On 6/24/2014 12:19 AM, Alan Bateman wrote:
On 24/06/2014 01:14, huizhe wang wrote:
:
One thing that concerns me a bit is that none of the .java files
that I looked at have any comments to say what the test does. Is
there anything that could be brought over from the original issue in
JIRA to explain what each of these tests is about?
It would have been nice if they had comments. But adding comments
would be too much work, unnecessary I would think. These tests serve
their purpose, that is, preventing regressions. In case any fails,
its bugid would allow us to easily find out what it was testing.
If there is a test failure then whoever runs into it will need to
decypher what the test is about. I would think that it would at least
be helpful to have a short summary at the top to give some indication
as to what it is doing. Are we even sure that all the issues are
accessible to all in JIRA?
If there is a test failure, we can narrow it down to the cause pretty
quickly since we have a context, that is, the current patch. Rather than
some comments, we generally rely on the test code itself to identify the
culprit or browse the original bug report to understand the whole story
(when change happened, the whole history and etc.).
Furthermore, they don't really fail often, in fact, some of the tests
have never failed because of the nature of related fixes. For someone
like Patrick who doesn't have knowledge of the original fixes or bug
reports to go into them and try to figure out a proper write-up, it may
not be an easy task. There are hundreds of tests, plus the function
tests that did not have comments either. I'm just not sure it's worth it
to spend all that time (and then, not used in most of the cases :-) ).
-Joe
-Alan