diff --git a/setup.py b/setup.py
index a924d4aba..795a8c75f 100644
--- a/setup.py
+++ b/setup.py
@@ -14,6 +14,7 @@ INSTALL_REQUIRES = [
     'pathlib2>=2.2.0;python_version<"3.6"',
     'colorama;sys_platform=="win32"',
     "pluggy>=0.9",
+    "wcwidth",
 ]
 
 
diff --git a/src/_pytest/skipping.py b/src/_pytest/skipping.py
index f7f085303..1013f4889 100644
--- a/src/_pytest/skipping.py
+++ b/src/_pytest/skipping.py
@@ -206,11 +206,13 @@ def pytest_terminal_summary(terminalreporter):
 
 def _get_line_with_reprcrash_message(config, rep, termwidth):
     """Get summary line for a report, trying to add reprcrash message."""
+    from wcwidth import wcswidth
+
     verbose_word = _get_report_str(config, rep)
     pos = _get_pos(config, rep)
 
     line = "%s %s" % (verbose_word, pos)
-    len_line = len(line)
+    len_line = wcswidth(line)
     ellipsis, len_ellipsis = "...", 3
     if len_line > termwidth - len_ellipsis:
         # No space for an additional message.
@@ -225,13 +227,17 @@ def _get_line_with_reprcrash_message(config, rep, termwidth):
         i = msg.find("\n")
         if i != -1:
             msg = msg[:i]
-        len_msg = len(msg)
+        len_msg = wcswidth(msg)
 
         sep, len_sep = " - ", 3
         max_len_msg = termwidth - len_line - len_sep
         if max_len_msg >= len_ellipsis:
             if len_msg > max_len_msg:
-                msg = msg[: (max_len_msg - len_ellipsis)] + ellipsis
+                max_len_msg -= len_ellipsis
+                msg = msg[:max_len_msg]
+                while wcswidth(msg) > max_len_msg:
+                    msg = msg[:-1]
+                msg += ellipsis
             line += sep + msg
     return line
 
diff --git a/testing/test_skipping.py b/testing/test_skipping.py
index e167d4477..952877d8e 100644
--- a/testing/test_skipping.py
+++ b/testing/test_skipping.py
@@ -1216,50 +1216,67 @@ def test_summary_list_after_errors(testdir):
 def test_line_with_reprcrash(monkeypatch):
     import _pytest.skipping
     from _pytest.skipping import _get_line_with_reprcrash_message
+    from wcwidth import wcswidth
+
+    mocked_verbose_word = "FAILED"
 
     def mock_get_report_str(*args):
-        return "FAILED"
-
-    def mock_get_pos(*args):
-        return "some::nodeid"
+        return mocked_verbose_word
 
     monkeypatch.setattr(_pytest.skipping, "_get_report_str", mock_get_report_str)
+
+    mocked_pos = "some::nodeid"
+
+    def mock_get_pos(*args):
+        return mocked_pos
+
     monkeypatch.setattr(_pytest.skipping, "_get_pos", mock_get_pos)
 
     class config:
         pass
 
-    class rep:
-        pass
-
-    f = _get_line_with_reprcrash_message
-    assert f(config, rep, 80) == "FAILED some::nodeid"
-
     class rep:
         class longrepr:
             class reprcrash:
-                message = "msg"
+                pass
 
-    assert f(config, rep, 80) == "FAILED some::nodeid - msg"
-    assert f(config, rep, 3) == "FAILED some::nodeid"
+    def check(msg, width, expected):
+        if msg:
+            rep.longrepr.reprcrash.message = msg
+        actual = _get_line_with_reprcrash_message(config, rep, width)
 
-    assert f(config, rep, 24) == "FAILED some::nodeid"
-    assert f(config, rep, 25) == "FAILED some::nodeid - msg"
+        assert actual == expected
+        if actual != "%s %s" % (mocked_verbose_word, mocked_pos):
+            assert len(actual) <= width
+            assert wcswidth(actual) <= width
 
-    rep.longrepr.reprcrash.message = "some longer message"
-    assert f(config, rep, 24) == "FAILED some::nodeid"
-    assert f(config, rep, 25) == "FAILED some::nodeid - ..."
-    assert f(config, rep, 26) == "FAILED some::nodeid - s..."
+    # AttributeError with message
+    check(None, 80, "FAILED some::nodeid")
 
-    rep.longrepr.reprcrash.message = "some\nmessage"
-    assert f(config, rep, 25) == "FAILED some::nodeid - ..."
-    assert f(config, rep, 26) == "FAILED some::nodeid - some"
-    assert f(config, rep, 80) == "FAILED some::nodeid - some"
+    check("msg", 80, "FAILED some::nodeid - msg")
+    check("msg", 3, "FAILED some::nodeid")
+
+    check("msg", 24, "FAILED some::nodeid")
+    check("msg", 25, "FAILED some::nodeid - msg")
+
+    check("some longer msg", 24, "FAILED some::nodeid")
+    check("some longer msg", 25, "FAILED some::nodeid - ...")
+    check("some longer msg", 26, "FAILED some::nodeid - s...")
+
+    check("some\nmessage", 25, "FAILED some::nodeid - ...")
+    check("some\nmessage", 26, "FAILED some::nodeid - some")
+    check("some\nmessage", 80, "FAILED some::nodeid - some")
 
     # Test unicode safety.
-    rep.longrepr.reprcrash.message = "😄😄😄😄😄\n2nd line"
-    assert f(config, rep, 26) == "FAILED some::nodeid - 😄..."
-    # XXX: this is actually wrong - since the character uses two terminal
-    # cells.
-    rep.longrepr.reprcrash.message = "😄😄😄😄\n2nd line"
-    assert f(config, rep, 26) == "FAILED some::nodeid - 😄😄😄😄"
+    check("😄😄😄😄😄\n2nd line", 25, "FAILED some::nodeid - ...")
+    check("😄😄😄😄😄\n2nd line", 26, "FAILED some::nodeid - ...")
+    check("😄😄😄😄😄\n2nd line", 27, "FAILED some::nodeid - 😄...")
+    check("😄😄😄😄😄\n2nd line", 28, "FAILED some::nodeid - 😄...")
+    check("😄😄😄😄😄\n2nd line", 29, "FAILED some::nodeid - 😄😄...")
+
+    mocked_pos = "nodeid::😄::withunicode"
+    check("😄😄😄😄😄\n2nd line", 29, "FAILED nodeid::😄::withunicode")
+    check("😄😄😄😄😄\n2nd line", 40, "FAILED nodeid::😄::withunicode - 😄😄...")
+    check("😄😄😄😄😄\n2nd line", 41, "FAILED nodeid::😄::withunicode - 😄😄...")
+    check("😄😄😄😄😄\n2nd line", 42, "FAILED nodeid::😄::withunicode - 😄😄😄...")
+    check("😄😄😄😄😄\n2nd line", 80, "FAILED nodeid::😄::withunicode - 😄😄😄😄😄")