crow input modes not changing?

[rwhaling]

I found that crow's input[1] would not respond to triggers in change mode, even though I could validate with input[1].query() that the input signal was rising above the threshold (and back below the hysteresis level); meanwhile, input[2] responds just fine to the same signals.

More details in this thread: https://llllllll.co/t/crow-help-norns/25863/23, including a few other folks with similar issues.

I think what is going on is - crow is disregarding the .mode() message that norns sends, because crow's method is called set_mode():
https://github.com/monome/crow/blob/master/lua/input.lua#L39
whereas the norns crow.input.mode() method definitely sends .mode() in the command text:
https://github.com/monome/norns/blob/master/lua/core/crow.lua#L56

So I think, as it is now, norns cannot set input modes at all for crow, but people have been masking it by relying on input 2 being set to change mode via First.

A fix might be as easy as adjust the command to run set_mode()? I don't know either codebase well but I can make an initial PR if that helps.

[tehn]

i am not able to replicate this bug. can you post a snippet?

also crow will report lua errors back to norns in the maiden window. mode is definitely correct--- see here: https://github.com/monome/crow/blob/master/lua/input.lua#L75

[rwhaling]

thanks for taking a look @tehn - sorry, I should have added a quick script to repro instead of a PR. try this:

-- counts gates at input 1 or 2
-- press key 2 to print volts at input 1
-- press key 3 to print volts at input 2

local count_1 = 0
local count_2 = 0

function init()
  crow.input[1].mode("change",1.0,0.1,"rising")
  crow.input[1].change = change_1
  crow.input[2].mode("change",1.0,0.1,"rising")
  crow.input[2].change = change_2
  screen.level(15)
  screen.aa(0)
  screen.line_width(1)
end

function change_1(s)
  count_1 = count_1 + 1
  redraw()
end

function change_2(s)
  count_2 = count_2 + 1
  redraw()
end

function redraw()
  screen.clear()
  screen.move(10,60)
  screen.text("count_1: "..count_1.." count_2:" .. count_2)
  screen.stroke()
  screen.update()
end

function key(n,z)
  if n==2 and z==1 then
    print(crow.input[1].query())
  elseif n==3 and z==1 then
    print(crow.input[2].query())
  end
end 

On my machine, only gates at input 2 get counted.
Despite that, I can query input 1 and see that the voltage is flipping between 0.05 and 7.64 - but the handler function never fires and the count does not increment

[tehn]

sortof good news is that your script works fine on my machine.

is your crow on v1.0.0 ?

and i just pushed some changes to the crow code on norns. i'll have an update out tomorrow, or you can ssh in, pull the changes and re-run ./waf (be sure to ;restart after in maiden)

[rwhaling]

mysterious! I did need to update crow, but no change in behavior from norns.
I also confirmed that I could set the modes of both inputs and assign change handlers via druid.
Thanks again for taking a look! I'll try again when the update is out tomorrow.

[tehn]

holy moly--- are you saying it's working as expected on druid?

[rwhaling]

yep -

function init()
    input[1].mode('change',1,0.1,'rising')
    input[2].mode('change',1,0.1,'rising')
end

input[1].change = function()
    print("BEEP")
    output[1].volts = math.random() * 10.0 - 5.0
end

input[2].change = function()
    print("BOOP")
    output[2].volts = math.random() * 10.0 - 5.0
end

behaves as I would expect on druid. a real head-scratcher!

[rwhaling]

I think I've resolved this?
The update did not fix the issue, but a crow.clear() did.

My engineering brain still wants to figure out what the underlying issue is, and how I managed to accidentally upload a script that disabled only one of the inputs, but I'm glad it's working.

Thanks so much for the help, and for making this amazing device! Closing the issue.