plugin low-level format

source files are always encoded in utf-8
identifiers for variables not special may include any unicode characters in groups Lu, Ll, Lt, Lm, Lo, Nd, as well as _
a token comprised only of digits is never treated as an identifier
eight registers are available as variables
special variables include >v (VVAR), >x (FTx), >y, >z, >c, <x (FPx), <y, <z, <c, >ma (TMa), >mb, ... >ml, >0 (register 0), ... >7


operation summary:
  (register manipulation)
  set v x - set v to immediate (floating-point) value x
  copy d s - set d to contents of s
  swap v w - switch contents of v and w (is this actually useful?)

  (binary operators)
  add d v w - set d to v+w
  sub d v w - set d to v-w
  mul d v w - set d to v*w
  div d v w - set d to v/w
  mod d v w - set d to v%w
  polar r t x y - set r to hypot(y, x) and t to atan2(y, x)
  rect x y r t - set x to r*cos(t) and y to r*sin(t)
  pow d v w - set d to pow(v, w)
  atan2 d v w - set d to atan2(w, v)
  hypot d v w - set d to hypot(v, w)

  (ternary)
  spherical r t p x y z - set r to sqrt(x*x+y*y+z*z), t to atan2(y, x), and p to acos(z/sqrt(x*x+y*y+z*z))
  rect3d x y z r t p - set x to r*cos(t)*sin(p), y to r*sin(t)*sin(p), and z to r*cos(p)
  fma d x y z - set d to x*y+z

  (unary)
  neg d v - d = -v
  acos d v - d = acos(v)
  asin d v - arcsine
  atan d v - arctangent
  cos d v - cosine
  cosh d v - hyperbolic cosine
  sin d v - sine
  sinh d v - hyperbolic sine
  tan d v - tangent
  tanh d v - hyperbolic tangent
  exp d v - exponential (e to the power of v)
  ln d v - natural logarithm
  sqrt d v - square root
  ceil d v - ceiling
  abs d v - absolute value
  floor d v - floor

  (shortcuts)
  addtoxy v w - add v to <x and w to <y
  addtoxyz u v w - add u to <x, v to <y, and w to <z
  vaddtoxy v w - fma >v * {v, w} into <x and <y, respectively
  vaddtoxyz u v w - fma >v * {u, v, w} into <x, <y, <z, respectively

  (control flow)
  beq v w n - branch forward n operations if v == w
  bne v w n - branch forward n operations if v != w
  bgt v w n - v > w
  ble v w n - v <= w
  bge v w n - v >= w
  blt v w n - v < w
  end - exit the function

  (randomness)
  randint d - set d to a random integer in [0, RAND_MAX]
  random d - set d to a random real in [0, 1)
  randintr d v w - set d to a random integer in [v, w]
  randomr d v w - set d to a random real in [v, w)

it is guaranteed that for each operation, the assignment to the destination register is the last thing to occur.

particular dot directives include:
  .name <variation name> - set the name of the variation; same rules as for identifiers
  .3d - indicate that the variation writes to FPx
  .dc - indicate that the variation writes to TC
  .vars {int | real | range | cyclic} <name> <init> [, ...] - list variables
  .privs {int | real} <name> [= <init>] [...] - private vars
  .prep - indicate an init() method
  .case <expr> - use the following to the next .case or .calc as the calc() method when expr evaluates to true. a .calc section must exist if any .case sections do.
  .calc - indicate start of calc() method. this is used as the default case; if the interpreter does not understand expressions, it will treat all cases as false and therefore always use .calc.

.name spherivoid // variation name is spherivoid
.3d // 3d variation
.vars real spherivoid_radius 0 // like VAR_REAL(spherivoid_radius, 0.0)
.calc // calculate method
spherical >0 >1 >2 >x >y >z // set registers 0-2 to spherical coordinates of input
add >0 >0 spherivoid_radius // add spherivoid_radius to radius in-place
rect3d >0 >1 >2 >0 >1 >2 // convert spherical coordinates to rectangular in-place
vaddtoxyz >0 >1 >2 // add coordinates scaled by >v to output
end

.name julian
.vars int julian_power 2, real julian_dist 1
.privs int absn, real cn
.prep // init() method
abs absn julian_power
div >0 julian_dist julian_power
set >1 0.5
mul cn >0 >1
end
.calc
atan2 >0 >x >y
set >1 0
randintr >1 >1 absn
set >2 6.283185308 // 2pi
fma >0 >1 >2 >0 // 2pi * randint + theta
div >0 julian_power >0 // (2pi * randint + theta) / power
mul >1 >x >x // sqr(x)
fma >1 >y >y >1 // sqr(x) + sqr(y)
pow >1 >1 cn // pow(sqr(x)+sqr(y), cn)
mul >1 >1 >v // vvar * pow(sqr(x)+sqr(y), cn)
rect >0 >1 >0 >1
addtoxy >0 >1
end