Does RBImplementedButNotSent result in any critics for any of you?

It seems like its not testing for the right conditions any more and therefore not finding any messages that are implemented but not sent.  


I think this and the problem I was having the other day is from removing the use of the RBSmallLintContext class from checking criitcs issues.  Why did you decide to do that?

> On 31 Jul 2015, at 16:00, Paul DeBruicker <[hidden email]> wrote:
>
> It seems like its not testing for the right conditions any more and therefore not finding any messages that are implemented but not sent.  
>
>
> I think this and the problem I was having the other day is from removing the use of the RBSmallLintContext class from checking criitcs issues.  Why did you decide to do that?


This happened during case 15724[1]. Shouldn’t be hard to fix but requires tests.

As for context removal: Because we have 100500 contexts in Pharo and everyone uses it’s own. Another thing is that “why should every rule care about the context?”. I wan to check a single class, and the rule needs some kind of context. I don’t have a context I have a class.

Another issues that we still have is that you cannot see whether your method/class is violating a rule. You have to go to a result, which is a context, query something…

 [1]: https://pharo.fogbugz.com/f/cases/15724/

Uko2 wrote

> On 31 Jul 2015, at 16:00, Paul DeBruicker <[hidden email]> wrote:
>
> It seems like its not testing for the right conditions any more and therefore not finding any messages that are implemented but not sent.  
>
>
> I think this and the problem I was having the other day is from removing the use of the RBSmallLintContext class from checking criitcs issues.  Why did you decide to do that?


This happened during case 15724[1]. Shouldn’t be hard to fix but requires tests.

As for context removal: Because we have 100500 contexts in Pharo and everyone uses it’s own. Another thing is that “why should every rule care about the context?”. I wan to check a single class, and the rule needs some kind of context. I don’t have a context I have a class.

Another issues that we still have is that you cannot see whether your method/class is violating a rule. You have to go to a result, which is a context, query something…

 [1]: https://pharo.fogbugz.com/f/cases/15724/

Thanks for the pointer to the bug.  I'll take a look.

Re contexts, there are 539,103 Arrays in Pharo and everyone uses its own. Are there plans to remove them?

Did you consider adding hooks #methodCheckWithContext: & #checkClassWithContext: for those that worked and used contexts so that for the ones that profited from the use of contexts could continue without continuing to be poorly named methods?  

For querying/viewing is there a way contexts could be made polymorphic with rules?

Paul DeBruicker wrote

Uko2 wrote

> On 31 Jul 2015, at 16:00, Paul DeBruicker <[hidden email]> wrote:
>
> It seems like its not testing for the right conditions any more and therefore not finding any messages that are implemented but not sent.  
>
>
> I think this and the problem I was having the other day is from removing the use of the RBSmallLintContext class from checking criitcs issues.  Why did you decide to do that?


This happened during case 15724[1]. Shouldn’t be hard to fix but requires tests.


 [1]: https://pharo.fogbugz.com/f/cases/15724/

Thanks for the pointer to the bug.  I'll take a look.

Backporting the changes from 15724 from Pharo 5 to Pharo 4 does not fix the bug.  The rule still finds no results, even on a package with one class and one method.  I've uploaded a slice that does fix the bug and includes a fix you may want to consider incorporating in Pharo 5.


https://pharo.fogbugz.com/f/cases/16081/Fix-RBImplementedButNotSent-rule-to-return-a-collection-of-methods-that-are-implemented-but-not-sent-in-Pharo-4

The slice is in the pharo 4 inbox