makeover_webview/lib.rs
1//! The webview renderer for [`makeover_layout`].
2//!
3//! <!-- wiki: makeover-webview -->
4//!
5//! # The renderer that needs no palette
6//!
7//! `makeover-immediate` and `makeover-tui` both take a `Palette`, because egui
8//! and a terminal need an actual colour before they can put anything on
9//! screen. A webview does not: `var(--surface-raised)` *is* the late binding,
10//! and the browser resolves it against whatever `themes.js` last wrote onto
11//! `:root`.
12//!
13//! So this crate emits text naming intents, and never learns a colour. It is
14//! the deferral rule with no adapter in the way, and it is why the webview was
15//! always the wrong renderer to derive a vocabulary from: it can express
16//! anything, so it never pushes back.
17//!
18//! # Phase A: the stylesheet
19//!
20//! This module emits component CSS and no markup, deliberately. GoingsOn has
21//! 145 `innerHTML` sites and Balanced Breakfast 175 `createElement` sites, so
22//! moving markup is a migration where adopting a generated stylesheet is not.
23//! The apps keep every line of their markup and gain the classes.
24//!
25//! It is not a deletion either, which this header claimed until the measurement
26//! came in. Adoption across goingson removed 49 declarations net and *added* 25
27//! lines: a rule loses its depth declarations and gains a variant selector next
28//! to it, so the file stays the same size. What phase A moves is where depth is
29//! defined, not how much CSS exists. Numbers and method in the wiki note under
30//! "The deletion test, run".
31//!
32//! The bevel properties are byte-identical to what both apps already
33//! hand-write, which is asserted below.
34//!
35//! # Phase B: the markup, one description at a time
36//!
37//! [`form`] renders [`makeover_layout::Field`], which is the half of phase B
38//! whose description is settled. It emits strings, because both apps
39//! interpolate their fields into larger string-built forms and returning nodes
40//! would rewrite those too. It owns its own escaping, on the reasoning in that
41//! module: a Rust encoder can cover element text and attribute values with one
42//! function, where the apps need four and have to choose correctly at every
43//! call site.
44//!
45//! [`facet`] renders `makeover_layout::Facet`: a dimension a set is narrowed by,
46//! and the one phase-B emitter whose markup an app is not keeping, because the
47//! markup it replaces was two mechanisms rather than one. A tag's selection and
48//! a tag's browse position were separate state on MNW's discover page, which is
49//! why every filter row there carries a tick box *and* a chevron; one gesture
50//! doing both is what lets the second one go.
51//!
52//! [`list`] is the other half: column tracks, the narrowing rules, and the cell
53//! containers a row is made of. It stops at the cell boundary and does not
54//! render what goes inside one, on the reasoning in that module. So phase B is
55//! now the frame around content in both directions, and what an app still owns
56//! is the content itself.
57//!
58//! # What phase A settled, and what it costs
59//!
60//! Decided 2026-07-29 against goingson's `styles.css` rather than against a
61//! component list. The useful finding there was that `.btn` (line 644),
62//! `.card` (768) and `.tag, .badge` (882) each hand-write the same
63//! composition, so three quarters of phase A is one rule with several names.
64//!
65//! Two of the four decisions change how goingson looks, and adoption should
66//! not be described as a pure deletion:
67//!
68//! - **Pressed carries its fill.** [`interactive_rules`] emits
69//! [`Depth::pressed`] whole. goingson presses to `--surface-sunken` today and
70//! will press to `--surface-well`, and hovers to `--surface-overlay` today
71//! and will hover to `--hover-surface`. Since `surface-well` inverts by theme
72//! where `surface-sunken` does not, a dark theme presses *lighter* than it
73//! hovers. That falls out of `makeover`'s own derivation, which says outright
74//! that `surface-sunken` cannot serve as a well, so if it reads wrong the
75//! answer is there and not here.
76//! - **Badges go flat.** See [`token_rules`].
77//!
78//! The other two: the progress trough is renderer-local and the scrollbar
79//! track was dropped ([`component_rules`]), and no class prefix ships by
80//! default, so adoption means deleting the app's hand-written rule in the same
81//! commit that adds the generated one. `.card`, `.badge` and the tab classes
82//! all already exist in goingson, and while both rules exist the cascade order
83//! decides which wins. That is the one real risk in adopting this, and it is
84//! why the migration lands per component rather than in one commit.
85//!
86//! # 0.10.0: the states this crate used to leave to its consumers
87//!
88//! [`interactive_rules`] emitted hover and pressed and stopped, because
89//! `makeover-layout` modelled no interaction state. Focus and disabled were
90//! therefore unsayable, and every app completed the primitive from outside the
91//! only way that works: by out-specifying a rule it does not own. goingson
92//! carries 19 such rules and the MNW server 21, and the three focus rings do
93//! not match each other.
94//!
95//! That also blocked the cascade-layer work outright. An app that declares
96//! `@layer` puts its own rules in a named layer, and unlayered declarations
97//! outrank every named layer regardless of specificity, so all of those
98//! overrides lose in the commit that adopts layers. They cannot simply be
99//! deleted, because they are the only thing supplying the missing states.
100//! Emitting the states here is what turns that adoption into a deletion.
101//!
102//! Four states now, in emission order, and the order is load-bearing: they are
103//! all specificity (0,2,0), so disabled beats hover by coming last and by
104//! nothing else. Nothing here reaches for `:not(:disabled)`, which would raise
105//! a selector this crate will shortly be wrapping in its own layer.
106//!
107//! Hover additionally sits inside a capability query now. `makeover-touch`
108//! answers whether a fingertip has hover and `makeover-geometry` spells the
109//! condition; this crate asks and does not decide. goingson's section 60 exists
110//! solely to take the hover state back on touch, which is a fight it should
111//! never have been handed.
112//!
113//! # 0.11.0: the layer contract
114//!
115//! [`stylesheet`] emits into the `makeover` cascade layer ([`CSS_LAYER`], which
116//! lives in `makeover-geometry` because that is the one crate every CSS emitter
117//! in the family already depends on). `makeover-geometry` 0.6.0 does the same
118//! for `geometry.css`.
119//!
120//! The cascade resolves origin and importance, then layer, then specificity,
121//! then source order, and **unlayered normal declarations outrank every named
122//! layer**. So before this, an app that declared `@layer base, components,
123//! responsive` put every rule it owns into a named layer and lost all of them to
124//! this unlayered file, regardless of specificity and regardless of loading
125//! last. Nothing errors when that happens: the CSS is valid, the minifier is
126//! happy, and buttons and badges look subtly wrong.
127//!
128//! That is why the layer belongs here rather than in each app. An app cannot fix
129//! it from its own stylesheet, because the fix is to layer the file it does not
130//! own.
131//!
132//! **What it flips**, and the reason each app wants a look when it bumps the
133//! pin: a generated rule that currently beats an app rule by being more specific
134//! stops beating it. The direction is always "the app wins", which is what the
135//! apps already assume, but a hand-written rule an app thought was dead can come
136//! back to life.
137//!
138//! An app should declare the order once, or the layer's position is decided by
139//! whichever generated file the browser happens to see first:
140//!
141//! ```css
142//! @layer makeover, base, components, responsive;
143//! ```
144//!
145//! [`in_css_layer`] is re-exported for an app that assembles its own stylesheet
146//! from this crate's pieces. goingson builds `tables.css` in its own `build.rs`
147//! out of [`list::narrowing_css`] and [`list::grid_template_columns`], and those
148//! rules are as generated as the ones here, so they belong in the same layer and
149//! this crate cannot put them there on the app's behalf.
150//!
151//! # 0.12.0: the ring gets its own width
152//!
153//! [`focus_rule`] reused [`Emit::border_width`] and emitted a 1px ring. That was
154//! an implementation convenience dressed as consistency with the invalid-field
155//! ring: a bevel and a focus indicator answer different questions, and only one
156//! of them has to be noticed from across a desk.
157//!
158//! Caught while adopting 0.11.0 into goingson, by the check the adoption tasks
159//! ask for. Every consumer had already written its own ring and all three chose
160//! at least 2px: the MNW server 2px across 10 rules, Balanced Breakfast 2px,
161//! goingson 2px on three rules and 3px on the one covering twelve selectors. The
162//! design system was the only thing in the tree saying 1px, so deleting the app
163//! rules in favour of it would have thinned the focus indicator everywhere.
164//!
165//! [`Emit::focus_width`] now carries it, defaulting to `2px`, and the offset is
166//! the same magnitude with its sign off the depth. Both values are the measured
167//! consensus rather than a new opinion.
168//!
169//! # 0.47.0: a range, a chooser's ghost text, and an option that cannot be
170//! picked yet
171//!
172//! `makeover-layout` 0.28.0's three form findings, all of them cheap here and
173//! none of them cheap in the app that found them.
174//!
175//! - `FieldKind::Range` emits `<input type="range">`, and `Field::step` emits
176//! `step`. The step is emitted only when the description carries one: the
177//! browser's own default is `step="1"`, which is what a description means by
178//! saying nothing, and is also what turns a 0-to-1 threshold into a
179//! two-position control.
180//! - A select with nothing chosen emits a disabled, selected, valueless first
181//! option carrying `Field::placeholder`. HTML has no placeholder attribute on
182//! `<select>`; this is the idiom, and `required` keeps working through it
183//! because the option's value is empty.
184//! - `Choice::unavailable` emits `disabled` plus the reason. Where it goes
185//! differs by control and the difference is forced: a radio group gets a
186//! `.form-option-reason` span beside the label, and a `<select>` option has
187//! room for no element at all, so the reason runs into its text.
188//!
189//! # 0.25.0: a cell says what it holds
190//!
191//! 0.23.0 gave a table its layout and left every cell the same. One `.cell`
192//! carried the whole thing, so a cell holding text and a cell holding a button
193//! were one class and one content colour, and a control in a cell was painted
194//! as text. That is the drift [`RowPart::intent`](makeover_layout::RowPart)
195//! has prevented for list rows since 0.2.0 and prevented for nothing here.
196//!
197//! makeover-layout 0.14.0's [`CellPart`](makeover_layout::CellPart) names the
198//! four things a cell holds, and [`table_rules`] turns them into
199//! `.cell-value`, `.cell-tokens`, `.cell-actions` and `.cell-link`. Only the
200//! first takes a colour: a token carries its own tone, an action is a control
201//! rather than text, and a link takes the action colour from the anchor it is.
202//!
203//! The colour going on `.cell-value` rather than on `.cell` is the fix rather
204//! than an implementation detail. On the container it cascades into the parts
205//! that are not text, which is the bug said in one rule.
206//!
207//! [`list::Cell::part`] is `Option<CellPart>` here, where it was
208//! `Option<RowPart>`. A table cell borrowing the list row's vocabulary was the
209//! drift with a type on it: the two answer different questions, and only one of
210//! them was ever about a cell.
211//!
212//! # 0.23.0: a table lays itself out, and a row shows its controls
213//!
214//! Three things a description could say and this renderer had no rule for,
215//! found together by rendering the MNW server's SSH-keys settings tab through
216//! `quasi` and preferring the hand-written Askama original.
217//!
218//! **A table had no layout at all.** [`list::narrowing_css`] emits the track
219//! list, and it has to be called with the columns, so it works where the
220//! columns are known at build time: goingson builds `tables.css` in its own
221//! `build.rs` and is untouched. A table a description produced knows its
222//! columns at render time, and the rules would have had to travel with the
223//! markup: a `<style>` element per table, which needs `style-src
224//! 'unsafe-inline'` that the MNW server is working to drop, or the head, which
225//! an htmx fragment swap does not carry. [`table_rules`] lays a table out with
226//! `display: table` instead, which aligns columns across rows knowing nothing
227//! about how many there are. [`Priority`](makeover_layout::Priority) hiding
228//! moves from a generated rule per dropped column to one rule per drop class,
229//! and [`list::column_classes`] is what puts those classes on a cell. A header
230//! row emitted by a renderer's own code should call it too, or the header and
231//! the body disagree about which column just dropped.
232//!
233//! **A destructive button had nowhere for its tone to land.** `.button` carried
234//! no tone, on the reading that a control's colour is its surface. Every
235//! consumer had written the danger rule itself. It joins the badge in taking
236//! the four tones as colour, off `data-tone`.
237//!
238//! **A list rendered as a bulleted list**, because nothing here reset the `ul`
239//! a renderer emits for one.
240//!
241//! `RowPart::revealed_on_hover` also stops being honoured, which its own doc
242//! sanctioned: a renderer decides. It was retired outright in makeover-layout
243//! 0.13.0, once this had been its only consumer for a release. The rule hid a
244//! row's actions until hover,
245//! and every escape it grew was a report that hiding was wrong for somebody:
246//! `focus-within` for the keyboard, the capability gate for a fingertip with no
247//! way to unhide. What survived hid the controls from pointer users alone, who
248//! are the ones scanning a list to learn what can be done to a row.
249//!
250//! # 0.17.0: the depth classes stop being controls
251//!
252//! [`depth_rules`] gave `.raised` the whole interactive set. A depth is a
253//! statement about shape, so that left the vocabulary with no raised surface
254//! that is merely an object, and an app wanting one had two moves: write its
255//! own class from tokens, or take a control class and cancel the control half.
256//! goingson took the second, in three variants over sixteen elements
257//! (`.card--static` at 14 call sites, `.card--muted` at 2, `.card--shell` at 1),
258//! each re-asserting the resting fill and bevel on `:hover` and `:active`.
259//!
260//! Measured before changing it: `.raised` is emitted into goingson, Balanced
261//! Breakfast and the MNW server, and none of the three has a single call site.
262//! The states were unasked-for everywhere at once, and dropping them costs no
263//! migration anywhere.
264//!
265//! `.card` and `.button` are unchanged. They are the same depth *and* controls,
266//! and they take their states from [`surface_rules`], which is where a state
267//! belongs: on the thing that claims to answer a pointer.
268//!
269//! # Substitution, three ways
270//!
271//! `Fill::Well` has no colour on makeover before 2.3.0, and each renderer
272//! answers that differently, which is the evidence that dropping
273//! `Fill::fallback` from the description was right:
274//!
275//! - `makeover-immediate` substitutes the page in Rust.
276//! - `makeover-tui` refuses to substitute and draws an edge instead, because a
277//! terminal would quantise the two together.
278//! - here, CSS already has the mechanism: `var(--surface-well,
279//! var(--surface-page))` falls back in the browser, and nothing in Rust
280//! decides anything.
281
282#![forbid(unsafe_code)]
283
284pub mod facet;
285pub mod figure;
286pub mod form;
287pub mod list;
288pub mod meter;
289pub mod placeholder;
290pub mod vocabulary;
291
292use crate::list::{cell_part_class, part_class};
293use makeover_geometry::{Density, SizeClass};
294// Re-exported rather than redefined. An app assembling its own stylesheet out
295// of this crate's pieces needs the same layer name, and most such apps depend
296// on this crate and not on `makeover-geometry` directly: goingson builds
297// `tables.css` in its own build.rs from [`list::narrowing_css`], and those
298// rules are as generated as the ones here.
299pub use makeover_geometry::{CSS_LAYER, in_css_layer};
300use makeover_layout::{
301 Bevel, CellPart, Depth, Fallback, Fill, Flow, Intent, RowPart, Selector, Sort, State, Token,
302 Tone,
303};
304use makeover_touch::Affordance;
305use std::fmt::Write as _;
306
307/// How the emitted CSS is shaped.
