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>

3 files changed, 93 insertions, 99 deletions

diff --git a/tests/PaintTest.cpp b/tests/PaintTest.cpp
index 647f8e2f23..0bf6df64de 100644
--- a/tests/PaintTest.cpp
+++ b/tests/PaintTest.cpp
@@ -23,7 +23,7 @@
 static size_t uni_to_utf8(const SkUnichar src[], void* dst, int count) {
     char* u8 = (char*)dst;
     for (int i = 0; i < count; ++i) {
-        int n = SkToInt(SkUTF8_FromUnichar(src[i], u8));
+        int n = SkToInt(SkUTF::ToUTF8(src[i], u8));
         u8 += n;
     }
     return u8 - (char*)dst;
@@ -32,7 +32,7 @@ static size_t uni_to_utf8(const SkUnichar src[], void* dst, int count) {
 static size_t uni_to_utf16(const SkUnichar src[], void* dst, int count) {
     uint16_t* u16 = (uint16_t*)dst;
     for (int i = 0; i < count; ++i) {
-        int n = SkToInt(SkUTF16_FromUnichar(src[i], u16));
+        int n = SkToInt(SkUTF::ToUTF16(src[i], u16));
         u16 += n;
     }
     return (char*)u16 - (char*)dst;
diff --git a/tests/UnicodeTest.cpp b/tests/UnicodeTest.cpp
index 5ed92d4b34..5dbcf08547 100644
--- a/tests/UnicodeTest.cpp
+++ b/tests/UnicodeTest.cpp
@@ -9,39 +9,10 @@
 #include "SkUtils.h"
 #include "Test.h"
 
-// Unicode Variation Selector ranges: inclusive
-#define UVS_MIN0    0x180B
-#define UVS_MAX0    0x180D
-#define UVS_MIN1    0xFE00
-#define UVS_MAX1    0xFE0F
-#define UVS_MIN2    0xE0100
-#define UVS_MAX2    0xE01EF
-
-static bool isUVS(SkUnichar uni) {
-    return (uni >= UVS_MIN0 && uni <= UVS_MAX0) ||
-           (uni >= UVS_MIN1 && uni <= UVS_MAX1) ||
-           (uni >= UVS_MIN2 && uni <= UVS_MAX2);
-}
-
-static void test_uvs(skiatest::Reporter* reporter) {
-    // [min, max], [min, max] ... inclusive
-    static const SkUnichar gRanges[] = {
-        UVS_MIN0, UVS_MAX0, UVS_MIN1, UVS_MAX1, UVS_MIN2, UVS_MAX2
-    };
-
-    for (size_t i = 0; i < SK_ARRAY_COUNT(gRanges); i += 2) {
-        for (SkUnichar uni = gRanges[i] - 8; uni <= gRanges[i+1] + 8; ++uni) {
-            bool uvs0 = isUVS(uni);
-            bool uvs1 = SkUnichar_IsVariationSelector(uni);
-            REPORTER_ASSERT(reporter, uvs0 == uvs1);
-        }
-    }
-}
-
 // Simple test to ensure that when we call textToGlyphs, we get the same
 // result (for the same text) when using UTF8, UTF16, UTF32.
 // TODO: make the text more complex (i.e. incorporate chars>7bits)
-static void test_textencodings(skiatest::Reporter* reporter) {
+DEF_TEST(Unicode_textencodings, reporter) {
     const char text8[] = "ABCDEFGabcdefg0123456789";
     uint16_t text16[sizeof(text8)];
     int32_t  text32[sizeof(text8)];
@@ -76,8 +47,3 @@ static void test_textencodings(skiatest::Reporter* reporter) {
     REPORTER_ASSERT(reporter, !memcmp(glyphs8, glyphs16, count8 * sizeof(uint16_t)));
     REPORTER_ASSERT(reporter, !memcmp(glyphs8, glyphs32, count8 * sizeof(uint16_t)));
 }
-
-DEF_TEST(Unicode, reporter) {
-    test_uvs(reporter);
-    test_textencodings(reporter);
-}
diff --git a/tests/UtilsTest.cpp b/tests/UtilsTest.cpp
index 6f8ec914bc..d621a8b12e 100644
--- a/tests/UtilsTest.cpp
+++ b/tests/UtilsTest.cpp
@@ -165,20 +165,19 @@ static void test_search(skiatest::Reporter* reporter) {
 }
 
