microsoft/playwright · error · Error

Invalid output dimensions

Error message

Invalid output dimensions

What it means

scaleImageToSize throws 'Invalid output dimensions' when the requested target width or height is <= 0, NaN, or +/- Infinity. The scaler needs a finite positive target to allocate the output buffer and compute sampling weights.

Source

Thrown at packages/isomorphic/imageUtils.ts:54

        buffer[to + 3] = 0;
      }
    }
  }
  return { data: Buffer.from(buffer), width: size.width, height: size.height };
}

export function scaleImageToSize(image: ImageData, size: { width: number; height: number }): ImageData {
  const { data: src, width: w1, height: h1 } = image;
  const w2 = Math.max(1, Math.floor(size.width));
  const h2 = Math.max(1, Math.floor(size.height));

  if (w1 === w2 && h1 === h2)
    return image;

  if (w1 <= 0 || h1 <= 0)
    throw new Error('Invalid input image');
  if (size.width <= 0 || size.height <= 0 || !isFinite(size.width) || !isFinite(size.height))
    throw new Error('Invalid output dimensions');

  const clamp = (v: number, lo: number, hi: number) => (v < lo ? lo : v > hi ? hi : v);

  // Catmull–Rom weights
  const weights = (t: number, o: Float32Array) => {
    const t2 = t * t;
    const t3 = t2 * t;
    o[0] = -0.5 * t + 1.0 * t2 - 0.5 * t3;
    o[1] = 1.0 - 2.5 * t2 + 1.5 * t3;
    o[2] = 0.5 * t + 2.0 * t2 - 1.5 * t3;
    o[3] = -0.5 * t2 + 0.5 * t3;
  };

  const srcRowStride = w1 * 4;
  const dstRowStride = w2 * 4;

  // Precompute X: indices, weights, and byte offsets (idx*4)
  const xOff = new Int32Array(w2 * 4); // byte offsets = xIdx*4

View on GitHub (pinned to c8fc3bf8d3)

Solutions

  1. Validate Number.isFinite(size.width) && Number.isFinite(size.height) && size.width > 0 && size.height > 0 before calling.
  2. Clamp targets to a minimum of 1 with Math.max(1, Math.floor(x)) and ensure the source of the number cannot be NaN.
  3. Default optional config fields to concrete positive values rather than 0/undefined.

Example fix

// before
const out = scaleImageToSize(img, { width: ratio, height: 0 }); // ratio may be NaN

// after
const w = Number.isFinite(ratio) && ratio > 0 ? Math.floor(ratio) : 1;
const out = scaleImageToSize(img, { width: w, height: 64 });
Defensive patterns

Strategy: validation

Validate before calling

function isValidTargetSize(s: {width:number;height:number}): boolean {
  return Number.isFinite(s.width) && Number.isFinite(s.height) && s.width > 0 && s.height > 0;
}

Type guard

function isPositiveSize(s: unknown): s is { width: number; height: number } {
  return !!s && typeof s === 'object'
    && Number.isFinite((s as any).width) && (s as any).width > 0
    && Number.isFinite((s as any).height) && (s as any).height > 0;
}

Try / catch

try { return scaleImageToSize(img, size); }
catch (e) { if (e.message === 'Invalid output dimensions') { size = { width: 1, height: 1 }; return scaleImageToSize(img, size); } throw e; }

Prevention

When it happens

Trigger: Calling scaleImageToSize with a target size where size.width or size.height is 0, negative, NaN, or Infinity. Common when the target size is derived from a computation that can yield NaN (e.g. division by zero) or from an unset/optional field defaulting to 0.

Common situations: Computing a thumbnail size from an aspect ratio that produced NaN/Infinity; reading target dimensions from config that was not populated; clamping logic that collapsed a dimension to 0; passing window innerWidth/Height before layout (0).

Related errors


AI-assisted analysis of microsoft/playwright@c8fc3bf8d3 (2026-08-12). Data as JSON: /api/errors/509cc9e804f5ff6a. Report an issue: GitHub.