308#[derive(Debug, Clone, Copy, PartialEq, Eq)]
309pub struct Emit {
310 /// Bevel thickness, as a CSS length.
311 ///
312 /// A value, so it arrives from the caller: border widths belong to
313 /// `makeover-geometry` and will come from there once it carries them.
314 pub border_width: &'static str,
315 /// Focus ring thickness, as a CSS length.
316 ///
317 /// Separate from [`border_width`](Self::border_width), which it reused
318 /// until 0.12.0. That reuse was an implementation convenience dressed as
319 /// consistency, and it emitted a 1px ring: a bevel and a focus indicator
320 /// are answering different questions, and only one of them has to be
321 /// noticed from across a desk.
322 ///
323 /// The default is the measured consensus rather than a new opinion. Every
324 /// consumer had already written its own ring and all three chose at least
325 /// 2px: the MNW server 2px across 10 rules, Balanced Breakfast 2px,
326 /// goingson 2px on three rules and 3px on the one covering twelve
327 /// selectors. The design system was the only thing in the tree saying 1px.
328 pub focus_width: &'static str,
329 /// Prefix for emitted class names, without the leading dot.
330 pub class_prefix: &'static str,
331}
332
333impl Default for Emit {
334 fn default() -> Self {
335 Self {
336 border_width: "1px",
337 focus_width: "2px",
338 class_prefix: "",
339 }
340 }
341}
342
343/// A string as CSS escapes, for a `content` value.
344///
345/// `\u{25B2}` becomes `\25B2`. Emitted escaped rather than literally so the
346/// stylesheet is ASCII whatever the description spells: a `content` string is
347/// read by whatever encoding the consumer serves the file as, and a caret that
348/// depends on that is a caret that works on one machine.
349///
350/// Terminated by the closing quote at every site here. A CSS hex escape takes
351/// up to six digits and ends at the first character that cannot be one, so an
352/// escape followed by more text would need a space that these do not.
353fn css_escape(text: &str) -> String {
354 text.chars().fold(String::new(), |mut out, c| {
355 let _ = write!(out, "\\{:X}", c as u32);
356 out
357 })
358}
359
360/// The CSS custom property holding a bevel's composition.
361#[must_use]
362pub fn bevel_var(bevel: Bevel) -> &'static str {
363 match bevel {
364 Bevel::Raised => "--bevel-raised",
365 Bevel::Inset => "--bevel-inset",
366 }
367}
368
369/// A `var()` reference to a fill intent, with the browser's own fallback where
370/// the intent may be absent.
371///
372/// The fallback is CSS syntax, not a decision made here. That is the whole
373/// difference between this renderer and the other two.
374#[must_use]
375pub fn fill_var(fill: Fill) -> String {
376 match fill {
377 Fill::Well => format!("var(--{}, var(--{}))", fill.token(), Fill::Page.token()),
378 other => format!("var(--{})", other.token()),
379 }
380}
381
382/// The two-tone edge as a `box-shadow` value.
383///
384/// Two inset shadows, one per corner pair: the light one offset down and
385/// right so it lands on the top and left edges, the dark one the other way.
386/// The same assignment `makeover-immediate` draws with polylines and
387/// `makeover-tui` draws with box-drawing characters.
388#[must_use]
389pub fn bevel_shadow(bevel: Bevel, opts: &Emit) -> String {
390 let (top_left, bottom_right) = bevel.edges();
391 let w = opts.border_width;
392 format!(
393 "inset {w} {w} 0 var(--{}), inset -{w} -{w} 0 var(--{})",
394 top_left.token(),
395 bottom_right.token()
396 )
397}
398
399/// The custom properties both bevels resolve through.
400///
401/// Emitted as properties rather than inlined into every rule because that is
402/// what the apps already do, and because a consumer that wants the edge
403/// without the fill reads the property directly.
404#[must_use]
405pub fn bevel_properties(opts: &Emit) -> String {
406 let mut css = String::new();
407 for bevel in [Bevel::Raised, Bevel::Inset] {
408 let _ = writeln!(
409 css,
410 " {}: {};",
411 bevel_var(bevel),
412 bevel_shadow(bevel, opts)
413 );
414 }
415 css.push_str(ELEVATION_PROPERTY);
416 css
417}
418
419/// The cast shadow of a surface that floats over the page.
420///
421/// Composed here for the reason the bevel pair is: `makeover` derives the tone,
422/// this crate owns the geometry, and neither has to know the other's numbers.
423///
424/// **Only for a surface that overlays the page.** A menu, a toast, a popover, a
425/// dropdown. A surface *in* the page takes `.raised` and its bevel, and a rule
426/// that reaches for this on a card or a plate has renamed a literal rather than
427/// replaced it.
428///
429/// Two lengths rather than one, because a single blur reads as a smudge at
430/// plate size and as a halo at menu size. The offset is small and downward: a
431/// Platinum-era menu sits just off the page rather than hovering above it.
432const ELEVATION_PROPERTY: &str =
433 " --elevation-overlay: 0 2px 4px var(--elevation), 0 8px 24px var(--elevation);\n";
434
435/// The class name for a depth.
436#[must_use]
437pub fn depth_class(depth: Depth, opts: &Emit) -> Option<String> {
438 let name = match depth {
439 Depth::Flat => return None,
440 Depth::Raised => "raised",
441 Depth::Well => "well",
442 Depth::Sunken => "sunken",
443 // A depth added to the description since this renderer was last
444 // built. No class, on the same footing as Flat: emitting a name
445 // whose rule body we cannot write would put a class in the markup
446 // that the stylesheet never defines.
447 _ => return None,
448 };
449 Some(format!("{}{name}", opts.class_prefix))
450}
451
452/// A prefixed class name.
453///
454/// Public since 0.27.0, for the renderers that emit markup this crate does not.
455/// A screen renderer writing `class="row"` has to prefix it the way the
456/// stylesheet half does or a prefixed app gets rules matching everything except
457/// the elements that renderer wrote, and the failure is invisible: the CSS
458/// stays valid and one element is unstyled. quasi-webview carried a byte
459/// identical copy of this function until it could call this one.
460#[must_use]
461pub fn class(name: &str, opts: &Emit) -> String {
462 let mut out = String::with_capacity(opts.class_prefix.len() + name.len());
463 push_class(&mut out, name, opts);
464 out
465}
466
467/// A prefixed class name, written into a buffer the caller already has.
468///
469/// The form the emitters use, and the reason it exists is [`escape_into`]'s:
470/// every class on every element went through a `format!` before 0.40.0,
471/// including the default case where the prefix is empty and the answer is the
472/// argument. A described table row carried roughly eighty transient
473/// allocations, and this and the escaper were most of them.
474///
475/// [`class`] stays for callers holding a name rather than a buffer.
476///
477/// [`escape_into`]: crate::form::escape_into
478pub fn push_class(out: &mut String, name: &str, opts: &Emit) {
479 out.push_str(opts.class_prefix);
480 out.push_str(name);
481}
482
483/// The class an option of a selector carries, which is what the rules key off.
484///
485/// Named for the option and not for the group: [`selector_rules`] styles the
486/// thing that gets picked, so `Selector::Tabs` is `tab` and not `tabs`. The
487/// distinction is not pedantry. quasi-webview spelled these `tabs`, `segmented`
488/// and `option`, put `toggle` on the wrapping element rather than on the
489/// buttons inside it, and every described selector in that renderer came out
490/// with no depth, no focus ring and no chosen state, while the toggle group got
491/// a bevel meant for its buttons.
492///
493/// The chosen option additionally carries `chosen`, the same way a latched chip
494/// carries `latched`. That name is this crate's too; there is no reason for a
495/// caller to spell it, and [`selector_rules`] is where it is written down.
496#[must_use]
497pub fn option_class(selector: Selector) -> &'static str {
498 match selector {
499 Selector::Tabs => "tab",
500 Selector::Segmented => "segment",
501 Selector::Toggle => "toggle",
502 }
503}
504
505/// The fill and edge declarations for a depth, as a rule body.
506///
507/// Empty for [`Depth::Flat`], which has neither and inherits what it sits on.
508/// Callers lean on the emptiness to skip the rule rather than emit a class that
509/// sets nothing: a class that sets no properties is a class that means "I
510/// thought about this", which is what comments are for.
511///
512/// The two halves are emitted independently because [`Depth::Sunken`] has a
513/// fill and no bevel. Requiring both, which this did before makeover-layout
514/// 0.3.0, silently dropped the fill for exactly that case. Independent does not
515/// mean unpaired: both halves still come off one `Depth`, so they cannot
516/// disagree about what the region is.
517#[must_use]
518pub fn depth_declarations(depth: Depth) -> String {
519 let mut css = String::new();
520 if let Some(fill) = depth.fill() {
521 let _ = writeln!(css, " background: {};", fill_var(fill));
522 }
523 if let Some(bevel) = depth.bevel() {
524 let _ = writeln!(css, " box-shadow: var({});", bevel_var(bevel));
525 }
526 css
527}
528
529/// One rule giving a selector a depth, or nothing when the depth declares
530/// nothing.
531#[must_use]
532pub fn depth_rule(selector: &str, depth: Depth) -> String {
533 let body = depth_declarations(depth);
534 if body.is_empty() {
535 return String::new();
536 }
537 format!(".{selector} {{\n{body}}}\n")
538}
539
540/// The media condition a hover rule has to sit inside, or `None` if hover is
541/// unconditional.
542///
543/// Two crates answer this and neither answer is made here. `makeover-touch`
544/// owns *whether* hover exists at a density, and `makeover-geometry` owns how
545/// that capability is spelled as a media condition. Asking both is what stops
546/// this renderer minting a third opinion, which is what all three apps did:
547/// goingson sniffed the user agent, Balanced Breakfast used `(hover: none)`
548/// alone, and the MNW server had no gate at all.
549///
550/// [`SizeClass`] is required by [`Affordance::available`] and ignored by this
551/// member, which reports as much through `reads_size`. Passing Compact is not
552/// a claim about width; the test below pins that every class agrees.
553fn hover_condition() -> Option<&'static str> {
554 if Affordance::Hover.available(Density::Touch, SizeClass::Compact) {
555 // A fingertip grew a hover state. Nothing to gate, and this renderer
556 // should not invent a reason to gate anyway.
557 None
558 } else {
559 Some(Density::Pointer.media_condition())
560 }
561}
562
563/// Put a rule inside a media query, or leave it alone.
564fn gated(condition: Option<&str>, rule: &str) -> String {
565 let Some(condition) = condition else {
566 return rule.to_string();
567 };
568 let mut css = format!("@media {condition} {{\n");
569 for line in rule.lines() {
570 // Blank lines stay blank. Indenting one leaves trailing whitespace,
571 // which is the sort of thing a formatter later reverts and calls a diff.
572 if line.is_empty() {
573 css.push('\n');
574 } else {
575 let _ = writeln!(css, " {line}");
576 }
577 }
578 css.push_str("}\n");
579 css
580}
581
582/// The keyboard focus ring, placed by the depth it lands on.
583///
584/// This is the webview's **focus ring** and nothing more. **Reach** and
585/// **focus** are both the browser's — the document decides what is reachable
586/// and `:focus-visible` decides which reached thing wears the ring — and no
587/// description states either. The three terms are defined once in
588/// `makeover_layout`'s crate header, "Reach, focus and the focus ring".
589///
590/// One ring for the whole system, because a focus ring's job is to be
591/// recognised and three apps having three of them is the failure. What varies
592/// is where it sits, and that comes off [`Depth`] rather than off a per-
593/// component choice: a well takes the ring inside its own edge, and anything
594/// standing proud of the page takes it outside.
595///
596/// `outline` rather than the composed `box-shadow` the invalid-field ring at
597/// [`field_rules`] uses, and deliberately the one place the two rings are built
598/// differently. A `box-shadow` ring has to restate the bevel beside it, because
599/// `box-shadow` is not additive and a lone ring silently drops the well out
600/// from under the element. That restatement is a second copy of the depth,
601/// living in a different function from the first, and it is exactly the
602/// duplication `Depth` exists to prevent. `outline` occupies its own property,
603/// so the bevel survives untouched and there is nothing to keep in agreement.
604/// They render the same: both are a flush ring one border-width wide.
605#[must_use]
606pub fn focus_rule(selector: &str, depth: Depth, opts: &Emit) -> String {
607 let w = opts.focus_width;
608 // Same magnitude either way, and only the sign comes off the depth. Both
609 // values are what the consumers had already converged on independently:
610 // 2px out is what all three wrote, and 2px in is the MNW server's own
611 // answer for the one inset ring it had.
612 let offset = match depth.bevel() {
613 // Inside the well, clear of its edge rather than painted over it.
614 Some(Bevel::Inset) => format!("calc(-1 * {w})"),
615 // Raised, or no edge at all. Outside, standing off by its own width.
616 _ => w.to_string(),
617 };
618 // The token by name. It is `makeover`'s, derived from the action colour,
619 // and reaching it through a description member was a second path to the
620 // same variable for as long as one existed.
621 format!(
622 ".{selector}:focus-visible {{\n outline: {w} solid var(--focus-ring);\n outline-offset: {offset};\n}}\n"
623 )
624}
625
626/// Present, visible, and not answering.
627///
628/// Matches the ARIA attribute as well as the pseudo-class, because `:disabled`
629/// only matches form elements and half the things this crate emits are not
630/// one: a `div` carrying `.chip` or `.tab` can never be `:disabled`. Keying on
631/// the accessible state is the pattern [`field_rules`] already establishes for
632/// `aria-invalid`, on the reasoning that one fact read by both the styling and
633/// the accessibility tree cannot drift from itself.
634///
635/// The rest depth is re-asserted rather than assumed, because this rule has to
636/// beat the hover and pressed rules above it. It does that on source order at
637/// equal specificity, not by out-specifying them: every rule this function's
638/// caller emits is (0,2,0), and adding a `:not(:disabled)` anywhere would raise
639/// one of them and have to be unpicked when this output moves inside its own
640/// cascade layer.
641#[must_use]
642pub fn disabled_rule(selector: &str, depth: Depth) -> String {
643 format!(
644 ".{selector}:disabled,\n.{selector}[aria-disabled=\"true\"] {{\n{} color: var(--{});\n cursor: not-allowed;\n}}\n",
645 depth_declarations(depth),
646 State::Disabled.token()
647 )
648}
649
650/// Every state a selector that answers a click implies: hover, pressed, focus
651/// and disabled, in that order.
652///
653/// Order is the whole cascade mechanism here. All four selectors are
654/// specificity (0,2,0), so disabled wins over hover and pressed by coming last
655/// and by nothing else.
656///
657/// Pressed emits [`Depth::pressed`] in full, fill and edge together. Emitting
658/// only the edge is what left goingson hand-writing `background:
659/// var(--surface-sunken)` on three separate rules, and a fill that does not
660/// travel with its edge is precisely the disagreement `Depth` exists to make
661/// unrepresentable. So the pressed fill comes from the description
662/// (`--surface-well`) rather than from whatever each app reached for.
663///
664/// Hover has no member in the description and is renderer policy: a terminal
665/// and an immediate-mode painter have no hover to express. It resolves against
666/// `--hover-surface`, which `makeover` already derives and which nothing
667/// consumed until now. What it *is* gated on is capability, via
668/// [`hover_condition`]. Before that gate existed the apps each wrote their own:
669/// goingson's section 60 exists solely to take back the hover state this
670/// function had just handed it, by out-specifying a rule it does not own.
