From 3bee9c60f013b8b7e99c39ee35699d132b330334 Mon Sep 17 00:00:00 2001 From: Miklos Vajna Date: Tue, 7 Aug 2018 21:45:34 +0000 Subject: Add FPDFTextObj_GetText() API Generalize CPDF_TextPage::GetTextByRect(), so that it's possible to get the text from a text page using a predicate, that way we can easily get the text that belongs to single text object as well. Change-Id: Ia457af0f41184694dc1481709be72b35685bce7f Reviewed-on: https://pdfium-review.googlesource.com/39530 Reviewed-by: Henrique Nakashima Reviewed-by: Lei Zhang Commit-Queue: Lei Zhang --- fpdfsdk/fpdf_text_embeddertest.cpp | 45 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 45 insertions(+) (limited to 'fpdfsdk/fpdf_text_embeddertest.cpp') diff --git a/fpdfsdk/fpdf_text_embeddertest.cpp b/fpdfsdk/fpdf_text_embeddertest.cpp index eafe1a2d11..bf064d672a 100644 --- a/fpdfsdk/fpdf_text_embeddertest.cpp +++ b/fpdfsdk/fpdf_text_embeddertest.cpp @@ -5,6 +5,7 @@ #include #include #include +#include #include "core/fxcrt/fx_memory.h" #include "public/cpp/fpdf_scopers.h" @@ -762,6 +763,50 @@ TEST_F(FPDFTextEmbeddertest, CountRects) { UnloadPage(page); } +TEST_F(FPDFTextEmbeddertest, GetText) { + ASSERT_TRUE(OpenDocument("hello_world.pdf")); + FPDF_PAGE page = LoadPage(0); + ASSERT_TRUE(page); + + FPDF_TEXTPAGE text_page = FPDFText_LoadPage(page); + ASSERT_TRUE(text_page); + + EXPECT_EQ(2, FPDFPage_CountObjects(page)); + FPDF_PAGEOBJECT text_object = FPDFPage_GetObject(page, 0); + ASSERT_TRUE(text_object); + + // Positive testing. + constexpr char kHelloText[] = "Hello, world!"; + // Return value includes the terminating NUL that is provided. + constexpr unsigned long kHelloUTF16Size = FX_ArraySize(kHelloText) * 2; + constexpr wchar_t kHelloWideText[] = L"Hello, world!"; + unsigned long size = FPDFTextObj_GetText(text_object, text_page, nullptr, 0); + ASSERT_EQ(kHelloUTF16Size, size); + + std::vector buffer(size); + ASSERT_EQ(size, + FPDFTextObj_GetText(text_object, text_page, buffer.data(), size)); + ASSERT_EQ(kHelloWideText, GetPlatformWString(buffer.data())); + + // Negative testing. + ASSERT_EQ(0U, FPDFTextObj_GetText(nullptr, text_page, nullptr, 0)); + ASSERT_EQ(0U, FPDFTextObj_GetText(text_object, nullptr, nullptr, 0)); + ASSERT_EQ(0U, FPDFTextObj_GetText(nullptr, nullptr, nullptr, 0)); + + // Buffer is too small, ensure it's not modified. + buffer.resize(2); + buffer[0] = 'x'; + buffer[1] = '\0'; + size = + FPDFTextObj_GetText(text_object, text_page, buffer.data(), buffer.size()); + ASSERT_EQ(kHelloUTF16Size, size); + ASSERT_EQ('x', buffer[0]); + ASSERT_EQ('\0', buffer[1]); + + FPDFText_ClosePage(text_page); + UnloadPage(page); +} + TEST_F(FPDFTextEmbeddertest, CroppedText) { static constexpr int kPageCount = 4; static constexpr FS_RECTF kBoxes[kPageCount] = { -- cgit v1.2.3