test: combine the asserts for output and errors

With the output and error asserts separate, the test case ends on the
first failing assertion, and we never know what was in errors. But the
clue to the output mismatch might be in the errors. For example, an
include file was not found. By combining the asserts into one, we get
all the differences in the pytest results.

test/testenv.py

@@ -104,8 +104,7 @@ def run_test(self):
 
         assert output_docs, 'empty output'
         assert expect_docs, 'empty expected'
-        assert output_docs == expect_docs
-        assert output_errors == expect_errors
+        assert output_docs + output_errors == expect_docs + expect_errors
 
 def get_testid(testcase):
     return testcase.get_testid()