My prior glitch prevention relied on previous pin configuration.  This update sets all values.

Also, I think I have found a way to reset the I2C hardware to the same state as a Power Cycle or Hardware reset.

These changes should remove the possibility of a signal glitch when starting or restarting the Wire() object.

Also, I think I have found a way to reset the i2c hardware to a Power On or Hardware Reset state.

After hardware reset, all i2c hardware register are back to zero, so the clockrate needs to be recovered.

I had manually applied these changes and misspelled two function names!

Continue with I2C transaction if bus_busy is cleared by i2cInit()

`SCL` and `SDA` need to be detached from the i2c hardware before the i2c hardware is init'd else a glitch is generated.  esp32-hal-i2c.c does not know which pins are assigned to the hardware.  Return the error and let `TwoWire()` do the reinit.

Also, '~' is not the same as '!'  

My Bad.

The i2c data bus must be disconnected from the i2c peripheral before it is init'd. Else it will generate a glitch on both pins.  Since the Hal layer does not remember which pins are assigned, just return the error and let `TwoWire()` handle re-initing.

had to move the hardware init code out of begin, so that I could call it without destroying queued transactions

Reinit i2c hardware when bus fault (busy) is detected