(year, month - 1, day)).toLocaleDateString('en-US', {
+		month: 'long',
+		day: 'numeric',
+		year: 'numeric',
+		timeZone: 'UTC'
+	});
+}
+
+/** Wrap text in a CDATA section, splitting any embedded `]]>` safely. */
+export function cdata(text: string): string {
+	return `<![CDATA[${text.replace(/]]>/g, ']]]]><![CDATA[>')}]]>`;
+}
+
+// Helper: Format verse text (VerseDisplay + Imposter unbalanced punctuation handling)
+export function formatVerseText(text: string): string {
+	let formatted = text;
+
+	// Handle unbalanced opening/closing punctuation (from Imposter.svelte)
+	const pairs: [string, string][] = [
+		['(', ')'],
+		['[', ']'],
+		['{', '}'],
+		['"', '"'],
+		["'", "'"],
+		['\u201C', '\u201D'], // " "
+		['\u2018', '\u2019'] // ' '
+	];
+
+	// Check if text starts with opening punctuation without closing
+	for (const [open, close] of pairs) {
+		if (formatted.startsWith(open) && !formatted.includes(close)) {
+			formatted += '...' + close;
+			break;
+		}
+	}
+
+	// Check if text ends with closing punctuation without opening
+	for (const [open, close] of pairs) {
+		if (formatted.endsWith(close) && !formatted.includes(open)) {
+			formatted = open + '...' + formatted;
+			break;
+		}
+	}
+
+	// Check if text contains unbalanced opening quotes (not at start) without closing
+	for (const [open, close] of pairs) {
+		const openCount = (formatted.match(new RegExp(`\\${open}`, 'g')) || []).length;
+		const closeCount = (formatted.match(new RegExp(`\\${close}`, 'g')) || []).length;
+		if (openCount > closeCount) {
+			formatted += close;
+			break;
+		}
+	}
+
+	// Capitalize first letter if lowercase (from VerseDisplay.svelte)
+	formatted = formatted.replace(/^([a-z])/, (c) => c.toUpperCase());
+
+	// Replace trailing punctuation with ellipsis
+	// Preserve closing quotes/brackets that may have been added
+	formatted = formatted.replace(/[,:;-—]([)\]}\"\'\u201D\u2019]*)$/, '...$1');
+
+	return formatted;
+}
+
+/**
+ * Content fingerprint used as a strong ETag. Any change to the included rows
+ * (new day, edited verse text, changed reference) produces a new hash, so
+ * cached clients revalidate correctly.
+ */
+export function feedFingerprint(
+	verses: Pick<DailyVerse, 'date' | 'reference' | 'verseText'>[]
+): string {
+	let hash = 5381;
+	for (const verse of verses) {
+		const row = `${verse.date}\u0000${verse.reference}\u0000${verse.verseText}`;
+		for (let i = 0; i < row.length; i++) {
+			hash = ((hash << 5) + hash + row.charCodeAt(i)) | 0;
+		}
+	}
+	return (hash >>> 0).toString(16);
+}
+
+/**
+ * The most recent `createdAt` across the given rows, used for `Last-Modified`.
+ * Falls back to `fallback` when no row carries a timestamp.
+ */
+export function latestModification(
+	verses: Pick<DailyVerse, 'createdAt'>[],
+	fallback: Date
+): Date {
+	return verses.reduce<Date | null>((latest, verse) => {
+		if (!verse.createdAt) return latest;
+		return !latest || verse.createdAt > latest ? verse.createdAt : latest;
+	}, null) ?? fallback;
+}

src/routes/+layout.svelte

@@ -31,12 +31,6 @@
 
 <svelte:head>
 	<link rel="icon" href={favicon} />
-	<link
-		rel="alternate"
-		type="application/rss+xml"
-		title="Bibdle RSS Feed"
-		href="/feed.xml"
-	/>
 	<meta name="description" content="A Wordle-inspired daily Bible game" />
 </svelte:head>
 

