twig/lib.rs
1mod error;
2
3// The raw FFI layer moved to the `twig-sys` crate. Alias it as `ffi` so every
4// `ffi::…` / `crate::ffi::…` reference in this crate keeps resolving unchanged,
5// and so `twig-sys`'s build script (via its `links = "twig"`) links `libtwig.a`
6// into this crate.
7pub(crate) use twig_sys as ffi;
8
9use std::marker::PhantomData;
10use std::ops::Range;
11use std::os::raw::{c_char, c_int};
12use std::ptr::NonNull;
13
14pub use error::Error;
15pub use ffi::TwigSpan as Span;
16
17/// Every format Twig can **parse** — the input axis, as opposed to [`Target`],
18/// which is where output bytes can go.
19///
20/// `#[non_exhaustive]` for the same reason [`Target`] is: Twig's parser list
21/// grows (reStructuredText is written and awaiting a registry entry), and a
22/// caller matching on this enum should not have to be recompiled to keep
23/// compiling. Match with a `_` arm.
24#[derive(Clone, Copy, Debug, Eq, PartialEq)]
25#[non_exhaustive]
26pub enum Format {
27 Djot,
28 Markdown,
29 Xml,
30 Html,
31 /// Parsed, rendered, serialized (`Target::Asciidoc`) and authored into:
32 /// every block gesture, the inline marks, a link and an image work over
33 /// an AsciiDoc document, while the footnote and table gestures report
34 /// [`Error::UnsupportedFormat`] — their AsciiDoc spellings have a shape
35 /// the gesture algorithms cannot write (see [`Format::supports`]).
36 ///
37 /// The parser covers the language as the AsciiDoc ASG schema enumerates
38 /// it. The few constructs it leaves unmodelled (`menu:`, `icon:`,
39 /// `include::`, the CSV table forms) survive as literal source text
40 /// rather than failing the parse.
41 Asciidoc,
42 /// Strict CommonMark 0.31.2: Markdown with every extension off. A
43 /// **dialect** of [`Format::Markdown`] — one parser under a different
44 /// preset, with a [`MarkdownExtensions`] laid over it — carried as a
45 /// format of its own so it can be named in one word, the way fig's
46 /// `json`/`jsonc`/`json5` are three formats over one language. It writes
47 /// as [`Target::Markdown`] (there is one Markdown serializer), and
48 /// [`Format::supports`] answers for it: strict CommonMark cannot author
49 /// the `~~x~~` the other two dialects can.
50 Commonmark,
51 /// GitHub-Flavored Markdown: the spec's four extensions and GFM's HTML
52 /// conventions (a cell's alignment as `align=` rather than `style=`). A
53 /// dialect of [`Format::Markdown`], as [`Format::Commonmark`] is.
54 /// [`Format::Markdown`] itself stays Twig's default flavor, CommonMark
55 /// plus the default-on extensions.
56 Gfm,
57 /// SVG: a **dialect** of [`Format::Xml`] — the same parser, serializer,
58 /// [`Target`] and spelling table, under a name of its own so a `.svg`
59 /// file is recognised on sight and a document records what it was
60 /// opened as. Twig knows nothing about what a `<rect>` means; what makes
61 /// an SVG an SVG is its consumer's business (a canvas editor reading the
62 /// elements it draws), and the variant exists so that consumer has a
63 /// format to name. It writes as [`Target::Xml`], and the one gesture it
64 /// supports is [`Gesture::SetNodeAttrs`], as XML does.
65 Svg,
66}
67
68impl From<Format> for ffi::TwigFormat {
69 fn from(value: Format) -> Self {
70 match value {
71 Format::Djot => ffi::TwigFormat::Djot,
72 Format::Markdown => ffi::TwigFormat::Markdown,
73 Format::Xml => ffi::TwigFormat::Xml,
74 Format::Html => ffi::TwigFormat::Html,
75 Format::Asciidoc => ffi::TwigFormat::Asciidoc,
76 Format::Commonmark => ffi::TwigFormat::Commonmark,
77 Format::Gfm => ffi::TwigFormat::Gfm,
78 Format::Svg => ffi::TwigFormat::Svg,
79 }
80 }
81}
82
83impl Format {
84 /// The language this format is a dialect of, or `None` for a language
85 /// itself: `Some(Format::Markdown)` for [`Format::Commonmark`] and
86 /// [`Format::Gfm`], `Some(Format::Xml)` for [`Format::Svg`], `None` for
87 /// everything else. A dialect shares its language's [`Target`]
88 /// (`Target::from`), which is what makes serializing a GFM document as
89 /// [`Target::Markdown`] a round trip rather than a conversion.
90 pub fn dialect_of(self) -> Option<Format> {
91 match self {
92 Format::Commonmark | Format::Gfm => Some(Format::Markdown),
93 Format::Svg => Some(Format::Xml),
94 _ => None,
95 }
96 }
97}
98
99/// Every format Twig can **write** — the output axis, as opposed to [`Format`],
100/// which is what Twig can **parse**.
101///
102/// Every [`Format`] is also a `Target` (use `Target::from(format)`), so the two
103/// lists coincide today and the distinction costs nothing to ignore. It exists
104/// because only one of them can grow freely: a [`Format`] must have a parser
105/// behind it, while a target only needs somewhere for bytes to go. That makes an
106/// *export-only* target — one Twig can write and no parser reads back, PDF being
107/// the motivating case — expressible here and nowhere else. See the two format
108/// axes in the Zig library's `DESIGN.md`.
109///
110/// `#[non_exhaustive]` for exactly that reason: a future export-only variant is
111/// then an additive change rather than a breaking one for callers that match on
112/// this enum.
113#[derive(Clone, Copy, Debug, Eq, PartialEq)]
114#[non_exhaustive]
115pub enum Target {
116 Djot,
117 Markdown,
118 Xml,
119 Html,
120 /// AsciiDoc source, in Asciidoctor's idiomatic spellings: `= Title`,
121 /// `*strong*` where the boundaries allow and `**strong**` where they do
122 /// not, `[source,lang]` listings, `|===` tables, `footnote:[]` macros.
123 Asciidoc,
124}
125
126impl Target {
127 /// The [`Format`] whose parser reads this target's own output back, or
128 /// `None` for an export-only target.
129 ///
130 /// Always `Some` today. It is the question to ask before assuming a target
131 /// can be round-tripped: `None` means bytes go out and nothing comes back,
132 /// so there is no "parse it again and compare" available for that target.
133 pub fn as_format(self) -> Option<Format> {
134 match self {
135 Target::Djot => Some(Format::Djot),
136 Target::Markdown => Some(Format::Markdown),
137 Target::Xml => Some(Format::Xml),
138 Target::Html => Some(Format::Html),
139 Target::Asciidoc => Some(Format::Asciidoc),
140 }
141 }
142}
143
144/// Total: every input format is also an output target, even the ones with no
145/// serializer yet (converting *into* XML reports [`Error::UnsupportedFormat`]
146/// rather than being unnameable). A dialect lands on its language's target —
147/// [`Format::Commonmark`] and [`Format::Gfm`] both write as
148/// [`Target::Markdown`] — so `Target` does not grow a row per dialect.
149impl From<Format> for Target {
150 fn from(value: Format) -> Self {
151 match value {
152 Format::Djot => Target::Djot,
153 Format::Markdown | Format::Commonmark | Format::Gfm => Target::Markdown,
154 Format::Xml | Format::Svg => Target::Xml,
155 Format::Html => Target::Html,
156 Format::Asciidoc => Target::Asciidoc,
157 }
158 }
159}
160
161impl From<Target> for ffi::TwigFormat {
162 fn from(value: Target) -> Self {
163 match value {
164 Target::Djot => ffi::TwigFormat::Djot,
165 Target::Markdown => ffi::TwigFormat::Markdown,
166 Target::Xml => ffi::TwigFormat::Xml,
167 Target::Html => ffi::TwigFormat::Html,
168 Target::Asciidoc => ffi::TwigFormat::Asciidoc,
169 }
170 }
171}
172
173/// A node's kind, as the shared vocabulary publishes it.
174///
175/// A typed enum rather than the `String` this used to be, because the string
176/// made a whole class of upstream change invisible here. When twig collapsed
177/// its four generic container kinds (`div`, `span`, `directive`, `element`)
178/// into one `container`, every site in this crate that compared a kind name
179/// kept compiling and started being wrong at runtime. With this, each of those
180/// sites is a compile error pointing at the exact line.
181///
182/// `#[non_exhaustive]`, and with an [`Other`](Kind::Other) arm, for the two
183/// different ways the vocabulary can outrun a given build of this crate:
184/// `#[non_exhaustive]` makes ADDING a variant here a non-breaking change for
185/// callers, and `Other` carries a name the linked library published that this
186/// crate has no variant for at all. Match with a `_` arm.
187///
188/// ## What is one variant here and two in the core
189///
190/// The nine inline marks share a single `inline_mark` kind in twig's own AST,
191/// and the nine text leaves share a single `text_leaf`; both publish their
192/// MEMBER name (`"superscript"`, not `"inline_mark"`). This enum follows the
193/// published vocabulary, so they are variants here — the grouping is an
194/// implementation detail of the core, not something a consumer should have to
195/// know.
196///
197/// ## No `PartialEq<&str>`
198///
199/// Deliberately absent, though it would be one impl and would keep every
200/// `node.kind == Kind::Image` in existing code compiling. That is precisely the
201/// property this type exists to remove: a comparison against a string literal
202/// is exactly what survived the container rename and went silently wrong.
203/// Compare against a variant; reach for [`as_str`](Kind::as_str) only when you
204/// genuinely want the name (logging it, or forwarding it to something that
205/// speaks the wire vocabulary).
206#[derive(Clone, Debug, Eq, PartialEq, Hash)]
207#[non_exhaustive]
208pub enum Kind {
209 // ── Document root ─────────────────────────────────────────────────────
210 Doc,
211 // ── Blocks ────────────────────────────────────────────────────────────
212 Para,
213 Heading,
214 ThematicBreak,
215 Section,
216 CodeBlock,
217 RawBlock,
218 Metadata,
219 BlockQuote,
220 BulletList,
221 OrderedList,
222 TaskList,
223 DefinitionList,
224 LineBlock,
225 Table,
226 // ── Structural children, and the document-level definitions ───────────
227 ListItem,
228 TaskListItem,
229 DefinitionListItem,
230 Term,
231 Definition,
232 Line,
233 Row,
234 Cell,
235 Column,
236 Caption,
237 Footnote,
238 Reference,
239 Citation,
240 Substitution,
241 // ── Inlines ───────────────────────────────────────────────────────────
242 Str,
243 SoftBreak,
244 HardBreak,
245 NonBreakingSpace,
246 RawInline,
247 SmartPunctuation,
248 Link,
249 Image,
250 // ── Inline marks — one `inline_mark` kind in the core, published apart
251 Emph,
252 Strong,
253 Mark,
254 Superscript,
255 Subscript,
256 Insert,
257 Delete,
258 DoubleQuoted,
259 SingleQuoted,
260 // ── Text leaves — one `text_leaf` kind in the core, published apart ───
261 Symb,
262 Verbatim,
263 InlineMath,
264 DisplayMath,
265 Url,
266 Email,
267 FootnoteReference,
268 CitationReference,
269 SubstitutionReference,
270 // ── Generic markup ────────────────────────────────────────────────────
271 Container,
272 ProcessingInstruction,
273 Comment,
274 Doctype,
275 Cdata,
276 /// A kind name the linked library published that this crate has no variant
277 /// for — a newer twig against an older binding.
278 ///
279 /// Deliberately not an error: a node whose kind this crate cannot name is
280 /// still a node with a span, children and attributes, and a renderer that
281 /// wants to pass it through unchanged should not be stopped from doing so.
282 Other(String),
283}
284
285impl Kind {
286 /// The name twig publishes for this kind — the exact string the C ABI's
287 /// `TwigFlatNode.kind` carries.
288 pub fn as_str(&self) -> &str {
289 match self {
290 Kind::Doc => "doc",
291 Kind::Para => "para",
292 Kind::Heading => "heading",
293 Kind::ThematicBreak => "thematic_break",
294 Kind::Section => "section",
295 Kind::CodeBlock => "code_block",
296 Kind::RawBlock => "raw_block",
297 Kind::Metadata => "metadata",
298 Kind::BlockQuote => "block_quote",
299 Kind::BulletList => "bullet_list",
300 Kind::OrderedList => "ordered_list",
301 Kind::TaskList => "task_list",
302 Kind::DefinitionList => "definition_list",
303 Kind::LineBlock => "line_block",
304 Kind::Table => "table",
305 Kind::ListItem => "list_item",
306 Kind::TaskListItem => "task_list_item",
307 Kind::DefinitionListItem => "definition_list_item",
308 Kind::Term => "term",
309 Kind::Definition => "definition",
310 Kind::Line => "line",
311 Kind::Row => "row",
312 Kind::Cell => "cell",
313 Kind::Column => "column",
314 Kind::Caption => "caption",
315 Kind::Footnote => "footnote",
316 Kind::Reference => "reference",
317 Kind::Citation => "citation",
318 Kind::Substitution => "substitution",
319 Kind::Str => "str",
320 Kind::SoftBreak => "soft_break",
321 Kind::HardBreak => "hard_break",
322 Kind::NonBreakingSpace => "non_breaking_space",
323 Kind::RawInline => "raw_inline",
324 Kind::SmartPunctuation => "smart_punctuation",
325 Kind::Link => "link",
326 Kind::Image => "image",
327 Kind::Container => "container",
328 Kind::ProcessingInstruction => "processing_instruction",
329 Kind::Emph => "emph",
330 Kind::Strong => "strong",
331 Kind::Mark => "mark",
332 Kind::Superscript => "superscript",
333 Kind::Subscript => "subscript",
334 Kind::Insert => "insert",
335 Kind::Delete => "delete",
336 Kind::DoubleQuoted => "double_quoted",
337 Kind::SingleQuoted => "single_quoted",
338 Kind::Symb => "symb",
339 Kind::Verbatim => "verbatim",
340 Kind::InlineMath => "inline_math",
341 Kind::DisplayMath => "display_math",
342 Kind::Url => "url",
343 Kind::Email => "email",
344 Kind::FootnoteReference => "footnote_reference",
345 Kind::CitationReference => "citation_reference",
346 Kind::SubstitutionReference => "substitution_reference",
347 Kind::Comment => "comment",
348 Kind::Doctype => "doctype",
349 Kind::Cdata => "cdata",
350 Kind::Other(name) => name.as_str(),
351 }
352 }
353
354 /// Whether this is a kind the linked library named and this crate could
355 /// not — the [`Other`](Kind::Other) case, and the one worth logging when a
356 /// renderer meets a node it has no arm for.
357 pub fn is_unknown(&self) -> bool {
358 matches!(self, Kind::Other(_))
359 }
360}
361
362impl From<&str> for Kind {
363 fn from(name: &str) -> Self {
364 match name {
365 "doc" => Kind::Doc,
366 "para" => Kind::Para,
367 "heading" => Kind::Heading,
368 "thematic_break" => Kind::ThematicBreak,
369 "section" => Kind::Section,
370 "code_block" => Kind::CodeBlock,
371 "raw_block" => Kind::RawBlock,
372 "metadata" => Kind::Metadata,
373 "block_quote" => Kind::BlockQuote,
374 "bullet_list" => Kind::BulletList,
375 "ordered_list" => Kind::OrderedList,
376 "task_list" => Kind::TaskList,
377 "definition_list" => Kind::DefinitionList,
378 "line_block" => Kind::LineBlock,
379 "table" => Kind::Table,
380 "list_item" => Kind::ListItem,
381 "task_list_item" => Kind::TaskListItem,
382 "definition_list_item" => Kind::DefinitionListItem,
383 "term" => Kind::Term,
384 "definition" => Kind::Definition,
385 "line" => Kind::Line,
386 "row" => Kind::Row,
387 "cell" => Kind::Cell,
388 "column" => Kind::Column,
389 "caption" => Kind::Caption,
390 "footnote" => Kind::Footnote,
391 "reference" => Kind::Reference,
392 "citation" => Kind::Citation,
393 "substitution" => Kind::Substitution,
394 "str" => Kind::Str,
395 "soft_break" => Kind::SoftBreak,
396 "hard_break" => Kind::HardBreak,
397 "non_breaking_space" => Kind::NonBreakingSpace,
398 "raw_inline" => Kind::RawInline,
399 "smart_punctuation" => Kind::SmartPunctuation,
400 "link" => Kind::Link,
401 "image" => Kind::Image,
402 "container" => Kind::Container,
403 "processing_instruction" => Kind::ProcessingInstruction,
404 "emph" => Kind::Emph,
405 "strong" => Kind::Strong,
406 "mark" => Kind::Mark,
407 "superscript" => Kind::Superscript,
408 "subscript" => Kind::Subscript,
409 "insert" => Kind::Insert,
410 "delete" => Kind::Delete,
411 "double_quoted" => Kind::DoubleQuoted,
412 "single_quoted" => Kind::SingleQuoted,
413 "symb" => Kind::Symb,
414 "verbatim" => Kind::Verbatim,
415 "inline_math" => Kind::InlineMath,
416 "display_math" => Kind::DisplayMath,
417 "url" => Kind::Url,
418 "email" => Kind::Email,
419 "footnote_reference" => Kind::FootnoteReference,
420 "citation_reference" => Kind::CitationReference,
421 "substitution_reference" => Kind::SubstitutionReference,
422 "comment" => Kind::Comment,
423 "doctype" => Kind::Doctype,
424 "cdata" => Kind::Cdata,
425 other => Kind::Other(other.to_string()),
426 }
427 }
428}
429
430impl std::fmt::Display for Kind {
431 fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
432 f.write_str(self.as_str())
433 }
434}
435
436/// One node returned by [`Document::query`]: its AST id, byte spans, and kind.
437#[derive(Clone, Debug, Eq, PartialEq)]
438pub struct QueryMatch {
439 /// The node's id in the shared AST.
440 pub node_id: u32,
441 /// The node's whole byte range in the source.
442 pub span: Range<usize>,
443 /// The node's interior byte range (between its delimiters), or `None` for
444 /// a leaf / a container with no known interior.
445 pub content_span: Option<Range<usize>>,
446 /// The node's kind. See [`Kind`] for why this is an enum and not the
447 /// name string the C ABI carries.
448 pub kind: Kind,
449}
450
451/// The byte-level effect of an [`Editor`] edit: `old` is the range of the
452/// pre-edit source that was replaced, `new` the range the replacement now
453/// occupies in the post-edit source (they share a start). An insertion has an
454/// empty `old`; a deletion an empty `new`. Everything a caret/selection needs
455/// to re-anchor across an edit without re-diffing: shift any offset `>= old.end`
456/// by `new.len() - old.len()`.
457#[derive(Clone, Debug, Eq, PartialEq)]
458pub struct Change {
459 pub old: Range<usize>,
460 pub new: Range<usize>,
461}
462
463impl Change {
464 /// The net change in source length (`new.len() - old.len()`).
465 pub fn delta(&self) -> isize {
466 self.new.len() as isize - self.old.len() as isize
467 }
468
469 fn from_ffi(c: ffi::TwigChange) -> Self {
470 Change {
471 old: c.old_span.start..c.old_span.end,
472 new: c.new_span.start..c.new_span.end,
473 }
474 }
475}
476
477/// One node of an [`Editor::nodes`] snapshot — the flat AST arena as owned Rust
478/// data (the JSON-free read path). `id` indexes the snapshot; `parent`,
479/// `first_child`, and `next_sibling` link the tree (`None` where absent).
480/// `text` is the node's primary payload (a `str`'s bytes, a `code_block`'s
481/// body, …) and `destination` a link/image target, each `None` when the kind
482/// carries no such payload.
483/// `#[non_exhaustive]`: a snapshot node is something twig *hands you*, never
484/// something you build, so it gains a field whenever a node kind's payload is
485/// surfaced (as `head`/`alignment` were for tables). Sealing construction here
486/// keeps every future addition a minor release instead of a major one.
487#[derive(Clone, Debug, Eq, PartialEq)]
488#[non_exhaustive]
489pub struct FlatNode {
490 pub id: NodeId,
491 pub parent: Option<NodeId>,
492 pub first_child: Option<NodeId>,
493 pub next_sibling: Option<NodeId>,
494 pub span: Range<usize>,
495 pub content_span: Option<Range<usize>>,
496 /// A heading's level; `None` for every other kind.
497 pub level: Option<u32>,
498 pub kind: Kind,
499 /// The node's text payload — a `str`'s bytes, a `code_block`'s body — and
500 /// `None` for a node whose content is its children. Since twig 3.5 this is
501 /// also `Some` for a [`Kind::Container`] whose body the HTML tokenizer read
502 /// as TEXT rather than markup (`<script>`, `<style>`, `<iframe>`, `<title>`,
503 /// `<textarea>`, …): such a node has no children, and this is how an editor
504 /// learns that the body is not prose without carrying the tokenizer's tag
505 /// lists itself. Test `text.is_some()`, not the tag name.
506 pub text: Option<String>,
507 pub destination: Option<String>,
508 /// Whether a `row`/`cell` belongs to the table head; `None` for every other
509 /// kind.
510 pub head: Option<bool>,
511 /// A `cell`'s column alignment; `None` for every other kind. The delimiter
512 /// row (`|:--|--:|`) that spells the alignment out is consumed by the parser
513 /// and has no node of its own, so this is the only way to recover it.
514 /// [`Alignment::Default`] is a real, unspecified alignment (a bare `---`) —
515 /// distinct from the `None` a non-cell node reports.
516 pub alignment: Option<Alignment>,
517 /// The name a generic container carries in its own payload rather than in
518 /// `kind`: an HTML/XML tag (`"picture"`, `"source"`, …) or a directive type
519 /// (`"note"`, `"embed"`, `"vis"`, …, no leading colons). `None` for every
520 /// semantic kind, whose identity is `kind` alone. With this an
521 /// `html_elements` parse's `<picture>`/`<source>` are distinguishable — both
522 /// report `kind == "container"` — and so are a `::embed` and a `::toc`.
523 ///
524 /// A tag and a directive type share one `kind` because they are one concept
525 /// in the core: a named container with attributes and children. `name` is
526 /// what tells them apart, which is why it is not optional in practice for
527 /// anything a renderer cares about.
528 pub name: Option<String>,
529 /// Which of the three generic-container SPELLINGS this node's producer
530 /// draws; `None` when it draws none. Pairs with [`name`](Self::name): the
531 /// name says *which* container, this says *how it is written*, and a
532 /// renderer needs both — the same type is a span inline
533 /// ([`DirectiveForm::Text`]), a standalone block with no body
534 /// ([`DirectiveForm::Leaf`]), and a wrapper around blocks
535 /// ([`DirectiveForm::Container`]).
536 ///
537 /// **This does not answer "is it a directive?"** — use
538 /// [`origin`](Self::origin). HTML's parser sets a form on `<div>` and
539 /// `<span>`, the two tags djot and Markdown have generic spellings for, so
540 /// this reports `Some(Container)` for a `<div>` and `None` for a
541 /// `<video>`: right often enough to look usable, wrong on the two tags you
542 /// meet first.
543 pub directive_form: Option<DirectiveForm>,
544 /// Whether a generic container was WRITTEN as a tag or as a directive;
545 /// `None` when nothing recorded it — the node is not a container, or no
546 /// parser produced it (a [`Builder`] tree).
547 ///
548 /// This is the field that separates an HTML `<div>` from a Markdown
549 /// `:::div`. Those two agree on [`kind`](Self::kind) (`"container"`), on
550 /// [`name`](Self::name) (`"div"`) and on
551 /// [`directive_form`](Self::directive_form) (`Container`), field for field,
552 /// so none of the three can tell you which one you have.
553 pub origin: Option<ContainerOrigin>,
554 /// The node's own MARKER — the leading bytes a rich view HIDES, on its
555 /// opening line: a heading's `#`s and the space after them, a list item's
556 /// `- ` / `1. `, a task item's marker plus its `[x] ` box, a block quote's
557 /// `> `. `None` for a node with no leading marker (every inline, a
558 /// paragraph, a SETEXT heading whose `---` sits *under* the block).
559 ///
560 /// **Not derivable from [`span`](Self::span) and
561 /// [`content_span`](Self::content_span).** For a heading it happens to be
562 /// `span.start..content_span.start`; for a marker-prefixed container it is
563 /// not, because those report `content_span == span` — a prefix repeating on
564 /// every line has no contiguous interior to point at. Before this field the
565 /// answer was recoverable only by a per-format rule (from the item's inner
566 /// paragraph in Markdown, from the item itself in Djot), which is the
567 /// "which parser produced this?" reasoning a shared AST exists to remove.
568 ///
569 /// Covers ONE LINE, and this node's own marker alone. For the whole prefix a
570 /// nested construct sits behind (`> 1. [ ] ` is four nodes' markers plus
571 /// the indent between them), call [`Document::line_prefix`].
572 pub marker_span: Option<Range<usize>>,
573 /// A task list item's checkbox state; `None` for every other kind.
574 ///
575 /// The parser has always known this — it is what decides
576 /// [`Kind::TaskListItem`] over [`Kind::ListItem`] in the first place — and
577 /// until now nothing surfaced it, so a consumer rendering a clickable
578 /// checkbox re-derived the state by scanning the source for `[x]`. That scan
579 /// is fooled by a `[` in prose, and it asks the bytes a question the tree
580 /// had already answered. Twig would WRITE a checkbox
581 /// ([`Editor::set_task_checked`]) and not read one back.
582 ///
583 /// `None` is distinct from `Some(false)`: a consumer treating "not a task
584 /// item" as unchecked draws an empty box beside every paragraph.
585 pub checked: Option<bool>,
586 /// The node's `{...}` / HTML attributes as `(key, value)` pairs in source
587 /// order (empty when it has none). A bare attribute (HTML `disabled`, or a
588 /// `<source media=…>` used as a flag) has a `None` value.
589 pub attrs: Vec<(String, Option<String>)>,
590}
591
592/// A synthesized line prefix — what [`Document::continuation_prefix`] and
593/// [`Document::blank_line_prefix`] build.
594///
595/// `columns` is deliberately not `text.len()`: a tab in a marker advances to a
596/// tab stop, so `-\tx` yields a four-column prefix from a two-byte marker. An
597/// editor sizing a Tab step, a caret's horizontal home, or an outdent wants the
598/// column count; one writing the prefix into the document wants the bytes.
599#[derive(Clone, Debug, Default, Eq, PartialEq)]
600pub struct LinePrefix {
601 /// The bytes to write at the head of the line.
602 pub text: String,
603 /// Their width in columns.
604 pub columns: usize,
605}
606
607/// An inline mark for [`Editor::wrap_range`] / [`Editor::toggle_inline`] — a
608/// rich editor's Bold / Italic / Code / … buttons. Djot spells all of them;
609/// Markdown spells [`InlineKind::Strong`], [`InlineKind::Emph`],
610/// [`InlineKind::Verbatim`] and [`InlineKind::Delete`] — GFM strikethrough is
611/// parsed by default, so every editor this crate creates authors it — plus
612/// [`InlineKind::Mark`] once [`MarkdownExtensions::highlight`] is set, since
613/// `==x==` is otherwise text the reparse hands back unchanged. An unsupported
614/// kind yields [`Error::UnsupportedFormat`]; [`Format::supports_with`] is the
615/// question asked ahead of the call.
616#[derive(Clone, Copy, Debug, Eq, PartialEq)]
617pub enum InlineKind {
618 Strong,
619 Emph,
620 Verbatim,
621 Mark,
622 Superscript,
623 Subscript,
624 Insert,
625 Delete,
626}
627
628impl InlineKind {
629 fn to_c(self) -> c_int {
630 match self {
631 InlineKind::Strong => 0,
632 InlineKind::Emph => 1,
633 InlineKind::Verbatim => 2,
634 InlineKind::Mark => 3,
635 InlineKind::Superscript => 4,
636 InlineKind::Subscript => 5,
637 InlineKind::Insert => 6,
638 InlineKind::Delete => 7,
639 }
640 }
641}
642
643/// A block target for [`Editor::set_block`] — the toolbar's H1…H6 / Body switch.
644#[derive(Clone, Copy, Debug, Eq, PartialEq)]
645pub enum BlockKind {
646 Paragraph,
647 /// A heading of the given level (1–6; out of range is [`Error::InvalidArgument`]).
648 Heading(u32),
649}
650
651impl BlockKind {
652 /// `(block_kind_code, level)` for the C ABI.
653 fn to_c(self) -> (c_int, u32) {
654 match self {
655 BlockKind::Paragraph => (0, 0),
656 BlockKind::Heading(level) => (1, level),
657 }
658 }
659}
660
661/// A block container for [`Editor::toggle_block_container`] — the toolbar's
662/// Quote / Bulleted list / Numbered list buttons. Where a [`BlockKind`] rewrites
663/// one block's leading marker, a container prefixes every line of a range and
664/// nests. Djot and Markdown spell all three; other formats yield
665/// [`Error::UnsupportedFormat`].
666#[derive(Clone, Copy, Debug, Eq, PartialEq)]
667pub enum BlockContainerKind {
668 BlockQuote,
669 BulletList,
670 OrderedList,
671}
672
673impl BlockContainerKind {
674 fn to_c(self) -> c_int {
675 match self {
676 BlockContainerKind::BlockQuote => 0,
677 BlockContainerKind::BulletList => 1,
678 BlockContainerKind::OrderedList => 2,
679 }
680 }
681}
682
683/// The colour of a highlight — the palette [`Editor::set_mark_color`] writes.
684///
685/// Obsidian's spelling, which is what Twig reads and writes: a large-circle
686/// emoji immediately after the opening `==`, so `==🔴 text==` is a highlight
687/// whose text is `text` and whose colour is [`MarkColor::Red`]. The emoji is
688/// **spelling**, not content — it is stripped from the highlighted text and
689/// carried as the mark's `data-color` attribute, which is where a
690/// [`Document::query`] for `mark[data-color=red]` finds it.
691///
692/// An enum rather than a string because the palette is closed: a name Twig has
693/// no emoji for is not a colour it can write, and an emoji it does not read
694/// back is text. The C ABI takes the name — [`MarkColor::as_str`] is it, and is
695/// exactly the attribute value.
696#[derive(Clone, Copy, Debug, Eq, PartialEq)]
697pub enum MarkColor {
698 Red,
699 Orange,
700 Yellow,
701 Green,
702 Blue,
703 Purple,
704 Brown,
705}
706
707impl MarkColor {
708 /// The `data-color` value — `"red"` — and what the C ABI is handed.
709 pub fn as_str(self) -> &'static str {
710 match self {
711 MarkColor::Red => "red",
712 MarkColor::Orange => "orange",
713 MarkColor::Yellow => "yellow",
714 MarkColor::Green => "green",
715 MarkColor::Blue => "blue",
716 MarkColor::Purple => "purple",
717 MarkColor::Brown => "brown",
718 }
719 }
720
721 /// The colour a `data-color` attribute names, or `None` for a value this
722 /// build has no spelling for.
723 pub fn from_str(s: &str) -> Option<Self> {
724 Some(match s {
725 "red" => MarkColor::Red,
726 "orange" => MarkColor::Orange,
727 "yellow" => MarkColor::Yellow,
728 "green" => MarkColor::Green,
729 "blue" => MarkColor::Blue,
730 "purple" => MarkColor::Purple,
731 "brown" => MarkColor::Brown,
732 _ => return None,
733 })
734 }
735}
736
737/// One authoring gesture, named with whatever kind it takes — the question
738/// [`Format::supports`] answers.
739///
740/// Twig's formats are **ragged**: Djot spells all eight inline marks and
741/// Markdown four (a fifth with [`MarkdownExtensions::highlight`]), HTML spells
742/// marks, headings, quotes, lists, code blocks, links and images but no task
743/// box, footnote or table edit, AsciiDoc spells everything but footnotes and
744/// tables, XML nothing. Every [`Editor`]
745/// method already reports that as
746/// [`Error::UnsupportedFormat`] — but only once called, which is too late for a
747/// UI that wants to *disable* the button rather than let it fail.
748///
749/// A variant carries a kind exactly where the [`Editor`] method takes one, so
750/// the query is spelled with the same value as the call:
751///
752/// ```no_run
753/// # use twig::{Format, Gesture, InlineKind};
754/// if Format::Markdown.supports(Gesture::ToggleInline(InlineKind::Mark)) {
755/// // never runs: `==mark==` is emit-only in Markdown.
756/// }
757/// ```
758///
759/// Only the gestures with a **format-level** gate appear — which, since the last
760/// nine were added, is every gesture the editor has. Those nine read no format
761/// spelling at all until an HTML document showed what that cost:
762///
763/// - The `Table*` variants. HTML's parser lowers `<table>/<tr>/<td>` to the same
764/// nodes a pipe table produces, so [`Editor::table_insert_row`] and its
765/// siblings extracted the grid and wrote pipe text over the elements. HTML
766/// reparses that as a paragraph —
767/// a document that still parses, so nothing rolled it back and no error was
768/// returned. The table was simply gone.
769/// - [`Gesture::SplitBlock`]. The blank line it writes means "two blocks" only
770/// where blank lines separate blocks; inside a `<p>` it is whitespace, so
771/// [`Editor::split_block`] reported success over an unchanged document.
772/// - [`Gesture::RenumberOrderedLists`]. A textual `N.` rewrite finds nothing in
773/// an `<ol>`, whose numbering is in the tag, and called the no-op a success.
774///
775/// `#[non_exhaustive]` for the reason [`Format`] is: the gesture list grows with
776/// the editor surface, and a caller matching on this should not need a rebuild.
777#[derive(Clone, Copy, Debug, Eq, PartialEq)]
778#[non_exhaustive]
779pub enum Gesture {
780 WrapRange(InlineKind),
781 ToggleInline(InlineKind),
782 SetBlock,
783 ToggleBlockContainer(BlockContainerKind),
784 InsertThematicBreak,
785 ToggleCodeBlock,
786 SetCodeLanguage,
787 ToggleTaskItem,
788 SetTaskChecked,
789 ToggleTaskChecked,
790 InsertLink,
791 InsertImage,
792 InsertFootnote,
793 InsertLiteral,
794 InsertLineBreak,
795 SplitBlock,
796 RenumberOrderedLists,
797 TableInsertRow,
798 TableDeleteRow,
799 TableInsertColumn,
800 TableDeleteColumn,
801 TableSetAlignment,
802 TableMoveRow,
803 TableMoveColumn,
804 /// Colour the highlight the caret is in — [`Editor::set_mark_color`].
805 ///
806 /// The one gesture whose support is a fact about the **parse extensions**
807 /// rather than about the format: it needs
808 /// [`MarkdownExtensions::highlight_colors`], so ask
809 /// [`Format::supports_with`] rather than [`Format::supports`], which
810 /// answers for default options and so always answers `false` here.
811 SetMarkColor,
812 /// Mint a fresh table — [`Editor::insert_table`]. Behind the same gate as
813 /// the seven table edits: a format that can re-spell a table can write one.
814 InsertTable,
815 /// Write a leaf directive — [`Editor::insert_directive`]. Supported where
816 /// the format's parser reads a printed named container back as one, which
817 /// for Markdown means [`MarkdownExtensions::directives`]: ask
818 /// [`Format::supports_with`] rather than [`Format::supports`], which
819 /// answers for default options and so answers `false` there.
820 InsertDirective,
821 /// Replace a block's attribute set — [`Editor::set_block_attrs`].
822 /// Supported where the format's parser reads a block's printed attributes
823 /// back, which for Markdown means [`MarkdownExtensions::html_elements`]:
824 /// ask [`Format::supports_with`] rather than [`Format::supports`].
825 SetBlockAttrs,
826 /// Wrap a range in an attributed span — [`Editor::wrap_range_attrs`].
827 /// Supported where the format reads the printed span back: djot and HTML,
828 /// and Markdown under [`MarkdownExtensions::html_elements`]; not AsciiDoc.
829 WrapRangeAttrs,
830 /// Join a block into the block before it — [`Editor::join_blocks`]. The
831 /// inverse of [`Gesture::SplitBlock`] and a **wider** gate than it: a join
832 /// writes a line break inside a block, which HTML spells, while a blank
833 /// line between two of its `<p>`s is not what separates them. Ask for this
834 /// one rather than reading the split's answer for both.
835 JoinBlocks,
836 /// Replace an element's attribute set by node id — [`Editor::set_node_attrs`].
837 /// Supported where the format keeps a node's attributes on the node's own
838 /// tag at a span its parser records: XML, and no prose format. The one
839 /// gesture a tree-shaped editor asks for and a caret never does, which is
840 /// why [`Format::is_authorable`] stays `false` for XML while this answers
841 /// `true`.
842 SetNodeAttrs,
843 /// Move a block to a boundary, re-spelling the line prefixes of the
844 /// container it lands in — [`Editor::move_block`]. Supported wherever a
845 /// caret can name a block: every prose format, HTML included; not XML.
846 MoveBlock,
847}
848
849impl Gesture {
850 /// `(gesture_code, kind_code)` for the C ABI. The kind rides in the
851 /// gesture's own space — an inline code for the two inline gestures, a
852 /// container code for the container one, and 0 where the gesture takes
853 /// none, which the C side *requires* rather than ignores.
854 fn to_c(self) -> (c_int, c_int) {
855 match self {
856 Gesture::WrapRange(k) => (0, k.to_c()),
857 Gesture::ToggleInline(k) => (1, k.to_c()),
858 Gesture::SetBlock => (2, 0),
859 Gesture::ToggleBlockContainer(k) => (3, k.to_c()),
860 Gesture::InsertThematicBreak => (4, 0),
861 Gesture::ToggleCodeBlock => (5, 0),
862 Gesture::SetCodeLanguage => (6, 0),
863 Gesture::ToggleTaskItem => (7, 0),
864 Gesture::SetTaskChecked => (8, 0),
865 Gesture::ToggleTaskChecked => (9, 0),
866 Gesture::InsertLink => (10, 0),
867 Gesture::InsertImage => (11, 0),
868 Gesture::InsertFootnote => (12, 0),
869 Gesture::InsertLiteral => (13, 0),
870 Gesture::InsertLineBreak => (14, 0),
871 Gesture::SplitBlock => (15, 0),
872 Gesture::RenumberOrderedLists => (16, 0),
873 Gesture::TableInsertRow => (17, 0),
874 Gesture::TableDeleteRow => (18, 0),
875 Gesture::TableInsertColumn => (19, 0),
876 Gesture::TableDeleteColumn => (20, 0),
877 Gesture::TableSetAlignment => (21, 0),
878 Gesture::TableMoveRow => (22, 0),
879 Gesture::TableMoveColumn => (23, 0),
880 Gesture::SetMarkColor => (24, 0),
881 Gesture::InsertTable => (25, 0),
882 Gesture::InsertDirective => (26, 0),
883 Gesture::SetBlockAttrs => (27, 0),
884 Gesture::WrapRangeAttrs => (28, 0),
885 Gesture::JoinBlocks => (29, 0),
886 Gesture::SetNodeAttrs => (30, 0),
887 Gesture::MoveBlock => (31, 0),
888 }
889 }
890}
891
892impl Format {
893 /// Whether this format can spell `gesture` — the toolbar's gray-out
894 /// question, answered **without a document**, so a caller can build its UI
895 /// before it has one. Pure and cheap: ask at startup and cache.
896 ///
897 /// `true` means the gesture will not fail with
898 /// [`Error::UnsupportedFormat`]. It is **not** a promise the call succeeds —
899 /// the caret still decides, so a supported gesture can still report
900 /// [`Error::NotFound`], [`Error::NotEditable`] or [`Error::EditConflict`] at
901 /// the position it is actually run. Gray out on `false`; do not read `true`
902 /// as "this will work here".
903 ///
904 /// Returns a plain `bool` rather than a `Result` because the two ways the C
905 /// query can fail — an unknown format code, a kind from the wrong
906 /// vocabulary — are both unrepresentable here: [`Format`] and [`Gesture`]
907 /// are enums, and a `Gesture` carries a kind only where one applies.
908 ///
909 /// Distinct from BOTH neighbouring questions:
910 ///
911 /// - [`Format::is_authorable`] is "is there a door in", true for
912 /// [`Format::Html`] on its inline marks alone.
913 /// - [`Warning::fidelity`] is "what survives a *conversion* to this target",
914 /// which is a different table with genuinely different answers — Djot
915 /// round-trips a smart-quote container faithfully while no editor gesture
916 /// may author one. Use that for a save-as warning, this for a button.
917 pub fn supports(self, gesture: Gesture) -> bool {
918 let (g, k) = gesture.to_c();
919 let mut supported: c_int = 0;
920 let status = unsafe {
921 ffi::twig_format_supports(ffi::TwigFormat::from(self) as c_int, g, k, &mut supported)
922 };
923 debug_assert!(
924 Error::from_status(status).is_ok(),
925 "twig_format_supports rejected a combination the Rust types make unrepresentable",
926 );
927 supported == 1
928 }
929
930 /// [`Format::supports`] for a document parsed with `extensions` — the same
931 /// question asked of the table an [`Editor`] created with them actually
932 /// holds.
933 ///
934 /// A Markdown extension can **widen** what may be authored, which is why
935 /// the format alone is not always the whole answer. `==x==` is literal text
936 /// under default options and a `mark` under
937 /// [`MarkdownExtensions::highlight`], so a toggle that wrote it without the
938 /// extension would mint bytes the reparse hands back as plain text — one
939 /// press that a second press cannot undo. [`Gesture::SetMarkColor`] needs
940 /// [`MarkdownExtensions::highlight_colors`] on top of that.
941 ///
942 /// ```no_run
943 /// # use twig::{Format, Gesture, InlineKind, MarkdownExtensions};
944 /// let exts = MarkdownExtensions { highlight: true, ..Default::default() };
945 /// assert!(!Format::Markdown.supports(Gesture::ToggleInline(InlineKind::Mark)));
946 /// assert!(Format::Markdown.supports_with(exts, Gesture::ToggleInline(InlineKind::Mark)));
947 /// ```
948 ///
949 /// Pass the extensions the editor was (or will be) created with in
950 /// [`Editor::new_ext`]; anything else answers a question about a
951 /// document you do not have. `extensions` is ignored for every
952 /// non-Markdown format, exactly as it is at creation.
