kernel: fix bug in MarkWeakPointerObj during copying

[fingolfin]

We did not mark the forwarding pointer when copying. As a result, a badly timed GC could have reaped the copy of the wpobj, and then when we try to finalize the copying, a crash (or, if we were really unlucky, corrupt data) could have resulted.

I never observed this manifest, though; I only discovered it while trying together with @rbehrends to track down a crash with Julia GC; while inspecting our code in PR #2688, I realized that the bug fixed here exists (and that I had "accidentally" fixed it in PR #2688 for Julia GC, because I changed the code there to not use MarkWeakPointerObj anymore).

This issue is also resolved by PR #2777, which completely refactors the copying code, and essentially makes this kind of bug impossible.

[codecov]

Codecov Report

Merging #2780 into master will increase coverage by 7.77%.
The diff coverage is 50%.

@@            Coverage Diff             @@
##           master    #2780      +/-   ##
==========================================
+ Coverage    68.1%   75.87%   +7.77%     
==========================================
  Files         480      481       +1     
  Lines      241225   241312      +87     
==========================================
+ Hits       164275   183099   +18824     
+ Misses      76950    58213   -18737

Impacted Files	Coverage Δ
src/weakptr.c	97.36% <50%> (-0.36%)	⬇️
hpcgap/lib/hpc/stdtasks.g	63.42% <0%> (-0.77%)	⬇️
lib/teachmod.g	12.94% <0%> (ø)
src/pperm.c	97.88% <0%> (+0.05%)	⬆️
src/vecgf2.c	72.83% <0%> (+0.05%)	⬆️
src/io.c	59.9% <0%> (+0.07%)	⬆️
grp/simple.gi	77.63% <0%> (+0.11%)	⬆️
src/gap.c	85.18% <0%> (+0.14%)	⬆️
lib/wordrep.gi	66.61% <0%> (+0.16%)	⬆️
src/exprs.c	96.41% <0%> (+0.17%)	⬆️
... and 182 more