Add support for the /Catalog Base-URI when resolving URLs (issue 14802)

As far as I can tell, this is actually the very first time that we've seen a PDF document with a Base-URI specified in the /Catalog; please refer to the specification:
https://web.archive.org/web/20220309040754if_/https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G11.2097122

To simplify the overall implementation, this new parameter is accessed via the existing `BasePdfManager.docBaseUrl`-getter and will thus override any user-specified `docBaseUrl` API-parameter.

src/core/annotation.js

@@ -72,9 +72,12 @@ class AnnotationFactory {
   static create(xref, ref, pdfManager, idFactory, collectFields) {
     return Promise.all([
       pdfManager.ensureCatalog("acroForm"),
+      // Only necessary to prevent the `pdfManager.docBaseUrl`-getter, used
+      // with certain Annotations, from throwing and thus breaking parsing:
+      pdfManager.ensureCatalog("baseUrl"),
       pdfManager.ensureDoc("xfaDatasets"),
       collectFields ? this._getPageIndex(xref, ref, pdfManager) : -1,
-    ]).then(([acroForm, xfaDatasets, pageIndex]) =>
+    ]).then(([acroForm, baseUrl, xfaDatasets, pageIndex]) =>
       pdfManager.ensure(this, "_create", [
         xref,
         ref,

src/core/catalog.js

@@ -1387,6 +1387,22 @@ class Catalog {
     return next(pageRef);
   }
 
+  get baseUrl() {
+    const uri = this._catDict.get("URI");
+    if (uri instanceof Dict) {
+      const base = uri.get("Base");
+      if (typeof base === "string") {
+        const absoluteUrl = createValidAbsoluteUrl(base, null, {
+          tryConvertEncoding: true,
+        });
+        if (absoluteUrl) {
+          return shadow(this, "baseUrl", absoluteUrl.href);
+        }
+      }
+    }
+    return shadow(this, "baseUrl", null);
+  }
+
   /**
    * @typedef {Object} ParseDestDictionaryParameters
    * @property {Dict} destDict - The dictionary containing the destination.
@@ -1464,8 +1480,6 @@ class Catalog {
             // Some bad PDFs do not put parentheses around relative URLs.
             url = "/" + url.name;
           }
-          // TODO: pdf spec mentions urls can be relative to a Base
-          // entry in the dictionary.
           break;
 
         case "GoTo":

src/core/pdf_manager.js

@@ -13,7 +13,12 @@
  * limitations under the License.
  */
 
-import { createValidAbsoluteUrl, unreachable, warn } from "../shared/util.js";
+import {
+  createValidAbsoluteUrl,
+  shadow,
+  unreachable,
+  warn,
+} from "../shared/util.js";
 import { ChunkedStreamManager } from "./chunked_stream.js";
 import { MissingDataException } from "./core_utils.js";
 import { PDFDocument } from "./document.js";
@@ -46,7 +51,8 @@ class BasePdfManager {
   }
 
   get docBaseUrl() {
-    return this._docBaseUrl;
+    const catalog = this.pdfDocument.catalog;
+    return shadow(this, "docBaseUrl", catalog.baseUrl || this._docBaseUrl);
   }
 
   onLoadedStream() {