Negative value corrections in calc_ftemp_kphio

[davidorme]

Hi Beni,

There's an issue with the correction to filter out negative values in calc_ftemp_kphio in version 1.1.2 . The max() function finds the global maximum of the inputs, which of course collapses matrix and array inputs down to a scalar (and gets hung up on missing values and NaN).

> tc <- array(c(NA, -9:28, NaN), dim=c(4,5,2))
> calc_ftemp_kphio(tc)
[1] NA

Substituting pmax(ftemp, 0) preserves the shape of ftemp and calculates maxima for each matrix/array entry.

[stineb]

This should have been fixed by v1.1.1 (see issue #3).
You can get this version through Github. It's submitted to CRAN (four months ago!) but hasn't been processed.

[davidorme]

I think this is a new bug. I'm using 1.1.2 installed using devtools and the problematic line is this one:

https://github.com/stineb/rpmodel/blob/master/R/calc_ftemp_kphio.R#L51

It does cause a vectorisation problem, but it is introduced by the recent commit (99f2654) that was specifically to handle negative values calc_ftemp_kphio.

I've had long waits for CRAN but at least one of them was me missing an email that required me to confirm something - worth checking the inbox!

[stineb]

This commit should solve the issue. Implemented proper vectorization.