Merge pull request #19095 from calixteman/issue17190

Rescale the image data when they're really too large

src/core/core_utils.js

@@ -27,6 +27,8 @@ import { Dict, isName, Ref, RefSet } from "./primitives.js";
 import { BaseStream } from "./base_stream.js";
 
 const PDF_VERSION_REGEXP = /^[1-9]\.\d$/;
+const MAX_INT_32 = 2 ** 31 - 1;
+const MIN_INT_32 = -(2 ** 31);
 
 function getLookupTableFactory(initializer) {
   let lookup;
@@ -713,6 +715,8 @@ export {
   lookupMatrix,
   lookupNormalRect,
   lookupRect,
+  MAX_INT_32,
+  MIN_INT_32,
   MissingDataException,
   numberToString,
   ParserEOFException,

src/core/image_resizer.js

@@ -14,6 +14,8 @@
  */
 
 import { FeatureTest, ImageKind, shadow, warn } from "../shared/util.js";
+import { convertToRGBA } from "../shared/image_utils.js";
+import { MAX_INT_32 } from "./core_utils.js";
 
 const MIN_IMAGE_DIM = 2048;
 
@@ -172,6 +174,18 @@ class ImageResizer {
   }
 
   async _createImage() {
+    const { _imgData: imgData } = this;
+    const { width, height } = imgData;
+
+    if (width * height * 4 > MAX_INT_32) {
+      // The resulting RGBA image is too large.
+      // We just rescale the data.
+      const result = this.#rescaleImageData();
+      if (result) {
+        return result;
+      }
+    }
+
     const data = this._encodeBMP();
     let decoder, imagePromise;
 
@@ -206,8 +220,6 @@ class ImageResizer {
     }
 
     const { MAX_AREA, MAX_DIM } = ImageResizer;
-    const { _imgData: imgData } = this;
-    const { width, height } = imgData;
     const minFactor = Math.max(
       width / MAX_DIM,
       height / MAX_DIM,
@@ -268,6 +280,91 @@ class ImageResizer {
     return imgData;
   }
 
+  #rescaleImageData() {
+    const { _imgData: imgData } = this;
+    const { data, width, height, kind } = imgData;
+    const rgbaSize = width * height * 4;
+    // K is such as width * height * 4 / 2 ** K <= 2 ** 31 - 1
+    const K = Math.ceil(Math.log2(rgbaSize / MAX_INT_32));
+    const newWidth = width >> K;
+    const newHeight = height >> K;
+    let rgbaData;
+    let maxHeight = height;
+
+    // We try to allocate the buffer with the maximum size but it can fail.
+    try {
+      rgbaData = new Uint8Array(rgbaSize);
+    } catch {
+      // n is such as 2 ** n - 1 > width * height * 4
+      let n = Math.floor(Math.log2(rgbaSize + 1));
+
+      while (true) {
+        try {
+          rgbaData = new Uint8Array(2 ** n - 1);
+          break;
+        } catch {
+          n -= 1;
+        }
+      }
+
+      maxHeight = Math.floor((2 ** n - 1) / (width * 4));
+      const newSize = width * maxHeight * 4;
+      if (newSize < rgbaData.length) {
+        rgbaData = new Uint8Array(newSize);
+      }
+    }
+
+    const src32 = new Uint32Array(rgbaData.buffer);
+    const dest32 = new Uint32Array(newWidth * newHeight);
+
+    let srcPos = 0;
+    let newIndex = 0;
+    const step = Math.ceil(height / maxHeight);
+    const remainder = height % maxHeight === 0 ? height : height % maxHeight;
+    for (let k = 0; k < step; k++) {
+      const h = k < step - 1 ? maxHeight : remainder;
+      ({ srcPos } = convertToRGBA({
+        kind,
+        src: data,
+        dest: src32,
+        width,
+        height: h,
+        inverseDecode: this._isMask,
+        srcPos,
+      }));
+
+      for (let i = 0, ii = h >> K; i < ii; i++) {
+        const buf = src32.subarray((i << K) * width);
+        for (let j = 0; j < newWidth; j++) {
+          dest32[newIndex++] = buf[j << K];
+        }
+      }
+    }
+
+    if (ImageResizer.needsToBeResized(newWidth, newHeight)) {
+      imgData.data = dest32;
+      imgData.width = newWidth;
+      imgData.height = newHeight;
+      imgData.kind = ImageKind.RGBA_32BPP;
+
+      return null;
+    }
+
+    const canvas = new OffscreenCanvas(newWidth, newHeight);
+    const ctx = canvas.getContext("2d", { willReadFrequently: true });
+    ctx.putImageData(
+      new ImageData(new Uint8ClampedArray(dest32.buffer), newWidth, newHeight),
+      0,
+      0
+    );
+    imgData.data = null;
+    imgData.bitmap = canvas.transferToImageBitmap();
+    imgData.width = newWidth;
+    imgData.height = newHeight;
+
+    return imgData;
+  }
+
   _encodeBMP() {
     const { width, height, kind } = this._imgData;
     let data = this._imgData.data;

