Hi
Johan and I were debugging why some of the GRSlimeTest fail in recent
Pharo versions. We found out that the code the tests compile contains
undeclareds. Does anybody know whether this is intentional or an
oversight? We believe it's an oversight.
The issue with #testUsesCanPerformOrUnderstand is that the code for
the 'valid' method references 'new' instead of '#new'.
The issue with #testBlockFormatting is that several times 'a' and 'b'
aren't declared as block arguments.
testUsesCanPerformOrUnderstand
| class |
class := self
defineClass: #GRSlimeMockObject
superclass: #GRObject.
class
compile: 'invalid1
Object class canUnderstand: #new'
classified: #(accessing).
class
compile: 'invalid2
Object class canPerform: #new'
classified: #(accessing).
class
compile: 'valid
Object respondsTo: new'
classified: #(accessing).
self
runRule: GRUsesCanPerformOrUnderstandRule
selectors: #(invalid1 invalid2)
testBlockFormatting
| class |
class := self
defineClass: #GRSlimeMockObject
superclass: #GRObject.
class
compile: 'invalid1
[a ]'
classified: #(accessing).
class
compile: 'invalid2
[:a | ]'
classified: #(accessing).
class
compile: 'invalid3
[ b]'
classified: #(accessing).
class
compile: 'invalid4
[ :a| ]'
classified: #(accessing).
class
compile: 'invalid5
[ :a |b ]'
classified: #(accessing).
self
runRule: GRBlockFormattingRule
selectors: #(invalid1 invalid2 invalid3 invalid4 invalid5)
Cheers
Philippe
_______________________________________________
seaside-dev mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
Locked
