Fix MPI buffer max_leading_slices

[AlexanderSinn]

As noticed by @huixingjian, the hipace.comms_buffer_max_leading_slices option doesn't work sometimes. This is because there was one slice where incoming metadata could be received too early, if the Isend of the same slice would happen instantly. Furthermore, I noticed that in some situations (all ranks compute slices faster or at the same speed as the head rank) it is possible to go into a deadlock where none of the leading slices are filled, so I adjusted the assert and documentation to only include the trailing slices.

• Small enough (< few 100s of lines), otherwise it should probably be split into smaller PRs
• Tested (describe the tests in the PR description)
• Runs on GPU (basic: the code compiles and run well with the new module)
• Contains an automated test (checksum and/or comparison with theory)
• Documented: all elements (classes and their members, functions, namespaces, etc.) are documented
• Constified (All that can be const is const)
• Code is clean (no unwanted comments, )
• Style and code conventions are respected at the bottom of https://github.com/Hi-PACE/hipace
• Proper label and GitHub project, if applicable

[MaxThevenet]

Thanks for the fix!