Enhancement: Report invalid unicode sequences encountered in PDF

[markpenny]

If a dotted circle character appears in the output PDF, it indicates the presence of dirty data, or broken output data. It would be extremely useful to alert the user to such problems, preferably by scanning the PDF for the glyph “◌” (U+25CC) and reporting which pages it occurs on. I would expect this to be an option on the Finishing tab, or possibly on the Advanced tab, which would be OFF by default, but when turned on would produce a status bar warning message, or even an error dialog.

[mhosken]

I wonder if we need a "Test" tab in which people can press fun buttons like: scan PDF for dotted circles, or search for missing fonts, or things like that and a process runs and gives a report in that tab.

[mhosken]

This is quite a bit of work since the fonts in the PDF may not include a 25CC cmap entry. We would need to parse the font files to get the GID and then hope that the GIDs are not recoded by xdvipdfm-x or xetex.

[markpenny]

@mhosken Given that you're now doing a lot of this kind of scanning of the XDV file, is this a good time to add something like this to the Finishing tab? "Check for cracks" / "Test" PDF.

[mhosken]

This is part of #995

[markpenny]

We should also parse the log file for "missing character in font" messages and present these to the user.