Bugfix: ns-name return type #44
Conversation
Since Named/getName:String and old Namespace/getName:Symbol
arrdem
force-pushed
the
bugfix/ns-name
branch
from
c75c3a7
to
f7369a1
Compare
| @@ -4037,7 +4037,7 @@ | |||
| :static true | |||
| :deprecated "1.9"} | |||
| [ns] | |||
| (.getName (the-ns ns))) | |||
| (symbol (.getName (the-ns ns)))) | |||
There was a problem hiding this comment.
I did wonder about changing the return type, but thought there was a good reason for it.
There was a problem hiding this comment.
I wanted Namespace and Var to participate in Named, because they do have names and we have an interface for that so it makes sense and I just got a Java method name conflict.
In retrospect I shouldn't have merged that change without thinking about it more and trying to analyze uses. Amusingly, nothing in core broke because of the type chance, since the uses were all (name (ns-name (the-ns ..))), so (name String) -> String and (ns-name Ns) -> String) same result type.
Deprecating ns-name was also probably a mistake in retrospect, since although name is mostly a drop-in replacement it does have a unique purpose. Maybe there's call here for a sym-name or something that's generic to Named. Need to think rather than just merging.
Was: Symbol
Became: String
Fixed to Symbol again.