Fix bug in compiler, and add tests

[ChrisJefferson]

This patch fixes #1781 (which was easy, as there was a fix already in that PR!), but more usefully adds a simple testing infrastructure for the GAP->C compiler, and adds some simple tests (including for the fixed bug).

Obviously, the number of tests here is very small, but it provides a basic sanity check and a place for people to add more tests if they want.

[codecov]

Codecov Report

Merging #1789 into master will increase coverage by 0.04%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1789      +/-   ##
==========================================
+ Coverage   62.93%   62.97%   +0.04%     
==========================================
  Files         969      969              
  Lines      293986   293986              
  Branches    12986    12962      -24     
==========================================
+ Hits       185013   185140     +127     
+ Misses     106162   106022     -140     
- Partials     2811     2824      +13

Impacted Files	Coverage Δ
src/intrprtr.c	71.34% <ø> (ø)	⬆️
lib/queue.g	66.4% <0%> (-3.2%)	⬇️
src/objset.c	82.36% <0%> (-0.24%)	⬇️
src/read.c	83.9% <0%> (+0.07%)	⬆️
src/lists.c	56.71% <0%> (+0.11%)	⬆️
src/funcs.c	72.1% <0%> (+0.14%)	⬆️
hpcgap/lib/hpc/stdtasks.g	38.87% <0%> (+0.25%)	⬆️
src/permutat.c	76.4% <0%> (+0.9%)	⬆️
src/exprs.c	79.59% <0%> (+1.68%)	⬆️
src/compiler.c	65.28% <0%> (+3.35%)	⬆️
... and 1 more

[fingolfin]

Instead of renaming .out to .output, you could also just adjust the rule in .gitignore... :-) E.g. replacing it by this:

/doc/*/main.out
/log/*.out

[ChrisJefferson]

Touching the .gitignore seems to lead to bikesheding, so I thought it was easier to just rename the files intead...

[fingolfin]

@ChrisJefferson not sure why you think changing entries in .gitignore leads to bike shedding? Those entries all are there to hide build artifacts. In some cases, they are pretty generic, for simplicity; but we can always replace a relatively "general" entry with more specialized ones, as is needed here. I see no room for bikeshedding whatsoever.

And even if, then we should actively combat that, instead of avoiding to make certain changes.

That said, if you prefer .output over .out, that's of course also OK, but I will now change the .gitignore rule for .out anyway...

[fingolfin]

Specifically, I already change .gitignore in PR #1780

[fingolfin]

@ChrisJefferson I am sorry but this now conflicts :/. Would you terribly mind updating it, then we can merge it ASAP.