953 pub fn supports_with(self, extensions: MarkdownExtensions, gesture: Gesture) -> bool {
954 let (g, k) = gesture.to_c();
955 let mut supported: c_int = 0;
956 let status = unsafe {
957 ffi::twig_format_supports_ext(
958 ffi::TwigFormat::from(self) as c_int,
959 extensions.to_flags(),
960 g,
961 k,
962 &mut supported,
963 )
964 };
965 debug_assert!(
966 Error::from_status(status).is_ok(),
967 "twig_format_supports_ext rejected a combination the Rust types make unrepresentable",
968 );
969 supported == 1
970 }
971
972 /// Whether this format can be authored into **at all** — `false` for a
973 /// parse-only format ([`Format::Xml`]), where every gesture refuses and
974 /// an editor should offer no toolbar. The open-read-only question.
975 ///
976 /// `true` is a **weaker** claim than it looks, and driving per-button state
977 /// from it is the mistake this doc exists to prevent: [`Format::Html`]
978 /// answers `true` — it spells the inline marks, a heading and a literal —
979 /// while the container, code-block, task, link and footnote gestures are
980 /// all still unsupported there. Use [`Format::supports`] per button.
981 pub fn is_authorable(self) -> bool {
982 let mut authorable: c_int = 0;
983 let status = unsafe {
984 ffi::twig_format_is_authorable(ffi::TwigFormat::from(self) as c_int, &mut authorable)
985 };
986 debug_assert!(Error::from_status(status).is_ok(), "unknown format code");
987 authorable == 1
988 }
989}
990
991#[derive(Clone, Copy, Debug, Eq, PartialEq)]
992pub struct Version {
993 pub major: u8,
994 pub minor: u8,
995 pub patch: u8,
996}
997
998pub fn version() -> Version {
999 let packed = unsafe { ffi::twig_version() };
1000 Version {
1001 major: (packed >> 16) as u8,
1002 minor: (packed >> 8) as u8,
1003 patch: packed as u8,
1004 }
1005}
1006
1007/// The C ABI contract version this crate was **compiled** against — the
1008/// compile-time counterpart to [`abi_version`] (which reports the **linked
1009/// library's**). This crate builds and links its own vendored copy of the Zig
1010/// source, so the two always agree; the pair is exposed so a consumer embedding
1011/// a separately-built library can verify layout compatibility at load time.
1012pub const ABI_VERSION: u32 = ffi::TWIG_ABI_VERSION;
1013
1014/// The C ABI contract version of the linked library. This crate is written
1015/// against [`ABI_VERSION`]; the two agreeing is what makes the `#[repr(C)]`
1016/// mirrors in `ffi` sound. It is bumped only on a breaking ABI change (a struct
1017/// layout change or a renumbered enum value), never on an additive one (a new
1018/// format code or a new function).
1019pub fn abi_version() -> u32 {
1020 unsafe { ffi::twig_abi_version() }
1021}
1022
1023pub fn version_string() -> &'static str {
1024 let ptr = unsafe { ffi::twig_version_string() };
1025 unsafe { std::ffi::CStr::from_ptr(ptr) }
1026 .to_str()
1027 .unwrap_or("")
1028}
1029
1030#[derive(Debug)]
1031pub struct Document {
1032 raw: NonNull<ffi::TwigDocument>,
1033}
1034
1035impl Document {
1036 pub fn parse(input: &[u8], format: Format) -> Result<Self, Error> {
1037 Self::parse_with(input, format, MarkdownExtensions::default())
1038 }
1039
1040 pub fn parse_str(input: &str, format: Format) -> Result<Self, Error> {
1041 Self::parse(input.as_bytes(), format)
1042 }
1043
1044 /// Like [`Document::parse`], plus Markdown `extensions` to enable (ignored
1045 /// for other formats) — the read-path counterpart of [`Editor::new_ext`].
1046 /// Enable [`MarkdownExtensions::html_elements`] here to make embedded HTML
1047 /// (`<img>`, `<picture>`, …) queryable via [`Document::query`] instead of
1048 /// arriving as opaque raw HTML.
1049 pub fn parse_with(
1050 input: &[u8],
1051 format: Format,
1052 extensions: MarkdownExtensions,
1053 ) -> Result<Self, Error> {
1054 let mut raw = std::ptr::null_mut();
1055 let ffi_format: ffi::TwigFormat = format.into();
1056 let status = unsafe {
1057 ffi::twig_parse_ext(
1058 input.as_ptr(),
1059 input.len(),
1060 ffi_format as i32,
1061 extensions.to_flags(),
1062 &mut raw,
1063 )
1064 };
1065 Error::from_status(status)?;
1066 let raw = NonNull::new(raw).ok_or(Error::Internal)?;
1067 Ok(Self { raw })
1068 }
1069
1070 /// [`Document::parse_with`] for a `&str`.
1071 pub fn parse_str_with(
1072 input: &str,
1073 format: Format,
1074 extensions: MarkdownExtensions,
1075 ) -> Result<Self, Error> {
1076 Self::parse_with(input.as_bytes(), format, extensions)
1077 }
1078
1079 /// Render the document to HTML. For Djot/Markdown this is the rich
1080 /// rendering path that resolves reference/footnote side tables.
1081 pub fn render_html(&mut self) -> Result<Vec<u8>, Error> {
1082 let raw = self.raw.as_ptr();
1083 collect_bytes(|ptr, len| unsafe { ffi::twig_document_render_html(raw, ptr, len) })
1084 }
1085
1086 /// Serialize the document to `target`'s own syntax: a round-trip when
1087 /// `target` names the document's own format, cross-format conversion
1088 /// otherwise (e.g. parse Markdown, serialize as Djot). Returns
1089 /// [`Error::UnsupportedFormat`] when the requested direction has no
1090 /// serializer (today: converting into XML from another format).
1091 ///
1092 /// Prefer this over [`Document::serialize`]: serializing is a question about
1093 /// where the bytes are going, so it takes a [`Target`]. The older spelling
1094 /// takes a [`Format`] and still works — every `Format` is a `Target` — but
1095 /// it cannot name an export-only target, and this one can.
1096 pub fn serialize_to(&mut self, target: Target) -> Result<Vec<u8>, Error> {
1097 let raw = self.raw.as_ptr();
1098 let ffi_target: ffi::TwigFormat = target.into();
1099 collect_bytes(|ptr, len| unsafe {
1100 ffi::twig_document_serialize(raw, ffi_target as i32, ptr, len)
1101 })
1102 }
1103
1104 /// Serialize the document to `format`'s own source syntax.
1105 ///
1106 /// The original spelling of [`Document::serialize_to`], kept for
1107 /// compatibility and defined in terms of it. It types the output axis as
1108 /// [`Format`], which is the input vocabulary; reach for `serialize_to` in
1109 /// new code.
1110 pub fn serialize(&mut self, format: Format) -> Result<Vec<u8>, Error> {
1111 self.serialize_to(format.into())
1112 }
1113
1114 /// Encode the document's AST as pretty-printed JSON (the same encoding as
1115 /// `twig convert -o ast`).
1116 pub fn ast_json(&mut self) -> Result<Vec<u8>, Error> {
1117 let raw = self.raw.as_ptr();
1118 collect_bytes(|ptr, len| unsafe { ffi::twig_document_ast_json(raw, ptr, len) })
1119 }
1120
1121 /// Resolve a CSS-lite selector (e.g. `heading[level=2]`,
1122 /// `link[dest^="http"]`, `code`, `list > item`) against the document,
1123 /// returning one [`QueryMatch`] per matching node in document order. A
1124 /// malformed selector yields [`Error::InvalidArgument`].
1125 ///
1126 /// This is the general replacement for scanning code spans by hand: a
1127 /// `verbatim` / `code_block` / `raw_inline` / `raw_block` selector recovers
1128 /// those, and every other node kind is reachable too.
1129 pub fn query(&mut self, selector: &str) -> Result<Vec<QueryMatch>, Error> {
1130 let raw = self.raw.as_ptr();
1131 collect_matches(|ptr, len| unsafe {
1132 ffi::twig_document_query(raw, selector.as_ptr(), selector.len(), ptr, len)
1133 })
1134 }
1135
1136 /// Return the whole source span of `node` without running a selector query.
1137 pub fn span(&mut self, node: NodeId) -> Result<Range<usize>, Error> {
1138 let mut span = ffi::TwigSpan { start: 0, end: 0 };
1139 let status = unsafe { ffi::twig_document_node_span(self.raw.as_ptr(), node.0, &mut span) };
1140 Error::from_status(status)?;
1141 Ok(span.start..span.end)
1142 }
1143
1144 /// Return the interior span of `node`, or `None` when the node has no
1145 /// recorded content span.
1146 pub fn content_span(&mut self, node: NodeId) -> Result<Option<Range<usize>>, Error> {
1147 let mut span = ffi::TwigSpan { start: 0, end: 0 };
1148 let status =
1149 unsafe { ffi::twig_document_node_content_span(self.raw.as_ptr(), node.0, &mut span) };
1150 match status.0 {
1151 ffi::TwigStatus::OK => Ok(Some(span.start..span.end)),
1152 ffi::TwigStatus::NOT_FOUND => Ok(None),
1153 _ => Err(Error::from_status(status).unwrap_err()),
1154 }
1155 }
1156
1157 /// The span of `node`'s own leading MARKER — the leading bytes a rich view
1158 /// HIDES on its opening line — or `None` when it has none. See
1159 /// [`FlatNode::marker_span`], which is the same answer inside a snapshot.
1160 pub fn marker_span(&mut self, node: NodeId) -> Result<Option<Range<usize>>, Error> {
1161 let mut span = ffi::TwigSpan { start: 0, end: 0 };
1162 let status =
1163 unsafe { ffi::twig_document_node_marker_span(self.raw.as_ptr(), node.0, &mut span) };
1164 match status.0 {
1165 ffi::TwigStatus::OK => Ok(Some(span.start..span.end)),
1166 ffi::TwigStatus::NOT_FOUND => Ok(None),
1167 _ => Err(Error::from_status(status).unwrap_err()),
1168 }
1169 }
1170
1171 /// The source span of the `{...}` attribute block attached to `node` — the
1172 /// bytes a lossless serializer re-emits instead of the flattened
1173 /// [`FlatNode::attrs`] projection, which a multi-line option block, or a
1174 /// nested or array-valued entry, says more than.
1175 ///
1176 /// `None` when the node has no attributes, or has some with no single
1177 /// recorded range: a synthesized set, or one merged from several source
1178 /// blocks. That is the case a caller has to handle rather than assume away
1179 /// — without this the only way to find an attribute block's extent was to
1180 /// scan the source for `{` near the node, which reads a `{` in prose as an
1181 /// attribute block and strands a real one that a heuristic missed.
1182 ///
1183 /// An accessor rather than a [`FlatNode`] field because the range is per
1184 /// attribute BLOCK, not per `(key, value)` pair.
1185 pub fn attrs_span(&mut self, node: NodeId) -> Result<Option<Range<usize>>, Error> {
1186 let mut span = ffi::TwigSpan { start: 0, end: 0 };
1187 let status =
1188 unsafe { ffi::twig_document_attrs_span(self.raw.as_ptr(), node.0, &mut span) };
1189 match status.0 {
1190 ffi::TwigStatus::OK => Ok(Some(span.start..span.end)),
1191 ffi::TwigStatus::NOT_FOUND => Ok(None),
1192 _ => Err(Error::from_status(status).unwrap_err()),
1193 }
1194 }
1195
1196 /// Everything HIDDEN before the content on the line byte `offset` sits on:
1197 /// every marker a node OPENS that line with, and the indentation between
1198 /// them, as one range running from the line start.
1199 ///
1200 /// This is the assembled form of [`FlatNode::marker_span`], which records
1201 /// each node's own marker alone. `> 1. [ ] ` is four nodes' markers plus
1202 /// the spaces between them, and the union is contiguous from the line start
1203 /// — so a caller gets one range to hide, or one width for a caret to step
1204 /// over, rather than a chain to walk and stitch together itself.
1205 ///
1206 /// `None` when nothing opens on this line — a CONTINUATION line, the second
1207 /// line of a wrapped paragraph or of a block quote. That is a real answer
1208 /// rather than a gap: what a continuation line repeats is a different
1209 /// question (a quote re-emits `> `, a list item re-emits spaces) and is not
1210 /// answerable from marker spans. [`Error::InvalidArgument`] if `offset`
1211 /// exceeds the source length.
1212 pub fn line_prefix(&mut self, offset: usize) -> Result<Option<Range<usize>>, Error> {
1213 let mut span = ffi::TwigSpan { start: 0, end: 0 };
1214 let status =
1215 unsafe { ffi::twig_document_line_prefix(self.raw.as_ptr(), offset, &mut span) };
1216 match status.0 {
1217 ffi::TwigStatus::OK => Ok(Some(span.start..span.end)),
1218 ffi::TwigStatus::NOT_FOUND => Ok(None),
1219 _ => Err(Error::from_status(status).unwrap_err()),
1220 }
1221 }
1222
1223 /// What a CONTINUATION LINE at `offset` must open with to stay inside every
1224 /// container holding it.
1225 ///
1226 /// The other half of [`Document::line_prefix`], and not derivable from it.
1227 /// That one reports the bytes ALREADY THERE on a line something opens, so it
1228 /// hands back a range into the source. This one reports the bytes that WOULD
1229 /// HAVE TO BE WRITTEN on a line nothing opens — a list item's continuation
1230 /// is spaces where its marker was, which is not source at all, so it is
1231 /// built rather than pointed at.
1232 ///
1233 /// A quote's `> ` is REPRODUCED (dropping it ends the quote); a list item's
1234 /// marker becomes its WIDTH IN SPACES (repeating it would open a second
1235 /// item). Each container on the caret's chain contributes the columns its
1236 /// own marker occupies, on its own opening line — which may be a different
1237 /// line for each of them, and is why this is a tree walk rather than a
1238 /// re-read of one line:
1239 ///
1240 /// ```text
1241 /// > - a quote "> " + item "- " as width -> "> "
1242 /// - a
1243 /// - b outer item + inner item -> " "
1244 /// ```
1245 ///
1246 /// Empty at the top level, which is the correct prefix there: none.
1247 /// [`Error::InvalidArgument`] if `offset` exceeds the source length.
1248 pub fn continuation_prefix(&mut self, offset: usize) -> Result<LinePrefix, Error> {
1249 self.prefix_via(offset, ffi::twig_document_continuation_prefix)
1250 }
1251
1252 /// What a BLANK line inside the containers at `offset` must carry.
1253 ///
1254 /// A quote's blank line still has to carry its `>` or the quote ENDS there;
1255 /// a list item's must carry nothing, because a blank line between two of an
1256 /// item's blocks is what makes its list loose and indenting it changes
1257 /// nothing about that. So this is [`Document::continuation_prefix`] with its
1258 /// trailing spaces cut back — which drops an item's indent entirely and
1259 /// leaves a quote marker standing.
1260 ///
1261 /// The quote form is `>` and not `> `, because the space after the marker is
1262 /// content indentation and a blank line has no content.
1263 pub fn blank_line_prefix(&mut self, offset: usize) -> Result<LinePrefix, Error> {
1264 self.prefix_via(offset, ffi::twig_document_blank_line_prefix)
1265 }
1266
1267 /// Shared marshalling for the two prefix builders above.
1268 fn prefix_via(
1269 &mut self,
1270 offset: usize,
1271 f: unsafe extern "C" fn(
1272 *mut ffi::TwigDocument,
1273 usize,
1274 *mut *const u8,
1275 *mut usize,
1276 *mut usize,
1277 ) -> ffi::TwigStatus,
1278 ) -> Result<LinePrefix, Error> {
1279 let mut ptr: *const u8 = std::ptr::null();
1280 let mut len = 0usize;
1281 let mut columns = 0usize;
1282 let status = unsafe { f(self.raw.as_ptr(), offset, &mut ptr, &mut len, &mut columns) };
1283 Error::from_status(status)?;
1284 let text = if ptr.is_null() || len == 0 {
1285 String::new()
1286 } else {
1287 let bytes = unsafe { std::slice::from_raw_parts(ptr, len) };
1288 String::from_utf8(bytes.to_vec()).map_err(|_| Error::Internal)?
1289 };
1290 Ok(LinePrefix { text, columns })
1291 }
1292
1293 /// The grid extent of the cell at `node` — how many `(columns, rows)` it
1294 /// occupies — or `None` when the node is not a cell. Both are at least 1,
1295 /// and `(1, 1)` is the ordinary one-square cell; anything larger is a merged
1296 /// cell from a format with a real grid (HTML's `colspan`/`rowspan`, an rST
1297 /// grid table). GFM and djot pipe tables always report `(1, 1)`.
1298 ///
1299 /// HTML's `rowspan="0"` ("to the end of the row group") is not a count and
1300 /// reports 1; the source spelling survives on the node's attributes.
1301 ///
1302 /// This is an accessor rather than a [`FlatNode`] field because the C struct
1303 /// it snapshots is ABI-frozen — see [`Document::span`] for the same shape.
1304 pub fn cell_extent(&mut self, node: NodeId) -> Result<Option<(u32, u32)>, Error> {
1305 let raw = self.raw.as_ptr();
1306 let mut colspan: u32 = 0;
1307 let status = unsafe { ffi::twig_document_cell_colspan(raw, node.0, &mut colspan) };
1308 match status.0 {
1309 ffi::TwigStatus::OK => {}
1310 ffi::TwigStatus::NOT_FOUND => return Ok(None),
1311 _ => return Err(Error::from_status(status).unwrap_err()),
1312 }
1313 let mut rowspan: u32 = 0;
1314 Error::from_status(unsafe { ffi::twig_document_cell_rowspan(raw, node.0, &mut rowspan) })?;
1315 Ok(Some((colspan, rowspan)))
1316 }
1317
1318 /// Snapshot the whole tree as a flat [`FlatNode`] array (the JSON-free read
1319 /// path for a renderer), indexed so `nodes[i].id == NodeId(i)`. Walk it via
1320 /// the `parent`/`first_child`/`next_sibling` links; the root is the node
1321 /// whose `parent` is `None`.
1322 pub fn nodes(&mut self) -> Result<Vec<FlatNode>, Error> {
1323 let raw = self.raw.as_ptr();
1324 collect_flat_nodes(|ptr, len| unsafe { ffi::twig_document_nodes(raw, ptr, len) })
1325 }
1326
1327 /// The document-level **definitions**: every node that hangs off no parent
1328 /// and is not the document root, in arena order. Usually empty.
1329 ///
1330 /// A parsed document is not one tree. Footnote definitions and
1331 /// link-reference definitions are resolved by LABEL rather than by
1332 /// position, so twig attaches them to nothing — walking from the root over
1333 /// [`FlatNode::first_child`] never reaches them, and a renderer that wants
1334 /// to resolve `[^1]` has to find the definition some other way. This is
1335 /// that way, and it replaces scanning the whole [`Document::nodes`] array
1336 /// for entries whose `parent` is `None`.
1337 ///
1338 /// Not filtered to a kind list: WHICH kinds end up detached is a property
1339 /// of how a format resolves its definitions (djot and Markdown detach
1340 /// [`Kind::Footnote`] and [`Kind::Reference`]; rST adds [`Kind::Citation`]
1341 /// and [`Kind::Substitution`]), not something a caller should enumerate.
1342 /// Read the [`kind`](QueryMatch::kind) on each match.
1343 pub fn definitions(&mut self) -> Result<Vec<QueryMatch>, Error> {
1344 let raw = self.raw.as_ptr();
1345 collect_matches(|ptr, len| unsafe { ffi::twig_document_definitions(raw, ptr, len) })
1346 }
1347
1348 /// What converting this document to `target` would silently **lose**: one
1349 /// [`Warning`] per lossy node — two at one path for a node whose
1350 /// attributes are lost as well as its kind — in document order. An empty vec means the
1351 /// conversion is lossless.
1352 ///
1353 /// Twig's serializers degrade or drop a node whenever the target has no
1354 /// spelling for it — a djot `{=mark=}` written into Markdown comes back as
1355 /// plain text, an HTML comment converted to djot vanishes entirely. None of
1356 /// it is an error, so all of it happens quietly. This is the call that makes
1357 /// it loud, and it replaces guessing from the outside: the answers are
1358 /// measured against the serializers by a round-trip probe in the Zig
1359 /// library, not asserted.
1360 ///
1361 /// The answer belongs to the (document, target) PAIR, not to the document —
1362 /// the same document has different answers for different targets, which is
1363 /// why this takes one and why nothing is cached on [`Document`] itself.
1364 ///
1365 /// [`Error::UnsupportedFormat`] for a target with no serializer at all
1366 /// ([`Target::Xml`]): "this cannot be written" is a capability answer,
1367 /// not a per-node diagnosis.
1368 pub fn diagnostics(&mut self, target: Target) -> Result<Vec<Warning>, Error> {
1369 let raw = self.raw.as_ptr();
1370 let code = ffi::TwigFormat::from(target) as c_int;
1371 let mut ptr: *const ffi::TwigWarning = std::ptr::null();
1372 let mut len = 0usize;
1373 let status = unsafe { ffi::twig_document_diagnostics(raw, code, &mut ptr, &mut len) };
1374 Error::from_status(status)?;
1375 if len == 0 || ptr.is_null() {
1376 return Ok(Vec::new());
1377 }
1378 let raw_warnings = unsafe { std::slice::from_raw_parts(ptr, len) };
1379 Ok(raw_warnings
1380 .iter()
1381 .map(|w| Warning {
1382 fidelity: Fidelity::from_c(w.fidelity),
1383 path: borrowed_bytes(w.path_ptr, w.path_len).unwrap_or_default(),
1384 kind: Kind::from(borrowed_cstr(w.kind).unwrap_or_default().as_str()),
1385 })
1386 .collect())
1387 }
1388
1389 /// The direct children of `node` as [`QueryMatch`]es (id, span, kind) —
1390 /// `None` enumerates the document root's children (the top-level blocks).
1391 /// The cheap enumeration an incremental renderer walks to decide which
1392 /// blocks to re-marshal with [`Document::subtree`]. A childless node yields
1393 /// an empty vec.
1394 pub fn children(&mut self, node: Option<NodeId>) -> Result<Vec<QueryMatch>, Error> {
1395 let raw = self.raw.as_ptr();
1396 let id = node.map_or(ffi::TWIG_NO_NODE, |n| n.0);
1397 collect_matches(|ptr, len| unsafe { ffi::twig_document_children(raw, id, ptr, len) })
1398 }
1399
1400 /// Snapshot the subtree rooted at `node` as a self-contained [`FlatNode`]
1401 /// array with *local* ids: `array[0]` is the root, every link is an index
1402 /// into the returned vec (or `None`), and spans stay absolute. The root's
1403 /// `parent` and `next_sibling` are `None`, so a walk from index 0 stays
1404 /// inside the subtree. [`Error::InvalidArgument`] if `node` is out of range.
1405 pub fn subtree(&mut self, node: NodeId) -> Result<Vec<FlatNode>, Error> {
1406 let raw = self.raw.as_ptr();
1407 collect_flat_nodes(|ptr, len| unsafe { ffi::twig_document_subtree(raw, node.0, ptr, len) })
1408 }
1409
1410 /// The deepest node whose span contains byte `offset` (with `offset` equal
1411 /// to the source length treated as inside the root) — hit-testing and
1412 /// cursor context. `Ok(None)` if no node covers the offset;
1413 /// [`Error::InvalidArgument`] if `offset` exceeds the source length.
1414 pub fn node_at(&mut self, offset: usize) -> Result<Option<QueryMatch>, Error> {
1415 let mut m = empty_ffi_match();
1416 let status = unsafe { ffi::twig_document_node_at(self.raw.as_ptr(), offset, &mut m) };
1417 match status.0 {
1418 ffi::TwigStatus::OK => Ok(Some(query_match_from_ffi(&m)?)),
1419 ffi::TwigStatus::NOT_FOUND => Ok(None),
1420 _ => Err(Error::from_status(status).unwrap_err()),
1421 }
1422 }
1423
1424 /// The chain of nodes containing byte `offset`, root-first down to the
1425 /// deepest (the node [`Document::node_at`] returns) — the ancestor path for
1426 /// a breadcrumb. Empty if no node covers the offset.
1427 pub fn ancestors_at(&mut self, offset: usize) -> Result<Vec<QueryMatch>, Error> {
1428 let raw = self.raw.as_ptr();
1429 let mut ptr: *const ffi::TwigQueryMatch = std::ptr::null();
1430 let mut len = 0usize;
1431 let status = unsafe { ffi::twig_document_nodes_at(raw, offset, &mut ptr, &mut len) };
1432 match status.0 {
1433 ffi::TwigStatus::OK => {}
1434 ffi::TwigStatus::NOT_FOUND => return Ok(Vec::new()),
1435 _ => return Err(Error::from_status(status).unwrap_err()),
1436 }
1437 if len == 0 || ptr.is_null() {
1438 return Ok(Vec::new());
1439 }
1440 let raw_matches = unsafe { std::slice::from_raw_parts(ptr, len) };
1441 raw_matches.iter().map(query_match_from_ffi).collect()
1442 }
1443
1444 /// [`Document::node_at`] under CARET containment — the same descent, under
1445 /// the rule an editing caret needs rather than the one a byte range needs.
1446 ///
1447 /// Two differences, both because a caret is a position BETWEEN bytes while a
1448 /// span is a range OF bytes:
1449 ///
1450 /// 1. **A block's end is inside it.** A caret after the last character of a
1451 /// paragraph is *in* that paragraph — it is where you stand to type the
1452 /// rest of it. Half-open containment puts it outside, which is why a
1453 /// consumer probing [`Document::ancestors_at`] ends up guessing at
1454 /// contrived offsets (the content start, `caret - 1`, a marker byte) to
1455 /// find the block it was plainly inside of.
1456 ///
1457 /// 2. **A trailing newline is not part of the block**, which is what makes
1458 /// the two authorable formats AGREE. Djot ends a paragraph's span after
1459 /// its newline and Markdown before it, so on `"a\n\nb\n"` the caret at
1460 /// offset 1 read as `para` through Djot and `doc` through Markdown — the
1461 /// same caret, two answers, decided by which parser produced the tree.
1462 ///
1463 /// Never `Ok(None)` for a non-empty document: a caret in the gap between two
1464 /// blocks reports the container holding the gap (usually the root) rather
1465 /// than nothing at all.
1466 pub fn node_at_caret(&mut self, offset: usize) -> Result<Option<QueryMatch>, Error> {
1467 let mut m = empty_ffi_match();
1468 let status = unsafe { ffi::twig_document_node_at_caret(self.raw.as_ptr(), offset, &mut m) };
1469 match status.0 {
1470 ffi::TwigStatus::OK => Ok(Some(query_match_from_ffi(&m)?)),
1471 ffi::TwigStatus::NOT_FOUND => Ok(None),
1472 _ => Err(Error::from_status(status).unwrap_err()),
1473 }
1474 }
1475
1476 /// [`Document::ancestors_at`] under caret containment — root-first down to
1477 /// the node [`Document::node_at_caret`] returns. See that method for the
1478 /// containment rule and why it differs.
1479 pub fn ancestors_at_caret(&mut self, offset: usize) -> Result<Vec<QueryMatch>, Error> {
1480 let raw = self.raw.as_ptr();
1481 let mut ptr: *const ffi::TwigQueryMatch = std::ptr::null();
1482 let mut len = 0usize;
1483 let status = unsafe { ffi::twig_document_nodes_at_caret(raw, offset, &mut ptr, &mut len) };
1484 match status.0 {
1485 ffi::TwigStatus::OK => {}
1486 ffi::TwigStatus::NOT_FOUND => return Ok(Vec::new()),
1487 _ => return Err(Error::from_status(status).unwrap_err()),
1488 }
1489 if len == 0 || ptr.is_null() {
1490 return Ok(Vec::new());
1491 }
1492 let raw_matches = unsafe { std::slice::from_raw_parts(ptr, len) };
1493 raw_matches.iter().map(query_match_from_ffi).collect()
1494 }
1495}
1496
1497/// A [`Document`] borrowed from an [`Editor`] (see [`Editor::document`]): the
1498/// editor's live tree behind the whole document read surface, without a parse.
1499///
1500/// It holds the editor mutably borrowed for as long as it lives, so the tree —
1501/// and every node id and span read out of it — cannot change underneath it.
1502/// Dropping it frees nothing; the editor owns the tree.
1503///
1504/// [`Document::render_html`] and [`Document::serialize`] are the two methods it
1505/// cannot serve ([`Error::UnsupportedFormat`] — they need a real parse's
1506/// language tag and side tables). Parse [`Editor::source`] for those.
1507#[derive(Debug)]
1508pub struct DocumentView<'a> {
1509 doc: Document,
1510 _editor: PhantomData<&'a mut Editor>,
1511}
1512
1513impl std::ops::Deref for DocumentView<'_> {
1514 type Target = Document;
1515
1516 fn deref(&self) -> &Document {
1517 &self.doc
1518 }
1519}
1520
1521impl std::ops::DerefMut for DocumentView<'_> {
1522 fn deref_mut(&mut self) -> &mut Document {
1523 &mut self.doc
1524 }
1525}
1526
1527impl Drop for Document {
1528 fn drop(&mut self) {
1529 unsafe { ffi::twig_document_destroy(self.raw.as_ptr()) }
1530 }
1531}
1532
1533/// Opt-in Markdown extensions to enable for a parse — for either the read path
1534/// ([`Document::parse_with`]) or the edit path ([`Editor::new_ext`]). Ignored
1535/// for non-Markdown formats. Every field defaults off, matching the library.
1536///
1537/// These lay **over** whichever Markdown dialect the [`Format`] named —
1538/// `Format::Gfm` with `math` is GFM plus math — and the default-on set
1539/// (tables, strikethrough, task lists, …) is the dialect's to decide, which is
1540/// why there is no field to turn one off: that is what [`Format::Commonmark`]
1541/// is.
1542#[derive(Clone, Copy, Debug, Default, Eq, PartialEq)]
1543pub struct MarkdownExtensions {
1544 /// Generic directives: `:name`, `::name`, `:::name`.
1545 pub directives: bool,
1546 /// `$...$` / `$$...$$` math.
1547 pub math: bool,
1548 /// Parse recognized raw HTML into semantic AST nodes — an `<img>` becomes an
1549 /// [`image` node](FlatNode) instead of an opaque `raw_block`/`raw_inline`, so
1550 /// it is addressable by [`Document::query`] and the tree read paths. Only
1551 /// tags that map verbatim onto the source are promoted; the rest stay raw.
1552 /// A bare `<div>` line paired with its bare `</div>` line, and a `<span>`
1553 /// paired with its `</span>` in the same run, become one container over
1554 /// what lies between — the spellings twig writes for a block's and a run's
1555 /// attributes, and what this flag reads back.
1556 pub html_elements: bool,
1557 /// `==text==` highlight, parsed as a `mark` node (the markdown-it-mark /
1558 /// Obsidian extension). Only a run of exactly two `=` delimits.
1559 pub highlight: bool,
1560 /// Coloured highlights on top of `highlight` (Obsidian 1.14): a circle
1561 /// emoji right after the opening `==` — `==🔴 text==` — is stripped from
1562 /// the content and recorded as the mark's `data-color` attribute
1563 /// (`red`, `orange`, `yellow`, `green`, `blue`, `purple`, `brown`). Inert
1564 /// unless `highlight` is also set.
1565 pub highlight_colors: bool,
1566}
1567
1568impl MarkdownExtensions {
1569 fn to_flags(self) -> u32 {
1570 let mut flags = 0;
1571 if self.directives {
1572 flags |= ffi::TWIG_MD_DIRECTIVES;
1573 }
1574 if self.math {
1575 flags |= ffi::TWIG_MD_MATH;
1576 }
1577 if self.html_elements {
1578 flags |= ffi::TWIG_MD_HTML_ELEMENTS;
1579 }
1580 if self.highlight {
1581 flags |= ffi::TWIG_MD_HIGHLIGHT;
1582 }
1583 if self.highlight_colors {
1584 flags |= ffi::TWIG_MD_HIGHLIGHT_COLORS;
1585 }
1586 flags
1587 }
1588}
1589
1590/// A span-splice editor over a document: applies lossless, in-place edits and
1591/// reparses after each one, so node addressing stays valid as the document
1592/// evolves. Every op is addressed by a `locator` — a dot-separated index path
1593/// (`"0.3.1"`) or a selector that must match exactly one node
1594/// (`heading("Status")`). A failed edit leaves the document unchanged.
1595#[derive(Debug)]
1596pub struct Editor {
1597 raw: NonNull<ffi::TwigEditor>,
1598}
1599
1600impl Editor {
1601 /// Create an editor over a private copy of `input`, parsed as `format` with
1602 /// default options.
1603 pub fn new(input: &[u8], format: Format) -> Result<Self, Error> {
1604 let mut raw = std::ptr::null_mut();
1605 let ffi_format: ffi::TwigFormat = format.into();
1606 let status = unsafe {
1607 ffi::twig_editor_create(input.as_ptr(), input.len(), ffi_format as i32, &mut raw)
1608 };
1609 Error::from_status(status)?;
1610 let raw = NonNull::new(raw).ok_or(Error::Internal)?;
1611 Ok(Self { raw })
1612 }
1613
1614 pub fn new_str(input: &str, format: Format) -> Result<Self, Error> {
1615 Self::new(input.as_bytes(), format)
1616 }
1617
1618 /// Like [`Editor::new`], plus Markdown `extensions` to enable (ignored for
1619 /// other formats). The editor reparses with these after every edit, so a
1620 /// directive-bearing document stays parseable — needed before
1621 /// [`Editor::filter`] can match `directive[...]` selectors.
1622 ///
1623 /// They also decide what the authoring gestures may **write**, since a
1624 /// gesture may only mint bytes this editor's own reparse reads back:
1625 /// [`MarkdownExtensions::highlight`] makes `==x==` a highlight
1626 /// [`Editor::toggle_inline`] can add and remove, and
1627 /// [`MarkdownExtensions::highlight_colors`] makes
1628 /// [`Editor::set_mark_color`] available on top of it. Without them those
1629 /// calls are [`Error::UnsupportedFormat`] — see [`Format::supports_with`],
1630 /// which answers for the extensions rather than for the format alone.
1631 pub fn new_ext(
1632 input: &[u8],
1633 format: Format,
1634 extensions: MarkdownExtensions,
1635 ) -> Result<Self, Error> {
1636 let mut raw = std::ptr::null_mut();
1637 let ffi_format: ffi::TwigFormat = format.into();
1638 let status = unsafe {
1639 ffi::twig_editor_create_ext(
1640 input.as_ptr(),
1641 input.len(),
1642 ffi_format as i32,
1643 extensions.to_flags(),
1644 &mut raw,
1645 )
1646 };
1647 Error::from_status(status)?;
1648 let raw = NonNull::new(raw).ok_or(Error::Internal)?;
1649 Ok(Self { raw })
1650 }
1651
1652 /// Replace the whole source of the located node with `text`.
1653 pub fn replace(&mut self, locator: &str, text: &str) -> Result<(), Error> {
1654 self.apply(locator, text, |ed, loc, loc_len, txt, txt_len| unsafe {
1655 ffi::twig_editor_replace(ed, loc, loc_len, txt, txt_len)
1656 })
1657 }
1658
1659 /// Replace the interior (between-delimiters content) of the located
1660 /// container.
1661 pub fn replace_content(&mut self, locator: &str, text: &str) -> Result<(), Error> {
1662 self.apply(locator, text, |ed, loc, loc_len, txt, txt_len| unsafe {
1663 ffi::twig_editor_replace_content(ed, loc, loc_len, txt, txt_len)
1664 })
1665 }
1666
1667 /// Insert `text` immediately before the located node.
1668 pub fn insert_before(&mut self, locator: &str, text: &str) -> Result<(), Error> {
1669 self.apply(locator, text, |ed, loc, loc_len, txt, txt_len| unsafe {
1670 ffi::twig_editor_insert_before(ed, loc, loc_len, txt, txt_len)
1671 })
1672 }
1673
1674 /// Insert `text` immediately after the located node.
1675 pub fn insert_after(&mut self, locator: &str, text: &str) -> Result<(), Error> {
1676 self.apply(locator, text, |ed, loc, loc_len, txt, txt_len| unsafe {
1677 ffi::twig_editor_insert_after(ed, loc, loc_len, txt, txt_len)
1678 })
1679 }
1680
1681 /// Insert `text` as the `index`-th child of the located container (an index
1682 /// at or past the child count appends).
1683 pub fn insert_child(&mut self, locator: &str, index: usize, text: &str) -> Result<(), Error> {
1684 let status = unsafe {
1685 ffi::twig_editor_insert_child(
1686 self.raw.as_ptr(),
1687 locator.as_ptr(),
1688 locator.len(),
1689 index,
1690 text.as_ptr(),
1691 text.len(),
1692 )
1693 };
1694 Error::from_status(status)
1695 }
1696
1697 /// Delete the located node (removes exactly its span; no whitespace
1698 /// cleanup).
1699 pub fn delete(&mut self, locator: &str) -> Result<(), Error> {
1700 let status =
1701 unsafe { ffi::twig_editor_delete(self.raw.as_ptr(), locator.as_ptr(), locator.len()) };
1702 Error::from_status(status)
1703 }
1704
1705 /// Delete the located node, tidying surrounding blank lines for a
1706 /// whole-line node — a block, or an element alone on an indented line,
1707 /// which goes with its indentation; an inline node degrades to the exact
1708 /// delete.
1709 pub fn delete_smart(&mut self, locator: &str) -> Result<(), Error> {
1710 let status = unsafe {
1711 ffi::twig_editor_delete_smart(self.raw.as_ptr(), locator.as_ptr(), locator.len())
1712 };
1713 Error::from_status(status)
1714 }
1715
1716 /// Unwrap the located node: replace it with its interior (drop the wrapper,
1717 /// keep the children) — e.g. peel a `:::vis{...}` container. A node with no
1718 /// interior (a leaf, or an empty container) is removed.
1719 pub fn unwrap_node(&mut self, locator: &str) -> Result<(), Error> {
1720 let status =
1721 unsafe { ffi::twig_editor_unwrap(self.raw.as_ptr(), locator.as_ptr(), locator.len()) };
1722 Error::from_status(status)
1723 }
1724
1725 /// Move the node `locator` names to immediately before the node `anchor`
1726 /// names — a canvas's "send backward", a list's reorder — in one splice
1727 /// and one undo step, the bytes between the two copied verbatim in their
1728 /// new order. The node travels with the whitespace run ahead of it (the
1729 /// line break and indentation a pretty-printed document separates
1730 /// siblings with), which lands after it here, so every sibling keeps its
1731 /// separator: `<g>\n <a/>\n <b/>\n</g>` reorders to
1732 /// `<g>\n <b/>\n <a/>\n</g>`, never to a line holding both. The rule is
1733 /// about bytes, not structure — a block quote's `> ` prefixes do not
1734 /// travel — and the anchor need not be a sibling: next to a node in
1735 /// another container is a reparent.
1736 ///
1737 /// [`Error::InvalidArgument`] when either node's span holds the other's,
1738 /// or the two are one node; [`Error::NotFound`] and [`Error::Ambiguous`]
1739 /// as every other tree op; [`Error::EditConflict`] when the moved document
1740 /// no longer parses, in which case nothing changed.
1741 pub fn move_before(&mut self, locator: &str, anchor: &str) -> Result<(), Error> {
1742 self.apply(locator, anchor, |ed, loc, loc_len, a, a_len| unsafe {
1743 ffi::twig_editor_move_before(ed, loc, loc_len, a, a_len)
1744 })
1745 }
1746
1747 /// Move the node `locator` names to immediately after the node `anchor`
1748 /// names — a canvas's "bring forward". The whitespace run ahead of the
1749 /// node travels with it and stays ahead of it. Otherwise
1750 /// [`Editor::move_before`].
1751 pub fn move_after(&mut self, locator: &str, anchor: &str) -> Result<(), Error> {
1752 self.apply(locator, anchor, |ed, loc, loc_len, a, a_len| unsafe {
1753 ffi::twig_editor_move_after(ed, loc, loc_len, a, a_len)
1754 })
1755 }
1756
1757 /// Prune the document in place: remove every node matching the `drop`
1758 /// selector except those also matching `keep` (`None` spares nothing),
1759 /// then — if `unwrap_kept` — unwrap the survivors. Read the result with
1760 /// [`Editor::source`].
1761 pub fn filter(
1762 &mut self,
1763 drop: &str,
1764 keep: Option<&str>,
1765 unwrap_kept: bool,
1766 ) -> Result<(), Error> {
1767 let (keep_ptr, keep_len) = match keep {
1768 Some(k) => (k.as_ptr(), k.len()),
1769 None => (std::ptr::null(), 0),
1770 };
1771 let status = unsafe {
1772 ffi::twig_editor_filter(
1773 self.raw.as_ptr(),
1774 drop.as_ptr(),
1775 drop.len(),
1776 keep_ptr,
1777 keep_len,
1778 unwrap_kept as i32,
1779 )
1780 };
1781 Error::from_status(status)
1782 }
1783
1784 /// The editor's current (edited) source bytes.
1785 pub fn source(&mut self) -> Result<Vec<u8>, Error> {
1786 let raw = self.raw.as_ptr();
1787 collect_bytes(|ptr, len| unsafe { ffi::twig_editor_source(raw, ptr, len) })
1788 }
1789
1790 /// The editor's current source bytes as a UTF-8 string.
1791 pub fn source_str(&mut self) -> Result<String, Error> {
1792 String::from_utf8(self.source()?).map_err(|_| Error::Internal)
1793 }
1794
1795 /// Encode the editor's current tree as pretty-printed JSON — the live
1796 /// counterpart of [`Document::ast_json`], for inspecting between edits.
1797 pub fn ast_json(&mut self) -> Result<Vec<u8>, Error> {
1798 let raw = self.raw.as_ptr();
1799 collect_bytes(|ptr, len| unsafe { ffi::twig_editor_ast_json(raw, ptr, len) })
1800 }
1801
1802 /// Resolve a selector against the editor's current tree — the live
1803 /// counterpart of [`Document::query`].
