Ingo Kresse
2009-06-25 17:33:28 UTC
Hi everybody,
I was putting some Hokuyo URG04LX lasers to work with ROS and fixed a
small glitch in the diagnostics system in the process.
I tried to create a ticket in the Trac System but it refused with:
Trac Error
Submission rejected as potential spam
Any hint what I was missing? The settings were:
Name: Ingo Kresse (kresse-xrfDFxQfymSzQB+***@public.gmane.org)
Priority: trivial
Component: ros
Type: defect
all the rest was empty, except the summary, description and the name.
I attached the patch to this mail, together with the bug description
and one other minor issue regarding diagnostics.
The patches are against the SVN repository from today.
Best regards,
Ingo Kresse
---
Bug Description:
The driver for the Hokuyo laser range scanner sends a DiagnosticMessage,
reporting (amongst other things) the actual and desired frequency of
laser scans.
Currently the desired frequency in this message is hard coded to 40Hz,
so for a URG-04LX (working at 10Hz) the driver reports an error (but
works fine otherwise).
Attached is a patch which makes the driver use the information that it
obtained from the laser.
---
One other (minor) issue: The frequency reported using DiagnosticUpdater
was always too high by 1 because the update interval was almost always
one iteration longer than one second (one second is the desired
diagnostic message period). However for calculating the frequency the
desired period (1.0 seconds) is used via DiagnosticUpdater::getPeriod().
How about if this method returns the time that has actually passed? We
would get much better measurements.
I did a grep over the personalrobots tree, and getPeriod() was only used
in this same way to compute a frequency. A proposal for a patch is
attached.
I was putting some Hokuyo URG04LX lasers to work with ROS and fixed a
small glitch in the diagnostics system in the process.
I tried to create a ticket in the Trac System but it refused with:
Trac Error
Submission rejected as potential spam
Any hint what I was missing? The settings were:
Name: Ingo Kresse (kresse-xrfDFxQfymSzQB+***@public.gmane.org)
Priority: trivial
Component: ros
Type: defect
all the rest was empty, except the summary, description and the name.
I attached the patch to this mail, together with the bug description
and one other minor issue regarding diagnostics.
The patches are against the SVN repository from today.
Best regards,
Ingo Kresse
---
Bug Description:
The driver for the Hokuyo laser range scanner sends a DiagnosticMessage,
reporting (amongst other things) the actual and desired frequency of
laser scans.
Currently the desired frequency in this message is hard coded to 40Hz,
so for a URG-04LX (working at 10Hz) the driver reports an error (but
works fine otherwise).
Attached is a patch which makes the driver use the information that it
obtained from the laser.
---
One other (minor) issue: The frequency reported using DiagnosticUpdater
was always too high by 1 because the update interval was almost always
one iteration longer than one second (one second is the desired
diagnostic message period). However for calculating the frequency the
desired period (1.0 seconds) is used via DiagnosticUpdater::getPeriod().
How about if this method returns the time that has actually passed? We
would get much better measurements.
I did a grep over the personalrobots tree, and getPeriod() was only used
in this same way to compute a frequency. A proposal for a patch is
attached.