Hydro tests recently broken

The MHD_API merge (83f57e77) seems to have broken the hydro tests in the planetary ((and minimal, see below)) hydro scheme (which are direct copies of the examples/HydroTests/ versions, just adding the particles' ideal-gas material IDs to the initial conditions), e.g. examples/Planetary/KelvinHelmholtz_2D:

./configure --with-hydro=planetary --with-equation-of-state=planetary --with-hydro-dimension=2 --enable-debug --enable-debugging-checks --disable-optimization
make -j 8
cd examples/Planetary/KelvinHelmholtz_2D/
python3 makeIC.py
../../../swift --hydro --threads=4 -e -v 1 kelvinHelmholtz.yml

Output:

...
#   Step           Time Scale-factor     Redshift      Time-step Time-bins      Updates    g-Updates    s-Updates sink-Updates    b-Updates  Wall-clock time [ms]  Props    Dead time [ms]
       0   0.000000e+00    1.0000000    0.0000000   0.000000e+00    1   56        97928            0            0            0            0               685.615     25             5.487
[00000.9] engine_step: Writing step info to files took 0.079 ms
[00000.9] engine_step: Updating general quantities took 0.003 ms
[00000.9] engine_prepare: Communicating rebuild flag took 0.002 ms.
[00000.9] engine_drift_all: Drifting all to t=1.716614e-05
[00000.9] cell_drift.c:cell_drift_part():275: Particle drifts by more than a box length! id 65289 xp->v_full 2.94912e+05 -5.89824e+05 0.00000e+00 p->v 3.68635e+04 -7.37280e+04 0.00000e+00
[00000.9] cell_drift.c:cell_drift_part():275: Particle drifts by more than a box length! id 7062 xp->v_full 1.13394e+08 9.43718e+06 0.00000e+00 p->v 1.41742e+07 1.17965e+06 0.00000e+00
Aborted

From git bisect, everything runs fine up to and including 83f57e77. The quick_update_planetary merge was needed to get the code to compile, but I noticed we were still missing some tweaks that were added to Minimal but not Planetary, now added on the planetary_update branch. But this doesn't change the error.