When you say "two of the fonts are names Montserrat, and one of them is named Montserrat[sic] Medium", is there any other distinguishing part of either of the fonts named "Montserrat" (is one of them "italic" and one of them "bold" or "semibold")? If not, the problem might be the actual embedded font files more than the error checker.
Unfortunately, some software (I've seen it primarily in the cloud-heaviest Adobe stuff), when embedding fonts, doesn't do a very good job of ensuring that the fonts being embedded have unique names, both of the files and internally... especially for fonts not originating in that software-vendor's ecosystem. For example, if one used the full ITC Galliard set when doing the original layout for print and then used the internal converter to create an epub or PDF, the font names all compress down to "ITC Galliard"... even though there are sixteen variants. (This can be a big problem trying to route large technical documents to overseas printers as PDFs or even raw PostScript.)
And if the document crossed the streams — moved between Apple and any other OS — font integrity gets even more fun. That's not Calibre's error checker...
|