1804 pub fn query(&mut self, selector: &str) -> Result<Vec<QueryMatch>, Error> {
1805 let raw = self.raw.as_ptr();
1806 collect_matches(|ptr, len| unsafe {
1807 ffi::twig_editor_query(raw, selector.as_ptr(), selector.len(), ptr, len)
1808 })
1809 }
1810
1811 // ── offset-addressed editing & read-back ────────────────────────────────
1812
1813 /// Splice `[start, end)` of the current source with `text`, reparse, and
1814 /// return the [`Change`] the edit produced — the offset-addressed primitive
1815 /// a caret editor is built on: a keystroke is `edit_range(c, c, "x")`,
1816 /// backspace `edit_range(c - 1, c, "")`, a selection replace
1817 /// `edit_range(a, b, s)`. `start <= end <= ` source length, else
1818 /// [`Error::InvalidArgument`]. A reparse-breaking edit is rolled back and
1819 /// returns [`Error::EditConflict`], leaving the document untouched.
1820 pub fn edit_range(&mut self, start: usize, end: usize, text: &str) -> Result<Change, Error> {
1821 let mut change = ffi::TwigChange {
1822 old_span: ffi::TwigSpan { start: 0, end: 0 },
1823 new_span: ffi::TwigSpan { start: 0, end: 0 },
1824 };
1825 let status = unsafe {
1826 ffi::twig_editor_edit_range(
1827 self.raw.as_ptr(),
1828 start,
1829 end,
1830 text.as_ptr(),
1831 text.len(),
1832 &mut change,
1833 )
1834 };
1835 Error::from_status(status)?;
1836 Ok(Change::from_ffi(change))
1837 }
1838
1839 /// The byte effect of the last successful edit — including the locator ops
1840 /// ([`Editor::replace`], [`Editor::delete_smart`], …), so any edit can
1841 /// re-anchor a caret without re-diffing. `None` before the first successful
1842 /// edit. (A multi-splice op such as [`Editor::filter`] reports only its
1843 /// final splice.)
1844 pub fn last_change(&mut self) -> Option<Change> {
1845 let mut change = ffi::TwigChange {
1846 old_span: ffi::TwigSpan { start: 0, end: 0 },
1847 new_span: ffi::TwigSpan { start: 0, end: 0 },
1848 };
1849 let status = unsafe { ffi::twig_editor_last_change(self.raw.as_ptr(), &mut change) };
1850 match status.0 {
1851 ffi::TwigStatus::OK => Some(Change::from_ffi(change)),
1852 _ => None,
1853 }
1854 }
1855
1856 /// Undo the last edit step, restoring the previous source and reparsing.
1857 /// Returns the [`Change`] the undo produced (current → restored) so a caret
1858 /// can re-anchor, or `None` when there's nothing to undo. History accrues
1859 /// across every successful edit that funnels through the splice primitive.
1860 pub fn undo(&mut self) -> Result<Option<Change>, Error> {
1861 let mut change = ffi::TwigChange {
1862 old_span: ffi::TwigSpan { start: 0, end: 0 },
1863 new_span: ffi::TwigSpan { start: 0, end: 0 },
1864 };
1865 let status = unsafe { ffi::twig_editor_undo(self.raw.as_ptr(), &mut change) };
1866 if status.0 == ffi::TwigStatus::NOT_FOUND {
1867 return Ok(None);
1868 }
1869 Error::from_status(status)?;
1870 Ok(Some(Change::from_ffi(change)))
1871 }
1872
1873 /// Redo the most recently undone edit step; the inverse of [`Editor::undo`].
1874 /// Returns `None` when the redo stack is empty (nothing undone, or a fresh
1875 /// edit has invalidated it).
1876 pub fn redo(&mut self) -> Result<Option<Change>, Error> {
1877 let mut change = ffi::TwigChange {
1878 old_span: ffi::TwigSpan { start: 0, end: 0 },
1879 new_span: ffi::TwigSpan { start: 0, end: 0 },
1880 };
1881 let status = unsafe { ffi::twig_editor_redo(self.raw.as_ptr(), &mut change) };
1882 if status.0 == ffi::TwigStatus::NOT_FOUND {
1883 return Ok(None);
1884 }
1885 Error::from_status(status)?;
1886 Ok(Some(Change::from_ffi(change)))
1887 }
1888
1889 /// Fold the most recent edit into the undo step before it, so a caret editor
1890 /// can coalesce a run of keystrokes into a single undo. Call right after an
1891 /// `edit_range` that continues a run (same kind, no intervening caret move);
1892 /// a no-op unless there are at least two steps to merge.
1893 pub fn coalesce_last_undo(&mut self) -> Result<(), Error> {
1894 let status = unsafe { ffi::twig_editor_coalesce_last(self.raw.as_ptr()) };
1895 Error::from_status(status)
1896 }
1897
1898 /// A monotonic change token, bumped once per successful mutation of the
1899 /// document (every edit and every undo/redo). Never decreases and never
1900 /// repeats for the life of the editor; the initial parse is revision 0.
1901 /// Equal revision means a byte-identical document, so it can key a cache
1902 /// instead of hand-tracking "did anything change?".
1903 pub fn revision(&mut self) -> u64 {
1904 unsafe { ffi::twig_editor_revision(self.raw.as_ptr()) }
1905 }
1906
1907 /// The cumulative dirty byte range since the last [`Editor::clear_dirty`]
1908 /// (or since the editor was created) — the union of every mutation's byte
1909 /// effect over that window, in current source coordinates — or `None` when
1910 /// the document is clean relative to the last clear.
1911 ///
1912 /// The incremental-rebuild companion to [`Editor::revision`]: `revision`
1913 /// says *whether* a cached view (glyph rows, syntax spans) needs rebuilding,
1914 /// this says *which bytes* changed, so a consumer rebuilds only the affected
1915 /// part instead of the whole document. A single conservative interval: it
1916 /// always covers every changed byte and may over-cover the gap between edits
1917 /// to disjoint regions, but never under-covers.
1918 ///
1919 /// It reports where *bytes* differ — exact, because twig splices losslessly
1920 /// and never reflows untouched bytes — not where the *parse* differs. An
1921 /// edit can reinterpret bytes outside the range (opening a code fence, a `#`
1922 /// promoting a paragraph to a heading), so a consumer rebuilding *structure*
1923 /// from it should widen the range to the enclosing block(s) itself (e.g. via
1924 /// [`Editor::node_at`] on each end). Typical loop: on a repaint, if
1925 /// [`Editor::revision`] moved, read this range, rebuild the rows it (widened)
1926 /// covers, then call [`Editor::clear_dirty`].
1927 pub fn dirty_range(&mut self) -> Option<Range<usize>> {
1928 let mut span = ffi::TwigSpan { start: 0, end: 0 };
1929 let status = unsafe { ffi::twig_editor_dirty_range(self.raw.as_ptr(), &mut span) };
1930 match status.0 {
1931 ffi::TwigStatus::OK => Some(span.start..span.end),
1932 _ => None,
1933 }
1934 }
1935
1936 /// Acknowledge the current dirty range: mark the document clean so a later
1937 /// [`Editor::dirty_range`] reports only mutations made after this call. Call
1938 /// it once you've consumed the range (rebuilt the affected view). Leaves the
1939 /// document, [`Editor::revision`], and [`Editor::last_change`] untouched.
1940 pub fn clear_dirty(&mut self) {
1941 unsafe { ffi::twig_editor_clear_dirty(self.raw.as_ptr()) };
1942 }
1943
1944 /// Attach an opaque, caller-owned blob (e.g. a serialized caret/selection)
1945 /// to the editor's current document state. Twig copies the bytes and never
1946 /// interprets them; it only carries them through the undo history so
1947 /// [`Editor::undo`]/[`Editor::redo`] hand back the caret matching the
1948 /// restored source (via [`Editor::caret_blob`]). Set it with the pre-edit
1949 /// caret *before* an edit so the retired undo step captures it. An empty
1950 /// blob clears the current caret.
1951 pub fn set_caret_blob(&mut self, blob: &[u8]) -> Result<(), Error> {
1952 let status = unsafe {
1953 ffi::twig_editor_set_caret_blob(self.raw.as_ptr(), blob.as_ptr(), blob.len())
1954 };
1955 Error::from_status(status)
1956 }
1957
1958 /// The opaque caret blob for the editor's current document state (see
1959 /// [`Editor::set_caret_blob`]). After [`Editor::undo`]/[`Editor::redo`] this
1960 /// is the restored state's caret; after an edit it is empty until set again.
1961 /// Returns an owned copy, so it outlives the next edit.
1962 pub fn caret_blob(&mut self) -> Result<Vec<u8>, Error> {
1963 let raw = self.raw.as_ptr();
1964 collect_bytes(|ptr, len| unsafe { ffi::twig_editor_caret_blob(raw, ptr, len) })
1965 }
1966
1967 /// The editor's current tree as a borrowed [`Document`], so the whole
1968 /// document read surface ([`Document::nodes`], [`Document::children`],
1969 /// [`Document::subtree`], [`Document::node_at`], [`Document::query`],
1970 /// [`Document::span`], …) applies to a document being edited.
1971 ///
1972 /// The view borrows the editor mutably, so no edit can land while it is
1973 /// alive and the ids it yields cannot go stale; drop it to edit again. See
1974 /// [`DocumentView`] for the two methods it cannot serve.
1975 pub fn document(&mut self) -> Result<DocumentView<'_>, Error> {
1976 let mut raw = std::ptr::null_mut();
1977 let status = unsafe { ffi::twig_editor_document(self.raw.as_ptr(), &mut raw) };
1978 Error::from_status(status)?;
1979 let raw = NonNull::new(raw).ok_or(Error::Internal)?;
1980 Ok(DocumentView {
1981 doc: Document { raw },
1982 _editor: PhantomData,
1983 })
1984 }
1985
1986 /// Snapshot the current tree as a flat [`FlatNode`] array (the JSON-free
1987 /// read path for a renderer), indexed so `nodes[i].id == NodeId(i)`. Walk it
1988 /// via the `parent`/`first_child`/`next_sibling` links; the root is the node
1989 /// whose `parent` is `None`.
1990 pub fn nodes(&mut self) -> Result<Vec<FlatNode>, Error> {
1991 let mut ptr: *const ffi::TwigFlatNode = std::ptr::null();
1992 let mut len = 0usize;
1993 let status = unsafe { ffi::twig_editor_nodes(self.raw.as_ptr(), &mut ptr, &mut len) };
1994 Error::from_status(status)?;
1995 if len == 0 {
1996 return Ok(Vec::new());
1997 }
1998 if ptr.is_null() {
1999 return Err(Error::Internal);
2000 }
2001 let raw = unsafe { std::slice::from_raw_parts(ptr, len) };
2002 raw.iter().map(flat_node_from_ffi).collect()
2003 }
2004
2005 /// The direct children of `node` as [`QueryMatch`]es (id, span, kind) —
2006 /// `None` enumerates the document root's children (the top-level blocks). The
2007 /// cheap top-level enumeration an incremental renderer walks to decide which
2008 /// blocks changed, without marshalling the whole arena; pair it with
2009 /// [`Editor::subtree`] to then re-marshal only those that did. A childless
2010 /// node yields an empty vec.
2011 pub fn child_spans(&mut self, node: Option<NodeId>) -> Result<Vec<QueryMatch>, Error> {
2012 let id = node.map_or(ffi::TWIG_NO_NODE, |n| n.0);
2013 let mut ptr: *const ffi::TwigQueryMatch = std::ptr::null();
2014 let mut len = 0usize;
2015 let status =
2016 unsafe { ffi::twig_editor_child_spans(self.raw.as_ptr(), id, &mut ptr, &mut len) };
2017 Error::from_status(status)?;
2018 if len == 0 || ptr.is_null() {
2019 return Ok(Vec::new());
2020 }
2021 let raw = unsafe { std::slice::from_raw_parts(ptr, len) };
2022 raw.iter().map(query_match_from_ffi).collect()
2023 }
2024
2025 /// Snapshot the subtree rooted at `node` as a self-contained [`FlatNode`]
2026 /// array with *local* ids: `array[0]` is the root, every link is an index
2027 /// into the returned vec (or `None`), and spans stay absolute. The
2028 /// incremental-render companion to [`Editor::nodes`] — re-marshal one edited
2029 /// block's subtree instead of the whole document. The root's `parent` and
2030 /// `next_sibling` are `None`, so a walk from index 0 stays inside the
2031 /// subtree. [`Error::InvalidArgument`] if `node` is out of range.
2032 pub fn subtree(&mut self, node: NodeId) -> Result<Vec<FlatNode>, Error> {
2033 let mut ptr: *const ffi::TwigFlatNode = std::ptr::null();
2034 let mut len = 0usize;
2035 let status =
2036 unsafe { ffi::twig_editor_subtree(self.raw.as_ptr(), node.0, &mut ptr, &mut len) };
2037 Error::from_status(status)?;
2038 if len == 0 || ptr.is_null() {
2039 return Ok(Vec::new());
2040 }
2041 let raw = unsafe { std::slice::from_raw_parts(ptr, len) };
2042 raw.iter().map(flat_node_from_ffi).collect()
2043 }
2044
2045 /// The deepest node whose span contains byte `offset` (with `offset` equal
2046 /// to the source length treated as inside the root) — mouse hit-testing and
2047 /// cursor context. `Ok(None)` if no node covers the offset;
2048 /// [`Error::InvalidArgument`] if `offset` exceeds the source length.
2049 pub fn node_at(&mut self, offset: usize) -> Result<Option<QueryMatch>, Error> {
2050 let mut m = ffi::TwigQueryMatch {
2051 node_id: 0,
2052 span: ffi::TwigSpan { start: 0, end: 0 },
2053 content_span: ffi::TwigSpan { start: 0, end: 0 },
2054 has_content_span: 0,
2055 kind: std::ptr::null(),
2056 };
2057 let status = unsafe { ffi::twig_editor_node_at(self.raw.as_ptr(), offset, &mut m) };
2058 match status.0 {
2059 ffi::TwigStatus::OK => Ok(Some(query_match_from_ffi(&m)?)),
2060 ffi::TwigStatus::NOT_FOUND => Ok(None),
2061 _ => Err(Error::from_status(status).unwrap_err()),
2062 }
2063 }
2064
2065 /// The chain of nodes containing byte `offset`, root-first down to the
2066 /// deepest (the node [`Editor::node_at`] returns) — the ancestor path for a
2067 /// breadcrumb or context-scoped edit. Empty if no node covers the offset.
2068 pub fn ancestors_at(&mut self, offset: usize) -> Result<Vec<QueryMatch>, Error> {
2069 let mut ptr: *const ffi::TwigQueryMatch = std::ptr::null();
2070 let mut len = 0usize;
2071 let status =
2072 unsafe { ffi::twig_editor_nodes_at(self.raw.as_ptr(), offset, &mut ptr, &mut len) };
2073 match status.0 {
2074 ffi::TwigStatus::OK => {}
2075 ffi::TwigStatus::NOT_FOUND => return Ok(Vec::new()),
2076 _ => return Err(Error::from_status(status).unwrap_err()),
2077 }
2078 if len == 0 || ptr.is_null() {
2079 return Ok(Vec::new());
2080 }
2081 let raw = unsafe { std::slice::from_raw_parts(ptr, len) };
2082 raw.iter().map(query_match_from_ffi).collect()
2083 }
2084
2085 // ── range-oriented rich-text ops (the toolbar) ──────────────────────────
2086
2087 /// Wrap `[start, end)` with `kind`'s delimiters — the unconditional half of
2088 /// the inline toolbar (always adds a mark; `*word*` → `**word**` stacks).
2089 /// [`Error::UnsupportedFormat`] if the document's format can't spell `kind`
2090 /// (e.g. a Markdown [`InlineKind::Mark`]); [`Error::InvalidArgument`] for a
2091 /// bad range; [`Error::EditConflict`] if the result doesn't reparse.
2092 ///
2093 /// A range crossing a **block boundary** gets one pair per block, in a
2094 /// single splice — so one undo step and one [`Change`]:
2095 ///
2096 /// ```text
2097 /// one two\n\nthree four -> **one two**\n\n**three four**
2098 /// ```
2099 ///
2100 /// rather than one pair straddling the blank line, which reparses as four
2101 /// literal asterisks and no mark at all. A block's own marker stays outside
2102 /// the pair (a heading keeps its `# `, a list item its `- `), and a code
2103 /// block inside the range is stepped over — `**` in a program is two
2104 /// asterisks. A code span the range cuts into is taken whole, so the pair
2105 /// closes around its backticks (`` **`word`** `` from a selection of
2106 /// `word`) rather than inside them. A range with no inline content
2107 /// anywhere in it, one wholly inside a fence, is [`Error::NotEditable`].
2108 /// A zero-width range is exempt
2109 /// from all of this: it crosses nothing, and opening an empty pair for the
2110 /// caret to type between is the gesture.
2111 pub fn wrap_range(
2112 &mut self,
2113 start: usize,
2114 end: usize,
2115 kind: InlineKind,
2116 ) -> Result<Change, Error> {
2117 self.change_op(|ed, out| unsafe {
2118 ffi::twig_editor_wrap_range(ed, start, end, kind.to_c(), out)
2119 })
2120 }
2121
2122 /// Toggle `kind` over `[start, end)`: remove the mark if the range already
2123 /// *is* a node of `kind` — covers its whole rendered interior and reaches
2124 /// no further than its own delimiters, or is a mark of another kind that
2125 /// is nothing but it (`***word***` selected whole is the strong inside the
2126 /// emphasis) — else wrap it — a rich editor's Cmd-B. Same error rules as
2127 /// [`Editor::wrap_range`], and the same per-block cutting: remove-or-wrap
2128 /// is decided once per block the range touches, so a second press over a
2129 /// multi-block selection takes off every mark the first one put on instead
2130 /// of nesting a second pair around each.
2131 pub fn toggle_inline(
2132 &mut self,
2133 start: usize,
2134 end: usize,
2135 kind: InlineKind,
2136 ) -> Result<Change, Error> {
2137 self.change_op(|ed, out| unsafe {
2138 ffi::twig_editor_toggle_inline(ed, start, end, kind.to_c(), out)
2139 })
2140 }
2141
2142 /// Set — or clear, with `None` — the colour of the highlight (a `mark`) the
2143 /// caret at `offset` is inside.
2144 ///
2145 /// Markdown only, and only for an editor created with
2146 /// [`MarkdownExtensions::highlight_colors`] (which needs
2147 /// [`MarkdownExtensions::highlight`] with it) — else
2148 /// [`Error::UnsupportedFormat`]. Ask [`Format::supports_with`] with
2149 /// [`Gesture::SetMarkColor`] and the same extensions.
2150 ///
2151 /// Setting a colour on an uncoloured highlight inserts the prefix, setting
2152 /// one on a coloured highlight replaces it, and `None` removes it — with
2153 /// the space after the emoji, which is part of the spelling. An existing
2154 /// prefix keeps its own spacing: `==🔴text==` recolours tight, because that
2155 /// is what its author wrote.
2156 ///
2157 /// [`Error::NotEditable`] when the caret is not inside a highlight.
2158 /// Clearing a colour a highlight does not have is a no-op that succeeds,
2159 /// and the [`Change`] it returns then describes the most recent **prior**
2160 /// edit (or an empty one), so it is not proof the source moved.
2161 ///
2162 /// Authoring a coloured highlight from nothing is two gestures — a colour
2163 /// is a property of a highlight that already exists:
2164 ///
2165 /// ```no_run
2166 /// # use twig::{Editor, Format, MarkColor, MarkdownExtensions};
2167 /// # fn main() -> Result<(), twig::Error> {
2168 /// let exts = MarkdownExtensions {
2169 /// highlight: true,
2170 /// highlight_colors: true,
2171 /// ..Default::default()
2172 /// };
2173 /// let mut ed = Editor::new_ext(b"a word b\n", Format::Markdown, exts)?;
2174 /// ed.toggle_inline(2, 6, twig::InlineKind::Mark)?; // a ==word== b
2175 /// ed.set_mark_color(4, Some(MarkColor::Red))?; // a ==🔴 word== b
2176 /// # Ok(())
2177 /// # }
2178 /// ```
2179 pub fn set_mark_color(
2180 &mut self,
2181 offset: usize,
2182 color: Option<MarkColor>,
2183 ) -> Result<Change, Error> {
2184 let name = color.map(MarkColor::as_str);
2185 let (ptr, len, has) = opt_str(name);
2186 self.change_op(|ed, out| unsafe {
2187 ffi::twig_editor_set_mark_color(ed, offset, ptr, len, has, out)
2188 })
2189 }
2190
2191 /// Convert the innermost heading/paragraph covering byte `offset` to `kind`
2192 /// (the toolbar's H1…H6 / Body switch). Where the format spells a heading
2193 /// with a leading marker (Djot, Markdown, AsciiDoc) that marker is
2194 /// rewritten and the inline content kept byte for byte; where it spells
2195 /// one as a tag pair (HTML) the block is rebuilt as a node of the new kind
2196 /// and printed by the format's own serializer, so `<p>a <em>b</em></p>`
2197 /// becomes `<h2>a <em>b</em></h2>` with its attributes along.
2198 /// [`Error::UnsupportedFormat`] for a format that can do neither (XML);
2199 /// [`Error::InvalidArgument`] for a heading level outside 1–6.
2200 ///
2201 /// On a BLANK LINE this OPENS the block rather than converting one, so
2202 /// "H2, then type" works from an empty line the way it works from a full
2203 /// one — there is no node there to rewrite, since no format spells an empty
2204 /// paragraph. The marker is blank-separated from whatever precedes it (Djot
2205 /// does not let a heading interrupt a paragraph, so a marker flush under one
2206 /// is read as that paragraph's text) and carries the line's quote markers,
2207 /// so a heading opened on a quote's blank line stays inside the quote.
2208 /// [`BlockKind::Paragraph`] there is a no-op: a blank line already holds no
2209 /// marker.
2210 ///
2211 /// [`Error::NotEditable`] when the blank line is INTERIOR to a block rather
2212 /// than between blocks — inside a fenced code block, or a table.
2213 pub fn set_block(&mut self, offset: usize, kind: BlockKind) -> Result<Change, Error> {
2214 let (block_kind, level) = kind.to_c();
2215 self.change_op(|ed, out| unsafe {
2216 ffi::twig_editor_set_block(ed, offset, block_kind, level, out)
2217 })
2218 }
2219
2220 /// Toggle a block container over the blocks `[start, end)` covers — the
2221 /// toolbar's Quote / Bulleted list / Numbered list buttons. Djot and Markdown
2222 /// only, else [`Error::UnsupportedFormat`]; [`Error::NotFound`] if the range
2223 /// covers no block; [`Error::InvalidArgument`] for a bad range.
2224 ///
2225 /// The range widens to whole lines of the blocks it touches (you cannot quote
2226 /// half a paragraph), and the prefix lands at column 0, so a container wraps
2227 /// the outermost structure on those lines.
2228 ///
2229 /// Whether this adds or removes is decided from the **AST** — the ancestors
2230 /// of `start` — not by looking for a `>` in the source. It removes the
2231 /// container only when the range covers every block that container holds, and
2232 /// then only one level (`> > a` → `> a`). A partly covered container **nests**
2233 /// instead, since removing it would drag its uncovered siblings out with it:
2234 /// selecting the first paragraph of `> a\n>\n> b\n` gives `> > a\n>\n> b\n`.
2235 /// Toggling one list kind while inside the other **converts** in place
2236 /// (`- a` → `1. a`) rather than nesting.
2237 ///
2238 /// Each covered block becomes one item, so an ordered list numbers a
2239 /// multi-block range `1.`, `2.`, `3.`… Removing a list inserts a blank line
2240 /// between items that lacked one, keeping them separate blocks (a tight
2241 /// `- a\n- b\n` stripped bare would be a single two-line paragraph).
2242 pub fn toggle_block_container(
2243 &mut self,
2244 start: usize,
2245 end: usize,
2246 kind: BlockContainerKind,
2247 ) -> Result<Change, Error> {
2248 self.change_op(|ed, out| unsafe {
2249 ffi::twig_editor_toggle_block_container(ed, start, end, kind.to_c(), out)
2250 })
2251 }
2252
2253 /// Renumber the ordered list at byte `offset` so its markers run `1, 2, 3, …`,
2254 /// each nesting level restarting at 1 — the numbering a caret editor keeps as
2255 /// items are inserted, deleted, and nested, where a raw splice leaves the
2256 /// source numbers stale (`1. 2. 2. 3.`). Djot and Markdown; the display of an
2257 /// ordered list is renumbered by any CommonMark renderer regardless, so this
2258 /// is source hygiene, not a render fix.
2259 ///
2260 /// [`Error::NotFound`] when `offset` is not inside an ordered list. When the
2261 /// numbering is already sequential this is a no-op that still returns `Ok` —
2262 /// the source is left byte-for-byte unchanged. The `Change` is not returned
2263 /// because a no-op has none; re-read [`Editor::source_str`] for the result.
2264 ///
2265 /// Only lines the PARSER reads as items are touched, so this never rewrites a
2266 /// digit the author wrote as prose. That is not a corner case across formats:
2267 /// Djot doesn't let a list marker interrupt a paragraph, so in
2268 /// `1. a\n 2. b` the second line is text inside item `a`, while Markdown
2269 /// reads it as a nested item — the same bytes, renumbered in one format and
2270 /// left alone in the other.
2271 pub fn renumber_ordered_lists(&mut self, offset: usize) -> Result<(), Error> {
2272 self.change_op(|ed, out| unsafe {
2273 ffi::twig_editor_renumber_ordered_lists(ed, offset, out)
2274 })?;
2275 Ok(())
2276 }
2277
2278 // ── Tables ───────────────────────────────────────────────────────────────
2279 // Structural editing of the pipe table at a byte `offset`: the caret's cell
2280 // is the anchor. The whole table is re-spelled and spliced in one edit, so a
2281 // caller re-reads [`Editor::source_str`] and re-places its caret rather than
2282 // leaning on the returned span. [`Error::NotFound`] when `offset` is not in a
2283 // table; [`Error::NotEditable`] for a refused (degenerate) edit.
2284
2285 /// Insert an empty row below (`below`) or above the caret's row.
2286 pub fn table_insert_row(&mut self, offset: usize, below: bool) -> Result<(), Error> {
2287 self.table_edit(offset, ffi::TWIG_TABLE_INSERT_ROW, below as c_int)
2288 }
2289
2290 /// Delete the caret's row. [`Error::NotEditable`] for the header row or the
2291 /// last remaining body row.
2292 pub fn table_delete_row(&mut self, offset: usize) -> Result<(), Error> {
2293 self.table_edit(offset, ffi::TWIG_TABLE_DELETE_ROW, 0)
2294 }
2295
2296 /// Insert an empty column right (`right`) or left of the caret's column.
2297 pub fn table_insert_column(&mut self, offset: usize, right: bool) -> Result<(), Error> {
2298 self.table_edit(offset, ffi::TWIG_TABLE_INSERT_COLUMN, right as c_int)
2299 }
2300
2301 /// Delete the caret's column. [`Error::NotEditable`] when it is the only one.
2302 pub fn table_delete_column(&mut self, offset: usize) -> Result<(), Error> {
2303 self.table_edit(offset, ffi::TWIG_TABLE_DELETE_COLUMN, 0)
2304 }
2305
2306 /// Set the caret's column to `alignment`.
2307 pub fn table_set_alignment(
2308 &mut self,
2309 offset: usize,
2310 alignment: Alignment,
2311 ) -> Result<(), Error> {
2312 self.table_edit(offset, ffi::TWIG_TABLE_SET_ALIGNMENT, alignment.to_c())
2313 }
2314
2315 /// Move the caret's row one place down (`down`) or up, within the body rows.
2316 pub fn table_move_row(&mut self, offset: usize, down: bool) -> Result<(), Error> {
2317 self.table_edit(offset, ffi::TWIG_TABLE_MOVE_ROW, down as c_int)
2318 }
2319
2320 /// Move the caret's column one place right (`right`) or left.
2321 pub fn table_move_column(&mut self, offset: usize, right: bool) -> Result<(), Error> {
2322 self.table_edit(offset, ffi::TWIG_TABLE_MOVE_COLUMN, right as c_int)
2323 }
2324
2325 fn table_edit(&mut self, offset: usize, op: c_int, arg: c_int) -> Result<(), Error> {
2326 self.change_op(|ed, out| unsafe { ffi::twig_editor_table_edit(ed, offset, op, arg, out) })?;
2327 Ok(())
2328 }
2329
2330 /// Insert a fresh table — one header row, `rows` body rows, `cols` columns,
2331 /// every cell empty — as its own block after the block `offset` sits in.
2332 ///
2333 /// The placement is [`Editor::insert_thematic_break`]'s, decision for
2334 /// decision: after the caret's block rather than at the caret, blank-line
2335 /// separated on both sides, carrying a block quote's prefix on every line,
2336 /// and at column zero after a list item. The blank above is load-bearing
2337 /// here too — GFM can read a table's header row out of the paragraph it
2338 /// follows. The bytes are the format's own table spelling, through the
2339 /// same emitter the `table_*` edits re-spell with, so the table this
2340 /// writes is one they can edit.
2341 ///
2342 /// There is no [`Error::NotFound`]: an empty document is a fine place for
2343 /// a table. [`Error::InvalidArgument`] for `rows == 0` or `cols == 0` — a
2344 /// header with nothing under it is the shape [`Editor::table_delete_row`]
2345 /// refuses to leave — or an `offset` past the source;
2346 /// [`Error::UnsupportedFormat`] where the format has no table spelling,
2347 /// before anything is read. [`Gesture::InsertTable`] answers ahead of time.
2348 pub fn insert_table(
2349 &mut self,
2350 offset: usize,
2351 rows: usize,
2352 cols: usize,
2353 ) -> Result<Change, Error> {
2354 self.change_op(|ed, out| unsafe {
2355 ffi::twig_editor_insert_table(ed, offset, rows, cols, out)
2356 })
2357 }
2358
2359 /// Insert a **leaf directive** — Markdown's `::name[label]{attrs}` — as its
2360 /// own block after the block `offset` sits in.
2361 ///
2362 /// The placement is [`Editor::insert_thematic_break`]'s, decision for
2363 /// decision: after the caret's block rather than at the caret, blank-line
2364 /// separated on both sides, a block quote's prefix on **every** line (djot
2365 /// spells this over two lines, and a marker on the opener alone would leave
2366 /// the closing fence outside the quote), and column zero after a list item.
2367 ///
2368 /// What a name *means* is the application's. Twig writes a named container
2369 /// and reads one back; `"page-break"` and `"embed"` are your words and
2370 /// nothing here interprets them. The bytes are the format's own spelling of
2371 /// that node: `::name{…}` in Markdown, an empty `::: name` fence in djot
2372 /// (whose div is anonymous, so the name comes back as a *class*),
2373 /// `<name>…</name>` in HTML.
2374 ///
2375 /// `label` is the bracketed text, `None` for none — a different document
2376 /// from `Some("")` (`::name` against `::name[]`). `attrs` is the
2377 /// `(key, Some(value))` / `(key, None)` pair list [`Builder::set_attrs`]
2378 /// takes.
2379 ///
2380 /// A `name` is an ASCII letter followed by letters, digits, `-` and `_` —
2381 /// the grammar every format reads one back by, checked before anything is
2382 /// written because a name goes where a delimiter would otherwise be.
2383 /// `"page-break"` and `"x-embed"` are names; `"a b"`, `"a:b"`, `"]{"` and
2384 /// `"1x"` are not, and each is a *different* wrong document per format
2385 /// (`::a b` is a paragraph holding an inline directive named `a`,
2386 /// `::: a:b` is a paragraph of colons). A `label` may not carry a line end
2387 /// or a square bracket, either of which closes the `[…]` early.
2388 ///
2389 /// [`Error::InvalidArgument`] for a name outside that grammar (an empty one
2390 /// included), such a label, or an `offset` past the source. There is no
2391 /// [`Error::NotFound`]: an empty document is a fine place for one.
2392 /// [`Error::UnsupportedFormat`] where the format would not read the printed
2393 /// bytes back as a container carrying the name, before anything is read —
2394 /// and for Markdown that is the parse config's answer, so
2395 /// [`Format::supports`] reports `false` while [`Format::supports_with`]
2396 /// reports `true`:
2397 ///
2398 /// ```no_run
2399 /// # use twig::{Format, Gesture, MarkdownExtensions};
2400 /// let exts = MarkdownExtensions { directives: true, ..Default::default() };
2401 /// assert!(!Format::Markdown.supports(Gesture::InsertDirective));
2402 /// assert!(Format::Markdown.supports_with(exts, Gesture::InsertDirective));
2403 /// ```
2404 ///
2405 /// The editor must have been created with the same extensions
2406 /// ([`Editor::new_ext`]) or the call itself refuses.
2407 pub fn insert_directive(
2408 &mut self,
2409 offset: usize,
2410 name: &str,
2411 label: Option<&str>,
2412 attrs: &[(&str, Option<&str>)],
2413 ) -> Result<Change, Error> {
2414 let kvs: Vec<ffi::TwigKeyVal> = attrs
2415 .iter()
2416 .map(|(k, v)| ffi::TwigKeyVal {
2417 key: k.as_ptr(),
2418 key_len: k.len(),
2419 value: v.map_or(std::ptr::null(), |s| s.as_ptr()),
2420 value_len: v.map_or(0, |s| s.len()),
2421 })
2422 .collect();
2423 // A null label pointer with a zero length is "no label" across this
2424 // boundary; a non-null one with a zero length is an empty label. Going
2425 // through the `Option` rather than through a `&str` is what keeps the
2426 // two distinct.
2427 let label_ptr = label.map_or(std::ptr::null(), |s| s.as_ptr());
2428 let label_len = label.map_or(0, |s| s.len());
2429 self.change_op(|ed, out| unsafe {
2430 ffi::twig_editor_insert_directive(
2431 ed,
2432 offset,
2433 name.as_ptr(),
2434 name.len(),
2435 label_ptr,
2436 label_len,
2437 kvs.as_ptr(),
2438 kvs.len(),
2439 out,
2440 )
2441 })
2442 }
2443
2444 /// Replace the attribute set of the block `offset` sits in — a paragraph
2445 /// or heading, the block [`Editor::set_block`] rewrites — with `attrs`,
2446 /// the `(key, Some(value))` list [`Builder::set_attrs`] takes. **Replace,
2447 /// not merge**: read the node's attributes, edit the list, pass it back
2448 /// whole; an empty list clears them.
2449 ///
2450 /// What a key *means* is yours, as a directive's name is: a centred
2451 /// paragraph is `set_block_attrs(off, &[("class", Some("center"))])` and
2452 /// twig spells the pair without interpreting either half. The spelling is
2453 /// the format's — djot's `{…}` line before the block, rewritten in place
2454 /// with the block's bytes untouched; HTML's tag and AsciiDoc's `[…]` line,
2455 /// the block re-printed; and in Markdown a `<div …>` around the block,
2456 /// blank-separated, which every Markdown renderer passes through and
2457 /// twig's own parser pairs back into a container only under
2458 /// [`MarkdownExtensions::html_elements`]. A block already the sole child
2459 /// of such a div has the div's attributes replaced instead, and an empty
2460 /// list unwraps it.
2461 ///
2462 /// An attribute must be one every format reads back: a key that is an
2463 /// ASCII letter or `_` followed by letters, digits, `-`, `_` and `:`, a
2464 /// `Some` value (djot has no bare attribute), and no line end or double
2465 /// quote in it. [`Error::InvalidArgument`] otherwise, or for an `offset`
2466 /// past the source. [`Error::NotFound`] when no paragraph or heading holds
2467 /// `offset`. [`Error::NotEditable`] where the spelling cannot be placed: a
2468 /// djot block starting on a list item's marker line, or whose attributes
2469 /// came from more than one `{…}` block; a Markdown block inside a list
2470 /// item. [`Error::UnsupportedFormat`] where the format would not read the
2471 /// printed attributes back, before anything is read — for Markdown the
2472 /// parse config's answer, so [`Format::supports`] reports `false` while
2473 /// [`Format::supports_with`] reports `true`:
2474 ///
2475 /// ```no_run
2476 /// # use twig::{Format, Gesture, MarkdownExtensions};
2477 /// let exts = MarkdownExtensions { html_elements: true, ..Default::default() };
2478 /// assert!(!Format::Markdown.supports(Gesture::SetBlockAttrs));
2479 /// assert!(Format::Markdown.supports_with(exts, Gesture::SetBlockAttrs));
2480 /// assert!(Format::Djot.supports(Gesture::SetBlockAttrs));
2481 /// ```
2482 pub fn set_block_attrs(
2483 &mut self,
2484 offset: usize,
2485 attrs: &[(&str, Option<&str>)],
2486 ) -> Result<Change, Error> {
2487 let kvs: Vec<ffi::TwigKeyVal> = attrs
2488 .iter()
2489 .map(|(k, v)| ffi::TwigKeyVal {
2490 key: k.as_ptr(),
2491 key_len: k.len(),
2492 value: v.map_or(std::ptr::null(), |s| s.as_ptr()),
2493 value_len: v.map_or(0, |s| s.len()),
2494 })
2495 .collect();
2496 self.change_op(|ed, out| unsafe {
2497 ffi::twig_editor_set_block_attrs(ed, offset, kvs.as_ptr(), kvs.len(), out)
2498 })
2499 }
2500
2501 /// Replace the attribute set of the element `node` — an id from
2502 /// [`Editor::nodes`], valid against the **current** tree, so read the tree
2503 /// again after any successful edit — with `attrs`, the same list
2504 /// [`Editor::set_block_attrs`] takes; an empty list clears them. Replace,
2505 /// not merge.
2506 ///
2507 /// The node-addressed sibling of [`Editor::set_block_attrs`], for the
2508 /// caller that holds a tree rather than a caret: a canvas editor over an
2509 /// SVG names the `<rect>` it is dragging, and no byte offset stands for
2510 /// it. The run is written on the element's own start tag, at the span
2511 /// [`Document::attrs_span`] reports, as ` key="value"` pairs with `&`,
2512 /// `<`, `>` and `"` as entities. Nothing else in the element moves, its
2513 /// children included — a `<g>` holding a thousand paths is not re-printed
2514 /// to change its `transform`. An element with no attributes yet has the
2515 /// run inserted right after its name.
2516 ///
2517 /// [`Error::UnsupportedFormat`] where the format keeps a node's attributes
2518 /// anywhere but on the node's own tag — every format but [`Format::Xml`]
2519 /// today; [`Format::supports`] with [`Gesture::SetNodeAttrs`] says which.
2520 /// [`Error::InvalidArgument`] for an id past the tree or an attribute no
2521 /// format reads back (the rule [`Editor::set_block_attrs`] states);
2522 /// [`Error::NotEditable`] for a node that is not an element — a text run,
2523 /// a comment.
2524 ///
2525 /// ```no_run
2526 /// # use twig::{Editor, Format, Kind};
2527 /// let mut ed = Editor::new_str("<svg><rect x=\"1\"/></svg>", Format::Xml)?;
2528 /// let rect = ed.nodes()?.into_iter().find(|n| n.name.as_deref() == Some("rect")).unwrap();
2529 /// ed.set_node_attrs(rect.id, &[("x", Some("10")), ("fill", Some("red"))])?;
2530 /// assert_eq!(ed.source_str()?, "<svg><rect x=\"10\" fill=\"red\"/></svg>");
2531 /// # Ok::<(), twig::Error>(())
2532 /// ```
2533 pub fn set_node_attrs(
2534 &mut self,
2535 node: NodeId,
2536 attrs: &[(&str, Option<&str>)],
2537 ) -> Result<Change, Error> {
2538 let kvs: Vec<ffi::TwigKeyVal> = attrs
2539 .iter()
2540 .map(|(k, v)| ffi::TwigKeyVal {
2541 key: k.as_ptr(),
2542 key_len: k.len(),
2543 value: v.map_or(std::ptr::null(), |s| s.as_ptr()),
2544 value_len: v.map_or(0, |s| s.len()),
2545 })
2546 .collect();
2547 self.change_op(|ed, out| unsafe {
2548 ffi::twig_editor_set_node_attrs(ed, node.0, kvs.as_ptr(), kvs.len(), out)
2549 })
2550 }
2551
2552 /// Wrap `[start, end)` in an anonymous inline container carrying `attrs`
2553 /// — djot's `[text]{…}`, HTML's and Markdown's `<span …>` — or, when the
2554 /// range already lies inside such a span, **replace** that span's
2555 /// attributes rather than nest a second; an empty list there unwraps it,
2556 /// keeping the content bytes. That is [`Editor::insert_link`]'s rule for
2557 /// a link covering the range, for the same reason. A span named `span`
2558 /// and an anonymous one are the same node here; a `:span[…]` the Markdown
2559 /// parser read as a directive is neither.
2560 ///
2561 /// The inline half of [`Editor::set_block_attrs`], with its vocabulary
2562 /// rule and its attribute grammar ([`Error::InvalidArgument`] for a key or
2563 /// value no format reads back, or a bad range). The covered inline nodes
2564 /// are printed under the container by the format's own serializer, so a
2565 /// mark inside the range rides along. [`Error::NotEditable`] for an empty
2566 /// range with no span to re-style, or a range cutting a node the gesture
2567 /// cannot slice. [`Error::UnsupportedFormat`] where the format would not
2568 /// read the printed span back: AsciiDoc, whose `[#id.role]#text#` keeps
2569 /// an id and a role and drops any other key, and Markdown without
2570 /// [`MarkdownExtensions::html_elements`] — ask [`Format::supports_with`].
2571 pub fn wrap_range_attrs(
2572 &mut self,
2573 start: usize,
2574 end: usize,
2575 attrs: &[(&str, Option<&str>)],
2576 ) -> Result<Change, Error> {
2577 let kvs: Vec<ffi::TwigKeyVal> = attrs
2578 .iter()
2579 .map(|(k, v)| ffi::TwigKeyVal {
2580 key: k.as_ptr(),
2581 key_len: k.len(),
2582 value: v.map_or(std::ptr::null(), |s| s.as_ptr()),
2583 value_len: v.map_or(0, |s| s.len()),
2584 })
2585 .collect();
2586 self.change_op(|ed, out| unsafe {
2587 ffi::twig_editor_wrap_range_attrs(ed, start, end, kvs.as_ptr(), kvs.len(), out)
2588 })
2589 }
2590
2591 /// Link `[start, end)` to `destination` — `[text](destination)`. Djot and
2592 /// Markdown only, else [`Error::UnsupportedFormat`];
2593 /// [`Error::InvalidArgument`] for a bad range or a destination containing a
2594 /// newline (neither format can carry one, and quietly rewriting the URL would
2595 /// be worse than refusing).
