lint: make output machine readable (#37)

Using Unix-style line number links and delimitors makes it easier to
parse the output using standard tools.

Specifically, popular editors are able to understand how to navigate
directly to a file's line number, and awk can be used to parse the
different fields.

lint/lint.py

@@ -408,8 +408,8 @@ def output_errors_text(errors):
     for error_type, description, path, line_number in errors:
         pos_string = path
         if line_number:
-            pos_string += " %s" % line_number
-        print("%s: %s %s" % (error_type, pos_string, description))
+            pos_string += ":%s" % line_number
+        print("%s: %s (%s)" % (pos_string, description, error_type))
 
 def output_errors_json(errors):
     for error_type, error, path, line_number in errors:

lint/tests/test_lint.py

@@ -151,7 +151,7 @@ def test_lint_failing(capsys):
             assert mocked_check_file_contents.call_count == 1
     out, err = capsys.readouterr()
     assert "TRAILING WHITESPACE" in out
-    assert "broken.html 1 " in out
+    assert "broken.html:1" in out
     assert err == ""
 
 
@@ -164,7 +164,7 @@ def test_lint_passing_and_failing(capsys):
             assert mocked_check_file_contents.call_count == 2
     out, err = capsys.readouterr()
     assert "TRAILING WHITESPACE" in out
-    assert "broken.html 1 " in out
+    assert "broken.html:1" in out
     assert "okay.html" not in out
     assert err == ""