JPEG(JFIF only) directly embedded into PDF

R=reed@google.com, djsollen@google.com, mtklein@google.com

Author: halcanary@google.com

Review URL: https://codereview.chromium.org/515493003
diff --git a/gyp/tests.gypi b/gyp/tests.gypi
index 91dedb3..c3b559e 100644
--- a/gyp/tests.gypi
+++ b/gyp/tests.gypi
@@ -143,6 +143,7 @@
     '../tests/OSPathTest.cpp',
     '../tests/ObjectPoolTest.cpp',
     '../tests/OnceTest.cpp',
+    '../tests/PDFJpegEmbedTest.cpp',
     '../tests/PDFPrimitivesTest.cpp',
     '../tests/PackBitsTest.cpp',
     '../tests/PaintTest.cpp',
diff --git a/src/pdf/SkPDFDevice.cpp b/src/pdf/SkPDFDevice.cpp
index 5946b9a..ec07a8f 100644
--- a/src/pdf/SkPDFDevice.cpp
+++ b/src/pdf/SkPDFDevice.cpp
@@ -2182,8 +2182,8 @@
         return;
     }
 
-    SkAutoTUnref<SkPDFImage> image(
-        SkPDFImage::CreateImage(*bitmap, subset, fEncoder));
+    SkAutoTUnref<SkPDFObject> image(
+            SkPDFCreateImageObject(*bitmap, subset, fEncoder));
     if (!image) {
         return;
     }
diff --git a/src/pdf/SkPDFImage.cpp b/src/pdf/SkPDFImage.cpp
index 715fb48..4946051 100644
--- a/src/pdf/SkPDFImage.cpp
+++ b/src/pdf/SkPDFImage.cpp
@@ -13,6 +13,7 @@
 #include "SkData.h"
 #include "SkFlate.h"
 #include "SkPDFCatalog.h"
+#include "SkPixelRef.h"
 #include "SkRect.h"
 #include "SkStream.h"
 #include "SkString.h"
@@ -628,3 +629,91 @@
     }
     return true;
 }
+
+namespace {
+/**
+ *  This PDFObject assumes that its constructor was handed
+ *  Jpeg-encoded data that can be directly embedded into a PDF.
+ */
+class PDFJPEGImage : public SkPDFObject {
+    SkAutoTUnref<SkData> fData;
+    int fWidth;
+    int fHeight;
+public:
+    PDFJPEGImage(SkData* data, int width, int height)
+        : fData(SkRef(data)), fWidth(width), fHeight(height) {}
+    virtual void getResources(const SkTSet<SkPDFObject*>&,
+                              SkTSet<SkPDFObject*>*) SK_OVERRIDE {}
+    virtual void emitObject(
+            SkWStream* stream,
+            SkPDFCatalog* catalog, bool indirect) SK_OVERRIDE {
+        if (indirect) {
+            this->emitIndirectObject(stream, catalog);
+            return;
+        }
+        SkASSERT(fData.get());
+        const char kPrefaceFormat[] =
+            "<<"
+            "/Type /XObject\n"
+            "/Subtype /Image\n"
+            "/Width %d\n"
+            "/Height %d\n"
+            "/ColorSpace /DeviceRGB\n"
+            "/BitsPerComponent 8\n"
+            "/Filter /DCTDecode\n"
+            "/ColorTransform 0\n"
+            "/Length " SK_SIZE_T_SPECIFIER "\n"
+            ">> stream\n";
+        SkString preface(
+                SkStringPrintf(kPrefaceFormat, fWidth, fHeight, fData->size()));
+        const char kPostface[] = "\nendstream";
+        stream->write(preface.c_str(), preface.size());
+        stream->write(fData->data(), fData->size());
+        stream->write(kPostface, sizeof(kPostface));
+    }
+};
+
+/**
+ *  If the bitmap is not subsetted, return its encoded data, if
+ *  availible.
+ */
+static inline SkData* ref_encoded_data(const SkBitmap& bm) {
+    if ((NULL == bm.pixelRef())
+        || !bm.pixelRefOrigin().isZero()
+        || (bm.info().dimensions() != bm.pixelRef()->info().dimensions())) {
+        return NULL;
+    }
+    return bm.pixelRef()->refEncodedData();
+}
+
+/*
+ *  This functions may give false negatives but no false positives.
+ */
+static bool is_jfif_jpeg(SkData* data) {
+    if (!data || (data->size() < 11)) {
+        return false;
+    }
+    const uint8_t bytesZeroToThree[] = {0xFF, 0xD8, 0xFF, 0xE0};
+    const uint8_t bytesSixToTen[] = {'J', 'F', 'I', 'F', 0};
+    // 0   1   2   3   4   5   6   7   8   9   10
+    // FF  D8  FF  E0  ??  ??  'J' 'F' 'I' 'F' 00 ...
+    return ((0 == memcmp(data->bytes(), bytesZeroToThree,
+                         sizeof(bytesZeroToThree)))
+            && (0 == memcmp(data->bytes() + 6, bytesSixToTen,
+                            sizeof(bytesSixToTen))));
+}
+}  // namespace
+
+SkPDFObject* SkPDFCreateImageObject(
+        const SkBitmap& bitmap,
+        const SkIRect& subset,
+        SkPicture::EncodeBitmap encoder) {
+    if (SkIRect::MakeWH(bitmap.width(), bitmap.height()) == subset) {
+        SkAutoTUnref<SkData> encodedData(ref_encoded_data(bitmap));
+        if (is_jfif_jpeg(encodedData)) {
+            return SkNEW_ARGS(PDFJPEGImage,
+                              (encodedData, bitmap.width(), bitmap.height()));
+        }
+    }
+    return SkPDFImage::CreateImage(bitmap, subset, encoder);
+}
diff --git a/src/pdf/SkPDFImage.h b/src/pdf/SkPDFImage.h
index 10f8be6..5c026da 100644
--- a/src/pdf/SkPDFImage.h
+++ b/src/pdf/SkPDFImage.h
@@ -17,9 +17,20 @@
 #include "SkRefCnt.h"
 
