Hi
When I select an instance method and use push down in a subclass without obvious conflicts. I get two DNUs. realClass ^baseClass realClass theMetaClass dnu: theMetaClass while baseClass is a class realClass -> nil the environement does not contain the class. Now I did a class rename before and I wonder if it is not the one that is broken :( Stef |
In fact I tried on another class that I did not renamed
And I get the same behavior. Now what is really strange is that I did AA renamed into CC and when I do a push down on XX which has nothing to do with AA or CC the ^baseClass realClass -> self name = CC so I do not get how this is possible. Stef Le 29/3/15 21:09, stepharo a écrit : > Hi > > When I select an instance method and use push down in a subclass > without obvious conflicts. > I get two DNUs. > > realClass > > ^baseClass realClass theMetaClass > > dnu: theMetaClass > > while baseClass is a class > realClass -> nil > > > the environement does not contain the class. > Now I did a class rename before and I wonder if it is not the one that > is broken :( > > Stef > > > |
In fact this is even worse (BTW I used the latest Pillar image)
I renamed AA into CC and then CC into DD and now self name = CC We broke something and I find surprising that we do not have test covering refactorings after all these years. We broke already two times refactorings and this is a bad signal. Stef Le 29/3/15 21:14, stepharo a écrit : > In fact I tried on another class that I did not renamed > And I get the same behavior. > Now what is really strange is that I did > > AA renamed into CC > > and when I do a push down on XX which has nothing to do with AA or CC > the > > ^baseClass realClass > -> self name = CC > > so I do not get how this is possible. > > Stef > > > Le 29/3/15 21:09, stepharo a écrit : >> Hi >> >> When I select an instance method and use push down in a subclass >> without obvious conflicts. >> I get two DNUs. >> >> realClass >> >> ^baseClass realClass theMetaClass >> >> dnu: theMetaClass >> >> while baseClass is a class >> realClass -> nil >> >> >> the environement does not contain the class. >> Now I did a class rename before and I wonder if it is not the one >> that is broken :( >> >> Stef >> >> >> > > > > |
Le 29/03/2015 21:18, stepharo a écrit :
> In fact this is even worse (BTW I used the latest Pillar image) > > I renamed AA into CC and then CC into DD > and now > self name = CC > We broke something and I find surprising that we do not have test > covering refactorings after all these years. Do you mean that RBRenameClassTest covers nothing? Thierry > We broke already two times refactorings and this is a bad signal. > Stef > > Le 29/3/15 21:14, stepharo a écrit : >> In fact I tried on another class that I did not renamed >> And I get the same behavior. >> Now what is really strange is that I did >> >> AA renamed into CC >> >> and when I do a push down on XX which has nothing to do with AA or CC >> the >> >> ^baseClass realClass >> -> self name = CC >> >> so I do not get how this is possible. >> >> Stef >> >> >> Le 29/3/15 21:09, stepharo a écrit : >>> Hi >>> >>> When I select an instance method and use push down in a subclass >>> without obvious conflicts. >>> I get two DNUs. >>> >>> realClass >>> >>> ^baseClass realClass theMetaClass >>> >>> dnu: theMetaClass >>> >>> while baseClass is a class >>> realClass -> nil >>> >>> >>> the environement does not contain the class. >>> Now I did a class rename before and I wonder if it is not the one >>> that is broken :( >>> >>> Stef >>> >>> >>> >> >> >> >> > > > > |
Le 29/3/15 21:30, Thierry Goubier a écrit : > Le 29/03/2015 21:18, stepharo a écrit : >> In fact this is even worse (BTW I used the latest Pillar image) >> >> I renamed AA into CC and then CC into DD >> and now >> self name = CC >> We broke something and I find surprising that we do not have test >> covering refactorings after all these years. > > Do you mean that RBRenameClassTest covers nothing? I do not know. But what is clear is that - Refactoring should not be difficult to test well - we break them regularly > > Thierry > >> We broke already two times refactorings and this is a bad signal. >> Stef >> >> Le 29/3/15 21:14, stepharo a écrit : >>> In fact I tried on another class that I did not renamed >>> And I get the same behavior. >>> Now what is really strange is that I did >>> >>> AA renamed into CC >>> >>> and when I do a push down on XX which has nothing to do with AA or CC >>> the >>> >>> ^baseClass realClass >>> -> self name = CC >>> >>> so I do not get how this is possible. >>> >>> Stef >>> >>> >>> Le 29/3/15 21:09, stepharo a écrit : >>>> Hi >>>> >>>> When I select an instance method and use push down in a subclass >>>> without obvious conflicts. >>>> I get two DNUs. >>>> >>>> realClass >>>> >>>> ^baseClass realClass theMetaClass >>>> >>>> dnu: theMetaClass >>>> >>>> while baseClass is a class >>>> realClass -> nil >>>> >>>> >>>> the environement does not contain the class. >>>> Now I did a class rename before and I wonder if it is not the one >>>> that is broken :( >>>> >>>> Stef >>>> >>>> >>>> >>> >>> >>> >>> >> >> >> >> > > > |
In reply to this post by stepharo
This bug is really severe
Once we have it it breaks loading MC files too. Le 29/3/15 21:09, stepharo a écrit : > Hi > > When I select an instance method and use push down in a subclass > without obvious conflicts. > I get two DNUs. > > realClass > > ^baseClass realClass theMetaClass > > dnu: theMetaClass > > while baseClass is a class > realClass -> nil > > > the environement does not contain the class. > Now I did a class rename before and I wonder if it is not the one that > is broken :( > > Stef > > > |
I can not reproduce this error, do you use refactoring from the Nautilus menu?2015-03-30 8:47 GMT+02:00 stepharo <[hidden email]>: This bug is really severe |
1. To AnnouncementsHelp I added method #y 2. Right-clicking on #y > Refactoring > Push down 3. Clicked <Ok> in the Changes Browser 4. Observed #y had moved from AnnouncementsHelp to both subclasses. Repeated for existing AnnouncementsHelp-class>>introduction. cheers -ben On Mon, Mar 30, 2015 at 3:03 PM, Nicolai Hess <[hidden email]> wrote:
|
Yes this is really strange because I tried to reproduce it in 4.0
and 4.0 plus pillar and I could not.
However I was really not doing anything special. So we can probably close this issue but it was annoying. Stef Le 30/3/15 15:48, Ben Coman a écrit :
|
Free forum by Nabble | Edit this page |