671///
672/// `depth` is the selector's **rest** depth, used to place the focus ring and
673/// to restore the surface under a disabled control. The pressed rule keeps
674/// inverting from [`Depth::Raised`] regardless, which is what every caller got
675/// before this parameter existed: a tab's unchosen depth is
676/// [`Depth::Sunken`], and `Sunken.pressed()` is `Sunken`, so deriving the press
677/// from the rest depth would leave a tab with no press at all.
678#[must_use]
679pub fn interactive_rules(selector: &str, depth: Depth, opts: &Emit) -> String {
680 let mut css = gated(
681 hover_condition(),
682 &format!(".{selector}:hover {{\n background: var(--hover-surface);\n}}\n"),
683 );
684 css.push_str(&depth_rule(
685 &format!("{selector}:active"),
686 Depth::Raised.pressed(),
687 ));
688 css.push_str(&focus_rule(selector, depth, opts));
689 css.push_str(&disabled_rule(selector, depth));
690 css
691}
692
693/// One rule per depth: its fill and its edge, together.
694///
695/// A depth and nothing else. `.raised` says a surface sits on what is behind
696/// it, which is a statement about the shape and not about what happens when a
697/// pointer arrives, so it emits no hover, press, focus or disabled rule. The
698/// named surfaces are where interaction lives: `.card` and `.button` are the
699/// same depth *and* controls, and they get their states from
700/// [`surface_rules`].
701///
702/// This class carried the interactive set until 0.17.0, which left the
703/// vocabulary with no raised surface that is merely an object. Consumers that
704/// needed one took a control class and cancelled half of it instead: sixteen
705/// elements in goingson across three `.card--*` variants, each re-asserting the
706/// resting fill and bevel on `:hover` and `:active`. Nothing anywhere used
707/// `.raised` itself, so the states were unasked-for in every consumer at once.
708#[must_use]
709pub fn depth_rules(opts: &Emit) -> String {
710 let mut css = String::new();
711 for depth in [Depth::Raised, Depth::Well] {
712 let Some(class) = depth_class(depth, opts) else {
713 continue;
714 };
715 css.push_str(&depth_rule(&class, depth));
716 }
717 css
718}
719
720/// The three surfaces that are a depth with a name.
721///
722/// `button` and `card` are both [`Depth::Raised`], and `field` is a
723/// [`Depth::Well`] because that is the reading `Depth`'s own documentation
724/// gives a text field. Their bodies come out identical by construction rather
725/// than by hand: three hand-written copies in goingson's stylesheet is what
726/// phase A deletes, and generating them from one call is what stops them
727/// drifting apart again.
728fn surface_rules(opts: &Emit) -> String {
729 let mut css = String::new();
730 for name in ["button", "card"] {
731 let c = class(name, opts);
732 css.push_str(&depth_rule(&c, Depth::Raised));
733 css.push_str(&interactive_rules(&c, Depth::Raised, opts));
734 }
735
736 let field = class("field", opts);
737 css.push_str(&depth_rule(&field, Depth::Well));
738
739 // A field takes focus and refuses input like everything else here, and got
740 // neither until now, which is why all three apps hand-write a focus ring
741 // for it and no two of them match. No hover or pressed: a text field does
742 // not light up under the pointer and does not invert when clicked, so the
743 // two states `interactive_rules` would add are the two it does not have.
744 css.push_str(&focus_rule(&field, Depth::Well, opts));
745 css.push_str(&disabled_rule(&field, Depth::Well));
746
747 // Keyed on the ARIA attribute rather than on a class, so the visual state
748 // and the accessible state cannot drift apart: there is one fact and both
749 // read it. goingson already drove its invalid styling this way and was
750 // right to; the `.invalid` class this emitted before 0.5.0 was a second
751 // place to forget.
752 //
753 // The ring composes *after* the bevel rather than replacing it. box-shadow
754 // is not additive, so a lone ring silently dropped the well out from under
755 // an invalid field. Flat and unlit: this edge is saying "wrong", and
756 // lighting one side would have it say "raised" at the same time.
757 let _ = writeln!(
758 css,
759 ".{field}[aria-invalid=\"true\"] {{\n box-shadow: var({}), 0 0 0 {} var(--danger);\n}}",
760 bevel_var(Bevel::Inset),
761 opts.border_width
762 );
763 css
764}
765
766/// Badges and chips.
767///
768/// The one place phase A changes how goingson looks rather than only where its
769/// rules live. [`Token::Badge`] is [`Depth::Flat`], so a badge emits no fill
770/// and no edge at all, where goingson ships `.tag, .badge` as a single rule
771/// carrying the raised bevel. Splitting that means reading every call site to
772/// decide which of the two it always was.
773///
774/// What a badge does carry is a [`Tone`], the intent family it shares with
775/// notices and nothing else. Neutral is the bare class rather than a variant,
776/// because it is the absence of a status and not a status called "none".
777/// Text that goes somewhere.
778///
779/// The one inline control, and the vocabulary had no word for it until a
780/// description layer needed one. A table cell has carried `cell-link` since
781/// 0.14.0's part list, deliberately unruled because the cell's own rule covers
782/// it; what was missing is the same thing outside a table, which is what a
783/// described run holds when a sentence contains a link.
784///
785/// Colour and underline only. Whether a link is inline in a sentence or sitting
786/// on its own line is the app's layout, and how much room it takes is
787/// `makeover-geometry`'s. What is here is the pair of signals that say "this
788/// goes somewhere" and nothing that says where it sits.
789///
790/// The visited arm is deliberately absent. A link inside an app points at the
791/// app's own screens, which the user is expected to have been to, so painting
792/// them differently marks almost everything and distinguishes nothing.
793fn link_rules(opts: &Emit) -> String {
794 let mut css = String::new();
795 let link = class("link", opts);
796
797 let _ = writeln!(
798 css,
799 ".{link} {{\n color: var(--action);\n \
800 text-decoration: underline;\n}}"
801 );
802 // The hover step is the same one every other control takes, and it is a
803 // colour rather than a surface: a link has no box to raise.
804 let _ = writeln!(
805 css,
806 "@media (hover: hover) and (pointer: fine) {{\n .{link}:hover \
807 {{\n color: var(--action-hover);\n }}\n}}"
808 );
809 let _ = writeln!(
810 css,
811 ".{link}:focus-visible {{\n outline: {} solid var(--focus-ring);\n \
812 outline-offset: 2px;\n}}",
813 opts.focus_width
814 );
815 // A link is often a `<button>` rather than an `<a>`: a renderer picks the
816 // element from the method, so a link that writes is a button that has to
817 // stop looking like one.
818 let _ = writeln!(
819 css,
820 "button.{link} {{\n background: none;\n border: none;\n \
821 padding: 0;\n font: inherit;\n cursor: pointer;\n}}"
822 );
823 css
824}
825
826fn token_rules(opts: &Emit) -> String {
827 let mut css = String::new();
828
829 // No `depth_rule` call here, deliberately: `Token::Badge.depth(_)` is Flat,
830 // and a label with an edge says it can be pressed.
831 let badge = class("badge", opts);
832 let _ = writeln!(
833 css,
834 ".{badge} {{\n color: var(--{});\n}}",
835 Tone::Neutral.token()
836 );
837 for tone in [Tone::Info, Tone::Success, Tone::Warning, Tone::Danger] {
838 let _ = writeln!(
839 css,
840 ".{badge}[data-tone=\"{0}\"] {{\n color: var(--{0});\n}}",
841 tone.token()
842 );
843 }
844
845 // A button carries the four tones a badge does. It had none, on the reading
846 // that a control's colour is its surface rather than its text, and that
847 // reading has one hole big enough to matter: the button that destroys
848 // something. Every consumer had written that rule itself, and a description
849 // that says `Tone::Danger` on an act had nowhere for it to land.
850 //
851 // Colour and not a fill, matching the badge. A red surface is a decision
852 // about emphasis that belongs to an app's own layer, and two of them
853 // fighting is worse than neither.
854 let button = class("button", opts);
855 for tone in [Tone::Info, Tone::Success, Tone::Warning, Tone::Danger] {
856 let _ = writeln!(
857 css,
858 ".{button}[data-tone=\"{0}\"] {{\n color: var(--{0});\n}}",
859 tone.token()
860 );
861 }
862
863 // A chip holds itself down, which is `Depth::pressed` arrived at
864 // independently by two apps. `removable` is a remove affordance, so it is
865 // markup and waits for phase B.
866 let chip = class("chip", opts);
867 let unlatched = Token::Chip { removable: false };
868 css.push_str(&depth_rule(&chip, unlatched.depth(false)));
869 css.push_str(&interactive_rules(&chip, unlatched.depth(false), opts));
870 css.push_str(&depth_rule(
871 &format!("{chip}.latched"),
872 unlatched.depth(true),
873 ));
874 css
875}
876
877/// The three selectors, each named by what it picks.
878///
879/// A tab comes *forward* to join the pane it opens, which is why
880/// [`Selector::Tabs`] chooses [`Depth::Raised`] where a segment and a toggle
881/// are held in. That is the folder semantic, and it is the whole reason the
882/// three are not one member with a flag.
883///
884/// [`Selector::abutting`] is not emitted: whether the options touch is
885/// spacing, and spacing is `makeover-geometry`'s question to answer.
886///
887/// Both states emit as of makeover-layout 0.3.0. Before it the description
888/// named only the chosen option, so an unchosen one fell through to
889/// [`Depth::Flat`] and nothing was drawn for it, which left goingson's tab
890/// strip hand-writing the recess that makes its chosen tab read as forward.
891fn selector_rules(opts: &Emit) -> String {
892 let mut css = String::new();
893 for selector in [Selector::Tabs, Selector::Segmented, Selector::Toggle] {
894 let c = class(option_class(selector), opts);
895 css.push_str(&depth_rule(&c, selector.unchosen()));
896 css.push_str(&interactive_rules(&c, selector.unchosen(), opts));
897 css.push_str(&depth_rule(&format!("{c}.chosen"), selector.chosen()));
898 }
899 css
900}
901
902/// The parts of a list row.
903///
904/// The list is written out rather than derived because `RowPart` is
905/// `#[non_exhaustive]` as of makeover-layout 0.9.0, so there is nothing to
906/// iterate. A member added upstream emits no rule until it is named here, which
907/// is the trade `non_exhaustive` makes: a silent gap instead of a build break.
908/// [`part_class`] carries the same list and the same obligation.
909fn row_rules(opts: &Emit) -> String {
910 let mut css = String::new();
911
912 // The container the rows sit in. Three zeroes and a keyword, all of them
913 // undoing a browser default the description never asked for: a described
914 // list of SSH keys is not a bulleted list, and it rendered as one because
915 // nothing here said otherwise. Not a size: there is no magnitude in it,
916 // which is the line this crate holds.
917 let _ = writeln!(
918 css,
919 ".{} {{\n list-style: none;\n margin: 0;\n padding: 0;\n}}",
920 class("list", opts)
921 );
922
923 for part in [
924 RowPart::Primary,
925 RowPart::Secondary,
926 RowPart::Meta,
927 RowPart::Actions,
928 RowPart::Tokens,
929 RowPart::Proportion,
930 ] {
931 let c = class(part_class(part), opts);
932
933 // Actions carry controls rather than text, and `RowPart::intent` says
934 // so by returning the same intent inheriting already gives. Pinning it
935 // would be louder than saying nothing. Tokens answer alike, for their
936 // own reason: each token carries its own tone, and a colour on the
937 // strip would fight the things sitting in it. A proportion is the same
938 // case again: the meter inside carries the tone.
939 if !matches!(
940 part,
941 RowPart::Actions | RowPart::Tokens | RowPart::Proportion
942 ) {
943 let _ = writeln!(css, ".{c} {{\n color: var(--{});\n}}", part.intent());
944 }
945
946 // A row's actions are shown at rest. `RowPart::revealed_on_hover` said
947 // otherwise and was not honoured here from 0.23.0; makeover-layout
948 // 0.13.0 retired the method, so there is no longer a description saying
949 // one thing and a renderer doing another.
950 //
951 // The rule was `opacity: 0` gated to pointer devices, revealed on
952 // `:hover` and on `:focus-within`. Each escape it needed was a report
953 // that hiding was wrong for somebody: `focus-within` because tabbing
954 // could never reach an action; the gate because a fingertip had no way
955 // to unhide, which both webview apps had already hand-written
956 // `opacity: 1` to undo. What was left was a control hidden from
957 // exactly one group: people using a pointer, who are also the group
958 // scanning a list to find out what can be done to a row.
959 //
960 // A settings screen is where that reads worst: the whole reason to be
961 // on it is to remove a key, and the button doing so was invisible
962 // until pointed at. A table cell's actions were never hidden, so the
963 // two arrangements now agree.
964 }
965
966 // A part that may take two lines. `Flow::Tight` gets no rule: one line is
967 // what a run already does, and restating it here would put a declaration on
968 // every part in every row to say nothing.
969 //
970 // This is the shape both webview apps had already written by hand and
971 // commented -- Balanced Breakfast on a feed row's title, goingson on a
972 // problem's body -- which is the whole argument for the description
973 // carrying it. `-webkit-` prefixed and unprefixed together: the prefixed
974 // trio is what every engine actually implements, and `line-clamp` is the
975 // standard property landing behind it.
976 let _ = writeln!(
977 css,
978 ".{} {{\n display: -webkit-box;\n -webkit-box-orient: vertical;\n -webkit-line-clamp: {lines};\n line-clamp: {lines};\n overflow: hidden;\n}}",
979 class("row-relaxed", opts),
980 lines = Flow::Relaxed.lines()
981 );
982
983 css
984}
985
986/// A row of things that share their space, and what each fallback gets here.
987///
988/// Ruling: wiki `layout-room-and-fallback`, Max 2026-08-18. The bug it answers
989/// is goingson's `styles.css:702`, which pinned every `.page-header` over the
990/// pill strip with `position: absolute` and so took the toolbar out of flow:
991/// zero width contributed to the row it shared, nothing able to collide with
992/// it, and therefore nothing preventing the collision. Rule 1 of the ruling is
993/// that every described member is in flow, and these rules are how that is kept
994/// rather than asked for.
995///
996/// # The floor, which is most of the fix
997///
998/// `.run > *` gets `min-width: min-content`. That is the derived minimum the
999/// ruling asks for, in this renderer's own unit and stated by the browser
1000/// rather than by anybody: a member cannot be squeezed narrower than what is
1001/// in it, so members in one flow push each other instead of overlapping. It
1002/// costs no query and no number, and it is what fixes all four measured widths
1003/// whichever fallback the group declared.
1004///
1005/// # What each fallback gets, exactly
1006///
1007/// [`Fallback::Wrap`] is `flex-wrap: wrap`, which is exact. The browser wraps
1008/// the run when the members no longer fit, deciding that from their own
1009/// intrinsic widths, which is the derived minimum doing the whole job.
1010///
1011/// [`Fallback::Stack`] is wrap plus `flex: 1 1 max-content` on the members, so
1012/// a member that cannot sit beside its sibling takes a line of its own and
1013/// fills it. For the two-member run this was ruled on -- a tab strip and a
1014/// band -- that is precisely "a row becomes a column".