 class SkBitmap;
+class SkData;
 class SkPDFCatalog;
 struct SkIRect;
 
+/**
+ *  Return the mose efficient availible encoding of the given bitmap.
+ *
+ *  If the bitmap has encoded JPEG data and that data can be embedded
+ *  into the PDF output stream directly, use that.  Otherwise, fall
+ *  back on SkPDFImage::CreateImage.
+ */
+SkPDFObject* SkPDFCreateImageObject(
+        const SkBitmap&, const SkIRect& subset, SkPicture::EncodeBitmap);
+
 /** \class SkPDFImage
 
     An image XObject.
diff --git a/tests/PDFJpegEmbedTest.cpp b/tests/PDFJpegEmbedTest.cpp
new file mode 100644
index 0000000..c1d0ea8
--- /dev/null
+++ b/tests/PDFJpegEmbedTest.cpp
@@ -0,0 +1,101 @@
+/*
+ * Copyright 2014 Google Inc.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE file.
+ */
+
+#include "SkDocument.h"
+#include "SkCanvas.h"
+#include "SkImageGenerator.h"
+#include "SkData.h"
+#include "SkStream.h"
+#include "SkDecodingImageGenerator.h"
+
+#include "Resources.h"
+#include "Test.h"
+
+// Returned bitmap is lazy.  Only lazy bitmaps hold onto the original data.
+static SkBitmap bitmap_from_data(SkData* data) {
+    SkASSERT(data);
+    SkBitmap bm;
+    SkInstallDiscardablePixelRef(
+            SkDecodingImageGenerator::Create(
+                    data, SkDecodingImageGenerator::Options()), &bm);
+    return bm;
+}
+
+static bool is_subset_of(SkData* smaller, SkData* larger) {
+    SkASSERT(smaller && larger);
+    if (smaller->size() > larger->size()) {
+        return false;
+    }
+    size_t size = smaller->size();
+    size_t size_diff = larger->size() - size;
+    for (size_t i = 0; i <= size_diff; ++i) {
+        if (0 == memcmp(larger->bytes() + i, smaller->bytes(), size)) {
+            return true;
+        }
+    }
+    return false;
+}
+
+
+static SkData* load_resource(
+        skiatest::Reporter* r, const char* test, const char* filename) {
+    SkString path(GetResourcePath(filename));
+    SkData* data = SkData::NewFromFileName(path.c_str());
+    if (!data && r->verbose()) {
+        SkDebugf("\n%s: Resource '%s' can not be found.\n",
+                 test, filename);
+    }
+    return data;  // May return NULL.
+}
+
+/**
+ *  Test that for Jpeg files that use the JFIF colorspace, they are
+ *  directly embedded into the PDF (without re-encoding) when that
+ *  makes sense.
+ */
+DEF_TEST(PDFJpegEmbedTest, r) {
+    const char test[] = "PDFJpegEmbedTest";
+    SkAutoTUnref<SkData> mandrillData(
+            load_resource(r, test, "mandrill_512_q075.jpg"));
+    SkAutoTUnref<SkData> cmykData(load_resource(r, test, "CMYK.jpg"));
+    if (!mandrillData || !cmykData) {
+        return;
+    }
+
+    SkDynamicMemoryWStream pdf;
+    SkAutoTUnref<SkDocument> document(SkDocument::CreatePDF(&pdf));
+    SkCanvas* canvas = document->beginPage(642, 1028);
+
+    canvas->clear(SK_ColorLTGRAY);
+
+    SkBitmap bm1(bitmap_from_data(mandrillData));
+    canvas->drawBitmap(bm1, 65.0, 0.0, NULL);
+    SkBitmap bm2(bitmap_from_data(cmykData));
+    canvas->drawBitmap(bm2, 0.0, 512.0, NULL);
+
+    canvas->flush();
+    document->endPage();
+    document->close();
+    SkAutoTUnref<SkData> pdfData(pdf.copyToData());
+    SkASSERT(pdfData);
+    pdf.reset();
+
+    REPORTER_ASSERT(r, is_subset_of(mandrillData, pdfData));
+
+    // This JPEG uses a nonstandard colorspace - it can not be
+    // embedded into the PDF directly.
+    REPORTER_ASSERT(r, !is_subset_of(cmykData, pdfData));
+
+    // The following is for debugging purposes only.
+    const char* outputPath = getenv("SKIA_TESTS_PDF_JPEG_EMBED_OUTPUT_PATH");
+    if (outputPath) {
+        SkFILEWStream output(outputPath);
+        if (output.isValid()) {
+            output.write(pdfData->data(), pdfData->size());
+        }
+    }
+}