We build for downlevels with USE(JPEGXL) but not USE(LCMS). We should port the existing color conversion code to use Core Graphics.

The way to do this is:
1. tryDecodeICCColorProfile() can use CGColorSpaceCreateWithICCData() with a CFData
2. Create a CGColorConversionInfo from the source and destination color space
3. Use vImageConverter_CreateWithCGColorConversionInfo() to create a converter
4. vImageConvert_AnyToAny()

We might need a custom refptr for the vImageConverter. We can probably do this with custom RefPtr traits, like I did in 264884@main