Validation on market and position decimals in market proposals

[davidsiska-vega]

Thanks to Jiajia we now have a feature test that can demonstrate that I can set up a market and drain someone’s monies by shifting price up slowly and then bringing it down quickly:
vegaprotocol/vega#10990

The problem is that the validation we have is market decimals >= asset decimals.
The validation we should have is: market decimals + position decimals >= asset decimals.

This would ensure that the smallest mark-to-market cashflow caused by the smallest price change on the smallest position is

10^(-mdp) x 10^(-pdp) = 10^(-(mdp+pdp)) >= 10^(-adp)

so it’s never possible to be in a situation where small price moves don’t register but big ones do.

[Jiajia-Cui]

The problem is that the validation we have is market decimals <= asset decimals.
The validation we should have is: market decimals + position decimals <= asset decimals.