One of his question is not answered. Chevereto is compressing the image when it is trying to fix the orientation by getting the data from EXIF. Any idea why ?
I wanted to know why it leads to compression of image.
Sure, I designed it that way. The "compression" is because the system needs to create a new image to fix the orientation, when it does that it needs to create a new image and surprise JPG is a lossy algorithm, it always have compression. This image orientation fix is because browsers back then (idk like 5 years ago) didn't fix orientation so in some browsers the images appeared sideways.
Nowadays I don't know if this is still required, according to this support is relatively new circa 2020.
I think that you should open an RFC requesting to stop fixing orientation, it is a useless step and we could add a polyfill for those browsers not supporting it.