2596 ///
2597 /// An existing link covering the range has its destination **replaced** and
2598 /// its text kept, so re-linking fixes a URL instead of nesting
2599 /// `[[t](a)](b)`; to unlink, use [`Editor::unwrap_node`].
2600 ///
2601 /// A **range inside an existing autolink** (`<https://x.dev>`) re-points it
2602 /// the same way, but there is no text to keep — an autolink's text *is* its
2603 /// destination — so the node is replaced whole, respelled canonically for the
2604 /// new destination. This covers a caret and any selection the autolink
2605 /// contains, including one covering it exactly: an autolink's URL is not
2606 /// editable text, so no part of it can host a `[`, and "link half this URL"
2607 /// has no spelling. A caret inside both an autolink and a link
2608 /// (`[<https://x.dev>](d)`) re-points the link, whose text is separable from
2609 /// its destination and so survives.
2610 ///
2611 /// A selection starting or ending strictly **inside** an autolink without
2612 /// being contained by it — running from ordinary text into the middle of a
2613 /// URL — is refused with [`Error::NotEditable`]: half of it is real text,
2614 /// so there is nothing to re-point, and any splice would rewrite the URL.
2615 /// A selection that *contains* an autolink whole is unaffected — it splices
2616 /// at the edges and wraps as usual.
2617 ///
2618 /// A link with **no text** — an empty range, or re-pointing an existing
2619 /// `[](old)` — is spelled canonically for the destination given, never as
2620 /// `[](destination)`: a childless link has nothing to render, so consumers
2621 /// fall back to showing the destination and a caret has nowhere to sit. A
2622 /// destination the format can autolink (an absolute URL or an email, by that
2623 /// format's own rules) yields `<destination>`; anything else yields
2624 /// `[destination](destination)`, the destination doubling as the text so it
2625 /// stays visible and editable. Which destinations autolink is not the
2626 /// caller's to guess — `<foo>` is raw HTML in Markdown, a relative path goes
2627 /// literal in both, and the formats disagree (`<mailto:a@b.dev>` is a url in
2628 /// Markdown, an email in Djot), so each is asked its own parser.
2629 ///
2630 /// The destination is escaped for the format, so a `)` or a space in it
2631 /// cannot break the markup — and the two formats genuinely differ: Markdown
2632 /// ends a destination at the first space (`[t](a b)` is not a link at all) so
2633 /// whitespace moves it into the `<…>` form, while Djot takes spaces literally
2634 /// and would read `<a b>` as the URL itself.
2635 pub fn insert_link(
2636 &mut self,
2637 start: usize,
2638 end: usize,
2639 destination: &str,
2640 ) -> Result<Change, Error> {
2641 self.change_op(|ed, out| unsafe {
2642 ffi::twig_editor_insert_link(
2643 ed,
2644 start,
2645 end,
2646 destination.as_ptr(),
2647 destination.len(),
2648 out,
2649 )
2650 })
2651 }
2652
2653 /// Spell `[start, end)` as an image pointing at `destination` —
2654 /// ``, the selected source becoming the alt text.
2655 ///
2656 /// The destination is escaped exactly as [`insert_link`](Self::insert_link)
2657 /// escapes one, because it is the same grammar production: Markdown moves a
2658 /// destination holding whitespace into the `<…>` form, Djot leaves it bare
2659 /// because `<…>` there would read as the URL itself. That is the reason this
2660 /// exists rather than being a `format!` at the call site — ``
2661 /// is not an image in Markdown at all, and no caller can fix that without
2662 /// reproducing twig's per-format escape table.
2663 ///
2664 /// Two ways it is simpler than a link. An empty range stays empty:
2665 /// `` is a perfectly good image, where the childless
2666 /// `[](destination)` that `insert_link` works to avoid has nothing to render
2667 /// or put a caret in. And there is no autolink or re-point reasoning — an
2668 /// image has no bare-URL spelling, and re-pointing an existing one is a read
2669 /// of its destination plus an insert, above this op.
2670 ///
2671 /// Returns [`Error::InvalidArgument`] for a destination holding a newline and
2672 /// [`Error::UnsupportedFormat`] for a parse-only format (XML, HTML).
2673 pub fn insert_image(
2674 &mut self,
2675 start: usize,
2676 end: usize,
2677 destination: &str,
2678 ) -> Result<Change, Error> {
2679 self.change_op(|ed, out| unsafe {
2680 ffi::twig_editor_insert_image(
2681 ed,
2682 start,
2683 end,
2684 destination.as_ptr(),
2685 destination.len(),
2686 out,
2687 )
2688 })
2689 }
2690
2691 /// Insert `text` at `offset` as a literal run: every byte the format reads as
2692 /// markup is escaped the format's way so the run reparses as exactly `text`
2693 /// — a typed `*`, `#` or `` ` `` stays that character rather than opening
2694 /// emphasis, a heading or a code span. This is the inverse of serialization
2695 /// (which writes an already-parsed run verbatim): it is what a WYSIWYG
2696 /// surface calls so that keyboard input can never mint markup, leaving
2697 /// formatting to explicit commands.
2698 ///
2699 /// The escaping is positional and per-format, and neither is the caller's to
2700 /// reproduce. In the backslash formats (Djot, Markdown, AsciiDoc) inline
2701 /// specials (`*`, `` ` ``, `[`, `<`…, and `$` under Markdown's `math`
2702 /// extension) are escaped anywhere on the line, while block markers (`#`, `>`, `-`…) are escaped only where `offset` sits
2703 /// in its line's leading whitespace — so an inserted "5 - 3" keeps its `-`
2704 /// but "- item" at column zero does not become a bullet — and an embedded
2705 /// newline in `text` re-enters that line-start zone. Inside a code span,
2706 /// code block or raw node the run is written as it is, since a backslash
2707 /// there would show. HTML escapes with entities (`<`, `&`) in every
2708 /// position.
2709 ///
2710 /// Two constructs a byte-alphabet cannot reach are left as typed: a GFM
2711 /// bare-URL autolink (`https://x.com`, with no delimiter to escape) and an
2712 /// ordered-list marker (`1.`, special only after a digit run). Returns
2713 /// [`Error::UnsupportedFormat`] for a parse-only format (XML) and
2714 /// [`Error::InvalidArgument`] when `offset` is past the source.
2715 pub fn insert_literal(&mut self, offset: usize, text: &str) -> Result<Change, Error> {
2716 self.change_op(|ed, out| unsafe {
2717 ffi::twig_editor_insert_literal(ed, offset, text.as_ptr(), text.len(), out)
2718 })
2719 }
2720
2721 /// Insert a hard line break *inside a table cell* at `offset`, spelled the
2722 /// format's way (`<br>` for Markdown). A table row is one source line, so the
2723 /// ordinary newline-based hard break can't appear there; the spliced `<br>`
2724 /// reparses as a semantic `hard_break` node — not opaque raw HTML — so the
2725 /// break reads back as structure. Like the other gestures it leans on the
2726 /// splice+reparse+rollback backstop: a break that would no longer parse as the
2727 /// same table yields [`Error::EditConflict`] and changes nothing.
2728 ///
2729 /// Returns [`Error::UnsupportedFormat`] for a format with no in-cell break
2730 /// spelling — djot (no idiomatic in-cell break), HTML and XML (parse-only);
2731 /// [`Error::NotFound`] when `offset` is not inside a table cell (only the
2732 /// in-cell gesture is spelled today); and [`Error::InvalidArgument`] when
2733 /// `offset` is past the source.
2734 pub fn insert_line_break(&mut self, offset: usize) -> Result<Change, Error> {
2735 self.change_op(|ed, out| unsafe { ffi::twig_editor_insert_line_break(ed, offset, out) })
2736 }
2737
2738 /// Insert a thematic break (a horizontal rule) as its own block, on the line
2739 /// after the block `offset` sits in. A rule is a block, so there is no
2740 /// spelling for one mid-paragraph.
2741 ///
2742 /// The rule is blank-line separated from its neighbours, and that is
2743 /// load-bearing rather than cosmetic: Markdown reads `---` on the line
2744 /// directly under a paragraph as a setext `<h2>` underline, so a rule written
2745 /// flush against its predecessor silently becomes a heading and swallows it.
2746 /// The blank below is added only when the next line isn't already blank. The
2747 /// spelling is the format's (`---` for Markdown, `* * *` for djot) and not
2748 /// the caller's to reproduce.
2749 ///
2750 /// Inside a block quote the rule inherits the quote's prefix and stays in the
2751 /// quote. Inside a list it lands at column zero after the caret's item, which
2752 /// splits the list in two with the rule between — a real document, nothing
2753 /// swallowed. There is no [`Error::NotFound`]: an empty document is a fine
2754 /// place for a rule. [`Error::UnsupportedFormat`] for a parse-only format
2755 /// (XML, HTML); [`Error::InvalidArgument`] when `offset` is past the source.
2756 pub fn insert_thematic_break(&mut self, offset: usize) -> Result<Change, Error> {
2757 self.change_op(|ed, out| unsafe { ffi::twig_editor_insert_thematic_break(ed, offset, out) })
2758 }
2759
2760 /// Split the block at `offset` in two at the caret, both halves the same
2761 /// kind — Enter in the middle of a paragraph, and the gesture
2762 /// [`Editor::insert_thematic_break`] deliberately is not. A host wanting
2763 /// "rule at the caret" calls this and then that.
2764 ///
2765 /// Nearly a pure insertion at `offset`: what is minted is the separator
2766 /// between the halves, and the only bytes removed are the second half's
2767 /// leading spaces and tabs, which are structure rather than content at the
2768 /// start of a block — a split at `- b| c` that kept its space would write
2769 /// `- c`, setting that item's content indent to three. A code block sheds
2770 /// nothing, because there leading whitespace *is* the content.
2771 ///
2772 /// * A **paragraph** gets a blank line. Inside a quote the blank carries the
2773 /// quote's marker and the second half its full prefix, so the split
2774 /// happens inside the quote rather than ending it.
2775 /// * A paragraph in a **list item** gets the item's marker instead of a
2776 /// blank, so the second half is a sibling item: `- this is |a list item`
2777 /// becomes `- this is ` and `- a list item`. The marker is repeated
2778 /// verbatim, ordered numbers included, so a split `1.` item yields two
2779 /// `1.` items — both formats renumber on render, and
2780 /// [`Editor::renumber_ordered_lists`] is the gesture for fixing the
2781 /// source. A **task** item's new half is an unchecked box whatever the
2782 /// original's state. A **nested** item's leading indent rides along with
2783 /// its marker, so the new sibling stays in its own list rather than
2784 /// dropping to column zero and joining the enclosing one.
2785 /// * A **heading** repeats its own marker at its own level;
2786 /// [`Editor::set_block`] is how a caller demotes the second half instead.
2787 /// * A **code block** becomes two code blocks, the opening fence line
2788 /// reproduced verbatim so its width and info string both survive. A
2789 /// consumer that doesn't want the gesture offered there can ask the tree
2790 /// what block the caret is in before calling.
2791 ///
2792 /// At a block boundary this still splits, which is what makes it Enter: at
2793 /// the end of a list item it opens an empty sibling item, which is the block
2794 /// the caller wants to type into. A paragraph is the one place that empty
2795 /// block cannot be spelled — no format has an empty paragraph — so the
2796 /// source gains a blank line and reparses as one paragraph; the node appears
2797 /// when there is text to hold.
2798 ///
2799 /// [`Error::NotEditable`] where a caret-split has no honest meaning: a
2800 /// **table** (a newline mid-cell destroys rather than divides; splitting one
2801 /// table into two is a table gesture, not this one), a **setext heading**
2802 /// (whose `---` underline would end up under the second half alone —
2803 /// [`Editor::set_block`] normalises one to ATX, which makes this work), and
2804 /// an **indented code block** (where a blank line is interior, so the split
2805 /// would parse back as one block). [`Error::NotFound`] when nothing covers
2806 /// `offset`; [`Error::InvalidArgument`] when `offset` is past the source.
2807 pub fn split_block(&mut self, offset: usize) -> Result<Change, Error> {
2808 self.change_op(|ed, out| unsafe { ffi::twig_editor_split_block(ed, offset, out) })
2809 }
2810
2811 /// Join the block at `offset` into the block **before** it — Backspace at
2812 /// the start of a block, and forward Delete at the end of the one above it.
2813 /// The inverse of [`Editor::split_block`], and the reason it is a gesture
2814 /// rather than a host's own delete: what joins two blocks is a fact about
2815 /// the **format**. Deleting the newline between them is right only for two
2816 /// Markdown paragraphs at the top level — in HTML that byte is the `>` of
2817 /// `</p>`, under a heading it leaves two blocks, and after a Markdown
2818 /// `<div>` it deletes the blank line the div needed and breaks the div.
2819 ///
2820 /// **B** is the innermost paragraph/heading covering `offset`. **A** is the
2821 /// leaf block immediately before it in **document order**, not the sibling
2822 /// before it: the block visually above `below` in
2823 /// `above` / `<div>` / `hello` / `</div>` / `below` is `hello`, three
2824 /// levels down, and joining into `above` would be the wrong paragraph.
2825 ///
2826 /// * What is **written** is a line break plus the container prefix A's own
2827 /// line sits behind — a quote's `> ` repeated, a list item's marker's
2828 /// *width* in spaces, nothing at the top level — which is what keeps the
2829 /// joined line inside its containers in a format with no lazy
2830 /// continuation. A heading A with a **leading marker** (`# Title`,
2831 /// AsciiDoc's `== Title`) is one line by its own spelling, so what joins
2832 /// there is a single **space**: `# Title` + `below` is `# Title below`.
2833 /// * What **travels** is A's own closing markup (an ATX closing `#` run, a
2834 /// setext underline, `</p>`) and the closers of every container A is in
2835 /// that B is not (a Markdown `</div>`, a djot `:::` fence), carried past
2836 /// the text that was pulled in — so the joined block keeps A's
2837 /// presentation and stays where it was.
2838 /// * What is **dropped** is everything between them: the blank line, B's
2839 /// markers, B's attribute line (djot's `{…}`, AsciiDoc's `[…]`), B's
2840 /// opening tags. B's attributes go on purpose — the joined text is A's
2841 /// block, so it takes A's presentation.
2842 /// * A **prefix** container (a quote, a list item, a list, a section) has
2843 /// no closing bytes, so whatever follows B inside one stays where it is:
2844 /// joining the first item's text out of a list leaves the rest a list.
2845 ///
2846 /// [`Error::NotFound`] when no block covers `offset`, and when B is the
2847 /// document's first block — the ordinary Backspace-at-the-top answer.
2848 /// [`Error::NotEditable`] when A is not a paragraph or a heading (a code
2849 /// block, a table, a rule: there is no text to join into), when either
2850 /// block is in a **table cell**, when B is a **setext heading** (whose
2851 /// underline is how it is spelled at all — [`Editor::set_block`] normalises
2852 /// one to ATX, which makes this work), when B would have to leave a
2853 /// **delimited** container that still has content after it, which is the
2854 /// one shape this refuses rather than guesses at, and when the **gap**
2855 /// between A and B holds anything but separation — an empty container, or
2856 /// a definition the splice would destroy and no tree walk could see
2857 /// (Markdown keeps a link reference and a footnote definition as a lookup
2858 /// table, not as a node).
2859 /// [`Error::InvalidArgument`] when `offset` is past the source.
2860 ///
2861 /// [`Error::UnsupportedFormat`] where a block cannot span lines at all —
2862 /// and note this is a **different, wider** gate than `split_block`'s. Ask
2863 /// [`Format::supports`] with [`Gesture::JoinBlocks`].
2864 pub fn join_blocks(&mut self, offset: usize) -> Result<Change, Error> {
2865 self.change_op(|ed, out| unsafe { ffi::twig_editor_join_blocks(ed, offset, out) })
2866 }
2867
2868 /// Move the block at `from` to the boundary `to` names, spelling the line
2869 /// prefixes of the container it lands in — a drag-and-drop, or Alt+↑/↓,
2870 /// in a host that has a caret and no tree.
2871 ///
2872 /// [`Editor::move_before`] and [`Editor::move_after`] move a node's
2873 /// **bytes** and say so: a quote's `> ` does not travel, a list item's
2874 /// continuation indent is not written. This is the gesture for a drop
2875 /// that crosses either — a paragraph dragged out of a quote arrives
2876 /// without its `> `, one dragged under a list item's text arrives behind
2877 /// the item's indent, and the blank lines a person would have typed are
2878 /// written and removed. Within one container the result is what
2879 /// `move_before` writes.
2880 ///
2881 /// * `from` names the deepest block owning the line it is on — the
2882 /// paragraph or heading [`Editor::set_block`] acts on, or the code
2883 /// block, table or rule where there is no text block — **widened to
2884 /// the list item** when it is the item's first block: a bullet's text
2885 /// is the bullet, and dragging it takes the item and everything under
2886 /// it. A block later in an item's tail moves alone.
2887 /// * `to` is a position **between** blocks: at or before a block's first
2888 /// content byte (the block lands before it), at or after its last
2889 /// (after it), on a blank line, or the source's length (the document's
2890 /// end, terminated or not). The block takes the prefixes of the
2891 /// container that boundary is inside — the innermost one, so `to` at
2892 /// the first content byte of a quote's first paragraph is inside the
2893 /// quote, while `to` at or before the quote's marker is before the
2894 /// quote, at its parent's level (`> > a` at 0 is before both quotes,
2895 /// at 2 before the inner one, at 4 inside both). A boundary the block
2896 /// already sits on is read at the container's level when the block is
2897 /// the container's first or last: a quote's last paragraph dropped at
2898 /// its own end, or on the blank line after the quote, leaves the
2899 /// quote; its first dropped at its own start leaves it upward. Two
2900 /// adjustments where a list item is involved, because a list holds
2901 /// items and nothing else: a
2902 /// boundary before an item's first block is before the **item**, at
2903 /// the list's level, while one after an item's last block is inside
2904 /// the item — how a block reaches an item's tail (`to` at the end of
2905 /// the item's text); and a moved item is always a sibling — dropped
2906 /// inside another item it lands after it, and elsewhere it stays a
2907 /// bullet (a one-item list between two paragraphs, a quoted bullet in
2908 /// a quote). An ordered list is not renumbered;
2909 /// [`Editor::renumber_ordered_lists`] is the call for that.
2910 /// * One splice, so one undo step and one [`Change`]. The block's lines
2911 /// go with the separator lines that kept them apart from their old
2912 /// neighbours, and a quote or list they were the only content of goes
2913 /// too; at the destination they are blank-separated from what they
2914 /// land beside (`>` inside a quote), except between items of a tight
2915 /// list. A delimited container (a `<div>`, a djot `:::` fence) that is
2916 /// emptied stands, since it may carry attributes.
2917 ///
2918 /// [`Error::NotFound`] when `from` is on a blank line.
2919 /// [`Error::NotEditable`] when `to` is interior to a block — inside a
2920 /// fence, a table, a paragraph's second line — or when the block shares
2921 /// a line with something else (an HTML `<p>` written beside another).
2922 /// [`Error::InvalidArgument`] when `to` is inside the block being moved
2923 /// or at a boundary it already sits on that is no container's edge,
2924 /// which would move nothing, and when either offset is past the source. [`Error::UnsupportedFormat`]
2925 /// where the format has no blocks a caret could name (XML); ask
2926 /// [`Format::supports`] with [`Gesture::MoveBlock`].
2927 pub fn move_block(&mut self, from: usize, to: usize) -> Result<Change, Error> {
2928 self.change_op(|ed, out| unsafe { ffi::twig_editor_move_block(ed, from, to, out) })
2929 }
2930
2931 /// Toggle a fenced code block over the blocks `[start, end)` covers: fence
2932 /// them if the caret is not in a code block, unfence the one it is in if it
2933 /// is. `language` tags the opening fence and is ignored when unfencing.
2934 ///
2935 /// `None` and `Some("")` are different requests: both write a bare fence, but
2936 /// the second says the caller asked for an empty info string. Reading the
2937 /// language back gives `None` either way — the distinction is in the ask, not
2938 /// the bytes. (Across the C ABI this rides as the `(ptr, len, has_*)` triple,
2939 /// the same spelling [`Builder::add_code_block`] uses for the same value.)
2940 ///
2941 /// Fencing *inserts* at the covered region's edges rather than rewriting its
2942 /// lines, so a body already carrying a quote's `> ` keeps it and the fence
2943 /// lines get the same prefix. The fence is **measured** — one character
2944 /// longer than the longest run of the fence character in the body — so
2945 /// fencing text that itself contains a fence nests instead of closing early.
2946 ///
2947 /// Unfencing peels the opening line and, when there is one, the closing fence
2948 /// line; a Markdown *indented* code block has no fence to peel and is
2949 /// dedented instead, so the toggle stays reversible on the older spelling.
2950 /// Note that unfencing can yield a different tree than the one that was
2951 /// fenced: a code body is by definition text the parser did not read as
2952 /// markup, so `# x` inside a fence becomes a heading once the fence is gone.
2953 ///
2954 /// [`Error::NotEditable`] **inside a list item**, in both directions: a
2955 /// quote's marker is on every line, a list item's is on its first line only,
2956 /// so a fence at column zero there would pull the `- ` into the code body and
2957 /// the item would stop being an item. [`Error::InvalidArgument`] for an info
2958 /// string the fence cannot carry (a line end, the fence character, or — in
2959 /// Markdown, whose info string ends at whitespace — a space);
2960 /// [`Error::UnsupportedFormat`] for a parse-only format;
2961 /// [`Error::NotFound`] when no block covers the range.
2962 pub fn toggle_code_block(
2963 &mut self,
2964 start: usize,
2965 end: usize,
2966 language: Option<&str>,
2967 ) -> Result<Change, Error> {
2968 let (ptr, len, has) = opt_str(language);
2969 self.change_op(|ed, out| unsafe {
2970 ffi::twig_editor_toggle_code_block(ed, start, end, ptr, len, has, out)
2971 })
2972 }
2973
2974 /// Retag the code block at `offset` with `language`, or clear its info string
2975 /// with `None` — the language dropdown beside a code block. Same
2976 /// `None`/`Some("")` distinction as [`Editor::toggle_code_block`].
2977 ///
2978 /// Only the info string is rewritten; the fence's own width is kept, because
2979 /// it was measured against a body this does not touch. [`Error::NotEditable`]
2980 /// for an *indented* Markdown code block, which has no fence and so nowhere
2981 /// to carry a language; [`Error::NotFound`] when `offset` is not in a code
2982 /// block.
2983 pub fn set_code_language(
2984 &mut self,
2985 offset: usize,
2986 language: Option<&str>,
2987 ) -> Result<Change, Error> {
2988 let (ptr, len, has) = opt_str(language);
2989 self.change_op(|ed, out| unsafe {
2990 ffi::twig_editor_set_code_language(ed, offset, ptr, len, has, out)
2991 })
2992 }
2993
2994 /// Add a checkbox to the list item at `offset`, or take one away — the
2995 /// gesture that converts between a plain list item and a task list item. A
2996 /// box is added unchecked; [`Editor::set_task_checked`] ticks it.
2997 ///
2998 /// The box is inline content of the item's first paragraph, not part of its
2999 /// marker, so adding or removing one leaves the item's continuation-line
3000 /// indentation alone. An item inside a quote is found past the quote markers.
3001 /// [`Error::NotFound`] when `offset` is in no list item;
3002 /// [`Error::NotEditable`] when the item's line carries no recognizable list
3003 /// marker; [`Error::UnsupportedFormat`] for a format with no checkbox.
3004 pub fn toggle_task_item(&mut self, offset: usize) -> Result<Change, Error> {
3005 self.change_op(|ed, out| unsafe { ffi::twig_editor_toggle_task_item(ed, offset, out) })
3006 }
3007
3008 /// Tick or untick the task item at `offset` — a checkbox click when the
3009 /// caller knows which way it should end up.
3010 ///
3011 /// Rewrites the box alone, never the space after it, so an item spelled with
3012 /// unusual spacing keeps it. A capital `[X]` is read as checked.
3013 ///
3014 /// When the box is already in the requested state this is a no-op that still
3015 /// returns `Ok` — the source is left byte-for-byte unchanged. The `Change` is
3016 /// not returned because a no-op has none; re-read [`Editor::source_str`].
3017 ///
3018 /// [`Error::NotEditable`] when the item has no box: minting one here would
3019 /// make "set checked" silently convert a bullet into a task, which is
3020 /// [`Editor::toggle_task_item`]'s job to do explicitly.
3021 pub fn set_task_checked(&mut self, offset: usize, checked: bool) -> Result<(), Error> {
3022 self.change_op(|ed, out| unsafe {
3023 ffi::twig_editor_set_task_checked(ed, offset, checked as c_int, out)
3024 })?;
3025 Ok(())
3026 }
3027
3028 /// Flip the task item at `offset` — what a checkbox click actually is when
3029 /// the caller does not already know the state. Always edits or fails, so
3030 /// unlike [`Editor::set_task_checked`] there is no silent no-op and the
3031 /// `Change` is always real.
3032 pub fn toggle_task_checked(&mut self, offset: usize) -> Result<Change, Error> {
3033 self.change_op(|ed, out| unsafe { ffi::twig_editor_toggle_task_checked(ed, offset, out) })
3034 }
3035
3036 /// Insert a footnote reference at `offset` and, unless the label is already
3037 /// defined, the matching definition at the end of the document.
3038 ///
3039 /// It writes **both halves**, because in neither format is half a footnote a
3040 /// footnote: a bare `[^a]` with nothing defining it renders as four literal
3041 /// characters. The definition body is left empty — that parses, and the
3042 /// caller then types into it like any other block. A label that is already
3043 /// defined gets only the reference, so referring to one footnote twice does
3044 /// not append a second, dead definition.
3045 ///
3046 /// It is **one** edit, spanning the caret to the end of the document even
3047 /// though the halves are far apart: two edits would take two undos to
3048 /// reverse, and the returned `Change` would describe only the second,
3049 /// omitting the reference the caret is sitting in.
3050 ///
3051 /// [`Error::InvalidArgument`] for a label that is empty or holds a line end
3052 /// or a reference bracket; [`Error::UnsupportedFormat`] for a format with no
3053 /// footnotes.
3054 pub fn insert_footnote(&mut self, offset: usize, label: &str) -> Result<Change, Error> {
3055 self.change_op(|ed, out| unsafe {
3056 ffi::twig_editor_insert_footnote(ed, offset, label.as_ptr(), label.len(), out)
3057 })
3058 }
3059
3060 /// Shared plumbing for the change-returning ops: run `op` (which fills a
3061 /// `TwigChange` out-param) and wrap the result.
3062 fn change_op(
3063 &mut self,
3064 op: impl FnOnce(*mut ffi::TwigEditor, *mut ffi::TwigChange) -> ffi::TwigStatus,
3065 ) -> Result<Change, Error> {
3066 let mut change = ffi::TwigChange {
3067 old_span: ffi::TwigSpan { start: 0, end: 0 },
3068 new_span: ffi::TwigSpan { start: 0, end: 0 },
3069 };
3070 let status = op(self.raw.as_ptr(), &mut change);
3071 Error::from_status(status)?;
3072 Ok(Change::from_ffi(change))
3073 }
3074
3075 /// Shared plumbing for the `(locator, text)` edit ops.
3076 fn apply(
3077 &mut self,
3078 locator: &str,
3079 text: &str,
3080 op: impl FnOnce(*mut ffi::TwigEditor, *const u8, usize, *const u8, usize) -> ffi::TwigStatus,
3081 ) -> Result<(), Error> {
3082 let status = op(
3083 self.raw.as_ptr(),
3084 locator.as_ptr(),
3085 locator.len(),
3086 text.as_ptr(),
3087 text.len(),
3088 );
3089 Error::from_status(status)
3090 }
3091}
3092
3093impl Drop for Editor {
3094 fn drop(&mut self) {
3095 unsafe { ffi::twig_editor_destroy(self.raw.as_ptr()) }
3096 }
3097}
3098
3099/// Run `call` (which writes a borrowed `(ptr, len)` byte buffer) and copy the
3100/// result into an owned `Vec` — the buffer is only valid until the next
3101/// same-accessor call on the handle, so we copy before returning. Shared by
3102/// [`Document`] and [`Editor`].
3103fn collect_bytes(
3104 call: impl FnOnce(*mut *const u8, *mut usize) -> ffi::TwigStatus,
3105) -> Result<Vec<u8>, Error> {
3106 let mut ptr = std::ptr::null();
3107 let mut len = 0usize;
3108 let status = call(&mut ptr, &mut len);
3109 Error::from_status(status)?;
3110 if len == 0 {
3111 return Ok(Vec::new());
3112 }
3113 if ptr.is_null() {
3114 return Err(Error::Internal);
3115 }
3116 let bytes = unsafe { std::slice::from_raw_parts(ptr, len) };
3117 Ok(bytes.to_vec())
3118}
3119
3120/// Run `call` (which writes a borrowed `(ptr, len)` match array) and copy each
3121/// match into an owned [`QueryMatch`]. Shared by [`Document`] and [`Editor`].
3122fn collect_matches(
3123 call: impl FnOnce(*mut *const ffi::TwigQueryMatch, *mut usize) -> ffi::TwigStatus,
3124) -> Result<Vec<QueryMatch>, Error> {
3125 let mut ptr = std::ptr::null();
3126 let mut len = 0usize;
3127 let status = call(&mut ptr, &mut len);
3128 Error::from_status(status)?;
3129 if len == 0 {
3130 return Ok(Vec::new());
3131 }
3132 if ptr.is_null() {
3133 return Err(Error::Internal);
3134 }
3135 let matches = unsafe { std::slice::from_raw_parts(ptr, len) };
3136 matches.iter().map(query_match_from_ffi).collect()
3137}
3138
3139/// `collect_matches` for the flat-node reads (`nodes` / `subtree`), which hand
3140/// back a borrowed [`ffi::TwigFlatNode`] array on the same contract.
3141fn collect_flat_nodes(
3142 call: impl FnOnce(*mut *const ffi::TwigFlatNode, *mut usize) -> ffi::TwigStatus,
3143) -> Result<Vec<FlatNode>, Error> {
3144 let mut ptr = std::ptr::null();
3145 let mut len = 0usize;
3146 let status = call(&mut ptr, &mut len);
3147 Error::from_status(status)?;
3148 if len == 0 {
3149 return Ok(Vec::new());
3150 }
3151 if ptr.is_null() {
3152 return Err(Error::Internal);
3153 }
3154 let nodes = unsafe { std::slice::from_raw_parts(ptr, len) };
3155 nodes.iter().map(flat_node_from_ffi).collect()
3156}
3157
3158/// The zeroed out-parameter the `node_at` hit-tests fill.
3159fn empty_ffi_match() -> ffi::TwigQueryMatch {
3160 ffi::TwigQueryMatch {
3161 node_id: 0,
3162 span: ffi::TwigSpan { start: 0, end: 0 },
3163 content_span: ffi::TwigSpan { start: 0, end: 0 },
3164 has_content_span: 0,
3165 kind: std::ptr::null(),
3166 }
3167}
3168
3169/// Copy a borrowed C ABI [`ffi::TwigQueryMatch`] into an owned [`QueryMatch`].
3170/// Shared by `collect_matches`, [`Editor::node_at`], and [`Editor::ancestors_at`].
3171fn query_match_from_ffi(m: &ffi::TwigQueryMatch) -> Result<QueryMatch, Error> {
3172 Ok(QueryMatch {
3173 node_id: m.node_id,
3174 span: m.span.start..m.span.end,
3175 content_span: if m.has_content_span != 0 {
3176 Some(m.content_span.start..m.content_span.end)
3177 } else {
3178 None
3179 },
3180 kind: Kind::from(borrowed_cstr(m.kind)?.as_str()),
3181 })
3182}
3183
3184/// Copy a borrowed C ABI [`ffi::TwigFlatNode`] into an owned [`FlatNode`].
3185fn flat_node_from_ffi(n: &ffi::TwigFlatNode) -> Result<FlatNode, Error> {
3186 let node_id = |v: u32| {
3187 if v == ffi::TWIG_NO_NODE {
3188 None
3189 } else {
3190 Some(NodeId(v))
3191 }
3192 };
3193 Ok(FlatNode {
3194 id: NodeId(n.id),
3195 parent: node_id(n.parent),
3196 first_child: node_id(n.first_child),
3197 next_sibling: node_id(n.next_sibling),
3198 span: n.span.start..n.span.end,
3199 content_span: if n.has_content_span != 0 {
3200 Some(n.content_span.start..n.content_span.end)
3201 } else {
3202 None
3203 },
3204 level: if n.level != 0 { Some(n.level) } else { None },
3205 kind: Kind::from(borrowed_cstr(n.kind)?.as_str()),
3206 text: borrowed_bytes(n.text_ptr, n.text_len),
3207 destination: borrowed_bytes(n.destination_ptr, n.destination_len),
3208 head: match n.head {
3209 ffi::TWIG_HEAD_NONE => None,
3210 v => Some(v != 0),
3211 },
3212 alignment: Alignment::from_c(n.alignment),
3213 name: borrowed_bytes(n.name_ptr, n.name_len),
3214 directive_form: DirectiveForm::from_c(n.directive_form),
3215 origin: ContainerOrigin::from_c(n.container_origin),
3216 marker_span: if n.has_marker_span != 0 {
3217 Some(n.marker_span.start..n.marker_span.end)
3218 } else {
3219 None
3220 },
3221 checked: match n.checked {
3222 ffi::TWIG_TASK_CHECKED_NONE => None,
3223 v => Some(v != 0),
3224 },
3225 attrs: borrowed_attrs(n.attrs_ptr, n.attrs_len),
3226 })
3227}
3228
3229/// Copy a borrowed `TwigKeyVal` array into owned `(key, value)` pairs, or an
3230/// empty vec for a NULL pointer (the node has no attributes). A bare attribute
3231/// (NULL `value`) maps to a `None` value, distinct from a present-but-empty one.
3232fn borrowed_attrs(ptr: *const ffi::TwigKeyVal, len: usize) -> Vec<(String, Option<String>)> {
3233 if ptr.is_null() || len == 0 {
3234 return Vec::new();
3235 }
3236 let kvs = unsafe { std::slice::from_raw_parts(ptr, len) };
3237 kvs.iter()
3238 .map(|kv| {
3239 let key = borrowed_bytes(kv.key, kv.key_len).unwrap_or_default();
3240 (key, borrowed_bytes(kv.value, kv.value_len))
3241 })
3242 .collect()
3243}
3244
3245/// Copy a NUL-terminated, library-owned C string into an owned `String`.
3246fn borrowed_cstr(ptr: *const c_char) -> Result<String, Error> {
3247 if ptr.is_null() {
3248 return Err(Error::Internal);
3249 }
3250 Ok(unsafe { std::ffi::CStr::from_ptr(ptr) }
3251 .to_str()
3252 .map_err(|_| Error::Internal)?
3253 .to_owned())
3254}
3255
3256/// Copy a borrowed `(ptr, len)` payload slice into an owned `String`, or `None`
3257/// for a NULL pointer (the kind carries no such payload). The bytes are a slice
3258/// of a UTF-8 document, so a lossy decode never actually substitutes.
3259fn borrowed_bytes(ptr: *const u8, len: usize) -> Option<String> {
3260 if ptr.is_null() {
3261 return None;
3262 }
3263 let bytes = unsafe { std::slice::from_raw_parts(ptr, len) };
3264 Some(String::from_utf8_lossy(bytes).into_owned())
3265}
3266
3267/// The id of a node added to a [`Builder`], returned by every `add*` method and
3268/// used to wire up the tree via [`Builder::set_children`] and to root a
3269/// render/serialize/query.
3270#[derive(Clone, Copy, Debug, Eq, PartialEq, Hash)]
3271pub struct NodeId(pub u32);
3272
3273/// The void-payload node kinds, addable via [`Builder::add`]. Kinds with a
3274/// payload have their own dedicated `add_*` method instead.
3275#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3276pub enum VoidKind {
3277 Doc,
3278 Para,
3279 ThematicBreak,
3280 Section,
3281 Div,
3282 BlockQuote,
3283 DefinitionList,
3284 Table,
3285 ListItem,
3286 DefinitionListItem,
3287 Term,
3288 Definition,
3289 Caption,
3290 SoftBreak,
3291 HardBreak,
3292 NonBreakingSpace,
3293 Emph,
3294 Strong,
3295 Span,
3296 Mark,
3297 Superscript,
3298 Subscript,
3299 Insert,
3300 Delete,
3301 DoubleQuoted,
3302 SingleQuoted,
3303}
3304
3305impl VoidKind {
3306 fn to_c(self) -> c_int {
3307 // Discriminants match `TwigNodeKind` in the C ABI.
3308 match self {
3309 VoidKind::Doc => 0,
3310 VoidKind::Para => 1,
3311 VoidKind::ThematicBreak => 3,
3312 VoidKind::Section => 4,
3313 VoidKind::Div => 5,
3314 VoidKind::BlockQuote => 9,
3315 VoidKind::DefinitionList => 13,
3316 VoidKind::Table => 14,
3317 VoidKind::ListItem => 15,
3318 VoidKind::DefinitionListItem => 17,
3319 VoidKind::Term => 18,
3320 VoidKind::Definition => 19,
3321 VoidKind::Caption => 22,
3322 VoidKind::SoftBreak => 26,
3323 VoidKind::HardBreak => 27,
3324 VoidKind::NonBreakingSpace => 28,
3325 VoidKind::Emph => 38,
3326 VoidKind::Strong => 39,
3327 VoidKind::Span => 42,
3328 VoidKind::Mark => 43,
3329 VoidKind::Superscript => 44,
3330 VoidKind::Subscript => 45,
3331 VoidKind::Insert => 46,
3332 VoidKind::Delete => 47,
3333 VoidKind::DoubleQuoted => 48,
3334 VoidKind::SingleQuoted => 49,
3335 }
3336 }
3337}
3338
3339/// The single-string-payload node kinds, addable via [`Builder::add_text`].
3340#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3341pub enum TextKind {
3342 Str,
3343 Symb,
3344 Verbatim,
3345 InlineMath,
3346 DisplayMath,
3347 Url,
3348 Email,
3349 FootnoteReference,
3350 /// reStructuredText's `[CIT2002]_` — a use of a citation definition. The
3351 /// payload is the label as WRITTEN, not the normalized name it resolves by.
3352 CitationReference,
3353 /// reStructuredText's `|name|` — a use of a substitution definition.
3354 SubstitutionReference,
3355 Comment,
3356 Doctype,
3357 Cdata,
3358}
3359
3360impl TextKind {
3361 fn to_c(self) -> c_int {
3362 match self {
3363 TextKind::Str => 25,
3364 TextKind::Symb => 29,
3365 TextKind::Verbatim => 30,
3366 TextKind::InlineMath => 32,
3367 TextKind::DisplayMath => 33,
3368 TextKind::Url => 34,
3369 TextKind::Email => 35,
3370 TextKind::FootnoteReference => 36,
3371 TextKind::CitationReference => 58,
3372 TextKind::SubstitutionReference => 59,
3373 TextKind::Comment => 52,
3374 TextKind::Doctype => 53,
3375 TextKind::Cdata => 55,
3376 }
3377 }
3378}
3379
3380/// Bullet marker style for [`Builder::add_bullet_list`].
3381#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3382pub enum BulletStyle {
3383 Dash,
3384 Plus,
3385 Star,
3386}
3387
3388impl BulletStyle {
3389 fn to_c(self) -> c_int {
3390 match self {
3391 BulletStyle::Dash => 0,
3392 BulletStyle::Plus => 1,
3393 BulletStyle::Star => 2,
3394 }
3395 }
3396}
3397
3398/// Numbering scheme for [`Builder::add_ordered_list`].
3399#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3400pub enum OrderedNumbering {
3401 Decimal,
3402 LowerAlpha,
3403 UpperAlpha,
3404 LowerRoman,
3405 UpperRoman,
3406}
3407
3408impl OrderedNumbering {
3409 fn to_c(self) -> c_int {
3410 match self {
3411 OrderedNumbering::Decimal => 0,
3412 OrderedNumbering::LowerAlpha => 1,
3413 OrderedNumbering::UpperAlpha => 2,
3414 OrderedNumbering::LowerRoman => 3,
3415 OrderedNumbering::UpperRoman => 4,
3416 }
3417 }
3418}
3419
3420/// Delimiter around an ordered-list number (`1.`, `1)`, `(1)`).
3421#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3422pub enum OrderedDelim {
3423 Period,
3424 ParenAfter,
3425 ParenBoth,
3426}
3427
3428impl OrderedDelim {
3429 fn to_c(self) -> c_int {
3430 match self {
3431 OrderedDelim::Period => 0,
3432 OrderedDelim::ParenAfter => 1,
3433 OrderedDelim::ParenBoth => 2,
3434 }
3435 }
3436}
3437
3438/// Table-cell alignment: written via [`Builder::add_cell`], read back on
3439/// [`FlatNode::alignment`].
3440#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3441pub enum Alignment {
3442 Default,
3443 Left,
3444 Right,
3445 Center,
3446}
3447
3448impl Alignment {
3449 fn to_c(self) -> c_int {
3450 match self {
3451 Alignment::Default => ffi::TWIG_ALIGN_DEFAULT,
3452 Alignment::Left => ffi::TWIG_ALIGN_LEFT,
3453 Alignment::Right => ffi::TWIG_ALIGN_RIGHT,
3454 Alignment::Center => ffi::TWIG_ALIGN_CENTER,
3455 }
3456 }
3457
3458 /// The inverse of [`Alignment::to_c`]; `None` for [`ffi::TWIG_ALIGN_NONE`]
3459 /// (the node isn't a cell) or any code this binding doesn't know.
3460 fn from_c(v: c_int) -> Option<Self> {
3461 match v {
3462 ffi::TWIG_ALIGN_DEFAULT => Some(Alignment::Default),
3463 ffi::TWIG_ALIGN_LEFT => Some(Alignment::Left),
3464 ffi::TWIG_ALIGN_RIGHT => Some(Alignment::Right),
3465 ffi::TWIG_ALIGN_CENTER => Some(Alignment::Center),
3466 _ => None,
3467 }
3468 }
3469}
3470
3471/// The smart-punctuation kind for [`Builder::add_smart_punctuation`].
