The default locale "" should be a UTF-8 locale.

"ls -q" (or "adb shell -tt ls") was mangling non-ASCII because mbrtowc
was returning multibyte characters as their individual bytes. This was
because toybox asks for "" rather than "C.UTF-8", and for some reason
we were interpreting that as "C" rather than "C.UTF-8".

Test: bionic tests, ls
Change-Id: Ic60e3b90cd5fe689e5489fad0d5d91062b9594ed

3 files changed, 18 insertions, 12 deletions

diff --git a/libc/bionic/locale.cpp b/libc/bionic/locale.cpp
index e51b38c2f..113118d22 100644
--- a/libc/bionic/locale.cpp
+++ b/libc/bionic/locale.cpp
@@ -37,7 +37,8 @@
 
 #include "private/bionic_macros.h"
 
-// We currently support a single locale, the "C" locale (also known as "POSIX").
+// We only support two locales, the "C" locale (also known as "POSIX"),
+// and the "C.UTF-8" locale (also known as "en_US.UTF-8").
 
 static bool __bionic_current_locale_is_utf8 = true;
 
@@ -100,12 +101,16 @@ static void __locale_init() {
   g_locale.int_n_sign_posn = CHAR_MAX;
 }
 
-static bool __is_supported_locale(const char* locale) {
-  return (strcmp(locale, "") == 0 ||
-          strcmp(locale, "C") == 0 ||
-          strcmp(locale, "C.UTF-8") == 0 ||
-          strcmp(locale, "en_US.UTF-8") == 0 ||
-          strcmp(locale, "POSIX") == 0);
+static bool __is_supported_locale(const char* locale_name) {
+  return (strcmp(locale_name, "") == 0 ||
+          strcmp(locale_name, "C") == 0 ||
+          strcmp(locale_name, "C.UTF-8") == 0 ||
+          strcmp(locale_name, "en_US.UTF-8") == 0 ||
+          strcmp(locale_name, "POSIX") == 0);
+}
+
+static bool __is_utf8_locale(const char* locale_name) {
+  return (*locale_name == '\0' || strstr(locale_name, "UTF-8"));
 }
 
 lconv* localeconv() {
@@ -133,7 +138,7 @@ locale_t newlocale(int category_mask, const char* locale_name, locale_t /*base*/
     return NULL;
   }
 
-  return new __locale_t(strstr(locale_name, "UTF-8") != NULL ? 4 : 1);
+  return new __locale_t(__is_utf8_locale(locale_name) ? 4 : 1);
 }
 
 char* setlocale(int category, const char* locale_name) {
@@ -150,7 +155,7 @@ char* setlocale(int category, const char* locale_name) {
       errno = ENOENT;
       return NULL;
     }
-    __bionic_current_locale_is_utf8 = (strstr(locale_name, "UTF-8") != NULL);
+    __bionic_current_locale_is_utf8 = __is_utf8_locale(locale_name);
   }
 
   return const_cast<char*>(__bionic_current_locale_is_utf8 ? "C.UTF-8" : "C");
diff --git a/libc/bionic/wchar.cpp b/libc/bionic/wchar.cpp
index d28888d36..7717e10c6 100644
--- a/libc/bionic/wchar.cpp
+++ b/libc/bionic/wchar.cpp
@@ -61,7 +61,7 @@ size_t mbrtowc(wchar_t* pwc, const char* s, size_t n, mbstate_t* ps) {
   static mbstate_t __private_state;
   mbstate_t* state = (ps == NULL) ? &__private_state : ps;
 
-  // Our wchar_t is UTF-32
+  // Our wchar_t is UTF-32.
   return mbrtoc32(reinterpret_cast<char32_t*>(pwc), s, n, state);
 }
 
diff --git a/tests/locale_test.cpp b/tests/locale_test.cpp
index f308af534..8b38c407e 100644
--- a/tests/locale_test.cpp
+++ b/tests/locale_test.cpp
@@ -59,9 +59,10 @@ TEST(locale, setlocale) {
   EXPECT_EQ(EINVAL, errno);
 
 #if defined(__BIONIC__)
-  // The "" locale is implementation-defined. For bionic, it's the C locale.
+  // The "" locale is implementation-defined. For bionic, it's the C.UTF-8 locale, which is
+  // pretty much all we support anyway.
   // glibc will give us something like "en_US.UTF-8", depending on the user's configuration.
-  EXPECT_STREQ("C", setlocale(LC_ALL, ""));
+  EXPECT_STREQ("C.UTF-8", setlocale(LC_ALL, ""));
 #endif
   EXPECT_STREQ("C", setlocale(LC_ALL, "C"));
   EXPECT_STREQ("C", setlocale(LC_ALL, "POSIX"));