Cropped cursors from hyprcursor-util

[v4u6h4n]

Hey there, I've been playing around with hyrpcursor-util, converting breeze cursors, and eventually some other major cursor themes, but I've found some of the cursors have cropping. Or there's some other rendering issue that unrelated to the creation.

[vaxerski]

I'd suspect an issue with xcur2png then. Are the png files in the working theme cropped too?

[v4u6h4n]

The cursors are good in the original theme, and in the png files created during the hyprcursor-util extraction, they are only cropped in the final working state. The default cursor, and a lot of other cursors are fine in the working state; it's only cursors that exceed a certain height and/or width that are cropped.

[Gigas002]

Hey, @v4u6h4n!
Can you check out the meta.hl files nearby cursors and compare the px sizes? I suspect you might've hit the same issue I did here: hyprwm/Hyprland#6517.
I've converted my xcursor theme with hyprcursor-util too, and didn't notice that difference in meta.hl files with actual cursor image sizes.

[v4u6h4n]

Hey @Gigas002 yeah sure, I'll have a look, but I can't see a meta.hl file; all I have is a manifest.hl file in the parent directory and a bunch of .hlc files in the hyprcursors subdirectory.

[Gigas002]

These .hlc files are an archives, you can open them and see the cursor png images and meta file

[vaxerski]

to be precise, .hlc is just a .zip

[v4u6h4n]

It appears I didn't know how to use peazip -_-

I don't know if this is the same issue exactly, all the images sizes for the cursor are correct (16, 24, 32, etc), but those images sizes are associated with the wrong define_size in meta.hl; my 24 cursor image is associated with the 32 size.

@vaxerski Is this on the KDE themes end? Or do you think hyprcursor isn't parsing and associating the cursor sizes correctly?

[vaxerski]

likely that hyprcursor-util is extracting them incorrectly from XCursor. X's spec allows for this mismatch and does... some weird stuff there. Anyways, this doesn't seem like a case I want to support. Manual intervention is part of life, util's X -> hc conversion is very approximate anyways. Closing :)