1015///
1016/// [`Fallback::Shed`] and [`Fallback::Menu`] get wrap, and this renderer is
1017/// honouring less than the description says. **CSS cannot express either one
1018/// without breaking the ruling's own first constraint.** Both need to know that
1019/// the run is out of room in order to take a member out of it, a container
1020/// query is the only construct that can ask, and `@container` compares against
1021/// a `<length>` -- there is no `@container (inline-size < min-content)`. So
1022/// every honest spelling of Shed here needs an authored breakpoint, which is
1023/// the thing the ruling exists to forbid, and the dishonest ones are worse: a
1024/// clamped height clips by document order rather than by [`Priority`], and
1025/// `display: none` under a viewport `@media` is the `nth-child(n+5)` bug the
1026/// vocabulary replaced.
1027///
1028/// Wrapping is the right thing to do instead. It keeps every member reachable,
1029/// which is the property that was actually broken -- goingson's new-contact
1030/// button left the viewport entirely at 560 -- and it keeps rule 1. A renderer
1031/// answering with less than was described is precedented and deliberate here:
1032/// [`makeover_layout::Region::Columns`] says a terminal stacking a board's
1033/// columns is honouring the description rather than degrading it.
1034///
1035/// The real mechanism needs the shed members to have somewhere to go, which is
1036/// markup and belongs to quasi-webview: an overflow control is a member of the
1037/// run, and the description does not yet say that a member *is* one. Filed
1038/// rather than guessed at.
1039fn run_rules(opts: &Emit) -> String {
1040 let run = class("run", opts);
1041 let mut css = String::new();
1042
1043 // `flex-wrap: nowrap` is stated rather than left to the default, because
1044 // the fallbacks below are read as overrides of this line and a reader
1045 // should not have to know which way flexbox leans to see that.
1046 //
1047 // No gap. Spacing between members is the app's, the same way this crate
1048 // states no margins anywhere else; a gap here would be a size, and the one
1049 // hardcoded size in the mechanism is makeover-geometry's contact patch.
1050 let _ = writeln!(
1051 css,
1052 ".{run} {{\n display: flex;\n flex-wrap: nowrap;\n align-items: center;\n}}"
1053 );
1054
1055 // The derived minimum, and the whole reason a member can no longer be
1056 // overlapped. `min-width: auto` is flexbox's default for a flex item and is
1057 // *not* the same thing: auto lets an item be compressed below its content
1058 // in a nowrap run, which is how a toolbar ends up drawn over a tab strip
1059 // even without anything leaving the flow.
1060 let _ = writeln!(css, ".{run} > * {{\n min-width: min-content;\n}}");
1061
1062 for fallback in [
1063 Fallback::Wrap,
1064 Fallback::Stack,
1065 Fallback::Shed,
1066 Fallback::Menu,
1067 ] {
1068 let name = fallback_class(fallback);
1069 let c = class(name, opts);
1070 let _ = writeln!(css, ".{c} {{\n flex-wrap: wrap;\n}}");
1071 if matches!(fallback, Fallback::Stack) {
1072 let _ = writeln!(css, ".{c} > * {{\n flex: 1 1 max-content;\n}}");
1073 }
1074 }
1075
1076 css
1077}
1078
1079/// Every class [`fallback_class`] can return, plus the run itself.
1080///
1081/// [`ROW_PART_CLASSES`](crate::list::ROW_PART_CLASSES)'s reasoning and the same
1082/// obligation: a `match` over a `#[non_exhaustive]` enum cannot be enumerated
1083/// from outside, so the list sits beside it and a test holds the two together.
1084/// `run` is in it because it is emitted in its own right rather than only as a
1085/// fallback's fallback.
1086pub const RUN_CLASSES: &[&str] = &["run", "run-wrap", "run-stack", "run-shed", "run-menu"];
1087
1088/// The class a run carries for what it does when it is tight.
1089///
1090/// A run always carries `.run` as well, so an unrecognised fallback -- the enum
1091/// is `#[non_exhaustive]` -- lands as a plain nowrap row with the min-content
1092/// floor still under it. That is the safe failure: every member in flow and
1093/// none overlapped, which is the property, with only the rearrangement missing.
1094#[must_use]
1095pub fn fallback_class(fallback: Fallback) -> &'static str {
1096 match fallback {
1097 Fallback::Wrap => "run-wrap",
1098 Fallback::Stack => "run-stack",
1099 Fallback::Shed => "run-shed",
1100 Fallback::Menu => "run-menu",
1101 _ => "run",
1102 }
1103}
1104
1105/// The progress trough these rules fill.
1106///
1107/// This was renderer-local chrome with nothing behind it until makeover-layout
1108/// 0.10.0, which is the unusual order: the tones below were emitted for every
1109/// bar in the tree while the only way to describe one was to concatenate the
1110/// numbers into a heading. `Meter` is the word that arrived late, and
1111/// [`meter::meter_html`](crate::meter::meter_html) is what now fills these.
1112///
1113/// The rules stay a superset of what a description can ask for. An app drawing
1114/// its own bar keeps these classes, which is what the four goingson grew
1115/// independently were adopted onto.
1116///
1117/// The trough is a [`Depth::Well`], the same reading a text field gets:
1118/// something with its content down inside it.
1119fn progress_rules(opts: &Emit) -> String {
1120 let progress = class("progress", opts);
1121 // `progress-fill` rather than a bare `fill`: an unprefixed build claims
1122 // these names in the app's own stylesheet, and `.fill` is grabby enough to
1123 // catch things that have nothing to do with progress. goingson already
1124 // calls it `.progress-fill`, so this is also the name that deletes.
1125 let fill = class("progress-fill", opts);
1126 let mut css = depth_rule(&progress, Depth::Well);
1127
1128 // The untoned bar is `--action`, not [`Tone::Neutral`]. That is the one
1129 // place this differs from the badge rules, and deliberately: a badge with
1130 // no status is a muted label, while a bar with no status is still
1131 // reporting progress, and `content-muted` would read as disabled.
1132 let _ = writeln!(
1133 css,
1134 ".{progress} > .{fill} {{\n background: var(--action);\n}}"
1135 );
1136
1137 // A bar can be saying something, same as a badge: goingson colours subtask
1138 // progress as success and an over-estimate as danger, which is real
1139 // information rather than decoration. Emitting the tones is what lets that
1140 // survive adoption instead of staying hand-written.
1141 for tone in [Tone::Info, Tone::Success, Tone::Warning, Tone::Danger] {
1142 let _ = writeln!(
1143 css,
1144 ".{progress} > .{fill}[data-tone=\"{0}\"] {{\n background: var(--{0});\n}}",
1145 tone.token()
1146 );
1147 }
1148 css
1149}
1150
1151/// A strip of figures, and the two spans inside each one.
1152///
1153/// Colour only, which is the deferral rule applied to a component that badly
1154/// wants to break it. A figure reads as a figure because the value is set large
1155/// over a small caption, and that is a size: `makeover-geometry` answers how
1156/// much space and this crate answers what the thing is. Emitting `font-size`
1157/// here would be this crate naming a value, which is the one thing it is defined
1158/// by not doing, and `progress_rules` is the precedent — it emits the tones and
1159/// never the width, because the width is not its to know.
1160///
1161/// So the arrangement and the type scale are the app's, and what is generated is
1162/// the part an app cannot get right by itself: which of the two spans carries
1163/// the tone.
1164fn figure_rules(opts: &Emit) -> String {
1165 let figure = class("figure", opts);
1166 let value = class("figure-value", opts);
1167 let caption = class("figure-caption", opts);
1168 let change = class("figure-change", opts);
1169 let mut css = String::new();
1170
1171 let _ = writeln!(
1172 css,
1173 ".{figure} > .{value} {{\n color: var(--{});\n}}",
1174 Tone::Neutral.token()
1175 );
1176 let _ = writeln!(
1177 css,
1178 ".{figure} > .{caption} {{\n color: var(--content-muted);\n}}"
1179 );
1180 let _ = writeln!(
1181 css,
1182 ".{figure} > .{change} {{\n color: var(--content-muted);\n}}"
1183 );
1184
1185 // A toned figure tones one part and never the caption. The caption is the
1186 // noun and stays muted.
1187 //
1188 // Which part depends on whether there is a change, and that is the whole of
1189 // what 0.13.0 changed here. A figure with a delta is an ordinary number that
1190 // has moved in a direction worth reading, so the delta takes the colour and
1191 // the number stays plain; a figure without one has nowhere else to put it.
1192 // `:has` is what lets one attribute mean both, and the alternative was the
1193 // emitter deciding by writing the attribute onto a different element, which
1194 // leaves two elements able to disagree about a figure's one meaning.
1195 for tone in [Tone::Info, Tone::Success, Tone::Warning, Tone::Danger] {
1196 let _ = writeln!(
1197 css,
1198 ".{figure}[data-tone=\"{0}\"] > .{change} {{\n color: var(--{0});\n}}",
1199 tone.token()
1200 );
1201 let _ = writeln!(
1202 css,
1203 ".{figure}[data-tone=\"{0}\"]:not(:has(> .{change})) > .{value} \
1204 {{\n color: var(--{0});\n}}",
1205 tone.token()
1206 );
1207 }
1208 css
1209}
1210
1211/// A picture, its frame and its caption.
1212///
1213/// `makeover_layout::Image` arrived at 0.21.0, found by trying to describe MNW's
1214/// carousel and discovering nothing named a picture.
1215///
1216/// # The frame is a border, and this is the one place a bevel is wrong
1217///
1218/// 0.36.0 emitted [`Depth::Raised`] here through [`depth_rule`], the same call
1219/// `button` and `card` make. It was wrong, and MNW's landing page is where it
1220/// showed: **the frames had no visible edge at all.**
1221///
1222/// `Depth::Raised`'s edge is `--bevel-raised`, which is an *inset* shadow — a
1223/// 1px light run at the top-left and a dark one at the bottom-right, drawn
1224/// **inside** the element's box. On a button or a card that box is a surface
1225/// this crate owns, so an inset edge reads as the surface catching the light.
1226/// On a picture it is drawn on top of the picture, over whatever pixels the
1227/// image happens to have at its border. MNW's screenshots are light-on-light
1228/// parchment, so the light half landed on a light image and the frame
1229/// disappeared.
1230///
1231/// **You cannot bevel a surface you do not own.** A picture's content is the
1232/// app's, arrives at request time, and can be any colour, so its edge has to
1233/// sit *outside* the content rather than on it. That is a border.
1234///
1235/// The fill stays, and it is not decoration: it is what shows through a
1236/// transparent PNG and what stands in the frame's place while the image is
1237/// still loading.
1238///
1239/// This is a real limit on [`Depth`] rather than a special case. Every other
1240/// consumer of a depth draws its own surface; a picture is the first member
1241/// whose surface belongs to someone else.
1242///
1243/// # No size
1244///
1245/// `width: 100%` and nothing else. How large a picture is depends on the box it
1246/// was put in, which is the app's arrangement and `makeover-geometry`'s scales,
1247/// and a renderer that picked one would be answering for every consumer at
1248/// once. This is where `figure_rules` landed for the same reason.
1249fn picture_rules(opts: &Emit) -> String {
1250 let picture = class("picture", opts);
1251 let img = class("picture-img", opts);
1252 let caption = class("picture-caption", opts);
1253 let mut css = String::new();
1254
1255 // Block, or an inline image sits on the text baseline and carries a
1256 // descender's worth of space under it that no app ever wants and every app
1257 // deletes by hand. The border is the frame; see the type docs for why it is
1258 // not the bevel every other surface here gets.
1259 let _ = writeln!(
1260 css,
1261 ".{img} {{\n display: block;\n width: 100%;\n height: auto;\n \
1262 background: var(--{});\n border: {} solid var(--border);\n}}",
1263 Fill::Raised.token(),
1264 opts.border_width
1265 );
1266
1267 // The two fits that need a rule. `Fit::Natural` emits no attribute at all,
1268 // so it is the bare rule above and needs nothing here.
1269 let _ = writeln!(
1270 css,
1271 ".{img}[data-fit=\"cover\"] {{\n height: 100%;\n object-fit: cover;\n}}"
1272 );
1273 let _ = writeln!(
1274 css,
1275 ".{img}[data-fit=\"contain\"] {{\n height: 100%;\n object-fit: contain;\n}}"
1276 );
1277
1278 // A caption reads back one step, which is `figure-caption`'s answer and the
1279 // same claim: it says what the thing above it is, and it is not the thing.
1280 let _ = writeln!(
1281 css,
1282 ".{picture} > .{caption} {{\n color: var(--content-muted);\n}}"
1283 );
1284
1285 css
1286}
1287
1288/// A region showing one child at a time, and the chrome that moves between them.
1289///
1290/// [`makeover_layout::Showing`] lets a description say that a region holds
1291/// several children and shows some of them. A renderer derives its own chrome
1292/// from that, which is what stops every renderer growing a `match` on a widget
1293/// name; these are the rules the derived chrome needs.
1294///
1295/// # Why the default is every child, and the enhancement takes them away
1296///
1297/// The controls are a lie until something binds them. A prev button rendered
1298/// into a document with no script is a control that looks live and answers
1299/// nothing, and the reader it lies to is exactly the one who cannot see the
1300/// other children either — the collapsing and the moving are the same half.
1301///
1302/// So the rules run in the direction the enhancement does. Nothing here hides a
1303/// child and nothing here shows a control. Whatever binds the region sets
1304/// `data-ready` on it, and that is what collapses the stack to one and reveals
1305/// the row that moves it. A reader with no script gets every child in order and
1306/// no controls, which is more content rather than less, and a reader with
1307/// script gets a settled page rather than a stack that jumps to one frame after
1308/// load.
1309///
1310/// MNW proved this shape by hand — a `<noscript>` stylesheet opening its
1311/// carousel back out — and it is here rather than there because the property is
1312/// the description's, not one app's.
1313///
1314/// # Not spacing
1315///
1316/// The row's gaps are `makeover-geometry`'s question and are absent for
1317/// [`row_rules`]'s reason. What is here is `display`, which carries no
1318/// magnitude, and the muted readout, which is the same claim `picture-caption`
1319/// makes: it says where you are among the children and it is not one of them.
1320fn showing_rules(opts: &Emit) -> String {
1321 let controls = class("showing", opts);
1322 let position = class("showing-position", opts);
1323 let frame = class("showing-frame", opts);
1324 let mut css = String::new();
1325
1326 // Hidden until something binds it, which is the whole argument above.
1327 let _ = writeln!(css, ".{controls} {{\n display: none;\n}}");
1328 // Block, and nothing about how the three sit in it. A button and a span are
1329 // inline already, so they make a row without this crate saying so, and
1330 // saying so is where `align-items` and a gap would follow -- both spacing,
1331 // both `makeover-geometry`'s, and `row_rules` refuses them for the same
1332 // reason.
1333 let _ = writeln!(
1334 css,
1335 "[data-ready] > .{controls} {{\n display: block;\n}}"
1336 );
1337
1338 // A child is in flow until the region is bound, and then only the current
1339 // one is. `.current` is a modifier for the reason `.chosen` and `.latched`
1340 // are: one name for the state, set by whoever knows it.
1341 let _ = writeln!(
1342 css,
1343 "[data-ready] > .{frame}:not(.current) {{\n display: none;\n}}"
1344 );
1345
1346 // Reads back one step. `picture-caption`'s rule and its reason.
1347 let _ = writeln!(css, ".{position} {{\n color: var(--content-muted);\n}}");
1348
1349 css
1350}
1351
1352/// A time axis: the container, its gridlines and ruler, and the placed things.
1353///
1354/// `makeover-layout` 0.24.0's [`Track`](makeover_layout::Track), and the first
1355/// member here whose whole point is *position*. That makes the magnitude line
1356/// this crate keeps worth restating rather than assuming.
