Skip to content

armgit / ente

Public Imported from GitHub
0

πŸ’š End-to-end encrypted cloud for everything.

Limit recovery sheet scanner resource use

Neeraj Gupta930a775b1a42
diff --git a/web/apps/legacy/src/features/legacy-kit/scan.ts b/web/apps/legacy/src/features/legacy-kit/scan.ts
index ae471e80d7..7d01d975e2 100644
--- a/web/apps/legacy/src/features/legacy-kit/scan.ts
+++ b/web/apps/legacy/src/features/legacy-kit/scan.ts
@@ -4,7 +4,20 @@ type DecodeQR = (typeof import("qr/decode.js"))["default"];
const legacyKitShareMetadataPrefix = "ente-legacy-kit-share-v1:";
const legacyKitBundleMetadataPrefix = "ente-legacy-kit-shares-v1:";
+const maxTextFileBytes = 64 * 1024;
+const maxPDFFileBytes = 32 * 1024 * 1024;
+const maxImageFileBytes = 32 * 1024 * 1024;
+const maxInputImagePixels = 64 * 1024 * 1024;
+const maxInputImageSide = 16_000;
+const maxCanvasPixels = 6 * 1024 * 1024;
+const maxCanvasSide = 2_400;
+const maxPDFPagesToScan = 5;
+const maxPDFTextChars = 128 * 1024;
const pdfFallbackTimeoutMS = 15_000;
+const pdfPageTimeoutMS = 5_000;
+
+const tooLargeMessage =
+ "This recovery sheet is too large to scan safely. Try the original PDF or a smaller, clearer photo.";
interface DecodableImage {
data: Uint8ClampedArray | number[];
@@ -56,6 +69,12 @@ const withTimeout = async <T>(
}
};
+const assertFileSize = (file: File, maxBytes: number) => {
+ if (file.size > maxBytes) {
+ throw new Error(tooLargeMessage);
+ }
+};
+
const findJsonObject = (value: string) => {
const start = value.indexOf('{"pv"');
if (start < 0) return undefined;
@@ -174,9 +193,26 @@ const loadDrawableFromFile = async (file: File): Promise<DrawableImage> => {
};
const imageDataFromDrawable = (drawable: DrawableImage) => {
+ if (
+ drawable.width <= 0 ||
+ drawable.height <= 0 ||
+ drawable.width > maxInputImageSide ||
+ drawable.height > maxInputImageSide ||
+ drawable.width * drawable.height > maxInputImagePixels
+ ) {
+ throw new Error(tooLargeMessage);
+ }
+
+ const scale = Math.min(
+ 1,
+ maxCanvasSide / Math.max(drawable.width, drawable.height),
+ Math.sqrt(maxCanvasPixels / (drawable.width * drawable.height)),
+ );
+ const width = Math.max(1, Math.floor(drawable.width * scale));
+ const height = Math.max(1, Math.floor(drawable.height * scale));
const canvas = document.createElement("canvas");
- canvas.width = drawable.width;
- canvas.height = drawable.height;
+ canvas.width = width;
+ canvas.height = height;
const context = canvas.getContext("2d", { willReadFrequently: true });
if (!context) {
throw new Error("Could not read that image.");
@@ -270,40 +306,92 @@ const decodeQrImage = async (image: DecodableImage) => {
const textFromPDF = async (pdf: PDFDocumentProxy) => {
const chunks: string[] = [];
- for (let pageNumber = 1; pageNumber <= pdf.numPages; pageNumber++) {
+ let charCount = 0;
+ const pageCount = Math.min(pdf.numPages, maxPDFPagesToScan);
+ for (let pageNumber = 1; pageNumber <= pageCount; pageNumber++) {
const page = await pdf.getPage(pageNumber);
- const textContent = await page.getTextContent();
- chunks.push(
- textContent.items
+ try {
+ const textContent = await withTimeout(
+ page.getTextContent(),
+ pdfPageTimeoutMS,
+ "Could not read that PDF page.",
+ );
+ const pageText = textContent.items
.map((item) => ("str" in item ? item.str : ""))
- .join(""),
- );
+ .join("")
+ .slice(0, maxPDFTextChars - charCount);
+ chunks.push(pageText);
+ charCount += pageText.length;
+ } finally {
+ page.cleanup();
+ }
+ if (charCount >= maxPDFTextChars) {
+ break;
+ }
}
return chunks.join("\n");
};
-const renderPDFPage = async (page: PDFPageProxy) => {
