Merge "Fix the printf family for non-ASCII."

4 files changed, 50 insertions, 2 deletions

diff --git a/libc/include/limits.h b/libc/include/limits.h
index dc45902..fb09657 100644
--- a/libc/include/limits.h
+++ b/libc/include/limits.h
@@ -112,7 +112,7 @@
 
 #define SSIZE_MAX LONG_MAX
 
-#define MB_LEN_MAX 6
+#define MB_LEN_MAX 4
 
 /* New code should use sysconf(_SC_PAGE_SIZE) instead. */
 #ifndef PAGE_SIZE
diff --git a/libc/include/stdlib.h b/libc/include/stdlib.h
index 834dcda..483aaf0 100644
--- a/libc/include/stdlib.h
+++ b/libc/include/stdlib.h
@@ -167,7 +167,7 @@ extern int      mbtowc(wchar_t *, const char *, size_t);
 extern int	wctomb(char *, wchar_t);
 extern size_t	wcstombs(char *, const wchar_t *, size_t);
 
-#define MB_CUR_MAX 1
+#define MB_CUR_MAX 4U
 
 #if 0 /* MISSING FROM BIONIC */
 extern int on_exit(void (*)(int, void *), void *);
diff --git a/tests/stdio_test.cpp b/tests/stdio_test.cpp
index 0ff85bf..e291f52 100644
--- a/tests/stdio_test.cpp
+++ b/tests/stdio_test.cpp
@@ -426,7 +426,26 @@ TEST(stdio, snprintf_negative_zero_5084292) {
   EXPECT_STREQ("-0.000000", buf);
 }
 
+TEST(stdio, snprintf_utf8_15439554) {
+  // http://b/15439554
+  char buf[BUFSIZ];
+
+  // 1-byte character.
+  snprintf(buf, sizeof(buf), "%dx%d", 1, 2);
+  EXPECT_STREQ("1x2", buf);
+  // 2-byte character.
+  snprintf(buf, sizeof(buf), "%d\xc2\xa2%d", 1, 2);
+  EXPECT_STREQ("1¢2", buf);
+  // 3-byte character.
+  snprintf(buf, sizeof(buf), "%d\xe2\x82\xac%d", 1, 2);
+  EXPECT_STREQ("1€2", buf);
+  // 4-byte character.
+  snprintf(buf, sizeof(buf), "%d\xf0\xa4\xad\xa2%d", 1, 2);
+  EXPECT_STREQ("1𤭢2", buf);
+}
+
 TEST(stdio, fprintf_failures_7229520) {
+  // http://b/7229520
   FILE* fp;
 
   // Unbuffered case where the fprintf(3) itself fails.
diff --git a/tests/wchar_test.cpp b/tests/wchar_test.cpp
index e76026f..a5f5f63 100644
--- a/tests/wchar_test.cpp
+++ b/tests/wchar_test.cpp
@@ -449,3 +449,32 @@ TEST(wchar, wmemmove) {
   wmemmove(wstr+5, wstr, sizeof(const_wstr)/sizeof(wchar_t) - 6);
   EXPECT_STREQ(L"This This is a test of something or other", wstr);
 }
+
+TEST(wchar, mbrtowc_15439554) {
+  // http://b/15439554
+  ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
+  uselocale(LC_GLOBAL_LOCALE);
+
+  ASSERT_GE(static_cast<size_t>(MB_LEN_MAX), MB_CUR_MAX);
+  ASSERT_GE(MB_CUR_MAX, 4U);
+
+  wchar_t wc;
+  size_t n;
+
+  // 1-byte character.
+  n = mbrtowc(&wc, "x", MB_CUR_MAX, NULL);
+  EXPECT_EQ(1U, n);
+  EXPECT_EQ(L'x', wc);
+  // 2-byte character.
+  n = mbrtowc(&wc, "\xc2\xa2", MB_CUR_MAX, NULL);
+  EXPECT_EQ(2U, n);
+  EXPECT_EQ(L'¢', wc);
+  // 3-byte character.
+  n = mbrtowc(&wc, "\xe2\x82\xac", MB_CUR_MAX, NULL);
+  EXPECT_EQ(3U, n);
+  EXPECT_EQ(L'€', wc);
+  // 4-byte character.
+  n = mbrtowc(&wc, "\xf0\xa4\xad\xa2", MB_CUR_MAX, NULL);
+  EXPECT_EQ(4U, n);
+  EXPECT_EQ(L'𤭢', wc);
+}