1357///
1358/// # Where the numbers come from
1359///
1360/// Every value that varies per item is a custom property the caller sets
1361/// inline, and every rule here reads one. `--track-at` and `--track-for` are
1362/// percentages of the span, which
1363/// [`Track::fraction`](makeover_layout::Track::fraction) computes once so three
1364/// renderers cannot disagree about it. Nothing here knows a pixel.
1365///
1366/// That is what lets the stylesheet stay static while the items move: an entry
1367/// carries `style="--track-at: 37.5%; --track-for: 4.166%"` and the rule below
1368/// turns it into a box. The alternative was emitting a rule per item, which is
1369/// a stylesheet that grows with the data.
1370///
1371/// **The height of the track is the app's**, not this crate's. 96 quarter-hour
1372/// slots at some slot height is a size, and a size is `makeover-geometry`'s
1373/// question -- the same refusal `figure_rules` and `placeholder_rules` make.
1374/// Percentages need a resolved height above them, so `.track` gets
1375/// `position: relative` and nothing else; the app says how tall a day is.
1376///
1377/// # Overlap
1378///
1379/// Two things at the same time is intrinsic to an axis and has no analogue in a
1380/// list. The description does not declare it -- `Placement::overlaps` derives it
1381/// from the times -- so what arrives here is a lane index and a lane count, and
1382/// the rule divides the width. A renderer that would rather stack them ignores
1383/// both properties and the defaults give it one full-width lane.
1384fn track_rules(opts: &Emit) -> String {
1385 let track = class("track", opts);
1386 let slot = class("track-slot", opts);
1387 let tick = class("track-tick", opts);
1388 let entry = class("track-entry", opts);
1389 let mut css = String::new();
1390
1391 // The positioning context every entry resolves against, and the whole of
1392 // what this crate says about the container. No height: see the doc above.
1393 let _ = writeln!(css, ".{track} {{\n position: relative;\n}}");
1394
1395 // Gridlines and the ruler are the axis reading itself back, which is
1396 // `picture-caption` and `showing-position`'s claim: about the thing rather
1397 // than one of the things.
1398 //
1399 // `border_width` rather than a literal, the way `depth_rule` writes its
1400 // edge. A gridline is the one place a timeline would most naturally reach
1401 // for a hardcoded 1px, and a hardcoded 1px is this crate naming a size.
1402 let _ = writeln!(
1403 css,
1404 ".{slot} {{\n border-top: {} solid var(--border);\n}}",
1405 opts.border_width
1406 );
1407 let _ = writeln!(css, ".{tick} {{\n color: var(--content-muted);\n}}");
1408
1409 // The one rule that does real work. Top and height are the placement;
1410 // left and width are the lane. Both lane properties default so an entry
1411 // that names neither is full width, which is the common case and the one a
1412 // renderer gets for free.
1413 let _ = writeln!(
1414 css,
1415 ".{entry} {{\n \
1416 position: absolute;\n \
1417 top: var(--track-at, 0%);\n \
1418 height: var(--track-for, 100%);\n \
1419 left: calc(var(--track-lane, 0) / var(--track-lanes, 1) * 100%);\n \
1420 width: calc(100% / var(--track-lanes, 1));\n\
1421 }}"
1422 );
1423
1424 css
1425}
1426
1427/// A region's stand-in, and the header of a table that can be reordered.
1428///
1429/// Both are 0.12.0 members and both are colour and affordance only, which is
1430/// where `figure_rules` landed after trying to emit a type scale. How much room
1431/// a stand-in gets is a size — goingson has the same one at three, as
1432/// `--compact`, `--dashboard` and `--padded` — and a size is
1433/// `makeover-geometry`'s question.
1434///
1435/// The caret is the one thing here that is neither colour nor affordance, and it
1436/// is a renderer's own expression rather than a value the description named:
1437/// `aria-sort` is what the table actually says, and this turns it into something
1438/// visible for everyone not using a screen reader. A terminal draws its own; an
1439/// immediate-mode painter draws its own.
1440fn state_rules(opts: &Emit) -> String {
1441 let placeholder = class("placeholder", opts);
1442 let text = class("placeholder-text", opts);
1443 let heading = class("table-heading", opts);
1444 let mut css = String::new();
1445
1446 let _ = writeln!(
1447 css,
1448 ".{placeholder} > .{text} {{\n color: var(--content-muted);\n}}"
1449 );
1450 // Only the failure is toned. An empty list is the normal state of a new
1451 // install, and `Readiness::tone` is what says so.
1452 let _ = writeln!(
1453 css,
1454 ".{placeholder}[data-tone=\"{0}\"] > .{text} {{\n color: var(--{0});\n}}",
1455 Tone::Danger.token()
1456 );
1457
1458 // A header that reorders the table is a control, and the pointer is the
1459 // only part of saying so that is not the app's own type and spacing.
1460 let _ = writeln!(
1461 css,
1462 ".{heading}[data-sortable] {{\n cursor: pointer;\n}}"
1463 );
1464 // The caret carries its own leading space, the way `makeover-tui` and
1465 // `makeover-immediate` both write `" \u{25B2}"`. It used to be emitted bare,
1466 // and both apps that adopted the vocabulary had to put the gap back in their
1467 // own stylesheets on the same afternoon -- each having to work out first that
1468 // app CSS outranks this crate's cascade layer, so adding the space the
1469 // obvious way, as `content`, silently wins over the glyph and leaves the
1470 // heading with no caret at all. A consumer should not have to know that, and
1471 // with the space emitted here there is nothing left for one to add.
1472 //
1473 // Three states, three tones, on the convention in wiki
1474 // `three-tone-convention`. A column in force is `content`; a column offering
1475 // to reorder and not doing it now is `content-secondary`, because it still
1476 // answers a press; a column that is not sortable emits no caret at all and
1477 // takes nothing. The idle arm used to hide its glyph and reserve the box,
1478 // which cost a reflow-free press and said nothing. It draws now, and the
1479 // reservation stops being a thing to get right.
1480 //
1481 // The glyph is `Sort::glyph`, escaped rather than written: a CSS `content`
1482 // string cannot carry the character literally through this file's own
1483 // escaping, and spelling it here as well would put the third copy back that
1484 // `makeover-layout` 0.27.5 exists to remove.
1485 let _ = writeln!(
1486 css,
1487 ".{heading}[data-sortable]::after \
1488 {{\n content: \" {}\";\n color: var(--content-secondary);\n}}",
1489 css_escape(Sort::Ascending.glyph())
1490 );
1491 for direction in [Sort::Ascending, Sort::Descending] {
1492 let _ = writeln!(
1493 css,
1494 ".{heading}[aria-sort=\"{}\"]::after \
1495 {{\n content: \" {}\";\n color: var(--content);\n}}",
1496 direction.as_str(),
1497 css_escape(direction.glyph())
1498 );
1499 }
1500 css
1501}
1502
1503/// The frame a table sits in.
1504///
1505/// # Why this is a CSS table and not the grid the rest of the module assumes
1506///
1507/// A grid row needs `grid-template-columns`, which has to name every column in
1508/// order, so it cannot be written without knowing the columns. That is what
1509/// [`list::narrowing_css`] is for, and it works: goingson builds `tables.css` in
1510/// its own `build.rs` out of it, and nothing here changes that.
1511///
1512/// It does not work for a table a *description* produced. Those columns are
1513/// known at render time rather than at build time, and the rules would have to
1514/// travel with the markup: a `<style>` element per table, which needs
1515/// `style-src 'unsafe-inline'` and so blocks the MNW server's standing plan to
1516/// drop it. The head is not an escape: a table swapped in by htmx after a
1517/// delete arrives as a fragment with no head at all.
1518///
1519/// A CSS table aligns its columns across rows knowing nothing about how many
1520/// there are, so there is no track list to emit and nothing per-table to carry.
1521/// The cost is that a described table cannot take a per-column fixed length,
1522/// which costs nothing today: [`Sizing`](list::Sizing) is looked up by column
1523/// name and a description carries no lengths to put in it, so every track a
1524/// described table could ask for is already content, fill or auto.
1525///
1526/// [`Priority`] hiding is unchanged in kind. It moves from a generated
1527/// `display: none` per dropped column to one rule per drop class, which is the
1528/// same fact addressed by class rather than by cutoff, and still never by
1529/// position.
1530fn table_rules(opts: &Emit) -> String {
1531 let table = class("table", opts);
1532 let head = class("table-head", opts);
1533 let row = class("table-row", opts);
1534 let heading = class("table-heading", opts);
1535 let cell = class("cell", opts);
1536 let mut css = String::new();
1537
1538 let _ = writeln!(
1539 css,
1540 ".{table} {{\n display: table;\n width: 100%;\n}}"
1541 );
1542 let _ = writeln!(css, ".{head},\n.{row} {{\n display: table-row;\n}}");
1543 let _ = writeln!(css, ".{heading},\n.{cell} {{\n display: table-cell;\n}}");
1544
1545 // A content column shrinks to what is in it. `width: 1%` is how a CSS table
1546 // is told that: auto layout hands the slack to the columns that asked for
1547 // room, and a column asking for almost none gets what it needs and no more.
1548 // The `nowrap` is what stops it being given less by wrapping.
1549 let _ = writeln!(
1550 css,
1551 ".{} {{\n white-space: nowrap;\n width: 1%;\n}}",
1552 class("cell-content", opts)
1553 );
1554
1555 // A fixed column has no length to be fixed to. The description carries none
1556 // and `Sizing` is not reachable from here, so it behaves as content: the
1557 // honest answer to a width nobody supplied, and the same one `Sizing::track`
1558 // gives it.
1559 let _ = writeln!(
1560 css,
1561 ".{} {{\n white-space: nowrap;\n}}",
1562 class("cell-fixed", opts)
1563 );
1564
1565 // Optional columns go at the narrowest class, secondary ones go with them,
1566 // which is the cutoff walk `kept_at` describes said as two media queries.
1567 // Essential columns have no rule at all, because never dropping is what not
1568 // being mentioned already means.
1569 for (size, drops) in [
1570 (
1571 SizeClass::Compact,
1572 &["cell-drops-first", "cell-drops-next"][..],
1573 ),
1574 (SizeClass::Medium, &["cell-drops-first"][..]),
1575 ] {
1576 let selectors: Vec<String> = drops
1577 .iter()
1578 .map(|drop| format!(".{}", class(drop, opts)))
1579 .collect();
1580 css.push_str(&gated(
1581 Some(&size.media_condition()),
1582 &format!("{} {{\n display: none;\n}}\n", selectors.join(",\n")),
1583 ));
1584 }
1585
1586 // What is inside a cell, which the table side could not say until
1587 // makeover-layout 0.14.0. Every cell was one `.cell` and one content
1588 // colour, so a button in a cell was painted as text -- the drift
1589 // `RowPart::intent` has prevented for list rows since 0.2.0 and prevented
1590 // for nothing here.
1591 //
1592 // The colour goes on `.cell-value` rather than on `.cell`, and that
1593 // placement is the whole fix. On the container it would cascade into the
1594 // tokens and the controls sitting beside the text, which is the bug said
1595 // in one rule; on the part that is text, it reaches text and stops.
1596 for part in [
1597 CellPart::Value,
1598 CellPart::Tokens,
1599 CellPart::Actions,
1600 CellPart::Link,
1601 ] {
1602 // Three of the four inherit, each for its own reason: a token carries
1603 // its own tone, an action is a control rather than text, and a link
1604 // takes the action colour from the anchor it is. `CellPart::intent`
1605 // says so by answering with the intent inheriting already gives, and
1606 // pinning that would be louder than saying nothing.
1607 //
1608 // Written as a skip-list rather than as a match on Value, so a member
1609 // added upstream gets its intent emitted rather than being silently
1610 // dropped. That is the same trade `part_class`'s fallback makes: land
1611 // plainly, never land as nothing.
1612 if !matches!(part, CellPart::Tokens | CellPart::Actions | CellPart::Link) {
1613 let _ = writeln!(
1614 css,
1615 ".{} {{\n color: var(--{});\n}}",
1616 class(cell_part_class(part), opts),
1617 part.intent()
1618 );
1619 }
1620 }
1621
1622 css
1623}
1624
1625/// The component layer: every named thing phase A emits.
1626///
1627/// No scrollbar track. It was on the phase A list and came off: eight lines of
1628/// `::-webkit-scrollbar` with no shape a terminal or an immediate-mode painter
1629/// would want handed to it, so it stays with the apps.
1630#[must_use]
1631pub fn component_rules(opts: &Emit) -> String {
1632 let mut css = String::new();
1633 css.push_str(&surface_rules(opts));
1634 css.push_str(&link_rules(opts));
1635 css.push_str(&token_rules(opts));
1636 css.push_str(&selector_rules(opts));
1637 css.push_str(&row_rules(opts));
1638 css.push_str(&run_rules(opts));
1639 css.push_str(&progress_rules(opts));
1640 css.push_str(&figure_rules(opts));
1641 css.push_str(&picture_rules(opts));
1642 css.push_str(&showing_rules(opts));
1643 css.push_str(&track_rules(opts));
1644 css.push_str(&state_rules(opts));
1645 css.push_str(&table_rules(opts));
1646 css.push_str(&facet::facet_rules(opts));
1647 css
1648}
1649
1650/// The whole phase-A stylesheet: properties, depth rules and components, in
1651/// [`CSS_LAYER`], under a generated-file banner.
1652///
1653/// The banner sits outside the layer, because a comment participates in no
1654/// cascade and a reader opening the file should see what it is before seeing
1655/// an at-rule.
1656#[must_use]
1657pub fn stylesheet(opts: &Emit) -> String {
1658 format!(
1659 "/* Generated by makeover-webview from makeover-layout. Do not edit.\n \
1660 Depth is a fill and an edge together; naming them apart is what let\n \
1661 them disagree. See the crate's README and wiki note makeover-layout.\n\n \
1662 Everything below is in the `{CSS_LAYER}` cascade layer. Declare the\n \
1663 order once in your own stylesheet, or this layer's position is decided\n \
1664 by whichever generated file the browser happens to see first:\n\n \
1665 @layer {CSS_LAYER}, base, components, responsive; */\n{}",
1666 in_css_layer(&format!(
1667 ":root {{\n{}}}\n\n{}\n{}",
1668 bevel_properties(opts),
1669 depth_rules(opts),
1670 component_rules(opts)
1671 ))
1672 )
1673}
1674
1675#[cfg(test)]
1676mod tests {
1677 use super::*;
1678 use makeover_layout::Edge;
1679
1680 #[test]
1681 fn every_fallback_class_is_one_a_checker_knows_about() {
1682 // The obligation ROW_PART_CLASSES carries, for the same reason: a class
1683 // this crate can write and the vocabulary list does not carry is
1684 // invisible to the dead-vocabulary seal and to the overlap check both.
1685 for fallback in [
1686 Fallback::Wrap,
1687 Fallback::Stack,
1688 Fallback::Shed,
1689 Fallback::Menu,
1690 ] {
1691 assert!(
1692 RUN_CLASSES.contains(&fallback_class(fallback)),
1693 "{fallback:?} is missing from RUN_CLASSES"
1694 );
1695 }
1696 let names = crate::vocabulary::names(&Emit::default());
1697 for name in RUN_CLASSES {
1698 assert!(names.contains(*name), "{name} is not in the vocabulary");
1699 }
1700 }
1701
1702 #[test]
1703 fn a_run_gives_every_member_a_floor_it_cannot_be_squeezed_below() {
1704 // The whole of what stops the overlap, and it is not a fallback: it
1705 // applies to every run whatever the group declared. flexbox's default
1706 // min-width is auto, which lets an item be compressed below its own
1707 // content in a nowrap row, and that is how a toolbar is drawn over a
1708 // tab strip even with nothing out of flow.
