Fix: set stroke and fill properly in PFont.getShape

[Junology]

Currently, PFont.getShape returns an instance of PShape which renders nothing.
This is because fill and storke are not properly set in the construction of the shape.
The commit fixes the issue by calling PShape.noStroke() and PShape.fill(0) at the begining of the construction.

Test code:

import java.awt.Font;
import java.util.stream.IntStream;

PFont symbols;
PShape[] symbolShapes;

void setup() {
  size(512, 512);
  
  symbols = createFont("winjs-symbols.ttf", 32);
  Font font = (Font) symbols.getNative();
  final int[] codes = IntStream.range(0, 0x10000).filter(font::canDisplay).toArray();
  symbolShapes = new PShape[codes.length];
  for(int i = 0; i < codes.length; ++i) {
    symbolShapes[i] = symbols.getShape((char) codes[i]);
    symbolShapes[i].width=32;
    symbolShapes[i].height=32;
  }
}

void draw() {
  background(#AABBCC);
  for(int i = 0; i < symbolShapes.length; ++i) {
    int x = i % 16;
    int y = i / 16;
    symbolShapes[i].setFill(#0000FF);
    shape(symbolShapes[i], 32 * x + 4, 32 * y + 36, 24, 24);
  }
}

winjs-symbols.ttf can be downladed here.

[Junology]

Another concern is the "origin" of the shape generated by PFont.getShape.
Namely, since the coordinate (0,0) is on the base line of the character, most of the part of the shape is above the origin, which may be inconsistent with the other shapes which are usually rendered below the origin.
However, the "height" of the character shapes completely depends on their use cases, so "on the base line" might be only a sane choice.
IMHO, PShape can have a functionality to set its origin after the creation (like we can for width and height); cf #788 .

[Junology]

I am really sorry that I mistakenly close the PR.
I didn't notice the PR was made based on main branch of my fork, so I heavily messed it.
I eventually remake this PR.