Status: FailingTest
Owner: [hidden email] Labels: Type-Bug Milestone-2.0 New issue 6680 by [hidden email]: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 testInstanceSideMethodsWithNilKeyInLastLiteral _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Updates:
Cc: [hidden email] Comment #1 on issue 6680 by marianopeck: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 Issue 6644 has been merged into this issue. _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Comment #2 on issue 6680 by marianopeck: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 BTw...I tried to ask in the mailng list some time ago...noone answer, and I couldn't find which was the problem... http://forum.world.st/nil-key-in-last-literal-in-instance-side-methods-reproducible-td4646398.html _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Comment #3 on issue 6680 by [hidden email]: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 More information: it seems to happen only when a class is changing shape and then for *all* methods. This means, something is going wrong when migrating methods to shape-changed classes. look at #binding of Class: binding "Answer a binding for the receiver, sharing if possible" | binding | binding := self environment associationAt: name ifAbsent: [nil -> self]. ^binding value == self ifTrue: [binding] ifFalse: [nil -> self] When a class is renamed, then it will return nil->self as the binding. If this is changed to binding "Answer a binding for the receiver, sharing if possible" ^ self environment associationAt: name ifAbsent: [ nil -> self ] The bug disappears... but I am not happy with that, as it is not the real binding of a class that has the same name yet is not in the environment. So next step: check what is done when class shape is changed. _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Updates:
Status: FixToInclude Comment #4 on issue 6680 by [hidden email]: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 so, yes, we should just change #binging. Binding has the meaning "give me the association in Smalltalk globals with your name". Attachments: Class-binding.st 314 bytes _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Updates:
Status: Integrated Comment #5 on issue 6680 by [hidden email]: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 in 2.0 294 _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Comment #6 on issue 6680 by marianopeck: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 Great catch Marcus. I did some research. Previously, we had that same version you have just commited. Then, we change it as part of this issue: http://code.google.com/p/pharo/issues/detail?id=1921 now I wonder what happens regarding that.... _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Comment #7 on issue 6680 by marianopeck: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 it says "Fix handling of Class>>binding: - Verify that the registered class in the environment is the receiver" _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Comment #8 on issue 6680 by [hidden email]: Failing test: testInstanceSideMethodsWithNilKeyInLastLiteral http://code.google.com/p/pharo/issues/detail?id=6680 yes, but it will always return the wrong binding... the binding is fixed in classbuilder due to the becomeForward: much later. _______________________________________________ Pharo-bugtracker mailing list [hidden email] http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker |
Free forum by Nabble | Edit this page |