src/routes/feed.xml/+server.ts

@@ -1,133 +1,99 @@
 import type { RequestHandler } from './$types';
 import { db } from '$lib/server/db';
 import { dailyVerses } from '$lib/server/db/schema';
-import { desc } from 'drizzle-orm';
-
-// Helper: Escape XML special characters
-function escapeXml(text: string): string {
-	return text
-		.replace(/&/g, '&')
-		.replace(/</g, '&lt;')
-		.replace(/>/g, '&gt;')
-		.replace(/"/g, '&quot;')
-		.replace(/'/g, '&apos;');
-}
-
-// Helper: Format YYYY-MM-DD to RFC 822 date string
-function formatRFC822(dateStr: string): string {
-	// Parse date in America/New_York timezone (EST/EDT)
-	// Assuming midnight ET
-	const date = new Date(dateStr + 'T00:00:00-05:00');
-	return date.toUTCString().replace('GMT', 'EST');
-}
-
-// Helper: Format YYYY-MM-DD to readable date
-function formatReadableDate(dateStr: string): string {
-	const date = new Date(dateStr + 'T00:00:00');
-	return date.toLocaleDateString('en-US', {
-		month: 'long',
-		day: 'numeric',
-		year: 'numeric',
-		timeZone: 'America/New_York'
-	});
-}
-
-// Helper: Format verse text (VerseDisplay + Imposter unbalanced punctuation handling)
-function formatVerseText(text: string): string {
-	let formatted = text;
-
-	// Handle unbalanced opening/closing punctuation (from Imposter.svelte)
-	const pairs: [string, string][] = [
-		['(', ')'],
-		['[', ']'],
-		['{', '}'],
-		['"', '"'],
-		["'", "'"],
-		['\u201C', '\u201D'], // " "
-		['\u2018', '\u2019'] // ' '
-	];
-
-	// Check if text starts with opening punctuation without closing
-	for (const [open, close] of pairs) {
-		if (formatted.startsWith(open) && !formatted.includes(close)) {
-			formatted += '...' + close;
-			break;
-		}
-	}
-
-	// Check if text ends with closing punctuation without opening
-	for (const [open, close] of pairs) {
-		if (formatted.endsWith(close) && !formatted.includes(open)) {
-			formatted = open + '...' + formatted;
-			break;
-		}
-	}
-
-	// Check if text contains unbalanced opening quotes (not at start) without closing
-	for (const [open, close] of pairs) {
-		const openCount = (formatted.match(new RegExp(`\\${open}`, 'g')) || []).length;
-		const closeCount = (formatted.match(new RegExp(`\\${close}`, 'g')) || []).length;
-		if (openCount > closeCount) {
-			formatted += close;
-			break;
-		}
-	}
-
-	// Capitalize first letter if lowercase (from VerseDisplay.svelte)
-	formatted = formatted.replace(/^([a-z])/, (c) => c.toUpperCase());
-
-	// Replace trailing punctuation with ellipsis
-	// Preserve closing quotes/brackets that may have been added
-	formatted = formatted.replace(/[,:;-—]([)\]}\"\'\u201D\u2019]*)$/, '...$1');
-
-	return formatted;
-}
+import { desc, lte } from 'drizzle-orm';
+import {
+	FEED_ITEM_LIMIT,
+	cdata,
+	escapeXml,
+	feedFingerprint,
+	formatRFC822,
+	formatReadableDate,
+	formatVerseText,
+	latestModification,
+	todayInTimeZone
+} from '$lib/server/rss';
 
 export const GET: RequestHandler = async ({ request }) => {
 	try {
-		// Query last 30 verses, ordered by date descending
+		const SITE_URL = process.env.PUBLIC_SITE_URL || 'https://bibdle.com';
+		const FEED_URL = `${SITE_URL}/feed.xml`;
+		const today = todayInTimeZone();
+
+		// Only publish verses whose calendar day has arrived. Pre-scheduled
+		// future verses (from the submission scheduling scan) and any stray
+		// future-dated rows must never leak into the feed early.
 		const verses = await db
 			.select()
 			.from(dailyVerses)
+			.where(lte(dailyVerses.date, today))
 			.orderBy(desc(dailyVerses.date))
-			.limit(30);
+			.limit(FEED_ITEM_LIMIT);
 
