Update GaugeIncrementsAsm.cpp

[veliusiidx]

Fixed hard gauge's behavior to accurately reflect how LR2 behaves by adding additional checks for when the gauge goes below 32 and if the total is at a certain value.

The changes made are based on exch-bms2/beatoraja#628

[MatVeiQaaa]

I'm not sure if the edge case for gauge<=32 you implemented is correct, as it was already in the GaugeIncrementsAsm.h file. So effectively now drain reduction is calculated twice and the end result is wrong. Not sure about total, i would have to test that, but in current form its based off of LR2 code and is correct according to my testing. The issue with total was that some charts have total value overrode, instead of calculated value its dictated by .bms file explicitly. It's only present for hard gauge because for other gauges the code gets total value from game memory, instead of calculating that, so it's always correct.

Also you made a pull request into your own repository, i think that might have not been intended :)