From 389a26c115de8e6de8df800ddaddaac9d59e3355 Mon Sep 17 00:00:00 2001
From: Jonas Jenwald <jonas.jenwald@gmail.com>
Date: Thu, 10 Aug 2023 11:59:40 +0200
Subject: [PATCH] Fallback to check all pages when getting the pageIndex of
FieldObjects
Given that the FieldObjects are parsed in parallel, in combination with the existing caching in the `getPage`-method and `annotations`-getter, adding additional caches for this fallback code-path doesn't seem entirely necessary.
---
src/core/annotation.js | 38 ++++++++++++++++++++++++++++++--------
test/unit/api_spec.js | 36 ++++++++++++++++++++++++++++++++++++
2 files changed, 66 insertions(+), 8 deletions(-)
diff --git a/src/core/annotation.js b/src/core/annotation.js
index 347fa3ed6..5674b1b75 100644
--- a/src/core/annotation.js
+++ b/src/core/annotation.js
@@ -26,6 +26,7 @@ import {
FeatureTest,
getModificationDate,
IDENTITY_MATRIX,
+ info,
LINE_DESCENT_FACTOR,
LINE_FACTOR,
OPS,
@@ -52,7 +53,7 @@ import {
parseAppearanceStream,
parseDefaultAppearance,
} from "./default_appearance.js";
-import { Dict, isName, Name, Ref, RefSet } from "./primitives.js";
+import { Dict, isName, isRefsEqual, Name, Ref, RefSet } from "./primitives.js";
import { Stream, StringStream } from "./stream.js";
import { writeDict, writeObject } from "./writer.js";
import { BaseStream } from "./base_stream.js";
@@ -245,17 +246,38 @@ class AnnotationFactory {
return -1;
}
const pageRef = annotDict.getRaw("P");
- if (!(pageRef instanceof Ref)) {
- return -1;
+ if (pageRef instanceof Ref) {
+ try {
+ const pageIndex = await pdfManager.ensureCatalog("getPageIndex", [
+ pageRef,
+ ]);
+ return pageIndex;
+ } catch (ex) {
+ info(`_getPageIndex -- not a valid page reference: "${ex}".`);
+ }
+ }
+ if (annotDict.has("Kids")) {
+ return -1; // Not an annotation reference.
+ }
+ // Fallback to, potentially, checking the annotations of all pages.
+ // PLEASE NOTE: This could force the *entire* PDF document to load,
+ // hence it absolutely cannot be done unconditionally.
+ const numPages = await pdfManager.ensureDoc("numPages");
+
+ for (let pageIndex = 0; pageIndex < numPages; pageIndex++) {
+ const page = await pdfManager.getPage(pageIndex);
+ const annotations = await pdfManager.ensure(page, "annotations");
+
+ for (const annotRef of annotations) {
+ if (annotRef instanceof Ref && isRefsEqual(annotRef, ref)) {
+ return pageIndex;
+ }
+ }
}
- const pageIndex = await pdfManager.ensureCatalog("getPageIndex", [
- pageRef,
- ]);
- return pageIndex;
} catch (ex) {
warn(`_getPageIndex: "${ex}".`);
- return -1;
}
+ return -1;
}
static generateImages(annotations, xref, isOffscreenCanvasSupported) {
diff --git a/test/unit/api_spec.js b/test/unit/api_spec.js
index ac68f80aa..7c14dbee7 100644
--- a/test/unit/api_spec.js
+++ b/test/unit/api_spec.js
@@ -1485,6 +1485,42 @@ describe("api", function () {
await loadingTask.destroy();
});
+ it("gets fieldObjects with missing /P-entries", async function () {
+ if (isNodeJS) {
+ pending("Linked test-cases are not supported in Node.js.");
+ }
+
+ const loadingTask = getDocument(buildGetDocumentParams("bug1847733.pdf"));
+ const pdfDoc = await loadingTask.promise;
+ const fieldObjects = await pdfDoc.getFieldObjects();
+
+ for (const name in fieldObjects) {
+ const pageIndexes = fieldObjects[name].map(o => o.page);
+ let expected;
+
+ switch (name) {
+ case "formID":
+ case "pdf_submission_new":
+ case "simple_spc":
+ case "adobeWarning":
+ expected = [0];
+ break;
+ case "typeA13":
+ expected = [0, 0, 0, 0];
+ break;
+ case "typeA15[0]":
+ case "typeA15[1]":
+ case "typeA15[2]":
+ case "typeA15[3]":
+ expected = [-1, 0, 0, 0, 0];
+ break;
+ }
+ expect(pageIndexes).toEqual(expected);
+ }
+
+ await loadingTask.destroy();
+ });
+
it("check field object for group of buttons", async function () {
if (isNodeJS) {
pending("Linked test-cases are not supported in Node.js.");