You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is actually a spec change; "bad" values should throw an error by design, not be ignored by design. Currently, a negative effort limit implies ignoring a joint limit...
For handling bad semantics, look into using min and max values in the spec descriptions
For non-scalar values like //joint/axis/xyz, libsdformat should enforce that the axis is a unit vector by either re-normalizing or requiring it to be normalized.
Also, spec should say that it will disregard non-unit vectors for axes.
We should do PR for libsdformat10 as an error.
If we have extra time, backport to libsdformat9, but as a warning. (With updates to Gazebo Classic + Ignition Gazebo.)
The text was updated successfully, but these errors were encountered:
scpeters
changed the title
Make parsing of bad values (syntactically and semantically) more strict
Parsing of bad values (syntactically and semantically) should be more strict
Apr 30, 2020
Example of bad syntax: #228
Example of bad semantics: RobotLocomotion/drake#12863 (review)
This is actually a spec change; "bad" values should throw an error by design, not be ignored by design. Currently, a negative effort limit implies ignoring a joint limit...
For handling bad semantics, look into using
min
andmax
values in the spec descriptionsFor non-scalar values like
//joint/axis/xyz
, libsdformat should enforce that the axis is a unit vector by either re-normalizing or requiring it to be normalized.Also, spec should say that it will disregard non-unit vectors for axes.
We should do PR for libsdformat10 as an error.
If we have extra time, backport to libsdformat9, but as a warning. (With updates to Gazebo Classic + Ignition Gazebo.)
The text was updated successfully, but these errors were encountered: