Two devices of the same type in the Bastode results in choosing JFETs only. Unfortunately I don't have a Pch JFET around, but it should work in theory.
According to the datasheet, J112 has 5mS transconductance if set for 1mA and ZVP2106a has around 10mS so they are quite comparable. I know the reality could be much worse but even if there's a 1:100 difference that only one is actively "transconducting" and the other one just passes the current, treat it as a single ended input stage and it's not the end of the world. It all depends on the linearty of these devices.
At least under my setup using FFT for THD measuring, open-loop distortion is around 1% and at 20kHz it still has more than 60dB of gain. This means a configuration of gain=10 has around 0.01% close-loop distortion across the audio band. I don't have a device to confirm this, and 0.01% isn't even close to the nowaday hi-fi standard. But for two not-the-same-type transistors this is already acceptable as a normal-fi circuit. If someone has a pair of complementary JFETs I'd like to hear how that influences the distortion.
Best regards.