SkUTF

Create new header and namespace, `SkUTF` where we are putting all of our
robust, well documented UTF-8, UTF-16, and UTF-32 functions:
`SkUTF::{Count,Next,To}UTF{8,16,32}()`.

SkUTF.h and SkUTF.cpp do not depend on the rest of Skia and are suitable
for re-use in other modules.

Some of the old UTF-{8,16} functions still live in SkUtils.h; their use
will be phased out in future CLs.

Also added more unit testing and cleaned up old tests.

Removed functions that were unused outside of tests or used only once.

Change-Id: Iaa59b8705abccf9c4ba082f855da368a0bad8380
Reviewed-on: https://skia-review.googlesource.com/143306
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>

1 files changed, 18 insertions, 9 deletions

diff --git a/modules/skshaper/src/SkShaper_harfbuzz.cpp b/modules/skshaper/src/SkShaper_harfbuzz.cpp
index 2077dedadc..5c49de04b4 100644
--- a/modules/skshaper/src/SkShaper_harfbuzz.cpp
+++ b/modules/skshaper/src/SkShaper_harfbuzz.cpp
@@ -83,6 +83,15 @@ HBFont create_hb_font(SkTypeface* tf) {
     return font;
 }
 
+/** this version replaces invalid utf-8 sequences with code point U+FFFD. */
+static inline SkUnichar utf8_next(const char** ptr, const char* end) {
+    SkUnichar val = SkUTF::NextUTF8(ptr, end);
+    if (val < 0) {
+        return 0xFFFD;  // REPLACEMENT CHARACTER
+    }
+    return val;
+}
+
 class RunIterator {
 public:
     virtual ~RunIterator() {}
@@ -138,16 +147,16 @@ public:
         SkASSERT(fUTF16LogicalPosition < ubidi_getLength(fBidi.get()));
         int32_t endPosition = ubidi_getLength(fBidi.get());
         fLevel = ubidi_getLevelAt(fBidi.get(), fUTF16LogicalPosition);
-        SkUnichar u = SkUTF8_NextUnichar(&fEndOfCurrentRun, fEndOfAllRuns);
-        fUTF16LogicalPosition += SkUTF16_FromUnichar(u);
+        SkUnichar u = utf8_next(&fEndOfCurrentRun, fEndOfAllRuns);
+        fUTF16LogicalPosition += SkUTF::ToUTF16(u);
         UBiDiLevel level;
         while (fUTF16LogicalPosition < endPosition) {
             level = ubidi_getLevelAt(fBidi.get(), fUTF16LogicalPosition);
             if (level != fLevel) {
                 break;
             }
-            u = SkUTF8_NextUnichar(&fEndOfCurrentRun, fEndOfAllRuns);
-            fUTF16LogicalPosition += SkUTF16_FromUnichar(u);
+            u = utf8_next(&fEndOfCurrentRun, fEndOfAllRuns);
+            fUTF16LogicalPosition += SkUTF::ToUTF16(u);
         }
     }
     const char* endOfCurrentRun() const override {
@@ -184,11 +193,11 @@ public:
     {}
     void consume() override {
         SkASSERT(fCurrent < fEnd);
-        SkUnichar u = SkUTF8_NextUnichar(&fCurrent, fEnd);
+        SkUnichar u = utf8_next(&fCurrent, fEnd);
         fCurrentScript = hb_unicode_script(fHBUnicode, u);
         while (fCurrent < fEnd) {
             const char* prev = fCurrent;
-            u = SkUTF8_NextUnichar(&fCurrent, fEnd);
+            u = utf8_next(&fCurrent, fEnd);
             const hb_script_t script = hb_unicode_script(fHBUnicode, u);
             if (script != fCurrentScript) {
                 if (fCurrentScript == HB_SCRIPT_INHERITED || fCurrentScript == HB_SCRIPT_COMMON) {
@@ -243,7 +252,7 @@ public:
     {}
     void consume() override {
         SkASSERT(fCurrent < fEnd);
-        SkUnichar u = SkUTF8_NextUnichar(&fCurrent, fEnd);
+        SkUnichar u = utf8_next(&fCurrent, fEnd);
         // If the starting typeface can handle this character, use it.
         if (fTypeface->charsToGlyphs(&u, SkTypeface::kUTF32_Encoding, nullptr, 1)) {
             fFallbackTypeface.reset();
@@ -265,7 +274,7 @@ public:
 
         while (fCurrent < fEnd) {
             const char* prev = fCurrent;
-            u = SkUTF8_NextUnichar(&fCurrent, fEnd);
+            u = utf8_next(&fCurrent, fEnd);
 
             // If using a fallback and the initial typeface has this character, stop fallback.
             if (fFallbackTypeface &&
@@ -554,7 +563,7 @@ SkPoint SkShaper::shape(SkTextBlobBuilder* builder,
         const char* utf8Current = utf8Start;
         while (utf8Current < utf8End) {
             unsigned int cluster = utf8Current - utf8Start;
-            hb_codepoint_t u = SkUTF8_NextUnichar(&utf8Current, utf8End);
+            hb_codepoint_t u = utf8_next(&utf8Current, utf8End);
             hb_buffer_add(buffer, u, cluster);
         }