Fix #211

[lihaoyi]

Ensure that transitive ivy jars are not directly aggregated under a module's runClasspath; only the ivy coordinates are aggregated, and every module resolves it's own set of ivy jars using coursier, which avoids duplicate jars entering the classpath through ivy

Fixes #211

Not sure why the tests are sad on Java 9

[rockjam]

Would be great to see what caused java.lang.reflect.InvocationTargetException in java 9 tests

all java.lang.reflect.InvocationTargetException
    java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    java.base/java.lang.reflect.Method.invoke(Method.java:564)
    mill.define.Module$Internal.$anonfun$reflect$2(Module.scala:67)
    scala.collection.TraversableLike$WithFilter.$anonfun$map$2(TraversableLike.scala:739)
    scala.collection.IndexedSeqOptimized.foreach(IndexedSeqOptimized.scala:32)
    scala.collection.IndexedSeqOptimized.foreach$(IndexedSeqOptimized.scala:29)
    scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:191)
    scala.collection.TraversableLike$WithFilter.map(TraversableLike.scala:738)
    mill.define.Module$Internal.reflect(Module.scala:61)
    mill.main.ResolveTasks$.endResolveLabel(Resolve.scala:125)

[robby-phd]

I managed to reduplicate the issue on my machine.
Based on @lihaoyi's comment here, I tried removing ~/.mill and it seems to workaround it (see https://travis-ci.org/lihaoyi/mill/builds/359079895). So, it seems the issue is Ammonite caching (no idea why the caching issue is exposed only in Java 9; I guess it's a ... feature 😆).

[robby-phd]

@lihaoyi I changed some ci test scripts and appveyor config in master to remove ~/.mill before running tests; if you rebase/merge, that should make this PR green. It can be reverted when the caching issue cause is identified and resolved.

[Ammonite-Bot]

@robby-phd great, will give it a shot. I have a tentative fix for the caching problem in the `test` branch of the Ammonite repo, so hopefully a fix isn’t too far away

…

On Tue, 27 Mar 2018 at 2:43 PM, Robby ***@***.***> wrote: @lihaoyi <https://github.com/lihaoyi> I changed some ci test scripts and appveyor config in master to remove ~/.mill before running tests; if you rebase/merge, that should make this PR green. It can be reverted when the caching issue cause is identified and resolved. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#261 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATpwjAXcAmcchY8RYs2yFgqgNSy9350_ks5tirJogaJpZM4S9Ry1> .