fix(studio-server): bound entity-quoted preview references (#5100)

packages/studio-server/src/helpers/previewReads.ts

@@ -1,11 +1,17 @@
 import { readlinkSync } from "node:fs";
 import { dirname, join, parse, relative, resolve } from "node:path";
+import { parseHTML } from "linkedom";
 import { STUDIO_SIGNATURE_MANIFEST_PATHS } from "./projectSignature.js";
 import { realFilePath } from "./safePath.js";
 
 const ALWAYS_AFFECTS = ["hyperframes.json", ...STUDIO_SIGNATURE_MANIFEST_PATHS];
+// A malformed inline URL must not turn ancestor tracking into an image-sized path walk.
+const MAX_PREVIEW_PATH_LENGTH = 4096;
+const SHOW_ELEMENT = 1;
 const REFERENCE =
   /\b(?:src|href|poster|data-composition-src)\s*=\s*(?:"([^"\n]*)"|'([^'\n]*)')|url\(\s*(?:"([^"\n]*)"|'([^'\n]*)'|([^"'\s)]+))/gi;
+const REFERENCE_ATTRIBUTE = /\b(?:src|href|poster|data-composition-src)$/i;
+const isElement = (node: Node): node is Element => node.nodeType === 1;
 // macOS and Windows volumes ignore letter case by default, so it is not part of a path's identity there.
 const pathKey =
   process.platform === "darwin" || process.platform === "win32"
@@ -39,6 +45,7 @@ function linksOnTheWay(path: string): string[] {
       continue;
     }
     const next = join(at, part);
+    if (next.length > MAX_PREVIEW_PATH_LENGTH) return passed;
     const target = hops < 40 ? readLink(next) : null;
     if (target === null) {
       at = next;
@@ -56,10 +63,12 @@ function linksOnTheWay(path: string): string[] {
 
 // A link's target can be created or retargeted later, so every read resolves it again.
 export function recordPreviewRead(projectDir: string, filePath: string): void {
+  if (filePath.length > MAX_PREVIEW_PATH_LENGTH) return;
   const key = projectKey(projectDir);
   let reads = readsByProject.get(key);
   if (!reads) readsByProject.set(key, (reads = new Set()));
   const read = inRealProject(projectDir, filePath);
+  if (read.length > MAX_PREVIEW_PATH_LENGTH) return;
   for (const found of [read, ...linksOnTheWay(read)]) {
     for (let path = pathKey(found); !reads.has(path); path = dirname(path)) reads.add(path);
   }
@@ -67,9 +76,9 @@ export function recordPreviewRead(projectDir: string, filePath: string): void {
 
 export function recordPreviewReferences(projectDir: string, html: string): void {
   const named = new Set<string>();
-  for (const match of html.matchAll(REFERENCE)) {
-    const url = (match[1] ?? match[2] ?? match[3] ?? match[4] ?? match[5] ?? "").trim();
-    if (!url || /^(?:[a-z][a-z0-9+.-]*:|[/#])/i.test(url)) continue;
+  const add = (reference: string) => {
+    const url = reference.trim();
+    if (!url || /^(?:[a-z][a-z0-9+.-]*:|[/#])/i.test(url)) return;
     const path = url.split(/[?#]/)[0] ?? "";
     let decoded = path;
     try {
@@ -77,7 +86,31 @@ export function recordPreviewReferences(projectDir: string, html: string): void
     } catch {
       // A malformed escape names the file literally.
     }
-    if (decoded) named.add(decoded);
+    if (decoded && decoded.length <= MAX_PREVIEW_PATH_LENGTH) named.add(decoded);
+  };
+  const scan = (text: string) => {
+    for (const match of text.matchAll(REFERENCE)) {
+      add(match[1] ?? match[2] ?? match[3] ?? match[4] ?? match[5] ?? "");
+    }
+  };
+  const scanElement = (element: Element) => {
+    for (const attribute of element.attributes) {
+      if (REFERENCE_ATTRIBUTE.test(attribute.name)) add(attribute.value);
+    }
+    const style = element.getAttribute("style");
+    if (style) scan(style);
+    if (element.localName === "style" || element.localName === "script") {
+      scan(element.textContent ?? "");
+    }
+  };
+  // The parser decodes attributes; walking also visits inert template descendants without cloning them.
+  const { document } = parseHTML(html);
+  for (const root of document.children) {
+    scanElement(root);
+    const elements = document.createTreeWalker(root, SHOW_ELEMENT);
+    for (let node = elements.nextNode(); node; node = elements.nextNode()) {
+      if (isElement(node)) scanElement(node);
+    }
   }
   for (const path of named) recordPreviewRead(projectDir, path);
 }

