fix(sitemap): Trailing slashes on root url (withastro#10772)

* add tests that reveal issue

* fix trailing slash root page issue

* add changeset

.changeset/late-bags-marry.md

@@ -0,0 +1,5 @@
+---
+"@astrojs/sitemap": patch
+---
+
+Fixes an issue where the root url does not follow the `trailingSlash` config option

packages/integrations/sitemap/package.json

@@ -34,6 +34,7 @@
   },
   "dependencies": {
     "sitemap": "^7.1.1",
+    "stream-replace-string": "^2.0.0",
     "zod": "^3.22.4"
   },
   "devDependencies": {

packages/integrations/sitemap/src/index.ts

@@ -2,11 +2,11 @@ import path from 'node:path';
 import { fileURLToPath } from 'node:url';
 import type { AstroConfig, AstroIntegration } from 'astro';
 import type { EnumChangefreq, LinkItem as LinkItemBase, SitemapItemLoose } from 'sitemap';
-import { simpleSitemapAndIndex } from 'sitemap';
 import { ZodError } from 'zod';
 
-import { generateSitemap } from './generate-sitemap.js';
 import { validateOptions } from './validate-options.js';
+import { generateSitemap } from './generate-sitemap.js';
+import { writeSitemap } from './write-sitemap.js';
 
 export { EnumChangefreq as ChangeFreqEnum } from 'sitemap';
 export type ChangeFreq = `${EnumChangefreq}`;
@@ -167,14 +167,13 @@ const createPlugin = (options?: SitemapOptions): AstroIntegration => {
 						}
 					}
 					const destDir = fileURLToPath(dir);
-					await simpleSitemapAndIndex({
+					await writeSitemap({
 						hostname: finalSiteUrl.href,
 						destinationDir: destDir,
 						publicBasePath: config.base,
 						sourceData: urlData,
-						limit: entryLimit,
-						gzip: false,
-					});
+						limit: entryLimit
+					}, config)
 					logger.info(`\`${OUTFILE}\` created at \`${path.relative(process.cwd(), destDir)}\``);
 				} catch (err) {
 					if (err instanceof ZodError) {

packages/integrations/sitemap/src/write-sitemap.ts

@@ -0,0 +1,69 @@
+import { normalize, resolve } from 'path';
+import { createWriteStream, type WriteStream } from 'fs'
+import { mkdir } from 'fs/promises';
+import { promisify } from 'util';
+import { Readable, pipeline } from 'stream';
+import replace from 'stream-replace-string'
+
+import { SitemapAndIndexStream, SitemapStream } from 'sitemap';
+
+import type { AstroConfig } from 'astro';
+import type { SitemapItem } from "./index.js";
+
+type WriteSitemapConfig = {
+   hostname: string;
+   sitemapHostname?: string;
+   sourceData: SitemapItem[];
+   destinationDir: string;
+   publicBasePath?: string;
+   limit?: number;
+}
+
+// adapted from sitemap.js/sitemap-simple
+export async function writeSitemap({ hostname, sitemapHostname = hostname,
+sourceData, destinationDir, limit = 50000, publicBasePath = './', }: WriteSitemapConfig, astroConfig: AstroConfig) {
+
+   await mkdir(destinationDir, { recursive: true })
+
+   const sitemapAndIndexStream = new SitemapAndIndexStream({
+      limit,
+      getSitemapStream: (i) => {
+        const sitemapStream = new SitemapStream({
+          hostname,
+        });
+        const path = `./sitemap-${i}.xml`;
+        const writePath = resolve(destinationDir, path);
+        if (!publicBasePath.endsWith('/')) {
+          publicBasePath += '/';
+        }
+        const publicPath = normalize(publicBasePath + path);
+
+        let stream: WriteStream
+        if (astroConfig.trailingSlash === 'never' || astroConfig.build.format === 'file') {
+          // workaround for trailing slash issue in sitemap.js: https://github.com/ekalinin/sitemap.js/issues/403
+          const host = hostname.endsWith('/') ? hostname.slice(0, -1) : hostname
+          const searchStr = `<loc>${host}/</loc>`
+          const replaceStr = `<loc>${host}</loc>`
+          stream = sitemapStream.pipe(replace(searchStr, replaceStr)).pipe(createWriteStream(writePath))
+        } else {
+          stream = sitemapStream.pipe(createWriteStream(writePath))
+        }
+
+        return [
+          new URL(
+            publicPath,
+            sitemapHostname
+          ).toString(),
+          sitemapStream,
+          stream,
+        ];
+      },
+    });
+
+    let src = Readable.from(sourceData)
+    const indexPath = resolve(
+      destinationDir,
+      `./sitemap-index.xml`
+    );
+   return promisify(pipeline)(src, sitemapAndIndexStream, createWriteStream(indexPath));
+}

