tangentstorm

@tangentstorm:

For the python wrapper, I want to either modify the cursor idea to support a "don't care" concept.

If the path down the bdd skips over a variable, it means we don't care about it. Right now, the solution iterators are designed to explicitly account for this, and do a "ripple carry" operation over the skipped variables. (Basically, we just treat that run of variables as a binary number and keep incrementing it until a variable above it flips.)

But, in the dd python wrapper for bdds, it prefers to skip over the duplicates by simply removing the variables we don't care about from the yielded assignments at each step. ... but it ALSO has a way to opt back in.

Study the code in bdd_sols.rs and come up with a way to do this while still supporting the current "iterate through every solution" behavior.

tangentstorm

{deck}
version:1
card:0
size:[512,342]
name:"itemlist.deck"

{card:home}
{widgets}
itemlist1:{"type":"contraption","size":[400,161],"pos":[13,30],"def":"itemlist","widgets":{"label":{},"item_text":{},"items":{},"bnw":{},"bgo":{},"run_event":{},"slider1":{},"bpr":{},"bnx":{},"brm":{},"bup":{},"dn":{}}}

tangentstorm

// sh.mjs: javascript shorthand

// array helpers (apl/j/k)
export const id=x=>x
export const af=(n,x)=>Array(n).fill(x) // TODO: make this 'afl' or 'fil' (aa?)
export const ii=(n,f)=>{for(let i=0;i<n;i++)f(i)}
export const im=(n,f)=>af(n,0).map((_,i)=>f(i))
export const ia=(n,f)=>im(n,id)
export const at=(a,ixs)=>ixs.map(i=>a[i])
export const io=(xs,ys)=>ys.map([].indexOf.bind(xs))

tangentstorm

customElements.define('ts-list', class extends HTMLElement {
  static observedAttributes = ['ix']
  constructor() { super(); this.ix=0 }
  connectedCallback() {
    let sh = this.attachShadow({mode: 'open'})
    sh.innerHTML=`
    <slot></slot>
    <button id="up">^</button>
    <button id="add">+</button>
    <button id="dn">v</button>`

tangentstorm

let set=(o,k,v)=>{let r={...o}; r[k]=v; return r}
let gs=k=>(x,y)=> y===undefined ? x[k] : set(y,k,x)
let mb=gs('mb'), // match bit
    ib=gs('ib'), // input buffer
    ix=gs('ix'), // index into input buffer
    ch=gs('ch'), // current character (ib[ix])
    mk=gs('mk'); // start index of current token
let s0=()=>({mb:0,ib:'',ch:'',ix:-1,mk:-1})
on=s=>ix(0,ch(s[0],ib(s,s0())))
m1=s=>mb(1,s) // match

tangentstorm

Class {
	#name : #GsEllipse,
	#superclass : #GsObject,
	#instVars : [
		'radius',
		'position'
	],
	#category : #GameSketchLib
}

tangentstorm

# i wrote this, and it works, but then i decided to just use Godot's own Expression parser.
# https://docs.godotengine.org/en/latest/tutorials/scripting/evaluating_expressions.html

func _examples():
	test('@title["the deck"]')
	test('@show["jp-editor"; 0]')
	test('@ed.xy[  0 5]')

func test(cmd:String):
	if cmd.begins_with('@'):

tangentstorm

NB. Core logic for a tiny editor in J.
NB. No select/copy/paste (in this gist), but it does support multiple cursors.
NB. This started as the code for editing a single line of text, but I'm now
NB. using three copies simultaneously: one for a single token, one for
NB. boxed tokens on a line, and one for boxed lines in a buffer.
coclass 'ed'

init =: {{
  B =: ''  NB. the buffer to edit.
  C =: 0   NB. cursor position(s)

tangentstorm

clear''
NB. Parser Combinators for J
NB.
NB. The semantics here are heavily inspired by 
NB. Allesandro Warth's ometa system:
NB.
NB.   http://tinlizzie.org/ometa/
NB.
NB. but implemented as parser combinators rather than a standalone language.
NB.

tangentstorm

const DATA=0, CALL=1, WORK=2, TEMP=3

class Worker {

  vm = null   // virtual machine
  w = 0       // worker number
  f = 0       // function pointer
  e = 0       // execution pointer
  t = 0       // current token in function