packages/studio-server/src/routes/preview.test.ts

@@ -1898,6 +1898,86 @@ describe("what the preview loaded", () => {
     expect(affectsPreview(projectDir, "im")).toBe(false);
   });
 
+  it("does not walk entity-quoted inline image data as a filesystem path", () => {
+    const projectDir = createProjectDir();
+    const data = `data:image/png;base64,${"A/".repeat(128)}`;
+    for (const quote of [""", "'", """, """, "'", "'"]) {
+      recordPreviewReferences(
+        projectDir,
+        `<i style="background-image:url(${quote}${data}${quote})"></i>`,
+      );
+      recordPreviewBuilt(projectDir);
+      expect(affectsPreview(projectDir, `${quote}${data}${quote}`)).toBe(false);
+    }
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, `&quot;${data}&quot;`)).toBe(false);
+  });
+
+  it("tracks entity-quoted local image paths after decoding their HTML entities", () => {
+    const projectDir = createProjectDir();
+    recordPreviewReferences(
+      projectDir,
+      '<i style="background-image:url(&quot;assets/a&amp;b.png&quot;)"></i>',
+    );
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, "assets/a&b.png")).toBe(true);
+  });
+
+  it("keeps entities literal in stylesheet URLs, as CSS raw text does", () => {
+    const projectDir = createProjectDir();
+    recordPreviewReferences(
+      projectDir,
+      "<style>i { background:url('assets/a&amp;b.png') }</style>",
+    );
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, "assets/a&amp;b.png")).toBe(true);
+    expect(affectsPreview(projectDir, "assets/a&b.png")).toBe(false);
+  });
+
+  it("does not mistake style-looking text for an HTML style attribute", () => {
+    const projectDir = createProjectDir();
+    recordPreviewReferences(
+      projectDir,
+      `<i title='style="'></i><style>/* style=" */ i { background:url('assets/a&amp;b.png') } /* " */</style><i title='"'></i>`,
+    );
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, "assets/a&amp;b.png")).toBe(true);
+    expect(affectsPreview(projectDir, "assets/a&b.png")).toBe(false);
+  });
+
+  it("tracks references in nested templates and script-created CSS", () => {
+    const projectDir = createProjectDir();
+    recordPreviewReferences(
+      projectDir,
+      '<template><template><img src="assets/nested.png"></template></template><script>el.style.background="url(assets/script.png)"</script>',
+    );
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, "assets/nested.png")).toBe(true);
+    expect(affectsPreview(projectDir, "assets/script.png")).toBe(true);
+  });
+
+  it("bounds filesystem paths after stripping long queries and decoding escapes", () => {
+    const projectDir = createProjectDir();
+    const path = `assets/${"x/".repeat(1200)}image.png`;
+    const encoded = path.replace(/x/g, "%78");
+    recordPreviewReferences(
+      projectDir,
+      `<img src="assets/image.png?${"x".repeat(5000)}"><img src="${encoded}">`,
+    );
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, "assets/image.png")).toBe(true);
+    expect(affectsPreview(projectDir, path)).toBe(true);
+  });
+
+  it("does not walk oversized preview references or direct read paths", () => {
+    const projectDir = createProjectDir();
+    const path = `assets/${"long/".repeat(850)}image.png`;
+    recordPreviewReferences(projectDir, `<img src="${path}">`);
+    recordPreviewRead(projectDir, path);
+    recordPreviewBuilt(projectDir);
+    expect(affectsPreview(projectDir, path)).toBe(false);
+  });
+
   it.skipIf(process.platform === "win32")(
     "counts an edit to the file a symlinked asset points at",
     async () => {