3472#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3473pub enum SmartPunctuation {
3474 LeftSingleQuote,
3475 RightSingleQuote,
3476 LeftDoubleQuote,
3477 RightDoubleQuote,
3478 Ellipses,
3479 EmDash,
3480 EnDash,
3481}
3482
3483impl SmartPunctuation {
3484 fn to_c(self) -> c_int {
3485 match self {
3486 SmartPunctuation::LeftSingleQuote => 0,
3487 SmartPunctuation::RightSingleQuote => 1,
3488 SmartPunctuation::LeftDoubleQuote => 2,
3489 SmartPunctuation::RightDoubleQuote => 3,
3490 SmartPunctuation::Ellipses => 4,
3491 SmartPunctuation::EmDash => 5,
3492 SmartPunctuation::EnDash => 6,
3493 }
3494 }
3495}
3496
3497/// Whether a generic container was written as a TAG or as a DIRECTIVE — the
3498/// axis [`DirectiveForm`] is repeatedly mistaken for and cannot answer.
3499///
3500/// `DirectiveForm` is a spelling hint: which of a directive-capable format's
3501/// three spellings fits this node. Twig's HTML parser sets one on `<div>` and
3502/// `<span>` because those are the two tags djot and Markdown have generic
3503/// spellings for — so a `<div>` and a Markdown `:::div` produce nodes that
3504/// agree on kind, name and form alike. Until this axis existed, the only way
3505/// to separate them was to re-read the source bytes under the node's span and
3506/// look at the first character.
3507///
3508/// Read-only: it records what a parser saw, and there is nothing to set on the
3509/// build path.
3510/// One thing a conversion would silently lose. See [`Document::diagnostics`].
3511#[derive(Clone, Debug, Eq, PartialEq)]
3512pub struct Warning {
3513 pub fidelity: Fidelity,
3514 /// A slash-separated child-index trail from the document root (`"1/0/2"`),
3515 /// EMPTY for the root itself.
3516 ///
3517 /// A path and not a byte span, because the output being described does not
3518 /// exist yet — there is nothing in it to point at. Resolve it against the
3519 /// tree you already have.
3520 pub path: String,
3521 /// The affected node's kind, with family members reported as themselves
3522 /// ([`Kind::Superscript`], not an `inline_mark`).
3523 pub kind: Kind,
3524}
3525
3526/// How much of a node survives a conversion.
3527#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3528#[non_exhaustive]
3529pub enum Fidelity {
3530 /// Something is emitted, but the target's parser reads it back as a
3531 /// DIFFERENT kind. The content survives; its meaning does not.
3532 Degraded,
3533 /// Nothing is emitted at all: the node and its subtree leave no trace.
3534 Dropped,
3535 /// Some of the node's **attributes** are written where the target's
3536 /// parser does not read them back as that node's — djot spells a
3537 /// heading's on its text, AsciiDoc moves a section title's onto the
3538 /// section. Independent of the node's own answer: a node that degrades
3539 /// and still has its attributes written — Markdown's `<div lang=…>` for
3540 /// a section it cannot spell — is reported twice at the same path,
3541 /// [`Degraded`](Fidelity::Degraded) and then this. Which keys is not carried across the C boundary: the node at
3542 /// [`Warning::path`] has them, and a consumer that wants the list reads
3543 /// them from the tree.
3544 AttrsDegraded,
3545 /// Some of the node's attributes are not written at all. Never reported
3546 /// beside [`Dropped`](Fidelity::Dropped): nothing of a dropped node is
3547 /// written, so that one variant is the whole answer.
3548 AttrsDropped,
3549}
3550
3551impl Fidelity {
3552 /// Only the lossy codes have a variant — a faithful node is never reported
3553 /// as a warning, so there is nothing for it to map to. An unknown code
3554 /// reads as [`Fidelity::Degraded`], the weakest of the claims.
3555 fn from_c(v: c_int) -> Self {
3556 match v {
3557 ffi::TWIG_FIDELITY_DROPPED => Fidelity::Dropped,
3558 ffi::TWIG_FIDELITY_ATTRS_DEGRADED => Fidelity::AttrsDegraded,
3559 ffi::TWIG_FIDELITY_ATTRS_DROPPED => Fidelity::AttrsDropped,
3560 _ => Fidelity::Degraded,
3561 }
3562 }
3563}
3564
3565#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3566#[non_exhaustive]
3567pub enum ContainerOrigin {
3568 /// An HTML or XML tag: `<div>`, `<video>`, `<svg:rect>`.
3569 Element,
3570 /// A lightweight-markup generic container: a djot fenced div or bracketed
3571 /// span, a Markdown `:::note` / `::name` / `:name`, an rST `.. note::`, an
3572 /// AsciiDoc delimited block.
3573 Directive,
3574}
3575
3576impl ContainerOrigin {
3577 /// `None` for [`ffi::TWIG_CONTAINER_ORIGIN_NONE`] (nothing recorded an
3578 /// origin) or any code this binding doesn't know.
3579 fn from_c(v: c_int) -> Option<Self> {
3580 match v {
3581 ffi::TWIG_CONTAINER_ORIGIN_ELEMENT => Some(ContainerOrigin::Element),
3582 ffi::TWIG_CONTAINER_ORIGIN_DIRECTIVE => Some(ContainerOrigin::Directive),
3583 _ => None,
3584 }
3585 }
3586}
3587
3588/// The surface form of a generic directive for [`Builder::add_directive`].
3589#[derive(Clone, Copy, Debug, Eq, PartialEq)]
3590pub enum DirectiveForm {
3591 Text,
3592 Leaf,
3593 Container,
3594}
3595
3596impl DirectiveForm {
3597 fn to_c(self) -> c_int {
3598 match self {
3599 DirectiveForm::Text => ffi::TWIG_DIRECTIVE_TEXT,
3600 DirectiveForm::Leaf => ffi::TWIG_DIRECTIVE_LEAF,
3601 DirectiveForm::Container => ffi::TWIG_DIRECTIVE_CONTAINER,
3602 }
3603 }
3604
3605 /// The inverse of [`DirectiveForm::to_c`]; `None` for
3606 /// [`ffi::TWIG_DIRECTIVE_NONE`] (the node isn't a directive) or any code
3607 /// this binding doesn't know.
3608 fn from_c(v: c_int) -> Option<Self> {
3609 match v {
3610 ffi::TWIG_DIRECTIVE_TEXT => Some(DirectiveForm::Text),
3611 ffi::TWIG_DIRECTIVE_LEAF => Some(DirectiveForm::Leaf),
3612 ffi::TWIG_DIRECTIVE_CONTAINER => Some(DirectiveForm::Container),
3613 _ => None,
3614 }
3615 }
3616}
3617
3618/// Decompose an optional string into `(ptr, len, has)` for the C ABI's
3619/// `(ptr, len, has_*)` optional-string triples. The pointer borrows `s` and is
3620/// only used within the same call.
3621fn opt_str(s: Option<&str>) -> (*const u8, usize, c_int) {
3622 match s {
3623 Some(x) => (x.as_ptr(), x.len(), 1),
3624 None => (std::ptr::null(), 0, 0),
3625 }
3626}
3627
3628/// Programmatic construction of a document — the write-path mirror of
3629/// [`Document::parse`]. Build the tree bottom-up (add children, then the
3630/// container, wiring them with [`Builder::set_children`]); every `add*` method
3631/// returns the new node's [`NodeId`]. Then render, serialize, query, or dump the
3632/// subtree rooted at any id, on demand, without consuming the builder. All input
3633/// strings are copied, so caller buffers need not outlive a call.
3634#[derive(Debug)]
3635pub struct Builder {
3636 raw: NonNull<ffi::TwigBuilder>,
3637}
3638
3639impl Builder {
3640 /// Create an empty builder.
3641 pub fn new() -> Result<Self, Error> {
3642 let mut raw = std::ptr::null_mut();
3643 let status = unsafe { ffi::twig_builder_create(&mut raw) };
3644 Error::from_status(status)?;
3645 let raw = NonNull::new(raw).ok_or(Error::Internal)?;
3646 Ok(Self { raw })
3647 }
3648
3649 /// Add a void-payload node (attach children later with
3650 /// [`Builder::set_children`]).
3651 pub fn add(&mut self, kind: VoidKind) -> Result<NodeId, Error> {
3652 self.emit(|b, out| unsafe { ffi::twig_builder_add(b, kind.to_c(), out) })
3653 }
3654
3655 /// Add a single-string-payload node (a `str`, code span, url, comment, …).
3656 pub fn add_text(&mut self, kind: TextKind, text: &str) -> Result<NodeId, Error> {
3657 self.emit(|b, out| unsafe {
3658 ffi::twig_builder_add_text(b, kind.to_c(), text.as_ptr(), text.len(), out)
3659 })
3660 }
3661
3662 /// Add a heading of the given level (attach its inline children afterward).
3663 pub fn add_heading(&mut self, level: u32) -> Result<NodeId, Error> {
3664 self.emit(|b, out| unsafe { ffi::twig_builder_add_heading(b, level, out) })
3665 }
3666
3667 /// Add a code block, with an optional info-string language.
3668 pub fn add_code_block(&mut self, lang: Option<&str>, text: &str) -> Result<NodeId, Error> {
3669 let (lp, ll, has) = opt_str(lang);
3670 self.emit(|b, out| unsafe {
3671 ffi::twig_builder_add_code_block(b, lp, ll, has, text.as_ptr(), text.len(), out)
3672 })
3673 }
3674
3675 /// Add a raw block targeting `format` (e.g. `"html"`).
3676 pub fn add_raw_block(&mut self, format: &str, text: &str) -> Result<NodeId, Error> {
3677 self.emit(|b, out| unsafe {
3678 ffi::twig_builder_add_raw_block(
3679 b,
3680 format.as_ptr(),
3681 format.len(),
3682 text.as_ptr(),
3683 text.len(),
3684 out,
3685 )
3686 })
3687 }
3688
3689 /// Add a document-metadata block written in config language `lang`.
3690 pub fn add_metadata(&mut self, lang: &str, text: &str) -> Result<NodeId, Error> {
3691 self.emit(|b, out| unsafe {
3692 ffi::twig_builder_add_metadata(
3693 b,
3694 lang.as_ptr(),
3695 lang.len(),
3696 text.as_ptr(),
3697 text.len(),
3698 out,
3699 )
3700 })
3701 }
3702
3703 /// Add a raw inline targeting `format`.
3704 pub fn add_raw_inline(&mut self, format: &str, text: &str) -> Result<NodeId, Error> {
3705 self.emit(|b, out| unsafe {
3706 ffi::twig_builder_add_raw_inline(
3707 b,
3708 format.as_ptr(),
3709 format.len(),
3710 text.as_ptr(),
3711 text.len(),
3712 out,
3713 )
3714 })
3715 }
3716
3717 /// Add a smart-punctuation node of `kind`. `text` is accepted for ABI
3718 /// compatibility but ignored by the underlying builder: the node's
3719 /// spelling is always the canonical one for `kind` (e.g. `"---"` for an
3720 /// em dash), never a caller-supplied one.
3721 pub fn add_smart_punctuation(
3722 &mut self,
3723 kind: SmartPunctuation,
3724 text: &str,
3725 ) -> Result<NodeId, Error> {
3726 self.emit(|b, out| unsafe {
3727 ffi::twig_builder_add_smart_punctuation(b, kind.to_c(), text.as_ptr(), text.len(), out)
3728 })
3729 }
3730
3731 /// Add a link with an optional destination and/or reference label (attach
3732 /// the link text as children).
3733 pub fn add_link(
3734 &mut self,
3735 destination: Option<&str>,
3736 reference: Option<&str>,
3737 ) -> Result<NodeId, Error> {
3738 let (dp, dl, hd) = opt_str(destination);
3739 let (rp, rl, hr) = opt_str(reference);
3740 self.emit(|b, out| unsafe { ffi::twig_builder_add_link(b, dp, dl, hd, rp, rl, hr, out) })
3741 }
3742
3743 /// Add an image — like [`Builder::add_link`], but children are the alt text.
3744 pub fn add_image(
3745 &mut self,
3746 destination: Option<&str>,
3747 reference: Option<&str>,
3748 ) -> Result<NodeId, Error> {
3749 let (dp, dl, hd) = opt_str(destination);
3750 let (rp, rl, hr) = opt_str(reference);
3751 self.emit(|b, out| unsafe { ffi::twig_builder_add_image(b, dp, dl, hd, rp, rl, hr, out) })
3752 }
3753
3754 /// Add a generic directive of the given form and name.
3755 pub fn add_directive(&mut self, form: DirectiveForm, name: &str) -> Result<NodeId, Error> {
3756 self.emit(|b, out| unsafe {
3757 ffi::twig_builder_add_directive(b, form.to_c(), name.as_ptr(), name.len(), out)
3758 })
3759 }
3760
3761 /// Add a generic named element (the escape hatch for HTML/XML tags).
3762 pub fn add_element(&mut self, name: &str) -> Result<NodeId, Error> {
3763 self.emit(|b, out| unsafe {
3764 ffi::twig_builder_add_element(b, name.as_ptr(), name.len(), out)
3765 })
3766 }
3767
3768 /// Add an XML processing instruction (`<?target data?>`).
3769 pub fn add_processing_instruction(
3770 &mut self,
3771 target: &str,
3772 data: &str,
3773 ) -> Result<NodeId, Error> {
3774 self.emit(|b, out| unsafe {
3775 ffi::twig_builder_add_processing_instruction(
3776 b,
3777 target.as_ptr(),
3778 target.len(),
3779 data.as_ptr(),
3780 data.len(),
3781 out,
3782 )
3783 })
3784 }
3785
3786 /// Add a footnote definition with the given label.
3787 pub fn add_footnote(&mut self, label: &str) -> Result<NodeId, Error> {
3788 self.emit(|b, out| unsafe {
3789 ffi::twig_builder_add_footnote(b, label.as_ptr(), label.len(), out)
3790 })
3791 }
3792
3793 /// Add a citation definition — reStructuredText's `.. [CIT2002] ...`. Holds
3794 /// blocks, like a footnote; the two differ in which name registry resolves
3795 /// them, which is why this is its own call and not a flag on
3796 /// [`Builder::add_footnote`].
3797 pub fn add_citation(&mut self, label: &str) -> Result<NodeId, Error> {
3798 self.emit(|b, out| unsafe {
3799 ffi::twig_builder_add_citation(b, label.as_ptr(), label.len(), out)
3800 })
3801 }
3802
3803 /// Add a substitution definition — reStructuredText's
3804 /// `.. |name| image:: p.png`. Unlike a footnote or citation, its children
3805 /// are INLINE nodes.
3806 pub fn add_substitution(&mut self, label: &str) -> Result<NodeId, Error> {
3807 self.emit(|b, out| unsafe {
3808 ffi::twig_builder_add_substitution(b, label.as_ptr(), label.len(), out)
3809 })
3810 }
3811
3812 /// Add a link/image reference definition (`label` → `destination`).
3813 pub fn add_reference(&mut self, label: &str, destination: &str) -> Result<NodeId, Error> {
3814 self.emit(|b, out| unsafe {
3815 ffi::twig_builder_add_reference(
3816 b,
3817 label.as_ptr(),
3818 label.len(),
3819 destination.as_ptr(),
3820 destination.len(),
3821 out,
3822 )
3823 })
3824 }
3825
3826 /// Add a bullet list.
3827 pub fn add_bullet_list(&mut self, style: BulletStyle, tight: bool) -> Result<NodeId, Error> {
3828 self.emit(|b, out| unsafe {
3829 ffi::twig_builder_add_bullet_list(b, style.to_c(), tight as c_int, out)
3830 })
3831 }
3832
3833 /// Add an ordered list, with an optional explicit start number.
3834 pub fn add_ordered_list(
3835 &mut self,
3836 numbering: OrderedNumbering,
3837 delim: OrderedDelim,
3838 tight: bool,
3839 start: Option<u32>,
3840 ) -> Result<NodeId, Error> {
3841 let (start_val, has_start) = match start {
3842 Some(s) => (s, 1),
3843 None => (0, 0),
3844 };
3845 self.emit(|b, out| unsafe {
3846 ffi::twig_builder_add_ordered_list(
3847 b,
3848 numbering.to_c(),
3849 delim.to_c(),
3850 tight as c_int,
3851 start_val,
3852 has_start,
3853 out,
3854 )
3855 })
3856 }
3857
3858 /// Add a task list.
3859 pub fn add_task_list(&mut self, tight: bool) -> Result<NodeId, Error> {
3860 self.emit(|b, out| unsafe { ffi::twig_builder_add_task_list(b, tight as c_int, out) })
3861 }
3862
3863 /// Add a task-list item with the given checkbox state.
3864 pub fn add_task_list_item(&mut self, checked: bool) -> Result<NodeId, Error> {
3865 self.emit(|b, out| unsafe {
3866 ffi::twig_builder_add_task_list_item(b, checked as c_int, out)
3867 })
3868 }
3869
3870 /// Add a table row (`head` marks a header row).
3871 pub fn add_row(&mut self, head: bool) -> Result<NodeId, Error> {
3872 self.emit(|b, out| unsafe { ffi::twig_builder_add_row(b, head as c_int, out) })
3873 }
3874
3875 /// Add a one-square table cell (`head` marks a header cell).
3876 pub fn add_cell(&mut self, head: bool, alignment: Alignment) -> Result<NodeId, Error> {
3877 self.emit(|b, out| unsafe {
3878 ffi::twig_builder_add_cell(b, head as c_int, alignment.to_c(), out)
3879 })
3880 }
3881
3882 /// Add a table cell occupying `colspan` columns and `rowspan` rows — a grid
3883 /// table's merged cell. Both must be at least 1
3884 /// ([`Error::InvalidArgument`] otherwise); `(1, 1)` is exactly
3885 /// [`Builder::add_cell`]. Read back with [`Document::cell_extent`].
3886 pub fn add_cell_spanning(
3887 &mut self,
3888 head: bool,
3889 alignment: Alignment,
3890 colspan: u32,
3891 rowspan: u32,
3892 ) -> Result<NodeId, Error> {
3893 self.emit(|b, out| unsafe {
3894 ffi::twig_builder_add_cell_spanning(
3895 b,
3896 head as c_int,
3897 alignment.to_c(),
3898 colspan,
3899 rowspan,
3900 out,
3901 )
3902 })
3903 }
3904
3905 /// Set `parent`'s children to `children` (in order), replacing any it had.
3906 /// Each child id should appear in exactly one `set_children` call.
3907 pub fn set_children(&mut self, parent: NodeId, children: &[NodeId]) -> Result<(), Error> {
3908 let ids: Vec<u32> = children.iter().map(|n| n.0).collect();
3909 let status = unsafe {
3910 ffi::twig_builder_set_children(self.raw.as_ptr(), parent.0, ids.as_ptr(), ids.len())
3911 };
3912 Error::from_status(status)
3913 }
3914
3915 /// Attach `{...}` attributes to `id` (`(key, Some(value))`, or
3916 /// `(key, None)` for a bare attribute), replacing any it had. An empty slice
3917 /// clears them.
3918 pub fn set_attrs(&mut self, id: NodeId, attrs: &[(&str, Option<&str>)]) -> Result<(), Error> {
3919 let kvs: Vec<ffi::TwigKeyVal> = attrs
3920 .iter()
3921 .map(|(k, v)| ffi::TwigKeyVal {
3922 key: k.as_ptr(),
3923 key_len: k.len(),
3924 value: v.map_or(std::ptr::null(), |s| s.as_ptr()),
3925 value_len: v.map_or(0, |s| s.len()),
3926 })
3927 .collect();
3928 let status = unsafe {
3929 ffi::twig_builder_set_attrs(self.raw.as_ptr(), id.0, kvs.as_ptr(), kvs.len())
3930 };
3931 Error::from_status(status)
3932 }
3933
3934 /// Render the subtree rooted at `root` to HTML (generic whole-vocabulary
3935 /// printer — a built tree has no djot/Markdown side tables).
3936 pub fn render_html(&mut self, root: NodeId) -> Result<Vec<u8>, Error> {
3937 let raw = self.raw.as_ptr();
3938 collect_bytes(|ptr, len| unsafe { ffi::twig_builder_render_html(raw, root.0, ptr, len) })
3939 }
3940
3941 /// Serialize the subtree rooted at `root` to `target`'s syntax. Returns
3942 /// [`Error::UnsupportedFormat`] when the target can't represent the built
3943 /// tree (e.g. semantic kinds into XML).
3944 ///
3945 /// Prefer this over [`Builder::serialize`], for the reason
3946 /// [`Document::serialize_to`] gives.
3947 pub fn serialize_to(&mut self, root: NodeId, target: Target) -> Result<Vec<u8>, Error> {
3948 let raw = self.raw.as_ptr();
3949 let ffi_target: ffi::TwigFormat = target.into();
3950 collect_bytes(|ptr, len| unsafe {
3951 ffi::twig_builder_serialize(raw, root.0, ffi_target as i32, ptr, len)
3952 })
3953 }
3954
3955 /// Serialize the subtree rooted at `root` to `format`'s source syntax.
3956 ///
3957 /// The original spelling of [`Builder::serialize_to`], kept for
3958 /// compatibility and defined in terms of it.
3959 pub fn serialize(&mut self, root: NodeId, format: Format) -> Result<Vec<u8>, Error> {
3960 self.serialize_to(root, format.into())
3961 }
3962
3963 /// Encode the subtree rooted at `root` as pretty-printed JSON.
3964 pub fn ast_json(&mut self, root: NodeId) -> Result<Vec<u8>, Error> {
3965 let raw = self.raw.as_ptr();
3966 collect_bytes(|ptr, len| unsafe { ffi::twig_builder_ast_json(raw, root.0, ptr, len) })
3967 }
3968
3969 /// Resolve a selector against the subtree rooted at `root` (same grammar as
3970 /// [`Document::query`]).
3971 pub fn query(&mut self, root: NodeId, selector: &str) -> Result<Vec<QueryMatch>, Error> {
3972 let raw = self.raw.as_ptr();
3973 collect_matches(|ptr, len| unsafe {
3974 ffi::twig_builder_query(raw, root.0, selector.as_ptr(), selector.len(), ptr, len)
3975 })
3976 }
3977
3978 /// Shared plumbing for the `add*` constructors: run `call` (which writes the
3979 /// new node's id) and wrap the result.
3980 fn emit(
3981 &mut self,
3982 call: impl FnOnce(*mut ffi::TwigBuilder, *mut u32) -> ffi::TwigStatus,
3983 ) -> Result<NodeId, Error> {
3984 let mut id: u32 = 0;
3985 let status = call(self.raw.as_ptr(), &mut id);
3986 Error::from_status(status)?;
3987 Ok(NodeId(id))
3988 }
3989}
3990
3991impl Drop for Builder {
3992 fn drop(&mut self) {
3993 unsafe { ffi::twig_builder_destroy(self.raw.as_ptr()) }
3994 }
3995}
3996
3997#[cfg(test)]
3998mod tests {
3999 use super::*;
4000
4001 #[test]
4002 fn abi_version_matches() {
4003 // The linked library must speak the exact ABI layout this crate's
4004 // `#[repr(C)]` mirrors assume. If this fails, the Zig `TWIG_ABI_VERSION`
4005 // was bumped without updating `ffi::TWIG_ABI_VERSION` (and the mirrors).
4006 assert_eq!(abi_version(), ffi::TWIG_ABI_VERSION);
4007 }
4008
4009 #[test]
4010 fn parses_and_renders_markdown_html() {
4011 let mut doc = Document::parse_str("# hi\n", Format::Markdown).expect("parse markdown");
4012 let html = doc.render_html().expect("render html");
4013 assert_eq!(String::from_utf8_lossy(&html), "<h1>hi</h1>\n");
4014 }
4015
4016 #[test]
4017 fn parses_html_input() {
4018 let mut doc = Document::parse_str("<p>hi</p>", Format::Html).expect("parse html");
4019 let html = doc.render_html().expect("render html");
4020 assert!(String::from_utf8_lossy(&html).contains("hi"));
4021 }
4022
4023 #[test]
4024 fn parses_renders_and_writes_asciidoc() {
4025 let mut doc = Document::parse_str("= Title\n\nsome *bold* text\n", Format::Asciidoc)
4026 .expect("parse asciidoc");
4027 let html = String::from_utf8_lossy(&doc.render_html().expect("render html")).into_owned();
4028 assert!(html.contains("<h1>Title</h1>"), "got {html:?}");
4029 assert!(html.contains("<strong>bold</strong>"), "got {html:?}");
4030
4031 // Round-trip, and a cross-format conversion from Markdown.
4032 let back = doc.serialize_to(Target::Asciidoc).expect("serialize asciidoc");
4033 assert_eq!(String::from_utf8_lossy(&back), "= Title\n\nsome *bold* text\n");
4034 let mut md = Document::parse_str("# Title\n\nsome **bold** text\n", Format::Markdown)
4035 .expect("parse markdown");
4036 let converted = md.serialize_to(Target::Asciidoc).expect("convert to asciidoc");
4037 assert_eq!(String::from_utf8_lossy(&converted), "= Title\n\nsome *bold* text\n");
4038 assert_eq!(Target::from(Format::Asciidoc), Target::Asciidoc);
4039 assert_eq!(Target::Asciidoc.as_format(), Some(Format::Asciidoc));
4040 }
4041
4042 #[test]
4043 fn markdown_dialects_are_formats_over_one_parser() {
4044 // Three formats, one parser: strict CommonMark reads `~~x~~` as text
4045 // and a pipe table as a paragraph; GFM and the default read both; an
4046 // extension laid over GFM adds what GFM proper leaves out.
4047 let src = "a ~~b~~ c\n\n| x |\n| - |\n| $m$ |\n";
4048 let count = |doc: &mut Document, sel: &str| doc.query(sel).expect("query").len();
4049 for (format, ext, delete, table, math) in [
4050 (Format::Commonmark, MarkdownExtensions::default(), 0, 0, 0),
4051 (Format::Markdown, MarkdownExtensions::default(), 1, 1, 0),
4052 (Format::Gfm, MarkdownExtensions::default(), 1, 1, 0),
4053 (Format::Gfm, MarkdownExtensions { math: true, ..Default::default() }, 1, 1, 1),
4054 ] {
4055 let mut doc = Document::parse_str_with(src, format, ext).expect("parse");
4056 assert_eq!(count(&mut doc, "delete"), delete, "{format:?} {ext:?}");
4057 assert_eq!(count(&mut doc, "table"), table, "{format:?} {ext:?}");
4058 assert_eq!(count(&mut doc, "inline_math"), math, "{format:?} {ext:?}");
4059 }
4060
4061 // A dialect writes as its language, and a GFM document serialized as
4062 // Markdown is a round trip, spelling intact.
4063 assert_eq!(Format::Gfm.dialect_of(), Some(Format::Markdown));
4064 assert_eq!(Format::Commonmark.dialect_of(), Some(Format::Markdown));
4065 assert_eq!(Format::Markdown.dialect_of(), None);
4066 // And xml's one dialect, which writes as xml and takes xml's gesture.
4067 assert_eq!(Format::Svg.dialect_of(), Some(Format::Xml));
4068 assert_eq!(Target::from(Format::Svg), Target::Xml);
4069 let mut svg =
4070 Document::parse_str("<svg><rect x=\"1\"/></svg>", Format::Svg).expect("parse svg");
4071 assert_eq!(
4072 String::from_utf8_lossy(&svg.serialize(Format::Xml).expect("serialize")),
4073 "<svg><rect x=\"1\"/></svg>"
4074 );
4075 assert!(Format::Svg.supports(Gesture::SetNodeAttrs));
4076 assert!(!Format::Svg.is_authorable());
4077 assert_eq!(Target::from(Format::Gfm), Target::Markdown);
4078 let mut gfm = Document::parse_str("* a ~~b~~\n", Format::Gfm).expect("parse gfm");
4079 let back = gfm.serialize(Format::Gfm).expect("serialize");
4080 assert_eq!(String::from_utf8_lossy(&back), "* a ~~b~~\n");
4081
4082 // The render follows the row: GFM spells alignment as an attribute.
4083 let mut table = Document::parse_str("| a |\n| :-: |\n| 1 |\n", Format::Gfm).expect("parse");
4084 let html = String::from_utf8_lossy(&table.render_html().expect("render")).into_owned();
4085 assert!(html.contains("align=\"center\""), "got {html:?}");
4086
4087 // And the capability query answers per dialect.
4088 assert!(!Format::Commonmark.supports(Gesture::ToggleInline(InlineKind::Delete)));
4089 assert!(Format::Gfm.supports(Gesture::ToggleInline(InlineKind::Delete)));
4090 }
4091
4092 #[test]
4093 fn serialize_round_trips_and_cross_converts() {
4094 let mut doc = Document::parse_str("# hi\n", Format::Markdown).expect("parse markdown");
4095
4096 let canonical = doc.serialize(Format::Markdown).expect("serialize markdown");
4097 assert!(String::from_utf8_lossy(&canonical).contains("# hi"));
4098
4099 // Cross-format Markdown -> XML has no serializer.
4100 assert_eq!(doc.serialize(Format::Xml), Err(Error::UnsupportedFormat));
4101 }
4102
4103 #[test]
4104 fn serialize_markdown_to_djot() {
4105 let mut doc =
4106 Document::parse_str("This is *markdown*.\n", Format::Markdown).expect("parse markdown");
4107 let djot = doc.serialize(Format::Djot).expect("serialize djot");
4108 assert!(String::from_utf8_lossy(&djot).contains("_markdown_"));
4109 }
4110
4111 #[test]
4112 fn serialize_to_takes_the_output_axis() {
4113 let mut doc =
4114 Document::parse_str("This is *markdown*.\n", Format::Markdown).expect("parse markdown");
4115
4116 let djot = doc.serialize_to(Target::Djot).expect("serialize djot");
4117 assert!(String::from_utf8_lossy(&djot).contains("_markdown_"));
4118
4119 // The capability answer is the target's, not the input's: converting
4120 // INTO XML has no serializer regardless of what parsed the document.
4121 assert_eq!(doc.serialize_to(Target::Xml), Err(Error::UnsupportedFormat));
4122 }
4123
4124 #[test]
4125 fn serialize_and_serialize_to_agree() {
4126 // `serialize` is defined in terms of `serialize_to`, so the older
4127 // spelling stays exact rather than merely similar.
4128 let mut a = Document::parse_str("# hi\n", Format::Markdown).expect("parse markdown");
4129 let mut b = Document::parse_str("# hi\n", Format::Markdown).expect("parse markdown");
4130 for format in [Format::Markdown, Format::Djot, Format::Html] {
4131 assert_eq!(a.serialize(format), b.serialize_to(Target::from(format)));
4132 }
4133 }
4134
4135 #[test]
4136 fn every_format_is_a_target_that_names_it_back() {
4137 // The subset invariant the Zig `targets` table enforces, restated at
4138 // this layer: `Target::from` is total, and `as_format` round-trips it.
4139 for format in [Format::Djot, Format::Markdown, Format::Xml, Format::Html] {
4140 assert_eq!(Target::from(format).as_format(), Some(format));
4141 }
4142 }
4143
4144 #[test]
4145 fn ast_json_dumps_the_tree() {
4146 let mut doc = Document::parse_str("hello\n", Format::Djot).expect("parse djot");
4147 let json = doc.ast_json().expect("ast json");
4148 assert!(String::from_utf8_lossy(&json).contains("\"kind\": \"doc\""));
4149 }
4150
4151 #[test]
4152 fn query_finds_nodes_by_selector() {
4153 let source = "# One\n\n## Two\n";
4154 let mut doc = Document::parse_str(source, Format::Markdown).expect("parse markdown");
4155 let matches = doc.query("heading").expect("query");
4156
4157 assert_eq!(matches.len(), 2);
4158 for m in &matches {
4159 assert_eq!(m.kind, Kind::Heading);
4160 assert!(m.span.start < m.span.end);
4161 }
4162 }
4163
4164 #[test]
4165 fn query_recovers_code_spans() {
4166 let source = "prose `code` more prose\n";
4167 let mut doc = Document::parse_str(source, Format::Markdown).expect("parse markdown");
4168 let matches = doc.query("verbatim").expect("query");
4169
4170 assert_eq!(matches.len(), 1);
4171 assert_eq!(&source[matches[0].span.clone()], "`code`");
4172 }
4173
4174 #[test]
4175 fn document_span_accessors_read_by_node_id() {
4176 let source = "# hi\n\ntext\n";
4177 let mut doc = Document::parse_str(source, Format::Markdown).expect("parse markdown");
4178 let heading = doc.query("heading").expect("query").pop().expect("heading");
4179
4180 assert_eq!(
4181 doc.span(NodeId(heading.node_id)).expect("span"),
4182 heading.span
4183 );
4184 assert_eq!(
4185 doc.content_span(NodeId(heading.node_id))
4186 .expect("content span"),
4187 heading.content_span
4188 );
4189 assert_eq!(doc.span(NodeId(u32::MAX)), Err(Error::InvalidArgument));
4190 }
4191
4192 #[test]
4193 fn document_walks_its_tree_without_an_editor() {
4194 let source = "# hi\n\ntext\n";
4195 let mut doc = Document::parse_str(source, Format::Markdown).expect("parse markdown");
4196
4197 let nodes = doc.nodes().expect("nodes");
4198 assert!(nodes.len() >= 3);
4199 for (i, n) in nodes.iter().enumerate() {
4200 assert_eq!(n.id, NodeId(i as u32));
4201 }
4202
4203 let kids = doc.children(None).expect("children");
4204 assert_eq!(kids.len(), 2);
4205 assert_eq!(kids[0].kind, Kind::Heading);
4206
4207 let sub = doc.subtree(NodeId(kids[0].node_id)).expect("subtree");
4208 assert_eq!(sub[0].id, NodeId(0));
4209 assert_eq!(sub[0].parent, None);
4210 assert_eq!(sub[0].span, kids[0].span);
4211
4212 let hit = doc.node_at(2).expect("node_at").expect("a node at 2");
4213 let chain = doc.ancestors_at(2).expect("ancestors");
4214 assert_eq!(chain.last().expect("deepest").node_id, hit.node_id);
4215 assert_eq!(chain[0].kind, Kind::Doc);
4216
4217 assert_eq!(doc.subtree(NodeId(u32::MAX)), Err(Error::InvalidArgument));
4218 }
4219
4220 #[test]
4221 fn editor_document_view_reads_the_live_tree() {
4222 let mut ed = Editor::new_str("# one\n\ntwo\n", Format::Markdown).expect("editor");
4223
4224 {
4225 let mut view = ed.document().expect("view");
4226 let kids = view.children(None).expect("children");
4227 assert_eq!(kids.len(), 2);
4228 assert_eq!(kids[0].kind, Kind::Heading);
4229 assert_eq!(view.span(NodeId(kids[0].node_id)).expect("span"), 0..5);
4230 // The two the view can't serve.
4231 assert_eq!(view.render_html(), Err(Error::UnsupportedFormat));
4232 assert_eq!(
4233 view.serialize(Format::Markdown),
4234 Err(Error::UnsupportedFormat)
4235 );
4236 }
4237
4238 ed.replace("0", "# one and a half").expect("replace");
4239 let mut view = ed.document().expect("view");
4240 let kids = view.children(None).expect("children");
4241 assert_eq!(view.span(NodeId(kids[0].node_id)).expect("span"), 0..16);
4242 }
4243
4244 #[test]
4245 fn query_rejects_a_malformed_selector() {
4246 let mut doc = Document::parse_str("hi\n", Format::Markdown).expect("parse markdown");
4247 assert_eq!(doc.query("list >"), Err(Error::InvalidArgument));
4248 }
4249
4250 #[test]
4251 fn editor_edits_by_index_path() {
4252 let mut ed = Editor::new_str("<a><b>hi</b></a>", Format::Xml).expect("editor");
4253 ed.replace_content("0.0", "bye").expect("replace_content");
4254 assert_eq!(ed.source_str().expect("source"), "<a><b>bye</b></a>");
4255 }
4256
4257 #[test]
4258 fn flat_nodes_expose_element_name_and_attrs() {
4259 // A `<picture>` with a theme-switching `<source>`: the dark alternative
4260 // lives only in the `<source>`'s attributes, which the snapshot now
4261 // surfaces (both `<picture>` and `<source>` report `kind == "container"`).
4262 let src = "<picture><source media=\"(prefers-color-scheme: dark)\" srcset=\"d.svg\"><img src=\"l.svg\" alt=\"x\"></picture>\n";
4263 let mut ed = Editor::new_ext(
4264 src.as_bytes(),
4265 Format::Markdown,
4266 MarkdownExtensions {
4267 html_elements: true,
4268 ..Default::default()
4269 },
4270 )
4271 .expect("editor");
4272 let nodes = ed.nodes().expect("nodes");
4273
4274 let source = nodes
4275 .iter()
4276 .find(|n| n.name.as_deref() == Some("source"))
4277 .expect("a <source> element node");
4278 assert_eq!(
4279 source.attrs,
4280 vec![
4281 (
4282 "media".to_string(),
4283 Some("(prefers-color-scheme: dark)".to_string())
4284 ),
4285 ("srcset".to_string(), Some("d.svg".to_string())),
4286 ]
4287 );
4288
4289 // The `<img>` fallback stays an `image` node (no element name), and its
4290 // `src` is the ordinary `destination`.
4291 let img = nodes
4292 .iter()
4293 .find(|n| n.kind == Kind::Image)
4294 .expect("an image node");
4295 assert!(img.name.is_none());
4296 assert_eq!(img.destination.as_deref(), Some("l.svg"));
4297
4298 // A semantic node carries neither an element name nor attributes.
4299 let picture_kids_str = nodes.iter().find(|n| n.kind == Kind::Str);
4300 if let Some(s) = picture_kids_str {
4301 assert!(s.name.is_none() && s.attrs.is_empty());
4302 }
4303 }
4304
4305 #[test]
4306 fn definitions_finds_what_a_walk_from_the_root_cannot() {
4307 // Both definitions are resolved by label, so neither is anybody's
4308 // child: the document root's subtree contains the paragraph and
4309 // nothing else.
4310 let mut doc = Document::parse_str(
4311 "text[^1] [x][a]\n\n[^1]: note\n\n[a]: /u\n",
4312 Format::Markdown,
4313 )
4314 .expect("parse markdown");
4315
4316 let defs = doc.definitions().expect("definitions");
4317 let mut kinds: Vec<Kind> = defs.iter().map(|m| m.kind.clone()).collect();
4318 kinds.sort_by(|a, b| a.as_str().cmp(b.as_str()));
4319 assert_eq!(kinds, vec![Kind::Footnote, Kind::Reference]);
4320
4321 // Each one knows where it stands. A link reference definition used
4322 // to answer `0..0`, which left an editor unable to tell its lines
4323 // from blank ones.
4324 for d in &defs {
4325 let want = match d.kind {
4326 Kind::Footnote => 17..27,
4327 Kind::Reference => 29..36,
4328 _ => unreachable!(),
4329 };
4330 assert_eq!(d.span, want, "{} stands on its own bytes", d.kind);
4331 }
4332
4333 // None of them is reachable from the root — the property that made a
4334 // whole-arena rescan the only way to find them.
4335 let all = doc.nodes().expect("nodes");
4336 let root = all
4337 .iter()
4338 .find(|n| n.kind == Kind::Doc)
4339 .expect("a doc root");
4340 let mut reachable = vec![root.id];
4341 let mut i = 0;
4342 while i < reachable.len() {
4343 let n = &all[reachable[i].0 as usize];
4344 let mut c = n.first_child;
4345 while let Some(cid) = c {
4346 reachable.push(cid);
4347 c = all[cid.0 as usize].next_sibling;
4348 }
4349 i += 1;
4350 }
4351 for d in &defs {
4352 assert!(
4353 !reachable.contains(&NodeId(d.node_id)),
4354 "{} should be unreachable from the root",
4355 d.kind
4356 );
4357 }
4358
4359 // A document that defines nothing gets an empty vec, not an error.
4360 let mut plain = Document::parse_str("just text\n", Format::Markdown).expect("parse");
4361 assert_eq!(plain.definitions().expect("definitions"), Vec::new());
4362 }
4363
4364 #[test]
4365 fn kind_round_trips_through_its_published_name() {
4366 // `as_str` is the wire vocabulary and `from` is its inverse, so any
4367 // variant whose spelling drifts from the C ABI's fails here rather
4368 // than quietly becoming `Other`.
4369 for k in [
4370 Kind::Doc,
4371 Kind::Para,
4372 Kind::Heading,
4373 Kind::Container,
4374 Kind::TaskListItem,
4375 Kind::Superscript,
4376 Kind::FootnoteReference,
4377 Kind::ProcessingInstruction,
4378 Kind::Cdata,
4379 ] {
4380 assert_eq!(Kind::from(k.as_str()), k, "{k} did not round-trip");
4381 assert!(!k.is_unknown());
4382 }
4383 }
4384
4385 #[test]
4386 fn an_unknown_kind_name_is_carried_rather_than_lost() {
4387 // A newer library against an older binding. The node is still a node,
4388 // and a renderer that passes it through unchanged should be able to.
4389 let k = Kind::from("some_future_kind");
4390 assert!(k.is_unknown());
4391 assert_eq!(k.as_str(), "some_future_kind");
4392 assert_eq!(k, Kind::Other("some_future_kind".to_string()));
4393 }
4394
4395 #[test]
4396 fn every_kind_the_library_publishes_has_a_variant() {
4397 // Walks documents covering every corner of the vocabulary this crate
4398 // can reach from Rust and asserts nothing arrives as `Other`. If twig
4399 // adds a kind, or renames one, this fails — which is the whole reason
4400 // the enum is here instead of a `String`.
