asl builder proof-of-concept

[pq]

food for thought.

caveat: i'm flummoxed as to why explicit unloading of asl is needed (where it isn't for other libraries brought in by a require, like, for example beat_clock).

anyway, the idea is a norns script with a line like this:

require 'asl'

will then be able to do things like:

crow.output[1].action = loop{ to(0,0), to(5,0.1), to(1,2) }
crow.output[1].action = lfo(1, 5) 

and have it behave like in druid.
...

the dynamic case is tricky of course and requires a low-level send. so, something like:

    crow.send[[
      output[1].action = ar(1, math.random(5))
    ]]

the bet is that a string-builder like this will be useful in the common-case and at least not confusing in the complex one...

it has the benefits of feedback from linters and IDEs (since it's valid lua) and scores bonus points for highlighting nicely in maiden

and generally in IDES (like vs code):

as for the global pollution issue (see #557), asl functions only get added to _G if asl is included and only hang around until the next script is loaded.

i think that's probably the right thing.

i should emphasize that i think of this purely as food for thought and i'm not hugely invested in it landing as is (or at all). i just wanted to see the idea through...

fixes: #890

/cc @tehn @catfact @trentgill

[tehn]

this is great! and perplexing about the asl require bug. i'll take a minute to look at it.

[tehn]

i'm thinking i'll merge this in parallel with a doc update. i like the approach, thanks for taking the time

[tehn]

ack, sorry for lagging on this. need to find to test. will get it in the next-next update (should be forthcoming)

[pq]

no worries. more food for thought than anything!