change ghc-config to use the path ghc instead of TEST_HC

[JoshMeredith]

This change uses the hard-coded system $PATH ghc to build ghc-config instead of the make variable TEST_HC, which brings the make behaviour in line Hadrian. Ideally we would use GHC_STAGE0 from config.mk, but I have been unable to get that to work via imports. Running the testsuite via make test before and after the changes (without a cross-compiler target) results in the same outputs, so it seems like this isn't causing any side-effect changes.

Alternately, I also considered matching the locators for other tools located in this same boilerplate.mk file, as follows, but this resulted in the same error for whatever reason.

ifeq "$(GHC_PKG)" ""
GHC_PKG := $(call find_tool,ghc-pkg)
endif

And these all result in an error message matching this:

make -C testsuite/tests SUMMARY_FILE=../../testsuite_summary.txt
make[1]: Entering directory '/home/josh/projects/ghc/config-test/ghc/testsuite/tests'
"" --make -o ../mk/ghc-config ../mk/ghc-config.hs
/bin/sh: 1: : Permission denied
PYTHON="python3" "python3" ../driver/runtests.py -e "ghc_compiler_always_flags='-dcore-lint -dstg-lint -dcmm-lint -no-user- -rtsopts  -Werror=compat -dno-debug-output'" -e "config.compiler_debugged=False" -e ghc_with_native_codegen=False -e "config.leading_underscore=False" -e config.have_vanilla=True -e config.have_dynamic=True -e config.have_profiling=False -e ghc_with_threaded_rts=False -e ghc_with_dynamic_rts=False -e config.have_interp=True -e config.unregisterised=False -e config.have_gdb=False -e config.have_readelf=True -e config.have_fast_bignum=False -e config.ghc_dynamic=False -e ghc_with_smp=False -e windows=False -e darwin=False -e config.in_tree_compiler=True -e config.cleanup=True -e config.local=True --rootdir=. --config-file=../config/ghc --top="/home/josh/projects/ghc/config-test/ghc/testsuite" -e 'config.platform=""' -e 'config.os=""' -e 'config.arch=""' -e 'config.wordsize=""' -e 'config.timeout=int() or config.timeout' -e 'config.exeext=""' --config 'compiler="/home/josh/projects/ghc/config-test/ghc/inplace/bin/ghc-stage2"' --config 'ghc_pkg="/home/josh/projects/ghc/config-test/ghc/inplace/bin/ghc-pkg"' --config 'haddock=' --config 'hp2ps="/home/josh/projects/ghc/config-test/ghc/inplace/bin/hp2ps"' --config 'hpc="/home/josh/projects/ghc/config-test/ghc/inplace/bin/hpc"' --config 'gs="gs"' --config 'timeout_prog="../timeout/install-inplace/bin/timeout"' --config 'stats_files_dir=../tests/perf/haddock' -e "config.stage=" --summary-file "../../testsuite_summary.txt"   --rootdir=../../libraries/array/tests  --rootdir=../../libraries/base/tests  --rootdir=../../libraries/binary/tests  --rootdir=../../libraries/bytestring/tests  --rootdir=../../libraries/deepseq/tests  --rootdir=../../libraries/directory/tests  --rootdir=../../libraries/exceptions/tests  --rootdir=../../libraries/filepath/tests  --rootdir=../../libraries/ghc-compact/tests  --rootdir=../../libraries/ghc-heap/tests  --rootdir=../../libraries/ghc-prim/tests  --rootdir=../../libraries/haskeline/tests  --rootdir=../../libraries/hpc/tests  --rootdir=../../libraries/pretty/tests  --rootdir=../../libraries/process/tests  --rootdir=../../libraries/stm/tests  --rootdir=../../libraries/template-haskell/tests  --rootdir=../../libraries/text/tests  --rootdir=../../libraries/unix/tests \
	 \
	 \
	 \
	 \
	 \
	 \
	
Traceback (most recent call last):
  File "../driver/runtests.py", line 90, in <module>
    exec(e)
  File "<string>", line 1
    config.stage=
                ^
SyntaxError: invalid syntax
make[1]: *** [../mk/test.mk:356: test] Error 1
make[1]: Leaving directory '/home/josh/projects/ghc/config-test/ghc/testsuite/tests'
make: *** [Makefile:224: test] Error 2

with this excerpt indicating that the variable isn't defined properly

"" --make -o ../mk/ghc-config ../mk/ghc-config.hs
/bin/sh: 1: : Permission denied

Given that the make builds aren't sticking around for very long, I don't think it's worth it to spend further time debugging when the system ghc bandaid appears to work.

[luite]

this is okay as a temporary fix, but should be changed to use the stage0 ghc before we try to merge upstream. Better add a comment that we really want the STAGE0 GHC here.