feat(language-server): add offset-preserving virtual HTML document

2026-08-18 20:34:24 +05:30
parent 7d481df652
commit 7301849a7f
2 changed files with 225 additions and 0 deletions
@@ -0,0 +1,116 @@
+import type { TextDocument } from "./index.ts";
+
+export interface HtmlRegion {
+  start: number;
+  end: number;
+}
+
+/**
+ * Byte ranges of the markup inside each `view { }` block.
+ *
+ * This is a tolerant scanner rather than the parser: completion fires while
+ * the document is being typed, which is exactly when it does not parse.
+ */
+export function viewRegions(text: string): HtmlRegion[] {
+  const regions: HtmlRegion[] = [];
+  const pattern = /\bview\s*\{/g;
+  let match: RegExpExecArray | null;
+
+  while ((match = pattern.exec(text))) {
+    const bodyStart = match.index + match[0].length;
+    const end = matchingBrace(text, bodyStart);
+    regions.push({ start: bodyStart, end });
+    pattern.lastIndex = end;
+  }
+  return regions;
+}
+
+/**
+ * Offset of the brace closing the block that starts at `from`, or the end of
+ * the text when it is never closed (an unterminated block is normal mid-edit).
+ *
+ * Quotes are only tracked inside a tag, never in text content: `<p>it's</p>`
+ * would otherwise open a string that never closes and swallow the rest of the
+ * file.
+ */
+function matchingBrace(text: string, from: number): number {
+  let depth = 1;
+  let inTag = false;
+  let quote: string | null = null;
+
+  for (let index = from; index < text.length; index += 1) {
+    const char = text[index]!;
+
+    if (quote) {
+      if (char === quote) quote = null;
+      continue;
+    }
+    if (inTag && (char === '"' || char === "'")) {
+      quote = char;
+      continue;
+    }
+    if (char === "<") inTag = true;
+    else if (char === ">") inTag = false;
+    else if (char === "{") depth += 1;
+    else if (char === "}") {
+      depth -= 1;
+      if (depth === 0) return index;
+    }
+  }
+  return text.length;
+}
+
+/**
+ * A parallel document containing only the markup.
+ *
+ * Everything outside a view block becomes whitespace of the same length, and
+ * newlines are preserved, so an offset in the source is the same offset here.
+ * That removes the need for a mapping table entirely.
+ */
+export function virtualHtmlDocument(document: TextDocument): {
+  uri: string;
+  languageId: "html";
+  text: string;
+} {
+  const source = document.text;
+  const keep = new Array<boolean>(source.length).fill(false);
+  for (const region of viewRegions(source)) {
+    for (let index = region.start; index < region.end; index += 1) keep[index] = true;
+  }
+
+  let text = "";
+  for (let index = 0; index < source.length; index += 1) {
+    const char = source[index]!;
+    text += keep[index] || char === "\n" ? char : char === "\r" ? "\r" : " ";
+  }
+
+  return { uri: `${document.uri}.html`, languageId: "html", text };
+}
+
+export function isInsideHtml(document: TextDocument, offset: number): boolean {
+  return regionsFor(document).some((region) => offset >= region.start && offset <= region.end);
+}
+
+/**
+ * Regions for a document, cached by uri and version.
+ *
+ * A single keystroke produces a burst of completion, hover, and tag-close
+ * requests; without this each one rescans the file.
