Fix gac on HPC-GAP and re-enable edim (WIP, DO NOT MERGE)

[fingolfin]

This is some work on getting gac to work for HPC-GAP, and to thus resolve issue #1295. Right now, it does not yet actually change gac, but at least it enables the test (namely, trying to build edim) which reveals this.

Help with this is welcome! :-).

[codecov]

Codecov Report

Merging #1444 into master will decrease coverage by 3.07%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1444      +/-   ##
==========================================
- Coverage   64.42%   61.34%   -3.08%     
==========================================
  Files        1002      909      -93     
  Lines      326709   266419   -60290     
  Branches    13197    13514     +317     
==========================================
- Hits       210477   163434   -47043     
+ Misses     113363   100073   -13290     
- Partials     2869     2912      +43

Impacted Files	Coverage Δ
extern/gmp/mpn/generic/gcdext_1.c	0% <0%> (-100%)	⬇️
extern/gmp/mpn/generic/toom_eval_dgr3_pm2.c	0% <0%> (-100%)	⬇️
extern/gmp/rand/randdef.c	0% <0%> (-100%)	⬇️
extern/gmp/mpn/generic/toom4_sqr.c	0% <0%> (-100%)	⬇️
extern/gmp/mpn/generic/toom_couple_handling.c	0% <0%> (-100%)	⬇️
...ern/gmp/mpn/generic/matrix22_mul1_inverse_vector.c	0% <0%> (-100%)	⬇️
extern/gmp/tal-reent.c	0% <0%> (-100%)	⬇️
extern/gmp/mpn/generic/toom33_mul.c	0% <0%> (-100%)	⬇️
extern/gmp/mpz/set_ui.c	0% <0%> (-100%)	⬇️
extern/gmp/mpn/generic/toom3_sqr.c	0% <0%> (-100%)	⬇️
... and 365 more

[fingolfin]

I'll do this differently in a new PR.