That's a great idea, and we can expand on it by adding an extra class for the clock based on the hour. For example

.clock-widget .icon.clock_23 {}
.clock-widget .label.clock_23 {}

so I can get a red clock at 23h :p

It's actually a good idea that I hadn't even thought of. Now it's done! =D

Cool thanks, can you clean up the commit history a bit before the merge? To make this more readable in the git log.

Oc ! I've squashed the commits into a single one for better readability. Let me know if there's anything else to adjust!

Thanks, looks good to me, but... I just thought, since this widget updates every 1 second, we can slightly optimize it. The idea is: what if we set self.current_hour = None inside __init__ and in the _update_label we have
current_hour = f'{now.hour:02d}'
Then, before doing any processing inside update_label we can compare

if self.current_hour != current_hour:
   self.current_hour = current_hour
   ...

and continue with reloading the CSS and changing icons.

I know this isn’t a big issue, the process is fast anyway, but if we can optimize it without much effort, why not do it?

Thanks for the suggestion! Don't worry, you're absolutely right, it's better to optimize what you can, especially if it's for like 3-4 lines of code.
I've just implemented it in the latest commit.
Now the ClockWidget skips icon and CSS updates when the time hasn't changed, which should avoid unnecessary QLabel refreshes during updates per second.

Let me know if there's anything else you'd like me to tweak or optimize, but I think this is good!

Looks good, thanks. There were some users who used two clock widgets on the same bar, one for time and another one just for date and calendar. And when you have it on 2 screens, this actually results in 4 updates per second. That's why I was thinking it would be better to have a simple check before processing any label update.