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
Sometimes, after leaving my rack powered on for a long time, the 2nd output of TB-3PO gets stuck high. Everything else appears to work fine, the pitch CV still comes out normally, I can switch to other applets and they work, but the problem persists with TB-3PO until I cycle power. It's hard to reproduce because it only happens after it's been sitting idle for probably >24 hours... but it's happened more than a few times.
My hunch is that it has something to do with the EndOfADCLag() function: bool EndOfADCLag(int ch) { return (--adc_lag_countdown[ch] == 0); }
If you overflow a regular signed int by subtracting 1, what happens? Will it eventually rollover and end up back at 0?
The text was updated successfully, but these errors were encountered:
oh hi, just stopping by to say I finally fixed this bug recently in my fork: djphazer@9457cb9
It was related to signed vs. unsigned type math errors, getting stuck after 36 hours... the tick counter rolls over after 72 hours, but a signed int32 will rollover at half that.
Sometimes, after leaving my rack powered on for a long time, the 2nd output of TB-3PO gets stuck high. Everything else appears to work fine, the pitch CV still comes out normally, I can switch to other applets and they work, but the problem persists with TB-3PO until I cycle power. It's hard to reproduce because it only happens after it's been sitting idle for probably >24 hours... but it's happened more than a few times.
My hunch is that it has something to do with the EndOfADCLag() function:
bool EndOfADCLag(int ch) { return (--adc_lag_countdown[ch] == 0); }
If you overflow a regular signed int by subtracting 1, what happens? Will it eventually rollover and end up back at 0?
The text was updated successfully, but these errors were encountered: