Closed Roughsketch closed 6 years ago
Throwing out this idea for now. The only image format that would need parsing would be JPEG, but since I believe the APPn and thumbnail bugs are fixed I don't think making a parser would be beneficial. Also for other image formats, the parser would be so minimal that the current implementation would probably be much simpler.
Might look into doing this. This would allow splitting each image type into separate crates (i.e., imagesize-jpeg, imagesize-png, etc) and also make the checking in general a lot cleaner.
The only potential downside I see is that nom may potentially be a lot slower, but that is something I will have to bench myself.