> As assetutil showed, there are multiple entries for ClarusSmooth2.pdf—and one of them is followed by data that starts with the PDF file format header (%PDF). But, unfortunately, extracting that data into a separate file seems to result in a blank PDF. And I don’t know enough about the format to figure out whether there is any vector data in it, or if it truly is empty.
I suspect this _is_ a vector format, and they’re using the Xcode 9+ features for a scalable image at runtime. I’ve never tried to extract the data back, it wouldn’t surprise me if they’re optimized for themselves, since in theory apple’s frameworks are the only “consumer” of the compiled asset catalog.
Ah, that definitely seems like what it is—CUINamedImage.isVectorBased is true for the Clarus image. Reverse engineering the format could be a fun project.
I suspect this _is_ a vector format, and they’re using the Xcode 9+ features for a scalable image at runtime. I’ve never tried to extract the data back, it wouldn’t surprise me if they’re optimized for themselves, since in theory apple’s frameworks are the only “consumer” of the compiled asset catalog.
ref: https://useyourloaf.com/blog/xcode-9-vector-images/