1709 let css = run_rules(&Emit::default());
1710 assert!(css.contains(".run > * {\n min-width: min-content;\n}"));
1711 // No number anywhere in it. The minimum is derived by the browser from
1712 // what the members contain, which is the ruling's own requirement.
1713 assert!(!css.contains("px"));
1714 assert!(!css.contains("rem"));
1715 assert!(!css.contains("@media"));
1716 }
1717
1718 #[test]
1719 fn room_is_never_asked_of_the_viewport() {
1720 // The 913 case: a window in SizeClass::Expanded holding a group out of
1721 // room. A viewport query answers about the window and would be wrong
1722 // about the group, which is why the table's @media walk is not the
1723 // precedent this follows.
1724 let css = run_rules(&Emit::default());
1725 for size in [SizeClass::Compact, SizeClass::Medium] {
1726 assert!(!css.contains(&size.media_condition()));
1727 }
1728 }
1729
1730 #[test]
1731 fn every_fallback_lands_as_a_class_and_an_unknown_one_lands_plainly() {
1732 let css = run_rules(&Emit::default());
1733 for fallback in [
1734 Fallback::Wrap,
1735 Fallback::Stack,
1736 Fallback::Shed,
1737 Fallback::Menu,
1738 ] {
1739 let class = fallback_class(fallback);
1740 assert!(css.contains(&format!(".{class} {{")), "{class} unemitted");
1741 }
1742 // Stack is the one that also says what a member does with the line it
1743 // took, which is what separates it from wrapping.
1744 assert!(css.contains(".run-stack > * {\n flex: 1 1 max-content;\n}"));
1745 }
1746
1747 #[test]
1748 fn the_emitted_bevel_matches_what_the_apps_already_hand_write() {
1749 // Balanced Breakfast's styles.css, verbatim. Adoption has to be a
1750 // deletion, not a redesign, or nobody will take it.
1751 let opts = Emit::default();
1752 assert_eq!(
1753 bevel_shadow(Bevel::Raised, &opts),
1754 "inset 1px 1px 0 var(--bevel-light), inset -1px -1px 0 var(--bevel-dark)"
1755 );
1756 assert_eq!(
1757 bevel_shadow(Bevel::Inset, &opts),
1758 "inset 1px 1px 0 var(--bevel-dark), inset -1px -1px 0 var(--bevel-light)"
1759 );
1760 }
1761
1762 #[test]
1763 fn no_colour_ever_reaches_the_output() {
1764 let css = stylesheet(&Emit::default());
1765 assert!(!css.contains('#'), "a hex literal escaped into the CSS");
1766 assert!(
1767 !css.contains("rgb"),
1768 "a colour function escaped into the CSS"
1769 );
1770 // Every colour is named, never resolved.
1771 assert!(css.contains("var(--surface-raised)"));
1772 assert!(css.contains("var(--bevel-light)"));
1773 }
1774
1775 #[test]
1776 fn a_well_falls_back_through_css_rather_than_through_rust() {
1777 assert_eq!(
1778 fill_var(Fill::Well),
1779 "var(--surface-well, var(--surface-page))"
1780 );
1781 // Nothing else needs one.
1782 assert_eq!(fill_var(Fill::Raised), "var(--surface-raised)");
1783 assert_eq!(fill_var(Fill::Page), "var(--surface-page)");
1784 }
1785
1786 #[test]
1787 fn raised_and_well_do_not_collapse_onto_each_other() {
1788 let css = depth_rules(&Emit::default());
1789 assert!(css.contains(".raised {"));
1790 assert!(css.contains(".well {"));
1791 assert!(css.contains("var(--bevel-raised)"));
1792 assert!(css.contains("var(--bevel-inset)"));
1793 }
1794
1795 /// The cast shadow is composed here from the tone `makeover` derives, so
1796 /// neither crate has to hold the other's numbers.
1797 ///
1798 /// It is a `:root` property and deliberately not a depth class. There is no
1799 /// `Depth::Overlay` in the description layer, and adding one would be a
1800 /// claim about what a screen means rather than about how it is painted;
1801 /// until something asks for it, a consumer names the property on the rule
1802 /// for the menu or the toast it already has.
1803 #[test]
1804 fn the_cast_shadow_is_a_root_property_not_a_depth() {
1805 let css = bevel_properties(&Emit::default());
1806 assert!(css.contains("--elevation-overlay:"));
1807 assert!(css.contains("var(--elevation)"));
1808 assert!(
1809 !depth_rules(&Emit::default()).contains("elevation"),
1810 "elevation is not a depth class"
1811 );
1812 }
1813
1814 #[test]
1815 fn the_cascade_carries_the_pressed_state() {
1816 let css = surface_rules(&Emit::default());
1817 // The one thing this renderer gets free that the other two resolve by
1818 // hand, eighteen call sites deep in audiofiles' case. Asserted on a
1819 // named surface: pressing belongs to the control, not to the depth.
1820 assert!(css.contains(".card:active {"));
1821 assert!(css.contains(".button:active {"));
1822 }
1823
1824 #[test]
1825 fn the_depth_class_is_a_surface_and_not_a_control() {
1826 let css = depth_rules(&Emit::default());
1827 // The static surface the vocabulary was missing. Sixteen goingson
1828 // elements wore .card and cancelled its hover and press to get this,
1829 // because a raised object that is not pressable had no other spelling.
1830 for state in [":hover", ":active", ":focus-visible", ":disabled"] {
1831 assert!(
1832 !css.contains(&format!(".raised{state}")),
1833 "the depth class claimed {state}: {css}"
1834 );
1835 }
1836 assert!(css.contains("var(--bevel-raised)"), "still raised: {css}");
1837 }
1838
1839 #[test]
1840 fn pressing_moves_the_fill_and_not_only_the_edge() {
1841 // The decision-1 guard, and the regression that mattered: emitting the
1842 // bevel flip alone is what left goingson hand-writing `background:
1843 // var(--surface-sunken)` on .btn, .card and .tag/.badge alike, so none
1844 // of the three could be deleted.
1845 let pressed = interactive_rules("button", Depth::Raised, &Emit::default());
1846 assert!(pressed.contains(".button:active {"));
1847 assert!(
1848 pressed.contains("background: var(--surface-well, var(--surface-page))"),
1849 "pressed dropped its fill: {pressed}"
1850 );
1851 assert!(pressed.contains("box-shadow: var(--bevel-inset)"));
1852 }
1853
1854 #[test]
1855 fn pressed_takes_its_fill_from_the_description_not_from_the_app() {
1856 // goingson presses to --surface-sunken. The description says a pressed
1857 // raised region reads as a well, and makeover says outright that
1858 // surface-sunken cannot serve as one, so the app is the thing that
1859 // moves.
1860 //
1861 // Scoped to the pressed rules rather than to the whole sheet: since
1862 // makeover-layout 0.3.0 an unchosen tab is legitimately
1863 // --surface-sunken, so the token appearing somewhere in the output no
1864 // longer means the app's choice leaked in.
1865 let css = stylesheet(&Emit::default());
1866 let mut checked = 0;
1867 for rule in css.split("}\n") {
1868 if !rule.contains(":active") {
1869 continue;
1870 }
1871 checked += 1;
1872 assert!(
1873 !rule.contains("surface-sunken"),
1874 "a pressed rule took the app's fill: {rule}"
1875 );
1876 }
1877 assert!(checked > 0, "no pressed rules found to check");
1878 assert_eq!(
1879 Depth::Raised.pressed().fill(),
1880 Some(Fill::Well),
1881 "the description changed under us"
1882 );
1883 }
1884
1885 #[test]
1886 fn the_whole_stylesheet_is_emitted_in_the_family_layer() {
1887 // The point of 0.11.0. Unlayered normal declarations outrank every
1888 // named layer, so an app declaring `@layer base, components` loses
1889 // every rule it owns to this file until this file is layered too.
1890 let css = stylesheet(&Emit::default());
1891 assert!(css.contains(&format!("@layer {CSS_LAYER} {{")));
1892
1893 // Exactly one layer block, and nothing outside it but the banner.
1894 assert_eq!(css.matches("@layer").count(), 2, "banner names it once");
1895 let opened = css.find("@layer makeover {").expect("layer opens");
1896 for (i, line) in css.lines().enumerate() {
1897 let before_layer = css.lines().take(i).map(str::len).sum::<usize>() < opened;
1898 if before_layer || line.is_empty() {
1899 continue;
1900 }
1901 assert!(
1902 line.starts_with(" ") || line == "}" || line.starts_with(" "),
1903 "line outside the layer: {line:?}"
1904 );
1905 }
1906 }
1907
1908 #[test]
1909 fn the_generated_sheet_carries_no_trailing_whitespace() {
1910 // A checked-in generated file that a formatter wants to rewrite is a
1911 // diff every time somebody saves it.
1912 let css = stylesheet(&Emit::default());
1913 for (i, line) in css.lines().enumerate() {
1914 assert_eq!(line, line.trim_end(), "trailing whitespace on line {i}");
1915 }
1916 }
1917
1918 #[test]
1919 fn the_banner_tells_an_app_how_to_order_the_layer() {
1920 // Without a declared order the layer's position depends on which
1921 // generated file the browser sees first, which is not a contract.
1922 let css = stylesheet(&Emit::default());
1923 assert!(css.contains("@layer makeover, base, components, responsive;"));
1924 // And the banner is outside the layer, not a rule inside it.
1925 assert!(css.starts_with("/* Generated by makeover-webview"));
1926 }
1927
1928 #[test]
1929 fn a_primitive_owns_every_state_it_implies() {
1930 // The whole point of 0.10.0. Anything emitting a hover rule owes the
1931 // other three, or the consuming app supplies them by out-specifying a
1932 // rule it does not own: 19 such rules in goingson, 21 in the MNW
1933 // server, and three focus rings that do not match.
1934 let css = stylesheet(&Emit::default());
1935 for selector in ["button", "card", "chip", "tab", "segment", "toggle"] {
1936 assert!(css.contains(&format!(".{selector}:hover {{")), "{selector}");
1937 assert!(
1938 css.contains(&format!(".{selector}:active {{")),
1939 "{selector}"
1940 );
1941 assert!(
1942 css.contains(&format!(".{selector}:focus-visible {{")),
1943 "{selector} has no focus ring"
1944 );
1945 assert!(
1946 css.contains(&format!(".{selector}:disabled,")),
1947 "{selector} has no disabled state"
1948 );
1949 }
1950 }
1951
1952 #[test]
1953 fn a_field_takes_focus_and_refuses_input_without_taking_a_hover() {
1954 // A text field does not light up under the pointer, so it gets the two
1955 // states it has and not the two it does not.
1956 let css = stylesheet(&Emit::default());
1957 assert!(css.contains(".field:focus-visible {"));
1958 assert!(css.contains(".field:disabled,"));
1959 assert!(!css.contains(".field:hover {"));
1960 assert!(!css.contains(".field:active {"));
1961 }
1962
1963 #[test]
1964 fn disabled_is_emitted_after_hover_so_source_order_settles_it() {
1965 // Every one of these selectors is specificity (0,2,0), so nothing but
1966 // order decides which wins. A disabled button taking the hover fill is
1967 // the exact bug goingson's `.button:disabled:hover` was written to fix,
1968 // and the reason it had to reach (0,3,0) to do it.
1969 let css = interactive_rules("button", Depth::Raised, &Emit::default());
1970 let hover = css.find(":hover").expect("hover");
1971 let active = css.find(":active").expect("active");
1972 let focus = css.find(":focus-visible").expect("focus");
1973 let disabled = css.find(":disabled").expect("disabled");
1974 assert!(hover < active && active < focus && focus < disabled);
1975
1976 // And it restores the surface, or the hover fill survives underneath.
1977 let tail = &css[disabled..];
1978 assert!(tail.contains("background: var(--surface-raised)"));
1979 }
1980
1981 #[test]
1982 fn a_disabled_state_reaches_things_that_cannot_be_disabled() {
1983 // `:disabled` matches form elements only, and a chip is a div. Keying
1984 // on the ARIA attribute too is the pattern the invalid field already
1985 // set: one fact, read by the styling and the accessibility tree alike.
1986 let css = disabled_rule("chip", Depth::Raised);
1987 assert!(css.contains(".chip:disabled,"));
1988 assert!(css.contains(".chip[aria-disabled=\"true\"]"));
1989 assert!(css.contains("cursor: not-allowed"));
1990 }
1991
1992 #[test]
1993 fn the_focus_ring_does_not_disturb_the_bevel_it_lands_on() {
1994 // `outline` has its own property, so unlike the invalid ring there is
1995 // no bevel to restate beside it and nothing to keep in agreement.
1996 let opts = Emit::default();
1997 let css = focus_rule("button", Depth::Raised, &opts);
1998 assert!(css.contains("outline: 2px solid var(--focus-ring)"));
1999 assert!(!css.contains("box-shadow"), "the ring restated the bevel");
2000 }
2001
2002 #[test]
2003 fn a_well_takes_the_ring_inside_and_a_raised_surface_outside() {
2004 // One ring, placed by depth. The offset comes off `Depth::bevel` and
2005 // not off a per-component choice, which is what gave three apps three
2006 // different rings.
2007 let opts = Emit::default();
2008 assert!(focus_rule("field", Depth::Well, &opts).contains("outline-offset: calc(-1 * 2px)"));
2009 assert!(focus_rule("button", Depth::Raised, &opts).contains("outline-offset: 2px"));
2010 // Nothing to sit inside of, so it sits outside.
2011 assert!(focus_rule("badge", Depth::Sunken, &opts).contains("outline-offset: 2px"));
2012
2013 // And the ring is not the bevel. Reusing border_width emitted a 1px
2014 // ring that every consumer had already overridden.
2015 assert_ne!(opts.focus_width, opts.border_width);
2016 }
2017
2018 #[test]
2019 fn hover_is_gated_on_capability_and_the_keyboard_path_is_not() {
2020 // goingson's section 60 exists only to take back the hover state this
2021 // crate handed it. Gating at the source is what deletes that section
2022 // in all three apps rather than having each fight for it.
2023 let css = stylesheet(&Emit::default());
2024 let condition = format!("@media {}", Density::Pointer.media_condition());
2025 assert!(css.contains(&condition));
2026
2027 // What is gated is every hover state the surfaces carry. The row's
2028 // actions used to be the other half of this test and are not gated any
2029 // more, because they are not hidden any more: a rule that reveals
2030 // nothing needs no capability answer.
2031 let gated: Vec<&str> = css.lines().filter(|line| line.contains(":hover")).collect();
2032 assert!(!gated.is_empty(), "{css}");
2033 for line in gated {
2034 let indent = line.len() - line.trim_start().len();
2035 assert!(indent > 4, "an ungated hover rule: {line}");
2036 }
2037 assert!(!css.contains(".row:hover"), "{css}");
2038 }
2039
2040 #[test]
2041 fn the_capability_answer_is_asked_for_and_not_assumed() {
2042 // Both halves come from the crates that own them. If `makeover-touch`
2043 // ever says a fingertip has hover, this stops gating on its own.
