Repository navigation
Replies: 2 comments 6 replies
|
Oh, i should say this doesn't happen at totally random times. It happens when a different port gets poked. So if I actually poke the center poke, there's like a 20% chance the left or right poke will immediately register a spurious poke, and an 80% chance it will function properly and detect no poke. |
|
Hm, also happens even if the GPIO inputs for the other pokes are tied directly to 3.3V, they can sometimes still pick up a spurious "poke". Starting to suspect a software issue. |
|
interesting, no I've never seen this, how are you seeing the pokes be triggered? like some callback registered to them? |
|
i'm running the free_water task and also the nafc task, same phenomenon in each. so i'm judging this by the response of the task, not inspecting the callbacks. I do have the pilot running with DEBUG loglevel, so I can see the events flying by, but not adept enough to interpret all of them yet |
|
another fascinating and totally critical bug. those register the which should only register callbacks for the particular pin the GPIO is set up for... the only thing I can think of would be something fishy happening if you have you might try adding a debug logging call inside handle_trigger to see if you get multiple calls to it when you trigger a single port?? |
|
it's odd because we also use those tasks regularly and trying to jam on kff-target ports is definitely part of my regular testing process |
|
lol will debug. I'm sure it's some weird interaction of some reasonable assumption and some innocuous decision I made in setting stuff up here. we'll learn a lot from whatever it turns out to be. On the plus side, I'm free water training mice here with autopilot!! Would like to upload a video but that no doubt violates some university policy. But you can imagine |
|
happy debugging, lmk if it doesn't reveal itseld. as always, the possibility of there being some preposterously badly written hunk of code is distinctly nonzero ;p |
Uh oh!
There was an error while loading. Please reload this page.
Does anyone have issues with the pokes triggering themselves even when nothing pokes them? This happens some fraction of the time for me, maybe 20%. I hooked up a voltmeter to the output of the IR poke detector, and it reads a stable 3.3V whether or not a poke is spuriously detected. The voltage does goes to zero during an actual poke, as expected. I'm not sure if it's a hardware or software issue.. any guesses?
All reactions