feat: regexRoutes stored as ordered array

RELEASE_NOTES.md

@@ -24,7 +24,7 @@ The following changes pertain to the imports in the default configs:
 - All default configurations with a file-based backend now use a file-based locker instead of a memory-based one,
   making them threadsafe.
 
-The following changes are relevant for v3 custom configs that replaced certain features.
+The following changes are relevant for v4 custom configs that replaced certain features.
 - `config/app/variables/cli.json` was changed to support the new `YargsCliExtractor` format.
 - `config/util/resource-locker/memory.json` had the locker @type changed from `SingleThreadedResourceLocker` to `MemoryResourceLocker`.
 - The content-length parser has been moved from the default configuration to the quota configurations.
@@ -33,6 +33,9 @@ The following changes are relevant for v3 custom configs that replaced certain f
    - `/storage/backend/quota/quota-file.json`
 - The structure of the init configs has changed significantly to support worker threads.
    - `/app/init/*`
+- RegexPathRouting has changed from a map datastructure to an array datastructure, allowing for fallthrough regex parsing. The change is reflected in the following default configs:
+   - `/storage/backend/regex.json`
+   - `/sparql-file-storage.json`
 
 ### Interface changes
 These changes are relevant if you wrote custom modules for the server that depend on existing interfaces.

config/sparql-file-storage.json

@@ -51,16 +51,18 @@
       "@id": "urn:solid-server:default:RouterRule",
       "@type": "RegexRouterRule",
       "base": { "@id": "urn:solid-server:default:variable:baseUrl" },
-      "storeMap": [
+      "rules": [
         {
           "comment": "Internal storage data",
-          "RegexRouterRule:_storeMap_key": "^/\\.internal/",
-          "RegexRouterRule:_storeMap_value": { "@id": "urn:solid-server:default:FileResourceStore" }
+          "@type": "RegexRule",
+          "regex": "^/\\.internal/",
+          "store": { "@id": "urn:solid-server:default:FileResourceStore" }
         },
         {
           "comment": "Send everything else to the SPARQL store.",
-          "RegexRouterRule:_storeMap_key": "^/(?!\\.internal/).*",
-          "RegexRouterRule:_storeMap_value": { "@id": "urn:solid-server:default:SparqlResourceStore" }
+          "@type": "RegexRule",
+          "regex": ".*",
+          "store": { "@id": "urn:solid-server:default:SparqlResourceStore" }
         }
       ]
     },

config/storage/backend/regex.json

@@ -21,22 +21,26 @@
       "@id": "urn:solid-server:default:RouterRule",
       "@type": "RegexRouterRule",
       "base": { "@id": "urn:solid-server:default:variable:baseUrl" },
-      "storeMap": [
+      "rules": [
         {
-          "RegexRouterRule:_storeMap_key": "^/(\\.acl)?$",
-          "RegexRouterRule:_storeMap_value": { "@id": "urn:solid-server:default:SparqlResourceStore" }
+          "@type": "RegexRule",
+          "regex": "^/(\\.acl)?$",
+          "store": { "@id": "urn:solid-server:default:SparqlResourceStore" }
         },
         {
-          "RegexRouterRule:_storeMap_key": "/file/",
-          "RegexRouterRule:_storeMap_value": { "@id": "urn:solid-server:default:FileResourceStore" }
+          "@type": "RegexRule",
+          "regex": "/file/",
+          "store": { "@id": "urn:solid-server:default:FileResourceStore" }
         },
         {
-          "RegexRouterRule:_storeMap_key": "/memory/",
-          "RegexRouterRule:_storeMap_value": { "@id": "urn:solid-server:default:MemoryResourceStore" }
+          "@type": "RegexRule",
+          "regex": "/memory/",
+          "store": { "@id": "urn:solid-server:default:MemoryResourceStore" }
         },
         {
-          "RegexRouterRule:_storeMap_key": "/sparql/",
-          "RegexRouterRule:_storeMap_value": { "@id": "urn:solid-server:default:SparqlResourceStore" }
+          "@type": "RegexRule",
+          "regex": "/sparql/",
+          "store": { "@id": "urn:solid-server:default:SparqlResourceStore" }
         }
       ]
     },

src/storage/routing/RegexRouterRule.ts

@@ -6,6 +6,19 @@ import { trimTrailingSlashes } from '../../util/PathUtil';
 import type { ResourceStore } from '../ResourceStore';
 import { RouterRule } from './RouterRule';
 
