fillPath consumes all memory when the path has an arc with out of range arguments

[kwhitefoot]

This program runs until killed with out of memory. Of course the arguments to the arc command are far out of range but I would expect a catchable exception instead of an out of memory crash. I suppose it is a similar problem to #403 and #392.

import pixie, chroma
let
  image = newImage(200, 200)
image.fill(rgba(255, 255, 255, 255))

let
  pathStr = """
L 3473901.0 1136732.75 
A 31888.0 31888.0 0 0 1 3493390.25 1076022.375 
L 32563.0 -2081.0"""

var
  paint = newPaint(SolidPaint)

paint.color = color(255, 255, 255, 255)

let
  path = parsePath(pathStr)
  
fillPath(image,
         path,
         paint,
         mat3(),
         NonZero)

Result:

$ ulimit -Sv unlimited; /usr/bin/time -v examples//test_mem6
Command terminated by signal 9
	Command being timed: "examples//test_mem6"
	User time (seconds): 154.55
	System time (seconds): 3.02
	Percent of CPU this job got: 97%
	Elapsed (wall clock) time (h:mm:ss or m:ss): 2:41.19
	Average shared text size (kbytes): 0
	Average unshared data size (kbytes): 0
	Average stack size (kbytes): 0
	Average total size (kbytes): 0
	Maximum resident set size (kbytes): 6226992
	Average resident set size (kbytes): 0
	Major (requiring I/O) page faults: 19
	Minor (reclaiming a frame) page faults: 2272969
	Voluntary context switches: 2120
	Involuntary context switches: 2021
	Swaps: 0
	File system inputs: 4192
	File system outputs: 0
	Socket messages sent: 0
	Socket messages received: 0
	Signals delivered: 0
	Page size (bytes): 4096
	Exit status: 0

[Ryan Oldenburg]

This is another float32 precision issue. Addressed in latest merged PR per the above test. Is it perfect though? Unclear, I'm sure there are adversarial paths out there that can still misbehave, we'll see tho.