src/core/jbig2.js

@@ -14,7 +14,14 @@
  */
 
 import { BaseException, shadow } from "../shared/util.js";
-import { log2, readInt8, readUint16, readUint32 } from "./core_utils.js";
+import {
+  log2,
+  MAX_INT_32,
+  MIN_INT_32,
+  readInt8,
+  readUint16,
+  readUint32,
+} from "./core_utils.js";
 import { ArithmeticDecoder } from "./arithmetic_decoder.js";
 import { CCITTFaxDecoder } from "./ccitt.js";
 
@@ -52,9 +59,6 @@ class DecodingContext {
   }
 }
 
-const MAX_INT_32 = 2 ** 31 - 1;
-const MIN_INT_32 = -(2 ** 31);
-
 // Annex A. Arithmetic Integer Decoding Procedure
 // A.2 Procedure for decoding values
 function decodeInteger(contextCache, procedure, decoder) {

src/shared/image_utils.js

@@ -77,7 +77,8 @@ function convertRGBToRGBA({
   height,
 }) {
   let i = 0;
-  const len32 = src.length >> 2;
+  const len = width * height * 3;
+  const len32 = len >> 2;
   const src32 = new Uint32Array(src.buffer, srcPos, len32);
 
   if (FeatureTest.isLittleEndian) {
@@ -94,7 +95,7 @@ function convertRGBToRGBA({
       dest[destPos + 3] = (s3 >>> 8) | 0xff000000;
     }
 
-    for (let j = i * 4, jj = src.length; j < jj; j += 3) {
+    for (let j = i * 4, jj = srcPos + len; j < jj; j += 3) {
       dest[destPos++] =
         src[j] | (src[j + 1] << 8) | (src[j + 2] << 16) | 0xff000000;
     }
@@ -110,13 +111,13 @@ function convertRGBToRGBA({
       dest[destPos + 3] = (s3 << 8) | 0xff;
     }
 
-    for (let j = i * 4, jj = src.length; j < jj; j += 3) {
+    for (let j = i * 4, jj = srcPos + len; j < jj; j += 3) {
       dest[destPos++] =
         (src[j] << 24) | (src[j + 1] << 16) | (src[j + 2] << 8) | 0xff;
     }
   }
 
-  return { srcPos, destPos };
+  return { srcPos: srcPos + len, destPos };
 }
 
 function grayToRGBA(src, dest) {

test/pdfs/issue17190.pdf.link

@@ -0,0 +1 @@
+https://github.com/mozilla/pdf.js/files/13180762/org_AVA89V01U0.Black.pdf

test/pdfs/issue17190_1.pdf.link

@@ -0,0 +1 @@
+https://github.com/mozilla/pdf.js/files/13670664/abc.pdf

test/test_manifest.json

@@ -10775,5 +10775,23 @@
     "forms": true,
     "talos": false,
     "type": "eq"
+  },
+  {
+    "id": "issue17190",
+    "file": "pdfs/issue17190.pdf",
+    "md5": "06e3ce6492dc0b5815a63965b5d7144c",
+    "rounds": 1,
+    "talos": false,
+    "type": "eq",
+    "link": true
+  },
+  {
+    "id": "issue17190_1",
+    "file": "pdfs/issue17190_1.pdf",
+    "md5": "63bbc71a6c2cdec11bb20c5744232c47",
+    "rounds": 1,
+    "talos": false,
+    "type": "eq",
+    "link": true
   }
 ]