4401 let cases: &[(&str, Format, MarkdownExtensions)] = &[
4402 (
4403 "# h\n\npara *emph* **strong** `code`\n\n- a\n- b\n\n1. c\n\n> q\n\n---\n\n```zig\nx\n```\n",
4404 Format::Markdown,
4405 MarkdownExtensions {
4406 directives: false,
4407 math: false,
4408 html_elements: false,
4409 highlight: false,
4410 highlight_colors: false,
4411 },
4412 ),
4413 (
4414 "| a | b |\n| --- | --- |\n| 1 | 2 |\n\n- [ ] task\n- [x] done\n\nfoot[^1]\n\n[^1]: note\n\n[l]: /u\n\n[x][l]\n",
4415 Format::Markdown,
4416 MarkdownExtensions::default(),
4417 ),
4418 (
4419 ":::note\nbody\n:::\n\n:role[x]\n\n$a+b$ ==h== ==🔴 r==\n",
4420 Format::Markdown,
4421 MarkdownExtensions {
4422 directives: true,
4423 math: true,
4424 html_elements: false,
4425 highlight: true,
4426 highlight_colors: true,
4427 },
4428 ),
4429 (
4430 "a^b^ c~d~ {=e=} {+f+} {-g-} 'q' \"dq\"\n\n\n\n<https://e.com>\n",
4431 Format::Djot,
4432 MarkdownExtensions::default(),
4433 ),
4434 (
4435 "<!-- c --><!DOCTYPE html><video controls><p>x</p></video>",
4436 Format::Html,
4437 MarkdownExtensions::default(),
4438 ),
4439 ];
4440
4441 let mut unknown: Vec<String> = Vec::new();
4442 let mut seen: Vec<String> = Vec::new();
4443 for (src, format, ext) in cases {
4444 let mut ed = Editor::new_ext(src.as_bytes(), *format, *ext).expect("editor");
4445 for n in ed.nodes().expect("nodes") {
4446 if n.kind.is_unknown() {
4447 unknown.push(n.kind.as_str().to_string());
4448 }
4449 seen.push(n.kind.as_str().to_string());
4450 }
4451 }
4452 unknown.sort();
4453 unknown.dedup();
4454 assert!(unknown.is_empty(), "kinds with no variant: {unknown:?}");
4455
4456 // And the sweep really swept: without this the assertion above passes
4457 // just as happily on an empty walk.
4458 seen.sort();
4459 seen.dedup();
4460 assert!(
4461 seen.len() >= 30,
4462 "only {} distinct kinds reached: {seen:?}",
4463 seen.len()
4464 );
4465 }
4466
4467 #[test]
4468 fn diagnostics_report_what_a_conversion_would_lose() {
4469 // A djot superscript has no Markdown spelling. The two answers below
4470 // are for the SAME document — fidelity is a property of the
4471 // (document, target) pair, which is why it is asked per target.
4472 let mut doc = Document::parse_str("a^b^ c\n", Format::Djot).expect("parse djot");
4473
4474 let to_md = doc
4475 .diagnostics(Target::Markdown)
4476 .expect("markdown diagnostics");
4477 assert_eq!(
4478 to_md,
4479 vec![Warning {
4480 fidelity: Fidelity::Degraded,
4481 path: "0/1".to_string(),
4482 kind: Kind::Superscript,
4483 }]
4484 );
4485
4486 // Lossless to djot: an empty vec is a real answer, not a failure.
4487 assert_eq!(
4488 doc.diagnostics(Target::Djot).expect("djot diagnostics"),
4489 Vec::new()
4490 );
4491 }
4492
4493 #[test]
4494 fn diagnostics_separate_a_droppable_node_from_a_degradable_one() {
4495 // An HTML comment converted to djot leaves NOTHING behind — a
4496 // different and worse answer than "comes back as something else", and
4497 // the distinction a consumer needs to decide whether to warn or refuse.
4498 let mut doc =
4499 Document::parse_str("<p>hi</p><!-- secret -->", Format::Html).expect("parse html");
4500 let warnings = doc.diagnostics(Target::Djot).expect("djot diagnostics");
4501 let comment = warnings
4502 .iter()
4503 .find(|w| w.kind == Kind::Comment)
4504 .expect("a warning about the comment");
4505 assert_eq!(comment.fidelity, Fidelity::Dropped);
4506 }
4507
4508 #[test]
4509 fn diagnostics_report_a_surviving_nodes_lost_attributes() {
4510 // A djot paragraph carrying a class converts to Markdown as a `<div>`
4511 // around the paragraph, which Markdown's default parser reads as raw
4512 // HTML beside it: the node survives, the class is written where it is
4513 // not read back, and the warning says which of the two it is about.
4514 let mut doc = Document::parse_str("{.center}\nhello\n", Format::Djot).expect("parse djot");
4515 assert_eq!(
4516 doc.diagnostics(Target::Markdown).expect("markdown diagnostics"),
4517 vec![Warning {
4518 fidelity: Fidelity::AttrsDegraded,
4519 path: "0".to_string(),
4520 kind: Kind::Para,
4521 }]
4522 );
4523 // An emphasis carrying one has nowhere to put it.
4524 let mut em = Document::parse_str("_x_{.big}\n", Format::Djot).expect("parse djot");
4525 assert_eq!(
4526 em.diagnostics(Target::Markdown).expect("markdown diagnostics"),
4527 vec![Warning {
4528 fidelity: Fidelity::AttrsDropped,
4529 path: "0/0".to_string(),
4530 kind: Kind::Emph,
4531 }]
4532 );
4533 // HTML and AsciiDoc keep a paragraph's class: nothing to report.
4534 assert_eq!(doc.diagnostics(Target::Html).expect("html"), Vec::new());
4535 assert_eq!(doc.diagnostics(Target::Asciidoc).expect("asciidoc"), Vec::new());
4536 }
4537
4538 #[test]
4539 fn diagnostics_report_a_degraded_nodes_attributes_beside_it() {
4540 // A Word paste: `<html>` and `<body>` are sections, which Markdown
4541 // degrades and still writes a `<div>` for each one's attributes.
4542 // Three `<div>`s, three attribute warnings — two of them at a path
4543 // that also carries the node's own.
4544 let mut doc = Document::parse_str(
4545 r#"<html xmlns:o="urn:x"><body lang="EN-US"><p class="MsoNormal">Word text</p></body></html>"#,
4546 Format::Html,
4547 )
4548 .expect("parse html");
4549 let w = |fidelity, path: &str, kind| Warning {
4550 fidelity,
4551 path: path.to_string(),
4552 kind,
4553 };
4554 assert_eq!(
4555 doc.diagnostics(Target::Markdown).expect("markdown diagnostics"),
4556 vec![
4557 w(Fidelity::Degraded, "0", Kind::Section),
4558 w(Fidelity::AttrsDegraded, "0", Kind::Section),
4559 w(Fidelity::Degraded, "0/0", Kind::Section),
4560 w(Fidelity::AttrsDegraded, "0/0", Kind::Section),
4561 w(Fidelity::AttrsDegraded, "0/0/0", Kind::Para),
4562 ]
4563 );
4564 }
4565
4566 #[test]
4567 fn diagnostics_say_nothing_of_what_a_link_or_image_already_spells() {
4568 // `href`, `src` and `alt` are the node's destination and alt text,
4569 // which Markdown writes in full.
4570 for html in [r#"<p>a <a href="https://x.dev">l</a></p>"#, r#"<p><img src="u" alt="p"></p>"#] {
4571 let mut doc = Document::parse_str(html, Format::Html).expect("parse html");
4572 assert_eq!(doc.diagnostics(Target::Markdown).expect("markdown"), Vec::new());
4573 assert_eq!(doc.diagnostics(Target::Djot).expect("djot"), Vec::new());
4574 }
4575 }
4576
4577 #[test]
4578 fn diagnostics_refuse_a_target_with_no_serializer() {
4579 // "This target cannot be written" is a capability answer, not a
4580 // per-node diagnosis of every node in the document.
4581 let mut doc = Document::parse_str("# hi\n", Format::Markdown).expect("parse markdown");
4582 assert_eq!(doc.diagnostics(Target::Xml), Err(Error::UnsupportedFormat));
4583 // AsciiDoc has a serializer now, so it gets a per-node answer instead.
4584 assert!(doc.diagnostics(Target::Asciidoc).is_ok());
4585 }
4586
4587 #[test]
4588 fn diagnostics_flag_a_header_less_table_and_leave_a_headed_one_alone() {
4589 // The instance-level answer, and the one a consumer cannot reach by
4590 // looking at kinds: both documents contain a `table`, and only one of
4591 // them costs anything to convert. GFM's delimiter row is mandatory, so
4592 // the header-less table gets an empty header synthesized above it.
4593 let mut headed = Document::parse_str(
4594 "<table><tr><th>H</th></tr><tr><td>a</td></tr></table>",
4595 Format::Html,
4596 )
4597 .expect("parse headed table");
4598 assert!(
4599 headed
4600 .diagnostics(Target::Markdown)
4601 .expect("diagnostics")
4602 .iter()
4603 .all(|w| w.kind != Kind::Table)
4604 );
4605
4606 let mut headless = Document::parse_str("<table><tr><td>a</td></tr></table>", Format::Html)
4607 .expect("parse header-less table");
4608 let table_warning = headless
4609 .diagnostics(Target::Markdown)
4610 .expect("diagnostics")
4611 .into_iter()
4612 .find(|w| w.kind == Kind::Table)
4613 .expect("a warning about the table");
4614 assert_eq!(table_warning.fidelity, Fidelity::Degraded);
4615 }
4616
4617 #[test]
4618 fn container_origin_separates_a_div_from_a_div() {
4619 // The collision this field exists for. These two documents produce
4620 // container nodes that agree on `kind`, on `name` AND on
4621 // `directive_form` — so a consumer holding one of them could not say
4622 // which syntax the author wrote without re-reading the source bytes.
4623 let mut html =
4624 Editor::new("<div>hi</div>\n".as_bytes(), Format::Html).expect("html editor");
4625 let mut md = Editor::new_ext(
4626 ":::div\nhi\n:::\n".as_bytes(),
4627 Format::Markdown,
4628 MarkdownExtensions {
4629 directives: true,
4630 ..Default::default()
4631 },
4632 )
4633 .expect("markdown editor");
4634
4635 let html_nodes = html.nodes().expect("html nodes");
4636 let md_nodes = md.nodes().expect("markdown nodes");
4637 let tag = html_nodes
4638 .iter()
4639 .find(|n| n.name.as_deref() == Some("div"))
4640 .expect("a <div> container");
4641 let directive = md_nodes
4642 .iter()
4643 .find(|n| n.name.as_deref() == Some("div"))
4644 .expect("a :::div container");
4645
4646 // Indistinguishable on every field that predates `origin`.
4647 assert_eq!(tag.kind, directive.kind);
4648 assert_eq!(tag.name, directive.name);
4649 assert_eq!(tag.directive_form, directive.directive_form);
4650 assert_eq!(tag.directive_form, Some(DirectiveForm::Container));
4651
4652 // And decidable now.
4653 assert_eq!(tag.origin, Some(ContainerOrigin::Element));
4654 assert_eq!(directive.origin, Some(ContainerOrigin::Directive));
4655 }
4656
4657 #[test]
4658 fn a_container_whose_body_is_text_says_so_in_text() {
4659 // `<script>` and `<span>` used to be the same shape — a container over
4660 // one `str` — and differ only in name, so an editor could not tell a
4661 // JavaScript body from prose without its own tag list. The tokenizer
4662 // knows: a raw-text or rcdata body is the node's `text`, and there
4663 // are no children to step into.
4664 let mut html = Editor::new(
4665 "<script>a < b</script><title>a & b</title><span>a & b</span>\n".as_bytes(),
4666 Format::Html,
4667 )
4668 .expect("html editor");
4669 let nodes = html.nodes().expect("html nodes");
4670 let by_name = |name: &str| {
4671 nodes
4672 .iter()
4673 .find(|n| n.name.as_deref() == Some(name))
4674 .unwrap_or_else(|| panic!("a <{name}> container"))
4675 };
4676 let script = by_name("script");
4677 assert_eq!(script.kind, Kind::Container);
4678 assert_eq!(script.text.as_deref(), Some("a < b"));
4679 assert_eq!(script.first_child, None);
4680 // rcdata is decoded, like any text.
4681 assert_eq!(by_name("title").text.as_deref(), Some("a & b"));
4682 // A markup body is children, and `text` stays `None`.
4683 let span = by_name("span");
4684 assert_eq!(span.text, None);
4685 assert!(span.first_child.is_some());
4686 }
4687
4688 /// Parse `src` as both authorable formats and run `check` over each — the
4689 /// shape every test below wants, because the point of these two APIs is
4690 /// that a consumer cannot tell which parser produced the tree.
4691 fn for_both_formats(src: &str, check: impl Fn(&mut Document, Format)) {
4692 for format in [Format::Markdown, Format::Djot] {
4693 let mut doc = Document::parse(src.as_bytes(), format).expect("parse");
4694 check(&mut doc, format);
4695 }
4696 }
4697
4698 #[test]
4699 fn marker_span_is_what_a_rich_view_hides() {
4700 for_both_formats("> - [x] done\n", |doc, format| {
4701 let nodes = doc.nodes().expect("nodes");
4702 let quote = nodes
4703 .iter()
4704 .find(|n| n.kind == Kind::BlockQuote)
4705 .expect("a block quote");
4706 let item = nodes
4707 .iter()
4708 .find(|n| n.kind == Kind::TaskListItem)
4709 .expect("a task item");
4710
4711 // The quote's `> ` and the item's `- [x] ` — the item's marker
4712 // takes its checkbox with it, because the rendered view draws a
4713 // checkbox in PLACE of those bytes rather than beside them.
4714 assert_eq!(quote.marker_span, Some(0..2), "{format:?}");
4715 assert_eq!(item.marker_span, Some(2..8), "{format:?}");
4716
4717 // Not derivable from the other two spans: a marker-prefixed
4718 // container reports its whole extent as its interior, so the
4719 // subtraction a caller might reach for yields nothing.
4720 assert_eq!(quote.content_span, Some(quote.span.clone()), "{format:?}");
4721
4722 // A paragraph has no marker of its own; only its ancestors do.
4723 let para = nodes
4724 .iter()
4725 .find(|n| n.kind == Kind::Para)
4726 .expect("a paragraph");
4727 assert_eq!(para.marker_span, None, "{format:?}");
4728 });
4729 }
4730
4731 #[test]
4732 fn attrs_span_locates_the_attribute_block_a_heuristic_had_to_guess_at() {
4733 // A Djot attribute line sits on its OWN line above the block it
4734 // attaches to, so a consumer dropping the block has to drop that line
4735 // too. Without this the extent was guessed at by scanning for `{`,
4736 // which strands the line as a published paragraph when the guess
4737 // misses — and the line names the audience.
4738 let src = "{.vis .family}\nheld back\n\nplain\n";
4739 let mut doc = Document::parse(src.as_bytes(), Format::Djot).expect("parse");
4740 let nodes = doc.nodes().expect("nodes");
4741 let paras: Vec<&FlatNode> = nodes.iter().filter(|n| n.kind == Kind::Para).collect();
4742 assert_eq!(paras.len(), 2);
4743
4744 let span = doc
4745 .attrs_span(paras[0].id)
4746 .expect("attrs span")
4747 .expect("the attributed paragraph has one");
4748 assert_eq!(&src[span.clone()], "{.vis .family}");
4749 // The block's own span starts AFTER the attribute line, which is why
4750 // dropping the block alone leaves the line behind.
4751 assert!(span.end <= paras[0].span.start);
4752
4753 // `None` is a real answer, not a failure: the second paragraph is
4754 // unattributed.
4755 assert_eq!(doc.attrs_span(paras[1].id).expect("attrs span"), None);
4756 }
4757
4758 #[test]
4759 fn line_prefix_assembles_every_marker_on_the_line() {
4760 for_both_formats("> - [x] done\n", |doc, format| {
4761 // Four nodes' worth of hidden width as one range, which is what a
4762 // caret stepping over it needs — not a chain to stitch together.
4763 assert_eq!(doc.line_prefix(9).expect("prefix"), Some(0..8), "{format:?}");
4764 });
4765 }
4766
4767 #[test]
4768 fn line_prefix_is_none_on_a_continuation_line() {
4769 // Line two continues the quote but OPENS nothing. `None` is the honest
4770 // answer: what a continuation line repeats is a different question with
4771 // a different answer, and guessing it from marker spans is how an
4772 // editor ends up restructuring a document that never had the shape it
4773 // inferred.
4774 for_both_formats("> c\n> d\n", |doc, format| {
4775 assert_eq!(doc.line_prefix(2).expect("prefix"), Some(0..2), "{format:?}");
4776 assert_eq!(doc.line_prefix(6).expect("prefix"), None, "{format:?}");
4777 });
4778 }
4779
4780 #[test]
4781 fn a_caret_at_a_blocks_end_is_in_that_block_in_both_formats() {
4782 // The divergence this API exists for. Djot ends a paragraph's span
4783 // AFTER its newline and Markdown BEFORE it, so under half-open
4784 // containment offset 1 — the caret you get by pressing End on line one,
4785 // the commonest caret position there is — resolved to the paragraph
4786 // through Djot and to the root through Markdown.
4787 for_both_formats("a\n\nb\n", |doc, format| {
4788 for offset in [0usize, 1, 3, 4] {
4789 let hit = doc
4790 .node_at_caret(offset)
4791 .expect("caret hit")
4792 .expect("some node");
4793 assert_eq!(hit.kind, Kind::Str, "{format:?} at {offset}");
4794 }
4795 // The blank line between the two blocks belongs to neither, and
4796 // neither does the empty line after the final newline.
4797 for offset in [2usize, 5] {
4798 let hit = doc
4799 .node_at_caret(offset)
4800 .expect("caret hit")
4801 .expect("some node");
4802 assert_eq!(hit.kind, Kind::Doc, "{format:?} at {offset}");
4803 }
4804 });
4805 }
4806
4807 #[test]
4808 fn the_caret_chain_ends_at_the_node_the_scalar_call_returns() {
4809 for_both_formats("- a\n", |doc, format| {
4810 let hit = doc.node_at_caret(3).expect("hit").expect("some node");
4811 let chain = doc.ancestors_at_caret(3).expect("chain");
4812 assert_eq!(chain.last().map(|m| m.node_id), Some(hit.node_id), "{format:?}");
4813 // And the chain passes through the item, which is what a gesture
4814 // scoped to "the block I'm in" needs at a caret sitting at its end.
4815 assert!(
4816 chain.iter().any(|m| m.kind == Kind::ListItem),
4817 "{format:?}: chain should reach the list item"
4818 );
4819 });
4820 }
4821
4822 #[test]
4823 fn continuation_prefix_repeats_a_quote_and_indents_past_an_item() {
4824 for_both_formats("> - a\n", |doc, format| {
4825 // The bytes already on the line, and the bytes a continuation would
4826 // need. They differ exactly where an editor gets it wrong by hand:
4827 // the item's `- ` is PRESENT and must not be repeated, or the
4828 // continuation opens a second item instead of continuing the first.
4829 assert_eq!(doc.line_prefix(4).expect("prefix"), Some(0..4), "{format:?}");
4830 let cont = doc.continuation_prefix(4).expect("continuation");
4831 assert_eq!(cont.text, "> ", "{format:?}");
4832 assert_eq!(cont.columns, 4, "{format:?}");
4833 });
4834 }
4835
4836 #[test]
4837 fn continuation_prefix_answers_on_a_line_that_opens_nothing() {
4838 // The case `line_prefix` declines. Each ancestor answers from its own
4839 // opening line, so the quote's marker is still found on line one.
4840 for_both_formats("> c\n> d\n", |doc, format| {
4841 assert_eq!(doc.line_prefix(6).expect("prefix"), None, "{format:?}");
4842 assert_eq!(
4843 doc.continuation_prefix(6).expect("continuation").text,
4844 "> ",
4845 "{format:?}"
4846 );
4847 });
4848 }
4849
4850 #[test]
4851 fn continuation_prefix_takes_an_ordered_markers_own_width() {
4852 // `10. ` is four columns where `1. ` is three. A fixed indent is the
4853 // assumption that makes Tab wrong on the tenth item.
4854 for_both_formats("10. x\n", |doc, format| {
4855 assert_eq!(
4856 doc.continuation_prefix(4).expect("continuation").columns,
4857 4,
4858 "{format:?}"
4859 );
4860 });
4861 for_both_formats("1. x\n", |doc, format| {
4862 assert_eq!(
4863 doc.continuation_prefix(3).expect("continuation").columns,
4864 3,
4865 "{format:?}"
4866 );
4867 });
4868 }
4869
4870 #[test]
4871 fn a_blank_line_keeps_a_quote_alive_and_drops_an_items_indent() {
4872 for_both_formats("> - a\n", |doc, format| {
4873 let blank = doc.blank_line_prefix(4).expect("blank");
4874 // `>` and not `> `: the space after the marker is content indent,
4875 // and a blank line has no content.
4876 assert_eq!(blank.text, ">", "{format:?}");
4877 assert_eq!(blank.columns, 1, "{format:?}");
4878 });
4879 // Inside a list alone there is nothing to keep alive, so a blank line
4880 // carries nothing at all.
4881 for_both_formats("- a\n", |doc, format| {
4882 assert_eq!(doc.blank_line_prefix(3).expect("blank").text, "", "{format:?}");
4883 });
4884 }
4885
4886 #[test]
4887 fn a_prefix_column_count_is_not_its_byte_length() {
4888 // A tab in a marker advances to a tab stop, so the two diverge — which
4889 // is why `columns` is carried rather than left to the caller to infer.
4890 let mut doc = Document::parse("- x
4891".as_bytes(), Format::Markdown).expect("parse");
4892 let cont = doc.continuation_prefix(2).expect("continuation");
4893 assert_eq!(cont.columns, 4);
4894 }
4895
4896 #[test]
4897 fn set_block_opens_a_heading_on_a_blank_line() {
4898 for format in [Format::Markdown, Format::Djot] {
4899 let mut ed = Editor::new("a\n\n".as_bytes(), format).expect("editor");
4900 ed.set_block(3, BlockKind::Heading(2)).expect("set_block");
4901 assert_eq!(ed.source().expect("source"), b"a\n\n## ", "{format:?}");
4902 // The reparse is the assertion that matters, not the bytes: Djot
4903 // does not let a heading interrupt a paragraph, so a marker written
4904 // without the separating blank would come back as literal text.
4905 let nodes = ed.nodes().expect("nodes");
4906 assert!(
4907 nodes.iter().any(|n| n.kind == Kind::Heading),
4908 "{format:?}: should have parsed a heading"
4909 );
4910 }
4911 }
4912
4913 #[test]
4914 fn set_block_refuses_a_blank_line_inside_a_code_block() {
4915 // `innermostBlock` reports nothing here exactly as it does between
4916 // blocks; only the line's owner tells them apart. Writing `# ` in would
4917 // add no heading and corrupt the listing.
4918 for format in [Format::Markdown, Format::Djot] {
4919 let src = "```\nx\n\ny\n```\n";
4920 let mut ed = Editor::new(src.as_bytes(), format).expect("editor");
4921 let blank = src.find("\n\n").expect("a blank line") + 1;
4922 assert!(
4923 matches!(
4924 ed.set_block(blank, BlockKind::Heading(1)),
4925 Err(Error::NotEditable)
4926 ),
4927 "{format:?}"
4928 );
4929 assert_eq!(ed.source().expect("source"), src.as_bytes(), "{format:?}");
4930 }
4931 }
4932
4933 #[test]
4934 fn task_items_report_their_checkbox_state() {
4935 // Twig would WRITE a checkbox and not read one back, so a consumer
4936 // rendering a clickable box re-derived the state by scanning for `[x]`.
4937 // A capital `[X]` is checked too, which that scan misses.
4938 for_both_formats("- [ ] a\n- [x] b\n- [X] c\n- d\n", |doc, format| {
4939 let nodes = doc.nodes().expect("nodes");
4940 let states: Vec<Option<bool>> = nodes
4941 .iter()
4942 .filter(|n| matches!(n.kind, Kind::TaskListItem | Kind::ListItem))
4943 .map(|n| n.checked)
4944 .collect();
4945 assert_eq!(
4946 states,
4947 vec![Some(false), Some(true), Some(true), None],
4948 "{format:?}"
4949 );
4950
4951 // `None` is not `Some(false)`: a consumer treating "not a task
4952 // item" as unchecked draws an empty box beside every paragraph.
4953 for n in nodes.iter().filter(|n| n.kind == Kind::Para) {
4954 assert_eq!(n.checked, None, "{format:?}");
4955 }
4956 });
4957 }
4958
4959 #[test]
4960 fn an_editor_reaches_the_caret_reads_through_its_document_view() {
4961 // The path an editing host actually takes. These reads are questions
4962 // about a TREE, not about an editing session, so they live on the
4963 // document surface and an editor borrows it — no `twig_editor_*` alias
4964 // to keep in step. See DESIGN.md, "The reads are not editor-specific."
4965 let mut ed = Editor::new("- a\n".as_bytes(), Format::Markdown).expect("editor");
4966 let mut view = ed.document().expect("document view");
4967
4968 assert_eq!(view.line_prefix(3).expect("prefix"), Some(0..2));
4969 let hit = view.node_at_caret(3).expect("hit").expect("some node");
4970 assert_eq!(hit.kind, Kind::Str);
4971 }
4972
4973 #[test]
4974 fn container_origin_is_none_for_non_containers() {
4975 // The field is a container's, so everything else reports `None` rather
4976 // than a default that would read as a real answer.
4977 let mut ed = Editor::new("# hi\n\npara\n".as_bytes(), Format::Markdown).expect("editor");
4978 for n in ed.nodes().expect("nodes") {
4979 assert_eq!(n.origin, None, "{} should carry no origin", n.kind);
4980 }
4981 }
4982
4983 #[test]
4984 fn flat_nodes_expose_directive_name_and_form() {
4985 // All three surface forms report `kind == "container"`, so the snapshot
4986 // has to carry both halves of a directive's identity: which type it is
4987 // (`name`) and how it was written (`directive_form`). Without them a
4988 // renderer can't tell an `::embed` from a `::toc`, nor an inline span
4989 // from a standalone block.
4990 let src = ":::note{.warning}\nBody\n:::\n\n::embed{src=\"demo.html\"}\n\nSee :abbr[HTML]{title=\"HyperText\"} inline.\n";
4991 let mut ed = Editor::new_ext(
4992 src.as_bytes(),
4993 Format::Markdown,
4994 MarkdownExtensions {
4995 directives: true,
4996 ..Default::default()
4997 },
4998 )
4999 .expect("editor");
5000 let nodes = ed.nodes().expect("nodes");
5001
5002 let forms: Vec<(Option<&str>, Option<DirectiveForm>)> = nodes
5003 .iter()
5004 .filter(|n| n.kind == Kind::Container)
5005 .map(|n| (n.name.as_deref(), n.directive_form))
5006 .collect();
5007 assert_eq!(
5008 forms,
5009 vec![
5010 (Some("note"), Some(DirectiveForm::Container)),
5011 (Some("embed"), Some(DirectiveForm::Leaf)),
5012 (Some("abbr"), Some(DirectiveForm::Text)),
5013 ]
5014 );
5015
5016 // The attributes still ride the ordinary side-table, and a non-directive
5017 // reports no form at all.
5018 let embed = nodes
5019 .iter()
5020 .find(|n| n.name.as_deref() == Some("embed"))
5021 .expect("embed");
5022 assert_eq!(
5023 embed.attrs,
5024 vec![("src".to_string(), Some("demo.html".to_string()))]
5025 );
5026 let para = nodes.iter().find(|n| n.kind == Kind::Para).expect("a para");
5027 assert!(para.directive_form.is_none() && para.name.is_none());
5028 }
5029
5030 #[test]
5031 fn editor_insert_child_and_delete() {
5032 let mut ed = Editor::new_str("<r><a/><c/></r>", Format::Xml).expect("editor");
5033 ed.insert_child("0", 1, "<b/>").expect("insert_child");
5034 assert_eq!(ed.source_str().expect("source"), "<r><a/><b/><c/></r>");
5035 ed.delete("0.1").expect("delete");
5036 assert_eq!(ed.source_str().expect("source"), "<r><a/><c/></r>");
5037 }
5038
5039 #[test]
5040 fn editor_edits_by_selector() {
5041 let mut ed = Editor::new_str("# One\n\n## Two\n", Format::Markdown).expect("editor");
5042 ed.replace("heading(\"Two\")", "## Renamed")
5043 .expect("replace");
5044 assert_eq!(ed.source_str().expect("source"), "# One\n\n## Renamed\n");
5045 }
5046
5047 #[test]
5048 fn editor_locator_errors_are_distinct() {
5049 let mut ed = Editor::new_str("<r><a/><a/></r>", Format::Xml).expect("editor");
5050 assert_eq!(ed.replace("0.9", "x"), Err(Error::NotFound));
5051 assert_eq!(ed.replace("element", "x"), Err(Error::Ambiguous));
5052 assert_eq!(ed.replace("element(", "x"), Err(Error::InvalidArgument));
5053 // Untouched by the failed edits.
5054 assert_eq!(ed.source_str().expect("source"), "<r><a/><a/></r>");
5055 }
5056
5057 #[test]
5058 fn editor_reparse_break_rolls_back() {
5059 let mut ed = Editor::new_str("<a>ok</a>", Format::Xml).expect("editor");
5060 assert_eq!(ed.replace_content("0", "<b>"), Err(Error::EditConflict));
5061 assert_eq!(ed.source_str().expect("source"), "<a>ok</a>");
5062 }
5063
5064 #[test]
5065 fn editor_leaf_content_is_not_editable() {
5066 let mut ed = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5067 assert_eq!(ed.replace_content("0.0", "x"), Err(Error::NotEditable));
5068 }
5069
5070 #[test]
5071 fn editor_query_reflects_current_tree() {
5072 let mut ed = Editor::new_str("<r><a/></r>", Format::Xml).expect("editor");
5073 ed.insert_child("0", 1, "<b/>").expect("insert_child");
5074 // Root <r> plus <a/> and <b/>.
5075 assert_eq!(ed.query("element").expect("query").len(), 3);
5076 let json = ed.ast_json().expect("ast_json");
5077 assert!(String::from_utf8_lossy(&json).contains("\"kind\": \"doc\""));
5078 }
5079
5080 // ── offset-addressed editing & read-back (P0–P3) ────────────────────────
5081
5082 #[test]
5083 fn editor_edit_range_types_backspaces_and_reports_change() {
5084 let mut ed = Editor::new_str("ab\n", Format::Markdown).expect("editor");
5085
5086 // Type "X" at offset 1 (a zero-width splice = an insertion).
5087 let c = ed.edit_range(1, 1, "X").expect("edit_range insert");
5088 assert_eq!(ed.source_str().unwrap(), "aXb\n");
5089 assert_eq!(c.old, 1..1);
5090 assert_eq!(c.new, 1..2);
5091 assert_eq!(c.delta(), 1);
5092
5093 // Backspace it (delete the "X").
5094 let c2 = ed.edit_range(1, 2, "").expect("edit_range delete");
5095 assert_eq!(ed.source_str().unwrap(), "ab\n");
5096 assert_eq!(c2.old, 1..2);
5097 assert_eq!(c2.new, 1..1);
5098 assert_eq!(c2.delta(), -1);
5099 }
5100
5101 #[test]
5102 fn editor_edit_range_rejects_bad_ranges() {
5103 let mut ed = Editor::new_str("hi\n", Format::Markdown).expect("editor");
5104 assert_eq!(ed.edit_range(0, 99, "x"), Err(Error::InvalidArgument)); // end past len
5105 assert_eq!(ed.edit_range(2, 1, "x"), Err(Error::InvalidArgument)); // start > end
5106 assert_eq!(ed.source_str().unwrap(), "hi\n"); // untouched
5107 }
5108
5109 #[test]
5110 fn editor_last_change_reports_locator_ops_too() {
5111 let mut ed = Editor::new_str("# One\n\n## Two\n", Format::Markdown).expect("editor");
5112 assert_eq!(ed.last_change(), None); // nothing edited yet
5113
5114 ed.replace("heading(\"Two\")", "## Renamed")
5115 .expect("replace");
5116 assert_eq!(ed.source_str().unwrap(), "# One\n\n## Renamed\n");
5117 let c = ed.last_change().expect("a change was recorded");
5118 // "## Two" occupied [7,13); "## Renamed" (10 bytes) now occupies [7,17).
5119 assert_eq!(c.old, 7..13);
5120 assert_eq!(c.new, 7..17);
5121 }
5122
5123 #[test]
5124 fn editor_nodes_is_a_walkable_flat_tree() {
5125 let mut ed = Editor::new_str("# Hi\n\ntext\n", Format::Markdown).expect("editor");
5126 let nodes = ed.nodes().expect("nodes");
5127 assert!(!nodes.is_empty());
5128
5129 // Dense, index-aligned ids.
5130 for (i, n) in nodes.iter().enumerate() {
5131 assert_eq!(n.id, NodeId(i as u32));
5132 }
5133 // Exactly one root (no parent), and it's the doc.
5134 let roots: Vec<_> = nodes.iter().filter(|n| n.parent.is_none()).collect();
5135 assert_eq!(roots.len(), 1);
5136 assert_eq!(roots[0].kind, Kind::Doc);
5137
5138 // The heading carries its level; the "Hi" text is reachable as a payload.
5139 let heading = nodes
5140 .iter()
5141 .find(|n| n.kind == Kind::Heading)
5142 .expect("a heading");
5143 assert_eq!(heading.level, Some(1));
5144 assert!(nodes.iter().any(|n| n.text.as_deref() == Some("Hi")));
5145
5146 // A kind with no row/cell payload reports neither.
5147 assert_eq!(heading.head, None);
5148 assert_eq!(heading.alignment, None);
5149
5150 // Every non-root node's parent links back to a node that lists it as a
5151 // child (via first_child/next_sibling).
5152 for n in nodes.iter().filter(|n| n.parent.is_some()) {
5153 let p = &nodes[n.parent.unwrap().0 as usize];
5154 let mut kid = p.first_child;
5155 let mut seen = false;
5156 while let Some(NodeId(k)) = kid {
5157 if k == n.id.0 {
5158 seen = true;
5159 break;
5160 }
5161 kid = nodes[k as usize].next_sibling;
5162 }
5163 assert!(
5164 seen,
5165 "node {:?} not found among its parent's children",
5166 n.id
5167 );
5168 }
5169 }
5170
5171 #[test]
5172 fn editor_child_spans_and_subtree_agree_with_nodes() {
5173 let src = "# Title\n\nHello **world** and more.\n\n- one\n- two\n";
5174 let mut ed = Editor::new_str(src, Format::Markdown).expect("editor");
5175 let all = ed.nodes().expect("nodes");
5176 let doc = all.iter().find(|n| n.kind == Kind::Doc).expect("doc");
5177
5178 // child_spans(None) == the doc root's children, same ids/kinds/spans and
5179 // in the same order.
5180 let top = ed.child_spans(None).expect("child_spans");
5181 let mut want = Vec::new();
5182 let mut c = doc.first_child;
5183 while let Some(id) = c {
5184 want.push(id);
5185 c = all[id.0 as usize].next_sibling;
5186 }
5187 assert_eq!(top.len(), want.len(), "top-level count");
5188 for (m, id) in top.iter().zip(&want) {
5189 assert_eq!(m.node_id, id.0, "child id");
5190 assert_eq!(m.kind, all[id.0 as usize].kind, "child kind");
5191 assert_eq!(m.span, all[id.0 as usize].span, "child span");
5192 }
5193 // The span addresses the block as written (absolute offsets).
5194 assert!(
5195 src[top[0].span.clone()].starts_with('#'),
5196 "first block is the heading"
5197 );
5198
5199 // child_spans works below the top level too.
5200 let list = top
5201 .iter()
5202 .find(|m| {
5203 matches!(
5204 m.kind,
5205 Kind::BulletList | Kind::OrderedList | Kind::TaskList
5206 )
5207 })
5208 .expect("a list");
5209 let items = ed.child_spans(Some(NodeId(list.node_id))).expect("items");
5210 assert_eq!(items.len(), 2);
5211 assert!(
5212 items.iter().all(|m| m.kind == Kind::ListItem),
5213 "items: {items:?}"
5214 );
5215
5216 // subtree(para) is self-contained, local-indexed, and spans stay absolute.
5217 let para = top
5218 .iter()
5219 .find(|m| m.kind == Kind::Para)
5220 .expect("a para")
5221 .node_id;
5222 let sub = ed.subtree(NodeId(para)).expect("subtree");
5223 assert_eq!(sub[0].id, NodeId(0), "root is local id 0");
5224 assert_eq!(sub[0].parent, None, "root has no parent inside the subtree");
5225 assert_eq!(sub[0].next_sibling, None, "root's sibling is severed");
5226 assert_eq!(sub[0].kind, Kind::Para);
5227 for (i, n) in sub.iter().enumerate() {
5228 assert_eq!(n.id, NodeId(i as u32), "dense local ids");
5229 for link in [n.parent, n.first_child, n.next_sibling]
5230 .into_iter()
5231 .flatten()
5232 {
5233 assert!(
5234 (link.0 as usize) < sub.len(),
5235 "link {link:?} escapes the subtree"
5236 );
5237 }
5238 }
5239 assert!(
5240 src[sub[0].span.clone()].starts_with("Hello"),
5241 "absolute span: {:?}",
5242 &src[sub[0].span.clone()]
5243 );
5244
5245 // Same multiset of node kinds as the paragraph's arena subtree.
5246 fn arena_kinds(all: &[FlatNode], root: NodeId) -> Vec<Kind> {
5247 let mut out = Vec::new();
5248 let mut stack = vec![root];
5249 while let Some(id) = stack.pop() {
5250 let n = &all[id.0 as usize];
5251 out.push(n.kind.clone());
5252 let mut c = n.first_child;
5253 while let Some(cid) = c {
5254 stack.push(cid);
5255 c = all[cid.0 as usize].next_sibling;
5256 }
5257 }
5258 out
5259 }
5260 let mut want_kinds = arena_kinds(&all, NodeId(para));
5261 let mut got_kinds: Vec<Kind> = sub.iter().map(|n| n.kind.clone()).collect();
5262 // Sorted by NAME: `Kind` is deliberately not `Ord` (there is no
5263 // meaningful order over a vocabulary), and this only needs a canonical
5264 // one to compare two multisets.
5265 want_kinds.sort_by(|a, b| a.as_str().cmp(b.as_str()));
5266 got_kinds.sort_by(|a, b| a.as_str().cmp(b.as_str()));
5267 assert_eq!(got_kinds, want_kinds, "subtree kinds match the arena");
5268
5269 // Out-of-range id is rejected.
5270 assert!(matches!(
5271 ed.subtree(NodeId(9999)),
5272 Err(Error::InvalidArgument)
5273 ));
5274 }
5275
5276 #[test]
5277 fn flat_nodes_carry_table_head_and_alignment() {
5278 // The delimiter row (`|:-----|----:|`) is consumed by the parser and has
5279 // no node of its own, so `alignment` on the cells is the only way a
5280 // consumer can recover the column alignment from a snapshot.
5281 let src = "| Name | Qty |\n|:-----|----:|\n| Pear | 3 |\n";
5282 let mut ed = Editor::new_str(src, Format::Markdown).expect("editor");
5283 let nodes = ed.nodes().expect("nodes");
5284
5285 let rows: Vec<_> = nodes.iter().filter(|n| n.kind == Kind::Row).collect();
5286 assert_eq!(rows.len(), 2, "a header row and one body row");
5287 assert_eq!(rows[0].head, Some(true), "first row is the header");
5288 assert_eq!(rows[1].head, Some(false), "second row is a body row");
5289
5290 let cells: Vec<_> = nodes.iter().filter(|n| n.kind == Kind::Cell).collect();
5291 assert_eq!(cells.len(), 4);
5292 // Alignment comes from the delimiter row and applies down the column.
5293 assert_eq!(cells[0].alignment, Some(Alignment::Left));
5294 assert_eq!(cells[1].alignment, Some(Alignment::Right));
5295 assert_eq!(cells[2].alignment, Some(Alignment::Left));
5296 assert_eq!(cells[3].alignment, Some(Alignment::Right));
5297 // Header cells are flagged too, not just their row.
5298 assert_eq!(cells[0].head, Some(true));
5299 assert_eq!(cells[2].head, Some(false));
5300
5301 // A table with no alignment spelled out reports Default — a real value,
5302 // distinct from the None a non-cell reports.
5303 let mut plain =
5304 Editor::new_str("| A |\n| --- |\n| b |\n", Format::Markdown).expect("editor");
5305 let pnodes = plain.nodes().expect("nodes");
5306 let pcell = pnodes
5307 .iter()
5308 .find(|n| n.kind == Kind::Cell)
5309 .expect("a cell");
5310 assert_eq!(pcell.alignment, Some(Alignment::Default));
5311 }
5312
5313 #[test]
5314 fn cell_extent_reports_merged_cells_and_nothing_else() {
5315 let src = "<table><tr><td colspan=\"2\" rowspan=\"3\">a</td><td>b</td></tr></table>";
5316 let mut doc = Document::parse_str(src, Format::Html).expect("parse");
5317 let cells: Vec<NodeId> = doc
5318 .nodes()
5319 .expect("nodes")
5320 .iter()
5321 .filter(|n| n.kind == Kind::Cell)
5322 .map(|n| n.id)
5323 .collect();
5324 assert_eq!(cells.len(), 2);
5325 assert_eq!(doc.cell_extent(cells[0]).expect("extent"), Some((2, 3)));
5326 // A plain cell is one square — 1, never 0.
5327 assert_eq!(doc.cell_extent(cells[1]).expect("extent"), Some((1, 1)));
5328
5329 // A pipe table cannot express a span at all, so every cell is (1, 1).
5330 let mut pipe =
5331 Document::parse_str("| a |\n| --- |\n| b |\n", Format::Markdown).expect("parse");
5332 let pipe_cell = pipe
5333 .nodes()
5334 .expect("nodes")
5335 .iter()
5336 .find(|n| n.kind == Kind::Cell)
5337 .expect("a cell")
5338 .id;
5339 assert_eq!(pipe.cell_extent(pipe_cell).expect("extent"), Some((1, 1)));
5340
5341 // Not a cell at all: None, distinct from any extent.
