Rangefinder arming checks #11078
Rangefinder arming checks #11078
Conversation
| hal.util->snprintf(failure_msg, failure_msg_len, "Rangefinder %d was not detected", i + 1); | ||
| return false; | ||
| } | ||
| } |
There was a problem hiding this comment.
you could add a second check that the status is ok ! see : https://github.com/ArduPilot/ardupilot/pull/9886/files#diff-03adf2804cc10b90bbfe953b5c07b33eR91
it would catch correct instanciation but late deconnexion !
There was a problem hiding this comment.
NoData is a valid state though if you are out of range, but don't know which direction you are out of range. (Which is common when on the ground for example and the rangefinder is mounted low on the vehicle).
WickedShell
force-pushed
the
wickedshell/rangefinder-arming
branch
from
ad62059
to
a613a4d
Compare
|
Outcome of the dev call was for @tridge to help @WickedShell with the parameter conversion magic. |
tridge
force-pushed
the
wickedshell/rangefinder-arming
branch
from
a613a4d
to
5eb88f0
Compare
|
issue with Sub is now resolved |
libraries/AP_Param/AP_Param.cpp
Outdated
| @@ -1096,7 +1096,9 @@ void AP_Param::save(bool force_save) | |||
| // when we are disarmed then loop waiting for a slot to become | |||
| // available. This guarantees completion for large parameter | |||
| // set loads | |||
| hal.scheduler->expect_delay_ms(500); | |||
There was a problem hiding this comment.
Unit mixup.
I suspect this kind of thing is going to be so common, that unless we wrap this internally into the delay() call in the HAL with a check if we are on the main thread, this is going to be a constant problem everywhere. Given the relative rarity of calling delay() I think it's worth the overhead.
WickedShell
force-pushed
the
wickedshell/rangefinder-arming
branch
from
5eb88f0
to
1758cd2
Compare
this prevents a race in sub where a set_default_by_name() directly follows a parameter conversion for ARMING_CHECK. Without the flush the default is written to storage by the IO queue draining after the set()
WickedShell
force-pushed
the
wickedshell/rangefinder-arming
branch
from
1758cd2
to
f388548
Compare
This starts on a global set of arming checks on rangefinders. For now it simply introduces the requirement that if a rangefinder was specified, then we must have created a driver for it.
This was actually pretty easy by itself, unfortunately it turns out we have consumed all the bits for
ARMING_CHECKSso I had to add param conversion for it, to get it to have 32 bits (we can still only use 24 of them sadly).This also makes it easier to dump the keys for conversion, just to make it a single define change in the future.
This is a draft as the conversion in Sub is not working correctly. For reference this is what the key dump prints on Sub:
Key 75: Index 0: GroupElement 2 : ARMING_CHECK: 17