fix windup issue

[LeanMeanBeanMachine4774]

No description provided.

[LeanMeanBeanMachine4774]

Also works if the setpoint time is updated immediatelly in the retract rather than _force_retract. I beleive we can expect the same behaviour in the wrist where doing it on enable works or in tilt_to but not _tilt_to

[LeanMeanBeanMachine4774]

Little bit of weirdness when implemented on wrist

[LeanMeanBeanMachine4774]

Weirdness is cause by removing the wrist execute from disabled not by the change itself. @LucienMorey, @auscompgeek what was the reason for that execute in the first place, isn't it a little dangerous?

[LucienMorey]

I am sceptical of this fix. I am not saying it doesn't do the bits, but moving the time reset effectively a few lines earlier doesn't make sense on its own. My observation of the robot is that it doesn't always happen, either, so maybe you're just getting lucky. Also, the bad behaviour happens a lot in test mode, which on_enable has no impact on.

[LucienMorey]

Weirdness is cause by removing the wrist execute from disabled not by the change itself. @LucienMorey, @auscompgeek what was the reason for that execute in the first place, isn't it a little dangerous?

it was added in this pr: #104 I think it was originally part of the zeroing processes when we had to index. It isn't really dangerous on its own but it is a bit smelly. It can go now.

you are also marked as a reviewer on this one :P

[LeanMeanBeanMachine4774]

I am sceptical of this fix. I am not saying it doesn't do the bits, but moving the time reset effectively a few lines earlier doesn't make sense on its own. My observation of the robot is that it doesn't always happen, either, so maybe you're just getting lucky. Also, the bad behaviour happens a lot in test mode, which on_enable has no impact on.

I understand, just testing in the sim, it seems to change the operation. Unsure why, but maybe we need to test and see

[LeanMeanBeanMachine4774]

I am sceptical of this fix. I am not saying it doesn't do the bits, but moving the time reset effectively a few lines earlier doesn't make sense on its own. My observation of the robot is that it doesn't always happen, either, so maybe you're just getting lucky. Also, the bad behaviour happens a lot in test mode, which on_enable has no impact on.

I would have to be very lucky for it to just be luck, given I tested it on off about 15 times. Although I do see no way of making it work without on enable.

[LucienMorey]

I am sceptical of this fix. I am not saying it doesn't do the bits, but moving the time reset effectively a few lines earlier doesn't make sense on its own. My observation of the robot is that it doesn't always happen, either, so maybe you're just getting lucky. Also, the bad behaviour happens a lot in test mode, which on_enable has no impact on.

I would have to be very lucky for it to just be luck, given I tested it on off about 15 times. Although I do see no way of making it work without on enable.

How are you testing this?

[LeanMeanBeanMachine4774]

I am sceptical of this fix. I am not saying it doesn't do the bits, but moving the time reset effectively a few lines earlier doesn't make sense on its own. My observation of the robot is that it doesn't always happen, either, so maybe you're just getting lucky. Also, the bad behaviour happens a lot in test mode, which on_enable has no impact on.

I would have to be very lucky for it to just be luck, given I tested it on off about 15 times. Although I do see no way of making it work without on enable.

How are you testing this?

In sim I enable in teleop, drop the intake, disable before it comes up then re enable after 5 seconds. I did this without the fix to confirm the extremely rapid retract. WIth the fix, I do the same and it retracts as expected

[LeanMeanBeanMachine4774]

Also, can confirm issue remains in test

[LeanMeanBeanMachine4774]

unknown_replay_2025.04.28-22.31.-.Trim.mp4

[LucienMorey]

Also, can confirm issue remains in test

yea this is what I was interested in.

FYI:

robotpy/robotpy-wpilib-utilities#217
robotpy/robotpy-wpilib-utilities#218

[LucienMorey]

this does the things properly now that my prs have been merged