spam: suspected: Re: [Mono-devel-list] patch for nunit support for xmlconf

[Atsushi Eno]

Andrew Skiba wrote:
> Atsushi Eno wrote:
>  > (Also note that you are saying "I gave you full Sunday for patch
>  > review" which is kinda zero day attack :-S)
> 
> Can't stop laughing. "Zero day attack" sounds good.

What if I posted changes on your works on Thursday evening for you,
while I know you guys have off days on Friday and Saturday? I know
you understood that representation as a joke (and that certainly is),
but the entire talk is not a joke.

> I thought, I have your permission to commit to the W3C directory without 
> review. And then, I did not say I will commit tomorrow morning, so 
> tomorrow people will have time to review.

Note that your evening would be still Monday afternoon for Boston
people (with somewhat insane timezone).

If you didn't need any comments, then you could just commit things.
But since you didn't do that, then what you asked is all readers'
comments. I'm just a reader here, without mentioning any kind of
"permission" and I think your way of answering is weird. No?

>> You need more explanation. No one would understand what
>> "add nunit support" exactly does.
> 
> It does exactly that - nunit support. You can see in the makefile there 
> is a new target - run-nunit-test. This target runs nunit-console with 
> this testsuite, and the results are stored in TestResults.xml and 
> TestResults.log
> 
> I made some trick to make minimal changes in existing code. Instead of 
> running test from nunit, I kept the existing code that runs the test. 
> Later instead of reporting results, it remembers them by creating 
> PredefinedTest object with already known outcome of the test, and when 
> nunit calls Run, it makes Success or Failure according to the result 
> already stored in _res field. I thought the code is self-explaining, but 
> if it's not so, I will explain more.

You should notice there are many possible interpretations
than you have in mind. People cannot identify whether "nunit
support" means:

	- if it is integrated into "make run-test" in the
	  containing directory (i.e. in mcs/class/System.XML).
	- if it is to generate a TestFixture class which contains
	  Test case methods.
	- if it reports "failure" on "you should remove fixed bugs
	  from list" cases (I mentioned above), since when it
	  is standalone tests it dictates us to do that.

Not all of readers would read the entire patch. They will first
check ChangeLog (and/or the post itself) and see if it is
significant for themselves or not. If that is not understandable,
then they will ask what it means. That's what I did.

See how Rafi answered to the question, which clarifies things.

Atsushi Eno