@parhelia/core 0.4.12825 → 0.4.12872

package/dist/editor/spellcheck/htmlSpellDocument.js

@@ -0,0 +1,181 @@
+import { decode, encode } from "html-entities";
+import { tokenizeHtml } from "./htmlSpellTokenizer";
+const ENTITY_PATTERN = /&(?:#[xX][0-9a-fA-F]+;?|#[0-9]+;?|[a-zA-Z][a-zA-Z0-9]*;?)/g;
+function decodeText(raw) {
+    return decode(raw, { level: "html5", scope: "body" });
+}
+/** `\r\n` and lone `\r` become `\n`, matching the HTML input stream preprocessing. */
+function normalizeNewlines(text) {
+    return text.replace(/\r\n?/g, "\n");
+}
+/**
+ * Decodes `raw` run by run (plain characters, entities, CRLF) and records the
+ * source range of every run. Returns undefined when the run-wise decode does
+ * not reproduce the authoritative decode, so callers never map offsets
+ * through an inconsistent table.
+ */
+function buildRuns(raw, decoded) {
+    const runs = [];
+    let out = "";
+    let src = 0;
+    let plainStart = -1;
+    const closePlain = (end) => {
+        if (plainStart < 0)
+            return;
+        runs.push({
+            decStart: out.length,
+            decEnd: out.length + (end - plainStart),
+            srcStart: plainStart,
+            srcEnd: end,
+        });
+        out += raw.slice(plainStart, end);
+        plainStart = -1;
+    };
+    while (src < raw.length) {
+        const code = raw.charCodeAt(src);
+        if (code === 13 /* \r */) {
+            closePlain(src);
+            const end = raw.charCodeAt(src + 1) === 10 ? src + 2 : src + 1;
+            runs.push({ decStart: out.length, decEnd: out.length + 1, srcStart: src, srcEnd: end });
+            out += "\n";
+            src = end;
+            continue;
+        }
+        if (code === 38 /* & */) {
+            ENTITY_PATTERN.lastIndex = src;
+            const match = ENTITY_PATTERN.exec(raw);
+            if (match && match.index === src) {
+                const entityText = decodeText(match[0]);
+                if (entityText !== match[0]) {
+                    closePlain(src);
+                    runs.push({
+                        decStart: out.length,
+                        decEnd: out.length + entityText.length,
+                        srcStart: src,
+                        srcEnd: src + match[0].length,
+                    });
+                    out += entityText;
+                    src += match[0].length;
+                    continue;
+                }
+            }
+        }
+        if (plainStart < 0)
+            plainStart = src;
+        src++;
+    }
+    closePlain(raw.length);
+    return out === decoded ? runs : undefined;
+}
+export function parseHtmlSpellDocument(source) {
+    const tokens = tokenizeHtml(source);
+    const segments = [];
+    const separators = [];
+    let text = "";
+    const separator = () => {
+        separators.push(text.length);
+        text += "\n";
+    };
+    for (const token of tokens) {
+        if (token.kind === "text") {
+            const raw = source.slice(token.start, token.end);
+            const decoded = normalizeNewlines(decodeText(raw));
+            const runs = buildRuns(raw, decoded);
+            segments.push({
+                kind: "text",
+                sourceStart: token.start,
+                sourceEnd: token.end,
+                decoded,
+                spellStart: text.length,
+                spellEnd: text.length + decoded.length,
+                runs,
+            });
+            text += decoded;
+            continue;
+        }
+        if (token.kind === "rawText") {
+            const raw = source.slice(token.start, token.end);
+            const content = normalizeNewlines(token.decodesEntities ? decodeText(raw) : raw);
+            const masked = " ".repeat(content.length);
+            segments.push({
+                kind: "masked",
+                sourceStart: token.start,
+                sourceEnd: token.end,
+                decoded: masked,
+                spellStart: text.length,
+                spellEnd: text.length + masked.length,
+            });
+            text += masked;
+            continue;
+        }
+        if (token.kind === "startTag") {
+            if (token.isBlock)
+                separator();
+            continue;
+        }
+        if (token.kind === "endTag") {
+            if (token.isBlock && !token.isVoid)
+                separator();
+        }
+    }
+    return { source, text, separators, segments, tokens };
+}
+/**
+ * Converts a spell-text offset to the matching `element.textContent` offset by
+ * discounting the separators inserted before it.
+ */
+export function toTextContentOffset(offset, separators) {
+    let count = 0;
+    for (const separator of separators) {
+        if (separator < offset)
+            count++;
+        else
+            break;
+    }
+    return offset - count;
+}
+/** Source offset of a decoded offset inside a segment, or undefined inside a multi-character decode. */
+function toSourceOffset(segment, decodedOffset) {
+    const runs = segment.runs;
+    if (!runs)
+        return undefined;
+    if (decodedOffset === segment.decoded.length)
+        return segment.sourceEnd;
+    // Run offsets are relative to the segment's raw text.
+    const base = segment.sourceStart;
+    for (const run of runs) {
+        if (decodedOffset < run.decStart || decodedOffset >= run.decEnd)
+            continue;
+        if (decodedOffset === run.decStart)
+            return base + run.srcStart;
+        const oneToOne = run.decEnd - run.decStart === run.srcEnd - run.srcStart;
+        return oneToOne ? base + run.srcStart + (decodedOffset - run.decStart) : undefined;
+    }
+    return decodedOffset === 0 ? segment.sourceStart : undefined;
+}
+/**
+ * Replaces `[start, end)` of the spell text, which must lie inside one text
+ * segment and equal `word`, and returns the new source value. Everything
+ * outside the replaced word is copied from the source byte for byte; the
+ * replacement itself is entity-encoded. Returns undefined when the range spans
+ * a tag boundary, does not match `word`, or falls inside an entity.
+ */
+export function replaceInDocument(document, start, end, word, replacement) {
+    if (end <= start)
+        return undefined;
+    const segment = document.segments.find((candidate) => candidate.kind === "text" && candidate.spellStart <= start && end <= candidate.spellEnd);
+    if (!segment)
+        return undefined;
+    const localStart = start - segment.spellStart;
+    const localEnd = end - segment.spellStart;
+    if (segment.decoded.slice(localStart, localEnd) !== word)
+        return undefined;
+    const sourceStart = toSourceOffset(segment, localStart);
+    const sourceEnd = toSourceOffset(segment, localEnd);
+    if (sourceStart === undefined || sourceEnd === undefined)
+        return undefined;
+    return (document.source.slice(0, sourceStart) +
+        encode(replacement, { mode: "specialChars" }) +
+        document.source.slice(sourceEnd));
+}
+//# sourceMappingURL=htmlSpellDocument.js.map

