Some PDF files report now non-integer pixel size.
Proofread page attempts to use this non-integer size for underlying image on book pages.
IMO, the retrieved size should be rounded up or down before further use.
This is a problem if the proofreadpage wiki configuration does not allow to alter the default scan size.
Example:
File:
https://commons.wikimedia.org/wiki/File:PL_Doyle_-_Zaginiony_footbalista_(zbi%C3%B3r).pdf
Page:
https://pl.wikisource.org/wiki/Strona:PL_Doyle_-_Zaginiony_footbalista_(zbi%C3%B3r).pdf/9
Image link:
https://upload.wikimedia.org/wikipedia/commons/thumb/7/72/PL_Doyle_-_Zaginiony_footbalista_%28zbi%C3%B3r%29.pdf/page9-772.91666666667px-PL_Doyle_-_Zaginiony_footbalista_%28zbi%C3%B3r%29.pdf.jpg