How to make two referencing sub-projection works?

I have below code, which doesn’t work. h and g are 2 projections defined inside f , and they are referencing to each other.

Theorectically, the answer should be:

f[10]=g[10]

=h[8] + 20

=g[5] + 30 + 20

=h[3] + 20 + 30 + 20

=g[0] + 30 + 20 + 30 + 20

=h[-2] +20 + 30 + 20 + 30 + 20

= -2 + 120 = 118

f:{[xF]
g:{[h;xG] $[xG < 0; xG; h[xG - 2] + 20]}[h]; // doesn't work, as h is not yet defined
h:{[g;xH] $[xH < 0; xH; g[xH - 3] + 30]}[g];
g[xF]
};
f[10];

 How to make it works?

This is a chicken and the egg situation. Define g and h without the projections first then define them with the projections. 

f:{[xF]
g:{[h;xG] $[xG < 0; xG; h[xG - 2] + 20]};
h:{[g;xH] $[xH < 0; xH; g[xH - 3] + 30]};
g:g[h];h:h[g];
g[xF]
};

 

f:{[xF] g:{[h;xG] $[xG < 0; xG; h[.z.s; xG - 2] + 20]}; h:{[g;xH] $[xH < 0; xH; g[.z.s; xH - 3] + 30]}; g[h; xF] }; f[10];

Another option is to use .z.s to refer to the calling function when calling the other function.

Thanks for your reply. But the above code doesn’t work. I got an “type: Mismatched types” error.

With f[10]? Can’t spot why and I didn’t run the above as on mobile. Not sure why it doesn’t work

yes, f[10] doesn’t work.

No worries. Kshephard’s solution is nicer anyway. Will take a look later when at a pc to see why.