package/dist/editor/spellcheck/htmlSpellDocument.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"htmlSpellDocument.js","sourceRoot":"","sources":["../../../src/editor/spellcheck/htmlSpellDocument.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,eAAe,CAAC;AAC/C,OAAO,EAAE,YAAY,EAAkB,MAAM,sBAAsB,CAAC;AAkDpE,MAAM,cAAc,GAAG,4DAA4D,CAAC;AAEpF,SAAS,UAAU,CAAC,GAAW;IAC7B,OAAO,MAAM,CAAC,GAAG,EAAE,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC;AACxD,CAAC;AAED,sFAAsF;AACtF,SAAS,iBAAiB,CAAC,IAAY;IACrC,OAAO,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;AACtC,CAAC;AAED;;;;;GAKG;AACH,SAAS,SAAS,CAAC,GAAW,EAAE,OAAe;IAC7C,MAAM,IAAI,GAAiB,EAAE,CAAC;IAC9B,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,IAAI,UAAU,GAAG,CAAC,CAAC,CAAC;IAEpB,MAAM,UAAU,GAAG,CAAC,GAAW,EAAE,EAAE;QACjC,IAAI,UAAU,GAAG,CAAC;YAAE,OAAO;QAC3B,IAAI,CAAC,IAAI,CAAC;YACR,QAAQ,EAAE,GAAG,CAAC,MAAM;YACpB,MAAM,EAAE,GAAG,CAAC,MAAM,GAAG,CAAC,GAAG,GAAG,UAAU,CAAC;YACvC,QAAQ,EAAE,UAAU;YACpB,MAAM,EAAE,GAAG;SACZ,CAAC,CAAC;QACH,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC;QAClC,UAAU,GAAG,CAAC,CAAC,CAAC;IAClB,CAAC,CAAC;IAEF,OAAO,GAAG,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACxB,MAAM,IAAI,GAAG,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;QACjC,IAAI,IAAI,KAAK,EAAE,CAAC,QAAQ,EAAE,CAAC;YACzB,UAAU,CAAC,GAAG,CAAC,CAAC;YAChB,MAAM,GAAG,GAAG,GAAG,CAAC,UAAU,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC;YAC/D,IAAI,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,GAAG,CAAC,EAAE,QAAQ,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;YACxF,GAAG,IAAI,IAAI,CAAC;YACZ,GAAG,GAAG,GAAG,CAAC;YACV,SAAS;QACX,CAAC;QACD,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO,EAAE,CAAC;YACxB,cAAc,CAAC,SAAS,GAAG,GAAG,CAAC;YAC/B,MAAM,KAAK,GAAG,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACvC,IAAI,KAAK,IAAI,KAAK,CAAC,KAAK,KAAK,GAAG,EAAE,CAAC;gBACjC,MAAM,UAAU,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;gBACxC,IAAI,UAAU,KAAK,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;oBAC5B,UAAU,CAAC,GAAG,CAAC,CAAC;oBAChB,IAAI,CAAC,IAAI,CAAC;wBACR,QAAQ,EAAE,GAAG,CAAC,MAAM;wBACpB,MAAM,EAAE,GAAG,CAAC,MAAM,GAAG,UAAU,CAAC,MAAM;wBACtC,QAAQ,EAAE,GAAG;wBACb,MAAM,EAAE,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM;qBAC9B,CAAC,CAAC;oBACH,GAAG,IAAI,UAAU,CAAC;oBAClB,GAAG,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;oBACvB,SAAS;gBACX,CAAC;YACH,CAAC;QACH,CAAC;QACD,IAAI,UAAU,GAAG,CAAC;YAAE,UAAU,GAAG,GAAG,CAAC;QACrC,GAAG,EAAE,CAAC;IACR,CAAC;IACD,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IAEvB,OAAO,GAAG,KAAK,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC;AAC5C,CAAC;AAED,MAAM,UAAU,sBAAsB,CAAC,MAAc;IACnD,MAAM,MAAM,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;IACpC,MAAM,QAAQ,GAAuB,EAAE,CAAC;IACxC,MAAM,UAAU,GAAa,EAAE,CAAC;IAChC,IAAI,IAAI,GAAG,EAAE,CAAC;IAEd,MAAM,SAAS,GAAG,GAAG,EAAE;QACrB,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC7B,IAAI,IAAI,IAAI,CAAC;IACf,CAAC,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC1B,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;YACjD,MAAM,OAAO,GAAG,iBAAiB,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC;YACnD,MAAM,IAAI,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YACrC,QAAQ,CAAC,IAAI,CAAC;gBACZ,IAAI,EAAE,MAAM;gBACZ,WAAW,EAAE,KAAK,CAAC,KAAK;gBACxB,SAAS,EAAE,KAAK,CAAC,GAAG;gBACpB,OAAO;gBACP,UAAU,EAAE,IAAI,CAAC,MAAM;gBACvB,QAAQ,EAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM;gBACtC,IAAI;aACL,CAAC,CAAC;YACH,IAAI,IAAI,OAAO,CAAC;YAChB,SAAS;QACX,CAAC;QACD,IAAI,KAAK,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;YACjD,MAAM,OAAO,GAAG,iBAAiB,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACjF,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;YAC1C,QAAQ,CAAC,IAAI,CAAC;gBACZ,IAAI,EAAE,QAAQ;gBACd,WAAW,EAAE,KAAK,CAAC,KAAK;gBACxB,SAAS,EAAE,KAAK,CAAC,GAAG;gBACpB,OAAO,EAAE,MAAM;gBACf,UAAU,EAAE,IAAI,CAAC,MAAM;gBACvB,QAAQ,EAAE,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM;aACtC,CAAC,CAAC;YACH,IAAI,IAAI,MAAM,CAAC;YACf,SAAS;QACX,CAAC;QACD,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU,EAAE,CAAC;YAC9B,IAAI,KAAK,CAAC,OAAO;gBAAE,SAAS,EAAE,CAAC;YAC/B,SAAS;QACX,CAAC;QACD,IAAI,KAAK,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;YAC5B,IAAI,KAAK,CAAC,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM;gBAAE,SAAS,EAAE,CAAC;QAClD,CAAC;IACH,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC;AACxD,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,MAAc,EAAE,UAAoB;IACtE,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,IAAI,SAAS,GAAG,MAAM;YAAE,KAAK,EAAE,CAAC;;YAC3B,MAAM;IACb,CAAC;IACD,OAAO,MAAM,GAAG,KAAK,CAAC;AACxB,CAAC;AAED,wGAAwG;AACxG,SAAS,cAAc,CAAC,OAAyB,EAAE,aAAqB;IACtE,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;IAC1B,IAAI,CAAC,IAAI;QAAE,OAAO,SAAS,CAAC;IAC5B,IAAI,aAAa,KAAK,OAAO,CAAC,OAAO,CAAC,MAAM;QAAE,OAAO,OAAO,CAAC,SAAS,CAAC;IACvE,sDAAsD;IACtD,MAAM,IAAI,GAAG,OAAO,CAAC,WAAW,CAAC;IACjC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,IAAI,aAAa,GAAG,GAAG,CAAC,QAAQ,IAAI,aAAa,IAAI,GAAG,CAAC,MAAM;YAAE,SAAS;QAC1E,IAAI,aAAa,KAAK,GAAG,CAAC,QAAQ;YAAE,OAAO,IAAI,GAAG,GAAG,CAAC,QAAQ,CAAC;QAC/D,MAAM,QAAQ,GAAG,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC,QAAQ,KAAK,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC;QACzE,OAAO,QAAQ,CAAC,CAAC,CAAC,IAAI,GAAG,GAAG,CAAC,QAAQ,GAAG,CAAC,aAAa,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACrF,CAAC;IACD,OAAO,aAAa,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,SAAS,CAAC;AAC/D,CAAC;AAED;;;;;;GAMG;AACH,MAAM,UAAU,iBAAiB,CAC/B,QAA2B,EAC3B,KAAa,EACb,GAAW,EACX,IAAY,EACZ,WAAmB;IAEnB,IAAI,GAAG,IAAI,KAAK;QAAE,OAAO,SAAS,CAAC;IACnC,MAAM,OAAO,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,CACpC,CAAC,SAAS,EAAE,EAAE,CACZ,SAAS,CAAC,IAAI,KAAK,MAAM,IAAI,SAAS,CAAC,UAAU,IAAI,KAAK,IAAI,GAAG,IAAI,SAAS,CAAC,QAAQ,CAC1F,CAAC;IACF,IAAI,CAAC,OAAO;QAAE,OAAO,SAAS,CAAC;IAC/B,MAAM,UAAU,GAAG,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC;IAC9C,MAAM,QAAQ,GAAG,GAAG,GAAG,OAAO,CAAC,UAAU,CAAC;IAC1C,IAAI,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,CAAC,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAC3E,MAAM,WAAW,GAAG,cAAc,CAAC,OAAO,EAAE,UAAU,CAAC,CAAC;IACxD,MAAM,SAAS,GAAG,cAAc,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;IACpD,IAAI,WAAW,KAAK,SAAS,IAAI,SAAS,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3E,OAAO,CACL,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,WAAW,CAAC;QACrC,MAAM,CAAC,WAAW,EAAE,EAAE,IAAI,EAAE,cAAc,EAAE,CAAC;QAC7C,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,CACjC,CAAC;AACJ,CAAC"}

