feat: draw block characters with background opacity

[Theaninova]

What kind of change does this PR introduce?

This fixes #2275 by letting box drawing, block elements, geometric shapes as well as the Nerdfont Powerline (+Extras) inherit the background opacity instead.

• (Fix?)
• Feature

Did this PR introduce a breaking change?

A breaking change includes anything that breaks backwards compatibility either at compile or run time.

• No

Main	This PR

[fredizzimo]

Thank you!

I think this is good for a quick workaround before we have a long-term proper solution like this or an alternative:

• [RFC] Proposal: handling of global transparency #2471

But it could also cause problems for other UI elements, there might be some UI element that is supposed to be opaque, like a border, which now becomes transparent with these changes. So, it's a two-edged sword.

But in any case, would you mind, rebasing on top of main, so that we get proper CI checks for this? After that I will discuss this with the other maintainers and decide what to do.

[fredizzimo]

Some of the clippy warnings will be fixed by:

• chore: Fix clippy warnings #2483

But the

RangeInclusive::contains implementation is caused by these changes.

As a sidenote, we need to improve the reporting of the clippy warnings:
they are not very readable currently, but you should get a better message locally if you run cargo clippy

[Theaninova]

Sorry about that, forgot to include rustfmt in my shell.nix

[github-actions]

Test Results

  6 files  ±0    6 suites  ±0   19s ⏱️ -1s
107 tests ±0  107 ✅ ±0  0 💤 ±0  0 ❌ ±0 
630 runs  ±0  630 ✅ ±0  0 💤 ±0  0 ❌ ±0 

Results for commit cfc9b09. ± Comparison against base commit 41622dc.

[Kethku]

This looks great! Thanks for the fix. @fredizzimo this is a small enough change that I'm merging it for now and we can revert it if and when the longer term fix gets implemented. In the meantime, I will do some testing to see how rough the issues you mentioned might be

[falcucci]

@Theaninova first of all, thank you for the contribution!

I had to revert this change in specific due to some issues to render the floating window borders. I'll share an example where the border disappear at top and the bottom:

and this one it's how it should be rendered:

I would be happy to help you with any further information to bring the fix.

[Theaninova]

I was debating in my head if it would be reasonable to include the box drawing set

Box Drawing

0

1

2

3

4

5

6

7

8

9

A

B

C

D

E

F

U+250x

─

━

│

┃

┄

┅

┆

┇

┈

┉

┊

┋

┌

┍

┎

┏

U+251x

┐

┑

┒

┓

└

┕

┖

┗

┘

┙

┚

┛

├

┝

┞

┟

U+252x

┠

┡

┢

┣

┤

┥

┦

┧

┨

┩

┪

┫

┬

┭

┮

┯

U+253x

┰

┱

┲

┳

┴

┵

┶

┷

┸

┹

┺

┻

┼

┽

┾

┿

U+254x

╀

╁

╂

╃

╄

╅

╆

╇

╈

╉

╊

╋

╌

╍

╎

╏

U+255x

═

║

╒

╓

╔

╕

╖

╗

╘

╙

╚

╛

╜

╝

╞

╟

U+256x

╠

╡

╢

╣

╤

╥

╦

╧

╨

╩

╪

╫

╬

╭

╮

╯

U+257x

╰

╱

╲

╳

╴

╵

╶

╷

╸

╹

╺

╻

╼

╽

╾

╿

In hindsight, that was probably a bad idea (who could have guessed). Removing that range should do the trick.

I also just realized that I messed up the ranges when fixing the clippy warnings, all of them should be inclusive

fn is_box_drawing_character(c: char) -> bool {
    // block elements
    ('\u{2580}'..='\u{259F}').contains(&c)
        // geometric shapes
        || ('\u{25A0}'..='\u{25FF}').contains(&c)
        // nerdfont powerline
        || ('\u{e0a0}'..='\u{e0a2}').contains(&c)
        || ('\u{e0b0}'..='\u{e0b3}').contains(&c)
        // nerdfont powerline extras
        || (c == '\u{e0a3}')
        || ('\u{e0b4}'..='\u{e0c8}').contains(&c)
        || (c == '\u{e0ca}')
        || ('\u{e0cc}'..='\u{e0d7}').contains(&c)
}

[Kethku]

This work is pretty related to #2491 which was opened just recently.