When does the sample fail completely?
It is incorrect to say that failure must occur when
G = R
There will be some cracking but complete failure (as in tension) also requires that
$${{{{\rm{d}}^2}U(c)} \over {{\rm{d}}{c^2}}} < 0$$
i.e. the energy is at a maximum, or
$${{{\rm{d}}G} \over {{\rm{d}}c}} > {{{\rm{d}}R} \over {{\rm{d}}c}}$$
In other words failure will be catastrophic when the rate of increase of the driving force with crack growth is greater than the change in R with crack growth, which we have taken as a constant.
Alternatively cracking will be stable when
$${{{{\rm{d}}^2}U(c)} \over {{\rm{d}}{c^2}}} > 0$$
i.e. the energy is at a minimum, or
$${{{\rm{d}}G} \over {{\rm{d}}c}} < {{{\rm{d}}R} \over {{\rm{d}}c}}$$
That is, as the crack grows, the resistance to cracking, R, increases faster than the driving force, G.