package/dist/editor/spellcheck/htmlSpellTokenizer.d.ts

@@ -0,0 +1,45 @@
+/**
+ * A small HTML tokenizer that preserves source locations. It follows the HTML
+ * tokenizer states for the subset spell checking needs: data, tags with
+ * double-quoted / single-quoted / unquoted attribute values, self-closing
+ * tags, comments, doctype / bogus comments and the raw-text (script, style)
+ * and RCDATA (textarea, title) elements. `>` inside a quoted attribute value
+ * never ends a tag, so attribute content can never leak into the text.
+ *
+ * Every token carries `[start, end)` into the original string, which lets
+ * callers rebuild the source byte-identically around a replaced text token.
+ */
+export type HtmlTagToken = {
+    kind: "startTag" | "endTag";
+    start: number;
+    end: number;
+    /** Lower-case element name. */
+    name: string;
+    selfClosing: boolean;
+    /** Block-level element: a text boundary when entered or left. */
+    isBlock: boolean;
+    /** Void element (`br`, `hr`, ...): no end tag follows. */
+    isVoid: boolean;
+};
+export type HtmlTextToken = {
+    kind: "text";
+    start: number;
+    end: number;
+};
+export type HtmlRawTextToken = {
+    kind: "rawText";
+    start: number;
+    end: number;
+    /** Element the raw text belongs to. */
+    name: string;
+    /** RCDATA elements (textarea, title) decode entities; raw text ones (script, style) do not. */
+    decodesEntities: boolean;
+};
+export type HtmlOtherToken = {
+    kind: "comment" | "doctype";
+    start: number;
+    end: number;
+};
+export type HtmlToken = HtmlTagToken | HtmlTextToken | HtmlRawTextToken | HtmlOtherToken;
+export declare function isBlockElementName(name: string): boolean;
+export declare function tokenizeHtml(html: string): HtmlToken[];

