Global fixes / Shared fixes

[andresmanelli]

If there's a fix that shares a run with another fix, this will cause problems, because for the moment the script looks for a fix and some commands associated with it until the next fix (or empty #OP, or end of the simulation).

The problem resides in the way the restart file is created. The search for the last operation will erase the previous.

Example:

#OP
fix FIX1 all nve

#OP
fix heat [...]
dump [...]
restart [...]
run 1000
unfix heat

#OP
fix cool [...]
dump [...]
restart [...]
run 2000
unfix cool

unfix FIX1

#OP
fix last [...]
[...]
run 1000

Here, If the simulation stops at 500 steps (only part of the second fix performed), the first fix will be ereased because it is not associated (as viewed by the script) to the fisrt run 1000. Then when unfix-ing it, lammps will stop due to a not found fix-ID.