Fall back to the /Catalog dictionary when a PDF file has no trailer

src/core/xref.js

@@ -21,7 +21,15 @@ import {
   InvalidPDFException,
   warn,
 } from "../shared/util.js";
-import { CIRCULAR_REF, Cmd, Dict, isCmd, Ref, RefSet } from "./primitives.js";
+import {
+  CIRCULAR_REF,
+  Cmd,
+  Dict,
+  isCmd,
+  isDict,
+  Ref,
+  RefSet,
+} from "./primitives.js";
 import { Lexer, Parser } from "./parser.js";
 import {
   MissingDataException,
@@ -701,8 +709,10 @@ class XRef {
     }
 
     // When no trailer dictionary candidate exists, try picking the first
-    // dictionary that contains a /Root entry (fixes issue18986.pdf).
+    // dictionary that contains a /Root entry (fixes issue18986.pdf), or else
+    // the first /Catalog dictionary (e.g. for files truncated before the end).
     if (!trailerDicts.length) {
+      let catalogRef = null;
       // In case, this.#entries is a sparse array we don't want to
       // iterate over empty entries so we use the `in` operator instead of
       // using for..of on entries() or a for with the array length.
@@ -716,7 +726,10 @@ class XRef {
 
         try {
           obj = this.fetch(ref);
-        } catch {
+        } catch (ex) {
+          if (ex instanceof MissingDataException) {
+            throw ex;
+          }
           continue;
         }
         if (obj instanceof BaseStream) {
@@ -725,6 +738,14 @@ class XRef {
         if (obj instanceof Dict && obj.has("Root")) {
           return obj;
         }
+        if (!catalogRef && isDict(obj, "Catalog")) {
+          catalogRef = ref;
+        }
+      }
+      if (catalogRef) {
+        const dict = new Dict(this);
+        dict.set("Root", catalogRef);
+        return dict;
       }
     }
 

test/unit/api_spec.js

@@ -453,6 +453,89 @@ describe("api", function () {
       await loadingTask.destroy();
     });
 
+    it("creates pdf doc from PDF file truncated before its trailer", async function () {
+      const compressedPixels =
+        "789c014000bfff00254a6f94b9de03284d7297bce1062b50759abfe4092e53789dc2e70c31567ba0c5ea0f34597ea3c8ed12375c81a6cbf0153a5f84a9cef3183d6287acd1f61ba08d1e61";
+      const streamObject = (num, content) =>
+        `${num} 0 obj\n<< /Length ${content.length} >>\nstream\n` +
+        `${content}\nendstream\nendobj\n`;
+      const imageObject = (num, data, complete) =>
+        `${num} 0 obj\n<< /Type /XObject /Subtype /Image /Width 4 ` +
+        `/Height 4 /ColorSpace /DeviceCMYK /BitsPerComponent 8 ` +
+        `/Filter [/ASCIIHexDecode /FlateDecode] ` +
+        `/Length ${compressedPixels.length + 1} >>\nstream\n` +
+        `${data}${complete ? ">\nendstream\nendobj\n" : ""}`;
+      const pageContents = [
+        "q 72 0 0 72 0 0 cm /Im0 Do /Im1 Do Q\n" +
+          "BT /F1 12 Tf 72 720 Td (Synthetic map page one) Tj ET",
+        "q 72 0 0 72 0 0 cm /Im0 Do /Im1 Do Q\n" +
+          "BT /F1 12 Tf 72 720 Td (Synthetic map page two) Tj ET",
+        "q 72 0 0 72 0 0 cm /Im0 Do /Im1 Do Q\n" +
+          "BT /F1 12 Tf 72 720 Td (Synthetic map page three) Tj ET",
+        "q 72 0 0 72 0 0 cm /Im0 Do /Im1 Do Q\n" +
+          "BT /F1 12 Tf 72 720 Td (Synthetic map page four) Tj ET",
+      ];
+      const pageObjects = [
+        [3, 4, 5],
+        [6, 15, 16],
+        [8, 25, 26],
+        [10, 32, 33],
+      ];
+      let pdf =
+        "%PDF-1.3\n" +
+        streamObject(4, pageContents[0]) +
+        "5 0 obj\n<< /ProcSet [/PDF /Text /ImageC] " +
+        "/XObject << /Im0 7 0 R /Im1 17 0 R >> " +
+        "/Font << /F1 11 0 R >> /ExtGState << /GS1 9 0 R >> >>\n" +
+        "endobj\n" +
+        "9 0 obj\n<< /Type /ExtGState /ca 1 /CA 1 >>\nendobj\n" +
+        "11 0 obj\n<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>\n" +
+        "endobj\n" +
+        streamObject(15, pageContents[1]) +
+        "16 0 obj\n<< /ProcSet [/PDF /Text /ImageC] " +
+        "/XObject << /Im0 7 0 R /Im1 17 0 R >> " +
+        "/Font << /F1 11 0 R >> /ExtGState << /GS1 9 0 R >> >>\n" +
+        "endobj\n" +
+        streamObject(25, pageContents[2]) +
+        "26 0 obj\n<< /ProcSet [/PDF /Text /ImageC] " +
+        "/XObject << /Im0 7 0 R /Im1 17 0 R >> " +
+        "/Font << /F1 11 0 R >> /ExtGState << /GS1 9 0 R >> >>\n" +
+        "endobj\n" +
+        streamObject(32, pageContents[3]) +
+        "33 0 obj\n<< /ProcSet [/PDF /Text /ImageC] " +
+        "/XObject << /Im0 7 0 R /Im1 17 0 R >> " +
+        "/Font << /F1 11 0 R >> /ExtGState << /GS1 9 0 R >> >>\n" +
+        "endobj\n" +
+        imageObject(7, compressedPixels, /* complete = */ true) +
+        "1 0 obj\n<< /Type /Catalog /Pages 2 0 R >>\nendobj\n" +
+        "2 0 obj\n<< /Type /Pages /Kids [3 0 R 6 0 R 8 0 R 10 0 R] " +
+        "/Count 4 >>\nendobj\n";
+      for (const [pageObj, contentObj, resourcesObj] of pageObjects) {
+        pdf +=
+          `${pageObj} 0 obj\n<< /Type /Page /Parent 2 0 R ` +
+          `/MediaBox [0 0 612 792] /Resources ${resourcesObj} 0 R ` +
+          `/Contents ${contentObj} 0 R >>\nendobj\n`;
+      }
+      // Keep the document tree intact, then truncate the final image stream.
+      pdf += imageObject(
+        17,
+        compressedPixels.slice(0, compressedPixels.length >> 1),
+        /* complete = */ false
+      );
+
+      const loadingTask = getDocument({
+        data: stringToBytes(pdf),
+      });
+      const pdfDocument = await loadingTask.promise;
+      expect(pdfDocument.numPages).toEqual(4);
+
+      const page = await pdfDocument.getPage(1);
+      const textContent = await page.getTextContent();
+      expect(mergeText(textContent.items)).toContain("Synthetic map page one");
+
+      await loadingTask.destroy();
+    });
+
     it("creates pdf doc from non-existent URL", async function () {
       const loadingTask = getDocument(
         buildGetDocumentParams("non-existent.pdf")