package/dist/editor/spellcheck/htmlSpellTokenizer.js

@@ -0,0 +1,277 @@
+/**
+ * A small HTML tokenizer that preserves source locations. It follows the HTML
+ * tokenizer states for the subset spell checking needs: data, tags with
+ * double-quoted / single-quoted / unquoted attribute values, self-closing
+ * tags, comments, doctype / bogus comments and the raw-text (script, style)
+ * and RCDATA (textarea, title) elements. `>` inside a quoted attribute value
+ * never ends a tag, so attribute content can never leak into the text.
+ *
+ * Every token carries `[start, end)` into the original string, which lets
+ * callers rebuild the source byte-identically around a replaced text token.
+ */
+const BLOCK_ELEMENTS = new Set([
+    "address",
+    "article",
+    "aside",
+    "blockquote",
+    "body",
+    "br",
+    "caption",
+    "center",
+    "colgroup",
+    "dd",
+    "details",
+    "dialog",
+    "dir",
+    "div",
+    "dl",
+    "dt",
+    "fieldset",
+    "figcaption",
+    "figure",
+    "footer",
+    "form",
+    "h1",
+    "h2",
+    "h3",
+    "h4",
+    "h5",
+    "h6",
+    "head",
+    "header",
+    "hgroup",
+    "hr",
+    "html",
+    "legend",
+    "li",
+    "main",
+    "menu",
+    "nav",
+    "ol",
+    "optgroup",
+    "option",
+    "p",
+    "pre",
+    "section",
+    "summary",
+    "table",
+    "tbody",
+    "td",
+    "tfoot",
+    "th",
+    "thead",
+    "tr",
+    "ul",
+]);
+const VOID_ELEMENTS = new Set([
+    "area",
+    "base",
+    "br",
+    "col",
+    "embed",
+    "hr",
+    "img",
+    "input",
+    "link",
+    "meta",
+    "param",
+    "source",
+    "track",
+    "wbr",
+]);
+/** Elements whose content the HTML tokenizer treats as raw text or RCDATA. */
+const RAW_TEXT_ELEMENTS = {
+    script: false,
+    style: false,
+    xmp: false,
+    iframe: false,
+    noembed: false,
+    noframes: false,
+    textarea: true,
+    title: true,
+};
+export function isBlockElementName(name) {
+    return BLOCK_ELEMENTS.has(name);
+}
+function isAsciiLetter(code) {
+    return (code >= 65 && code <= 90) || (code >= 97 && code <= 122);
+}
+function isWhitespace(code) {
+    // Space, tab, LF, FF, CR.
+    return code === 32 || code === 9 || code === 10 || code === 12 || code === 13;
+}
+/** True when the `<` at `index` opens a tag, comment or doctype (per the HTML tag-open state). */
+function isTagStart(html, index) {
+    const next = html.charCodeAt(index + 1);
+    if (Number.isNaN(next))
+        return false;
+    if (next === 33 /* ! */ || next === 63 /* ? */)
+        return true;
+    if (next === 47 /* / */)
+        return isAsciiLetter(html.charCodeAt(index + 2));
+    return isAsciiLetter(next);
+}
+function readName(html, from) {
+    let index = from;
+    while (index < html.length) {
+        const code = html.charCodeAt(index);
+        if (isWhitespace(code) || code === 47 || code === 62)
+            break;
+        index++;
+    }
+    return index;
+}
+/**
+ * Scans a start tag's attributes from `from` (right after the tag name) and
+ * returns the index after the closing `>` (or the end of input) plus whether
+ * the tag was self-closing.
+ */
+function scanAttributes(html, from) {
+    let index = from;
+    let selfClosing = false;
+    while (index < html.length) {
+        const code = html.charCodeAt(index);
+        if (isWhitespace(code)) {
+            index++;
+            continue;
+        }
+        if (code === 62 /* > */) {
+            return { end: index + 1, selfClosing };
+        }
+        if (code === 47 /* / */) {
+            selfClosing = html.charCodeAt(index + 1) === 62;
+            index++;
+            continue;
+        }
+        selfClosing = false;
+        // Attribute name.
+        while (index < html.length) {
+            const c = html.charCodeAt(index);
+            if (isWhitespace(c) || c === 61 /* = */ || c === 62 || c === 47)
+                break;
+            index++;
+        }
+        while (index < html.length && isWhitespace(html.charCodeAt(index)))
+            index++;
+        if (html.charCodeAt(index) !== 61)
+            continue;
+        index++;
+        while (index < html.length && isWhitespace(html.charCodeAt(index)))
+            index++;
+        const quote = html.charCodeAt(index);
+        if (quote === 34 /* " */ || quote === 39 /* ' */) {
+            const close = html.indexOf(String.fromCharCode(quote), index + 1);
+            index = close < 0 ? html.length : close + 1;
+            continue;
+        }
+        // Unquoted value: up to whitespace or `>`.
+        while (index < html.length) {
+            const c = html.charCodeAt(index);
+            if (isWhitespace(c) || c === 62)
+                break;
+            index++;
+        }
+    }
+    return { end: html.length, selfClosing };
+}
+/** Index of `</name` (case-insensitive) followed by whitespace, `/`, `>` or EOF, from `from`. */
+function findRawTextEnd(html, name, from) {
+    const lower = html.toLowerCase();
+    const needle = "</" + name;
+    let index = lower.indexOf(needle, from);
+    while (index >= 0) {
+        const after = lower.charCodeAt(index + needle.length);
+        if (Number.isNaN(after) || isWhitespace(after) || after === 47 || after === 62) {
+            return index;
+        }
+        index = lower.indexOf(needle, index + 1);
+    }
+    return -1;
+}
+export function tokenizeHtml(html) {
+    const tokens = [];
+    const length = html.length;
+    let index = 0;
+    let textStart = 0;
+    const flushText = (end) => {
+        if (end > textStart)
+            tokens.push({ kind: "text", start: textStart, end });
+    };
+    while (index < length) {
+        const lt = html.indexOf("<", index);
+        if (lt < 0 || !isTagStart(html, lt)) {
+            if (lt < 0)
+                break;
+            index = lt + 1;
+            continue;
+        }
+        flushText(lt);
+        // Comment.
+        if (html.startsWith("<!--", lt)) {
+            const close = html.indexOf("-->", lt + 4);
+            const end = close < 0 ? length : close + 3;
+            tokens.push({ kind: "comment", start: lt, end });
+            index = textStart = end;
+            continue;
+        }
+        // Doctype, CDATA and bogus comments (`<!...>`, `<?...>`).
+        const second = html.charCodeAt(lt + 1);
+        if (second === 33 || second === 63) {
+            const close = html.indexOf(">", lt + 2);
+            const end = close < 0 ? length : close + 1;
+            tokens.push({ kind: second === 33 ? "doctype" : "comment", start: lt, end });
+            index = textStart = end;
+            continue;
+        }
+        // End tag.
+        if (second === 47) {
+            const nameEnd = readName(html, lt + 2);
+            const name = html.slice(lt + 2, nameEnd).toLowerCase();
+            const close = html.indexOf(">", nameEnd);
+            const end = close < 0 ? length : close + 1;
+            tokens.push({
+                kind: "endTag",
+                start: lt,
+                end,
+                name,
+                selfClosing: false,
+                isBlock: BLOCK_ELEMENTS.has(name),
+                isVoid: VOID_ELEMENTS.has(name),
+            });
+            index = textStart = end;
+            continue;
+        }
+        // Start tag.
+        const nameEnd = readName(html, lt + 1);
+        const name = html.slice(lt + 1, nameEnd).toLowerCase();
+        const { end, selfClosing } = scanAttributes(html, nameEnd);
+        tokens.push({
+            kind: "startTag",
+            start: lt,
+            end,
+            name,
+            selfClosing,
+            isBlock: BLOCK_ELEMENTS.has(name),
+            isVoid: VOID_ELEMENTS.has(name),
+        });
+        index = textStart = end;
+        const decodesEntities = RAW_TEXT_ELEMENTS[name];
+        if (decodesEntities !== undefined && !selfClosing && end < length) {
+            const rawEnd = findRawTextEnd(html, name, end);
+            const contentEnd = rawEnd < 0 ? length : rawEnd;
+            if (contentEnd > end) {
+                tokens.push({
+                    kind: "rawText",
+                    start: end,
+                    end: contentEnd,
+                    name,
+                    decodesEntities,
+                });
+            }
+            index = textStart = contentEnd;
+        }
+    }
+    flushText(length);
+    return tokens;
+}
+//# sourceMappingURL=htmlSpellTokenizer.js.map

