fix(uploads): bound the HEIF fallback decode by declared pixels, not just bytes

apps/sim/lib/uploads/server/heic-pixel-guard.test.ts

@@ -0,0 +1,124 @@
+/**
+ * @vitest-environment node
+ *
+ * The pixel ceiling in `transcodeHeicToJpeg`, tested against a stubbed decoder.
+ *
+ * Separate from `heic.test.ts` so that file keeps exercising the real WebAssembly
+ * decoder — mocking it there would retire the one test proving the dynamic import
+ * resolves. Reaching the guard for real would mean hand-building a HEVC-coded HEIF,
+ * which needs an encoder this repo does not ship; stubbing the declared dimensions
+ * tests the decision the guard actually makes.
+ */
+import { beforeEach, describe, expect, it, vi } from 'vitest'
+
+const { mockAll, mockConvert } = vi.hoisted(() => ({
+  mockAll: vi.fn(),
+  mockConvert: vi.fn(),
+}))
+
+vi.mock('heic-decode', () => ({ all: mockAll, default: Object.assign(vi.fn(), { all: mockAll }) }))
+vi.mock('heic-convert', () => ({ default: mockConvert }))
+
+import { transcodeHeicToJpeg } from '@/lib/uploads/server/heic'
+
+/** An ISO-BMFF `ftyp` box declaring a HEVC-coded HEIF still. */
+function heifHeader(): Buffer {
+  const header = Buffer.alloc(16)
+  header.writeUInt32BE(16, 0)
+  header.write('ftyp', 4, 'ascii')
+  header.write('heic', 8, 'ascii')
+  return header
+}
+
+const MAX_TRANSCODE_INPUT_PIXELS = 100_000_000
+
+/** `all()` returns live libheif handles plus the `dispose` that frees them. */
+function handles(sizes: Array<{ width: number; height: number }>) {
+  const dispose = vi.fn()
+  const list = sizes.map((size) => ({ ...size, decode: vi.fn() }))
+  return Object.assign(list, { dispose })
+}
+
+describe('transcodeHeicToJpeg pixel ceiling', () => {
+  beforeEach(() => {
+    vi.clearAllMocks()
+    mockConvert.mockResolvedValue(Buffer.from('jpeg-bytes'))
+  })
+
+  it.each([
+    ['refused', [{ width: 30_000, height: 30_000 }]],
+    ['transcoded', [{ width: 8064, height: 6048 }]],
+  ])('frees the decoder handles when the image is %s', async (_outcome, sizes) => {
+    // `all()` leaves freeing to the caller, so skipping it leaks the libheif
+    // context on the WebAssembly heap once per preview.
+    const list = handles(sizes)
+    mockAll.mockResolvedValue(list)
+
+    await transcodeHeicToJpeg(heifHeader())
+
+    expect(list.dispose).toHaveBeenCalledTimes(1)
+  })
+
+  it('frees the decoder handles even when reading dimensions throws', async () => {
+    const list = handles([{ width: 100, height: 100 }])
+    Object.defineProperty(list[0], 'width', {
+      get() {
+        throw new Error('handle went away')
+      },
+    })
+    mockAll.mockResolvedValue(list)
+
+    await transcodeHeicToJpeg(heifHeader())
+
+    expect(list.dispose).toHaveBeenCalledTimes(1)
+  })
+
+  it('refuses a container declaring more pixels than the ceiling', async () => {
+    // 30000x30000 is ~900MP — the decoder would allocate ~3.4GB before the codec
+    // is asked for anything, so the refusal has to happen on the declared size.
+    mockAll.mockResolvedValue(handles([{ width: 30_000, height: 30_000 }]))
+
+    expect(await transcodeHeicToJpeg(heifHeader())).toBeNull()
+    expect(mockConvert).not.toHaveBeenCalled()
+  })
+
+  it('refuses when any image in a multi-image container is oversized', async () => {
+    mockAll.mockResolvedValue(
+      handles([
+        { width: 100, height: 100 },
+        { width: 30_000, height: 30_000 },
+      ])
+    )
+
+    expect(await transcodeHeicToJpeg(heifHeader())).toBeNull()
+    expect(mockConvert).not.toHaveBeenCalled()
+  })
+
+  it('transcodes a container at the ceiling', async () => {
+    mockAll.mockResolvedValue(
+      handles([{ width: MAX_TRANSCODE_INPUT_PIXELS / 10_000, height: 10_000 }])
+    )
+
+    expect(await transcodeHeicToJpeg(heifHeader())).toEqual(Buffer.from('jpeg-bytes'))
+    expect(mockConvert).toHaveBeenCalledTimes(1)
+  })
+
+  it('transcodes an ordinary phone photo', async () => {
+    // A 48MP iPhone still, which must stay well inside the ceiling.
+    mockAll.mockResolvedValue(handles([{ width: 8064, height: 6048 }]))
+
+    expect(await transcodeHeicToJpeg(heifHeader())).toEqual(Buffer.from('jpeg-bytes'))
+    expect(mockConvert).toHaveBeenCalledTimes(1)
+  })
+
+  it('never asks the stubbed handle to decode', async () => {
+    // The whole point of `all()` over `one()`: the decision is made before the
+    // raster is allocated.
+    const list = handles([{ width: 30_000, height: 30_000 }])
+    mockAll.mockResolvedValue(list)
+
+    await transcodeHeicToJpeg(heifHeader())
+
+    expect(list[0].decode).not.toHaveBeenCalled()
+  })
+})