 static void test_utf16(skiatest::Reporter* reporter) {
+    // Test non-basic-multilingual-plane unicode.
     static const SkUnichar gUni[] = {
         0x10000, 0x18080, 0x20202, 0xFFFFF, 0x101234
     };
-
-    uint16_t buf[2];
-
-    for (size_t i = 0; i < SK_ARRAY_COUNT(gUni); i++) {
-        size_t count = SkUTF16_FromUnichar(gUni[i], buf);
+    for (SkUnichar uni : gUni) {
+        uint16_t buf[2];
+        size_t count = SkUTF::ToUTF16(uni, buf);
         REPORTER_ASSERT(reporter, count == 2);
-        size_t count2 = SkUTF16_CountUnichars(buf, 2 * sizeof(uint16_t));
+        size_t count2 = SkUTF::CountUTF16(buf, sizeof(buf));
         REPORTER_ASSERT(reporter, count2 == 1);
         const uint16_t* ptr = buf;
-        SkUnichar c = SkUTF16_NextUnichar(&ptr, buf + SK_ARRAY_COUNT(buf));
-        REPORTER_ASSERT(reporter, c == gUni[i]);
+        SkUnichar c = SkUTF::NextUTF16(&ptr, buf + SK_ARRAY_COUNT(buf));
+        REPORTER_ASSERT(reporter, c == uni);
         REPORTER_ASSERT(reporter, ptr - buf == 2);
     }
 }
@@ -204,8 +203,8 @@ DEF_TEST(Utils, reporter) {
     for (size_t i = 0; i < SK_ARRAY_COUNT(gTest); i++) {
         const char* p = gTest[i].fUtf8;
         const char* stop = p + strlen(p);
-        int         n = SkUTF8_CountUnichars(p, strlen(p));
-        SkUnichar   u1 = SkUTF8_NextUnichar(&p, stop);
+        int         n = SkUTF::CountUTF8(p, strlen(p));
+        SkUnichar   u1 = SkUTF::NextUTF8(&p, stop);
 
         REPORTER_ASSERT(reporter, n == 1);
         REPORTER_ASSERT(reporter, u1 == gTest[i].fUni);
@@ -220,62 +219,91 @@ DEF_TEST(Utils, reporter) {
 }
 
