fix(core,producer): stamp render ids on empty-src media and pair the snapshot by them

packages/core/src/compiler/mediaRenderIds.test.ts

@@ -75,10 +75,25 @@ describe("assignMediaRenderIds", () => {
     expect(ids[1]).toBe("clip__hf2");
   });
 
-  it("leaves media with no source at all alone", () => {
+  it("stamps empty-src media with colliding author ids", () => {
+    // `src=""` used to skip the stamp. The snapshot then keyed by raw id and
+    // colliding scenes collapsed. Runtime assignment is why the src is empty,
+    // not a second path this function sees.
+    const { document } = parseHTML(
+      '<video id="clip" src=""></video><video id="clip" src=""></video>',
+    );
+    assignMediaRenderIds(document as unknown as Parameters<typeof assignMediaRenderIds>[0]);
+    expect(
+      Array.from(document.querySelectorAll("video")).map((el) =>
+        el.getAttribute(MEDIA_RENDER_ID_ATTR),
+      ),
+    ).toEqual(["clip", "clip__hf2"]);
+  });
+
+  it("stamps a video with no source attribute at all", () => {
     const { document } = parseHTML('<video id="no-src"></video>');
     assignMediaRenderIds(document as unknown as Parameters<typeof assignMediaRenderIds>[0]);
-    expect(document.querySelector("video")?.hasAttribute(MEDIA_RENDER_ID_ATTR)).toBe(false);
+    expect(document.querySelector("video")?.getAttribute(MEDIA_RENDER_ID_ATTR)).toBe("no-src");
   });
 
   it("stamps media whose source is a <source> child rather than a src attribute", () => {
@@ -103,10 +118,10 @@ describe("assignMediaRenderIds", () => {
     expect(document.querySelector("audio")?.getAttribute(MEDIA_RENDER_ID_ATTR)).toBe("bed");
   });
 
-  it("ignores a <source> child that carries no src", () => {
+  it("stamps a video whose <source> child carries no src", () => {
     const { document } = parseHTML('<video id="empty"><source type="video/mp4"></video>');
     assignMediaRenderIds(document as unknown as Parameters<typeof assignMediaRenderIds>[0]);
-    expect(document.querySelector("video")?.hasAttribute(MEDIA_RENDER_ID_ATTR)).toBe(false);
+    expect(document.querySelector("video")?.getAttribute(MEDIA_RENDER_ID_ATTR)).toBe("empty");
   });
 });
 

packages/core/src/compiler/mediaRenderIds.ts

@@ -43,32 +43,23 @@ export const AUDIO_GROUP_RENDER_ID_ATTR = "data-hf-group-render-id";
 /**
  * Elements the render pipeline addresses by id.
  *
- * `<video>`/`<audio>` are matched whether the source is a `src` attribute or a
- * `<source>` child. Matching only `[src]` left the `<source>`-child shape
- * unstamped, so two scenes each declaring `<video id="clip"><source …></video>`
- * kept colliding ids in the render document, which is exactly the failure this
- * module exists to prevent.
+ * `<video>`/`<audio>` are matched even with an empty `src`. Authors assign the
+ * URL from the scene script (`el.src = url`); the static parse then skips them
+ * and the browser snapshot has to pair the clips. Without a render id on those
+ * elements the snapshot keys by raw id and colliding scenes collapse. `<img>`
+ * still requires a `src` attribute (empty is enough) so we do not stamp every
+ * decorative image.
  */
 const MEDIA_SELECTOR = "video, audio, img[src]";
 
 /** Buses, which are addressed by id in exactly the same way and collide the
  *  same way. Only an id'd bus can be joined at all. */
 const AUDIO_GROUP_SELECTOR = "hf-audio-group[id]";
 
-/** A `<source>`-bearing media element is addressable even without its own `src`. */
-function hasPlayableSource(el: MediaElementLike): boolean {
-  if (el.getAttribute("src")) return true;
-  const sources = el.querySelectorAll?.("source[src]");
-  if (!sources) return false;
-  for (const _ of sources) return true;
-  return false;
-}
-
 interface MediaElementLike {
   readonly tagName?: string;
   getAttribute(name: string): string | null;
   setAttribute(name: string, value: string): void;
-  querySelectorAll?(selector: string): Iterable<unknown>;
 }
 
 /** A bus or member, which additionally needs subtree scoping to be paired up. */
@@ -108,7 +99,6 @@ export function assignMediaRenderIds(document: DocumentLike): void {
   const pending: MediaElementLike[] = [];
 
   for (const el of document.querySelectorAll(MEDIA_SELECTOR)) {
-    if (!hasPlayableSource(el)) continue;
     const existing = el.getAttribute(MEDIA_RENDER_ID_ATTR);
     if (existing) {
       taken.add(existing);

