Integration methods (summary issue)

[mstimberg]

This is a placeholder issue to not forget about a few potential improvements/fixes to our integration methods.
For exponential Euler:

• Staggered updates
• [WIP/discuss] More strict check that equation is conditionally linear #1143 (side note)

For exact integration:

• minor annoyance: method fails for constant (e.g. dv/dt = 0*Hz : 1)

Others:

• Add a backward Euler state updater #479