@pitlane/content 0.2.0 → 0.2.1

package/CHANGELOG.md

@@ -1,5 +1,17 @@
 # @pitlane/content
 
+## 0.2.1
+
+### Patch Changes
+
+- f2fa73e: Read a heading's `text`, and the `id` made from it, from what the heading shows on the page. Three kinds of heading change, and an anchor written by hand against one of them needs checking once:
+
+    - Inline HTML now contributes its words and not its tags. `## <span>Visible</span> text` was `<span>Visible</span> text` with the `id` `spanvisiblespan-text`, and is now `Visible text` with `visible-text`.
+    - An image no longer contributes its alt text, which matches GitHub and Astro. `## [A](url) ![Cat photo](cat.png)` was `a-cat-photo` and is now `a-`, keeping the space before the image as GitHub does.
+    - In MDX, an expression that is a single string literal now contributes its value. `## Hello {"world"}` was `hello-` and is now `hello-world`, and `## The {"{"} key` reads `The { key`. Any other expression, such as `{name}`, still contributes nothing.
+
+    Headings without inline HTML, images, or expressions keep their `text` and `id`.
+
 ## 0.2.0
 
 ### Minor Changes

package/dist/satteri.d.mts

@@ -3,7 +3,8 @@ import { HastPluginEntry, MdastPluginEntry } from "satteri";
 /**
  * Collects every heading of a document as `{ depth, slug, text }`, publishes the
  * list as `data.headings`, and gives each heading an `id` matching its slug so an
- * anchor link lands on it. On MDX the list is also appended to the tree as
+ * anchor link lands on it. A heading's text is the text it shows on the page,
+ * as `visibleText` reads it. On MDX the list is also appended to the tree as
  * `export const headings`, so the compiled module carries its own table of
  * contents.
  *

package/dist/satteri.mjs

@@ -47,9 +47,73 @@ function uniqueSlug(text, taken) {
 	return slug;
 }
 /**
+* The text a heading shows on the page, which is what GitHub and Astro slug.
+* Sätteri's `textContent` differs in three places: it keeps raw HTML tags as
+* text, includes image alt text, and reads every MDX expression as nothing.
+* Its options drop the first two, but an expression only carries its source,
+* so the walk is done here.
+*
+* Raw HTML nodes hold only the tags; the text between them is already its own
+* sibling node, so dropping `html` leaves the visible words. Images render no
+* text.
+*/
+function visibleText(node) {
+	switch (node.type) {
+		case "html":
+		case "image":
+		case "imageReference": return "";
+		case "mdxTextExpression": return stringLiteralValue(node.value ?? "") ?? "";
+	}
+	if (node.children) return node.children.map(visibleText).join("");
+	return node.value ?? "";
+}
+/** A lone JavaScript string or template literal, surrounded by nothing but whitespace. */
+const STRING_LITERAL = /^\s*(?:"((?:[^"\\\n\r]|\\[^])*)"|'((?:[^'\\\n\r]|\\[^])*)'|`((?:[^`\\$]|\\[^]|\$(?!\{))*)`)\s*$/;
+/**
+* One escape sequence: a code point in either `\u` form, a `\x` byte, `\0` not
+* followed by a digit, a line continuation, or any other escaped character.
+*/
+const ESCAPE = /\\(?:u\{([\da-fA-F]+)\}|u([\da-fA-F]{4})|x([\da-fA-F]{2})|(0)(?!\d)|(\r\n|[\n\r\u2028\u2029])|([^]))/g;
+const SINGLE_CHARACTER_ESCAPES = {
+	b: "\b",
+	f: "\f",
+	n: "\n",
+	r: "\r",
+	t: "	",
+	v: "\v"
+};
+/**
+* The value of an MDX expression whose source is a lone string literal, such
+* as `{"{"}`, the usual way to write a brace in MDX text. `undefined` for any
+* other expression, whose value is only known once the module runs.
+*
+* Escapes decode as in a module, which is strict code: a legacy octal escape
+* such as `\1`, a `\8`, or a malformed `\x` or `\u` makes the literal invalid.
+*/
+function stringLiteralValue(source) {
+	let match = STRING_LITERAL.exec(source);
+	if (!match) return void 0;
+	let [, double, single, template] = match;
+	let body = double ?? single ?? template.replace(/\r\n?/g, "\n");
+	let valid = true;
+	let value = body.replace(ESCAPE, (_, braced, unit, byte, nul, lineBreak, other) => {
+		if (braced !== void 0) {
+			let codePoint = Number.parseInt(braced, 16);
+			if (codePoint <= 1114111) return String.fromCodePoint(codePoint);
+		} else if (unit !== void 0 || byte !== void 0) return String.fromCharCode(Number.parseInt(unit ?? byte, 16));
+		else if (nul !== void 0) return "\0";
+		else if (lineBreak !== void 0) return "";
+		else if (!/[ux\d]/.test(other)) return SINGLE_CHARACTER_ESCAPES[other] ?? other;
+		valid = false;
+		return "";
+	});
+	return valid ? value : void 0;
+}
+/**
 * Collects every heading of a document as `{ depth, slug, text }`, publishes the
 * list as `data.headings`, and gives each heading an `id` matching its slug so an
-* anchor link lands on it. On MDX the list is also appended to the tree as
+* anchor link lands on it. A heading's text is the text it shows on the page,
+* as `visibleText` reads it. On MDX the list is also appended to the tree as
 * `export const headings`, so the compiled module carries its own table of
 * contents.
 *
@@ -72,7 +136,7 @@ function headings() {
 		return {
 			name: "pitlane-headings",
 			heading(node, context) {
-				let text = context.textContent(node);
+				let text = visibleText(node);
 				let slug = uniqueSlug(text, taken);
 				collected.push({
 					depth: node.depth,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pitlane/content",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "description": "Schema-validated content collections for Remix.",
   "keywords": [
     "content",