Overview

This page lists identified bugs against the date at which they were identified and/or fixed, and describes any impacts and workarounds. Bug fixes are published as part of MDK releases—to establish whether your platform has a given bug, establish whether the firmware element specified in the bug report installed on your platform has a release tag that comes after the bug fix date.

Bugs in release 170616

170706a (Two copies of miro_bridge can be started)

Module
P3.
Bug
A second copy of miro_bridge can be started when one is already running (easy to do accidentally if autostart is enabled).
Fix
The fix modifies miro_bridge so that it cannot be started twice.
Impact
Two copies of bridge share the USB driver and both copies will not function correctly (network may be lost). Various warning messages are emitted, typically of the type shown below.
[007.250] **** WARNING **** partial frame not assembled (C0, F97, L74-98) yf/yi mismatch [007.251] **** WARNING **** released reassembly buffer without delivering it

170626a (Frictionless floor in simulator pet_shop model)

Module
Simulator.
Bug
Floor element of pet_shop model used in default miro.world is near-frictionless and MIROs sat on top of it can't drive around successfully.
Fix
The fix corrects the friction parameters, mu and mu2, from 0.1 to 50, in element visual_5 of model pet_shop. Find lines 233 and 234 of file mdk/sim/gazebo/models/pet_shop/model.sdf and change the value on both lines.
Impact
MIROs are not controllable when in the pet_shop. Workaround is to manually correct the friction parameter or to not use the pet_shop model.

Bugs in release 170423

170601a (P2 RTC skew)

Module
P2.
Bug
RTC skew driven by circadian model and immediate sensory input is enabled if the flag AFFECT_UPDATE_RTC is NOT set, which is semantically and functionally incorrect.
Fix
The fix changes behaviour so that RTC skew is enabled if the flags AFFECT_ENABLE and AFFECT_UPDATE_RTC are both set.
Impact
There is no impact unless the user is making use of the RTC clock provided by P2. Workarounds might include ignoring the RTC clock in P2 (there is one available in P3, as an alternative, which is not affected), or using the flags in the indicated (incorrect) way to control the RTC skew in P2.