Replace int flags with struct FXDIB_ResampleOptions.

Using bit values in an int may not be reliable, since different parts of
the code can interpret the bits differently. e.g. FXDIB_DOWNSAMPLE and
RENDER_FORCE_DOWNSAMPLE are defined in different places, but can be used
interchangeably because they just happen to have the same value. It
works but is rather fragile.

Instead, use a struct of bools to explicitly define what different bits
mean. Remove FXDIB_DOWNSAMPLE and friends.

Change-Id: I9cf0c8f94d1ed27edf8dba22b0ab0ee67f2722cc
Reviewed-on: https://pdfium-review.googlesource.com/c/44650
Commit-Queue: Lei Zhang <thestig@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>

1 files changed, 7 insertions, 2 deletions

diff --git a/core/fxge/dib/cstretchengine_unittest.cpp b/core/fxge/dib/cstretchengine_unittest.cpp
index 8169cae328..a79945c67d 100644
--- a/core/fxge/dib/cstretchengine_unittest.cpp
+++ b/core/fxge/dib/cstretchengine_unittest.cpp
@@ -27,6 +27,11 @@ TEST(CStretchEngine, OverflowInCtor) {
   auto dib_source = pdfium::MakeRetain<CPDF_DIBBase>();
   dib_source->Load(nullptr, stream.get());
   CStretchEngine engine(nullptr, FXDIB_8bppRgb, 500, 500, clip_rect, dib_source,
-                        0);
-  EXPECT_EQ(FXDIB_INTERPOL, engine.m_Flags);
+                        FXDIB_ResampleOptions());
+  EXPECT_FALSE(engine.m_ResampleOptions.bInterpolateDownsample);
+  EXPECT_TRUE(engine.m_ResampleOptions.bInterpolateBilinear);
+  EXPECT_FALSE(engine.m_ResampleOptions.bInterpolateBicubic);
+  EXPECT_FALSE(engine.m_ResampleOptions.bHalftone);
+  EXPECT_FALSE(engine.m_ResampleOptions.bNoSmoothing);
+  EXPECT_FALSE(engine.m_ResampleOptions.bLossy);
 }