(76d10d7) Strike through a relation to a discarded ticket, and bump to 0.31.5

The relations block marked a target `○ ◐ ●` or `?`, and none of those can say
the target was discarded: the status mark is one cell in a column of five
states, and `●` in grey against `●` in green is a colour a pipe does not get.
A row pointing at abandoned work therefore read as a live dependency, which is
exactly what it is not.

The row is struck through instead, the whole of it, gutter and padding
included, so the line runs unbroken from the label to the end of the title.
The relation stays recorded, for completeness and for the case the target is
moved back out of discarded, and reads at a glance as no longer standing for
anything. The mark column is untouched: 451fb83 already gave a discarded
target the full circle of done in the grey of the untouched pile, and the
strike is additive on top of that.

`Cell` learns a modifier, the first style on it that is not a colour. Same
trick as the palette, one step further: `Modifier::ansi()` is SGR 9 and
`From<Modifier> for ratatui::style::Modifier` is `CROSSED_OUT`, so a style
that exists for one renderer and not the other is a compile error. Bold stays
out of it, because bold rides inside `Color::Magenta.ansi()` and `span_from`
adds it back from the colour alone; a second modifier goes beside this one.

`paint_cell` opens the modifier ahead of the colour and closes both with one
reset, which is why the reset is now a constant the two writers share rather
than a literal in `paint`. `span_from` takes the modifier too and has no early
return for the uncoloured case any more: a struck plain run is a styled span
with no foreground. `Cell::struck` is a builder, because the strike is decided
a whole row at a time by the layout that already chose the colours.

`Body::push_cell` merges on the whole style rather than the colour alone.
Nothing outside the relations block carries a modifier today, and a merge
keyed on the colour would silently strike a plain run through the moment one
did.

The strike is under the same flag every other escape is, so a piped `show`
carries none of it. That is the bargain committed and reviewing already take
for being told apart by colour alone, and `assets/SKILL.md` says so outright
rather than leaving an agent to infer that a plain-looking row may point at
abandoned work.

Tests are paired both ways, and both halves are load bearing: a renderer that
struck nothing fails the presence, and one that struck every row fails the
absence. `the_status_marks_are_pinned` had never listed discarded, so it does.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Ansho Rei
2026-08-25 11:53:45 +02:00
co-authored by Claude Opus 5
parent d1a0f0ee27
commit 7994d98913
6 changed files with 359 additions and 32 deletions
+89 -15
View File
@@ -2030,13 +2030,13 @@ fn detail_body(parts: &[BodyPart]) -> Text<'static> {
let mut rest = cell.text.as_str();
while let Some((line, tail)) = rest.split_once('\n') {
if !line.is_empty() {
open.push(span_from(line, cell.color));
open.push(span_from(line, cell.color, cell.modifier));
}
lines.push(Line::from(std::mem::take(&mut open)));
rest = tail;
}
if !rest.is_empty() {
open.push(span_from(rest, cell.color));
open.push(span_from(rest, cell.color, cell.modifier));
}
}
}
@@ -2058,28 +2058,45 @@ fn detail_body(parts: &[BodyPart]) -> Text<'static> {
Text::from(lines)
}
/// A cell as the board draws it: the same text in the same colour, as a span.
/// A cell as the board draws it: the same text in the same style, as a span.
///
/// The other half of `paint_cells`, which is the CLI's. `From<Color>` is what makes the
/// two agree about what the colour is.
/// The other half of `paint_cells`, which is the CLI's. `From<Color>` and
/// `From<crate::util::Modifier>` are what make the two agree about what the style is.
fn span_of(cell: Cell) -> Span<'static> {
span_from(cell.text, cell.color)
span_from(cell.text, cell.color, cell.modifier)
}
/// The same, for a renderer holding the two halves of a cell rather than the cell.
/// The same, for a renderer holding the parts of a cell rather than the cell.
