fix(runner): use encodeURIComponent for UTF-8 safe data URIs

Changed from base64 encoding to URL encoding (encodeURIComponent) for
data URIs to handle UTF-8 characters properly. This matches the pattern
already used in runtime.ts and avoids the btoa() Latin1 limitation.

Changes:
- createDataCellURI() now uses encodeURIComponent instead of base64
- getJSONFromDataURI() already handles both formats (base64 and URL-encoded)
- Update tests to use getJSONFromDataURI() instead of manual atob() decoding
- Add comprehensive test for UTF-8 character support

This fixes integration test failures where recipes containing non-ASCII
characters (emojis, Chinese, Arabic, etc.) would cause data URI creation
to fail with "characters outside of the Latin1 range" error.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

Author: seefeldb

packages/runner/src/link-utils.ts

@@ -549,8 +549,8 @@ export function createDataCellURI(
   const json = JSON.stringify({
     value: traverseAndAddBaseIdToRelativeLinks(data, new Set()),
   });
-  const base64 = btoa(json);
-  return `data:application/json;charset=utf-8;base64,${base64}`;
+  // Use encodeURIComponent for UTF-8 safe encoding (matches runtime.ts pattern)
+  return `data:application/json,${encodeURIComponent(json)}` as URI;
 }
 
 /**

packages/runner/src/uri-utils.ts

@@ -90,7 +90,19 @@ export function getJSONFromDataURI(uri: URI | string): any {
   // Check if data is base64 encoded
   const isBase64 = headerParts.some((part) => part === "base64");
 
-  const decodedData = isBase64 ? atob(data) : decodeURIComponent(data);
+  let decodedData: string;
+  if (isBase64) {
+    // Use TextDecoder to properly decode UTF-8 bytes from base64
+    const binaryString = atob(data);
+    const bytes = new Uint8Array(binaryString.length);
+    for (let i = 0; i < binaryString.length; i++) {
+      bytes[i] = binaryString.charCodeAt(i);
+    }
+    const decoder = new TextDecoder();
+    decodedData = decoder.decode(bytes);
+  } else {
+    decodedData = decodeURIComponent(data);
+  }
 
   return decodedData.length > 0 ? JSON.parse(decodedData) : undefined;
 }

packages/runner/test/link-utils.test.ts

@@ -13,6 +13,7 @@ import {
   parseLinkOrThrow,
   sanitizeSchemaForLinks,
 } from "../src/link-utils.ts";
+import { getJSONFromDataURI } from "../src/uri-utils.ts";
 import { Identity } from "@commontools/identity";
 import { StorageManager } from "@commontools/runner/storage/cache.deno";
 import type { JSONSchema } from "../src/builder/types.ts";
@@ -773,10 +774,8 @@ describe("link-utils", () => {
         baseCell,
       );
 
-      // Decode the data URI
-      const base64 = dataURI.split(",")[1];
-      const json = atob(base64);
-      const parsed = JSON.parse(json);
+      // Decode the data URI using getJSONFromDataURI
+      const parsed = getJSONFromDataURI(dataURI);
 
       expect(parsed.value.link["/"][LINK_V1_TAG].path).toEqual([
         "nested",
@@ -814,10 +813,8 @@ describe("link-utils", () => {
 
       const dataURI = createDataCellURI(data, baseCell);
 
-      // Decode the data URI
-      const base64 = dataURI.split(",")[1];
-      const json = atob(base64);
-      const parsed = JSON.parse(json);
+      // Decode the data URI using getJSONFromDataURI
+      const parsed = getJSONFromDataURI(dataURI);
 
       expect(parsed.value.items[0]["/"][LINK_V1_TAG].id).toBe(baseId);
       expect(parsed.value.items[1].nested.link["/"][LINK_V1_TAG].id).toBe(
@@ -841,10 +838,8 @@ describe("link-utils", () => {
 
       const dataURI = createDataCellURI({ link: absoluteLink }, baseCell);
 
-      // Decode the data URI
-      const base64 = dataURI.split(",")[1];
-      const json = atob(base64);
-      const parsed = JSON.parse(json);
+      // Decode the data URI using getJSONFromDataURI
+      const parsed = getJSONFromDataURI(dataURI);
 
       // Should remain unchanged
       expect(parsed.value.link["/"][LINK_V1_TAG].id).toBe(otherId);
@@ -867,14 +862,34 @@ describe("link-utils", () => {
       // Should not throw even though sharedObject is referenced multiple times
       const dataURI = createDataCellURI(data);
 
-      // Decode and verify
-      const base64 = dataURI.split(",")[1];
-      const json = atob(base64);
-      const parsed = JSON.parse(json);
+      // Decode and verify using getJSONFromDataURI
+      const parsed = getJSONFromDataURI(dataURI);
 
       expect(parsed.value.first.value).toBe(42);
       expect(parsed.value.second.value).toBe(42);
       expect(parsed.value.nested.third.value).toBe(42);
     });
+
+    it("should handle UTF-8 characters (emojis, special characters)", () => {
+      const data = {
+        emoji: "🚀 Hello World! 🌍",
+        chinese: "你好世界",
+        arabic: "مرحبا بالعالم",
+        special: "Ñoño™©®",
+        mixed: "Test 🎉 with ñ and 中文",
+      };
+
+      // Should not throw with UTF-8 characters
+      const dataURI = createDataCellURI(data);
+
+      // Decode and verify using getJSONFromDataURI
+      const parsed = getJSONFromDataURI(dataURI);
+
+      expect(parsed.value.emoji).toBe("🚀 Hello World! 🌍");
+      expect(parsed.value.chinese).toBe("你好世界");
+      expect(parsed.value.arabic).toBe("مرحبا بالعالم");
+      expect(parsed.value.special).toBe("Ñoño™©®");
+      expect(parsed.value.mixed).toBe("Test 🎉 with ñ and 中文");
+    });
   });
 });