aren't extension functions a forward compatibility...
# announcementsb
Bernhard
08/10/2018, 7:45 AMaren't extension functions a forward compatibility hazard? e.g. you add a function to a class and the lib adds the same method with a different behavior later on. Now your code breaks
a
Andreas Sinz
08/10/2018, 8:07 AMYes, that could happen
e
edwardwongtl
08/10/2018, 8:19 AMDoesn’t your own extension function shadows the original function?
k
karelpeeters
08/10/2018, 8:40 AMNo, member functions always have precedence, https://kotlinlang.org/docs/reference/extensions.html#extensions-are-resolved-statically:
If a class has a member function, and an extension function is defined which has the same receiver type, the same name and is applicable to given arguments, the member always wins.
karelpeeters
08/10/2018, 8:41 AMIt's strange it isn't a compilation error then, since you'd never be able to call the extension function.
a
alex2069
08/10/2018, 8:41 AMI believe there is a flag for it
alex2069
08/10/2018, 8:41 AMAnd it does show a warning in the source (but yea - IMO it should be a compile error)
e
edwardwongtl
08/10/2018, 8:42 AMI think at least should be a warning saying the extension function never called.
a
alex2069
08/10/2018, 8:42 AMBut it might be called depending on the visibility/scope
alex2069
08/10/2018, 8:43 AMIf you shadow a package-private (Java) function then else where outside of package it might work fine? (not 100% sure on that)
k
karelpeeters
08/10/2018, 8:43 AMAlso an extension function on base class that's shadowed by a child's member function.
a
alex2069
08/10/2018, 8:45 AMIt should be noted this only breaks at the source - compiled code wont break (since it's compiled to a static function with the receiver as the first arg)
k
karelpeeters
08/10/2018, 8:52 AMWhich doesn't really help much because how often are dependencies updated without a compile cycle...
a
Andreas Sinz
08/10/2018, 8:58 AMIf it is compiled and the dependency is updated, it still uses the extension function
đź‘Ś 1
4 Views