packages/integrations/sitemap/test/fixtures/trailing-slash/src/pages/index.astro

@@ -0,0 +1,8 @@
+<html>
+	<head>
+		<title>Index</title>
+	</head>
+	<body>
+		<h1>Index</h1>
+	</body>
+</html>

packages/integrations/sitemap/test/trailing-slash.test.js

@@ -22,7 +22,10 @@ describe('Trailing slash', () => {
 			it('URLs end with trailing slash', async () => {
 				const data = await readXML(fixture.readFile('/sitemap-0.xml'));
 				const urls = data.urlset.url;
-				assert.equal(urls[0].loc[0], 'http://example.com/one/');
+
+				assert.equal(urls[0].loc[0], 'http://example.com/');
+				assert.equal(urls[1].loc[0], 'http://example.com/one/');
+				assert.equal(urls[2].loc[0], 'http://example.com/two/');
 			});
 		});
 
@@ -41,7 +44,10 @@ describe('Trailing slash', () => {
 			it('URLs do not end with trailing slash', async () => {
 				const data = await readXML(fixture.readFile('/sitemap-0.xml'));
 				const urls = data.urlset.url;
-				assert.equal(urls[0].loc[0], 'http://example.com/one');
+
+				assert.equal(urls[0].loc[0], 'http://example.com');
+				assert.equal(urls[1].loc[0], 'http://example.com/one');
+				assert.equal(urls[2].loc[0], 'http://example.com/two');
 			});
 		});
 	});
@@ -55,10 +61,13 @@ describe('Trailing slash', () => {
 			await fixture.build();
 		});
 
-		it('URLs do no end with trailing slash', async () => {
+		it('URLs do not end with trailing slash', async () => {
 			const data = await readXML(fixture.readFile('/sitemap-0.xml'));
 			const urls = data.urlset.url;
-			assert.equal(urls[0].loc[0], 'http://example.com/one');
+
+			assert.equal(urls[0].loc[0], 'http://example.com');
+			assert.equal(urls[1].loc[0], 'http://example.com/one');
+			assert.equal(urls[2].loc[0], 'http://example.com/two');
 		});
 		describe('with base path', () => {
 			before(async () => {
@@ -73,7 +82,9 @@ describe('Trailing slash', () => {
 			it('URLs do not end with trailing slash', async () => {
 				const data = await readXML(fixture.readFile('/sitemap-0.xml'));
 				const urls = data.urlset.url;
-				assert.equal(urls[0].loc[0], 'http://example.com/base/one');
+				assert.equal(urls[0].loc[0], 'http://example.com/base');
+				assert.equal(urls[1].loc[0], 'http://example.com/base/one');
+				assert.equal(urls[2].loc[0], 'http://example.com/base/two');
 			});
 		});
 	});
@@ -90,7 +101,9 @@ describe('Trailing slash', () => {
 		it('URLs end with trailing slash', async () => {
 			const data = await readXML(fixture.readFile('/sitemap-0.xml'));
 			const urls = data.urlset.url;
-			assert.equal(urls[0].loc[0], 'http://example.com/one/');
+			assert.equal(urls[0].loc[0], 'http://example.com/');
+			assert.equal(urls[1].loc[0], 'http://example.com/one/');
+			assert.equal(urls[2].loc[0], 'http://example.com/two/');
 		});
 		describe('with base path', () => {
 			before(async () => {
@@ -105,7 +118,9 @@ describe('Trailing slash', () => {
 			it('URLs end with trailing slash', async () => {
 				const data = await readXML(fixture.readFile('/sitemap-0.xml'));
 				const urls = data.urlset.url;
-				assert.equal(urls[0].loc[0], 'http://example.com/base/one/');
+				assert.equal(urls[0].loc[0], 'http://example.com/base/');
+				assert.equal(urls[1].loc[0], 'http://example.com/base/one/');
+				assert.equal(urls[2].loc[0], 'http://example.com/base/two/');
 			});
 		});
 	});