Remnants of previous strokes infecting current stroke

[morinted]

It seems like strokes can intrude with each other. I'm using the repeatandhold branch, but I tested on master as well.

If you press strokes in sequence, like SK-S THEPL SK-S THEPL… etc., sometimes, the output will be TKHEPL or THEPLS, despite very deliberately not hitting the misstroke.

[morinted]

Charley helped me look into this and compare it with his Forth code...

Seems like it's a debouncing issue caused by not clearing the bytes in the debounce loop. Additionally, there's a possible issue when setting the pins to INPUT instead of INPUT_PULLUP

diff --git a/stenomod.ino b/stenomod.ino
index 11ee736..d926034 100644
--- a/stenomod.ino
+++ b/stenomod.ino
@@ -45,7 +45,7 @@ void set_output(uint8_t p) {

 // Set to input mode
 void set_input(uint8_t p) {
-  pinMode(pin[p], INPUT);
+  pinMode(pin[p], INPUT_PULLUP);
 }

 // Read the current byte, where
@@ -90,6 +90,7 @@ void scan_keys() {
   b[0] = b[1] = b[2] = b[3] = 0;

   while(key_pressed == false) {
+    b[0] = b[1] = b[2] = b[3] = 0;
     while(look() == false);
     delay(20);
     key_pressed = look();
@@ -121,4 +122,3 @@ void loop() {
     send_stroke();
   }
 }
-

[morinted]

Okay, that's not quite right, Charley walked me through some of the finer points. This commit outlines the solution/improvements: morinted@230acad

I'd be happy to make PRs to both master and repeatandhold if you'd like.