-		// Generate ETag based on latest verse date
-		const etag = verses[0]?.date ? `"bibdle-feed-${verses[0].date}"` : '"bibdle-feed-empty"';
+		// Content-addressed ETag + Last-Modified let aggregators revalidate
+		// cheaply with If-None-Match / If-Modified-Since.
+		const etag = `"bibdle-feed-${feedFingerprint(verses)}"`;
+		const lastModifiedDate = latestModification(
+			verses,
+			new Date(`${today}T00:00:00Z`)
+		);
+		const lastModified = lastModifiedDate.toUTCString();
 
-		// Check if client has cached version
-		if (request.headers.get('If-None-Match') === etag) {
-			return new Response(null, { status: 304 });
+		const notModified =
+			request.headers.get('If-None-Match') === etag ||
+			(() => {
+				const since = request.headers.get('If-Modified-Since');
+				if (!since) return false;
+				const sinceMs = Date.parse(since);
+				return !Number.isNaN(sinceMs) && lastModifiedDate.getTime() <= sinceMs;
+			})();
+
+		if (notModified) {
+			return new Response(null, {
+				status: 304,
+				headers: {
+					ETag: etag,
+					'Last-Modified': lastModified,
+					'Cache-Control': 'public, max-age=3600, s-maxage=3600'
+				}
+			});
 		}
 
-		// Get site URL from environment or use default
-		const SITE_URL = process.env.PUBLIC_SITE_URL || 'https://bibdle.com';
-
-		// Build RSS XML
 		const lastBuildDate = verses[0] ? formatRFC822(verses[0].date) : new Date().toUTCString();
 
 		const items = verses
-			.map(
-				(verse) => `
+			.map((verse) => {
+				const body = formatVerseText(verse.verseText);
+				const escaped = escapeXml(body);
+				return `
     <item>
       <title>Bibdle verse for ${formatReadableDate(verse.date)}</title>
-      <description>${escapeXml(formatVerseText(verse.verseText))}</description>
+      <description>${escaped}</description>
+      <content:encoded>${cdata(`<p>${escaped}</p>`)}</content:encoded>
       <link>${SITE_URL}</link>
       <guid isPermaLink="false">bibdle-verse-${verse.date}</guid>
       <pubDate>${formatRFC822(verse.date)}</pubDate>
-    </item>`
-			)
+    </item>`;
+			})
 			.join('');
 
 		const xml = `<?xml version="1.0" encoding="UTF-8" ?>
-<rss version="2.0">
+<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
   <channel>
     <title>Bibdle</title>
     <link>${SITE_URL}</link>
-    <description>A daily Bible game</description>
+    <description>A daily Bible verse guessing game</description>
     <language>en-us</language>
     <lastBuildDate>${lastBuildDate}</lastBuildDate>
-    <ttl>720</ttl>${items}
+    <pubDate>${lastBuildDate}</pubDate>
+    <generator>SvelteKit</generator>
+    <ttl>720</ttl>
+    <atom:link href="${escapeXml(FEED_URL)}" rel="self" type="application/rss+xml" />
+    <image>
+      <url>${SITE_URL}/favicon.ico</url>
+      <title>Bibdle</title>
+      <link>${SITE_URL}</link>
+    </image>${items}
   </channel>
 </rss>`;
 
@@ -135,11 +101,12 @@ export const GET: RequestHandler = async ({ request }) => {
 			headers: {
 				'Content-Type': 'application/rss+xml; charset=utf-8',
 				'Cache-Control': 'public, max-age=3600, s-maxage=3600',
-				ETag: etag
+				ETag: etag,
+				'Last-Modified': lastModified
 			}
 		});
 	} catch (error) {
 		console.error('RSS feed generation error:', error);
 		return new Response('Internal Server Error', { status: 500 });
 	}
-};
+}

