test_zinc_compile_integration.py?

How about a sibling new test file dedicated to zinc+cache

Or does that file above have utils you want to leverage?

it doesn't

a new one makes more sense

so, the new way of overriding tool classpaths is just defining the target in the root, and if the target exists it is used?

no

so, its just like the old way - except you can only pass 1 spec not a list

zinc: any/random/spec/pointing/to/my/override

if you want that spec to be at the root in BUILD.tools for nostalgia or convention - you can

and the default spec == old, not a list, so //:zinc

Apologies for verbose output, but is this expected - in particular the M2Coordinate for hamcrest being the junit coordinate?:

Found cp tuples for target JvmBinary(BuildFileAddress(FilesystemBuildFile(/home/jsirois/dev/3rdparty/jsirois-pants/src/java/org/pantsbuild/tools/junit/BUILD), main)):
	(u'default', ClasspathEntry(path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/compile/jvm/java/classes'))
	(u'default', ArtifactClasspathEntry(coordinate=M2Coordinate(org='args4j', name='args4j', rev='2.32', classifier=None, jar=u'jar'), path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/ivy/jars/args4j/args4j/bundles/args4j-2.32.jar'))
	(u'default', ArtifactClasspathEntry(coordinate=M2Coordinate(org='com.google.guava', name='guava', rev='18.0', classifier=None, jar=u'jar'), path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/ivy/jars/com.google.guava/guava/bundles/guava-18.0.jar'))
	(u'default', ArtifactClasspathEntry(coordinate=M2Coordinate(org='junit', name='junit', rev='4.12', classifier=None, jar=u'jar'), path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/ivy/jars/junit/junit/jars/junit-4.12.jar'))
	(u'default', ArtifactClasspathEntry(coordinate=M2Coordinate(org='junit', name='junit', rev='4.12', classifier=None, jar=u'jar'), path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/ivy/jars/org.hamcrest/hamcrest-core/jars/hamcrest-core-1.3.jar'))
	(u'default', ArtifactClasspathEntry(coordinate=M2Coordinate(org='commons-io', name='commons-io', rev='2.4', classifier=None, jar=u'jar'), path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/ivy/jars/commons-io/commons-io/jars/commons-io-2.4.jar'))
	(u'default', ArtifactClasspathEntry(coordinate=M2Coordinate(org='com.google.code.findbugs', name='jsr305', rev='3.0.0', classifier=None, jar=u'jar'), path=u'/home/jsirois/dev/3rdparty/jsirois-pants/.pants.d/ivy/jars/com.google.code.findbugs/jsr305/jars/jsr305-3.0.0.jar'))

I think its not expected, I'll have an RB with the fix and reasoning this afternoon.

ok, sounds good.

and the deletion of all of the

zinc: ['//:zinc']

stuff was primarily because they were all defaulted anyway, ya

thanks john.

Yes - I deleted all BUILD.tools except for name_[scala platform] ones

That will take a bit more work to do right

yea. i should upstream our platform stuff anyway

Aha - yes, that would be great

I haven't started on that fix, but I'm close

(not much there, but it makes a world of difference)

The hard part on that one is there is no natural forum for injecting the scala library dep for the active ScalaPlatform

the current solution injects a spec

the new one needs to inject a live synthetic JarLibrary

oh. i see.

@witty-crayon-22786: @bored-art-40741 @enough-analyst-54434 howdy, just joined. let me catch up on chat history to learn what's been the progress on zinc memory issue and performance, like @witty-crayon-22786 said, we are very focused on improving performance

@limited-country-20626 is actively profiling right now

excellent, thanks guys.

very eager to knock this out.