mux readouts suffer from "IQ offsets" that vary depending on the filter-bank channel, which was not accounted for by the workaround in #254. This would mostly only bite you if your readout frequency was precisely at a channel's center frequency, but now that case is also covered.
if you used pulse() with a list of channels, and set t='auto', your pulse times would be weird (the t parameter would get overwritten by the computed pulse time for the first channel, and then the other channels would use that value)
the QickConfig printout was just broken for firmwares with constant-IQ generators, since #209.
t='auto', your pulse times would be weird (thetparameter would get overwritten by the computed pulse time for the first channel, and then the other channels would use that value)