Unify parsing of cross ref table v4.

We can use 'Prev' value of first-page cross-reference table trailer
for load main cross-reference table,
instead of 'T' value of Linearized header (Offset of first entry in main cross-reference table).

This is better solution, because this is allow us check entry count in main cross-ref table and
unify loading of main cross-ref table with loading methods of non linearized document.

See PDF specification:
http://wwwimages.adobe.com/content/dam/Adobe/en/devnet/pdf/pdfs/PDF32000_2008.pdf#678
(page 678, Example 3/part 3)

Change-Id: I59dcf3c73a0fb561221ded78e827e40535dbd717
Reviewed-on: https://pdfium-review.googlesource.com/13810
Commit-Queue: Art Snake <art-snake@yandex-team.ru>
Reviewed-by: dsinclair <dsinclair@chromium.org>

1 files changed, 16 insertions, 0 deletions

diff --git a/core/fpdfapi/parser/cpdf_parser_embeddertest.cpp b/core/fpdfapi/parser/cpdf_parser_embeddertest.cpp
index 99bc2c2d42..6aa3e2785a 100644
--- a/core/fpdfapi/parser/cpdf_parser_embeddertest.cpp
+++ b/core/fpdfapi/parser/cpdf_parser_embeddertest.cpp
@@ -57,3 +57,19 @@ TEST_F(CPDFParserEmbeddertest, Bug_602650) {
 TEST_F(CPDFParserEmbeddertest, Bug_757705) {
   EXPECT_TRUE(OpenDocument("bug_757705.pdf"));
 }
+
+TEST_F(CPDFParserEmbeddertest, LoadMainCrossRefTable) {
+  EXPECT_TRUE(OpenDocument("feature_linearized_loading.pdf", nullptr, true));
+  // To check that main cross ref table is loaded correctly,will be enough to
+  // check that the second page was correctly loaded. Because it is contains
+  // crossrefs for second page.
+  EXPECT_EQ(2, GetPageCount());
+  FPDF_PAGE page = LoadPage(1);
+  EXPECT_NE(nullptr, page);
+  FPDF_TEXTPAGE text_page = FPDFText_LoadPage(page);
+  EXPECT_NE(nullptr, text_page);
+  // The page should not be blank.
+  EXPECT_LT(0, FPDFText_CountChars(text_page));
+  FPDFText_ClosePage(text_page);
+  UnloadPage(page);
+}