fix: respect raw media fetch cap (#6639) (thanks @davidiach)

steipete · steipete · commit 31b196852014 · 2026-02-02T00:47:37.000-08:00
diff --git a/src/web/media.test.ts b/src/web/media.test.ts
@@ -4,7 +4,7 @@ import path from "node:path";
 import sharp from "sharp";
 import { afterEach, describe, expect, it, vi } from "vitest";
 import { optimizeImageToPng } from "../media/image-ops.js";
-import { loadWebMedia, optimizeImageToJpeg } from "./media.js";
+import { loadWebMedia, loadWebMediaRaw, optimizeImageToJpeg } from "./media.js";
 
 const tmpFiles: string[] = [];
 
@@ -106,6 +106,22 @@ describe("web media loading", () => {
     fetchMock.mockRestore();
   });
 
+  it("respects maxBytes for raw URL fetches", async () => {
+    const fetchMock = vi.spyOn(globalThis, "fetch").mockResolvedValueOnce({
+      ok: true,
+      body: true,
+      arrayBuffer: async () => Buffer.alloc(2048).buffer,
+      headers: { get: () => "image/png" },
+      status: 200,
+    } as Response);
+
+    await expect(loadWebMediaRaw("https://example.com/too-big.png", 1024)).rejects.toThrow(
+      /exceeds maxBytes 1024/i,
+    );
+
+    fetchMock.mockRestore();
+  });
+
   it("uses content-disposition filename when available", async () => {
     const fetchMock = vi.spyOn(globalThis, "fetch").mockResolvedValueOnce({
       ok: true,
diff --git a/src/web/media.ts b/src/web/media.ts
@@ -201,9 +201,14 @@ async function loadWebMediaInternal(
 
   if (/^https?:\/\//i.test(mediaUrl)) {
     // Enforce a download cap during fetch to avoid unbounded memory usage.
-    // Use the largest per-kind cap (documents) to allow oversize images to
-    // be fetched and then compressed under a smaller maxBytes.
-    const fetchCap = Math.max(maxBytes ?? 0, maxBytesForKind("unknown"));
+    // For optimized images, allow fetching larger payloads before compression.
+    const defaultFetchCap = maxBytesForKind("unknown");
+    const fetchCap =
+      maxBytes === undefined
+        ? defaultFetchCap
+        : optimizeImages
+          ? Math.max(maxBytes, defaultFetchCap)
+          : maxBytes;
     const fetched = await fetchRemoteMedia({ url: mediaUrl, maxBytes: fetchCap });
     const { buffer, contentType, fileName } = fetched;
     const kind = mediaKindFromMime(contentType);
