run tests. Made some progress but bogged down in scalacheck when
I realized things were farther from working than I'd thought. Please
forgive this brief expression of frustration, because the despair is
palpable: the weeks of work I did on partest was apparently of no
interest, and now I'm in here having to fix the same partest
bugs again, in a more virulent form. They've evolved into some
kind of ant-partest-sbt artificial creature, more fearsome than any
one of them could be under its own power.
Did we not know we have a whole directory full of pretty code like this?
List(
"Scala compiler classes in: " + testBuildDir,
"Scala version is: " + nsc.Properties.versionMsg,
"Scalac options are: " + universe.scalacOpts,
"Java binaries in: " + javaBin,
"Java runtime is: " + javaInfoString,
"Java runtime options: " + (Process.javaVmArguments mkString " "),
"Javac options are: " + universe.javacOpts,
"Java options are: " + universe.javaOpts,
"Source directory is: " + src,
"Selected categories: " + (selectedCategories mkString " "),
""
) mkString "\n"
It's src/partest-alternative in case anyone wants to consider any of it.
At the very least there are individual files or functions which could be
cherry picked.
-- This patch contains: --
Readability efforts. Twelve parameter functions whose last six
parameters are "true,true,failedOnly,true,false,log" are pretty
challenging to work with.
Changes to the environment variables sbt was setting to look a bit
more like what partest expects and will crash without.
Sets partest.debug=true if the sbt logger level is at Debug.
On my machine all test groups except scalacheck appear to run
to completion.
Review by moix. (Hello, moix! I am available for assistance if you
are in the market for it. You have brought us long suffering ant-ers
to the brink of sbt goodness: and now, we enter the fray!)
git-svn-id: http://lampsvn.epfl.ch/svn-repos/scala/scala/trunk@22976 5e8d7ff9-d8ef-0310-90f0-a4852d11357a
acf4abf883