2044 assert!(!Affordance::Hover.available(Density::Touch, SizeClass::Compact));
2045 assert!(Affordance::Hover.available(Density::Pointer, SizeClass::Compact));
2046 assert_eq!(hover_condition(), Some(Density::Pointer.media_condition()));
2047
2048 // And the size class passed to that call is not a claim about width.
2049 assert!(Affordance::Hover.reads_density());
2050 for size in [SizeClass::Compact, SizeClass::Medium, SizeClass::Expanded] {
2051 assert!(!Affordance::Hover.available(Density::Touch, size));
2052 }
2053 }
2054
2055 #[test]
2056 fn hover_resolves_against_the_token_makeover_already_derives() {
2057 let css = interactive_rules("card", Depth::Raised, &Emit::default());
2058 assert!(css.contains(".card:hover {"));
2059 assert!(css.contains("background: var(--hover-surface)"));
2060 // Not the app's choice, which was --surface-overlay.
2061 assert!(!css.contains("surface-overlay"));
2062 }
2063
2064 #[test]
2065 fn a_badge_gets_no_edge_and_no_fill() {
2066 // Decision 2, and the one visible redesign in phase A. Token::Badge is
2067 // Flat: an edge on a label says it can be pressed.
2068 let css = token_rules(&Emit::default());
2069 let badge = css
2070 .lines()
2071 .skip_while(|l| !l.starts_with(".badge {"))
2072 .take_while(|l| !l.starts_with('}'))
2073 .collect::<Vec<_>>()
2074 .join("\n");
2075 assert!(!badge.contains("box-shadow"), "badge kept an edge: {badge}");
2076 assert!(!badge.contains("background"), "badge kept a fill: {badge}");
2077 assert_eq!(Token::Badge.depth(false), Depth::Flat);
2078 assert_eq!(Token::Badge.depth(true), Depth::Flat);
2079 }
2080
2081 #[test]
2082 fn a_badge_carries_a_tone_and_neutral_is_the_bare_class() {
2083 let css = token_rules(&Emit::default());
2084 // Neutral is the absence of a status, not a status named "none".
2085 assert!(css.contains(".badge {\n color: var(--content-muted);"));
2086 assert!(!css.contains("data-tone=\"content-muted\""));
2087 for tone in ["info", "success", "warning", "danger"] {
2088 assert!(
2089 css.contains(&format!(".badge[data-tone=\"{tone}\"]")),
2090 "missing tone {tone}"
2091 );
2092 assert!(css.contains(&format!("color: var(--{tone})")));
2093 }
2094 }
2095
2096 #[test]
2097 fn a_chip_is_raised_and_latches_into_a_well() {
2098 let css = token_rules(&Emit::default());
2099 assert!(css.contains(".chip {"));
2100 assert!(css.contains(".chip.latched {"));
2101 assert!(css.contains(".chip:active {"));
2102 // The whole difference from a badge: it answers a click.
2103 assert!(Token::Chip { removable: false }.interactive());
2104 assert!(!Token::Badge.interactive());
2105 }
2106
2107 #[test]
2108 fn only_a_tab_comes_forward_when_chosen() {
2109 // The folder semantic. Collapsing the three selectors would lose it.
2110 let css = selector_rules(&Emit::default());
2111 assert!(css.contains(".tab.chosen {"));
2112 assert!(css.contains(".segment.chosen {"));
2113 assert!(css.contains(".toggle.chosen {"));
2114 assert_eq!(Selector::Tabs.chosen(), Depth::Raised);
2115 assert_eq!(Selector::Segmented.chosen(), Depth::Well);
2116 assert_eq!(Selector::Toggle.chosen(), Depth::Well);
2117
2118 let tab = css
2119 .lines()
2120 .skip_while(|l| !l.starts_with(".tab.chosen {"))
2121 .take_while(|l| !l.starts_with('}'))
2122 .collect::<Vec<_>>()
2123 .join("\n");
2124 assert!(
2125 tab.contains("var(--bevel-raised)"),
2126 "tab was held in: {tab}"
2127 );
2128 }
2129
2130 #[test]
2131 fn an_unchosen_tab_recedes_without_looking_picked() {
2132 let css = selector_rules(&Emit::default());
2133 // Recessed by colour and given no edge. An edge would make every option
2134 // look picked; flat would leave the chosen one nothing to come forward
2135 // from, which is the gap makeover-layout 0.3.0 closed.
2136 assert!(
2137 css.contains(".tab {\n background: var(--surface-sunken);\n}"),
2138 "unchosen tab is not recessed: {css}"
2139 );
2140 assert_eq!(Selector::Tabs.unchosen(), Depth::Sunken);
2141 assert!(css.contains(".tab:hover {"));
2142 }
2143
2144 #[test]
2145 fn a_segment_stands_up_so_the_chosen_one_can_be_held_in() {
2146 // The inverse of the tab, and why the three selectors are not one
2147 // member with a flag.
2148 let css = selector_rules(&Emit::default());
2149 assert!(css.contains(".segment {\n background: var(--surface-raised);"));
2150 assert_eq!(Selector::Segmented.unchosen(), Depth::Raised);
2151 assert_eq!(Selector::Segmented.chosen(), Depth::Well);
2152 }
2153
2154 #[test]
2155 fn a_rows_actions_are_shown_at_rest() {
2156 let css = row_rules(&Emit::default());
2157
2158 // The hover reveal is gone, and with it every escape it needed. What
2159 // it hid was hidden from pointer users alone, who are the ones
2160 // scanning a list to learn what can be done to a row.
2161 assert!(!css.contains("opacity"), "{css}");
2162 assert!(!css.contains("pointer-events"), "{css}");
2163 assert!(!css.contains(":hover"), "{css}");
2164 assert!(!css.contains(":focus-within"), "{css}");
2165
2166 // Nor is it hidden any other way. `display: none` would reflow the row
2167 // and `visibility: hidden` would take the actions out of the focus
2168 // order; the point is that neither is reached for.
2169 assert!(!css.contains("display: none"), "{css}");
2170 assert!(!css.contains("visibility:"), "{css}");
2171 }
2172
2173 #[test]
2174 fn a_figures_tone_lands_on_the_delta_when_there_is_one() {
2175 // 0.13.0. The delta is the part that reads as good or bad; the number
2176 // itself is an ordinary fact. A figure with no delta has nowhere else to
2177 // put the colour, so the value takes it, and `:has` is what lets one
2178 // attribute mean both without the emitter choosing an element.
2179 let css = stylesheet(&Emit::default());
2180
2181 assert!(
2182 css.contains(".figure[data-tone=\"success\"] > .figure-change"),
2183 "{css}"
2184 );
2185 assert!(
2186 css.contains(
2187 ".figure[data-tone=\"success\"]:not(:has(> .figure-change)) > .figure-value"
2188 ),
2189 "{css}"
2190 );
2191 // The caption is the noun and never takes the tone.
2192 assert!(
2193 !css.contains("[data-tone=\"success\"] > .figure-caption"),
2194 "{css}"
2195 );
2196 }
2197
2198 #[test]
2199 fn the_three_text_parts_take_their_intents_and_actions_inherits() {
2200 let css = row_rules(&Emit::default());
2201 assert!(css.contains(".row-primary {\n color: var(--content);"));
2202 assert!(css.contains(".row-secondary {\n color: var(--content-secondary);"));
2203 assert!(css.contains(".row-meta {\n color: var(--content-muted);"));
2204 // Actions carry controls, not text. Pinning the colour it would inherit
2205 // anyway is louder than saying nothing.
2206 assert!(!css.contains(".row-actions {\n color:"));
2207 }
2208
2209 #[test]
2210 fn the_token_strip_takes_no_colour_of_its_own() {
2211 // makeover-layout 0.9.0. A token carries its own tone, so a colour on
2212 // the strip would be a rule fighting the things sitting in it -- the
2213 // same reasoning as actions, reached for a different reason.
2214 let css = row_rules(&Emit::default());
2215 assert!(!css.contains(".row-tokens {\n color:"));
2216 }
2217
2218 #[test]
2219 fn an_unknown_row_part_renders_plainly_rather_than_failing_to_build() {
2220 // What `#[non_exhaustive]` bought and what it cost. `part_class` can no
2221 // longer be exhaustive, so a member added upstream lands as a bare
2222 // class with no rule instead of stopping the build. Asserting the
2223 // fallback exists is what keeps it from being written as `unreachable!`
2224 // by someone who reads the match as closed.
2225 assert_eq!(part_class(RowPart::Tokens), "row-tokens");
2226 assert_eq!(part_class(RowPart::Meta), "row-meta");
2227 }
2228
2229 #[test]
2230 fn a_link_takes_the_action_colour_the_theme_actually_defines() {
2231 // `--action-primary` shipped here for months and no theme has ever
2232 // defined it, so every `.link` dropped its colour declaration outright
2233 // and fell back to inherited text. Nothing caught it because the sheet
2234 // is valid CSS either way; MNW's no-undefined-token lint is what found
2235 // it, 2026-08-14. The hover arm two lines below was always `--action-hover`,
2236 // which is what makes the typo legible in hindsight.
2237 let css = link_rules(&Emit::default());
2238 assert!(css.contains("color: var(--action);"));
2239 assert!(!css.contains("--action-primary"));
2240 assert!(css.contains("color: var(--action-hover);"));
2241 }
2242
2243 #[test]
2244 fn the_progress_trough_is_a_well() {
2245 let css = progress_rules(&Emit::default());
2246 assert!(css.contains(".progress {"));
2247 assert!(css.contains("box-shadow: var(--bevel-inset)"));
2248 assert!(css.contains(".progress > .progress-fill {"));
2249 assert!(css.contains("background: var(--action)"));
2250 // A bare `.fill` would catch things that have nothing to do with
2251 // progress once the sheet lands unprefixed.
2252 assert!(!css.contains("> .fill "));
2253 }
2254
2255 #[test]
2256 fn a_progress_bar_can_carry_a_tone_and_defaults_to_action() {
2257 let css = progress_rules(&Emit::default());
2258 // Untoned is --action, not Tone::Neutral's content-muted: a bar with no
2259 // status is still reporting progress, and muted would read as disabled.
2260 assert!(css.contains(".progress > .progress-fill {\n background: var(--action);"));
2261 assert!(!css.contains("progress-fill {\n color: var(--content-muted)"));
2262 for tone in ["info", "success", "warning", "danger"] {
2263 assert!(
2264 css.contains(&format!(".progress > .progress-fill[data-tone=\"{tone}\"]")),
2265 "missing progress tone {tone}"
2266 );
2267 }
2268 // goingson's two live cases, which is why the tones are emitted at all.
2269 assert!(css.contains("[data-tone=\"success\"] {\n background: var(--success);"));
2270 assert!(css.contains("[data-tone=\"danger\"] {\n background: var(--danger);"));
2271 }
2272
2273 #[test]
2274 fn no_scrollbar_track_is_emitted() {
2275 // Decision 3's negative half. It was on the phase A list and came off;
2276 // this is what stops it drifting back in.
2277 let css = stylesheet(&Emit::default());
2278 assert!(!css.contains("scrollbar"));
2279 assert!(!css.contains("::-webkit"));
2280 }
2281
2282 #[test]
2283 fn an_invalid_field_is_ringed_without_being_lit() {
2284 let css = surface_rules(&Emit::default());
2285 assert!(css.contains(".field {"));
2286 // The ARIA attribute, not a class: one fact, read by both the visual
2287 // and the accessible state, so they cannot drift.
2288 assert!(css.contains(".field[aria-invalid=\"true\"] {"));
2289 assert!(!css.contains(".field.invalid"));
2290 // A flat ring: this edge says "wrong", and a two-tone bevel would have
2291 // it say "raised" at the same time.
2292 assert!(css.contains("0 0 0 1px var(--danger)"));
2293 }
2294
2295 #[test]
2296 fn an_invalid_field_keeps_the_well_underneath_it() {
2297 // box-shadow is not additive. A lone ring replaces the bevel and drops
2298 // the well out from under the field, which is what this emitted before
2299 // 0.5.0 and is the whole reason the rule composes.
2300 let css = surface_rules(&Emit::default());
2301 let invalid = css
2302 .lines()
2303 .skip_while(|l| !l.starts_with(".field[aria-invalid"))
2304 .take_while(|l| !l.starts_with('}'))
2305 .collect::<Vec<_>>()
2306 .join("\n");
2307 assert!(
2308 invalid.contains("var(--bevel-inset)"),
2309 "the well was dropped: {invalid}"
2310 );
2311 assert!(invalid.contains("var(--danger)"));
2312 }
2313
2314 #[test]
2315 fn button_and_card_come_out_identical_by_construction() {
2316 // The duplication phase A deletes. They are the same composition, so
2317 // the only honest way to emit both is from one call.
2318 let opts = Emit::default();
2319 let css = surface_rules(&opts);
2320 assert_eq!(
2321 depth_declarations(Depth::Raised),
2322 depth_declarations(Depth::Raised)
2323 );
2324 assert!(css.contains(".button {"));
2325 assert!(css.contains(".card {"));
2326 assert_eq!(
2327 interactive_rules("button", Depth::Raised, &Emit::default()).replace("button", "card"),
2328 interactive_rules("card", Depth::Raised, &Emit::default())
2329 );
2330 }
2331
2332 #[test]
2333 fn a_prefix_reaches_the_component_classes_too() {
2334 let opts = Emit {
2335 class_prefix: "mo-",
2336 ..Emit::default()
2337 };
2338 let css = stylesheet(&opts);
2339 for name in [
2340 "mo-button",
2341 "mo-card",
2342 "mo-field",
2343 "mo-badge",
2344 "mo-chip",
2345 "mo-tab",
2346 "mo-row-primary",
2347 "mo-progress",
2348 "mo-progress-fill",
2349 ] {
2350 assert!(css.contains(&format!(".{name}")), "unprefixed: {name}");
2351 }
2352 // The bare names must be gone entirely, or a prefixed build still
2353 // collides with the app's own stylesheet.
2354 assert!(!css.contains(".button {"));
2355 assert!(!css.contains(".card {"));
2356 assert!(!css.contains(".badge {"));
2357 }
2358
2359 #[test]
2360 fn the_class_a_renderer_puts_on_an_option_is_the_one_the_rules_key_off() {
2361 // `option_class` is the contract a screen renderer writes markup
2362 // against, and the rules below are the other half of it. They come off
2363 // one mapping now, so this asserts the mapping is the one that reaches
2364 // the stylesheet rather than that two lists still agree.
2365 let css = stylesheet(&Emit::default());
2366 for selector in [Selector::Tabs, Selector::Segmented, Selector::Toggle] {
2367 let name = option_class(selector);
2368 assert!(css.contains(&format!(".{name} {{")), "{name}: {css}");
2369 assert!(css.contains(&format!(".{name}.chosen {{")), "{name}: {css}");
2370 }
2371 assert_eq!(option_class(Selector::Tabs), "tab");
2372 }
2373
2374 #[test]
2375 fn the_caret_brings_its_own_gap_and_is_the_glyph_the_description_names() {
2376 let css = stylesheet(&Emit::default());
2377
2378 // The space is inside the glyph, which is what the other two renderers
2379 // write. Emitted bare, every consumer has to add it back, and the
2380 // obvious way to add it -- `content` in an app stylesheet, which is
2381 // unlayered and so outranks this sheet -- deletes the caret instead.
