Delete dead, name-forcing code - fixing nightlies #13901
Conversation
dwijnand
force-pushed
the
random-nightly-fix
branch
from
bcc56c8 to
f83886f
Compare
|
This fixes CI but breaks locally running the test, at least for me. |
| assert(1234567890123456789L.round == 1234567890123456789L) // error | ||
| assert(math.round(1234567890123456789L) == 1234567890123456789L) // error | ||
| assert(123456789.round == 123456789) // error (Int to Flaot) | ||
| assert(math.round(123456789) == 123456789) |
There was a problem hiding this comment.
That behavior seems correct. round is overloaded, so there is no expected type yet. This seems to indicate that the tests are done in the wrong phase. Maybe we should test this only after Typer?
There was a problem hiding this comment.
Maybe test assert(math.round(123456789) == 123456789) in another test file to test the other error/warning.
There was a problem hiding this comment.
Yes, the test was correct. I just wanted a pull request ready that we could merge today to unbreak the nightly builds. I spent some time with Lukas and I think we should be able to merge #13911 instead. If something comes up I'll PR moving the test instead.
There was a problem hiding this comment.
Ironically, @nicolasstucki, it's thanks to the fact that these were in the same file that we even found that the wrong warning was being emitted. I'm still not sure I like this new "// error" with no message assertion practice that scala 3 has introduced..
|
Duplicate of #13911 |
[test_java8]