Moderator: Mmiscool
If anybody's found out otherwise, I'd like to know, so I could re-edit my list of gotchas.
A thought on your gpio03 interrupt inconsistency - floating inputs tend to float high, so when used for something, may have had a pulldown resistor added (typically 10K), in which case adding a similar value pullup would probably still leave the input in an indeterminate midway condition. Try using a 1K or 2K pullup and you might make the interrupt dependable.
I don't have time to waste unnecessarilly chasing shadows, so I'd be interested in having a peek at your Gotcha's list if that was possible?
I have indeed been using a pullup of 10k. I'll try lowering that today.
But how does that explain polling still working reliably, I wonder?