package/dist/editor/spellcheck/htmlSpellTokenizer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"htmlSpellTokenizer.js","sourceRoot":"","sources":["../../../src/editor/spellcheck/htmlSpellTokenizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AA2CH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC;IAC7B,SAAS;IACT,SAAS;IACT,OAAO;IACP,YAAY;IACZ,MAAM;IACN,IAAI;IACJ,SAAS;IACT,QAAQ;IACR,UAAU;IACV,IAAI;IACJ,SAAS;IACT,QAAQ;IACR,KAAK;IACL,KAAK;IACL,IAAI;IACJ,IAAI;IACJ,UAAU;IACV,YAAY;IACZ,QAAQ;IACR,QAAQ;IACR,MAAM;IACN,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,MAAM;IACN,QAAQ;IACR,QAAQ;IACR,IAAI;IACJ,MAAM;IACN,QAAQ;IACR,IAAI;IACJ,MAAM;IACN,MAAM;IACN,KAAK;IACL,IAAI;IACJ,UAAU;IACV,QAAQ;IACR,GAAG;IACH,KAAK;IACL,SAAS;IACT,SAAS;IACT,OAAO;IACP,OAAO;IACP,IAAI;IACJ,OAAO;IACP,IAAI;IACJ,OAAO;IACP,IAAI;IACJ,IAAI;CACL,CAAC,CAAC;AAEH,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC;IAC5B,MAAM;IACN,MAAM;IACN,IAAI;IACJ,KAAK;IACL,OAAO;IACP,IAAI;IACJ,KAAK;IACL,OAAO;IACP,MAAM;IACN,MAAM;IACN,OAAO;IACP,QAAQ;IACR,OAAO;IACP,KAAK;CACN,CAAC,CAAC;AAEH,8EAA8E;AAC9E,MAAM,iBAAiB,GAA4B;IACjD,MAAM,EAAE,KAAK;IACb,KAAK,EAAE,KAAK;IACZ,GAAG,EAAE,KAAK;IACV,MAAM,EAAE,KAAK;IACb,OAAO,EAAE,KAAK;IACd,QAAQ,EAAE,KAAK;IACf,QAAQ,EAAE,IAAI;IACd,KAAK,EAAE,IAAI;CACZ,CAAC;AAEF,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC7C,OAAO,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;AAClC,CAAC;AAED,SAAS,aAAa,CAAC,IAAY;IACjC,OAAO,CAAC,IAAI,IAAI,EAAE,IAAI,IAAI,IAAI,EAAE,CAAC,IAAI,CAAC,IAAI,IAAI,EAAE,IAAI,IAAI,IAAI,GAAG,CAAC,CAAC;AACnE,CAAC;AAED,SAAS,YAAY,CAAC,IAAY;IAChC,0BAA0B;IAC1B,OAAO,IAAI,KAAK,EAAE,IAAI,IAAI,KAAK,CAAC,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,KAAK,EAAE,CAAC;AAChF,CAAC;AAED,kGAAkG;AAClG,SAAS,UAAU,CAAC,IAAY,EAAE,KAAa;IAC7C,MAAM,IAAI,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;IACxC,IAAI,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC;IACrC,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAC5D,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO;QAAE,OAAO,aAAa,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC;IAC1E,OAAO,aAAa,CAAC,IAAI,CAAC,CAAC;AAC7B,CAAC;AAED,SAAS,QAAQ,CAAC,IAAY,EAAE,IAAY;IAC1C,IAAI,KAAK,GAAG,IAAI,CAAC;IACjB,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;QACpC,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,KAAK,EAAE;YAAE,MAAM;QAC5D,KAAK,EAAE,CAAC;IACV,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;;GAIG;AACH,SAAS,cAAc,CACrB,IAAY,EACZ,IAAY;IAEZ,IAAI,KAAK,GAAG,IAAI,CAAC;IACjB,IAAI,WAAW,GAAG,KAAK,CAAC;IACxB,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;QACpC,IAAI,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC;YACvB,KAAK,EAAE,CAAC;YACR,SAAS;QACX,CAAC;QACD,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO,EAAE,CAAC;YACxB,OAAO,EAAE,GAAG,EAAE,KAAK,GAAG,CAAC,EAAE,WAAW,EAAE,CAAC;QACzC,CAAC;QACD,IAAI,IAAI,KAAK,EAAE,CAAC,OAAO,EAAE,CAAC;YACxB,WAAW,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC;YAChD,KAAK,EAAE,CAAC;YACR,SAAS;QACX,CAAC;QACD,WAAW,GAAG,KAAK,CAAC;QACpB,kBAAkB;QAClB,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YAC3B,MAAM,CAAC,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YACjC,IAAI,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,OAAO,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE;gBAAE,MAAM;YACvE,KAAK,EAAE,CAAC;QACV,CAAC;QACD,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,IAAI,YAAY,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YAAE,KAAK,EAAE,CAAC;QAC5E,IAAI,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,EAAE;YAAE,SAAS;QAC5C,KAAK,EAAE,CAAC;QACR,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,IAAI,YAAY,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YAAE,KAAK,EAAE,CAAC;QAC5E,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;QACrC,IAAI,KAAK,KAAK,EAAE,CAAC,OAAO,IAAI,KAAK,KAAK,EAAE,CAAC,OAAO,EAAE,CAAC;YACjD,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;YAClE,KAAK,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;YAC5C,SAAS;QACX,CAAC;QACD,2CAA2C;QAC3C,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YAC3B,MAAM,CAAC,GAAG,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;YACjC,IAAI,YAAY,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,EAAE;gBAAE,MAAM;YACvC,KAAK,EAAE,CAAC;QACV,CAAC;IACH,CAAC;IACD,OAAO,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,WAAW,EAAE,CAAC;AAC3C,CAAC;AAED,iGAAiG;AACjG,SAAS,cAAc,CAAC,IAAY,EAAE,IAAY,EAAE,IAAY;IAC9D,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;IACjC,MAAM,MAAM,GAAG,IAAI,GAAG,IAAI,CAAC;IAC3B,IAAI,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IACxC,OAAO,KAAK,IAAI,CAAC,EAAE,CAAC;QAClB,MAAM,KAAK,GAAG,KAAK,CAAC,UAAU,CAAC,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;QACtD,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,YAAY,CAAC,KAAK,CAAC,IAAI,KAAK,KAAK,EAAE,IAAI,KAAK,KAAK,EAAE,EAAE,CAAC;YAC/E,OAAO,KAAK,CAAC;QACf,CAAC;QACD,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;IAC3C,CAAC;IACD,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,MAAM,GAAgB,EAAE,CAAC;IAC/B,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;IAC3B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,MAAM,SAAS,GAAG,CAAC,GAAW,EAAE,EAAE;QAChC,IAAI,GAAG,GAAG,SAAS;YAAE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,SAAS,EAAE,GAAG,EAAE,CAAC,CAAC;IAC5E,CAAC,CAAC;IAEF,OAAO,KAAK,GAAG,MAAM,EAAE,CAAC;QACtB,MAAM,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QACpC,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,EAAE,CAAC,EAAE,CAAC;YACpC,IAAI,EAAE,GAAG,CAAC;gBAAE,MAAM;YAClB,KAAK,GAAG,EAAE,GAAG,CAAC,CAAC;YACf,SAAS;QACX,CAAC;QAED,SAAS,CAAC,EAAE,CAAC,CAAC;QAEd,WAAW;QACX,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC,EAAE,CAAC;YAChC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;YAC1C,MAAM,GAAG,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;YACjD,KAAK,GAAG,SAAS,GAAG,GAAG,CAAC;YACxB,SAAS;QACX,CAAC;QAED,0DAA0D;QAC1D,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;QACvC,IAAI,MAAM,KAAK,EAAE,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;YACnC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;YACxC,MAAM,GAAG,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,EAAE,KAAK,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;YAC7E,KAAK,GAAG,SAAS,GAAG,GAAG,CAAC;YACxB,SAAS;QACX,CAAC;QAED,WAAW;QACX,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;YAClB,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;YACvC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,EAAE,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;YACvD,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YACzC,MAAM,GAAG,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;YAC3C,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,QAAQ;gBACd,KAAK,EAAE,EAAE;gBACT,GAAG;gBACH,IAAI;gBACJ,WAAW,EAAE,KAAK;gBAClB,OAAO,EAAE,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;gBACjC,MAAM,EAAE,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC;aAChC,CAAC,CAAC;YACH,KAAK,GAAG,SAAS,GAAG,GAAG,CAAC;YACxB,SAAS;QACX,CAAC;QAED,aAAa;QACb,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;QACvC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,EAAE,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;QACvD,MAAM,EAAE,GAAG,EAAE,WAAW,EAAE,GAAG,cAAc,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;QAC3D,MAAM,CAAC,IAAI,CAAC;YACV,IAAI,EAAE,UAAU;YAChB,KAAK,EAAE,EAAE;YACT,GAAG;YACH,IAAI;YACJ,WAAW;YACX,OAAO,EAAE,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;YACjC,MAAM,EAAE,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC;SAChC,CAAC,CAAC;QACH,KAAK,GAAG,SAAS,GAAG,GAAG,CAAC;QAExB,MAAM,eAAe,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,eAAe,KAAK,SAAS,IAAI,CAAC,WAAW,IAAI,GAAG,GAAG,MAAM,EAAE,CAAC;YAClE,MAAM,MAAM,GAAG,cAAc,CAAC,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC;YAC/C,MAAM,UAAU,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC;YAChD,IAAI,UAAU,GAAG,GAAG,EAAE,CAAC;gBACrB,MAAM,CAAC,IAAI,CAAC;oBACV,IAAI,EAAE,SAAS;oBACf,KAAK,EAAE,GAAG;oBACV,GAAG,EAAE,UAAU;oBACf,IAAI;oBACJ,eAAe;iBAChB,CAAC,CAAC;YACL,CAAC;YACD,KAAK,GAAG,SAAS,GAAG,UAAU,CAAC;QACjC,CAAC;IACH,CAAC;IAED,SAAS,CAAC,MAAM,CAAC,CAAC;IAClB,OAAO,MAAM,CAAC;AAChB,CAAC"}

