[p5.js 2.0 Bug Report]: mouseX and mouseY not updated for touches

[bojidar-bg]

Most appropriate sub-area of p5.js?

• Accessibility
• Color
• Core/Environment/Rendering
• Data
• DOM
• Events
• Image
• IO
• Math
• Typography
• Utilities
• WebGL
• Build process
• Unit testing
• Internationalization
• Friendly errors
• Other (specify if possible)

p5.js version

2.0.0

Web browser and version

Firefox Android 137.0.2 / Chrome Android 135.0.7049.100

Operating system

Android

Steps to reproduce this

Steps:

1. Open a sketch that uses mouseX or mouseY on Android, e.g. an example from https://beta.p5js.org/reference/p5/mouseX/
2. Tap somewhere on the canvas
3. mouseX and mouseY remain 0. Expected: mouseX and mouseY "hold the x/y-coordinate of the most recent touch point" as described in the documentation.

Snippet:

function draw() {
  background(200);
  text(`x: ${mouseX} y: ${mouseY}`, 0, 50);
}

[bojidar-bg]

This behavior seems to stem from the following lines (observe that mouse* is not updated for touch events):

p5.js/src/events/pointer.js

Lines 919 to 933 in e1cc2e0

	if (e.pointerType == 'touch') {
	const touches = [];
	for (const touch of this._activePointers.values()) {
	touches.push(getTouchInfo(canvas, sx, sy, touch));
	}
	this.touches = touches;
	} else {
	const mousePos = getMouseInfo(canvas, sx, sy, e);
	this.movedX = e.movementX || 0;
	this.movedY = e.movementY || 0;
	this.mouseX = mousePos.x;
	this.mouseY = mousePos.y;
	this.winMouseX = mousePos.winX;
	this.winMouseY = mousePos.winY;
	}

Which were introduced by #7378, CC @diyaayay

Technically, this probably should have been caught by this test, but the whole file is marked as suite.todo and thus might not have been executed since:

p5.js/test/unit/events/mouse.js

Line 84 in e1cc2e0

test('mouseX should be current horizontal position of touch relative to the canvas', function() {

[limzykenneth]

The mouse coordinate updates probably can be moved out of the conditional and from my test it works, I'll file a PR towards it. The events unit test I wanted to use a different testing API that is more robust, if I have the time to do it I'll do a proof of concept in the PR otherwise I'll do later.