Bug in C4 calculation

[davidorme]

There is a problem with the test for assimilation == GPP when using C4=TRUE:

> ret <- rpmodel(tc = 20, vpd = 1000, co2 = 400, ppfd = 30, elv = 0, c4=TRUE, fapar=1)
Error in rpmodel(tc = 20, vpd = 1000, co2 = 400, ppfd = 30, elv = 0, c4 = TRUE,  : 
  rpmodel(): Assimilation and GPP are not identical.
> ret <- rpmodel(tc = 20, vpd = 1000, co2 = 400, ppfd = 30, elv = 0, c4=FALSE, fapar=1)
>

The lines are:

    assim <- ifelse(a_j < a_c, a_j, a_c)
    if (any(abs(assim - gpp/c_molmass) > 0.001)) 
        stop("rpmodel(): Assimilation and GPP are not identical.")

I can't immediately see what is wrong - it isn't NA issues as in #11 and the values genuinely are unequal. Having said that, I think all.equal is more robust here as well.

Also - the first line there creating assim: the previous code has just tested that a_j == a_c, so presumably a_j or a_c could be used directly rather than going to the computational expense of combining them?