utf-8 support

2 files changed, 46 insertions, 1 deletions

diff --git a/src/library.js b/src/library.js
index bb73c48a..a763bd9c 100644
--- a/src/library.js
+++ b/src/library.js
@@ -367,12 +367,44 @@ LibraryManager.library = {
           return input.cache.shift();
         };
       }
+
+      var utf8Buffer = [];
+      var utf8Needed = 0;
+      function utf8Processor(code) {
+        code = code & 0xff;
+        if (utf8Needed) {
+          utf8Buffer.push(code);
+          utf8Needed--;
+        }
+        if (utf8Buffer.length == 0) {
+          if (code < 128) return String.fromCharCode(code);
+          utf8Buffer.push(code);
+          if (code > 191 && code < 224) {
+            utf8Needed = 1;
+          } else {
+            utf8Needed = 2;
+          }
+          return '';
+        }
+        if (utf8Needed > 0) return '';
+        var c1 = utf8Buffer[0];
+        var c2 = utf8Buffer[1];
+        var c3 = utf8Buffer[2];
+        var ret;
+        if (c1 > 191 && c1 < 224) {
+          ret = String.fromCharCode(((c1 & 31) << 6) | (c2 & 63));
+        } else {
+          ret = String.fromCharCode(((c1 & 15) << 12) | ((c2 & 63) << 6) | (c3 & 63));
+        }
+        utf8Buffer.length = 0;
+        return ret;
+      }
       function simpleOutput(val) {
         if (val === null || val === '\n'.charCodeAt(0)) {
           output.printer(output.buffer.join(''));
           output.buffer = [];
         } else {
-          output.buffer.push(String.fromCharCode(val));
+          output.buffer.push(utf8Processor(val));
         }
       }
       if (!output) {
diff --git a/tests/runner.py b/tests/runner.py
index 01a4fb61..c7d1ff28 100755
--- a/tests/runner.py
+++ b/tests/runner.py
@@ -1,4 +1,5 @@
 #!/usr/bin/env python
+# This Python file uses the following encoding: utf-8
 
 '''
 Simple test runner
@@ -4478,6 +4479,18 @@ def process(filename):
       '''
       self.do_run(src, re.sub('(^|\n)\s+', '\\1', expected), post_build=add_pre_run_and_checks)
 
+    def test_utf(self):
+      self.banned_js_engines = [SPIDERMONKEY_ENGINE] # only node handles utf well
+      src = r'''
+        #include <stdio.h>
+
+        int main() {
+          char *c = "μ†ℱ ╋ℯ╳╋";
+          printf("%d %d %d %d %s\n", c[0]&0xff, c[1]&0xff, c[2]&0xff, c[3]&0xff, c);
+        }
+      '''
+      self.do_run(src, '206 188 226 128 μ†ℱ ╋ℯ╳╋\n');
+
     def test_direct_string_constant_usage(self):
       if self.emcc_args is None: return self.skip('requires libcxx')