///
/// The bold on magenta is not a flourish. `Color::Magenta.ansi()` is `\x1b[1;35m`, so
/// the CLI draws every bold run bold as well as magenta, and a board that dropped the
/// modifier would be painting the same cell differently from the CLI. The colour is the
/// only thing `From<Color>` can carry, so the attribute has to be added on this side.
fn span_from(text: impl Into<String>, color: Option<Color>) -> Span<'static> {
let Some(color) = color else {
return Span::raw(text.into());
};
///
/// A cell's own modifier is the other one, and it arrives on cells with no colour too, so
/// there is no early return for the uncoloured case any more: a struck plain run is a
/// styled span with no foreground rather than a raw one. `crate::util::Modifier` is spelled
/// out because `Modifier` in this file is ratatui's, which is what the style asks for.
fn span_from(
text: impl Into<String>,
color: Option<Color>,
modifier: Option<crate::util::Modifier>,
) -> Span<'static> {
let mut style = Style::default();
let mut style = Style::default().fg(color.into());
if color == Color::Magenta {
style = style.add_modifier(Modifier::BOLD);
if let Some(color) = color {
style = style.fg(color.into());
if color == Color::Magenta {
style = style.add_modifier(Modifier::BOLD);
}
}
if let Some(modifier) = modifier {
style = style.add_modifier(modifier.into());
}
if style == Style::default() {
return Span::raw(text.into());
}
Span::styled(text.into(), style)
@@ -6952,6 +6969,12 @@ mod tests {
/// `write_ticket`'s trailer: the block has to be on screen for a test to read the
/// colour off the drawn buffer.
fn related_board() -> Fixture {
related_board_targeting("reviewing")
}
/// The same, with the target in a status of the caller's choosing: how the row is
/// drawn depends on it, so a test about a discarded target needs one there.
fn related_board_targeting(status: &str) -> Fixture {
let dir = board_dir();
let holder = "11111111-0000-4000-8000-000000000000";
let target = "22222222-0000-4000-8000-000000000000";
@@ -6965,7 +6988,7 @@ mod tests {
),
)
.unwrap();
write_ticket(&dir, "reviewing", target, "The target", "");
write_ticket(&dir, status, target, "The target", "");
dir
}
@@ -7025,6 +7048,57 @@ mod tests {
);
}
/// The strike reaches the board as a ratatui modifier, not as `\x1b[9m` in the
/// paragraph, which is what a `Paragraph` would draw as literal bytes. `Cell` carries
/// the modifier for the same reason it carries the colour: one layout, two painters.
///
/// Paired the way the guideline demands, and both halves are load bearing. A renderer
/// that struck nothing would pass the absence, and one that went back to painting
/// strings would pass the presence.
#[test]
fn a_discarded_target_is_struck_in_the_overlay_and_no_escape_reached_it() {
let dir = related_board_targeting(DISCARDED);
let mut board = one_board(&dir);
board.toggle_detail();
let detail = board.detail.as_ref().expect("no detail");
let row = body_line(detail, "blocked by");
for span in &row.spans {
assert!(
span.style.add_modifier.contains(Modifier::CROSSED_OUT),
"{:?} is not struck, so the line breaks mid row: {row:?}",
span.content
);
}
// The pair: the strike arrived as a style, and no escape arrived with it.
assert!(!row.spans.is_empty(), "the row has no spans, so the loop above is vacuous");
assert!(
!detail.plain().contains('\x1b'),
"escapes reached the body: {:?}",
detail.plain()
);
}
/// The other side of it, so the assertion above cannot be satisfied by a board that
/// strikes every relation row through: a live target is drawn as it always was.
#[test]
fn a_live_target_is_not_struck_in_the_overlay() {
let dir = related_board();
let mut board = one_board(&dir);
board.toggle_detail();
let detail = board.detail.as_ref().expect("no detail");
let row = body_line(detail, "blocked by");
for span in &row.spans {
assert!(
!span.style.add_modifier.contains(Modifier::CROSSED_OUT),
"{:?} is struck through: {row:?}",
span.content
);
}
}
/// The body is no longer one string, so the clamp and the draw could measure two
/// different things. They take the same `Text`, and this is what says so: a ticket
/// whose relations run past the overlay scrolls to its last row.
+257 -14
View File
@@ -245,13 +245,55 @@ impl From<Color> for ratatui::style::Color {
}
}
/// The escape that closes a run, whatever was opened: a colour, a modifier or both.
