Improve find cursor

[ChangeCaps]

This pr does two things. When load_cursor now splits the $XCURSOR_PATH variable on both ':' and ' ', on my system $XCURSOR_PATH is separated by spaces. Second, cursor::Handle::new now checks $XCURSOR_THEME if getting Xcursor.theme from the resource_database fails.

[psychon]

First: Due to #934 asking for all of this code to be replaced by the xcursor crate, I don't feel like we should spend too much time on this here. Sorry.

When load_cursor now splits the $XCURSOR_PATH variable on both ':' and ' ', on my system $XCURSOR_PATH is separated by spaces

Why is it that way on your system? I am pretty sure that xcursor-rs also only splits on :. But that's just another random Rust crate. IMO the "source of truth" is libXcursor. I find its source code always hard to navigate, but I am pretty sure that it also just splits on :.

So, does your space-separates $XCURSOR_PATH actually work with anything? Did I miss something?

Second, cursor::Handle::new now checks $XCURSOR_THEME if getting Xcursor.theme from the resource_database.

That seems to be done by libXcursor (it prefers the env var over the resource database), but does not seem to be done by xcursor-rs.

[ChangeCaps]

Why is it that way on your system? I am pretty sure that xcursor-rs also only splits on :. But that's just another random Rust crate. IMO the "source of truth" is libXcursor. I find its source code always hard to navigate, but I am pretty sure that it also just splits on :.

So, does your space-separates $XCURSOR_PATH actually work with anything? Did I miss something?

You didn't, it was a brain-fart on my part... I ran echo $XCURSOR_PATH which apparently replaces semicolons with spaces in the output. I have reverted the change.

Second, cursor::Handle::new now checks $XCURSOR_THEME if getting Xcursor.theme from the resource_database.

That seems to be done by libXcursor (it prefers the env var over the resource database), but does not seem to be done by xcursor-rs.

As of 070e2bb $XCURSOR_THEME is now tried first.

[psychon]

Could you squash this into a single commit and perhaps rebase onto master? Thanks.

#934 is now dealt with and the code you are touching is still there. Sorry, previously I thought that this code would go away with that change.

[ChangeCaps]

Could you squash this into a single commit and perhaps rebase onto master? Thanks.

All done!

[psychon]

Thanks!