test: pass correct objcopy and ar paths to the test runner

[labath]

Summary

The test runner has code to autodetect this, but it's not very smart --
in particular, it fails in the case where we build the test executables
with the just-built clang. Since cmake already has the knowledge about
the right toolchain, we can just have it pass the appropriate flags to
the test runner.

This also removes the "temporary" cache-scrubbing hack added a couple
months ago.

lldb/trunk/test/CMakeLists.txt

if( LLDB_DEFAULT_TEST_ARCH STREQUAL "host" )
   string(REGEX MATCH "^[^-]*" LLDB_DEFAULT_TEST_ARCH ${LLVM_HOST_TRIPLE})
 endif ()
 
 # Allow the user to override the default by setting LLDB_TEST_ARCH
 set(LLDB_TEST_ARCH
   ${LLDB_DEFAULT_TEST_ARCH}
   CACHE STRING "Specify the architecture to run LLDB tests as (x86|x64).  Determines whether tests are compiled with -m32 or -m64")
 
-# Scrub LLDB_TEST_COMPILER out of the CMake caches
-# TODO: remove the replace lines and the FORCE parameter in a few days once the
-# change has made its way through bots to clean up their CMake caches.
-string(REPLACE "-C;${LLDB_TEST_COMPILER}" "" LLDB_TEST_USER_ARGS_New "${LLDB_TEST_USER_ARGS}")
-
 if(LLDB_TEST_CLANG)
   set(LLDB_TEST_COMPILER $<TARGET_FILE:clang>)
 endif()
 
 # Users can override LLDB_TEST_USER_ARGS to specify arbitrary arguments to pass to the script
 set(LLDB_TEST_USER_ARGS
   "${LLDB_TEST_USER_ARGS_New}"
   CACHE STRING "Specify additional arguments to pass to test runner. For example: '-C gcc -C clang -A i386 -A x86_64'" FORCE)

 if(LLDB_CODESIGN_IDENTITY)
   list(APPEND LLDB_TEST_COMMON_ARGS --codesign-identity "${LLDB_CODESIGN_IDENTITY}")
 endif()
 
 if(LLDB_BUILD_FRAMEWORK)
   list(APPEND LLDB_TEST_COMMON_ARGS --framework $<TARGET_FILE_DIR:liblldb>)
 endif()
 
+if (NOT ${CMAKE_SYSTEM_NAME} MATCHES "Windows|Darwin")
+  list(APPEND LLDB_TEST_COMMON_ARGS
+    --env ARCHIVER=${CMAKE_AR} --env OBJCOPY=${CMAKE_OBJCOPY})
+endif()
+
 add_python_test_target(check-lldb-single
   ${LLDB_SOURCE_DIR}/test/dotest.py
   "--no-multiprocess;${LLDB_TEST_COMMON_ARGS};${LLDB_TEST_USER_ARGS}"
   "Testing LLDB with args: ${LLDB_TEST_COMMON_ARGS};${LLDB_TEST_USER_ARGS}"
   )
 
 set(LLDB_DOTEST_ARGS -q;${LLDB_TEST_COMMON_ARGS};${LLDB_TEST_USER_ARGS})
 
 # If tests crash cause LLDB to crash, or things are otherwise unstable, or if machine-parsable
 # output is desired (i.e. in continuous integration contexts) check-lldb-single is a better target.
 add_python_test_target(check-lldb
   ${LLDB_SOURCE_DIR}/test/dotest.py
   "${LLDB_DOTEST_ARGS}"
   "Testing LLDB (parallel execution, with a separate subprocess per test)"
   )