feat(cell): add EMPTY and (const) new method

[EdJoPaTo]

This simplifies calls to Buffer::filled in tests.

[codecov]

Codecov Report

Attention: Patch coverage is 98.07692% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 94.2%. Comparing base (4f77910) to head (def27fa).
Report is 3 commits behind head on main.

Files	Patch %	Lines
src/buffer/cell.rs	92.8%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##            main   #1143     +/-   ##
=======================================
- Coverage   94.2%   94.2%   -0.1%     
=======================================
  Files         60      60             
  Lines      14623   14511    -112     
=======================================
- Hits       13780   13672    -108     
+ Misses       843     839      -4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[EdJoPaTo]

A (breaking) follow up to this will be the following performance improvement (removes one needless clone / drop):

     /// Returns a Buffer with all cells initialized with the attributes of the given Cell
     #[must_use]
-    pub fn filled(area: Rect, cell: &Cell) -> Self {
+    pub fn filled(area: Rect, cell: Cell) -> Self {
         let size = area.area() as usize;
-        let content = vec![cell.clone(); size];
+        let content = vec![cell; size];
         Self { area, content }
     }

As this would result in a bunch of conflicts (and is breaking) I'll wait with that until this one is sorted out.

[joshka]

A (breaking) follow up to this will be the following performance improvement (removes one needless clone / drop):

     /// Returns a Buffer with all cells initialized with the attributes of the given Cell
     #[must_use]
-    pub fn filled(area: Rect, cell: &Cell) -> Self {
+    pub fn filled(area: Rect, cell: Cell) -> Self {
         let size = area.area() as usize;
-        let content = vec![cell.clone(); size];
+        let content = vec![cell; size];
         Self { area, content }
     }

+1 to this - looks like a good idea. It's breaking but in a function that's only really useful for testing, so that's not too bad.
https://github.com/search?q=lang%3Arust+ratatui+%22Buffer%3A%3Afilled%22+AND+%28NOT+org%3Aratatui-org%29+AND+%28NOT+is%3Afork%29&type=code

[joshka]

Another idea that springs from this - doing a quick search suggests perhaps a Cell::styled() constructor could be useful

// instead of:
buf.get_mut(x,y)
  .set_symbol("x")
  .set_style(style);
// what about:
buf[(x,y)] = Cell::styled("x", style);

I'm not sure what impact this would have on allocations (likely many) / perf (likely little). Doesn't have to be part of this PR at all, but I'm curious to hear your thoughts on this.

[EdJoPaTo]

Another idea that springs from this - doing a quick search suggests perhaps a Cell::styled() constructor could be useful

// instead of:
buf.get_mut(x,y)
  .set_symbol("x")
  .set_style(style);
// what about:
buf[(x,y)] = Cell::styled("x", style);

I'm not sure what impact this would have on allocations (likely many) / perf (likely little). Doesn't have to be part of this PR at all, but I'm curious to hear your thoughts on this.

set_styled could be a neat idea for simpler code. For performance it would be the same as the two methods set_symbol / set_style.
Using a constructor and overriding it seems weird, this should stick with &mut.

The constructor in this PR is mainly simplifying tests.