Change Test() to reject unexpected or invalid inputs

[fingolfin]

This is a somewhat experimental change to fix #2506 -- let's see if the GAP test suite passes with it.

Assuming this works for the test suite, I wonder if there are any reasonable scenarios where the new Error call is triggered and where this is undesirable. If anybody knows such an example, please bring it up. In that case, we could add an option to turn off that error; or restrict the cases in which it is triggered.

[codecov]

Codecov Report

Merging #2519 into master will increase coverage by 0.04%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master    #2519      +/-   ##
==========================================
+ Coverage   74.41%   74.46%   +0.04%     
==========================================
  Files         481      481              
  Lines      242659   245614    +2955     
==========================================
+ Hits       180586   182907    +2321     
- Misses      62073    62707     +634

Impacted Files	Coverage Δ
lib/test.gi	61.66% <0%> (-0.15%)	⬇️
src/io.c	59.88% <0%> (-13.08%)	⬇️
src/opers.c	92.1% <0%> (-1.91%)	⬇️
src/code.c	92.85% <0%> (-0.87%)	⬇️
src/calls.c	94.24% <0%> (-0.43%)	⬇️
src/hpc/tls.h	66.66% <0%> (ø)	⬆️
src/trans.c	99.81% <0%> (ø)	⬆️
src/pperm.c	97.9% <0%> (+0.01%)	⬆️
src/integer.c	92.24% <0%> (+0.04%)	⬆️
src/rational.c	96.87% <0%> (+0.05%)	⬆️
... and 11 more

[ChrisJefferson]

I suggest we merge this, and then see if any packages fail

[frankluebeck]

Agree, that there is probably some proper error if this is triggered.

[olexandr-konovalov]

Agree. Cf. also #2524 which is somehow complementary to this.