packages/producer/src/services/htmlCompiler.test.ts

@@ -107,6 +107,16 @@ describe("discoverMediaFromBrowser", () => {
     );
     expect(media[0]).toMatchObject({ start: 0, end: 2, duration: 2, mediaStart: 1 });
   });
+
+  it("reports colliding empty-src videos by render id, not author id", async () => {
+    const media = await discover(
+      `<video id="clip" data-hf-render-id="clip" src="" data-start="0" data-end="4" data-media-start="10"></video>` +
+        `<video id="clip" data-hf-render-id="clip__hf2" src="" data-start="4" data-end="8" data-media-start="40"></video>`,
+      {},
+    );
+    expect(media.map((entry) => entry.id)).toEqual(["clip", "clip__hf2"]);
+    expect(media.map((entry) => entry.mediaStart)).toEqual([10, 40]);
+  });
 });
 
 function validTestMediaResponse(): Response {
@@ -2811,6 +2821,31 @@ describe("duplicate media ids across nested compositions", () => {
     expect(compiled.audios[0]).toMatchObject({ start: 0, end: 3, mediaStart: 5 });
     expect(compiled.audios[1]).toMatchObject({ start: 3, end: 6, mediaStart: 50 });
   });
+
+  it("stamps unique render ids on empty-src videos across two scenes", async () => {
+    const { projectDir, indexPath } = writeTwoSceneProject(
+      "scene-a.html",
+      "scene-b.html",
+      (label, mediaStart) =>
+        `<div data-composition-id="${label}" data-start="0" data-duration="3"
+     data-width="640" data-height="360">
+  <video id="clip" src="" data-start="0" data-duration="3"
+         data-media-start="${mediaStart}" data-track-index="0"></video>
+</div>`,
+    );
+
+    const compiled = await compileForRender(projectDir, indexPath, projectDir);
+
+    // Static parse still omits empty src from the media list; the stamp is
+    // what the snapshot uses to keep the two clips distinct.
+    expect(compiled.videos).toHaveLength(0);
+    const { document } = parseHTML(compiled.html);
+    expect(
+      Array.from(document.querySelectorAll("video")).map((el) =>
+        el.getAttribute("data-hf-render-id"),
+      ),
+    ).toEqual(["clip", "clip__hf2"]);
+  });
 });
 
 describe("STUDIO-5433 — ffprobe failure includes src URL for attribution", () => {

packages/producer/src/services/htmlCompiler.ts

@@ -2090,7 +2090,9 @@ export async function compileForRender(
  * Discover media elements from the browser DOM after JavaScript has run.
  * This catches videos/audios whose `src` is set dynamically via JS
  * (e.g. `document.getElementById("pip-video").src = URL`), which the
- * static regex parsers miss because the HTML has `src=""`.
+ * static regex parsers miss because the HTML has `src=""`. Clips are keyed
+ * by `data-hf-render-id` when present — author ids collide across inlined
+ * scenes, and this snapshot is the only identity those empty-src elements get.
  */
 export interface BrowserMediaElement {
   id: string;
@@ -2152,7 +2154,14 @@ export async function discoverMediaFromBrowser(page: Page): Promise<BrowserMedia
         : htmlEl.tagName.toLowerCase() === "video"
           ? "video"
           : "audio";
-      const id = htmlEl.id || (isImage ? autoImageIds.get(htmlEl) : undefined);
+      // Render id is document-unique after inlining; author id is only unique
+      // per composition file. Empty-src media is skipped by the static parse
+      // and lives or dies on this snapshot — keying by author id collapses
+      // colliding scenes onto one clip (residual of #3340).
+      const id =
+        htmlEl.getAttribute("data-hf-render-id") ||
+        htmlEl.id ||
+        (isImage ? autoImageIds.get(htmlEl) : undefined);
       if (!id) return;
 
       // currentSrc is authoritative for <video>/<audio><source> and responsive images.
@@ -2213,7 +2222,10 @@ export async function discoverAudioVolumeAutomationFromTimeline(
   const sampleStep = 1 / Math.min(60, Math.max(1, sampleFps));
   const rawWindows = await page.evaluate((ids: string[]) => {
     return ids.flatMap((id) => {
-      const el = document.getElementById(id) ?? document.getElementById(id.replace(/-audio$/, ""));
+      const el =
+        window.__hfMediaEl?.(id) ??
+        document.getElementById(id) ??
+        document.getElementById(id.replace(/-audio$/, ""));
       if (!(el instanceof HTMLAudioElement) && !(el instanceof HTMLVideoElement)) return [];
       return [
         {
@@ -2302,7 +2314,9 @@ export async function discoverAudioVolumeAutomationFromTimeline(
 
       for (const { id, start, end } of clips) {
         const el =
-          document.getElementById(id) ?? document.getElementById(id.replace(/-audio$/, ""));
+          window.__hfMediaEl?.(id) ??
+          document.getElementById(id) ??
+          document.getElementById(id.replace(/-audio$/, ""));
         if (!(el instanceof HTMLAudioElement) && !(el instanceof HTMLVideoElement)) continue;
 
         const sampleStart = Math.max(0, start);
@@ -2419,7 +2433,7 @@ export async function discoverVideoVisibilityFromTimeline(
       lastVisible: number | null;
     }[] = [];
     for (const videoEl of videos) {
-      const id = videoEl.id;
+      const id = videoEl.getAttribute?.("data-hf-render-id") || videoEl.id;
       if (!id) continue;
       entries.push({
         id,