package/dist/editor/spellcheck/replacementGuard.d.ts

@@ -0,0 +1,27 @@
+import type { Misspelling } from "./spellCheckTypes";
+/**
+ * What a surface captures when it opens the suggestion menu, so applying a
+ * suggestion later can prove the text is still the one that was checked.
+ */
+export type SpellReplacementCapture = {
+    checkedText: string;
+    start: number;
+    end: number;
+    word: string;
+};
+export type SpellReplacementVerdict = {
+    ok: true;
+} | {
+    ok: false;
+    reason: "changed" | "mismatch";
+};
+export declare const STALE_REPLACEMENT_MESSAGE = "The text changed, re-checking spelling\u2026";
+export declare function captureReplacement(checkedText: string, misspelling: Misspelling): SpellReplacementCapture;
+/**
+ * A replacement is applied only when the surface's current text is exactly
+ * the captured text and the captured range still spells the captured word.
+ * There is no fallback search: a stale capture is rejected and re-checked.
+ */
+export declare function verifyReplacement(capture: SpellReplacementCapture, currentText: string): SpellReplacementVerdict;
+/** Plain-text replacement of a verified capture. */
+export declare function applyReplacementToText(capture: SpellReplacementCapture, replacement: string): string;

package/dist/editor/spellcheck/replacementGuard.js

@@ -0,0 +1,29 @@
+export const STALE_REPLACEMENT_MESSAGE = "The text changed, re-checking spelling…";
+export function captureReplacement(checkedText, misspelling) {
+    return {
+        checkedText,
+        start: misspelling.start,
+        end: misspelling.end,
+        word: misspelling.word,
+    };
+}
+/**
+ * A replacement is applied only when the surface's current text is exactly
+ * the captured text and the captured range still spells the captured word.
+ * There is no fallback search: a stale capture is rejected and re-checked.
+ */
+export function verifyReplacement(capture, currentText) {
+    if (currentText !== capture.checkedText)
+        return { ok: false, reason: "changed" };
+    if (currentText.slice(capture.start, capture.end) !== capture.word) {
+        return { ok: false, reason: "mismatch" };
+    }
+    return { ok: true };
+}
+/** Plain-text replacement of a verified capture. */
+export function applyReplacementToText(capture, replacement) {
+    return (capture.checkedText.slice(0, capture.start) +
+        replacement +
+        capture.checkedText.slice(capture.end));
+}
+//# sourceMappingURL=replacementGuard.js.map

