Summoning covariant class leads to Nothing in type argument
              
              #17708
            
            
                  
                    
                      soronpo
                    
                  
                
                  started this conversation in
                Feature Requests
              
            Replies: 3 comments
-
| I think this should be a feature request then. | 
Beta Was this translation helpful? Give feedback.
                  
                    0 replies
                  
                
            -
| OK, I wasn't sure this isn't a bug. Tried to transfer it to lampepfl/dotty-feature-requests but I have no privileges to so. | 
Beta Was this translation helpful? Give feedback.
                  
                    0 replies
                  
                
            -
| OK, I transferred it now. | 
Beta Was this translation helpful? Give feedback.
                  
                    0 replies
                  
                
            
  
    Sign up for free
    to join this conversation on GitHub.
    Already have an account?
    Sign in to comment
  
        
    
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Removing covariance of
TinFuzzyresults in successful compilation.Not sure that this is a bug, since this could be relevant to Scala's eager widening of non-invariant parameters, since Scala 2 has a similar issue.
Compiler version
v3.2.0-RC2
Minimized code
https://scastie.scala-lang.org/DB2gSGGzT5SLtkbSkskmNA
The following code fails similarly in Scala 2 (and succeeds when covariance is removed):
https://scastie.scala-lang.org/hRFQ9LLCQf2Jbxi9jRy76g
Output
Expectation
No error, if this is indeed a bug. If it's expected result from the widening mechanism, then we need a future language feature to specify the type argument must be precise/exact/narrow.
Beta Was this translation helpful? Give feedback.
All reactions