apps/sim/lib/uploads/server/heic.test.ts

@@ -124,4 +124,51 @@ describe('transcodeHeicToJpeg', () => {
     // amount of type-checking establishes for a lazily loaded WebAssembly module.
     expect(await transcodeHeicToJpeg(ftypHeader('heic'))).toBeNull()
   })
+
+  it('reports dimensions from `all()` without decoding, which the pixel check relies on', async () => {
+    // The guard is only worth anything if `all()` exposes the declared size up front:
+    // were dimensions to move behind `decode()` (as they are on the default export),
+    // `width * height` would silently become NaN and the check would never reject.
+    // Driven with a stub libheif so the real mapping runs without a HEVC encoder.
+    const buildDecoder = (await import('heic-decode/lib.js')).default as (lib: unknown) => {
+      all: (options: {
+        buffer: Buffer
+      }) => Promise<Array<{ width: number; height: number }> & { dispose: () => void }>
+    }
+    let decoded = false
+    const { all } = buildDecoder({
+      ready: Promise.resolve(),
+      HeifDecoder: class {
+        decoder = { delete: () => {} }
+        decode() {
+          return [
+            {
+              get_width: () => 30_000,
+              get_height: () => 20_000,
+              free: () => {},
+              display: (target: unknown, cb: (t: unknown) => void) => {
+                decoded = true
+                cb(target)
+              },
+            },
+          ]
+        }
+      },
+    })
+
+    const handles = await all({ buffer: ftypHeader('heic') })
+
+    expect(handles[0].width * handles[0].height).toBe(600_000_000)
+    expect(typeof handles.dispose).toBe('function')
+    expect(decoded).toBe(false)
+  })
+
+  it('exposes `all` as a named export, which the pixel check destructures', async () => {
+    // A CJS `module.exports = one; module.exports.all = all` need not surface `all`
+    // as a named ESM export. If it stopped doing so the pixel check would throw,
+    // get swallowed by the catch, and quietly stop guarding — with mocked tests
+    // still green. Pin the real shape.
+    const { all } = await import('heic-decode')
+    expect(typeof all).toBe('function')
+  })
 })

apps/sim/lib/uploads/server/heic.ts

@@ -25,11 +25,26 @@ const HEIF_BRANDS = new Set([...HEVC_HEIF_BRANDS, 'mif1', 'msf1', 'avif', 'avis'
  * generous headroom over any phone photo — a 12MP iPhone HEIC is 1-4MB — while
  * bounding what one read can cost.
  *
- * This bounds file size, not pixel count. A small file declaring enormous
- * dimensions is rejected during parse by libheif's own security limits.
+ * This bounds file size only; {@link MAX_TRANSCODE_INPUT_PIXELS} bounds the raster,
+ * which a small file can still declare to be enormous.
  */
 const MAX_TRANSCODE_INPUT_BYTES = 20 * 1024 * 1024
 
