Engine exchange strays

Tested this locally with and without MPI. Didn't actually test sending around gparts, because we can't really deal with them quite yet.

See merge request !117

Conflicts:
	src/engine.c
	src/engine.h
	src/part.c
	src/part.h
	src/proxy.h
	src/scheduler.c

This reverts commit 4ab23609

Redistribute fixes



See merge request !131

Redistribute fixes

Fixes #130 and makes `engine_split()` work for gparts.

That lead to some issues in common_io.c but the master works if you don't run with -g. (And -g does not work beyond one time-step anyway as we don't have the MPI communications for gparts yet.)

See merge request !130

Redistribute gparts

Initial version of `engine_redistribute()` which sends around the gparts as well as the parts + preserves their links.

@nnrw56 what do you think ?
I'll need to update some bits later to take into account your changes to the MPI_types and the new version of `space_part_sort()`

See merge request !124

Rewait fixes

Fixes the deadlock described in #125.

See merge request !122

Support for XMF file descriptors with multiple particle types.

This fixes #127 and adds support for all particle types within the XMF files. Both in single node and MPI modes.

See merge request !126

Local variable cleanup

Yet another batch of cleanups.

See merge request !123

Save a sqrtf() call per particle in drift task

Any reason why this was not written like this in the first place ? Looks like we can save one `sqrtf()` call per particle in the drift tasks.

See merge request !127

Initialize some variables that MemorySanitizer suggested

Also reported issues in MPI and METIS, the METIS issue doesn't seem to have an effect (zeroed in a local build).



See merge request !125

Also reported issues in MPI and METIS, the METIS issue doesn't seem to have an effect (zeroed in a local build).