Cover fpdftext.h API with embeddertests.

fpdfsdk/src/fpdftext_embeddertest.cpp

 // Copyright 2015 PDFium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "../../testing/embedder_test.h"
 #include "../../fpdfsdk/include/fpdfview.h"
 #include "../../fpdfsdk/include/fpdftext.h"
 #include "testing/gtest/include/gtest/gtest.h"
 
+namespace {
+
+static bool check_unsigned_shorts(const char* chars,

[Lei Zhang, 2015/01/28 02:27:52]

Maybe rename this to "expected" and rename "buffer" to |actual| ?

[Tom Sepez, 2015/01/28 18:12:12]

Done.

+                                  const unsigned short* buffer,
+                                  int length) {

[Lei Zhang, 2015/01/28 02:27:52]

Why not make |length| a size_t and avoid the cast below?

[Tom Sepez, 2015/01/28 18:12:13]

Done. 

+  if (length > static_cast<int>(strlen(chars)) + 1) {
+    return false;
+  }
+  for (int i = 0; i < length; ++i) {
+    if (buffer[i] != static_cast<unsigned short>(chars[i])) {
+      return false;
+    }
+  }
+  return true;
+}
+
+static int three_places(double value) {
+  return static_cast<int>(1000.0 * value);
+}
+
+}  // namespace
+
 class FPDFTextEmbeddertest : public EmbedderTest {
 };
 
+TEST_F(FPDFTextEmbeddertest, Text) {
+  EXPECT_TRUE(OpenDocument("testing/resources/hello_world.pdf"));
+  FPDF_FORMHANDLE form_handle = SetFormFillEnvironment();

[Lei Zhang, 2015/01/28 02:27:51]

Maybe the FPDFTextEmbeddertest class should have SetUp() and TearDown() 
overrides that gets |form_handle| and cleans it up?

[Tom Sepez, 2015/01/28 18:12:12]

Problem is that creating a form_handle requires a document, which isn't present
at setup() time (until we call OpenDocument()). There probably is some cleanup
here; if we assume form_handle is 1:1 with document, then we can have
OpenDocument() also create it and TearDown() close it out for us.  That would
save some boilerplate.  I'll do that as a follow-up.

+  FPDF_PAGE page = LoadPage(0, form_handle);
+  EXPECT_NE(nullptr, page);
+
+  FPDF_TEXTPAGE textpage = FPDFText_LoadPage(page);
+  EXPECT_NE(nullptr, textpage);
+  EXPECT_EQ(30, FPDFText_CountChars(textpage));

[Lei Zhang, 2015/01/28 02:27:52]

Can we use sizeof(expected) instead of 30 and 31?

[Tom Sepez, 2015/01/28 18:12:13]

Done.

+
+  const char* expected = "Hello, world!\r\nGoodbye, world!";

[Lei Zhang, 2015/01/28 02:27:52]

const char expected[] = "...";

[Tom Sepez, 2015/01/28 18:12:13]

Ok, was trying to avoid a copy into the stack var, but its worth it to get
sizeof.

+  unsigned short fixed_buffer[128];
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(31, FPDFText_GetText(textpage, 0, 128, fixed_buffer));
+  EXPECT_TRUE(check_unsigned_shorts(expected, fixed_buffer, 31));
+
+  for (int i = 0; i < 31; ++i) {
+    EXPECT_EQ(expected[i], FPDFText_GetUnicode(textpage, i)) << " at " << i;
+  }
+
+  EXPECT_EQ(12.0, FPDFText_GetFontSize(textpage, 0));
+  EXPECT_EQ(16.0, FPDFText_GetFontSize(textpage, 15));
+
+  double left = 0.0;
+  double right = 0.0;
+  double bottom = 0.0;
+  double top = 0.0;
+  FPDFText_GetCharBox(textpage, 4, &left, &right, &bottom, &top);
+  EXPECT_EQ(three_places(41.071), three_places(left));

[Lei Zhang, 2015/01/28 02:27:52]

Why not just use EXPECT_FLOAT_EQ() and drop three_places()?

[Tom Sepez, 2015/01/28 18:12:13]

Ah, that's almost the magic I was looking for. EXPECT_NEAR will do the trick. 
Note that I'm just guessing here about the precision we'd tolerate from changes
to pixel layout computations.

