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