Fix HPC-GAP's testinstall

[fingolfin]

We currently do not run resp. ignore testinstall on HPC-GAP, as there are various diffs in there. Some diffs are caused by the kernel unification process. Others need to be addressed separately.

[olexandr-konovalov]

From what I now see in Jenkins testing hpc-merge-full branch, only two diffs remain:

########> Diff in /Volumes/hudson-fs/workspace/GAP-merge-compilers/GAPCOPTS/64\
build/GAPREADLINE/noreadline/HPCMODE/hpcgap/label/clang-xcode/GAP-merge-snapsh\
ot/hpcgap/tst/testinstall/oprt.tst:39
# Input is:
MaximalBlocks(eo);
# Expected output:
[ [ 1, 3, 5, 7, 9, 11 ], [ 2, 4, 6, 8, 10, 12 ] ]
# But found:
[ [ 1, 3 .. 11 ], [ 2, 4 .. 12 ] ]
########

and

########> Diff in /Volumes/hudson-fs/workspace/GAP-merge-compilers/GAPCOPTS/64\
build/GAPREADLINE/noreadline/HPCMODE/hpcgap/label/clang-xcode/GAP-merge-snapsh\
ot/hpcgap/tst/testinstall/trans.tst:2629
# Input is:
TransformationNumber(5, 2);
# Expected output:
Error, TransformationNumber: usage, the first argument must be at most 4,
# But found:
Error, TransformationNumber: usage, the first argument must be at most 4,
Error, AppendList: <list2> must be a small list (not a integer)
########