Support decoding opaque to *premul

src/codec/SkCodecPriv.h

Index: src/codec/SkCodecPriv.h
diff --git a/src/codec/SkCodecPriv.h b/src/codec/SkCodecPriv.h
index 27e2a63722d8ff095911846ece0d45d46cc7479f..4ae049744c3017b4a1a87da9800845c2404df04c 100644
--- a/src/codec/SkCodecPriv.h
+++ b/src/codec/SkCodecPriv.h
@@ -75,11 +75,14 @@ inline bool is_coord_necessary(int srcCoord, int sampleFactor, int scaledDim) {
 }
 
 inline bool valid_alpha(SkAlphaType dstAlpha, SkAlphaType srcAlpha) {
-    // Check for supported alpha types
+    if (kUnknown_SkAlphaType == dstAlpha) {
+        return false;
+    }
+
     if (srcAlpha != dstAlpha) {
         if (kOpaque_SkAlphaType == srcAlpha) {
-            // If the source is opaque, we must decode to opaque
-            return false;
+            // If the source is opaque, we can support any.

[msarett, 2016/01/29 14:50:38]

Can we print a "Warning: This is a bad idea..." or something that like that
here?

[scroggo, 2016/01/29 17:19:54]

You'll invoke Cary's ire for adding print statements to DM (since we test this
to make sure it works). We could instead add an SkError (a more powerful feature
for errors which never caught on), which would allow us to silence it in the
test.

[msarett, 2016/01/29 19:17:05]

What about SkCodecPrintf?  I believe that's silent in DM, but turned on in
Android?

[scroggo, 2016/01/29 19:31:22]

Done.

+            return true;
         }
 
         // The source is not opaque
@@ -99,7 +102,8 @@ inline bool valid_alpha(SkAlphaType dstAlpha, SkAlphaType srcAlpha) {
 /*
  * Most of our codecs support the same conversions:
  * - profileType must be the same
- * - opaque only to opaque (and 565 only if opaque)
+ * - opaque to any alpha type
+ * - 565 only if opaque
  * - premul to unpremul and vice versa
  * - always support N32
  * - otherwise match the src color type