Add method to get tagged type

This CL adds a FPDF_StructElement_GetType method to get the type of a given
tagged struct. This corresponds to the /S key in the structure element
dictionary. See PDF1.7 Table 10.10.

Bug: pdfium:672
Change-Id: I313eb886dc9623715995d73a76bce01ba9985e4b
Reviewed-on: https://pdfium-review.googlesource.com/3623
Commit-Queue: dsinclair <dsinclair@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>

4 files changed, 75 insertions, 15 deletions

diff --git a/fpdfsdk/fpdf_structtree.cpp b/fpdfsdk/fpdf_structtree.cpp
index 06713fba1d..8a93d2299d 100644
--- a/fpdfsdk/fpdf_structtree.cpp
+++ b/fpdfsdk/fpdf_structtree.cpp
@@ -21,6 +21,19 @@ IPDF_StructElement* ToStructTreeElement(FPDF_STRUCTELEMENT struct_element) {
   return reinterpret_cast<IPDF_StructElement*>(struct_element);
 }
 
+unsigned long WideStringToBuffer(const CFX_WideString& str,
+                                 void* buffer,
+                                 unsigned long buflen) {
+  if (str.IsEmpty())
+    return 0;
+
+  CFX_ByteString encodedStr = str.UTF16LE_Encode();
+  const unsigned long len = encodedStr.GetLength();
+  if (buffer && len <= buflen)
+    memcpy(buffer, encodedStr.c_str(), len);
+  return len;
+}
+
 }  // namespace
 
 DLLEXPORT FPDF_STRUCTTREE STDCALL FPDF_StructTree_GetForPage(FPDF_PAGE page) {
@@ -54,22 +67,19 @@ FPDF_StructElement_GetAltText(FPDF_STRUCTELEMENT struct_element,
                               void* buffer,
                               unsigned long buflen) {
   IPDF_StructElement* elem = ToStructTreeElement(struct_element);
-  if (!elem)
-    return 0;
-
-  CPDF_Dictionary* dict = elem->GetDict();
-  if (!dict)
-    return 0;
-
-  CFX_WideString str = elem->GetDict()->GetUnicodeTextFor("Alt");
-  if (str.IsEmpty())
-    return 0;
+  return (elem && elem->GetDict())
+             ? WideStringToBuffer(elem->GetDict()->GetUnicodeTextFor("Alt"),
+                                  buffer, buflen)
+             : 0;
+}
 
-  CFX_ByteString encodedStr = str.UTF16LE_Encode();
-  const unsigned long len = encodedStr.GetLength();
-  if (buffer && len <= buflen)
-    memcpy(buffer, encodedStr.c_str(), len);
-  return len;
+DLLEXPORT unsigned long STDCALL
+FPDF_StructElement_GetType(FPDF_STRUCTELEMENT struct_element,
+                           void* buffer,
+                           unsigned long buflen) {
+  IPDF_StructElement* elem = ToStructTreeElement(struct_element);
+  return elem ? WideStringToBuffer(elem->GetType().UTF8Decode(), buffer, buflen)
+              : 0;
 }
 
 DLLEXPORT int STDCALL
diff --git a/fpdfsdk/fpdf_structtree_embeddertest.cpp b/fpdfsdk/fpdf_structtree_embeddertest.cpp
index 8ddde5317c..3110988d82 100644
--- a/fpdfsdk/fpdf_structtree_embeddertest.cpp
+++ b/fpdfsdk/fpdf_structtree_embeddertest.cpp
@@ -68,3 +68,32 @@ TEST_F(FPDFStructTreeEmbeddertest, GetAltText) {
   FPDF_StructTree_Close(struct_tree);
   FPDF_ClosePage(page);
 }
+
+TEST_F(FPDFStructTreeEmbeddertest, GetType) {
+  ASSERT_TRUE(OpenDocument("tagged_alt_text.pdf"));
+  FPDF_PAGE page = LoadPage(0);
+  ASSERT_TRUE(page);
+
+  FPDF_STRUCTTREE struct_tree = FPDF_StructTree_GetForPage(page);
+  ASSERT_TRUE(struct_tree);
+  ASSERT_EQ(1, FPDF_StructTree_CountChildren(struct_tree));
+
+  FPDF_STRUCTELEMENT element = FPDF_StructTree_GetChildAtIndex(struct_tree, 0);
+  ASSERT_NE(nullptr, element);
+
+  unsigned short buffer[12];
+  memset(buffer, 0, sizeof(buffer));
+  // Deliberately pass in a small buffer size to make sure |buffer| remains
+  // untouched.
+  ASSERT_EQ(18U, FPDF_StructElement_GetType(element, buffer, 1));
+  for (size_t i = 0; i < FX_ArraySize(buffer); ++i)
+    EXPECT_EQ(0U, buffer[i]);
+
+  ASSERT_EQ(18U, FPDF_StructElement_GetType(element, buffer, sizeof(buffer)));
+  const wchar_t kExpected[] = L"Document";
+  EXPECT_EQ(CFX_WideString(kExpected),
+            CFX_WideString::FromUTF16LE(buffer, FXSYS_len(kExpected)));
+
+  FPDF_StructTree_Close(struct_tree);
+  FPDF_ClosePage(page);
+}
diff --git a/fpdfsdk/fpdfview_c_api_test.c b/fpdfsdk/fpdfview_c_api_test.c
index fec91efeca..9afbdd479e 100644
--- a/fpdfsdk/fpdfview_c_api_test.c
+++ b/fpdfsdk/fpdfview_c_api_test.c
@@ -178,6 +178,7 @@ int CheckPDFiumCApi() {
     CHK(FPDF_StructTree_CountChildren);
     CHK(FPDF_StructTree_GetChildAtIndex);
     CHK(FPDF_StructElement_GetAltText);
+    CHK(FPDF_StructElement_GetType);
     CHK(FPDF_StructElement_CountChildren);
     CHK(FPDF_StructElement_GetChildAtIndex);
 
diff --git a/public/fpdf_structtree.h b/public/fpdf_structtree.h
index 3d4da402aa..82156365d9 100644
--- a/public/fpdf_structtree.h
+++ b/public/fpdf_structtree.h
@@ -73,6 +73,26 @@ FPDF_StructElement_GetAltText(FPDF_STRUCTELEMENT struct_element,
                               void* buffer,
                               unsigned long buflen);
 
+// Function: FPDF_StructElement_GetType
+//           Get the type (/S) for a given element.
+// Parameters:
+//           struct_element - Handle to the struct element.
+//           buffer        - A buffer for output. May be NULL.
+//           buflen        - The length of the buffer, in bytes. May be 0.
+// Return value:
+//           The number of bytes in the type, including the terminating NUL
+//           character. The number of bytes is returned regardless of the
+//           |buffer| and |buflen| parameters.
+// Comments:
+//           Regardless of the platform, the |buffer| is always in UTF-16LE
+//           encoding. The string is terminated by a UTF16 NUL character. If
+//           |buflen| is less than the required length, or |buffer| is NULL,
+//           |buffer| will not be modified.
+DLLEXPORT unsigned long STDCALL
+FPDF_StructElement_GetType(FPDF_STRUCTELEMENT struct_element,
+                           void* buffer,
+                           unsigned long buflen);
+
 // Function: FPDF_StructElement_CountChildren
 //          Count the number of children for the structure element.
 // Parameters: