Ergonomic, hermetic, user-friendly software build system for Python, Go, Java, Kotlin, Protobuf, Scala, Shell. Pants lets you fearlessly scale up your codebase!

Pants

not in BuildFileAliases. sorry. too easy.

<@U06A03HV1>: Did the cache double check logging/behavior change since 0.0.49?

I'm no longer seeing "double checking the cache" in the logs

but to answer your question: i don't think so, but am not sure

it logs at debug when it misses, and logs at info when it hits

<@U04S45AHA>: so, using `2.11` by default causes a bunch of tests to fail while since we don't override it there, and then try to resolve things from the root of the repo

mind if i postpone making that the default until after we can guarantee those will match?

example: ```:: org.scalastyle#scalastyle_2.11;0.3.2: not found```

Hey <@U0AH2MXCK> (who is the sbt expert right?): If I'm in a compiler plugin, how do I figure out the output directory where the classfiles are going for a given compilation unit?  Does scalac just always use cwd?  Or does zinc control this somehow?

We want the plugin to output a text file alongside the classfiles so it gets picked up as a resource later on in the pipeline

Aha, I think it's just `settings.outdir.value` because of the way we invoke zinc/scalac

<@U053HNE90>: I know little about sbt internals, <@U06A03HV1> may know

How close are we to being able to get rid of group task?

Now that zinc does all our jvm compilation?

I don't think we have any other uses of it

Also I'd love to be able to bootstrap compiler plugins in-repo with a group

No, apt is handled since these are just java targets

I'd say kill it if bootstrapping plugins is only at the convenience level

To be more clear - extra-ordering for apt is only needed when compiling &gt;1 target at a time.  Otherwise a target needing the apt could be scheduled in the same 2 or more target chunk, which would mean the apt would not run in that round

javac only runs an apt if its in classfile form

I think scalac is the same - it only runs plugins in classfile form

So these have the exact same scheduking problem in 2+ target compilation schemes