Hey @zep, in case you see this:

If double pixel mode (0x5f2c) and devkit more are turned on, stat(32) and stat(33) values are correctly halved when running in PICO-8, so the mouse treats the screen as having a 64x64 resolution.

In the web player, however, pixel values are not halved and stat(32)/stat(33) treat the screen as 128x128.

My solution is to just manually divide stat(32) and stat(33) by 2 if stat(102)!=0, but this still seems like a bug.

See https://www.lexaloffle.com/bbs/?tid=31596

Tobiasvl:

• I don't think this is a bug.

Remember when you turn on double-pixel mode, all you are doing is changing the way pixels appear.

It does not actually take your current image and double it in size. If you read pixels, you will see while you can read 0-63, only 0-63 are available.

Now I suspect you can draw off-screen in this mode too (>=64) and as long as you are viewing double-pixel size, the data will retain.

This is also proof that reading from the mouse is unaffected by the "visual" change. Because that's all that's doubling is what you see - not actual recorded pixel data.

I think it is a bug since reading the mouse position is affected by double-pixel mode when running PICO-8 locally, but not in the web player.

The inconsistent behavior is the reason I consider it a bug. Sorry if that was unclear.