Fix crash computing zero(ZZ[t]) - zero(ZZ)

[fingolfin]

The following lead to a crash (uncovered by Nemocas/AbstractAlgebra.jl#1867)

using Nemo
R, t = ZZ[:t]
zero(R) - ZZ(0)

The backtrace is always the same:

[57675] signal 11 (2): Segmentation fault: 11
in expression starting at REPL[4]:1
fmpz_neg at /Users/mhorn/.julia/artifacts/f76e8c291f81b2f313a6368345a2c17de479af1e/lib/libflint.19.0.dylib (unknown line)
sub! at /Users/mhorn/Projekte/OSCAR/Nemo.jl/src/flint/fmpz_poly.jl:878
- at /Users/mhorn/Projekte/OSCAR/Nemo.jl/src/flint/fmpz_poly.jl:153
...

This is due to a bug in FLINT's fmpz_poly_sub_fmpz, which does this:

void fmpz_poly_sub_fmpz(fmpz_poly_t res, const fmpz_poly_t poly, fmpz_t c)
{
    if (poly->length == 0)
    {
        fmpz_poly_set_fmpz(res, c);
        fmpz_neg(res->coeffs + 0, res->coeffs + 0);   // <- CRASH IS HERE
    }
    else
    {
        fmpz_poly_set(res, poly);
        fmpz_sub(res->coeffs + 0, res->coeffs + 0, c);
        _fmpz_poly_normalise(res);
    }
}

The problem is that even after the fmpz_poly_set_fmpz(res, c) we still have res->length == 0.

Fix for FLINT: flintlib/flint#2102

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.38%. Comparing base (9dd324d) to head (5544d49).
Report is 5 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1930      +/-   ##
==========================================
- Coverage   87.40%   87.38%   -0.03%     
==========================================
  Files          97       97              
  Lines       35642    35521     -121     
==========================================
- Hits        31154    31040     -114     
+ Misses       4488     4481       -7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.