 #define ASCII_BYTE         "X"
-#define CONTINUATION_BYTE  "\x80"
-#define LEADING_TWO_BYTE   "\xC4"
-#define LEADING_THREE_BYTE "\xE0"
+#define CONTINUATION_BYTE  "\xA1"
+#define LEADING_TWO_BYTE   "\xC2"
+#define LEADING_THREE_BYTE "\xE1"
 #define LEADING_FOUR_BYTE  "\xF0"
 #define INVALID_BYTE       "\xFC"
-static bool valid_utf8(const char* p, size_t l) {
-    return SkUTF8_CountUnichars(p, l) >= 0;
-}
-DEF_TEST(Utils_UTF8_ValidLength, r) {
-    const char* goodTestcases[] = {
-        "",
-        ASCII_BYTE,
-        ASCII_BYTE ASCII_BYTE,
-        LEADING_TWO_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE LEADING_TWO_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE ASCII_BYTE LEADING_TWO_BYTE CONTINUATION_BYTE,
-        LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE ASCII_BYTE LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        LEADING_FOUR_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE LEADING_FOUR_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE ASCII_BYTE LEADING_FOUR_BYTE CONTINUATION_BYTE CONTINUATION_BYTE
-            CONTINUATION_BYTE,
+DEF_TEST(SkUTF_CountUTF8, r) {
+    struct {
+        int expectedCount;
+        const char* utf8String;
+    } testCases[] = {
+        { 0, "" },
+        { 1, ASCII_BYTE },
+        { 2, ASCII_BYTE ASCII_BYTE },
+        { 1, LEADING_TWO_BYTE CONTINUATION_BYTE },
+        { 2, ASCII_BYTE LEADING_TWO_BYTE CONTINUATION_BYTE },
+        { 3, ASCII_BYTE ASCII_BYTE LEADING_TWO_BYTE CONTINUATION_BYTE },
+        { 1, LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { 2, ASCII_BYTE LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { 3, ASCII_BYTE ASCII_BYTE LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { 1, LEADING_FOUR_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { 2, ASCII_BYTE LEADING_FOUR_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { 3, ASCII_BYTE ASCII_BYTE LEADING_FOUR_BYTE CONTINUATION_BYTE CONTINUATION_BYTE
+             CONTINUATION_BYTE },
+        { -1, INVALID_BYTE },
+        { -1, INVALID_BYTE CONTINUATION_BYTE },
+        { -1, INVALID_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { -1, INVALID_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { -1, LEADING_TWO_BYTE },
+        { -1, CONTINUATION_BYTE },
+        { -1, CONTINUATION_BYTE CONTINUATION_BYTE },
+        { -1, LEADING_THREE_BYTE CONTINUATION_BYTE },
+        { -1, CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { -1, LEADING_FOUR_BYTE CONTINUATION_BYTE },
+        { -1, CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
     };
-    for (const char* testcase : goodTestcases) {
-        REPORTER_ASSERT(r, valid_utf8(testcase, strlen(testcase)));
+    for (auto testCase : testCases) {
+        const char* str = testCase.utf8String;
+        REPORTER_ASSERT(r, testCase.expectedCount == SkUTF::CountUTF8(str, strlen(str)));
     }
-    const char* badTestcases[] = {
-        INVALID_BYTE,
-        INVALID_BYTE CONTINUATION_BYTE,
-        INVALID_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        INVALID_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        LEADING_TWO_BYTE,
-        CONTINUATION_BYTE,
-        CONTINUATION_BYTE CONTINUATION_BYTE,
-        LEADING_THREE_BYTE CONTINUATION_BYTE,
-        CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        LEADING_FOUR_BYTE CONTINUATION_BYTE,
-        CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-
-        ASCII_BYTE INVALID_BYTE,
-        ASCII_BYTE INVALID_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE INVALID_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE INVALID_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE LEADING_TWO_BYTE,
-        ASCII_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE LEADING_THREE_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE LEADING_FOUR_BYTE CONTINUATION_BYTE,
-        ASCII_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE CONTINUATION_BYTE,
-
-        // LEADING_FOUR_BYTE LEADING_TWO_BYTE CONTINUATION_BYTE,
+}
+
+DEF_TEST(SkUTF_NextUTF8_ToUTF8, r) {
+    struct {
+        SkUnichar expected;
+        const char* utf8String;
+    } testCases[] = {
+        { -1, INVALID_BYTE },
+        { -1, "" },
+        { 0x0058, ASCII_BYTE },
+        { 0x00A1, LEADING_TWO_BYTE CONTINUATION_BYTE },
+        { 0x1861, LEADING_THREE_BYTE CONTINUATION_BYTE CONTINUATION_BYTE },
+        { 0x010330, LEADING_FOUR_BYTE "\x90\x8C\xB0" },
     };
-    for (const char* testcase : badTestcases) {
-        REPORTER_ASSERT(r, !valid_utf8(testcase, strlen(testcase)));
+    for (auto testCase : testCases) {
+        const char* str = testCase.utf8String;
+        SkUnichar uni = SkUTF::NextUTF8(&str, str + strlen(str));
+        REPORTER_ASSERT(r, str == testCase.utf8String + strlen(testCase.utf8String));
+        REPORTER_ASSERT(r, uni == testCase.expected);
+        char buff[5] = {0, 0, 0, 0, 0};
+        size_t len = SkUTF::ToUTF8(uni, buff);
+        if (buff[len] != 0) {
+            ERRORF(r, "unexpected write");
+            continue;
+        }
+        if (uni == -1) {
+            REPORTER_ASSERT(r, len == 0);
+            continue;
+        }
+        if (len == 0) {
+           ERRORF(r, "unexpected failure.");
+           continue;
+        }
+        if (len > 4) {
+           ERRORF(r, "wrote too much");
+           continue;
+        }
+        str = testCase.utf8String;
+        REPORTER_ASSERT(r, len == strlen(buff));
+        REPORTER_ASSERT(r, len == strlen(str));
+        REPORTER_ASSERT(r, 0 == strcmp(str, buff));
     }
-
 }
+#undef ASCII_BYTE
+#undef CONTINUATION_BYTE
+#undef LEADING_TWO_BYTE
+#undef LEADING_THREE_BYTE
+#undef LEADING_FOUR_BYTE
+#undef INVALID_BYTE