IssueTracker

ImageDecoder allows callers, via OnHeaderDecodedListener, to make decisions based on the decoded metadata of an image. By this point, however, callers of the API have already committed to decoding the image, but they may choose to go a different route based on the information received. To solve this, I'd like to see support in ImageDecoder for decoding just the metadata of an image.

A use case for this is to handle images that exceed the devices max texture size. The documentation details one possible way to handle this case, by adjusting the sample size of the decoder to decode a degraded image, but an equally valid alternative could be to halt image decoding and instead load the image using a BitmapRegionDecoder to display the image in a tiled manner, in order to maintain image fidelity.

With the current implementation of ImageDecoder, the only way to achieve that would be to attempt to decode the image and throw an exception in the OnHeaderDecodedListener if the image exceeds your threshold for subsampling, which you'd then have to catch to branch into your subsampling approach.