2382 assert!(css.contains("content: \" \\25B2\";"), "{css}");
2383 assert!(css.contains("content: \" \\25BC\";"), "{css}");
2384 assert!(!css.contains("content: \"\\2"), "{css}");
2385
2386 // The arrows this renderer used to draw alone are gone. Composition
2387 // rather than agreement: the glyph comes from `Sort::glyph`, so a
2388 // fourth spelling cannot appear here without appearing everywhere.
2389 assert!(!css.contains("2191") && !css.contains("2193"), "{css}");
2390 assert!(css.contains(&css_escape(Sort::Ascending.glyph())), "{css}");
2391
2392 // Three states, three tones. An idle sortable heading draws its caret
2393 // now rather than reserving a hidden box for it, so there is no
2394 // visibility to order and no reflow left to guard against; what
2395 // separates the states is the colour, and the sorted arms come after
2396 // the idle one because the specificity is the same.
2397 let idle = css
2398 .find(".table-heading[data-sortable]::after")
2399 .expect("the idle caret is emitted");
2400 let sorted = css
2401 .find(".table-heading[aria-sort=\"ascending\"]::after")
2402 .expect("the ascending caret is emitted");
2403 assert!(idle < sorted, "{css}");
2404 assert!(
2405 css[idle..sorted].contains("color: var(--content-secondary);"),
2406 "{css}"
2407 );
2408 assert!(css[sorted..].contains("color: var(--content);"), "{css}");
2409 assert!(!css.contains("visibility: hidden;"), "{css}");
2410 }
2411
2412 #[test]
2413 fn a_destructive_button_has_somewhere_for_its_tone_to_land() {
2414 let css = component_rules(&Emit::default());
2415
2416 for tone in [Tone::Info, Tone::Success, Tone::Warning, Tone::Danger] {
2417 assert!(
2418 css.contains(&format!(".button[data-tone=\"{}\"]", tone.token())),
2419 "{css}"
2420 );
2421 }
2422 // Colour, not a fill. A red surface is an app's decision about emphasis.
2423 assert!(!css.contains(".button[data-tone=\"danger\"] {\n background"));
2424 }
2425
2426 #[test]
2427 fn a_described_list_is_not_a_bulleted_list() {
2428 let css = component_rules(&Emit::default());
2429 assert!(css.contains(".list {\n list-style: none;"), "{css}");
2430 }
2431
2432 #[test]
2433 fn a_table_lays_itself_out_without_being_told_its_columns() {
2434 // The whole point of the CSS table. A described table's columns are
2435 // known at render time, so anything the stylesheet has to be told about
2436 // them would have to travel with the markup.
2437 let css = component_rules(&Emit::default());
2438
2439 assert!(css.contains(".table {\n display: table;"), "{css}");
2440 assert!(css.contains("display: table-row;"), "{css}");
2441 assert!(css.contains("display: table-cell;"), "{css}");
2442 assert!(!css.contains("grid-template-columns"), "{css}");
2443 }
2444
2445 #[test]
2446 fn a_control_in_a_cell_is_not_painted_as_text() {
2447 // The point of makeover-layout 0.14.0's CellPart, and the table-side
2448 // twin of `the_three_text_parts_take_their_intents_and_actions_inherits`
2449 // above. One `.cell` and one content colour meant a button in a cell
2450 // inherited it.
2451 let css = table_rules(&Emit::default());
2452
2453 assert!(
2454 css.contains(".cell-value {\n color: var(--content);"),
2455 "{css}"
2456 );
2457 assert!(!css.contains(".cell-actions {\n color:"), "{css}");
2458 assert!(!css.contains(".cell-tokens {\n color:"), "{css}");
2459 assert!(!css.contains(".cell-link {\n color:"), "{css}");
2460
2461 // The colour is on the part that is text, never on the container. On
2462 // `.cell` it would cascade into the three parts that are not text,
2463 // which is the bug written as one rule.
2464 assert!(!css.contains(".cell {\n color:"), "{css}");
2465 }
2466
2467 #[test]
2468 fn an_unknown_cell_part_renders_plainly_rather_than_failing_to_build() {
2469 // `part_class`'s obligation, taken on for the table side too. CellPart
2470 // is `#[non_exhaustive]`, so a member added upstream must land as a
2471 // bare class rather than as a build that stops.
2472 assert_eq!(cell_part_class(CellPart::Value), "cell-value");
2473 assert_eq!(cell_part_class(CellPart::Actions), "cell-actions");
2474 }
2475
2476 #[test]
2477 fn a_column_drops_by_its_priority_and_never_by_its_position() {
2478 let css = component_rules(&Emit::default());
2479
2480 // Optional goes at the narrowest class and secondary goes with it,
2481 // which is `kept_at`'s cutoff walk said as two queries.
2482 let compact = css
2483 .find(&format!("@media {}", SizeClass::Compact.media_condition()))
2484 .expect("a compact query");
2485 let medium = css
2486 .find(&format!("@media {}", SizeClass::Medium.media_condition()))
2487 .expect("a medium query");
2488 assert!(css[compact..].contains(".cell-drops-next"), "{css}");
2489 assert!(!css[medium..].contains(".cell-drops-next"), "{css}");
2490
2491 // Essential columns are never mentioned, because not being mentioned is
2492 // already what never dropping means.
2493 assert!(!css.contains(".cell-keeps"), "{css}");
2494
2495 // And nothing counts. `nth-child` is the bug the priority vocabulary
2496 // exists to end.
2497 assert!(!css.contains("nth-child"), "{css}");
2498 }
2499
2500 #[test]
2501 fn a_track_places_by_custom_property_and_never_by_a_size() {
2502 let css = track_rules(&Emit::default());
2503
2504 // Placement arrives from the caller, computed once by Track::fraction.
2505 // If either of these becomes a literal, three renderers have started
2506 // disagreeing about where 09:30 is.
2507 assert!(css.contains("top: var(--track-at"), "{css}");
2508 assert!(css.contains("height: var(--track-for"), "{css}");
2509
2510 // Overlap lanes default so an entry naming neither is full width.
2511 assert!(css.contains("--track-lane, 0"), "{css}");
2512 assert!(css.contains("--track-lanes, 1"), "{css}");
2513
2514 // The refusal that matters. A slot height here would be this crate
2515 // deciding how tall a quarter of an hour is, which is the thing
2516 // makeover-geometry owns and the reason `.track` gets no height at all.
2517 assert!(!css.contains("height: var(--track-slot"), "{css}");
2518 for size in ["px", "rem", "em", "vh"] {
2519 let bare = css
2520 .lines()
2521 .filter(|l| !l.contains("var(--"))
2522 .any(|l| l.contains(size));
2523 assert!(!bare, "track_rules named a {size} outside a var(): {css}");
2524 }
2525 }
2526
2527 #[test]
2528 fn a_relaxed_part_clamps_and_a_tight_one_says_nothing() {
2529 let css = stylesheet(&Emit::default());
2530 assert!(css.contains(".row-relaxed {"));
2531 assert!(css.contains("-webkit-line-clamp: 2;"));
2532 // The count is `Flow`'s, not this crate's. If the tier ever means three
2533 // lines, this fails here rather than in an app.
2534 assert!(css.contains(&format!("line-clamp: {};", Flow::Relaxed.lines())));
2535 // Tight gets no rule at all: one line is what a run already does, and a
2536 // class per part saying so is a declaration that changes nothing.
2537 assert!(!css.contains("row-tight"));
2538 assert_eq!(crate::list::flow_class(Flow::Relaxed), Some("row-relaxed"));
2539 assert_eq!(crate::list::flow_class(Flow::Tight), None);
2540 // Emitted, therefore checkable: an app's dead-vocabulary seal and the
2541 // overlap check both read `vocabulary::names`, so a class the renderer
2542 // can write and that list does not carry is invisible to both.
2543 assert!(crate::vocabulary::names(&Emit::default()).contains("row-relaxed"));
2544 }
2545
2546 #[test]
2547 fn the_whole_sheet_still_names_every_colour() {
2548 // The crate's founding property, asserted over the component layer and
2549 // not only the primitives.
2550 let css = stylesheet(&Emit::default());
2551 assert!(!css.contains('#'));
2552 assert!(!css.contains("rgb"));
2553 for line in css.lines() {
2554 // Declarations only: a selector or an at-rule can carry a colon of
2555 // its own (`:root`, `:hover`, `@media (hover: hover)`) and declares
2556 // nothing. Keyed on the trailing semicolon rather than on leading
2557 // indentation, which only ever worked as a proxy for nesting depth
2558 // and stopped when the sheet gained a cascade layer around it.
2559 let trimmed = line.trim();
2560 if !trimmed.ends_with(';') {
2561 continue;
2562 }
2563 let Some((_, value)) = trimmed.split_once(": ") else {
2564 continue;
2565 };
2566 if value.contains("var(--") {
2567 continue;
2568 }
2569 // Everything left has to be a keyword, a number or a
2570 // caller-supplied length, never a colour.
2571 //
2572 // The length arm is what the comment above always claimed and the
2573 // list never covered: `border_width` arrives from `Emit` and lands
2574 // bare in the focus ring's offset, where the bevel had only ever
2575 // used it inside an `inset` shadow.
2576 let opts = Emit::default();
2577 assert!(
2578 value.contains("inset")
2579 || value.contains(opts.border_width)
2580 || value.contains(opts.focus_width)
2581 // 0.12.0's two: a sortable header is a control and says so
2582 // with the pointer, and the caret is this renderer's own
2583 // expression of `aria-sort`. Neither is a colour, which is
2584 // what this test is actually about, and neither is a size,
2585 // which is the other thing this crate must not name. The
2586 // leading space inside the glyph is the same thing the other
2587 // two renderers write into theirs, so it is part of the
2588 // caret rather than spacing this crate decided on.
2589 || value
2590 .trim_start_matches('"')
2591 .trim_start()
2592 .starts_with("\\2")
2593 || matches!(
2594 value.trim_end_matches(';'),
2595 "0" | "1"
2596 | "none"
2597 | "auto"
2598 | "not-allowed"
2599 | "pointer"
2600 // The caret's reserved box. Visibility is presence,
2601 // not magnitude and not colour.
2602 | "hidden"
2603 | "visible"
2604 // The link's two signals. `underline` is a line and
2605 // `inherit` defers to whatever the app set, so
2606 // neither names a colour or a magnitude.
2607 | "underline"
2608 | "inherit"
2609 // The table frame. `display` is structure and not a
2610 // size; `nowrap` is what makes a content column
2611 // content. The two widths are the awkward pair and
2612 // they are still not sizes: `100%` is "all of
2613 // whatever you were given" and `1%` is the CSS
2614 // table idiom for "shrink to fit", which is a
2615 // behaviour spelled as a number because CSS has no
2616 // keyword for it. Neither names a magnitude, which
2617 // is the thing this crate leaves to
2618 // makeover-geometry.
2619 | "table"
2620 | "table-row"
2621 | "table-cell"
2622 | "nowrap"
2623 | "100%"
2624 | "1%"
2625 // The time axis, 0.42.0. `position` is the one
2626 // property whose whole job is where a thing sits,
2627 // which is exactly what this crate spent its life
2628 // refusing to say -- so it is worth being exact
2629 // about why these two are not that refusal
2630 // breaking.
2631 //
2632 // Neither names a magnitude. `relative` says the
2633 // track is what its entries resolve against, and
2634 // `absolute` says an entry is placed rather than
2635 // flowed. *Where* each entry lands is
2636 // `--track-at` and `--track-for`, custom
2637 // properties the caller sets from
2638 // `Track::fraction`, and they are skipped by the
2639 // `var(--` arm above like every other value this
2640 // crate refuses to decide.
2641 //
2642 // The rule that would break the refusal is a slot
2643 // height, and there is none: the track's height is
2644 // the app's, so the percentages have something to
2645 // resolve against and this crate still never says
2646 // how tall a day is.
2647 | "relative"
2648 | "absolute"
2649 // A run's five, 0.49.0. `flex`, `wrap` and
2650 // `center` are structure and alignment, the same
2651 // reading `table` gets: which way members are laid
2652 // out and how they line up, never how much of
2653 // anything.
2654 //
2655 // The two intrinsic keywords are the interesting
2656 // pair and they are the opposite of a size. A
2657 // magnitude is a number somebody chose;
2658 // `min-content` and `max-content` are the browser
2659 // being asked what the members themselves come to,
2660 // which is the derived minimum the room ruling
2661 // requires and the reason no breakpoint appears
2662 // anywhere in these rules. `1 1 max-content` is
2663 // grow, shrink and that basis, so its two digits
2664 // are ratios rather than lengths.
2665 | "flex"
2666 | "wrap"
2667 | "center"
2668 | "min-content"
2669 | "1 1 max-content"
2670 // A picture's three, 0.36.0. `block` is structure
2671 // for the reason `table` is: an inline image sits
2672 // on the baseline and carries a descender's worth
2673 // of space under it, which is a fact about
2674 // replaced elements rather than a size this crate
2675 // chose. `cover` and `contain` are `Fit`'s two
2676 // named members reaching CSS unchanged, which is
2677 // an intent arriving rather than a value being
2678 // picked.
2679 | "block"
2680 | "cover"
2681 | "contain"
2682 // A relaxed part's three, and the third is the
2683 // awkward one. `-webkit-box` and `vertical` are
2684 // structure: they say the part is a box of lines
2685 // stacked downward, which is the only way CSS lets
2686 // anyone ask for a clamp at all.
2687 //
2688 // `2` is a count of lines, not a length. The
2689 // distinction this crate holds is between naming a
2690 // magnitude -- a padding, a height, a font size,
2691 // all of which belong to makeover-geometry -- and
2692 // naming how many of something there are. A line's
2693 // height is still the app's, so two lines is
2694 // whatever two of the app's lines come to, and
2695 // nothing here decides how tall that is. It is also
2696 // not a value picked here: it is `Flow::Relaxed`'s
2697 // own answer arriving unchanged, the same way
2698 // `cover` and `contain` are `Fit`'s.
2699 | "-webkit-box"
2700 | "vertical"
2701 | "2"
2702 ),
2703 "unrecognised literal value: {line}"
2704 );
2705 }
2706 }
2707
2708 #[test]
2709 fn flat_emits_nothing_at_all() {
2710 assert_eq!(depth_class(Depth::Flat, &Emit::default()), None);
2711 assert!(!depth_rules(&Emit::default()).contains("flat"));
2712 }
2713
2714 #[test]
2715 fn a_prefix_namespaces_every_class() {
2716 let opts = Emit {
2717 class_prefix: "mo-",
2718 ..Emit::default()
2719 };
2720 let css = depth_rules(&opts);
2721 assert!(css.contains(".mo-raised {"));
2722 assert!(css.contains(".mo-well {"));
2723 assert!(!css.contains(".raised {"));
2724 }
2725
2726 #[test]
2727 fn the_border_width_is_the_callers() {
2728 let opts = Emit {
2729 border_width: "2px",
2730 ..Emit::default()
2731 };
2732 assert!(bevel_shadow(Bevel::Raised, &opts).contains("inset 2px 2px 0"));
2733 }
2734
2735 #[test]
2736 fn edges_agree_with_the_description() {
2737 // Not a tautology: it is the guard that a CSS-shaped convenience never
2738 // quietly reverses which side is lit.
2739 let (tl, br) = Bevel::Raised.edges();
2740 assert_eq!(tl.token(), Edge::Light.token());
2741 assert_eq!(br.token(), Edge::Dark.token());
2742 }
2743}