///
/// One spelling for the two writers. `paint` opens a colour and `paint_cell` may open a
/// modifier ahead of it, and a reset that only knew about the colour would leave the
/// strikethrough running to the end of the line.
pub const RESET: &str = "\x1b[0m";
/// A style that is not a colour, which is the other half of what a [`Cell`] carries.
///
/// An enum with one variant rather than a `bool`, for the reason the palette is an enum:
/// the CLI needs an SGR sequence and the board needs a `ratatui` modifier, so a style
/// that exists for one renderer and not the other is a compile error. Bold is not here,
/// because bold rides inside `Color::Magenta.ansi()` and is added back on the board's
/// side from the colour alone; a second modifier would be added here beside this one.
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum Modifier {
/// What a row about a discarded ticket is drawn with: the relation is still
/// recorded, and no longer stands for anything.
Strikethrough,
}
impl Modifier {
/// The SGR sequence the CLI writes. Closed by [`RESET`], like a colour.
pub fn ansi(self) -> &'static str {
match self {
Modifier::Strikethrough => "\x1b[9m",
}
}
}
/// The same modifier as ratatui sees it, the other half of [`Modifier::ansi`].
///
/// SGR 9 and `CROSSED_OUT` are the same request, so the two renderers cannot come to
/// draw one struck row and one plain one.
impl From<Modifier> for ratatui::style::Modifier {
fn from(modifier: Modifier) -> Self {
match modifier {
Modifier::Strikethrough => ratatui::style::Modifier::CROSSED_OUT,
}
}
}
/// `color_text` with the decision passed in rather than asked for.
///
/// The seam matters for anything that aligns columns: a caller has to be able to
/// render both ways in one process to prove the padding does not depend on it.
pub fn paint(text: &str, color: Color, enabled: bool) -> String {
if enabled {
format!("{}{}{}", color.ansi(), text, "\x1b[0m")
format!("{}{}{}", color.ansi(), text, RESET)
} else {
text.to_string()
}
@@ -722,11 +764,15 @@ impl Body {
self.push_cell(Cell::plain(c.to_string()));
}
/// One run, merged into the trailing cell when it is the same colour.
/// One run, merged into the trailing cell when it is drawn the same way.
///
/// Merging is not an optimisation. The board draws a span per cell, and the body is
/// pushed a character at a time in places, so without this a plain line arrives as
/// forty single-character spans for ratatui to wrap.
///
/// On the whole style and not the colour alone: nothing outside the relations block
/// carries a modifier today, and a merge keyed on the colour would silently strike a
/// plain run through the moment one did.
fn push_cell(&mut self, cell: Cell) {
if cell.text.is_empty() {
return;
@@ -734,7 +780,9 @@ impl Body {
match self.parts.last_mut() {
Some(BodyPart::Cells(cells)) => match cells.last_mut() {
Some(last) if last.color == cell.color => last.text.push_str(&cell.text),
Some(last) if last.color == cell.color && last.modifier == cell.modifier => {
last.text.push_str(&cell.text)
}
_ => cells.push(cell),
},
_ => self.parts.push(BodyPart::Cells(vec![cell])),
@@ -2268,13 +2316,38 @@ pub fn reference_cells(uuid: &str, name: Option<&str>) -> Vec<Cell> {
/// Cells as the CLI prints them: each one painted, or carried through as it stands.
pub fn paint_cells(cells: &[Cell], colored: bool) -> String {
cells
.iter()
.map(|cell| match cell.color {
Some(color) => paint(&cell.text, color, colored),
None => cell.text.clone(),
})
.collect()
cells.iter().map(|cell| paint_cell(cell, colored)).collect()
}
/// One cell as an escape sequence, its modifier ahead of its colour and one [`RESET`]
/// closing both.
///
/// The modifier is under the same flag the colour is, so a piped `show` and the board's
/// own unpainted pass carry no escape at all, per the guideline: a function that takes
/// an `enabled` flag has to use it all the way down. The cost is that a pipe cannot see
/// a struck row, the same bargain committed and reviewing already take for being told
/// apart by colour alone.
fn paint_cell(cell: &Cell, colored: bool) -> String {
if !colored {
return cell.text.clone();
}
// Built rather than matched over the four combinations: either half may be absent,
// and a cell with neither has to come back byte for byte as it went in, since the
// CLI's output is pinned against what it printed before the cells conversion.
let mut opened = String::new();
if let Some(modifier) = cell.modifier {
opened.push_str(modifier.ansi());
}
if let Some(color) = cell.color {
opened.push_str(color.ansi());
}
if opened.is_empty() {
return cell.text.clone();
}
format!("{opened}{}{RESET}", cell.text)
}
/// The mark a blocking relation carries, and a card waiting on one.
