@bettercms-ai/richtext 0.8.0 → 0.8.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bettercms-ai/richtext",
-  "version": "0.8.0",
+  "version": "0.8.1",
   "type": "module",
   "sideEffects": false,
   "description": "The one Portable Text contract for BetterCMS: schema, converters, and rendering. Shared by the backend, the dashboard and every delivery SDK.",

package/src/ingest.ts

@@ -187,6 +187,35 @@ function tableRule(key: KeyGenerator, ingestCell: (html: string) => PortableText
   };
 }
 
+/** A code block's text, line breaks included: Lexical writes them as `<br>`, not newlines. */
+const codeText = (node: Node): string =>
+  node.nodeName.toLowerCase() === "br"
+    ? "\n"
+    : node.nodeType === 3
+      ? (node.textContent ?? "")
+      : [...node.childNodes].map(codeText).join("");
+
+/** `language-x` on the `<code>` (Markdown, our renderer) or `data-language` on the `<pre>` (Lexical, Shiki). */
+const codeLanguage = (pre: Element): string | undefined =>
+  /(?:^|\s)language-([\w+#-]+)/.exec(pre.querySelector("code")?.getAttribute("class") ?? "")?.[1] ??
+  (pre.getAttribute("data-language") || undefined);
+
+/**
+ * `<pre>` → the `bcmsCode` block the renderer turns back into `<pre><code>`. Without it the
+ * deserialiser's default flattens a code block to a paragraph with a `code` mark and its
+ * newlines collapsed: every imported code fence became one line of inline code.
+ */
+function codeRule(key: KeyGenerator): DeserializerRule {
+  return {
+    deserialize(el: Node, _next, createBlock) {
+      if (!isElement(el) || el.nodeName.toLowerCase() !== "pre") return undefined;
+      const language = codeLanguage(el);
+      const code = codeText(el).replace(/\n$/, "");
+      return createBlock({ _type: PT_TYPES.code, _key: key(), code, ...(language ? { language } : {}) });
+    },
+  };
+}
+
 /**
  * An inline element Portable Text does not model → the `styled` annotation.
  *
@@ -384,9 +413,9 @@ function mediaRule(key: KeyGenerator, schema: PtSchema | undefined): Deserialize
 const allowsStyled = (schema: PtSchema | undefined): boolean =>
   !schema || schema.annotations.some((a) => a.name === PT_TYPES.styled);
 
-/** Does this field's schema carry the table block object? Asked by name, not by capability. */
-const allowsTables = (schema: PtSchema | undefined): boolean =>
-  !schema || schema.blockObjects.some((o) => o.name === PT_TYPES.table);
+/** Does this field's schema carry the named block object? Asked by name, not by capability. */
+const allowsBlockObject = (schema: PtSchema | undefined, name: string): boolean =>
+  !schema || schema.blockObjects.some((o) => o.name === name);
 
 const isElement = (n: unknown): n is Element =>
   !!n && typeof n === "object" && (n as Node).nodeType === 1;
@@ -413,7 +442,8 @@ export function htmlToPortableText(html: string, opts: IngestOptions = {}): Port
      * accepts only images.
      */
     rules: [
-      ...(allowsTables(opts.schema) ? [tableRule(key, (cell) => htmlToPortableText(cell, opts))] : []),
+      ...(allowsBlockObject(opts.schema, PT_TYPES.table) ? [tableRule(key, (cell) => htmlToPortableText(cell, opts))] : []),
+      ...(allowsBlockObject(opts.schema, PT_TYPES.code) ? [codeRule(key)] : []),
       mediaRule(key, opts.schema),
       ...(allowsStyled(opts.schema) ? [styledRule(key)] : []),
       ...(allowsLinks(opts.schema) ? [classedLinkRule(key)] : []),