Re: xfail: expected failures

[Garrett Cooper <yanegomi%gmail.com@localhost>]

On Sun, Jun 27, 2010 at 3:06 PM, Julio Merino <jmmv%netbsd.org@localhost> wrote:
> On Sun, Jun 27, 2010 at 10:49 PM, Garrett Cooper 
> <yanegomi%gmail.com@localhost> wrote:
>>>
>>> The rationale I have in mind is: you are always expecting a test case
>>> to pass; correct? So that's the default behavior.  You can change the
>>> "expect" behavior to something else (like expect failure, expect
>>> crash, etc.).  But if you want to go back to the regular behavior (for
>>> cleanup steps, or for some othe reason), then you just "expect pass".
>>
>>    Maybe there should be an ability to reset the expectations to the
>> default :) (or maybe just change the expectations to any arbitrary
>> state)?
>
> Resetting the expectations to the default is what atf_tc_expect_pass
> does.  And to set them to arbitrary state, you use the other
> atf_tc_expect_* functions.  Take a look at
> https://www.julipedia.org/projects/atf/trac/wiki/DesignXFail for
> details on all cases and examples.

Yes, but I think that what Antti was trying to convey was the fact
that the act of specifying that function call doesn't intuitively
reset the default expectations.

BTW, this comment from the wiki looks wrong:

   atf_tc_expect_pass();  /* If reached, immediately exits with failure. */

> And, by the way, I have just finished implementing what is documented
> in that page; you can check out the repository and give it a look
> (surprisingly, I even remembered to document the new features :-P).  I
> need to add to add one extra scenario, "expected_timeout", and 0.10
> will be ready for release -- it has way too many changes already :-)

    :)...
Thanks,
-Garrett