From 37e98e39f60bd15c8954a88572ccb0a2842e9dd0 Mon Sep 17 00:00:00 2001
From: Jonas Jenwald <jonas.jenwald@gmail.com>
Date: Mon, 12 Feb 2024 15:31:08 +0100
Subject: [PATCH] Skip any whitespace after the first object in linearized PDFs
 (issue 17665)

This way the code is now consistent with the non-linearized branch in the `PDFDocument.startXRef` getter.
---
 src/core/document.js  |  9 ++++++++-
 src/core/worker.js    |  3 +++
 src/display/api.js    |  7 +++++++
 test/unit/api_spec.js | 12 ++++++++++++
 4 files changed, 30 insertions(+), 1 deletion(-)

diff --git a/src/core/document.js b/src/core/document.js
index 37d81d9bc..f783ef31d 100644
--- a/src/core/document.js
+++ b/src/core/document.js
@@ -930,7 +930,14 @@ class PDFDocument {
       // Find the end of the first object.
       stream.reset();
       if (find(stream, ENDOBJ_SIGNATURE)) {
-        startXRef = stream.pos + 6 - stream.start;
+        stream.skip(6);
+
+        let ch = stream.peekByte();
+        while (isWhiteSpace(ch)) {
+          stream.pos++;
+          ch = stream.peekByte();
+        }
+        startXRef = stream.pos - stream.start;
       }
     } else {
       // Find `startxref` by checking backwards from the end of the file.
diff --git a/src/core/worker.js b/src/core/worker.js
index fe699b978..70fa1c249 100644
--- a/src/core/worker.js
+++ b/src/core/worker.js
@@ -880,6 +880,9 @@ class WorkerMessageHandler {
           .ensureXRef("trailer")
           .then(trailer => trailer.get("Prev"));
       });
+      handler.on("GetStartXRefPos", function (data) {
+        return pdfManager.ensureDoc("startXRef");
+      });
       handler.on("GetAnnotArray", function (data) {
         return pdfManager.getPage(data.pageIndex).then(function (page) {
           return page.annotations.map(a => a.toString());
diff --git a/src/display/api.js b/src/display/api.js
index faebddbea..2ae61ea29 100644
--- a/src/display/api.js
+++ b/src/display/api.js
@@ -768,6 +768,9 @@ class PDFDocumentProxy {
       Object.defineProperty(this, "getXRefPrevValue", {
         value: () => this._transport.getXRefPrevValue(),
       });
+      Object.defineProperty(this, "getStartXRefPos", {
+        value: () => this._transport.getStartXRefPos(),
+      });
       Object.defineProperty(this, "getAnnotArray", {
         value: pageIndex => this._transport.getAnnotArray(pageIndex),
       });
@@ -2349,6 +2352,10 @@ class WorkerTransport {
         value: () =>
           this.messageHandler.sendWithPromise("GetXRefPrevValue", null),
       });
+      Object.defineProperty(this, "getStartXRefPos", {
+        value: () =>
+          this.messageHandler.sendWithPromise("GetStartXRefPos", null),
+      });
       Object.defineProperty(this, "getAnnotArray", {
         value: pageIndex =>
           this.messageHandler.sendWithPromise("GetAnnotArray", { pageIndex }),
diff --git a/test/unit/api_spec.js b/test/unit/api_spec.js
index 7662c405e..0035c0c65 100644
--- a/test/unit/api_spec.js
+++ b/test/unit/api_spec.js
@@ -511,6 +511,18 @@ describe("api", function () {
       await loadingTask.destroy();
     });
 
+    it("checks the `startxref` position of a linearized pdf doc (issue 17665)", async function () {
+      const loadingTask = getDocument(buildGetDocumentParams("empty.pdf"));
+      expect(loadingTask instanceof PDFDocumentLoadingTask).toEqual(true);
+
+      const pdfDocument = await loadingTask.promise;
+
+      const startXRefPos = await pdfDocument.getStartXRefPos();
+      expect(startXRefPos).toEqual(116);
+
+      await loadingTask.destroy();
+    });
+
     it("checks that `docId`s are unique and increasing", async function () {
       const loadingTask1 = getDocument(basicApiGetDocumentParams);
       expect(loadingTask1 instanceof PDFDocumentLoadingTask).toEqual(true);