Hi there. This is not broken at all. Just as life=0 spng can never drain tmp=1 gel, tmp=100 gel can never fully saturate life=99 spng.

Gel and spng are not "equals", rather, gel has a *slightly* higher affinity for water, so if 1 spng and 1 gel are together, if you put 1 pixel of water on the spng, then if the gel gets near it will steal this saturation.

The relevant code is shown below. I have bolded the relevant sections. Note that the exchanges between gel and spng only happen when life[spng]+1=/=tmp[gel]
So when tmp is 100 and life is 99, these are equal. When tmp is 1 and life is 0, they are also equal. 

(From GEL.cpp) https://github.com/The-Powder-Toy/The-Powder-Toy/blob/master/src/simulation/elements/GEL.cpp

case PT_SPNG:
// Concentration diffusion
if (parts[ID(r)].life>0 && parts[i].tmp<100 && ((parts[ID(r)].life+1)>parts[i].tmp))
{
// SPNG -> GEL
parts[ID(r)].life--;
parts[i].tmp++;
}
else if (parts[i].tmp>0 && (parts[ID(r)].life+1)<parts[i].tmp)
{
// SPNG <- GEL (saturation limit of SPNG is ignored here)
parts[ID(r)].life++;
parts[i].tmp--;
}

Tldr its not a bug its a feature