On IRC, eikke asked me if concatenating Inputs and Outputs was efficient (i.e. whether or not it can wait on multiple ones efficiently). It's worth investigating if the underlying STM machinery is efficient enough for this purpose or if it could be made more efficient by overriding mconcat for each type.
On IRC,
eikkeasked me if concatenatingInputs andOutputs was efficient (i.e. whether or not it can wait on multiple ones efficiently). It's worth investigating if the underlying STM machinery is efficient enough for this purpose or if it could be made more efficient by overridingmconcatfor each type.