tests/rss.test.ts

@@ -0,0 +1,94 @@
+import { describe, test, expect } from 'bun:test';
+import {
+	cdata,
+	escapeXml,
+	feedFingerprint,
+	formatRFC822,
+	formatReadableDate,
+	formatVerseText,
+	latestModification,
+	todayInTimeZone
+} from '$lib/server/rss';
+
+describe('RSS date formatting', () => {
+	test('formatReadableDate uses the calendar day, independent of server TZ', () => {
+		expect(formatReadableDate('2026-09-18')).toBe('September 18, 2026');
+		expect(formatReadableDate('2026-01-01')).toBe('January 1, 2026');
+	});
+
+	test('formatRFC822 emits a valid GMT timestamp', () => {
+		expect(formatRFC822('2026-01-15')).toBe('Thu, 15 Jan 2026 05:00:00 GMT');
+		expect(formatRFC822('2026-09-18')).toBe('Fri, 18 Sep 2026 04:00:00 GMT');
+	});
+
+	test('formatRFC822 handles daylight saving (EST -05:00 vs EDT -04:00)', () => {
+		// Winter: midnight New York is 05:00 UTC.
+		expect(formatRFC822('2026-01-15')).toContain('05:00:00 GMT');
+		// Summer: midnight New York is 04:00 UTC.
+		expect(formatRFC822('2026-07-15')).toContain('04:00:00 GMT');
+		// Never a bare, ambiguous abbreviation like EST/EDT.
+		expect(formatRFC822('2026-07-15')).not.toMatch(/\bE[SD]T\b/);
+	});
+
+	test('todayInTimeZone respects the reference timezone', () => {
+		// 02:00 UTC on the 23rd is still the evening of the 22nd in New York.
+		expect(
+			todayInTimeZone('America/New_York', new Date('2026-09-23T02:00:00Z'))
+		).toBe('2026-09-22');
+		// 04:00 UTC on the 23rd is midnight in New York.
+		expect(
+			todayInTimeZone('America/New_York', new Date('2026-09-23T04:00:00Z'))
+		).toBe('2026-09-23');
+	});
+});
+
+describe('RSS content fingerprint', () => {
+	const row = { date: '2026-09-18', reference: 'Leviticus 19:9-11', verseText: 'a' };
+
+	test('is stable for identical content', () => {
+		expect(feedFingerprint([row])).toBe(feedFingerprint([{ ...row }]));
+	});
+
+	test('changes when verse text changes', () => {
+		expect(feedFingerprint([row])).not.toBe(feedFingerprint([{ ...row, verseText: 'b' }]));
+	});
+
+	test('changes when the set of days changes', () => {
+		expect(feedFingerprint([row])).not.toBe(
+			feedFingerprint([row, { ...row, date: '2026-09-17' }])
+		);
+	});
+});
+
+describe('latestModification', () => {
+	const fallback = new Date('2026-09-18T00:00:00Z');
+
+	test('returns the newest createdAt', () => {
+		const older = new Date('2026-09-16T10:00:00Z');
+		const newer = new Date('2026-09-17T10:00:00Z');
+		expect(
+			latestModification([{ createdAt: older }, { createdAt: newer }], fallback)
+		).toEqual(newer);
+	});
+
+	test('falls back when rows have no timestamp', () => {
+		expect(latestModification([{ createdAt: null }], fallback)).toEqual(fallback);
+		expect(latestModification([], fallback)).toEqual(fallback);
+	});
+});
+
+describe('XML serialization helpers', () => {
+	test('escapeXml escapes all predefined entities', () => {
+		expect(escapeXml(`<a href="x">&'</a>`)).toBe(
+			'&lt;a href=&quot;x&quot;&gt;&amp;&apos;&lt;/a&gt;'
+		);
+	});
+
+	test('cdata splits embedded CDATA terminators', () => {
+		expect(cdata('a]]>b')).toBe('<![CDATA[a]]]]><![CDATA[>b]]>');
+	});
+
+	test('formatVerseText capitalizes and replaces trailing punctuation', () => {
+		expect(formatVerseText('for God so loved the world,')).toBe('For God so loved the world...');
+	});
+});