+/**
+ * Utility class to easily configure Regex to ResourceStore mappings in the config files.
+ */
+export class RegexRule {
+  public readonly regex: RegExp;
+  public readonly store: ResourceStore;
+
+  public constructor(regex: string, store: ResourceStore) {
+    this.regex = new RegExp(regex, 'u');
+    this.store = store;
+  }
+}
+
 /**
  * Routes requests to a store based on the path of the identifier.
  * The identifier will be stripped of the base URI after which regexes will be used to find the correct store.
@@ -16,16 +29,15 @@ import { RouterRule } from './RouterRule';
  */
 export class RegexRouterRule extends RouterRule {
   private readonly base: string;
-  private readonly regexes: Map<RegExp, ResourceStore>;
+  private readonly rules: RegexRule[];
 
   /**
    * The keys of the `storeMap` will be converted into actual RegExp objects that will be used for testing.
    */
-  public constructor(base: string, storeMap: Record<string, ResourceStore>) {
+  public constructor(base: string, rules: RegexRule[]) {
     super();
     this.base = trimTrailingSlashes(base);
-    this.regexes = new Map(Object.keys(storeMap).map((regex): [ RegExp, ResourceStore ] =>
-      [ new RegExp(regex, 'u'), storeMap[regex] ]));
+    this.rules = rules;
   }
 
   public async canHandle(input: { identifier: ResourceIdentifier; representation?: Representation }): Promise<void> {
@@ -42,9 +54,9 @@ export class RegexRouterRule extends RouterRule {
    */
   private matchStore(identifier: ResourceIdentifier): ResourceStore {
     const path = this.toRelative(identifier);
-    for (const regex of this.regexes.keys()) {
+    for (const { regex, store } of this.rules) {
       if (regex.test(path)) {
-        return this.regexes.get(regex)!;
+        return store;
       }
     }
     throw new NotImplementedHttpError(`No stored regexes match ${identifier.path}`);

test/unit/storage/routing/RegexRouterRule.test.ts

@@ -1,5 +1,5 @@
 import type { ResourceStore } from '../../../../src/storage/ResourceStore';
-import { RegexRouterRule } from '../../../../src/storage/routing/RegexRouterRule';
+import { RegexRouterRule, RegexRule } from '../../../../src/storage/routing/RegexRouterRule';
 import { BadRequestHttpError } from '../../../../src/util/errors/BadRequestHttpError';
 import { NotImplementedHttpError } from '../../../../src/util/errors/NotImplementedHttpError';
 
@@ -7,29 +7,36 @@ describe('A RegexRouterRule', (): void => {
   const base = 'http://test.com/';
   const store: ResourceStore = 'resourceStore' as any;
 
+  it('can construct a RegexRule utility object.', (): void => {
+    const regex = '/myPath/';
+    const rule = new RegexRule(regex, store);
+    expect(rule.regex).toEqual(new RegExp(regex, 'u'));
+    expect(rule.store).toEqual(store);
+  });
+
   it('rejects identifiers not containing the base.', async(): Promise<void> => {
-    const router = new RegexRouterRule(base, {});
+    const router = new RegexRouterRule(base, []);
     const result = router.canHandle({ identifier: { path: 'http://notTest.com/apple' }});
     await expect(result).rejects.toThrow(BadRequestHttpError);
     await expect(result).rejects.toThrow('Identifiers need to start with http://test.com');
   });
 
   it('rejects identifiers not matching any regex.', async(): Promise<void> => {
-    const router = new RegexRouterRule(base, { pear: store });
+    const router = new RegexRouterRule(base, [ new RegexRule('pear', store) ]);
     const result = router.canHandle({ identifier: { path: `${base}apple/` }});
     await expect(result).rejects.toThrow(NotImplementedHttpError);
     await expect(result).rejects.toThrow('No stored regexes match http://test.com/apple/');
   });
 
   it('accepts identifiers matching any regex.', async(): Promise<void> => {
-    const router = new RegexRouterRule(base, { '^/apple': store });
+    const router = new RegexRouterRule(base, [ new RegexRule('^/apple', store) ]);
     await expect(router.canHandle({ identifier: { path: `${base}apple/` }}))
       .resolves.toBeUndefined();
   });
 
   it('returns the corresponding store.', async(): Promise<void> => {
     const store2: ResourceStore = 'resourceStore2' as any;
-    const router = new RegexRouterRule(base, { '^/apple': store2, '/pear/': store });
+    const router = new RegexRouterRule(base, [ new RegexRule('^/apple', store2), new RegexRule('/pear/', store) ]);
     await expect(router.handle({ identifier: { path: `${base}apple/` }})).resolves.toBe(store2);
   });
 });