We have noticed a potential issue with the newest VEXnet v2.0 firmware and master code. Program uploads over wireless from the PROS IDE seem to fail much more often than before, and sometimes we notice invalid code is uploaded. We urge teams using to verify that any code uploaded wirelessly using VEXnet 2.0 keys is functional before proceeding.
Due to the timing of the firmware release, we cannot debug and test an update to the PROS system in time for worlds. We recommend that teams use the tether or VEXnet 1.0 keys for uploading. After code is uploaded, we have not noticed any changes to PROS system performance, so teams using VEXnet 2.0 should perform the upgrade anyways and simply exercise caution when uploading wirelessly.
I only did minimal testing on the firmware that was released on Friday. Did you notice these issues only with V4.22 and not V4.21? Are you saying that sometimes the upload is a success (all acks received from the cortex for each packet) but the code is corrupted? I will try and duplicate this later today, I hadn’t noticed any issues but only tried your latest version a couple of times over the weekend.
One difference I have noticed that may effect this, switching to the download channel can take longer in V4.22 than V4.21, the firmware is choosing the radio channels in a different way in this version.
We have not been able to do a detailed byte analysis on the returns. From the diagnostic logs of the flasher, we have determined that the download switching delay causes most of the loud failures; downloading again solves these. However, the silent failures appear to occur when the VEX Cortex acknowledges the packets (we would get a loud failure if this was not the case) but fails to exit download mode. When cycling power, code behavior can be incorrect.