5342 let root = NodeId(0);
5343 assert_eq!(pipe.cell_extent(root).expect("extent"), None);
5344 }
5345
5346 #[test]
5347 fn builder_add_cell_spanning_renders_colspan_and_rowspan() {
5348 let mut b = Builder::new().expect("builder");
5349 let wide_text = b.add_text(TextKind::Str, "wide").expect("str");
5350 let wide = b
5351 .add_cell_spanning(false, Alignment::Default, 2, 3)
5352 .expect("cell");
5353 b.set_children(wide, &[wide_text]).expect("children");
5354 let plain_text = b.add_text(TextKind::Str, "one").expect("str");
5355 let plain = b.add_cell(false, Alignment::Default).expect("cell");
5356 b.set_children(plain, &[plain_text]).expect("children");
5357 let row = b.add_row(false).expect("row");
5358 b.set_children(row, &[wide, plain]).expect("children");
5359 let table = b.add(VoidKind::Table).expect("table");
5360 b.set_children(table, &[row]).expect("children");
5361
5362 let html = String::from_utf8(b.render_html(table).expect("html")).expect("utf-8");
5363 assert!(
5364 html.contains("<td colspan=\"2\" rowspan=\"3\">wide</td>"),
5365 "{html}"
5366 );
5367 // `add_cell` is the one-square case: the default extent writes nothing.
5368 assert!(html.contains("<td>one</td>"), "{html}");
5369
5370 // A zero extent is no cell anyone can lay out.
5371 assert!(matches!(
5372 b.add_cell_spanning(false, Alignment::Default, 0, 1),
5373 Err(Error::InvalidArgument)
5374 ));
5375 }
5376
5377 #[test]
5378 fn editor_node_at_and_ancestors_hit_test_offsets() {
5379 let mut ed = Editor::new_str("# Hi\n\ntext\n", Format::Markdown).expect("editor");
5380
5381 // Offset 2 is the "H" of the heading "# Hi" [0,4).
5382 let m = ed
5383 .node_at(2)
5384 .expect("node_at")
5385 .expect("a node covers offset 2");
5386 assert!(m.span.contains(&2));
5387
5388 // The ancestor chain is root-first and ends at the deepest (== node_at).
5389 let chain = ed.ancestors_at(2).expect("ancestors_at");
5390 assert!(!chain.is_empty());
5391 assert_eq!(chain[0].kind, Kind::Doc);
5392 assert_eq!(chain.last().unwrap().node_id, m.node_id);
5393
5394 // An out-of-range offset is an error; a gap covers nothing deeper than doc.
5395 assert_eq!(ed.node_at(999), Err(Error::InvalidArgument));
5396 }
5397
5398 // ── range-oriented rich-text ops (P5) ───────────────────────────────────
5399
5400 #[test]
5401 fn editor_wrap_and_toggle_inline_round_trip() {
5402 let mut ed = Editor::new_str("a word b\n", Format::Markdown).expect("editor");
5403
5404 // Bold "word" [2,6); the Change reports the new "**word**" region.
5405 let c = ed.wrap_range(2, 6, InlineKind::Strong).expect("wrap");
5406 assert_eq!(ed.source_str().unwrap(), "a **word** b\n");
5407 assert_eq!(&ed.source_str().unwrap()[c.new.clone()], "**word**");
5408
5409 // Toggle it off by selecting the strong node's interior [4,8).
5410 ed.toggle_inline(4, 8, InlineKind::Strong)
5411 .expect("toggle off");
5412 assert_eq!(ed.source_str().unwrap(), "a word b\n");
5413
5414 // Toggle emphasis on when the range isn't already marked.
5415 ed.toggle_inline(2, 6, InlineKind::Emph).expect("toggle on");
5416 assert_eq!(ed.source_str().unwrap(), "a *word* b\n");
5417 }
5418
5419 #[test]
5420 fn editor_inline_marks_cut_at_block_boundaries() {
5421 // One pair per block, not one pair straddling the blank line — which
5422 // would reparse as four literal asterisks and no mark.
5423 let mut ed = Editor::new_str("one two\n\nthree four\n", Format::Markdown)
5424 .expect("editor");
5425 let c = ed.toggle_inline(0, 19, InlineKind::Strong).expect("toggle on");
5426 assert_eq!(
5427 ed.source_str().unwrap(),
5428 "**one two**\n\n**three four**\n"
5429 );
5430
5431 // One splice, so one Change spanning the lot and one undo step — the
5432 // whole reason the pieces are assembled before anything is written.
5433 assert_eq!(&ed.source_str().unwrap()[c.new.clone()], "**one two**\n\n**three four**");
5434 ed.undo().expect("undo");
5435 assert_eq!(ed.source_str().unwrap(), "one two\n\nthree four\n");
5436
5437 // And the second press removes both, rather than nesting a second pair
5438 // around each.
5439 ed.toggle_inline(0, 19, InlineKind::Strong).expect("toggle on");
5440 ed.toggle_inline(0, 27, InlineKind::Strong).expect("toggle off");
5441 assert_eq!(ed.source_str().unwrap(), "one two\n\nthree four\n");
5442
5443 // A range with nowhere in it to put a mark says so.
5444 let mut fenced = Editor::new_str("```\nx y\n```\n", Format::Markdown).expect("editor");
5445 assert_eq!(
5446 fenced.toggle_inline(4, 7, InlineKind::Strong),
5447 Err(Error::NotEditable)
5448 );
5449 }
5450
5451 #[test]
5452 fn editor_inline_kind_support_is_format_specific() {
5453 // Markdown has no highlight/mark spelling.
5454 let mut md = Editor::new_str("a word b\n", Format::Markdown).expect("editor");
5455 assert_eq!(
5456 md.wrap_range(2, 6, InlineKind::Mark),
5457 Err(Error::UnsupportedFormat)
5458 );
5459
5460 // Djot spells it {=…=}.
5461 let mut dj = Editor::new_str("a word b\n", Format::Djot).expect("editor");
5462 dj.wrap_range(2, 6, InlineKind::Mark).expect("djot mark");
5463 assert_eq!(dj.source_str().unwrap(), "a {=word=} b\n");
5464 }
5465
5466 #[test]
5467 fn editor_authors_gfm_strikethrough_out_of_the_box() {
5468 // The extension that defaults ON, so the default editor is the one
5469 // that can write it — the opposite direction from `highlight` below,
5470 // and no flag on this side turns it off.
5471 assert!(Format::Markdown.supports(Gesture::ToggleInline(InlineKind::Delete)));
5472 let mut ed = Editor::new_str("a word b\n", Format::Markdown).expect("editor");
5473 ed.toggle_inline(2, 6, InlineKind::Delete).expect("strike");
5474 assert_eq!(ed.source_str().unwrap(), "a ~~word~~ b\n");
5475 ed.toggle_inline(4, 8, InlineKind::Delete).expect("unstrike");
5476 assert_eq!(ed.source_str().unwrap(), "a word b\n");
5477 }
5478
5479 #[test]
5480 fn editor_highlight_is_authorable_with_the_extension_on() {
5481 let exts = MarkdownExtensions {
5482 highlight: true,
5483 ..Default::default()
5484 };
5485 // The same format and the same gesture, answered two ways: `==x==` is
5486 // text under default options and a mark under `highlight`, so the
5487 // toggle refuses in one and reverses in the other.
5488 assert!(!Format::Markdown.supports(Gesture::ToggleInline(InlineKind::Mark)));
5489 assert!(Format::Markdown.supports_with(exts, Gesture::ToggleInline(InlineKind::Mark)));
5490
5491 let mut ed =
5492 Editor::new_ext(b"a word b\n", Format::Markdown, exts).expect("editor");
5493 ed.toggle_inline(2, 6, InlineKind::Mark).expect("highlight");
5494 assert_eq!(ed.source_str().unwrap(), "a ==word== b\n");
5495 ed.toggle_inline(4, 8, InlineKind::Mark).expect("unhighlight");
5496 assert_eq!(ed.source_str().unwrap(), "a word b\n");
5497 }
5498
5499 #[test]
5500 fn editor_set_mark_color_writes_reads_and_clears_the_colour() {
5501 let exts = MarkdownExtensions {
5502 highlight: true,
5503 highlight_colors: true,
5504 ..Default::default()
5505 };
5506 assert!(Format::Markdown.supports_with(exts, Gesture::SetMarkColor));
5507 // The narrower gate: highlights alone do not buy a palette.
5508 let hi_only = MarkdownExtensions {
5509 highlight: true,
5510 ..Default::default()
5511 };
5512 assert!(!Format::Markdown.supports_with(hi_only, Gesture::SetMarkColor));
5513 assert!(!Format::Markdown.supports(Gesture::SetMarkColor));
5514 assert!(!Format::Djot.supports_with(exts, Gesture::SetMarkColor));
5515
5516 let mut ed =
5517 Editor::new_ext("a ==word== b\n".as_bytes(), Format::Markdown, exts).expect("editor");
5518 ed.set_mark_color(6, Some(MarkColor::Red)).expect("colour");
5519 assert_eq!(ed.source_str().unwrap(), "a ==\u{1F534} word== b\n");
5520
5521 // And it is queryable as the attribute it is, not as text.
5522 let mut doc =
5523 Document::parse_with(ed.source_str().unwrap().as_bytes(), Format::Markdown, exts)
5524 .expect("parse");
5525 assert_eq!(doc.query("mark[data-color=red]").expect("query").len(), 1);
5526
5527 ed.set_mark_color(9, Some(MarkColor::Blue)).expect("recolour");
5528 assert_eq!(ed.source_str().unwrap(), "a ==\u{1F535} word== b\n");
5529 ed.set_mark_color(9, None).expect("clear");
5530 assert_eq!(ed.source_str().unwrap(), "a ==word== b\n");
5531
5532 // A caret outside any highlight edits nothing.
5533 assert_eq!(
5534 ed.set_mark_color(0, Some(MarkColor::Red)),
5535 Err(Error::NotEditable)
5536 );
5537 assert_eq!(ed.source_str().unwrap(), "a ==word== b\n");
5538
5539 // The names are the attribute values, both ways.
5540 for c in [
5541 MarkColor::Red,
5542 MarkColor::Orange,
5543 MarkColor::Yellow,
5544 MarkColor::Green,
5545 MarkColor::Blue,
5546 MarkColor::Purple,
5547 MarkColor::Brown,
5548 ] {
5549 assert_eq!(MarkColor::from_str(c.as_str()), Some(c));
5550 }
5551 assert_eq!(MarkColor::from_str("pink"), None);
5552 }
5553
5554 #[test]
5555 fn editor_toggle_strips_verbatim_via_content_span() {
5556 let mut ed = Editor::new_str("a `code` b\n", Format::Markdown).expect("editor");
5557 // The verbatim node [2,8) reports content_span [3,7); toggle peels it.
5558 ed.toggle_inline(2, 8, InlineKind::Verbatim)
5559 .expect("toggle code off");
5560 assert_eq!(ed.source_str().unwrap(), "a code b\n");
5561
5562 // A multi-backtick span peels BOTH runs via content_span, not by
5563 // stripping a single delimiter (which would corrupt it to "`x`").
5564 let mut ed2 = Editor::new_str("a ``x`` b\n", Format::Markdown).expect("editor");
5565 ed2.toggle_inline(2, 7, InlineKind::Verbatim)
5566 .expect("toggle multi off");
5567 assert_eq!(ed2.source_str().unwrap(), "a x b\n");
5568 }
5569
5570 #[test]
5571 fn editor_set_block_switches_para_and_heading_levels() {
5572 let mut ed = Editor::new_str("Title\n\nbody text\n", Format::Markdown).expect("editor");
5573
5574 // Paragraph -> H2 (offset 0 is inside "Title").
5575 ed.set_block(0, BlockKind::Heading(2)).expect("to h2");
5576 assert_eq!(ed.source_str().unwrap(), "## Title\n\nbody text\n");
5577
5578 // H2 -> H1 (offset now inside "## Title").
5579 ed.set_block(3, BlockKind::Heading(1)).expect("to h1");
5580 assert_eq!(ed.source_str().unwrap(), "# Title\n\nbody text\n");
5581
5582 // Heading -> paragraph, dropping the marker.
5583 ed.set_block(2, BlockKind::Paragraph).expect("to para");
5584 assert_eq!(ed.source_str().unwrap(), "Title\n\nbody text\n");
5585 }
5586
5587 #[test]
5588 fn editor_set_block_rejects_bad_level_and_format() {
5589 let mut md = Editor::new_str("hi\n", Format::Markdown).expect("editor");
5590 assert_eq!(
5591 md.set_block(0, BlockKind::Heading(9)),
5592 Err(Error::InvalidArgument)
5593 );
5594
5595 let mut xml = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5596 assert_eq!(
5597 xml.set_block(1, BlockKind::Heading(1)),
5598 Err(Error::UnsupportedFormat)
5599 );
5600 }
5601
5602 #[test]
5603 fn editor_toggle_block_container_round_trips() {
5604 let mut ed = Editor::new_str("a\n", Format::Djot).expect("editor");
5605
5606 let c = ed
5607 .toggle_block_container(0, 1, BlockContainerKind::BlockQuote)
5608 .expect("quote on");
5609 assert_eq!(ed.source_str().unwrap(), "> a\n");
5610 assert_eq!(&ed.source_str().unwrap()[c.new.clone()], "> a\n");
5611
5612 ed.toggle_block_container(2, 3, BlockContainerKind::BlockQuote)
5613 .expect("quote off");
5614 assert_eq!(ed.source_str().unwrap(), "a\n");
5615 }
5616
5617 #[test]
5618 fn editor_toggle_block_container_nests_a_partial_selection() {
5619 let mut ed = Editor::new_str("> a\n>\n> b\n", Format::Djot).expect("editor");
5620
5621 // Only the first paragraph is covered, so the quote is not fully
5622 // selected: nest rather than drag `b` out of the quote too.
5623 ed.toggle_block_container(2, 3, BlockContainerKind::BlockQuote)
5624 .expect("nest");
5625 assert_eq!(ed.source_str().unwrap(), "> > a\n>\n> b\n");
5626
5627 // Peel the inner level back off, leaving the outer quote intact.
5628 ed.toggle_block_container(4, 5, BlockContainerKind::BlockQuote)
5629 .expect("peel");
5630 assert_eq!(ed.source_str().unwrap(), "> a\n>\n> b\n");
5631 }
5632
5633 #[test]
5634 fn editor_toggle_block_container_numbers_and_converts_lists() {
5635 let mut ed = Editor::new_str("a\n\nb\n", Format::Djot).expect("editor");
5636
5637 // Each covered block becomes its own numbered item.
5638 ed.toggle_block_container(0, 4, BlockContainerKind::OrderedList)
5639 .expect("ordered on");
5640 assert_eq!(ed.source_str().unwrap(), "1. a\n\n2. b\n");
5641
5642 // The other list kind converts in place instead of nesting.
5643 ed.toggle_block_container(3, 9, BlockContainerKind::BulletList)
5644 .expect("convert");
5645 assert_eq!(ed.source_str().unwrap(), "- a\n\n- b\n");
5646 }
5647
5648 #[test]
5649 fn editor_toggle_block_container_rejects_unspellable_format() {
5650 let mut xml = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5651 assert_eq!(
5652 xml.toggle_block_container(3, 5, BlockContainerKind::BlockQuote),
5653 Err(Error::UnsupportedFormat)
5654 );
5655 }
5656
5657 #[test]
5658 fn editor_insert_link_wraps_and_repoints() {
5659 let mut ed = Editor::new_str("a word b\n", Format::Djot).expect("editor");
5660
5661 ed.insert_link(2, 6, "http://x.dev").expect("link");
5662 assert_eq!(ed.source_str().unwrap(), "a [word](http://x.dev) b\n");
5663
5664 // A caret inside the existing link re-points it rather than nesting.
5665 ed.insert_link(3, 7, "http://y.dev").expect("re-point");
5666 assert_eq!(ed.source_str().unwrap(), "a [word](http://y.dev) b\n");
5667 }
5668
5669 #[test]
5670 fn editor_insert_link_repoints_an_autolink() {
5671 // The regression: an autolink is a `url`/`email` node whose text IS its
5672 // destination. Read as ordinary text, a caret inside it spliced a whole
5673 // new link into the middle of the old URL —
5674 // `see <https<https://y.dev>://x.dev> ok`.
5675 for format in [Format::Markdown, Format::Djot] {
5676 let mut ed = Editor::new_str("see <https://x.dev> ok\n", format).expect("editor");
5677 ed.insert_link(10, 10, "https://y.dev").expect("re-point");
5678 assert_eq!(ed.source_str().unwrap(), "see <https://y.dev> ok\n");
5679
5680 // Source that looks right can still parse wrong: assert the reparse.
5681 let nodes = ed.nodes().expect("nodes");
5682 let url = nodes
5683 .iter()
5684 .find(|n| n.kind == Kind::Url)
5685 .expect("still an autolink");
5686 assert_eq!(url.text.as_deref(), Some("https://y.dev"));
5687 assert!(!nodes.iter().any(|n| n.kind == Kind::Link));
5688 }
5689 }
5690
5691 #[test]
5692 fn editor_insert_link_escapes_the_destination() {
5693 // Unescaped, the `)` would close the link early and spill `b` into the
5694 // paragraph as literal text.
5695 let mut dj = Editor::new_str("w\n", Format::Djot).expect("editor");
5696 dj.insert_link(0, 1, "a)b").expect("link");
5697 assert_eq!(dj.source_str().unwrap(), "[w](a\\)b)\n");
5698
5699 // Whitespace is where the formats part ways: Markdown needs the angle
5700 // form (a bare space ends the destination and kills the link outright),
5701 // Djot must NOT use it (it would link to the literal text `<a b>`).
5702 let mut md = Editor::new_str("w\n", Format::Markdown).expect("editor");
5703 md.insert_link(0, 1, "a b").expect("link");
5704 assert_eq!(md.source_str().unwrap(), "[w](<a b>)\n");
5705
5706 let mut dj2 = Editor::new_str("w\n", Format::Djot).expect("editor");
5707 dj2.insert_link(0, 1, "a b").expect("link");
5708 assert_eq!(dj2.source_str().unwrap(), "[w](a b)\n");
5709 }
5710
5711 #[test]
5712 fn editor_insert_image_escapes_the_destination_per_format() {
5713 // The whole point of the op: a caller's `` is not an image
5714 // in Markdown, and the correct repair differs by format.
5715 let mut md = Editor::new_str("w\n", Format::Markdown).expect("editor");
5716 md.insert_image(0, 1, "my cat.png").expect("image");
5717 assert_eq!(md.source_str().unwrap(), "\n");
5718
5719 let mut dj = Editor::new_str("w\n", Format::Djot).expect("editor");
5720 dj.insert_image(0, 1, "my cat.png").expect("image");
5721 assert_eq!(dj.source_str().unwrap(), "\n");
5722
5723 // A `)` would close the image early and spill the rest as literal text.
5724 let mut paren = Editor::new_str("w\n", Format::Djot).expect("editor");
5725 paren.insert_image(0, 1, "a)b.png").expect("image");
5726 assert_eq!(paren.source_str().unwrap(), "b.png)\n");
5727 }
5728
5729 #[test]
5730 fn editor_insert_image_keeps_an_empty_alt_empty() {
5731 // Unlike a link, where an empty range spells an autolink or doubles the
5732 // destination as text — an image with no alt is ordinary.
5733 let mut ed = Editor::new_str("ab\n", Format::Markdown).expect("editor");
5734 ed.insert_image(1, 1, "cat.png").expect("image");
5735 assert_eq!(ed.source_str().unwrap(), "ab\n");
5736 }
5737
5738 #[test]
5739 fn editor_insert_image_rejects_a_newline_destination() {
5740 let mut ed = Editor::new_str("w\n", Format::Djot).expect("editor");
5741 assert_eq!(
5742 ed.insert_image(0, 1, "a\nb.png"),
5743 Err(Error::InvalidArgument)
5744 );
5745
5746 let mut xml = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5747 assert_eq!(
5748 xml.insert_image(3, 5, "x.png"),
5749 Err(Error::UnsupportedFormat)
5750 );
5751 }
5752
5753 #[test]
5754 fn editor_insert_link_rejects_a_newline_destination() {
5755 let mut ed = Editor::new_str("w\n", Format::Djot).expect("editor");
5756 assert_eq!(ed.insert_link(0, 1, "a\nb"), Err(Error::InvalidArgument));
5757
5758 let mut xml = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5759 assert_eq!(xml.insert_link(3, 5, "u"), Err(Error::UnsupportedFormat));
5760 }
5761
5762 #[test]
5763 fn editor_insert_literal_keeps_typed_specials_literal() {
5764 for format in [Format::Markdown, Format::Djot] {
5765 let mut ed = Editor::new_str("z\n", format).expect("editor");
5766 // A `*` at a line start would open emphasis unescaped.
5767 ed.insert_literal(0, "*hi*").expect("literal");
5768
5769 // Source that looks right can still parse wrong: assert the reparse.
5770 let nodes = ed.nodes().expect("nodes");
5771 assert!(
5772 !nodes
5773 .iter()
5774 .any(|n| n.kind == Kind::Emph || n.kind == Kind::Strong)
5775 );
5776 let text: String = nodes
5777 .iter()
5778 .filter(|n| n.kind == Kind::Str)
5779 .filter_map(|n| n.text.clone())
5780 .collect();
5781 assert_eq!(text, "*hi*z");
5782 }
5783 }
5784
5785 #[test]
5786 fn editor_insert_literal_escapes_block_markers_only_at_line_start() {
5787 // Mid-line, a `#` opens nothing and is left as typed.
5788 let mut ed = Editor::new_str("az\n", Format::Markdown).expect("editor");
5789 ed.insert_literal(1, "# ").expect("literal");
5790 assert_eq!(ed.source_str().unwrap(), "a# z\n");
5791
5792 // At a line start it would open a heading, so it is escaped.
5793 let mut ed2 = Editor::new_str("z\n", Format::Markdown).expect("editor");
5794 ed2.insert_literal(0, "# ").expect("literal");
5795 assert_eq!(ed2.source_str().unwrap(), "\\# z\n");
5796 assert!(
5797 !ed2.nodes()
5798 .expect("nodes")
5799 .iter()
5800 .any(|n| n.kind == Kind::Heading)
5801 );
5802 }
5803
5804 #[test]
5805 fn editor_insert_literal_escapes_a_dollar_only_under_math() {
5806 // Under the math extension `$` opens a formula, so it is escaped.
5807 let exts = MarkdownExtensions { math: true, ..Default::default() };
5808 let mut ed = Editor::new_ext(b"a \n", Format::Markdown, exts).expect("editor");
5809 ed.insert_literal(2, "$x$ and $$y$$").expect("literal");
5810 assert_eq!(ed.source_str().unwrap(), "a \\$x\\$ and \\$\\$y\\$\\$\n");
5811
5812 // Without it a `$` is text, and stays bare.
5813 let mut plain = Editor::new_str("a \n", Format::Markdown).expect("editor");
5814 plain.insert_literal(2, "$x$ and $$y$$").expect("literal");
5815 assert_eq!(plain.source_str().unwrap(), "a $x$ and $$y$$\n");
5816 }
5817
5818 #[test]
5819 fn editor_insert_literal_rejects_bad_offset_and_parse_only_format() {
5820 let mut ed = Editor::new_str("ab\n", Format::Markdown).expect("editor");
5821 assert_eq!(ed.insert_literal(99, "x"), Err(Error::InvalidArgument));
5822
5823 let mut xml = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5824 assert_eq!(xml.insert_literal(3, "x"), Err(Error::UnsupportedFormat));
5825 }
5826
5827 #[test]
5828 fn editor_insert_line_break_splices_in_cell_br() {
5829 let mut ed =
5830 Editor::new_str("| a | b |\n| --- | --- |\n", Format::Markdown).expect("editor");
5831 // Caret just after `a` in the header cell.
5832 ed.insert_line_break(3).expect("line break");
5833 assert_eq!(ed.source_str().unwrap(), "| a<br> | b |\n| --- | --- |\n");
5834 // The break reads back as a semantic node, not raw HTML.
5835 let nodes = ed.nodes().expect("nodes");
5836 assert!(nodes.iter().any(|n| n.kind == Kind::HardBreak));
5837 assert!(!nodes.iter().any(|n| n.kind == Kind::RawInline));
5838 }
5839
5840 #[test]
5841 fn editor_insert_line_break_rejects_off_cell_off_format_and_bad_offset() {
5842 // Not inside a cell → NotFound.
5843 let mut para = Editor::new_str("just text\n", Format::Markdown).expect("editor");
5844 assert_eq!(para.insert_line_break(3), Err(Error::NotFound));
5845
5846 // Djot has no in-cell break spelling → UnsupportedFormat.
5847 let mut dj = Editor::new_str("| a | b |\n| --- | --- |\n", Format::Djot).expect("editor");
5848 assert_eq!(dj.insert_line_break(3), Err(Error::UnsupportedFormat));
5849
5850 // Out-of-range offset → InvalidArgument.
5851 let mut ed =
5852 Editor::new_str("| a | b |\n| --- | --- |\n", Format::Markdown).expect("editor");
5853 assert_eq!(ed.insert_line_break(9999), Err(Error::InvalidArgument));
5854 }
5855
5856 #[test]
5857 fn editor_insert_thematic_break_is_blank_separated_per_format() {
5858 // The blank line above is load-bearing, not cosmetic: flush against the
5859 // paragraph, Markdown's `---` is a setext underline and the paragraph
5860 // becomes an <h2>. So assert the reparsed KIND, not just the bytes.
5861 let mut md = Editor::new_str("a\n", Format::Markdown).expect("editor");
5862 md.insert_thematic_break(0).expect("rule");
5863 assert_eq!(md.source_str().unwrap(), "a\n\n---\n");
5864 let nodes = md.nodes().expect("nodes");
5865 assert!(nodes.iter().any(|n| n.kind == Kind::ThematicBreak));
5866 assert!(!nodes.iter().any(|n| n.kind == Kind::Heading));
5867
5868 // Djot spells the same construct differently — the reason the spelling
5869 // is the library's and not the caller's.
5870 let mut dj = Editor::new_str("a\n", Format::Djot).expect("editor");
5871 dj.insert_thematic_break(0).expect("rule");
5872 assert_eq!(dj.source_str().unwrap(), "a\n\n* * *\n");
5873
5874 let mut xml = Editor::new_str("<a>hi</a>", Format::Xml).expect("editor");
5875 assert_eq!(xml.insert_thematic_break(3), Err(Error::UnsupportedFormat));
5876 }
5877
5878 #[test]
5879 fn editor_insert_table_writes_an_editable_table_after_the_block() {
5880 let mut md = Editor::new_str("a\n", Format::Markdown).expect("editor");
5881 md.insert_table(0, 1, 2).expect("table");
5882 assert_eq!(md.source_str().unwrap(), "a\n\n| | |\n| --- | --- |\n| | |\n");
5883 // What was minted is what the table edits read back.
5884 md.table_insert_row(4, true).expect("row");
5885 assert_eq!(
5886 md.source_str().unwrap(),
5887 "a\n\n| | |\n| --- | --- |\n| | |\n| | |\n"
5888 );
5889
5890 let mut dj = Editor::new_str("a\n", Format::Djot).expect("editor");
5891 dj.insert_table(0, 1, 2).expect("table");
5892 assert_eq!(dj.source_str().unwrap(), "a\n\n| | |\n|---|---|\n| | |\n");
5893
5894 let mut md = Editor::new_str("a\n", Format::Markdown).expect("editor");
5895 assert_eq!(md.insert_table(0, 0, 2), Err(Error::InvalidArgument));
5896 assert_eq!(md.insert_table(0, 1, 0), Err(Error::InvalidArgument));
5897 assert_eq!(md.source_str().unwrap(), "a\n");
5898
5899 let mut html = Editor::new_str("<p>ab</p>\n", Format::Html).expect("editor");
5900 assert_eq!(html.insert_table(4, 1, 1), Err(Error::UnsupportedFormat));
5901 assert!(!Format::Html.supports(Gesture::InsertTable));
5902 assert!(Format::Markdown.supports(Gesture::InsertTable));
5903 }
5904
5905 #[test]
5906 fn editor_split_block_keeps_both_halves_the_same_kind() {
5907 // A list item's halves are both items — the marker is repeated, so the
5908 // second half doesn't fall out of the list as a paragraph.
5909 let mut item = Editor::new_str("- this is a list item\n", Format::Markdown).expect("editor");
5910 item.split_block(10).expect("split");
5911 assert_eq!(item.source_str().unwrap(), "- this is \n- a list item\n");
5912 let nodes = item.nodes().expect("nodes");
5913 assert_eq!(nodes.iter().filter(|n| n.kind == Kind::ListItem).count(), 2);
5914
5915 // At the item's end the empty sibling IS the point — that is Enter.
5916 let mut tail = Editor::new_str("- a\n", Format::Markdown).expect("editor");
5917 tail.split_block(3).expect("split");
5918 assert_eq!(tail.source_str().unwrap(), "- a\n- \n");
5919
5920 // A paragraph divides on a blank line instead.
5921 let mut para = Editor::new_str("ab\n", Format::Markdown).expect("editor");
5922 para.split_block(1).expect("split");
5923 assert_eq!(para.source_str().unwrap(), "a\n\nb\n");
5924
5925 // A table has no honest caret-split: a newline mid-cell destroys it.
5926 let mut table =
5927 Editor::new_str("| a | b |\n|---|---|\n| c | d |\n", Format::Markdown).expect("editor");
5928 assert_eq!(table.split_block(3), Err(Error::NotEditable));
5929
5930 let mut empty = Editor::new_str("", Format::Markdown).expect("editor");
5931 assert_eq!(empty.split_block(0), Err(Error::NotFound));
5932 }
5933
5934 #[test]
5935 fn editor_join_blocks_carries_the_markup_that_has_to_travel() {
5936 // The Markdown `<div>` case, which is the whole reason this is a
5937 // gesture: `below` joins the div's last paragraph, and the `</div>` is
5938 // carried past the text that came in rather than being left above it.
5939 let src = "above\n\n<div class=\"center\">\n\nhello\n\n</div>\n\nbelow\n";
5940 let exts = MarkdownExtensions {
5941 html_elements: true,
5942 ..Default::default()
5943 };
5944 let mut into = Editor::new_ext(src.as_bytes(), Format::Markdown, exts).expect("editor");
5945 into.join_blocks(44).expect("join");
5946 assert_eq!(
5947 into.source_str().unwrap(),
5948 "above\n\n<div class=\"center\">\n\nhello\nbelow\n\n</div>\n"
5949 );
5950 let nodes = into.nodes().expect("nodes");
5951 assert_eq!(nodes.iter().filter(|n| n.kind == Kind::Para).count(), 2);
5952
5953 // Out of the div instead: B leaves it, so the div's own closers go with
5954 // B's markup — and the paragraph's attributes go with them, because the
5955 // joined text is the other block's.
5956 let mut out_of = Editor::new_ext(src.as_bytes(), Format::Markdown, exts).expect("editor");
5957 out_of.join_blocks(29).expect("join");
5958 assert_eq!(out_of.source_str().unwrap(), "above\nhello\n\nbelow\n");
5959
5960 // A list item's continuation is spaces, not its marker: two items in,
5961 // one out.
5962 let mut item = Editor::new_str("- a\n- b\n", Format::Markdown).expect("editor");
5963 item.join_blocks(6).expect("join");
5964 assert_eq!(item.source_str().unwrap(), "- a\n b\n");
5965 let nodes = item.nodes().expect("nodes");
5966 assert_eq!(nodes.iter().filter(|n| n.kind == Kind::ListItem).count(), 1);
5967
5968 // The document's first block has nothing above it.
5969 let mut first = Editor::new_str("above\n", Format::Markdown).expect("editor");
5970 assert_eq!(first.join_blocks(0), Err(Error::NotFound));
5971 }
5972
5973 #[test]
5974 fn editor_move_block_spells_the_destination_prefixes() {
5975 // Out of a quote: the `> ` and the `>` separator line go with the
5976 // block, and it arrives blank-separated at the top level.
5977 let mut out = Editor::new_str("> a\n>\n> y\n\nb\n", Format::Markdown).expect("editor");
5978 let change = out.move_block(8, 13).expect("move");
5979 assert_eq!(out.source_str().unwrap(), "> a\n\nb\n\ny\n");
5980 assert_eq!(change.old.start, 4);
5981
5982 // Into a list item's tail: the continuation indent and the blank line
5983 // that makes it a second block of the item.
5984 let mut tail = Editor::new_str("- x\n\ny\n", Format::Markdown).expect("editor");
5985 tail.move_block(5, 3).expect("move");
5986 assert_eq!(tail.source_str().unwrap(), "- x\n\n y\n");
5987 let nodes = tail.nodes().expect("nodes");
5988 assert_eq!(nodes.iter().filter(|n| n.kind == Kind::ListItem).count(), 1);
5989
5990 // A bullet's text is the bullet: the item moves whole, and between
5991 // two items of a tight list no blank is written.
5992 let mut items = Editor::new_str("- a\n - b\n- c\n", Format::Markdown).expect("editor");
5993 items.move_block(12, 0).expect("move");
5994 assert_eq!(items.source_str().unwrap(), "- c\n- a\n - b\n");
5995
5996 // Where it already is moves nothing; inside a fence is no boundary.
5997 let mut same = Editor::new_str("a\n\n```\nx\n```\n", Format::Markdown).expect("editor");
5998 assert_eq!(same.move_block(0, 1), Err(Error::InvalidArgument));
5999 assert_eq!(same.move_block(0, 8), Err(Error::NotEditable));
6000 assert_eq!(same.move_block(2, 0), Err(Error::NotFound));
6001
6002 // XML has no blocks a caret could name.
6003 let mut xml = Editor::new_str("<r><a/><b/></r>", Format::Xml).expect("editor");
6004 assert_eq!(xml.move_block(3, 11), Err(Error::UnsupportedFormat));
6005 }
6006
6007 #[test]
6008 fn editor_join_blocks_is_spelled_in_html_where_the_split_is_not() {
6009 // The pair that motivates the separate gate. A blank line between two
6010 // `<p>`s is not what separates them, so the split refuses; a newline
6011 // inside one is exactly the break the join needs, and the reparse hands
6012 // back the single paragraph the gesture claims to have made.
6013 let mut ed =
6014 Editor::new_str("<p>a</p>\n<p class=\"x\">b</p>\n", Format::Html).expect("editor");
6015 assert_eq!(ed.split_block(4), Err(Error::UnsupportedFormat));
6016 assert!(Format::Html.supports(Gesture::JoinBlocks));
6017 ed.join_blocks(22).expect("join");
6018 assert_eq!(ed.source_str().unwrap(), "<p>a\nb</p>\n");
6019 let nodes = ed.nodes().expect("nodes");
6020 assert_eq!(nodes.iter().filter(|n| n.kind == Kind::Para).count(), 1);
6021
6022 // XML spells nothing at all, so there the refusal is the format's.
6023 let mut xml = Editor::new_str("<r>ab</r>", Format::Xml).expect("editor");
6024 assert_eq!(xml.join_blocks(4), Err(Error::UnsupportedFormat));
6025 }
6026
6027 #[test]
6028 fn editor_toggle_code_block_round_trips_and_measures_the_fence() {
6029 let mut ed = Editor::new_str("a\n", Format::Markdown).expect("editor");
6030 ed.toggle_code_block(0, 1, Some("zig")).expect("fence");
6031 assert_eq!(ed.source_str().unwrap(), "```zig\na\n```\n");
6032 let nodes = ed.nodes().expect("nodes");
6033 assert!(nodes.iter().any(|n| n.kind == Kind::CodeBlock));
6034
6035 ed.toggle_code_block(0, 0, None).expect("unfence");
6036 assert_eq!(ed.source_str().unwrap(), "a\n");
6037
6038 // Three backticks in the body would close a three-backtick fence, so the
6039 // fence is measured against the body rather than fixed.
6040 let mut runs = Editor::new_str("a ``` b\n", Format::Markdown).expect("editor");
6041 runs.toggle_code_block(0, 7, None).expect("fence");
6042 assert_eq!(runs.source_str().unwrap(), "````\na ``` b\n````\n");
6043 }
6044
6045 #[test]
6046 fn editor_toggle_code_block_refuses_inside_a_list_item() {
6047 // A fence at column zero here would pull the item's `- ` into the code
6048 // body and the item would stop being an item.
6049 let mut ed = Editor::new_str("- a\n- b\n", Format::Markdown).expect("editor");
6050 assert_eq!(ed.toggle_code_block(2, 3, None), Err(Error::NotEditable));
6051 assert_eq!(ed.source_str().unwrap(), "- a\n- b\n");
6052 }
6053
6054 #[test]
6055 fn editor_set_code_language_retags_clears_and_refuses() {
6056 let mut ed = Editor::new_str("```zig\na\n```\n", Format::Markdown).expect("editor");
6057 ed.set_code_language(0, Some("rust")).expect("retag");
6058 assert_eq!(ed.source_str().unwrap(), "```rust\na\n```\n");
6059
6060 // `None` clears the info string; `Some("")` writes the same bytes but is
6061 // a different request.
6062 ed.set_code_language(0, None).expect("clear");
6063 assert_eq!(ed.source_str().unwrap(), "```\na\n```\n");
6064 ed.set_code_language(0, Some("")).expect("empty");
6065 assert_eq!(ed.source_str().unwrap(), "```\na\n```\n");
6066
6067 // Markdown's info string ends at whitespace, so a space would come back
6068 // truncated — refused rather than silently clipped.
6069 assert_eq!(
6070 ed.set_code_language(0, Some("a b")),
6071 Err(Error::InvalidArgument)
6072 );
6073 // Djot's runs to the end of the line, so the same string is fine there.
6074 let mut dj = Editor::new_str("```\na\n```\n", Format::Djot).expect("editor");
6075 dj.set_code_language(0, Some("a b"))
6076 .expect("djot info string");
6077 assert_eq!(dj.source_str().unwrap(), "```a b\na\n```\n");
6078
6079 let mut para = Editor::new_str("x\n", Format::Markdown).expect("editor");
6080 assert_eq!(para.set_code_language(0, Some("zig")), Err(Error::NotFound));
6081 }
6082
6083 #[test]
6084 fn editor_task_checkbox_gestures() {
6085 let mut ed = Editor::new_str("- a\n", Format::Markdown).expect("editor");
6086
6087 // The box is added by one gesture and ticked by another — adding
6088 // converts the item's kind, ticking only changes what the box holds.
6089 ed.toggle_task_item(2).expect("add box");
6090 assert_eq!(ed.source_str().unwrap(), "- [ ] a\n");
6091 assert!(
6092 ed.nodes()
6093 .unwrap()
6094 .iter()
6095 .any(|n| n.kind == Kind::TaskListItem)
6096 );
6097
6098 ed.set_task_checked(6, true).expect("tick");
6099 assert_eq!(ed.source_str().unwrap(), "- [x] a\n");
6100 // Already checked: a no-op that still succeeds and moves nothing.
6101 ed.set_task_checked(6, true).expect("no-op");
6102 assert_eq!(ed.source_str().unwrap(), "- [x] a\n");
6103
6104 ed.toggle_task_checked(6).expect("flip");
6105 assert_eq!(ed.source_str().unwrap(), "- [ ] a\n");
6106
6107 ed.toggle_task_item(6).expect("remove box");
6108 assert_eq!(ed.source_str().unwrap(), "- a\n");
6109
6110 // A plain bullet has no box to tick; `toggle_task_item` is how a caller
6111 // asks for one.
6112 assert_eq!(ed.set_task_checked(2, true), Err(Error::NotEditable));
6113 // And a caret in no list item has no item at all.
6114 let mut para = Editor::new_str("a\n", Format::Markdown).expect("editor");
6115 assert_eq!(para.toggle_task_item(0), Err(Error::NotFound));
6116 }
6117
6118 #[test]
6119 fn editor_insert_footnote_writes_both_halves_as_one_edit() {
6120 for format in [Format::Markdown, Format::Djot] {
6121 let mut ed = Editor::new_str("see\n", format).expect("editor");
6122 ed.insert_footnote(3, "a").expect("footnote");
6123 assert_eq!(ed.source_str().unwrap(), "see[^a]\n\n[^a]: \n");
6124
6125 // Half a footnote is not a footnote, so assert both nodes exist.
6126 let nodes = ed.nodes().expect("nodes");
6127 assert!(nodes.iter().any(|n| n.kind == Kind::FootnoteReference));
6128 assert!(nodes.iter().any(|n| n.kind == Kind::Footnote));
6129
6130 // One edit, so one undo takes both halves back.
6131 ed.undo().expect("undo");
6132 assert_eq!(ed.source_str().unwrap(), "see\n");
6133 }
6134 }
6135
6136 #[test]
6137 fn editor_insert_footnote_reuses_an_existing_definition() {
6138 let mut ed = Editor::new_str("see\n", Format::Markdown).expect("editor");
6139 ed.insert_footnote(3, "a").expect("first");
6140 ed.insert_footnote(7, "a").expect("second reference");
6141 assert_eq!(ed.source_str().unwrap(), "see[^a][^a]\n\n[^a]: \n");
6142 let defs = ed
6143 .nodes()
6144 .unwrap()
6145 .iter()
6146 .filter(|n| n.kind == Kind::Footnote)
6147 .count();
6148 assert_eq!(defs, 1);
6149
6150 assert_eq!(ed.insert_footnote(3, ""), Err(Error::InvalidArgument));
6151 assert_eq!(ed.insert_footnote(3, "a]b"), Err(Error::InvalidArgument));
6152 }
6153
6154 #[test]
6155 fn editor_undo_redo_round_trip() {
6156 let mut ed = Editor::new_str("hello\n", Format::Markdown).expect("editor");
6157 ed.edit_range(5, 5, "!").expect("edit");
6158 assert_eq!(ed.source_str().unwrap(), "hello!\n");
6159
6160 let change = ed.undo().expect("undo ok").expect("something to undo");
6161 assert_eq!(ed.source_str().unwrap(), "hello\n");
6162 assert_eq!(change.new.end, 5);
6163 assert!(ed.undo().expect("undo ok").is_none(), "history exhausted");
6164
6165 ed.redo().expect("redo ok").expect("something to redo");
6166 assert_eq!(ed.source_str().unwrap(), "hello!\n");
6167 }
6168
6169 #[test]
6170 fn editor_coalesce_folds_a_run() {
6171 let mut ed = Editor::new_str("\n", Format::Markdown).expect("editor");
6172 ed.edit_range(0, 0, "a").expect("edit");
6173 ed.edit_range(1, 1, "b").expect("edit");
6174 ed.coalesce_last_undo().expect("coalesce");
6175 assert_eq!(ed.source_str().unwrap(), "ab\n");
6176 // One undo removes the whole coalesced run.
