Mouse click accuracy

[obivan]

Hyprland Version

System/Version info

Hyprland, built from branch main at commit a66cfe0fbed7fb4dc69383e107c2bf3b1e7cd80a  ().
Date: 2024-05-16
Tag: , commits: 

flags: (if any)


System Information:
System name: Linux
Node name: nixos-pc
Release: 6.9.0
Version: #1-NixOS SMP PREEMPT_DYNAMIC Sun May 12 21:12:29 UTC 2024


GPU information: 
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation AD102 [GeForce RTX 4090] [10de:2684] (rev a1) (prog-if 00 [VGA controller])


os-release: ANSI_COLOR="1;34"
BUG_REPORT_URL="https://github.com/NixOS/nixpkgs/issues"
BUILD_ID="24.05.20240515.33d1e75"
DOCUMENTATION_URL="https://nixos.org/learn.html"
HOME_URL="https://nixos.org/"
ID=nixos
IMAGE_ID=""
IMAGE_VERSION=""
LOGO="nix-snowflake"
NAME=NixOS
PRETTY_NAME="NixOS 24.05 (Uakari)"
SUPPORT_URL="https://nixos.org/community.html"
VERSION="24.05 (Uakari)"
VERSION_CODENAME=uakari
VERSION_ID="24.05"


plugins:

Bug or Regression?

Bug

Description

When selecting text or clicking on links/buttons, the click is registered a few pixels to the right of the actual cursor position. Similar to old issue #4691, this only seems to happen when the cursor shape is changed.

There is also a strange jump to the right when changing the cursor shape (see video).

This happens starting from commit ed411f5

I'm using xcursor. Enabling or disabling hardware cursors using the WLR_NO_HARDWARE_CURSORS environment variable and/or the cursor:no_hardware_cursors option has no visible effect whatsoever.

How to reproduce

Move the cursor to the boundary of the link.

Crash reports, logs, images, videos

issue.mp4

normal.mp4

[vaxerski]

doesn't happen to me so I'll need more info (does it happen if you disable scaling? if you use hyprcursor? if you change the cursor size?)

[obivan]

does it happen if you disable scaling?

Nope, with scaling 1 everything works fine. I normally use scaling 2.

if you use hyprcursor?

It takes me a while to figure out how to do it right. I guess I'll try converting the theme I'm using to hyprcursor format and set the corresponding environment variables.

if you change the cursor size?

I've tried two cursor themes (breeze_cursors from KDE and phinger-cursors-light) in two size options: 24 and 32. The issue persists.

Oh, one more thing - it looks like the issue is not reproducible inside XWayland apps.

[vaxerski]

likely a math error with xcursor hotspots somewhere then

[obivan]

I have checked and can reproduce this with hyprcursor as well.

[obivan]

Well, when I change this code (in main)

CBox CPointerManager::getCursorBoxGlobal() {
    return CBox{pointerPos, currentCursorImage.size / currentCursorImage.scale}.translate(-currentCursorImage.hotspot / currentCursorImage.scale);
}

to this one

CBox CPointerManager::getCursorBoxGlobal() {
    return CBox{pointerPos, currentCursorImage.size / currentCursorImage.scale}.translate(-currentCursorImage.hotspot);
}

it fixes the issue.

I don't know any shit about how valid this is.
But it works

[vaxerski]

does it happen if you disable hardware cursors...?

[obivan]

I see no change in behavior from enabling/disabling hardware cursors.

I mean, if I apply the nano-patch above, the issue disappears regardless of the kind of cursors used.

[vaxerski]

ok then

[obivan]

Thank you! In the latest main (553232a3e4c112c8511c8511309e6b685cb614895e714) the issue is still reproducible, but only if hardware cursors are enabled. I don't know why, maybe because of 0d6eae0, which was not present at the time of initial testing.

[vaxerski]

bing chilling

[vaxerski]

patch.txt
try

[obivan]

Nope, still doesn't work with hardware cursors enabled (and works with software cursors on)

[obivan]

This change fixes the hardware cursors for me (while not breaking the software cursors). Found it tinkering around your patch.
patch2.txt

[vaxerski]

thanks, can you make a MR?

[obivan]

Sure. Please take a look.