quotation mark characters messed up when importing certain PDF files

[Sam S.]

Inkscape 0.47 fails to correctly interpret LEFT/RIGHT SINGLE QUOTATION MARK characters in some PDF files created with Microsoft Word 2007.

I've managed to narrow it down and found that this happens when in addition to the LEFT/RIGHT SINGLE QUOTATION MARK characters, the document also contains LEFT/RIGHT DOUBLE QUOTATION MARK characters. (There migth also be other scenarios, of course, but this is the one I've found so far.)

I've created a small sample document to demonstrate this, and attached it to this bug report (quotation_marks.pdf).
The file gets rendered 100% correct in Acrobat Reader, Evince, Okular, and Gimp - but in Inkscape, the single quotation marks are replaced by weird characters.

I've confirmed this bug with Inkscape under both Arch Linux and Windows 7.

[Sam S. (smls)]

For comparison, here is the same document, but with the last two lines (the ones containing the double quotation marks) deleted before exporting to PDF in Microsoft Word.

[jazzynico (jazzynico)]

Reproduced on Ubuntu 10.10, Inkscape trunk revision 10190.

[Beluga (buovjaga)]

Problem is only seen with internal, not with Poppler/Cairo import.

Arch Linux 64-bit, KDE Plasma 5
Inkscape 0.92pre1 15054 (GTK3)