+  EXPECT_EQ(three_places(46.243), three_places(right));
+  EXPECT_EQ(three_places(49.844), three_places(bottom));
+  EXPECT_EQ(three_places(55.520), three_places(top));
+
+  EXPECT_EQ(4, FPDFText_GetCharIndexAtPos(
+      textpage, 42.0, 50.0, 1.0, 1.0));
+  EXPECT_EQ(-1, FPDFText_GetCharIndexAtPos(
+      textpage, 0.0, 0.0, 1.0, 1.0));
+  EXPECT_EQ(-1, FPDFText_GetCharIndexAtPos(
+      textpage, 199.0, 199.0, 1.0, 1.0));
+
+  // Test out of range indicies.
+  EXPECT_EQ(-1, FPDFText_GetCharIndexAtPos(
+      textpage, 42.0, 10000000.0, 1.0, 1.0));
+  EXPECT_EQ(-1, FPDFText_GetCharIndexAtPos(
+      textpage, -1.0, 50.0, 1.0, 1.0));
+
+  EXPECT_EQ(2, FPDFText_CountRects(textpage, 0, 30));
+
+  left = 0.0;
+  right = 0.0;
+  bottom = 0.0;
+  top = 0.0;
+  FPDFText_GetRect(textpage, 1, &left, &top, &right, &bottom);
+  EXPECT_EQ(three_places(20.847), three_places(left));
+  EXPECT_EQ(three_places(135.167), three_places(right));
+  EXPECT_EQ(three_places(96.655), three_places(bottom));
+  EXPECT_EQ(three_places(116.000), three_places(top));
+
+  // Test out of range indicies.
+  left = 0.0;
+  right = 0.0;
+  bottom = 0.0;
+  top = 0.0;
+  FPDFText_GetRect(textpage, -1, &left, &top, &right, &bottom);
+  EXPECT_EQ(0.0, left);
+  EXPECT_EQ(0.0, right);
+  EXPECT_EQ(0.0, bottom);
+  EXPECT_EQ(0.0, top);
+
+  FPDFText_GetRect(textpage, 2, &left, &top, &right, &bottom);
+  EXPECT_EQ(0.0, left);
+  EXPECT_EQ(0.0, right);
+  EXPECT_EQ(0.0, bottom);
+  EXPECT_EQ(0.0, top);
+
+  EXPECT_EQ(9, FPDFText_GetBoundedText(
+      textpage, 41.0, 56.0, 82.0, 48.0, 0, 0));
+
+  // Extract starting at character 4 as above.
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(1, FPDFText_GetBoundedText(
+      textpage, 41.0, 56.0, 82.0, 48.0, fixed_buffer, 1));
+  EXPECT_TRUE(check_unsigned_shorts(expected + 4, fixed_buffer, 1));
+  EXPECT_EQ(0xbdbd, fixed_buffer[1]);
+
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(9, FPDFText_GetBoundedText(
+      textpage, 41.0, 56.0, 82.0, 48.0, fixed_buffer, 9));
+  EXPECT_TRUE(check_unsigned_shorts(expected + 4, fixed_buffer, 9));
+  EXPECT_EQ(0xbdbd, fixed_buffer[9]);
+
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(10, FPDFText_GetBoundedText(
+      textpage, 41.0, 56.0, 82.0, 48.0, fixed_buffer, 128));
+  EXPECT_TRUE(check_unsigned_shorts(expected + 4, fixed_buffer, 9));
+  EXPECT_EQ(0u, fixed_buffer[9]);
+  EXPECT_EQ(0xbdbd, fixed_buffer[10]);
+
+  FPDFText_ClosePage(textpage);
+  ClearFormFillEnvironment(form_handle);
+}
+
+TEST_F(FPDFTextEmbeddertest, TextSearch) {
+  EXPECT_TRUE(OpenDocument("testing/resources/hello_world.pdf"));
+  FPDF_FORMHANDLE form_handle = SetFormFillEnvironment();
+  FPDF_PAGE page = LoadPage(0, form_handle);
+  EXPECT_NE(nullptr, page);
+
+  FPDF_TEXTPAGE textpage = FPDFText_LoadPage(page);
+  EXPECT_NE(nullptr, textpage);
+
+  // Avoid issues with system wchar_t width vs. FPDF_WideString.
+  const unsigned short nope[] = { 'n', 'o', 'p', 'e', '\0' };
+  const unsigned short world[] = { 'w', 'o', 'r', 'l', 'd', '\0' };
+  const unsigned short world_caps[] = { 'W', 'O', 'R', 'L', 'D', '\0' };
+  const unsigned short world_substr[] = { 'o', 'r', 'l', 'd', '\0' };
+
+  // No occurances of "nope" in test page.
+  FPDF_SCHHANDLE search = FPDFText_FindStart(textpage, nope, 0, 0);
+  EXPECT_NE(nullptr, search);
+  EXPECT_EQ(0, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(0, FPDFText_GetSchCount(search));
+
+  // Advancing finds nothing.
+  EXPECT_FALSE(FPDFText_FindNext(search));
+  EXPECT_EQ(0, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(0, FPDFText_GetSchCount(search));
+
+  // Retreating finds nothing.
+  EXPECT_FALSE(FPDFText_FindPrev(search));
+  EXPECT_EQ(0, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(0, FPDFText_GetSchCount(search));
+  FPDFText_FindClose(search);
+
+  // Two occurances of "world" in test page.

