Simplify Name should not simplify RQA

[forki]

Technically it's correct that the Type Name of the DU is redundant inside of the DU itself, but it's there for a reason and the simplifier should not suggest this case.

/cc @vasily-kirichenko

[vasily-kirichenko]

The analyzer works like this: it finds all qualified names (i.e. ones with dots), cuts leftmost part and check if it's still resolvable at the place it is. If it is, the analyzer cuts the next leftmost part, check and so on. So, to make it work as you want, it's necessary to add a special case (like "if we are in a DU definition, don't take RQA attribute into account for this DU"). I think it's something about style, not about core functionality.

[forki]

Yes the analyzer should be in style ;-) Am 19.10.2017 13:15 schrieb "Vasily Kirichenko" <notifications@github.com>:

…

The analyzer works like this: it finds all qualified names (i.e. ones with dots), cuts leftmost part and check if it's still resolvable at the place it is. If it is, the analyzer cuts the next leftmost part, check and so on. So, to make it work as you want, it's necessary to add a special case (line "if we are in a DU definition, don't take RQA attribute into account for this DU). I think it's something about style, not about core functionality. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#3778 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AADgNA8-RGWjwQ-raHB6f2g5_Fd7oCltks5sty9QgaJpZM4P_Ds9> .

[charlesroddie]

I find the behavior of RQA (where RQA requires qualification outside the DU but not inside) convenient, and to change it would be breaking.

Given that that is the behavior, it's logical that qualification inside the DU gets a "name can be simplified" message, because... the name can be simplified.