@@ -2395,19 +2468,39 @@ pub struct RelationTarget {
pub struct Cell {
pub text: String,
pub color: Option<Color>,
/// The style that is not the colour, `None` for text drawn as it stands. Kept
/// beside the colour rather than folded into it: a struck run keeps whatever colour
/// its column already gave it, so the two are independent.
pub modifier: Option<Modifier>,
}
impl Cell {
pub fn plain(text: impl Into<String>) -> Self {
Cell { text: text.into(), color: None }
Cell {
text: text.into(),
color: None,
modifier: None,
}
}
pub(crate) fn colored(text: impl Into<String>, color: Color) -> Self {
Cell {
text: text.into(),
color: Some(color),
modifier: None,
}
}
/// The same cell, struck through.
///
/// A builder rather than a third constructor, because the strike is decided a whole
/// row at a time and the cells of that row are built by the layout that already
/// chose their colours. `relation_cells` maps this over the row it has just built,
/// so nothing has to thread the decision through every cell it makes.
pub(crate) fn struck(mut self) -> Self {
self.modifier = Some(Modifier::Strikethrough);
self
}
}
/// The relations of one ticket, both directions, in the order they are shown.
@@ -2468,8 +2561,8 @@ pub fn relation_cells(rows: &[RelationRow]) -> Vec<Vec<Cell>> {
rows.iter()
.map(|row| {
let (kind, kind_colour) = relation_mark(row.kind);
let (status, status_colour) =
status_mark(row.target.as_ref().map(|target| target.status.as_str()));
let status_of = row.target.as_ref().map(|target| target.status.as_str());
let (status, status_colour) = status_mark(status_of);
let padding = " ".repeat(width - row.label.chars().count());
// Kind, then state, then who: what sort of edge this is, how far the ticket
@@ -2489,7 +2582,16 @@ pub fn relation_cells(rows: &[RelationRow]) -> Vec<Vec<Cell>> {
row.target.as_ref().map(|target| target.title.as_str()),
));
cells
// The whole row, gutter and padding included, so the line runs unbroken from
// the label to the end of the title. The strike is what says the target was
// discarded, since the status mark cannot: it is one cell in a column of five
// states, and a row pointing at abandoned work otherwise reads as a live
// dependency, which is exactly what it is not. The relation stays recorded,
// for completeness and for the case the target is moved back out.
match status_of {
Some(DISCARDED) => cells.into_iter().map(Cell::struck).collect(),
_ => cells,
}
})
.collect()
}
@@ -5236,9 +5338,150 @@ Plain line with **bold** and a [link](http://x.test/a).
assert_eq!(status_mark(Some("committed")), ("\u{25d0}", Color::Orange));
assert_eq!(status_mark(Some("reviewing")), ("\u{25d0}", Color::Cyan));
assert_eq!(status_mark(Some("done")), ("\u{25cf}", Color::Green));
assert_eq!(status_mark(Some("discarded")), ("\u{25cf}", Color::Grey));
assert_eq!(status_mark(None), ("?", Color::Grey));
}
/// The strike is what says the target will never be done, because the status mark
/// cannot: it is one cell in a column of five states, and a row pointing at abandoned
/// work otherwise reads as a live dependency, which is exactly what it is not.
///
/// On every cell of the row, gutter and padding included, so the line runs unbroken
/// from the label to the end of the title. A cell left out would draw a gap.
