These are chat archives for sbt/sbt

29th
Jun 2017
Rob Norris
@tpolecat
Jun 29 06:41
So, if I have compiling code and make a change that breaks something, then I undo my changes sometimes sbt just instantly says [success] Total time: ... without any [info] Compiling message. It's as if it has cached the previous good state. Am I imagining something?
It may only happen with ~compile, not sure.
@eed3si9n hate to bug you but it smells like the kind of thing nobody else would know ^
Dale Wijnand
@dwijnand
Jun 29 08:04
@tpolecat yeah if the reverted state hashes to the same value, the incremental compiler with match it with the already existing class files and not do any unnecessary work.
That said, of course incremental compiler bugs are a thing, so please report if you see errors.
Rob Norris
@tpolecat
Jun 29 10:10
@dwijnand cool, thanks. No errors, I was chasing down an issue with an editor plugin that seemed to correspond with that phenomenon.
Anthony Garo
@agaro1121
Jun 29 14:23
hey guys, when you call “reload” from within the sbt repl, does it discard the -Dparams passed in originally?
Dale Wijnand
@dwijnand
Jun 29 14:26
no
Anthony Garo
@agaro1121
Jun 29 14:31
thank you
Dale Wijnand
@dwijnand
Jun 29 14:31
no worries
eugene yokota
@eed3si9n
Jun 29 15:45
@tpolecat i think you are witnessing "transactional" classfile manager, which saves the old *.class filles related to the changed source in a temporary place before embarking on incremental compilation, and if the round of incremantal compilation results in an error, it places the old *.class files back. so if you undo right away, the round 2 of incremental would be no-op.
Rob Norris
@tpolecat
Jun 29 16:11
Cool, thanks.