SkUTF8_CountUnichars(s,l) and SkUTF8_NextUnichar(s,l) now safe.

Theory:  We will accept blobs of data as utf-8 text without validation,
but when it comes time to process it: count code poits or convert to
code points, be careful to check for errors.

TODO: SkTypeface::charsToGlyphs() needs to take a length.

Change-Id: Id8110ab43dbffce96faffdda1e0bdaa39cad40e4
Reviewed-on: https://skia-review.googlesource.com/6849
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Herb Derby <herb@google.com>

1 files changed, 61 insertions, 0 deletions

diff --git a/tests/UtilsTest.cpp b/tests/UtilsTest.cpp
index b6f90e03b5..ed861c69a0 100644
--- a/tests/UtilsTest.cpp
+++ b/tests/UtilsTest.cpp
@@ -219,3 +219,64 @@ DEF_TEST(Utils, reporter) {
     test_autounref(reporter);
     test_autostarray(reporter);
 }
+
+#define ASCII_BYTE         "X"
+#define CONTINUATION_BYTE  "\x80"
+#define LEADING_TWO_BYTE   "\xC4"
+#define LEADING_THREE_BYTE "\xE0"
+#define LEADING_FOUR_BYTE  "\xF0"
+#define INVALID_BYTE       "\xFC"
+static bool valid_utf8(const char* p, size_t l) {
+    return SkUTF8_CountUnicharsWithError(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,
+    };
+    for (const char* testcase : goodTestcases) {
+        REPORTER_ASSERT(r, valid_utf8(testcase, strlen(testcase)));
+    }
+    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,
+    };
+    for (const char* testcase : badTestcases) {
+        REPORTER_ASSERT(r, !valid_utf8(testcase, strlen(testcase)));
+    }
+
+}