From 18e3a98c2bbfc63e0603b66a49f9e8a2ff58e09c Mon Sep 17 00:00:00 2001
From: Calixte Denizet <calixte.denizet@gmail.com>
Date: Sun, 13 Feb 2022 19:39:40 +0100
Subject: [PATCH] [api-minor] Don't add in the text content the chars which are
 out-of-page (bug 1755201) - it aims to fix
 https://bugzilla.mozilla.org/show_bug.cgi?id=1755201; - if the glyph position
 is not within the view then skip it.

---
 src/core/document.js          |  1 +
 src/core/evaluator.js         | 46 +++++++++++++++++++++++++----------
 test/pdfs/bug1755201.pdf.link |  1 +
 test/test_manifest.json       |  7 ++++++
 test/unit/api_spec.js         | 16 ++++++++++++
 5 files changed, 58 insertions(+), 13 deletions(-)
 create mode 100644 test/pdfs/bug1755201.pdf.link

diff --git a/src/core/document.js b/src/core/document.js
index 848e05512..1ebff3ab3 100644
--- a/src/core/document.js
+++ b/src/core/document.js
@@ -471,6 +471,7 @@ class Page {
         includeMarkedContent,
         combineTextItems,
         sink,
+        viewBox: this.view,
       });
     });
   }
diff --git a/src/core/evaluator.js b/src/core/evaluator.js
index b510e4a93..3cfb89ec8 100644
--- a/src/core/evaluator.js
+++ b/src/core/evaluator.js
@@ -2167,6 +2167,7 @@ class PartialEvaluator {
     includeMarkedContent = false,
     sink,
     seenStyles = new Set(),
+    viewBox,
   }) {
     // Ensure that `resources`/`stateManager` is correctly initialized,
     // even if the provided parameter is e.g. `null`.
@@ -2393,22 +2394,35 @@ class PartialEvaluator {
     }
 
     function compareWithLastPosition() {
+      const currentTransform = getCurrentTextTransform();
+      let posX = currentTransform[4];
+      let posY = currentTransform[5];
+
+      const shiftedX = posX - viewBox[0];
+      const shiftedY = posY - viewBox[1];
+
+      if (
+        shiftedX < 0 ||
+        shiftedX > viewBox[2] ||
+        shiftedY < 0 ||
+        shiftedY > viewBox[3]
+      ) {
+        return false;
+      }
+
       if (
         !combineTextItems ||
         !textState.font ||
         !textContentItem.prevTransform
       ) {
-        return;
+        return true;
       }
 
-      const currentTransform = getCurrentTextTransform();
-      let posX = currentTransform[4];
-      let posY = currentTransform[5];
       let lastPosX = textContentItem.prevTransform[4];
       let lastPosY = textContentItem.prevTransform[5];
 
       if (lastPosX === posX && lastPosY === posY) {
-        return;
+        return true;
       }
 
       let rotate = -1;
@@ -2473,16 +2487,16 @@ class PartialEvaluator {
             0.5 * textContentItem.width /* not the same column */
           ) {
             appendEOL();
-            return;
+            return true;
           }
 
           flushTextContentItem();
-          return;
+          return true;
         }
 
         if (Math.abs(advanceX) > textContentItem.width) {
           appendEOL();
-          return;
+          return true;
         }
         if (advanceY <= textOrientation * textContentItem.trackingSpaceMin) {
           textContentItem.height += advanceY;
@@ -2508,7 +2522,7 @@ class PartialEvaluator {
           }
         }
 
-        return;
+        return true;
       }
 
       const advanceX = (posX - lastPosX) / textContentItem.textAdvanceScale;
@@ -2523,15 +2537,15 @@ class PartialEvaluator {
           0.5 * textContentItem.height /* not the same line */
         ) {
           appendEOL();
-          return;
+          return true;
         }
         flushTextContentItem();
-        return;
+        return true;
       }
 
       if (Math.abs(advanceY) > textContentItem.height) {
         appendEOL();
-        return;
+        return true;
       }
 
       if (advanceX <= textOrientation * textContentItem.trackingSpaceMin) {
@@ -2553,6 +2567,8 @@ class PartialEvaluator {
           textContentItem.width += advanceX;
         }
       }
+
+      return true;
     }
 
     function buildTextContentItem({ chars, extraSpacing }) {
@@ -2617,7 +2633,10 @@ class PartialEvaluator {
           continue;
         }
 
-        compareWithLastPosition();
+        if (!compareWithLastPosition()) {
+          // The glyph is not in page so just skip it.
+          continue;
+        }
 
         // Must be called after compareWithLastPosition because
         // the textContentItem could have been flushed.
@@ -3026,6 +3045,7 @@ class PartialEvaluator {
                     includeMarkedContent,
                     sink: sinkWrapper,
                     seenStyles,
+                    viewBox,
                   })
                   .then(function () {
                     if (!sinkWrapper.enqueueInvoked) {
diff --git a/test/pdfs/bug1755201.pdf.link b/test/pdfs/bug1755201.pdf.link
new file mode 100644
index 000000000..54b18c659
--- /dev/null
+++ b/test/pdfs/bug1755201.pdf.link
@@ -0,0 +1 @@
+https://bugzilla.mozilla.org/attachment.cgi?id=9263657
diff --git a/test/test_manifest.json b/test/test_manifest.json
index b9302e61d..41f7f7765 100644
--- a/test/test_manifest.json
+++ b/test/test_manifest.json
@@ -1,4 +1,11 @@
 [
+    {  "id": "bug1755201",
+      "file": "pdfs/bug1755201.pdf",
+      "md5": "cece14097812d8a1f69e86a51e4a3804",
+      "rounds": 1,
+      "link": true,
+      "type": "other"
+    },
     {  "id": "filled-background-range",
       "file": "pdfs/filled-background.pdf",
       "md5": "2e3120255d9c3e79b96d2543b12d2589",
diff --git a/test/unit/api_spec.js b/test/unit/api_spec.js
index fc56ae9a2..95cc8461c 100644
--- a/test/unit/api_spec.js
+++ b/test/unit/api_spec.js
@@ -2219,6 +2219,22 @@ Caron Broadcasting, Inc., an Ohio corporation (“Lessee”).`)
       await loadingTask.destroy();
     });
 
+    it("gets text content, and check that out-of-page text is not present (bug 1755201)", async function () {
+      if (isNodeJS) {
+        pending("Linked test-cases are not supported in Node.js.");
+      }
+
+      const loadingTask = getDocument(buildGetDocumentParams("bug1755201.pdf"));
+      const pdfDoc = await loadingTask.promise;
+      const pdfPage = await pdfDoc.getPage(6);
+      const { items } = await pdfPage.getTextContent();
+      const text = mergeText(items);
+
+      expect(/win aisle/.test(text)).toEqual(false);
+
+      await loadingTask.destroy();
+    });
+
     it("gets empty structure tree", async function () {
       const tree = await page.getStructTree();