[Lei Zhang, 2015/01/28 02:27:51]

occurrences, ditto below

[Tom Sepez, 2015/01/28 18:12:13]

Done.

+  search = FPDFText_FindStart(textpage, world, 0, 2);
+  EXPECT_NE(nullptr, search);
+
+  // Remains not found until advanced.
+  EXPECT_EQ(0, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(0, FPDFText_GetSchCount(search));
+
+  // First occurance of "world" in this test page.
+  EXPECT_TRUE(FPDFText_FindNext(search));
+  EXPECT_EQ(7, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+
+  // Last occurance of "world" in this test page.
+  EXPECT_TRUE(FPDFText_FindNext(search));
+  EXPECT_EQ(24, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+
+  // Found position unchanged when fails to advance.
+  EXPECT_FALSE(FPDFText_FindNext(search));
+  EXPECT_EQ(24, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+
+  // Back to first occurance.
+  EXPECT_TRUE(FPDFText_FindPrev(search));
+  EXPECT_EQ(7, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+
+  // Found position unchanged when fails to retreat.
+  EXPECT_FALSE(FPDFText_FindPrev(search));
+  EXPECT_EQ(7, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+  FPDFText_FindClose(search);
+
+  // Exact search unaffected by case sensitiity and whole word flags.
+  search = FPDFText_FindStart(
+      textpage, world, FPDF_MATCHCASE | FPDF_MATCHWHOLEWORD, 0);
+  EXPECT_NE(nullptr, search);
+  EXPECT_TRUE(FPDFText_FindNext(search));
+  EXPECT_EQ(7, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+  FPDFText_FindClose(search);
+
+  // Default is case-insensitive, so matching agaist caps works.
+  search = FPDFText_FindStart(textpage, world_caps, 0, 0);
+  EXPECT_NE(nullptr, search);
+  EXPECT_TRUE(FPDFText_FindNext(search));
+  EXPECT_EQ(7, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(5, FPDFText_GetSchCount(search));
+  FPDFText_FindClose(search);
+
+  // But can be made case sensitive, in which case this fails.
+  search = FPDFText_FindStart(textpage, world_caps, FPDF_MATCHCASE, 0);
+  EXPECT_FALSE(FPDFText_FindNext(search));
+  EXPECT_EQ(0, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(0, FPDFText_GetSchCount(search));
+  FPDFText_FindClose(search);
+
+  // Default is match anywhere within word, so matching substirng works.
+  search = FPDFText_FindStart(textpage, world_substr, 0, 0);
+  EXPECT_TRUE(FPDFText_FindNext(search));
+  EXPECT_EQ(8, FPDFText_GetSchResultIndex(search));
+  EXPECT_EQ(4, FPDFText_GetSchCount(search));
+  FPDFText_FindClose(search);
+
+  // But can be made to mach word boundaries, in which case this fails.
+  search = FPDFText_FindStart(textpage, world_substr, FPDF_MATCHWHOLEWORD, 0);
+  EXPECT_FALSE(FPDFText_FindNext(search));
+  // TODO(tsepez): investigate strange index/count values in this state.
+  FPDFText_FindClose(search);
+
+  FPDFText_ClosePage(textpage);
+  ClearFormFillEnvironment(form_handle);
+}
+
 // Test that the page has characters despite a bad stream length.
 TEST_F(FPDFTextEmbeddertest, StreamLengthPastEndOfFile) {
   EXPECT_TRUE(OpenDocument("testing/resources/bug_57.pdf"));
   FPDF_FORMHANDLE form_handle = SetFormFillEnvironment();
   FPDF_PAGE page = LoadPage(0, form_handle);
   EXPECT_NE(nullptr, page);
+
   FPDF_TEXTPAGE textpage = FPDFText_LoadPage(page);
   EXPECT_NE(nullptr, textpage);
   EXPECT_EQ(13, FPDFText_CountChars(textpage));
+
+  FPDFText_ClosePage(textpage);
   ClearFormFillEnvironment(form_handle);
 }
+
+TEST_F(FPDFTextEmbeddertest, WebLinks) {
+  EXPECT_TRUE(OpenDocument("testing/resources/weblinks.pdf"));
+  FPDF_FORMHANDLE form_handle = SetFormFillEnvironment();
+  FPDF_PAGE page = LoadPage(0, form_handle);
+  EXPECT_NE(nullptr, page);
+
+  FPDF_TEXTPAGE textpage = FPDFText_LoadPage(page);
+  EXPECT_NE(nullptr, textpage);
+
+  FPDF_PAGELINK pagelink = FPDFLink_LoadWebLinks(textpage);
+  EXPECT_NE(nullptr, pagelink);
+
+  // Page contains two HTTP-style URLs.
+  EXPECT_EQ(2, FPDFLink_CountWebLinks(pagelink));
+
+  // Only a terminating NUL required for bogus links.
+  EXPECT_EQ(1, FPDFLink_GetURL(pagelink, 2, nullptr, 0));
+  EXPECT_EQ(1, FPDFLink_GetURL(pagelink, 1400, nullptr, 0));
+  EXPECT_EQ(1, FPDFLink_GetURL(pagelink, -1, nullptr, 0));
+
+  // Query the number of characters required for each link (incl NUL).
+  EXPECT_EQ(25, FPDFLink_GetURL(pagelink, 0, nullptr, 0));
+  EXPECT_EQ(26, FPDFLink_GetURL(pagelink, 1, nullptr, 0));
+
+  const char* expected_url = "http://example.com?q=foo";

