Merge pull request #3733 from yglukhov/js-unicode

Fixed unicode handling in JS. Fixes #3714.

compiler/jsgen.nim

@@ -163,8 +163,26 @@ proc mangleName(s: PSym): Rope =
     add(result, rope(s.id))
     s.loc.r = result
 
+proc escapeJSString(s: string): string =
+  result = newStringOfCap(s.len + s.len shr 2)
+  result.add("\"")
+  for c in items(s):
+    case c
+    of '\l': result.add("\\n")
+    of '\r': result.add("\\r")
+    of '\t': result.add("\\t")
+    of '\b': result.add("\\b")
+    of '\a': result.add("\\a")
+    of '\e': result.add("\\e")
+    of '\v': result.add("\\v")
+    of '\\': result.add("\\\\")
+    of '\'': result.add("\\'")
+    of '\"': result.add("\\\"")
+    else: add(result, c)
+  result.add("\"")
+
 proc makeJSString(s: string): Rope =
-  (if s.isNil: "null".rope else: strutils.escape(s).rope)
+  (if s.isNil: "null".rope else: escapeJSString(s).rope)
 
 include jstypes
 

lib/system/jssys.nim

@@ -166,14 +166,33 @@ proc SetConstr() {.varargs, asmNoStackFrame, compilerproc.} =
   """
 
 proc cstrToNimstr(c: cstring): string {.asmNoStackFrame, compilerproc.} =
-  asm """
+  {.emit: """
-    var result = [];
+  var ln = `c`.length;
-    for (var i = 0; i < `c`.length; ++i) {
+  var result = new Array(ln);
-      result[i] = `c`.charCodeAt(i);
+  var r = 0;
+  for (var i = 0; i < ln; ++i) {
+    var ch = `c`.charCodeAt(i);
+
+    if (ch < 128) {
+      result[r] = ch;
     }
-    result[result.length] = 0; // terminating zero
+    else if((ch > 127) && (ch < 2048)) {
-    return result;
+      result[r] = (ch >> 6) | 192;
-  """
+      ++r;
+      result[r] = (ch & 63) | 128;
+    }
+    else {
+      result[r] = (ch >> 12) | 224;
+      ++r;
+      result[r] = ((ch >> 6) & 63) | 128;
+      ++r;
+      result[r] = (ch & 63) | 128;
+    }
+    ++r;
+  }
+  result[r] = 0; // terminating zero
+  return result;
+  """.}
 
 proc toJSStr(s: string): cstring {.asmNoStackFrame, compilerproc.} =
   asm """