Hashtags may contain unicode letters

Change-Id: Iddca4c676e0b76657ecebae9bab79197bf4804ca
diff --git a/test/test.js b/test/test.js
index 5177948..3014c90 100644
--- a/test/test.js
+++ b/test/test.js
@@ -66,6 +66,27 @@
     done();
   });
 
+  test('Regression test for hashtags with Unicode letters: emit HST', (done) => {
+    const testInput = [
+      '# foundry = base',
+      '# text_id = test-hashtag-unicode',
+      '# text = #okeichhörejetztauf #schön #10',
+      ['1', '#okeichhörejetztauf', '_', '_', '_', '_', '_', '_', '_', '_'].join('\t'),
+      ['2', '#schön', '_', '_', '_', '_', '_', '_', '_', '_'].join('\t'),
+      ['3', '#10', '_', '_', '_', '_', '_', '_', '_', '_'].join('\t'),
+      ''
+    ].join('\n');
+    const stdout = execSync('node src/index.js', { input: testInput }).toString();
+
+    expect(stdout).toContain('#okeichhörejetztauf\t_\t_\tHST');
+    expect(stdout).toContain('#schön\t_\t_\tHST');
+    expect(stdout).not.toContain('#10\t_\t_\tHST');
+
+    var hst_count = (stdout.match(/\tHST\t/g) || []).length;
+    expect(hst_count).toBe(2);
+    done();
+  });
+
   test('Regression test for addresses: emit ADR regardless of existing POS values', (done) => {
     const testInput = [
       '# foundry = base',