feat: Decode URI in target extractor

src/ldp/http/BasicTargetExtractor.ts

@@ -1,6 +1,7 @@
 import type { TLSSocket } from 'tls';
 import { format } from 'url';
 import type { HttpRequest } from '../../server/HttpRequest';
+import { toCanonicalUrl } from '../../util/Util';
 import type { ResourceIdentifier } from '../representation/ResourceIdentifier';
 import { TargetExtractor } from './TargetExtractor';
 
@@ -28,6 +29,6 @@ export class BasicTargetExtractor extends TargetExtractor {
       pathname: input.url,
     });
 
-    return { path: url };
+    return { path: toCanonicalUrl(url) };
   }
 }

src/util/Util.ts

@@ -66,3 +66,16 @@ export const pipeStreamsAndErrors = <T extends Writable>(readable: Readable, des
   readable.on('error', (error): boolean => destination.emit('error', new UnsupportedHttpError(error.message)));
   return destination;
 };
+
+/**
+ * Converts a URL string to the "canonical" version that should be used internally for consistency.
+ * Decodes all percent encodings and then makes sure only the necessary characters are encoded again.
+ */
+export const toCanonicalUrl = (url: string): string => {
+  const match = /(\w+:\/\/[^/]+\/)(.*)/u.exec(url);
+  if (!match) {
+    throw new UnsupportedHttpError(`Invalid URL ${url}`);
+  }
+  const [ , domain, path ] = match;
+  return encodeURI(domain + path.split('/').map(decodeURIComponent).join('/'));
+};

test/unit/ldp/http/BasicTargetExtractor.test.ts

@@ -16,7 +16,8 @@ describe('A BasicTargetExtractor', (): void => {
   });
 
   it('returns the input URL.', async(): Promise<void> => {
-    await expect(extractor.handle({ url: 'url', headers: { host: 'test.com' }} as any)).resolves.toEqual({ path: 'http://test.com/url' });
+    await expect(extractor.handle({ url: 'url', headers: { host: 'test.com' }} as any))
+      .resolves.toEqual({ path: 'http://test.com/url' });
   });
 
   it('uses https protocol if the connection is secure.', async(): Promise<void> => {
@@ -24,4 +25,9 @@ describe('A BasicTargetExtractor', (): void => {
       { url: 'url', headers: { host: 'test.com' }, connection: { encrypted: true } as any } as any,
     )).resolves.toEqual({ path: 'https://test.com/url' });
   });
+
+  it('decodes relevant percent encodings.', async(): Promise<void> => {
+    await expect(extractor.handle({ url: '/a%20path%26/name', headers: { host: 'test.com' }} as any))
+      .resolves.toEqual({ path: 'http://test.com/a%20path&/name' });
+  });
 });

test/unit/util/Util.test.ts

@@ -1,5 +1,6 @@
 import streamifyArray from 'streamify-array';
-import { ensureTrailingSlash, matchingMediaType, readableToString } from '../../../src/util/Util';
+import { UnsupportedHttpError } from '../../../src/util/errors/UnsupportedHttpError';
+import { ensureTrailingSlash, matchingMediaType, readableToString, toCanonicalUrl } from '../../../src/util/Util';
 
 describe('Util function', (): void => {
   describe('ensureTrailingSlash', (): void => {
@@ -31,4 +32,15 @@ describe('Util function', (): void => {
       expect(matchingMediaType('text/plain', 'text/turtle')).toBeFalsy();
     });
   });
+
+  describe('toCanonicalUrl', (): void => {
+    it('makes sure only the necessary parts are encoded.', async(): Promise<void> => {
+      expect(toCanonicalUrl('http://test.com/a%20path%26/name'))
+        .toEqual('http://test.com/a%20path&/name');
+    });
+
+    it('errors on invalid URLs.', async(): Promise<void> => {
+      expect((): any => toCanonicalUrl('notAnUrl')).toThrow(new UnsupportedHttpError('Invalid URL notAnUrl'));
+    });
+  });
 });