package/dist/editor/spellcheck/replacementGuard.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"replacementGuard.js","sourceRoot":"","sources":["../../../src/editor/spellcheck/replacementGuard.ts"],"names":[],"mappings":"AAiBA,MAAM,CAAC,MAAM,yBAAyB,GACpC,yCAAyC,CAAC;AAE5C,MAAM,UAAU,kBAAkB,CAChC,WAAmB,EACnB,WAAwB;IAExB,OAAO;QACL,WAAW;QACX,KAAK,EAAE,WAAW,CAAC,KAAK;QACxB,GAAG,EAAE,WAAW,CAAC,GAAG;QACpB,IAAI,EAAE,WAAW,CAAC,IAAI;KACvB,CAAC;AACJ,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,iBAAiB,CAC/B,OAAgC,EAChC,WAAmB;IAEnB,IAAI,WAAW,KAAK,OAAO,CAAC,WAAW;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,SAAS,EAAE,CAAC;IACjF,IAAI,WAAW,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,KAAK,OAAO,CAAC,IAAI,EAAE,CAAC;QACnE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,UAAU,EAAE,CAAC;IAC3C,CAAC;IACD,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC;AACtB,CAAC;AAED,oDAAoD;AACpD,MAAM,UAAU,sBAAsB,CACpC,OAAgC,EAChC,WAAmB;IAEnB,OAAO,CACL,OAAO,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,CAAC;QAC3C,WAAW;QACX,OAAO,CAAC,WAAW,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CACvC,CAAC;AACJ,CAAC"}

package/dist/editor/spellcheck/spellCheckCanvasStore.d.ts

@@ -0,0 +1,17 @@
+import type { SpellCheckCanvasMatch, SpellCheckFieldTarget } from "./spellCheckCanvasTargets";
+export type SpellCheckCanvasState = {
+    matches: SpellCheckCanvasMatch[];
+    /** Latest targets by `targetKey`, for the replacement guard. */
+    targets: Map<string, SpellCheckFieldTarget>;
+};
+type Listener = () => void;
+export declare const spellCheckCanvasStore: {
+    getState: () => SpellCheckCanvasState;
+    getMatch: (key: string) => SpellCheckCanvasMatch | undefined;
+    getTarget: (targetKey: string) => SpellCheckFieldTarget | undefined;
+    subscribe: (listener: Listener) => () => void;
+    set: (matches: SpellCheckCanvasMatch[], targets: SpellCheckFieldTarget[]) => void;
+    reset: () => void;
+};
+export declare function useSpellCheckCanvasState(): SpellCheckCanvasState;
+export {};

package/dist/editor/spellcheck/spellCheckCanvasStore.js

@@ -0,0 +1,37 @@
+import { useSyncExternalStore } from "react";
+/**
+ * Tiny external store bridging the canvas overlay (which computes matches)
+ * and the page frame's context menu (which looks a right-clicked match up).
+ */
+function createSpellCheckCanvasStore() {
+    let state = { matches: [], targets: new Map() };
+    const listeners = new Set();
+    const notify = () => listeners.forEach((listener) => listener());
+    return {
+        getState: () => state,
+        getMatch: (key) => state.matches.find((match) => match.key === key),
+        getTarget: (targetKey) => state.targets.get(targetKey),
+        subscribe: (listener) => {
+            listeners.add(listener);
+            return () => {
+                listeners.delete(listener);
+            };
+        },
+        set: (matches, targets) => {
+            state = {
+                matches,
+                targets: new Map(targets.map((target) => [target.targetKey, target])),
+            };
+            notify();
+        },
+        reset: () => {
+            state = { matches: [], targets: new Map() };
+            notify();
+        },
+    };
+}
+export const spellCheckCanvasStore = createSpellCheckCanvasStore();
+export function useSpellCheckCanvasState() {
+    return useSyncExternalStore(spellCheckCanvasStore.subscribe, spellCheckCanvasStore.getState, spellCheckCanvasStore.getState);
+}
+//# sourceMappingURL=spellCheckCanvasStore.js.map

package/dist/editor/spellcheck/spellCheckCanvasStore.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"spellCheckCanvasStore.js","sourceRoot":"","sources":["../../../src/editor/spellcheck/spellCheckCanvasStore.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,OAAO,CAAC;AAc7C;;;GAGG;AACH,SAAS,2BAA2B;IAClC,IAAI,KAAK,GAA0B,EAAE,OAAO,EAAE,EAAE,EAAE,OAAO,EAAE,IAAI,GAAG,EAAE,EAAE,CAAC;IACvE,MAAM,SAAS,GAAG,IAAI,GAAG,EAAY,CAAC;IACtC,MAAM,MAAM,GAAG,GAAG,EAAE,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,EAAE,CAAC,CAAC;IAEjE,OAAO;QACL,QAAQ,EAAE,GAAG,EAAE,CAAC,KAAK;QACrB,QAAQ,EAAE,CAAC,GAAW,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,GAAG,CAAC;QAC3E,SAAS,EAAE,CAAC,SAAiB,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC;QAC9D,SAAS,EAAE,CAAC,QAAkB,EAAE,EAAE;YAChC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;YACxB,OAAO,GAAG,EAAE;gBACV,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;YAC7B,CAAC,CAAC;QACJ,CAAC;QACD,GAAG,EAAE,CAAC,OAAgC,EAAE,OAAgC,EAAE,EAAE;YAC1E,KAAK,GAAG;gBACN,OAAO;gBACP,OAAO,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC,CAAC;aACtE,CAAC;YACF,MAAM,EAAE,CAAC;QACX,CAAC;QACD,KAAK,EAAE,GAAG,EAAE;YACV,KAAK,GAAG,EAAE,OAAO,EAAE,EAAE,EAAE,OAAO,EAAE,IAAI,GAAG,EAAE,EAAE,CAAC;YAC5C,MAAM,EAAE,CAAC;QACX,CAAC;KACF,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,MAAM,qBAAqB,GAAG,2BAA2B,EAAE,CAAC;AAEnE,MAAM,UAAU,wBAAwB;IACtC,OAAO,oBAAoB,CACzB,qBAAqB,CAAC,SAAS,EAC/B,qBAAqB,CAAC,QAAQ,EAC9B,qBAAqB,CAAC,QAAQ,CAC/B,CAAC;AACJ,CAAC"}