Unify object type detection code inside cpdf_document.cpp.
Use the code added by https://pdfium-review.googlesource.com/113892 in
CountPages() as well.
Change-Id: If11b435b07bd11e9f468b851c8d0a018b6fd2542
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/113991
Commit-Queue: Lei Zhang <thestig@chromium.org>
Reviewed-by: Nigi <nigi@chromium.org>
diff --git a/core/fpdfapi/parser/cpdf_document.cpp b/core/fpdfapi/parser/cpdf_document.cpp
index e12b7af..b22c704 100644
--- a/core/fpdfapi/parser/cpdf_document.cpp
+++ b/core/fpdfapi/parser/cpdf_document.cpp
@@ -35,6 +35,30 @@
const int kMaxPageLevel = 1024;
+enum class NodeType : bool {
+ kBranch, // /Type /Pages, AKA page tree node.
+ kLeaf, // /Type /Page, AKA page object.
+};
+
+// Note that this function may modify `kid_dict` to correct PDF spec violations.
+// Same reasoning as CountPages() below.
+NodeType GetNodeType(RetainPtr<CPDF_Dictionary> kid_dict) {
+ const ByteString kid_type_value = kid_dict->GetNameFor("Type");
+ if (kid_type_value == "Pages") {
+ return NodeType::kBranch;
+ }
+ if (kid_type_value == "Page") {
+ return NodeType::kLeaf;
+ }
+
+ // Even though /Type is required for page tree nodes and page objects, PDFs
+ // may not have them or have the wrong type. Tolerate these errors and guess
+ // the type. Then fix the in-memory representation.
+ const bool has_kids = kid_dict->KeyExist("Kids");
+ kid_dict->SetNewFor<CPDF_Name>("Type", has_kids ? "Pages" : "Page");
+ return has_kids ? NodeType::kBranch : NodeType::kLeaf;
+}
+
// Returns a value in the range [0, `CPDF_Document::kPageMaxNum`), or nullopt on
// error. Note that this function may modify `pages_dict` to correct PDF spec
// violations. By normalizing the in-memory representation, other code that
@@ -61,7 +85,9 @@
if (!kid_dict || pdfium::Contains(*visited_pages, kid_dict)) {
continue;
}
- if (kid_dict->KeyExist("Kids")) {
+
+ NodeType kid_type = GetNodeType(kid_dict);
+ if (kid_type == NodeType::kBranch) {
// Use |visited_pages| to help detect circular references of pages.
ScopedSetInsertion<RetainPtr<CPDF_Dictionary>> local_add(visited_pages,
kid_dict);
@@ -72,9 +98,10 @@
}
count += local_count.value();
} else {
- // This page is a leaf node.
+ CHECK_EQ(kid_type, NodeType::kLeaf);
count++;
}
+
if (count >= CPDF_Document::kPageMaxNum) {
return absl::nullopt; // Error: too many pages.
}
@@ -468,27 +495,8 @@
}
for (size_t i = 0; i < kids_list->size(); i++) {
- enum class NodeType : bool {
- kBranch, // /Type /Pages, AKA page tree node.
- kLeaf, // /Type /Page, AKA page object.
- };
-
RetainPtr<CPDF_Dictionary> kid_dict = kids_list->GetMutableDictAt(i);
- const ByteString kid_type_value = kid_dict->GetNameFor("Type");
- NodeType kid_type;
- if (kid_type_value == "Pages") {
- kid_type = NodeType::kBranch;
- } else if (kid_type_value == "Page") {
- kid_type = NodeType::kLeaf;
- } else {
- // Even though /Type is required for page tree nodes and page objects,
- // PDFs may not have them or have the wrong type. Tolerate these errors
- // and guess the type. Then fix the in-memory representation.
- const bool has_kids = kid_dict->KeyExist("Kids");
- kid_type = has_kids ? NodeType::kBranch : NodeType::kLeaf;
- kid_dict->SetNewFor<CPDF_Name>("Type", has_kids ? "Pages" : "Page");
- }
-
+ NodeType kid_type = GetNodeType(kid_dict);
if (kid_type == NodeType::kLeaf) {
if (pages_to_go != 0) {
pages_to_go--;