+ */
+const regionCache = new Map<string, { version: number; regions: HtmlRegion[] }>();
+
+function regionsFor(document: TextDocument): HtmlRegion[] {
+  const version = document.version ?? -1;
+  const cached = regionCache.get(document.uri);
+  if (cached && cached.version === version) return cached.regions;
+
+  const regions = viewRegions(document.text);
+  regionCache.set(document.uri, { version, regions });
+  return regions;
+}
+
+/** Drops a document's cached regions. Call when a document closes. */
+export function clearHtmlRegionCache(uri?: string): void {
+  if (uri) regionCache.delete(uri);
+  else regionCache.clear();
+}
@@ -0,0 +1,109 @@
+import { expect, test } from "bun:test";
+import { isInsideHtml, viewRegions, virtualHtmlDocument } from "../src/html-regions.ts";
+
+function doc(text: string): { uri: string; text: string; version?: number } {
+  return { uri: `file:///${Math.random()}.wrn`, text };
+}
+
+const PAGE = `page Home {
+  view {
+    <div class="card">hello</div>
+  }
+}
+`;
+
+test("the virtual document preserves length and newline offsets", () => {
+  // This is what makes position mapping unnecessary. If it breaks, every
+  // feature reports positions off by some amount instead of failing loudly.
+  const source = doc(PAGE);
+  const virtual = virtualHtmlDocument(source);
+
+  expect(virtual.text.length).toBe(source.text.length);
+  expect(virtual.languageId).toBe("html");
+  for (let i = 0; i < source.text.length; i += 1) {
+    if (source.text[i] === "\n") expect(virtual.text[i]).toBe("\n");
+  }
+});
+
+test("markup survives into the virtual document and everything else is blanked", () => {
+  const virtual = virtualHtmlDocument(doc(PAGE));
+  expect(virtual.text).toContain('<div class="card">hello</div>');
+  expect(virtual.text).not.toContain("page Home");
+  expect(virtual.text).not.toContain("view");
+});
+
+test("an apostrophe in text content does not swallow later regions", () => {
+  // A scanner treating ' as a string delimiter anywhere considers the rest of
+  // the file one open string and loses every later region.
+  const source = `page A {
+  view {
+    <p>it's fine</p>
+  }
+}
+component B {
+  view {
+    <span>second</span>
+  }
+}
+`;
+  expect(viewRegions(source)).toHaveLength(2);
+  expect(virtualHtmlDocument(doc(source)).text).toContain("<span>second</span>");
+});
+
+test("interpolation braces nest without ending the region early", () => {
+  const source = `page A {
+  view {
+    <div class={cond ? "a" : "b"} data-x={{ a: 1 }}>after</div>
+  }
+}
+`;
+  const regions = viewRegions(source);
+  expect(regions).toHaveLength(1);
+  expect(virtualHtmlDocument(doc(source)).text).toContain("after</div>");
+});
+
+test("unparseable mid-edit markup still yields a region", () => {
+  // Completion fires exactly when the document does not parse.
+  const source = `page A {
+  view {
+    <div class="
+  }
+}
+`;
+  expect(viewRegions(source).length).toBe(1);
+});
+
+test("a file with no view block yields no regions and a fully blank document", () => {
+  const source = `page A {
+  functions {
+    function go() {}
+  }
+}
+`;
+  expect(viewRegions(source)).toEqual([]);
+  expect(virtualHtmlDocument(doc(source)).text.trim()).toBe("");
+});
+
+test("regions are cached per document version", () => {
+  // One keystroke fans out into completion, hover, and tag-close requests.
+  const first = doc(PAGE);
+  first.version = 1;
+  expect(isInsideHtml(first, PAGE.indexOf("<div"))).toBe(true);
+
+  // Same version, mutated text: the cached regions are reused, proving the
+  // scan did not run again.
+  first.text = "page A { }";
+  expect(isInsideHtml(first, PAGE.indexOf("<div"))).toBe(true);
+
+  first.version = 2;
+  expect(isInsideHtml(first, PAGE.indexOf("<div"))).toBe(false);
+});
+
+test("isInsideHtml distinguishes markup from surrounding code", () => {
+  const source = doc(PAGE);
+  const markupOffset = PAGE.indexOf("<div");
+  const keywordOffset = PAGE.indexOf("page");
+
+  expect(isInsideHtml(source, markupOffset)).toBe(true);
+  expect(isInsideHtml(source, keywordOffset)).toBe(false);
+});