You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
TL;DR in theory IntelliJ's sbt compiler should not be working in Scala 2 because generatedNonLocalClass is not called, but in reality it does work, which might mean there is a bug (e.g. maybe assuming all classes are Local classes?)
Problem
it used to be the case that generatedNonLocalClass is called in "xsbt-analyzer" phase, which always runs when the bridge wraps the compiler - even if AnalysisCallback is disabled (enabled == false).
in sbt/zinc#582 (in 2018) this call was moved to the "xsbt-api" phase, which only runs when the AnalysisCallback is enabled (enabled == true).
however it seems that perhaps classes that should be generatedNonLocalClass are actually being reported as generatedLocalClass - cc @lrytz, @vasilmkd and @Gedochao
this investigation is prompted by discussion on scala/scala3#21179 where I raised the point that in theory Scala 3 copied the behaviour of Scala 2 by moving generatedNonLocalClass calls to ExtractAPI, but in practice it behaves differently
The text was updated successfully, but these errors were encountered:
TL;DR in theory IntelliJ's sbt compiler should not be working in Scala 2 because
generatedNonLocalClass
is not called, but in reality it does work, which might mean there is a bug (e.g. maybe assuming all classes are Local classes?)Problem
it used to be the case that
generatedNonLocalClass
is called in"xsbt-analyzer"
phase, which always runs when the bridge wraps the compiler - even ifAnalysisCallback
is disabled (enabled == false
).in sbt/zinc#582 (in 2018) this call was moved to the
"xsbt-api"
phase, which only runs when theAnalysisCallback
is enabled (enabled == true
).however it seems that perhaps classes that should be
generatedNonLocalClass
are actually being reported asgeneratedLocalClass
- cc @lrytz, @vasilmkd and @Gedochaothis investigation is prompted by discussion on scala/scala3#21179 where I raised the point that in theory Scala 3 copied the behaviour of Scala 2 by moving
generatedNonLocalClass
calls to ExtractAPI, but in practice it behaves differentlyThe text was updated successfully, but these errors were encountered: