From 46633b1f00204a310e630009f0420218186439e2 Mon Sep 17 00:00:00 2001
From: Ciro Santilli <ciro.santilli@gmail.com>
Date: Sun, 21 Dec 2014 14:33:31 +0100
Subject: [PATCH] Don't raise exception on invalid UTF-8 output

This could be reproduced with:

    ./test/spec_tests.py --program kramdown --pattern Enti

with Kramdown version 1.5.0.

The main problem is that it would stop future tests from being carried out.

After this commit it is just counted as a failure.

The error message printed is of the form:

    Unicode error: 'utf8' codec can't decode byte 0xfc in position 8: invalid start byte
    '<p># \xd3\x92 \xcf\xa0 \xef\xbf\xbd</p>\n'
    '<p># \xd3\x92 \xcf\xa0 \xfc\x85\xb8\xb0\xa9\xb8</p>\n'
---
 test/spec_tests.py | 22 ++++++++++++++++------
 1 file changed, 16 insertions(+), 6 deletions(-)

diff --git a/test/spec_tests.py b/test/spec_tests.py
index 23ae5029..99ac0dda 100755
--- a/test/spec_tests.py
+++ b/test/spec_tests.py
@@ -37,8 +37,13 @@ def do_test(test, normalize):
     [retcode, actual_html, err] = cmark.to_html(test['markdown'])
     if retcode == 0:
         expected_html = test['html']
+        unicode_error = None
         if normalize:
-            passed = normalize_html(actual_html) == normalize_html(expected_html)
+            try:
+                passed = normalize_html(actual_html) == normalize_html(expected_html)
+            except UnicodeDecodeError, e:
+                unicode_error = e
+                passed = False
         else:
             passed = actual_html == expected_html
         if passed:
@@ -46,11 +51,16 @@ def do_test(test, normalize):
         else:
             print_test_header(test['section'], test['example'], test['start_line'], test['end_line'])
             sys.stdout.write(test['markdown'])
-            expected_html_lines = expected_html.splitlines(True)
-            actual_html_lines = actual_html.splitlines(True)
-            for diffline in unified_diff(expected_html_lines, actual_html_lines,
-                            "expected HTML", "actual HTML"):
-                sys.stdout.write(diffline)
+            if unicode_error:
+                print "Unicode error: " + str(unicode_error)
+                print repr(expected_html)
+                print repr(actual_html)
+            else:
+                expected_html_lines = expected_html.splitlines(True)
+                actual_html_lines = actual_html.splitlines(True)
+                for diffline in unified_diff(expected_html_lines, actual_html_lines,
+                                "expected HTML", "actual HTML"):
+                    sys.stdout.write(diffline)
             sys.stdout.write('\n')
             return 'fail'
     else: