in show, use StringView not String

[ThomasBreuer]

I think that the show method for GAP objects fits better to GAP's ViewObj than to String.
Ideally, ViewString should be called but this is not (yet) supported for enough GAP objects.
Thus StringView is a reasonable temporary alternative.
(Cf. gap-packages/JupyterKernel#104.)

[codecov]

Codecov Report

Merging #248 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master     #248   +/-   ##
=======================================
  Coverage   70.07%   70.07%           
=======================================
  Files          50       50           
  Lines        3262     3262           
=======================================
  Hits         2286     2286           
  Misses        976      976

Impacted Files	Coverage Δ
src/gap2.jl	46% <100%> (ø)	⬆️

[ThomasBreuer]

One of the Travis tests failed, due to a change of a type name in AbstractAlgebra:
MatSpaceElem instead of Mat.
The other tests succeeded, they seem to use an older version of AbstractAlgebra.
At the moment, I see no way to get a consistent output.
(Should I better avoid showing type names in tests?)

[mohamed-barakat]

(Should I better avoid showing type names in tests?)

@ThomasBreuer: Yes, maybe this is a good idea for the moment.

[ThomasBreuer]

@mohamed-barakat
Note that @fingolfin has already disabled the tests in question, as a part of pull request #253.

For the future, I would still like to have some check whether a Julia object created in a specific conversion is reasonable.
If the typeof string is not reliable then a compromise would be to check for a supertype, using isa, and to hope that the names of these supertypes are stable.
(This has nothing to do with type stability, just type name stability.)

For example, a matrix m in AbstractAlgebra and Nemo can be detected with isa( m, MatElem ).
There seem to be no general functions such as ismatrix.

[mohamed-barakat]

I think that the show method for GAP objects fits better to GAP's ViewObj than to String.
Ideally, ViewString should be called but this is not (yet) supported for enough GAP objects.
Thus StringView is a reasonable temporary alternative.
(Cf. gap-packages/JupyterKernel#104.)

I was unaware of StringView and for me it seems like the correct choice anyway.

[mohamed-barakat]

I rebased this PR locally and it works perfectly for me.

[mohamed-barakat]

Approved once rebased.

[mohamed-barakat]

May I resolve by rebasing and force-pushing?