Object 39 is an image; but even Object 4 (the content stream of the first page; obj 93 *is* used on that page) has such a decompression error.
Ghostscript prints: warning: ignoring zlib error: incorrect data check
So probably libqpdf (which is used for the very low-level pdf stuff in pdftopdf) should also handle this as warning instead of as error?
Object 39 is an image; but even Object 4 (the content stream of the first page; obj 93 *is* used on that page) has such a decompression error.
Ghostscript prints:
warning: ignoring zlib error: incorrect data check
So probably libqpdf (which is used for the very low-level pdf stuff in pdftopdf) should also handle this as warning instead of as error?