+const pdfRenderScale = (page: PDFPageProxy) => {
const baseViewport = page.getViewport({ scale: 1 });
- const scale = Math.min(
+ if (
+ !Number.isFinite(baseViewport.width) ||
+ !Number.isFinite(baseViewport.height) ||
+ baseViewport.width <= 0 ||
+ baseViewport.height <= 0
+ ) {
+ throw new Error(tooLargeMessage);
+ }
+ const pixelScale = Math.sqrt(
+ maxCanvasPixels / (baseViewport.width * baseViewport.height),
+ );
+ return Math.min(
3,
- Math.max(2, 1800 / Math.max(baseViewport.width, baseViewport.height)),
+ maxCanvasSide / Math.max(baseViewport.width, baseViewport.height),
+ pixelScale,
);
- const viewport = page.getViewport({ scale });
+};
+
+const renderPDFPage = async (page: PDFPageProxy) => {
+ const viewport = page.getViewport({ scale: pdfRenderScale(page) });
+ const width = Math.max(1, Math.ceil(viewport.width));
+ const height = Math.max(1, Math.ceil(viewport.height));
+ if (width * height > maxCanvasPixels) {
+ throw new Error(tooLargeMessage);
+ }
+
const canvas = document.createElement("canvas");
- canvas.width = Math.ceil(viewport.width);
- canvas.height = Math.ceil(viewport.height);
+ canvas.width = width;
+ canvas.height = height;
const context = canvas.getContext("2d", { willReadFrequently: true });
if (!context) {
throw new Error("Could not render that PDF.");
}
- await page.render({ canvas, canvasContext: context, viewport }).promise;
+ const renderTask = page.render({
+ canvas,
+ canvasContext: context,
+ viewport,
+ });
+ try {
+ await withTimeout(
+ renderTask.promise,
+ pdfPageTimeoutMS,
+ "Could not render that PDF page.",
+ );
+ } catch (error) {
+ renderTask.cancel();
+ throw error;
+ }
return context.getImageData(0, 0, canvas.width, canvas.height);
};
const decodeQrFromPDFBytes = async (bytes: Uint8Array) => {
const pdfjs =
(await import("pdfjs-dist/legacy/webpack.mjs")) as unknown as typeof import("pdfjs-dist");
- const loadingTask = pdfjs.getDocument({ data: bytes.slice() });
+ const loadingTask = pdfjs.getDocument({
+ data: bytes.slice(),
+ disableFontFace: true,
+ });
let pdf: PDFDocumentProxy;
try {
pdf = await withTimeout(
@@ -317,17 +405,25 @@ const decodeQrFromPDFBytes = async (bytes: Uint8Array) => {
}
try {
- const textCode = findJsonObject(await textFromPDF(pdf));
+ let textCode: string | undefined;
+ try {
+ textCode = findJsonObject(await textFromPDF(pdf));
+ } catch {
+ // Text extraction is a fast path; scanned PDFs should still use QR rendering.
+ }
if (textCode) {
return textCode;
}
- for (let pageNumber = 1; pageNumber <= pdf.numPages; pageNumber++) {
+ const pageCount = Math.min(pdf.numPages, maxPDFPagesToScan);
+ for (let pageNumber = 1; pageNumber <= pageCount; pageNumber++) {
const page = await pdf.getPage(pageNumber);
try {
return await decodeQrImage(await renderPDFPage(page));
} catch {
// Try the next page before surfacing a failure.
+ } finally {
+ page.cleanup();
}
}
} finally {
@@ -338,6 +434,7 @@ const decodeQrFromPDFBytes = async (bytes: Uint8Array) => {
};
const decodeQrFromPDF = async (file: File) => {
+ assertFileSize(file, maxPDFFileBytes);
const bytes = new Uint8Array(await file.arrayBuffer());
const metadataCode = codeFromPDFMetadata(bytes);
if (metadataCode) return metadataCode;
@@ -347,8 +444,10 @@ const decodeQrFromPDF = async (file: File) => {
export const readLegacyKitCodeFromFile = async (file: File) => {
if (isPDF(file)) return decodeQrFromPDF(file);
if (isPlainText(file)) {
+ assertFileSize(file, maxTextFileBytes);
const value = await file.text();
return findJsonObject(value) ?? value.trim();
}
+ assertFileSize(file, maxImageFileBytes);
return decodeQrImage(await imageDataFromFile(file));
};