#[test]
fn a_row_pointing_at_a_discarded_target_is_struck_through() {
let rows = vec![row(
RelationKind::BlockedBy,
"blocked by",
TARGET,
"Dropped",
DISCARDED,
)];
let cells = relation_cells(&rows);
assert_eq!(cells.len(), 1, "{cells:?}");
for cell in &cells[0] {
assert_eq!(
cell.modifier,
Some(Modifier::Strikethrough),
"{:?} is not struck, so the line breaks mid row",
cell.text
);
}
// Every run opened, since a cell is closed by its own reset: the CLI writes one
// `\x1b[9m` per cell or the strike stops at the first colour change.
let rendered = render_relations(&rows, true);
assert_eq!(
rendered.matches(Modifier::Strikethrough.ansi()).count(),
cells[0].len(),
"the CLI did not open the strike on every run:\n{rendered:?}"
);
}
/// Only discarded. A row to a live target is drawn as it always was, done included:
/// the strike says the relation no longer stands for anything, and a blocker that
/// landed still stands for the fact that it was one.
///
/// Over `status_dirs` rather than the four names, so a status added later is decided
/// here rather than inheriting the strike or the plain row by accident.
#[test]
fn a_row_to_a_live_target_is_not_struck() {
let live = status_dirs().into_iter().filter(|status| *status != DISCARDED);
for status in live {
let rows = vec![row(
RelationKind::BlockedBy,
"blocked by",
TARGET,
"Alive",
status,
)];
for cell in relation_cells(&rows).concat() {
assert_eq!(cell.modifier, None, "{status} is struck: {:?}", cell.text);
}
assert!(
!render_relations(&rows, true).contains(Modifier::Strikethrough.ansi()),
"{status} is struck through in show"
);
}
}
/// A target the board no longer holds is not struck either. `?` and `_missing_` say
/// the row cannot be resolved, which is a different fact from a target that was
/// resolved and abandoned, and striking both would fold the two together.
#[test]
fn a_row_to_a_missing_target_is_not_struck() {
let rows = vec![RelationRow {
kind: RelationKind::BlockedBy,
label: "blocked by",
uuid: TARGET.to_string(),
target: None,
}];
for cell in relation_cells(&rows).concat() {
assert_eq!(cell.modifier, None, "{:?}", cell.text);
}
}
/// Paired the way the guideline demands, and both halves are load bearing. The strike
/// is an escape sequence like a colour, so it is under the same flag: a piped `show`
/// carries none of it, the same bargain committed and reviewing already take for
/// being told apart by colour alone. A renderer that never struck anything would pass
/// the absence on its own.
#[test]
fn the_strike_is_under_the_colour_flag_like_every_other_escape() {
let rows = vec![row(
RelationKind::BlockedBy,
"blocked by",
TARGET,
"Dropped",
DISCARDED,
)];
assert!(
render_relations(&rows, true).contains(Modifier::Strikethrough.ansi()),
"nothing was struck at all, so the guard below is vacuous"
);
assert!(
!render_relations(&rows, false).contains('\x1b'),
"an escape survived the colour being off: {:?}",
render_relations(&rows, false)
);
}
/// Both halves of the modifier, the way the palette is pinned from both ends. SGR 9
/// and `CROSSED_OUT` are the same request, so one renderer cannot come to draw a
/// struck row plain while the other strikes it.
#[test]
fn the_strikethrough_is_the_same_request_in_both_renderers() {
assert_eq!(Modifier::Strikethrough.ansi(), "\x1b[9m");
assert_eq!(
ratatui::style::Modifier::from(Modifier::Strikethrough),
ratatui::style::Modifier::CROSSED_OUT
);
}
/// The modifier opens ahead of the colour and one reset closes both, which is why the
/// reset is a constant the two writers share. A cell carrying neither comes back byte
/// for byte as it went in, since the CLI's output is pinned against what it printed
/// before the cells conversion.
#[test]
fn a_cell_opens_its_modifier_ahead_of_its_colour_and_closes_once() {
let struck_and_coloured = Cell::colored("x", Color::Red).struck();
assert_eq!(
paint_cells(&[struck_and_coloured], true),
format!("\x1b[9m\x1b[31mx{RESET}")
);
let struck_only = Cell::plain("x").struck();
assert_eq!(paint_cells(&[struck_only], true), format!("\x1b[9mx{RESET}"));
assert_eq!(paint_cells(&[Cell::plain("x")], true), "x");
assert_eq!(paint_cells(&[Cell::plain("x").struck()], false), "x");
}
/// Every status the board holds is told apart from the other three. The shape does
/// not do it alone, since committed and reviewing are both half circles, so it is the
/// pair that has to be unique. Pinned against `status_dirs` rather than the four