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