6177 ed.undo().expect("undo ok").expect("something to undo");
6178 assert_eq!(ed.source_str().unwrap(), "\n");
6179 assert!(ed.undo().expect("undo ok").is_none());
6180 }
6181
6182 #[test]
6183 fn editor_revision_bumps_per_successful_mutation() {
6184 let mut ed = Editor::new_str("x\n", Format::Markdown).expect("editor");
6185 assert_eq!(ed.revision(), 0);
6186 ed.edit_range(1, 1, "y").expect("edit");
6187 assert_eq!(ed.revision(), 1);
6188
6189 // A reparse-breaking edit is rolled back and must not bump the revision.
6190 let mut xml = Editor::new_str("<a>ok</a>", Format::Xml).expect("editor");
6191 assert_eq!(xml.revision(), 0);
6192 assert!(xml.replace_content("0", "<b>").is_err());
6193 assert_eq!(xml.revision(), 0);
6194
6195 // undo and redo are mutations too.
6196 ed.undo().expect("undo ok").expect("something to undo");
6197 assert_eq!(ed.revision(), 2);
6198 ed.redo().expect("redo ok").expect("something to redo");
6199 assert_eq!(ed.revision(), 3);
6200 }
6201
6202 #[test]
6203 fn editor_dirty_range_tracks_and_clears() {
6204 let mut ed = Editor::new_str("abcdefgh\n", Format::Markdown).expect("editor");
6205 // Clean to start.
6206 assert_eq!(ed.dirty_range(), None);
6207
6208 // One insertion of two bytes at offset 2 dirties exactly [2, 4).
6209 ed.edit_range(2, 2, "XY").expect("edit");
6210 assert_eq!(ed.dirty_range(), Some(2..4));
6211
6212 // A second, disjoint edit near the end accumulates conservatively: the
6213 // reported range is a superset covering both edits.
6214 ed.edit_range(9, 9, "Z").expect("edit"); // source is now "abXYcdefgZh\n"
6215 let d = ed.dirty_range().expect("dirty");
6216 assert!(
6217 d.start <= 2 && d.end >= 10,
6218 "range {d:?} must cover both edits"
6219 );
6220
6221 // clear_dirty acknowledges without moving the revision.
6222 let rev = ed.revision();
6223 ed.clear_dirty();
6224 assert_eq!(ed.dirty_range(), None);
6225 assert_eq!(ed.revision(), rev);
6226
6227 // Post-clear, only new mutations show up — and undo counts as one.
6228 ed.undo().expect("undo ok").expect("something to undo");
6229 assert!(ed.dirty_range().is_some());
6230 }
6231
6232 #[test]
6233 fn editor_caret_blob_follows_undo_and_redo() {
6234 let mut ed = Editor::new_str("hello\n", Format::Markdown).expect("editor");
6235 assert!(ed.caret_blob().unwrap().is_empty());
6236
6237 // Set the pre-edit caret, then edit: the retired undo step captures it.
6238 ed.set_caret_blob(b"before").expect("set caret");
6239 ed.edit_range(5, 5, "!").expect("edit");
6240 // A fresh state starts caret-less until the host sets one.
6241 assert!(ed.caret_blob().unwrap().is_empty());
6242 ed.set_caret_blob(b"after").expect("set caret");
6243
6244 // Undo restores the pre-edit source AND the pre-edit caret.
6245 ed.undo().expect("undo ok").expect("something to undo");
6246 assert_eq!(ed.source_str().unwrap(), "hello\n");
6247 assert_eq!(ed.caret_blob().unwrap(), b"before");
6248
6249 // Redo restores the post-edit source AND the post-edit caret.
6250 ed.redo().expect("redo ok").expect("something to redo");
6251 assert_eq!(ed.source_str().unwrap(), "hello!\n");
6252 assert_eq!(ed.caret_blob().unwrap(), b"after");
6253 }
6254
6255 #[test]
6256 fn editor_coalesced_run_keeps_the_pre_run_caret() {
6257 let mut ed = Editor::new_str("\n", Format::Markdown).expect("editor");
6258 ed.set_caret_blob(b"c0").expect("set caret");
6259 ed.edit_range(0, 0, "a").expect("edit");
6260 ed.set_caret_blob(b"c1").expect("set caret");
6261 ed.edit_range(1, 1, "b").expect("edit");
6262 ed.coalesce_last_undo().expect("coalesce");
6263 ed.set_caret_blob(b"c2").expect("set caret");
6264
6265 // One undo folds the run and restores the caret from before it began.
6266 ed.undo().expect("undo ok").expect("something to undo");
6267 assert_eq!(ed.source_str().unwrap(), "\n");
6268 assert_eq!(ed.caret_blob().unwrap(), b"c0");
6269 }
6270
6271 #[test]
6272 fn editor_renumber_ordered_lists_fixes_a_stale_sequence() {
6273 let mut ed = Editor::new_str("1. a\n2. x\n2. b\n3. c\n", Format::Markdown).expect("editor");
6274 ed.renumber_ordered_lists(0).expect("renumber ok");
6275 assert_eq!(ed.source_str().unwrap(), "1. a\n2. x\n3. b\n4. c\n");
6276 }
6277
6278 #[test]
6279 fn editor_renumber_ordered_lists_leaves_djot_prose_alone() {
6280 // Djot reads ` 2. b` as text inside item `a`; Markdown reads the same
6281 // bytes as a nested item. The author's digit survives in the one case.
6282 let src = "1. a\n 2. b\n2. c\n";
6283 let mut dj = Editor::new_str(src, Format::Djot).expect("editor");
6284 dj.renumber_ordered_lists(0).expect("renumber ok");
6285 assert_eq!(dj.source_str().unwrap(), src);
6286
6287 let mut md = Editor::new_str(src, Format::Markdown).expect("editor");
6288 md.renumber_ordered_lists(0).expect("renumber ok");
6289 assert_eq!(md.source_str().unwrap(), "1. a\n 1. b\n2. c\n");
6290 }
6291
6292 #[test]
6293 fn editor_renumber_ordered_lists_off_a_list_is_not_found() {
6294 let mut ed = Editor::new_str("a paragraph\n", Format::Markdown).expect("editor");
6295 assert!(matches!(ed.renumber_ordered_lists(2), Err(Error::NotFound)));
6296 }
6297
6298 #[test]
6299 fn editor_table_insert_row_and_set_alignment() {
6300 let src = "| a | b |\n| --- | --- |\n| 1 | 2 |\n";
6301 let mut ed = Editor::new_str(src, Format::Markdown).expect("editor");
6302 ed.table_insert_row(24, true).expect("insert row"); // caret in body `1`
6303 assert_eq!(
6304 ed.source_str().unwrap(),
6305 "| a | b |\n| --- | --- |\n| 1 | 2 |\n| | |\n"
6306 );
6307 ed.table_set_alignment(6, Alignment::Center).expect("align"); // column `b`
6308 assert!(ed.source_str().unwrap().contains("| --- | :---: |"));
6309 }
6310
6311 #[test]
6312 fn editor_table_edit_off_a_table_is_not_found() {
6313 let mut ed = Editor::new_str("nope\n", Format::Markdown).expect("editor");
6314 assert!(matches!(ed.table_delete_row(2), Err(Error::NotFound)));
6315 }
6316
6317 #[test]
6318 fn editor_set_block_converts_setext_heading() {
6319 // A setext heading rebuilt from its content_span collapses the underline.
6320 let mut ed = Editor::new_str("Title\n=====\n\nbody\n", Format::Markdown).expect("editor");
6321 ed.set_block(0, BlockKind::Heading(1))
6322 .expect("setext to atx");
6323 assert_eq!(ed.source_str().unwrap(), "# Title\n\nbody\n");
6324 }
6325
6326 #[test]
6327 fn editor_moves_a_node_next_to_another_in_one_step() {
6328 let mut ed = Editor::new_str("<svg>\n <rect/>\n <circle/>\n</svg>\n", Format::Xml)
6329 .expect("editor");
6330 ed.move_after("element[name=\"rect\"]", "element[name=\"circle\"]")
6331 .expect("after");
6332 assert_eq!(
6333 ed.source_str().unwrap(),
6334 "<svg>\n <circle/>\n <rect/>\n</svg>\n"
6335 );
6336 ed.move_before("element[name=\"rect\"]", "element[name=\"circle\"]")
6337 .expect("before");
6338 assert_eq!(
6339 ed.source_str().unwrap(),
6340 "<svg>\n <rect/>\n <circle/>\n</svg>\n"
6341 );
6342 // One undo step per move.
6343 ed.undo().expect("undo");
6344 assert_eq!(
6345 ed.source_str().unwrap(),
6346 "<svg>\n <circle/>\n <rect/>\n</svg>\n"
6347 );
6348 assert_eq!(
6349 ed.move_after("element[name=\"rect\"]", "element[name=\"svg\"]"),
6350 Err(Error::InvalidArgument)
6351 );
6352 assert_eq!(
6353 ed.move_after("element[name=\"rect\"]", "element[name=\"path\"]"),
6354 Err(Error::NotFound)
6355 );
6356 // Markdown's blank-line separator travels the same way.
6357 let mut md = Editor::new_str("A\n\nB\n\nC\n", Format::Markdown).expect("editor");
6358 md.move_after("1", "2").expect("after");
6359 assert_eq!(md.source_str().unwrap(), "A\n\nC\n\nB\n");
6360 }
6361
6362 #[test]
6363 fn editor_unwrap_and_smart_delete() {
6364 let mut ed = Editor::new_str("<r><box><b/><c/></box></r>", Format::Xml).expect("editor");
6365 ed.unwrap_node("0.0").expect("unwrap"); // <box>
6366 assert_eq!(ed.source_str().expect("source"), "<r><b/><c/></r>");
6367
6368 let mut md = Editor::new_str("A\n\nB\n\nC\n", Format::Markdown).expect("editor");
6369 md.delete_smart("1").expect("delete_smart"); // the "B" paragraph
6370 assert_eq!(md.source_str().expect("source"), "A\n\nC\n");
6371 }
6372
6373 #[test]
6374 fn editor_directives_require_the_extension_flag() {
6375 let src = ":::vis{.public}\nhi\n:::\n";
6376 // Without the flag, the colon-fence lines are plain paragraph text —
6377 // no directive node.
6378 let mut plain = Editor::new_str(src, Format::Markdown).expect("editor");
6379 assert_eq!(plain.query("directive").expect("query").len(), 0);
6380 // With it enabled, the container directive is recognized.
6381 let mut ext = Editor::new_ext(
6382 src.as_bytes(),
6383 Format::Markdown,
6384 MarkdownExtensions {
6385 directives: true,
6386 ..Default::default()
6387 },
6388 )
6389 .expect("editor");
6390 assert_eq!(ext.query("directive").expect("query").len(), 1);
6391 }
6392
6393 #[test]
6394 fn editor_set_node_attrs_rewrites_an_xml_tag_by_id_and_nothing_else() {
6395 assert!(Format::Xml.supports(Gesture::SetNodeAttrs));
6396 assert!(!Format::Xml.is_authorable());
6397 assert!(!Format::Html.supports(Gesture::SetNodeAttrs));
6398 assert!(!Format::Djot.supports(Gesture::SetNodeAttrs));
6399
6400 let src = "<svg>\n <g id=\"a\">\n <rect x=\"1\" y=\"2\"/>\n </g>\n</svg>\n";
6401 let mut ed = Editor::new_str(src, Format::Xml).expect("editor");
6402 let rect = |ed: &mut Editor| {
6403 ed.nodes()
6404 .expect("nodes")
6405 .into_iter()
6406 .find(|n| n.name.as_deref() == Some("rect"))
6407 .expect("a rect")
6408 .id
6409 };
6410 let id = rect(&mut ed);
6411 let change = ed
6412 .set_node_attrs(
6413 id,
6414 &[("x", Some("10")), ("y", Some("2")), ("fill", Some("#f00"))],
6415 )
6416 .expect("set");
6417 assert_eq!(
6418 ed.source_str().unwrap(),
6419 "<svg>\n <g id=\"a\">\n <rect x=\"10\" y=\"2\" fill=\"#f00\"/>\n </g>\n</svg>\n"
6420 );
6421 // The change is the tag's interior alone: the `<g>` around it and the
6422 // whitespace runs are not re-printed.
6423 assert_eq!(change.old, 28..40);
6424 assert_eq!(change.new, 28..53);
6425 // Ids are the current tree's: read them again after an edit.
6426 let id = rect(&mut ed);
6427 assert_eq!(
6428 ed.nodes()
6429 .unwrap()
6430 .iter()
6431 .find(|n| n.id == id)
6432 .unwrap()
6433 .attrs,
6434 vec![
6435 ("x".to_string(), Some("10".to_string())),
6436 ("y".to_string(), Some("2".to_string())),
6437 ("fill".to_string(), Some("#f00".to_string())),
6438 ]
6439 );
6440 ed.set_node_attrs(id, &[]).expect("clear");
6441 assert_eq!(
6442 ed.source_str().unwrap(),
6443 "<svg>\n <g id=\"a\">\n <rect/>\n </g>\n</svg>\n"
6444 );
6445 // A text run is no element; an id past the tree is no node; a bare
6446 // attribute is one no format reads back; and a prose format refuses
6447 // before looking.
6448 let text = ed
6449 .nodes()
6450 .unwrap()
6451 .into_iter()
6452 .find(|n| n.kind == Kind::Str)
6453 .expect("a text run")
6454 .id;
6455 assert_eq!(
6456 ed.set_node_attrs(text, &[("a", Some("b"))]),
6457 Err(Error::NotEditable)
6458 );
6459 let past = NodeId(ed.nodes().unwrap().len() as u32 + 5);
6460 assert_eq!(
6461 ed.set_node_attrs(past, &[("a", Some("b"))]),
6462 Err(Error::InvalidArgument)
6463 );
6464 assert_eq!(
6465 ed.set_node_attrs(id, &[("hidden", None)]),
6466 Err(Error::InvalidArgument)
6467 );
6468 let mut md = Editor::new_str("hello\n", Format::Markdown).expect("editor");
6469 assert_eq!(
6470 md.set_node_attrs(NodeId(0), &[("a", Some("b"))]),
6471 Err(Error::UnsupportedFormat)
6472 );
6473 }
6474
6475 #[test]
6476 fn editor_set_block_attrs_spells_per_format_and_replaces_rather_than_merging() {
6477 // djot: the line before the block, rewritten in place; the block's
6478 // bytes are never touched.
6479 let mut dj = Editor::new_str("hello _em_\n", Format::Djot).expect("editor");
6480 dj.set_block_attrs(0, &[("class", Some("center"))]).expect("class");
6481 assert_eq!(dj.source_str().unwrap(), "{.center}\nhello _em_\n");
6482 dj.set_block_attrs(12, &[("data-size", Some("large"))]).expect("size");
6483 assert_eq!(dj.source_str().unwrap(), "{data-size=\"large\"}\nhello _em_\n");
6484 dj.set_block_attrs(22, &[]).expect("clear");
6485 assert_eq!(dj.source_str().unwrap(), "hello _em_\n");
6486
6487 // HTML: the tag; AsciiDoc: its attribute line.
6488 let mut html = Editor::new_str("<p>a</p>\n", Format::Html).expect("editor");
6489 html.set_block_attrs(4, &[("class", Some("c"))]).expect("class");
6490 assert_eq!(html.source_str().unwrap(), "<p class=\"c\">a</p>\n");
6491 let mut adoc = Editor::new_str("hello\n", Format::Asciidoc).expect("editor");
6492 adoc.set_block_attrs(0, &[("class", Some("c"))]).expect("class");
6493 assert_eq!(adoc.source_str().unwrap(), "[.c]\nhello\n");
6494
6495 // Markdown: refused without the flag; a div with it, rewritten rather
6496 // than nested on a second call, unwrapped by an empty list.
6497 assert!(!Format::Markdown.supports(Gesture::SetBlockAttrs));
6498 let exts = MarkdownExtensions {
6499 html_elements: true,
6500 ..Default::default()
6501 };
6502 assert!(Format::Markdown.supports_with(exts, Gesture::SetBlockAttrs));
6503 let mut plain = Editor::new_str("hello\n", Format::Markdown).expect("editor");
6504 assert_eq!(
6505 plain.set_block_attrs(0, &[("class", Some("c"))]),
6506 Err(Error::UnsupportedFormat)
6507 );
6508 let mut md = Editor::new_ext(b"hello\n", Format::Markdown, exts).expect("editor");
6509 md.set_block_attrs(0, &[("class", Some("c"))]).expect("class");
6510 assert_eq!(md.source_str().unwrap(), "<div class=\"c\">\n\nhello\n\n</div>\n");
6511 md.set_block_attrs(20, &[("class", Some("d"))]).expect("reclass");
6512 assert_eq!(md.source_str().unwrap(), "<div class=\"d\">\n\nhello\n\n</div>\n");
6513 md.set_block_attrs(20, &[]).expect("clear");
6514 assert_eq!(md.source_str().unwrap(), "hello\n");
6515
6516 // An attribute no format reads back is refused before anything is
6517 // written.
6518 assert_eq!(
6519 dj.set_block_attrs(0, &[("hidden", None)]),
6520 Err(Error::InvalidArgument)
6521 );
6522 assert_eq!(
6523 dj.set_block_attrs(0, &[("a b", Some("x"))]),
6524 Err(Error::InvalidArgument)
6525 );
6526 assert_eq!(dj.source_str().unwrap(), "hello _em_\n");
6527 }
6528
6529 #[test]
6530 fn editor_wrap_range_attrs_spells_a_span_and_re_styles_rather_than_nesting() {
6531 let mut dj = Editor::new_str("a big b\n", Format::Djot).expect("editor");
6532 dj.wrap_range_attrs(2, 5, &[("class", Some("large"))]).expect("wrap");
6533 assert_eq!(dj.source_str().unwrap(), "a [big]{.large} b\n");
6534 dj.wrap_range_attrs(3, 6, &[("class", Some("small"))]).expect("re-style");
6535 assert_eq!(dj.source_str().unwrap(), "a [big]{.small} b\n");
6536 dj.wrap_range_attrs(3, 6, &[]).expect("unwrap");
6537 assert_eq!(dj.source_str().unwrap(), "a big b\n");
6538
6539 let mut html = Editor::new_str("<p>a big b</p>\n", Format::Html).expect("editor");
6540 html.wrap_range_attrs(5, 8, &[("class", Some("large"))]).expect("wrap");
6541 assert_eq!(html.source_str().unwrap(), "<p>a <span class=\"large\">big</span> b</p>\n");
6542
6543 // Markdown reads the span back only under `html_elements`; AsciiDoc
6544 // has no spelling that keeps every key.
6545 assert!(!Format::Markdown.supports(Gesture::WrapRangeAttrs));
6546 assert!(!Format::Asciidoc.supports(Gesture::WrapRangeAttrs));
6547 let exts = MarkdownExtensions {
6548 html_elements: true,
6549 ..Default::default()
6550 };
6551 assert!(Format::Markdown.supports_with(exts, Gesture::WrapRangeAttrs));
6552 let mut md = Editor::new_ext(b"a big b\n", Format::Markdown, exts).expect("editor");
6553 md.wrap_range_attrs(2, 5, &[("class", Some("large"))]).expect("wrap");
6554 assert_eq!(md.source_str().unwrap(), "a <span class=\"large\">big</span> b\n");
6555 let mut plain = Editor::new_str("a big b\n", Format::Markdown).expect("editor");
6556 assert_eq!(
6557 plain.wrap_range_attrs(2, 5, &[("class", Some("large"))]),
6558 Err(Error::UnsupportedFormat)
6559 );
6560 }
6561
6562 #[test]
6563 fn editor_insert_directive_writes_a_directive_the_reparse_reads_back() {
6564 let exts = MarkdownExtensions {
6565 directives: true,
6566 ..Default::default()
6567 };
6568 let mut md = Editor::new_ext(b"a\n\nb\n", Format::Markdown, exts).expect("editor");
6569 md.insert_directive(0, "page-break", None, &[])
6570 .expect("directive");
6571 assert_eq!(md.source_str().unwrap(), "a\n\n::page-break\n\nb\n");
6572 // A directive to the parser, not a paragraph that starts with colons.
6573 assert_eq!(md.query("directive").expect("query").len(), 1);
6574
6575 // A label and an attribute, in the serializer's own spelling — the
6576 // value is quoted because that is what the Markdown serializer writes.
6577 let mut em = Editor::new_ext(b"a\n", Format::Markdown, exts).expect("editor");
6578 em.insert_directive(0, "embed", Some("Contents"), &[("src", Some("x.html"))])
6579 .expect("directive");
6580 assert_eq!(
6581 em.source_str().unwrap(),
6582 "a\n\n::embed[Contents]{src=\"x.html\"}\n"
6583 );
6584 assert_eq!(em.query("directive").expect("query").len(), 1);
6585
6586 // Djot's div is anonymous, so its spelling is an empty `:::` fence
6587 // carrying the name — which is why the gate asks about the NAME
6588 // surviving rather than about the syntax.
6589 let mut dj = Editor::new_str("a\n", Format::Djot).expect("editor");
6590 dj.insert_directive(0, "page-break", None, &[])
6591 .expect("directive");
6592 assert_eq!(dj.source_str().unwrap(), "a\n\n::: page-break\n:::\n");
6593
6594 // `None` is no label and `Some("")` is an empty one — two different
6595 // documents, which is the whole reason the parameter is an `Option`
6596 // rather than a `&str`.
6597 let mut none = Editor::new_ext(b"", Format::Markdown, exts).expect("editor");
6598 none.insert_directive(0, "x", None, &[]).expect("directive");
6599 assert_eq!(none.source_str().unwrap(), "::x\n");
6600 let mut empty = Editor::new_ext(b"", Format::Markdown, exts).expect("editor");
6601 empty
6602 .insert_directive(0, "x", Some(""), &[])
6603 .expect("directive");
6604 assert_eq!(empty.source_str().unwrap(), "::x[]\n");
6605 }
6606
6607 #[test]
6608 fn editor_insert_directive_needs_the_extension_it_will_be_read_back_with() {
6609 // Without the flag `::page-break` is a paragraph of colons, so the
6610 // gesture refuses rather than minting bytes one press cannot undo.
6611 let mut plain = Editor::new_str("a\n", Format::Markdown).expect("editor");
6612 assert_eq!(
6613 plain.insert_directive(0, "page-break", None, &[]),
6614 Err(Error::UnsupportedFormat)
6615 );
6616 assert_eq!(plain.source_str().unwrap(), "a\n");
6617
6618 assert!(!Format::Markdown.supports(Gesture::InsertDirective));
6619 assert!(Format::Markdown.supports_with(
6620 MarkdownExtensions {
6621 directives: true,
6622 ..Default::default()
6623 },
6624 Gesture::InsertDirective
6625 ));
6626 // Djot needs no extension; XML spells nothing at all.
6627 assert!(Format::Djot.supports(Gesture::InsertDirective));
6628 assert!(!Format::Xml.supports(Gesture::InsertDirective));
6629
6630 // An empty name is the caller's error, not a nameless directive.
6631 let mut ext = Editor::new_ext(
6632 b"a\n",
6633 Format::Markdown,
6634 MarkdownExtensions {
6635 directives: true,
6636 ..Default::default()
6637 },
6638 )
6639 .expect("editor");
6640 assert_eq!(
6641 ext.insert_directive(0, "", None, &[]),
6642 Err(Error::InvalidArgument)
6643 );
6644 assert_eq!(
6645 ext.insert_directive(0, "a\nb", None, &[]),
6646 Err(Error::InvalidArgument)
6647 );
6648 // A name goes where a delimiter would otherwise be, so the grammar is
6649 // checked rather than left to the serializer: `::a b` would reparse as
6650 // a paragraph holding an inline directive named `a`.
6651 assert_eq!(
6652 ext.insert_directive(0, "a b", None, &[]),
6653 Err(Error::InvalidArgument)
6654 );
6655 // And a bracket in the label closes the `[…]` early.
6656 assert_eq!(
6657 ext.insert_directive(0, "page-break", Some("]"), &[]),
6658 Err(Error::InvalidArgument)
6659 );
6660 assert_eq!(ext.source_str().unwrap(), "a\n");
6661 }
6662
6663 #[test]
6664 fn document_html_elements_make_embedded_img_queryable() {
6665 let src = "text <img src=\"a.png\" alt=\"x\"> more\n";
6666 // Without the flag, the `<img>` is opaque raw HTML — no `image` node.
6667 let mut plain = Document::parse_str(src, Format::Markdown).expect("parse");
6668 assert_eq!(plain.query("image").expect("query").len(), 0);
6669 // With it enabled on the read path, the promoted image is queryable.
6670 let mut ext = Document::parse_str_with(
6671 src,
6672 Format::Markdown,
6673 MarkdownExtensions {
6674 html_elements: true,
6675 ..Default::default()
6676 },
6677 )
6678 .expect("parse");
6679 let images = ext.query("image").expect("query");
6680 assert_eq!(images.len(), 1);
6681 assert_eq!(images[0].kind, Kind::Image);
6682 }
6683
6684 #[test]
6685 fn editor_filter_public_audience_view() {
6686 let src = "# Archive\n\n:::vis{.public}\nPublic.\n:::\n\n:::vis{.family}\nPrivate.\n:::\n";
6687 let mut ed = Editor::new_ext(
6688 src.as_bytes(),
6689 Format::Markdown,
6690 MarkdownExtensions {
6691 directives: true,
6692 ..Default::default()
6693 },
6694 )
6695 .expect("editor");
6696 // Drop every vis block except the public one, then unwrap it.
6697 ed.filter(
6698 "directive[name=vis]",
6699 Some("directive[class~=public]"),
6700 true,
6701 )
6702 .expect("filter");
6703 assert_eq!(ed.source_str().expect("source"), "# Archive\n\nPublic.\n");
6704 }
6705
6706 #[test]
6707 fn editor_filter_rejects_a_malformed_selector() {
6708 let mut ed = Editor::new_str("hi\n", Format::Markdown).expect("editor");
6709 assert_eq!(
6710 ed.filter("list >", None, false),
6711 Err(Error::InvalidArgument)
6712 );
6713 }
6714
6715 #[test]
6716 fn builder_builds_and_renders_a_document() {
6717 let mut b = Builder::new().expect("builder");
6718
6719 // # Title\n\nhello *world*
6720 let title = b.add_text(TextKind::Str, "Title").unwrap();
6721 let heading = b.add_heading(1).unwrap();
6722 b.set_children(heading, &[title]).unwrap();
6723
6724 let hello = b.add_text(TextKind::Str, "hello ").unwrap();
6725 let world = b.add_text(TextKind::Str, "world").unwrap();
6726 let emph = b.add(VoidKind::Emph).unwrap();
6727 b.set_children(emph, &[world]).unwrap();
6728 let para = b.add(VoidKind::Para).unwrap();
6729 b.set_children(para, &[hello, emph]).unwrap();
6730
6731 let doc = b.add(VoidKind::Doc).unwrap();
6732 b.set_children(doc, &[heading, para]).unwrap();
6733
6734 let html = String::from_utf8(b.render_html(doc).unwrap()).unwrap();
6735 assert!(html.contains("<h1>Title</h1>"), "{html}");
6736 assert!(html.contains("<em>world</em>"), "{html}");
6737
6738 let md = String::from_utf8(b.serialize(doc, Format::Markdown).unwrap()).unwrap();
6739 assert!(md.contains("# Title"), "{md}");
6740 assert!(md.contains("*world*"), "{md}");
6741
6742 let matches = b.query(doc, "heading").unwrap();
6743 assert_eq!(matches.len(), 1);
6744 assert_eq!(matches[0].kind, Kind::Heading);
6745
6746 let json = String::from_utf8(b.ast_json(doc).unwrap()).unwrap();
6747 assert!(json.contains("\"kind\": \"doc\""), "{json}");
6748 }
6749
6750 #[test]
6751 fn builder_element_with_attributes() {
6752 let mut b = Builder::new().expect("builder");
6753 let inner = b.add_text(TextKind::Str, "hi").unwrap();
6754 let el = b.add_element("section").unwrap();
6755 b.set_children(el, &[inner]).unwrap();
6756 b.set_attrs(el, &[("class", Some("note")), ("hidden", None)])
6757 .unwrap();
6758
6759 let html = String::from_utf8(b.render_html(el).unwrap()).unwrap();
6760 assert!(html.contains("<section"), "{html}");
6761 assert!(html.contains("class=\"note\""), "{html}");
6762 assert!(html.contains("hidden"), "{html}");
6763 }
6764
6765 #[test]
6766 fn builder_lists_round_trip_to_markdown() {
6767 let mut b = Builder::new().expect("builder");
6768
6769 // An ordered list: 1. one / 2. two
6770 let one_txt = b.add_text(TextKind::Str, "one").unwrap();
6771 let one_para = b.add(VoidKind::Para).unwrap();
6772 b.set_children(one_para, &[one_txt]).unwrap();
6773 let one = b.add(VoidKind::ListItem).unwrap();
6774 b.set_children(one, &[one_para]).unwrap();
6775
6776 let two_txt = b.add_text(TextKind::Str, "two").unwrap();
6777 let two_para = b.add(VoidKind::Para).unwrap();
6778 b.set_children(two_para, &[two_txt]).unwrap();
6779 let two = b.add(VoidKind::ListItem).unwrap();
6780 b.set_children(two, &[two_para]).unwrap();
6781
6782 let list = b
6783 .add_ordered_list(
6784 OrderedNumbering::Decimal,
6785 OrderedDelim::Period,
6786 true,
6787 Some(1),
6788 )
6789 .unwrap();
6790 b.set_children(list, &[one, two]).unwrap();
6791 let doc = b.add(VoidKind::Doc).unwrap();
6792 b.set_children(doc, &[list]).unwrap();
6793
6794 let md = String::from_utf8(b.serialize(doc, Format::Markdown).unwrap()).unwrap();
6795 assert!(md.contains("1. one"), "{md}");
6796 assert!(md.contains("2. two"), "{md}");
6797 }
6798
6799 #[test]
6800 fn builder_rejects_invalid_kind_and_id() {
6801 let b = Builder::new().expect("builder");
6802 // `heading` (code 2) carries a payload, so the void-kind `add` rejects it
6803 // — the safe `VoidKind` enum has no such variant, so we go through the raw
6804 // ABI to prove the guard.
6805 let mut id = 0u32;
6806 let status = unsafe { ffi::twig_builder_add(b.raw.as_ptr(), 2, &mut id) };
6807 assert_eq!(Error::from_status(status), Err(Error::InvalidArgument));
6808
6809 // A root id past the end can't be rendered.
6810 let mut ptr = std::ptr::null();
6811 let mut len = 0usize;
6812 let status =
6813 unsafe { ffi::twig_builder_render_html(b.raw.as_ptr(), 4242, &mut ptr, &mut len) };
6814 assert_eq!(Error::from_status(status), Err(Error::InvalidArgument));
6815 }
6816
6817 // ── Format capability ───────────────────────────────────────────────────
6818
6819 /// Every gesture with a format-level gate, both kind vocabularies in full.
6820 fn all_gestures() -> Vec<Gesture> {
6821 let inline = [
6822 InlineKind::Strong,
6823 InlineKind::Emph,
6824 InlineKind::Verbatim,
6825 InlineKind::Mark,
6826 InlineKind::Superscript,
6827 InlineKind::Subscript,
6828 InlineKind::Insert,
6829 InlineKind::Delete,
6830 ];
6831 let mut all: Vec<Gesture> = Vec::new();
6832 for k in inline {
6833 all.push(Gesture::WrapRange(k));
6834 all.push(Gesture::ToggleInline(k));
6835 }
6836 for k in [
6837 BlockContainerKind::BlockQuote,
6838 BlockContainerKind::BulletList,
6839 BlockContainerKind::OrderedList,
6840 ] {
6841 all.push(Gesture::ToggleBlockContainer(k));
6842 }
6843 all.extend([
6844 Gesture::SetMarkColor,
6845 Gesture::SetBlock,
6846 Gesture::InsertThematicBreak,
6847 Gesture::ToggleCodeBlock,
6848 Gesture::SetCodeLanguage,
6849 Gesture::ToggleTaskItem,
6850 Gesture::SetTaskChecked,
6851 Gesture::ToggleTaskChecked,
6852 Gesture::InsertLink,
6853 Gesture::InsertImage,
6854 Gesture::InsertFootnote,
6855 Gesture::InsertLiteral,
6856 Gesture::InsertLineBreak,
6857 Gesture::SplitBlock,
6858 Gesture::RenumberOrderedLists,
6859 Gesture::TableInsertRow,
6860 Gesture::TableDeleteRow,
6861 Gesture::TableInsertColumn,
6862 Gesture::TableDeleteColumn,
6863 Gesture::TableSetAlignment,
6864 Gesture::TableMoveRow,
6865 Gesture::TableMoveColumn,
6866 Gesture::InsertTable,
6867 Gesture::InsertDirective,
6868 Gesture::SetBlockAttrs,
6869 Gesture::WrapRangeAttrs,
6870 Gesture::JoinBlocks,
6871 Gesture::SetNodeAttrs,
6872 Gesture::MoveBlock,
6873 ]);
6874 all
6875 }
6876
6877 #[test]
6878 fn the_wire_space_ends_where_the_sweep_does() {
6879 // `all_gestures` is hand-written and, unlike the Zig union it mirrors,
6880 // has no compile-time cross-check: a variant added to the enum and to
6881 // `to_c` can silently miss the sweep below. So pin the space from both
6882 // ends — the sweep must cover a contiguous range of codes, every one of
6883 // them must decode C-side, and one past the end must not.
6884 let mut codes: Vec<c_int> = all_gestures().iter().map(|g| g.to_c().0).collect();
6885 codes.sort_unstable();
6886 codes.dedup();
6887 assert_eq!(codes, (0..=31).collect::<Vec<c_int>>());
6888
6889 let mut supported = -1;
6890 for code in &codes {
6891 let status = unsafe {
6892 ffi::twig_format_supports(
6893 ffi::TwigFormat::from(Format::Markdown) as c_int,
6894 *code,
6895 0,
6896 &mut supported,
6897 )
6898 };
6899 assert_eq!(Error::from_status(status), Ok(()), "code {code} did not decode");
6900 }
6901 // One past the end is not a gesture, which is what makes appending safe.
6902 let status = unsafe {
6903 ffi::twig_format_supports(
6904 ffi::TwigFormat::from(Format::Markdown) as c_int,
6905 32,
6906 0,
6907 &mut supported,
6908 )
6909 };
6910 assert_eq!(Error::from_status(status), Err(Error::InvalidArgument));
6911 }
6912
6913 #[test]
6914 fn supports_answers_per_gesture_where_authorable_cannot() {
6915 // HTML is why the per-gesture query exists. `is_authorable` is true for
6916 // it — it spells the inline marks, and every block its parser reads
6917 // back through a renderer — while a toolbar built on that predicate
6918 // would show a task-box button and a footnote button that both fail.
6919 assert!(Format::Html.is_authorable());
6920 assert!(Format::Html.supports(Gesture::ToggleInline(InlineKind::Strong)));
6921 assert!(Format::Html.supports(Gesture::SetBlock));
6922 assert!(Format::Html.supports(Gesture::InsertLiteral));
6923 assert!(Format::Html.supports(Gesture::ToggleBlockContainer(
6924 BlockContainerKind::BlockQuote
6925 )));
6926 assert!(Format::Html.supports(Gesture::ToggleCodeBlock));
6927 assert!(Format::Html.supports(Gesture::InsertLink));
6928 assert!(!Format::Html.supports(Gesture::ToggleTaskItem));
6929 assert!(!Format::Html.supports(Gesture::InsertFootnote));
6930 // The nine that used to answer nothing at all: HTML has a table its
6931 // parser reads and no spelling to write one back with, no blank-line
6932 // block separation, and no numbered list marker.
6933 assert!(!Format::Html.supports(Gesture::TableInsertRow));
6934 assert!(!Format::Html.supports(Gesture::TableSetAlignment));
6935 assert!(!Format::Html.supports(Gesture::SplitBlock));
6936 // And the one that goes the other way, which is why the join has a gate
6937 // of its own: HTML cannot be split at a blank line and *can* be joined
6938 // at a newline inside its `<p>`.
6939 assert!(Format::Html.supports(Gesture::JoinBlocks));
6940 // And the block move, whose lines HTML has like any other format.
6941 assert!(Format::Html.supports(Gesture::MoveBlock));
6942 assert!(!Format::Html.supports(Gesture::RenumberOrderedLists));
6943 assert!(Format::Markdown.supports(Gesture::TableInsertRow));
6944 assert!(Format::Djot.supports(Gesture::SplitBlock));
6945
6946 // A format that spells no prose answers false to every caret gesture,
6947 // so the coarse predicate agrees there — it only misleads in the middle
6948 // of the range. The one gesture XML (and its svg dialect) supports is
6949 // the node-addressed one, which no caret asks for and `is_authorable`
6950 // deliberately does not count.
6951 for fmt in [Format::Xml, Format::Svg] {
6952 assert!(!fmt.is_authorable());
6953 for g in all_gestures() {
6954 assert_eq!(
6955 fmt.supports(g),
6956 g == Gesture::SetNodeAttrs,
6957 "{fmt:?} on {g:?}"
6958 );
6959 }
6960 }
6961 // AsciiDoc is in the middle of the range the other way round from
6962 // HTML: the block gestures work, a link prints through its renderer,
6963 // the footnote/table shapes don't.
6964 assert!(Format::Asciidoc.is_authorable());
6965 assert!(Format::Asciidoc.supports(Gesture::SetBlock));
6966 assert!(Format::Asciidoc.supports(Gesture::ToggleInline(InlineKind::Mark)));
6967 assert!(Format::Asciidoc.supports(Gesture::InsertLink));
6968 assert!(!Format::Asciidoc.supports(Gesture::InsertFootnote));
6969 assert!(!Format::Asciidoc.supports(Gesture::TableInsertRow));
6970
6971 // And the two authorable formats differ from each other, which is the
6972 // other half of why one boolean can't serve.
6973 assert!(Format::Djot.supports(Gesture::ToggleInline(InlineKind::Mark)));
6974 assert!(!Format::Markdown.supports(Gesture::ToggleInline(InlineKind::Mark)));
6975 assert!(Format::Markdown.supports(Gesture::InsertLineBreak));
6976 assert!(!Format::Djot.supports(Gesture::InsertLineBreak));
6977 }
6978
6979 #[test]
6980 fn supports_agrees_with_what_the_editor_then_does() {
6981 // The pin at this layer: for the gestures whose refusal an `Editor`
6982 // can be made to demonstrate, the query's answer is the call's answer.
6983 // Zig covers the full (format x gesture) sweep; what's checked here is
6984 // that the Rust decode reaches the same question.
6985 for fmt in [Format::Djot, Format::Markdown, Format::Html] {
6986 let mut ed = Editor::new_str("ab\n", fmt).expect("editor");
6987 let claimed = fmt.supports(Gesture::ToggleInline(InlineKind::Mark));
6988 let observed = ed.toggle_inline(0, 2, InlineKind::Mark);
6989 assert_eq!(
6990 claimed,
6991 !matches!(observed, Err(Error::UnsupportedFormat)),
6992 "{fmt:?}: supports said {claimed}, gesture said {observed:?}",
6993 );
6994
6995 let mut ed = Editor::new_str("ab\n", fmt).expect("editor");
6996 let claimed = fmt.supports(Gesture::SetBlock);
6997 let observed = ed.set_block(0, BlockKind::Heading(1));
6998 assert_eq!(
6999 claimed,
7000 !matches!(observed, Err(Error::UnsupportedFormat)),
7001 "{fmt:?}: supports said {claimed}, gesture said {observed:?}",
7002 );
7003 }
7004
7005 // The destructive one, spelled out: an HTML `<table>` extracts as a grid
7006 // and cannot be written back, so the refusal has to arrive before
7007 // anything is spliced. A `Ok(())` here once meant a destroyed table.
7008 let src = "<table><tr><td>a</td></tr></table>";
7009 let mut ed = Editor::new_str(src, Format::Html).expect("editor");
7010 assert!(!Format::Html.supports(Gesture::TableInsertRow));
7011 assert_eq!(ed.table_insert_row(15, true), Err(Error::UnsupportedFormat));
7012 assert_eq!(ed.renumber_ordered_lists(15), Err(Error::UnsupportedFormat));
7013 assert!(matches!(ed.split_block(15), Err(Error::UnsupportedFormat)));
7014 // The join is spelled in HTML, so its refusal here is about the caret
7015 // being in a table — a different answer, from a different gate.
7016 assert_eq!(ed.join_blocks(15), Err(Error::NotEditable));
7017 assert_eq!(ed.source().expect("source"), src.as_bytes());
7018 }
7019
7020 #[test]
7021 fn supports_rides_the_gestures_own_kind_space() {
7022 // The same integer means different things per gesture on the wire (1 is
7023 // `emph` inline and `bullet_list` container). The Rust types make that
7024 // unrepresentable, which is why `supports` returns a bare bool — but
7025 // the raw call underneath still has to be handed the right pair.
7026 let (g, k) = Gesture::ToggleBlockContainer(BlockContainerKind::BulletList).to_c();
7027 assert_eq!((g, k), (3, 1));
7028 let (g, k) = Gesture::ToggleInline(InlineKind::Emph).to_c();
7029 assert_eq!((g, k), (1, 1));
7030 // A kindless gesture sends 0, which the C side requires rather than
7031 // ignores.
7032 assert_eq!(Gesture::InsertLink.to_c(), (10, 0));
7033
7034 // And the C side does reject the combinations Rust can't build.
7035 let mut out: c_int = 0;
7036 let status = unsafe {
7037 ffi::twig_format_supports(ffi::TwigFormat::Markdown as c_int, 10, 3, &mut out)
7038 };
7039 assert_eq!(Error::from_status(status), Err(Error::InvalidArgument));
7040 let status = unsafe {
7041 ffi::twig_format_supports(ffi::TwigFormat::Markdown as c_int, 9999, 0, &mut out)
7042 };
7043 assert_eq!(Error::from_status(status), Err(Error::InvalidArgument));
7044 }
7045}