+/**
+ * Pixel ceiling for the fallback decode, checked against the container's declared
+ * dimensions before any raster exists.
+ *
+ * Needed because the decoder allocates `width * height * 4` up front — the size is
+ * taken straight from the `ispe` box and the buffer is built before the codec is
+ * asked for anything, so a malformed file never has to decode to cost the memory.
+ * libheif's own default ceiling is ~1.07e9 pixels (~4.3GB as RGBA), which is far too
+ * loose to be the only guard.
+ *
+ * 100MP caps that allocation near 400MB and clears every phone camera — a 48MP
+ * iPhone still is 8064x6048.
+ */
+const MAX_TRANSCODE_INPUT_PIXELS = 100_000_000
+
 /** A real `ftyp` box holds a handful of brands; anything larger is malformed or hostile. */
 const MAX_FTYP_BOX_BYTES = 512
 
@@ -97,6 +112,36 @@ export async function transcodeHeicToJpeg(buffer: Buffer): Promise<Buffer | null
   }
 
   try {
+    // Read the declared dimensions first. `all()` parses the container and reports
+    // each image's size while leaving the decode — and therefore the allocation —
+    // for `decode()`, which is what makes refusing an oversized one cheap. The
+    // container gets parsed twice as a result; that is a header parse against a
+    // ceiling this path exists to enforce, and only on the HEVC fallback.
+    const { all } = await import('heic-decode')
+    const images = await all({ buffer })
+    // `all()` hands back live libheif handles and, unlike the default export, leaves
+    // freeing them to the caller — skipping this leaks the decoder context on the
+    // WebAssembly heap once per preview. The dimensions are plain numbers, so they
+    // outlive the handles safely.
+    let oversized: { width: number; height: number } | undefined
+    try {
+      oversized = images
+        .map(({ width, height }) => ({ width, height }))
+        .find((image) => image.width * image.height > MAX_TRANSCODE_INPUT_PIXELS)
+    } finally {
+      images.dispose()
+    }
+    if (oversized) {
+      logger.warn('Skipped HEIC transcode above the pixel ceiling', {
+        width: oversized.width,
+        height: oversized.height,
+        pixels: oversized.width * oversized.height,
+        ceiling: MAX_TRANSCODE_INPUT_PIXELS,
+        bytes: buffer.length,
+      })
+      return null
+    }
+
     const convert = (await import('heic-convert')).default
     const jpeg = await convert({ buffer, format: 'JPEG' })
     logger.info('Transcoded HEIC image', {

apps/sim/package.json

@@ -169,6 +169,7 @@
     "gray-matter": "^4.0.3",
     "groq-sdk": "^0.15.0",
     "heic-convert": "2.1.0",
+    "heic-decode": "2.1.0",
     "html-to-text": "^9.0.5",
     "http-proxy-agent": "7.0.2",
     "https-proxy-agent": "7.0.6",

apps/sim/types/heic-decode.d.ts

@@ -0,0 +1,35 @@
+/**
+ * `heic-decode` ships no types. Only the surface we use is declared: `all()`
+ * reports each image's declared dimensions and defers the decode, which is what
+ * lets a caller refuse an oversized one before any raster is allocated.
+ */
+declare module 'heic-decode' {
+  interface DecodedHeifImage {
+    width: number
+    height: number
+    data: Uint8ClampedArray
+  }
+
+  interface HeifImageHandle {
+    width: number
+    height: number
+    decode: () => Promise<DecodedHeifImage>
+  }
+
+  /**
+   * `dispose` is non-enumerable on the returned array and is NOT optional: it frees
+   * the image handles and the libheif context, which `all()` — unlike the default
+   * export — leaves to the caller. Declared required so a caller cannot forget it.
+   */
+  interface HeifImageHandles extends Array<HeifImageHandle> {
+    dispose: () => void
+  }
+
+  function decode(options: { buffer: Buffer }): Promise<DecodedHeifImage>
+
+  namespace decode {
+    function all(options: { buffer: Buffer }): Promise<HeifImageHandles>
+  }
+
+  export = decode
+}