[Lei Zhang, 2015/01/28 02:27:52]

also const char foo[]

[Tom Sepez, 2015/01/28 18:12:13]

Done.

+  unsigned short fixed_buffer[128];
+
+  // Retrieve a link with too small a buffer.  Buffer will not be
+  // NUL-terminated, but must not be modified past indicated length,
+  // so pre-fill with a pattern to check write bounds.
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(1, FPDFLink_GetURL(pagelink, 0, fixed_buffer, 1));
+  EXPECT_TRUE(check_unsigned_shorts(expected_url, fixed_buffer, 1));
+  EXPECT_EQ(0xbdbd, fixed_buffer[1]);
+
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(24, FPDFLink_GetURL(pagelink, 0, fixed_buffer, 24));
+  EXPECT_TRUE(check_unsigned_shorts(expected_url, fixed_buffer, 24));
+  EXPECT_EQ(0xbdbd, fixed_buffer[24]);
+
+  // Retreive link with ample-sized-buffer.
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(25, FPDFLink_GetURL(pagelink, 0, fixed_buffer, 25));
+  EXPECT_TRUE(check_unsigned_shorts(expected_url, fixed_buffer, 25));
+  EXPECT_EQ(0u, fixed_buffer[24]);
+  EXPECT_EQ(0xbdbd, fixed_buffer[25]);
+
+  memset(fixed_buffer, 0xbd, sizeof(fixed_buffer));
+  EXPECT_EQ(25, FPDFLink_GetURL(pagelink, 0, fixed_buffer, 128));
+  EXPECT_TRUE(check_unsigned_shorts(expected_url, fixed_buffer, 25));
+  EXPECT_EQ(0u, fixed_buffer[24]);
+  EXPECT_EQ(0xbdbd, fixed_buffer[25]);
+
+  // Each link rendered in a single rect in this test page.
+  EXPECT_EQ(1, FPDFLink_CountRects(pagelink, 0));
+  EXPECT_EQ(1, FPDFLink_CountRects(pagelink, 1));
+
+  // Each link rendered in a single rect in this test page.
+  EXPECT_EQ(0, FPDFLink_CountRects(pagelink, -1));
+  EXPECT_EQ(0, FPDFLink_CountRects(pagelink, 2));
+  EXPECT_EQ(0, FPDFLink_CountRects(pagelink, 10000));
+
+  // Check boundary of valid link with valid rect index.
+  double left = 0.0;
+  double right = 0.0;
+  double top = 0.0;
+  double bottom = 0.0;
+  FPDFLink_GetRect(pagelink, 0, 0, &left, &top, &right, &bottom);
+  EXPECT_EQ(three_places(50.791), three_places(left));
+  EXPECT_EQ(three_places(187.963), three_places(right));
+  EXPECT_EQ(three_places(97.624), three_places(bottom));
+  EXPECT_EQ(three_places(108.736), three_places(top));
+
+  // Check boundary of valid link with invalid rect index.
+  left = 0.0;
+  right = 0.0;
+  top = 0.0;
+  bottom = 0.0;
+  FPDFLink_GetRect(pagelink, 0, 1, &left, &top, &right, &bottom);
+  EXPECT_EQ(0.0, left);
+  EXPECT_EQ(0.0, right);
+  EXPECT_EQ(0.0, bottom);
+  EXPECT_EQ(0.0, top);
+
+  // Check bounardy of invalid link.
+  left = 0.0;
+  right = 0.0;
+  top = 0.0;
+  bottom = 0.0;
+  FPDFLink_GetRect(pagelink, -1, 0, &left, &top, &right, &bottom);
+  EXPECT_EQ(0.0, left);
+  EXPECT_EQ(0.0, right);
+  EXPECT_EQ(0.0, bottom);
+  EXPECT_EQ(0.0, top);
+
+  FPDFLink_CloseWebLinks(pagelink);
+  FPDFText_ClosePage(textpage);
+  ClearFormFillEnvironment(form_handle);
+}