Improve string generation in Java

Summary:
Remove all the code for dealing with escape sequences from Java codegens. It is no longer needed because escape sequences are now handled in the lexer/parser. Instead add previously missing escaping of control characters.

In particular, this fixes handling of multiline strings.

Before:

```
 public static final String multi_line_string = "This
is a
multi line string.
";
```

After:

```

  public static final String multi_line_string = "This\nis a\nmulti line string.\n";
```

Reviewed By: bahadird

Differential Revision: D44559991

fbshipit-source-id: c079aad9ba69d08fd0576bd14d73bd2ea93b5147

Author: vitaut

thrift/compiler/generate/t_java_deprecated_generator.cc

@@ -422,62 +422,30 @@ string t_java_deprecated_generator::render_const_value(
     t_base_type::t_base tbase = ((t_base_type*)type)->get_base();
     switch (tbase) {
       case t_base_type::TYPE_STRING:
-      case t_base_type::TYPE_BINARY: {
+      case t_base_type::TYPE_BINARY:
         render << '"';
-        auto& rawValue = value->get_string();
-        for (std::string::size_type i = 0; i < rawValue.size();) {
-          switch (rawValue[i]) {
-            case '\\': {
-              render << rawValue[i];
-              ++i;
-              assert(i <= rawValue.size());
-              if (i == rawValue.size()) {
-                throw std::string(
-                    "compiler error: leading backslash missing "
-                    "escape sequence: ") +
-                    rawValue;
-              }
-              if (rawValue[i] == 'x') {
-                auto end =
-                    rawValue.find_first_not_of("0123456789abcdefABCDEF", ++i);
-                if (end == std::string::npos) {
-                  end = rawValue.size();
-                }
-                if (end == i) {
-                  throw std::string(
-                      "compiler error: missing hexadecimal "
-                      "character code in escape sequence: ") +
-                      rawValue;
-                }
-                assert(i < end);
-                if (end > i + 2) {
-                  end = i + 2;
-                }
-                render << 'u';
-                for (auto n = 4 - (end - i); n--;) {
-                  render << '0';
-                }
-                render.write(std::next(rawValue.data(), i), end - i);
-                i = end;
-              } else {
-                render << rawValue[i++];
-              }
+        for (unsigned char c : value->get_string()) {
+          switch (c) {
+            case '\n':
+              render << "\\n";
               break;
-            }
             case '"':
-              render << '\\';
-              // intentional fallback
+              render << "\\\"";
+              break;
             default:
-              render << rawValue[i];
-              ++i;
+              if (c < 0x20) {
+                render << fmt::format("\\x{:02x}", c);
+              } else {
+                render << c;
+              }
+              break;
           }
         }
         render << '"';
         if (tbase == t_base_type::TYPE_BINARY) {
           render << ".getBytes()";
         }
         break;
-      }
       case t_base_type::TYPE_BOOL:
         render << ((value->get_integer() > 0) ? "true" : "false");
         break;

thrift/compiler/lib/java/util.cc

@@ -80,33 +80,22 @@ std::string mangle_java_constant_name(const std::string& ref) {
 std::string quote_java_string(const std::string& unescaped) {
   std::ostringstream quoted;
   quoted << '\"';
-  for (std::string::size_type i = 0; i < unescaped.size();) {
-    switch (unescaped[i]) {
-      case '\\': {
-        quoted << unescaped[i];
-        ++i;
-        assert(i <= unescaped.size());
-        if (i == unescaped.size()) {
-          throw std::runtime_error(
-              "compiler error: leading backslash missing escape sequence: " +
-              unescaped);
-        }
-        assert(unescaped[i] != 'x');
-        quoted << unescaped[i++];
+  for (unsigned char c : unescaped) {
+    switch (c) {
+      case '\n':
+        quoted << "\\n";
         break;
-      }
       case '"':
-        quoted << '\\' << unescaped[i];
-        ++i;
+        quoted << "\\\"";
         break;
       default: {
-        unsigned char c = unescaped[i];
-        if (c >= 0xf8) {
+        if (c < 0x20) {
+          quoted << fmt::format("\\x{:02x}", c);
+        } else if (c >= 0xf8) {
           quoted << fmt::format("\\u{:04x}", c);
         } else {
-          quoted << unescaped[i];
+          quoted << c;
         }
-        ++i;
         break;
       }
     }