Down-integrate from internal branch.

Change-Id: Ieb7a2c2fbf35bc2a8fa65b915a5ecb68c83863e4

1 files changed, 24 insertions, 0 deletions

diff --git a/src/google/protobuf/stubs/common.h b/src/google/protobuf/stubs/common.h
index de866e14..88e7084f 100644
--- a/src/google/protobuf/stubs/common.h
+++ b/src/google/protobuf/stubs/common.h
@@ -35,6 +35,7 @@
 #ifndef GOOGLE_PROTOBUF_COMMON_H__
 #define GOOGLE_PROTOBUF_COMMON_H__
 
+#include <string>
 
 #include <google/protobuf/stubs/port.h>
 #include <google/protobuf/stubs/macros.h>
@@ -137,12 +138,35 @@ std::string LIBPROTOBUF_EXPORT VersionString(int version);
 // ===================================================================
 // from google3/util/utf8/public/unilib.h
 
+class StringPiece;
 namespace internal {
 
 // Checks if the buffer contains structurally-valid UTF-8.  Implemented in
 // structurally_valid.cc.
 LIBPROTOBUF_EXPORT bool IsStructurallyValidUTF8(const char* buf, int len);
 
+inline bool IsStructurallyValidUTF8(const std::string& str) {
+  return IsStructurallyValidUTF8(str.data(), str.length());
+}
+
+// Returns initial number of bytes of structually valid UTF-8.
+LIBPROTOBUF_EXPORT int UTF8SpnStructurallyValid(const StringPiece& str);
+
+// Coerce UTF-8 byte string in src_str to be
+// a structurally-valid equal-length string by selectively
+// overwriting illegal bytes with replace_char (typically ' ' or '?').
+// replace_char must be legal printable 7-bit Ascii 0x20..0x7e.
+// src_str is read-only.
+//
+// Returns pointer to output buffer, src_str.data() if no changes were made,
+//  or idst if some bytes were changed. idst is allocated by the caller
+//  and must be at least as big as src_str
+//
+// Optimized for: all structurally valid and no byte copying is done.
+//
+LIBPROTOBUF_EXPORT char* UTF8CoerceToStructurallyValid(
+    const StringPiece& str, char* dst, char replace_char);
+
 }  // namespace internal