feat [UI]: Refined ctrl+tab panel

[gordonn]

This PR aims to implement an Arc-like ctrl+tab panel with ability to cycle/scroll through entire tab list.

ctrl-tab.mp4

Windows:

• Enable the panel to cycle/scroll through ordered tab list (including unloaded tabs, excluding hidden tabs) like default Ctrl+Tab behavior
• Ensure panel positioning/UI works consistently across MacOS, Windows and Linux
• Add about:config pref to cycle through recently used workspace tabs like Arc/Firefox/alt-tabbing
• Remove default Firefox Ctrl+Tab panel and add option for this panel in settings

[12th-devs]

@gordonn I figured I would share this pr also being developed. #12383

[CosmoCreeper]

A whole lot of commits for something that could be as simple as a few tweaks to the existing ctrl+tab panel in my opinion. Not sure why there's a whole separate ctrl+tab panel component for these changes but something tells me it's AI's fault (the markdown file in the first commit).

[gordonn]

A whole lot of commits for something that could be as simple as a few tweaks to the existing ctrl+tab panel in my opinion. Not sure why there's a whole separate ctrl+tab panel component for these changes but something tells me it's AI's fault (the markdown file in the first commit).

It's quite different from Firefox's panel implementation and has extra functionality such as scrolling throught entire tab list in order, seeing unloaded tabs, better thumbnail updating and utilizing native panel styling for each OS. Maybe review the most recent files instead of the initial commits which were vibe coded because I just was just prototyping. Since them I have tested it thoroughly and optimized it by hand coding it myself to remove any mistakes done by AI.

[CosmoCreeper]

Okay, no problem, but I just wasn't sure why these couldn't just be improvements to the initial Firefox ctrl+tab panel. Like generic tweaks to the panel that would be beneficial regardless of the browser could be packaged in a merge request to the initial Firefox code, while the styling for Zen could just be applied here.

Only reason I say this is cause the improvements you're mentioning sound like they could at least partially be pushed into Firefox's source.