I tried today to use Okular from the 4.2.0 (stable) mingw binary package. I noticed it wasn't showing some text in a PDF file, and it turned out it was because the fonts for that text were not included in the PDF. Apparently, fontconfig was not able to find the font files in the C:WINDOWSFonts directory. I tried "fc-list.exe", but it showed a message that the configuration file could not be found. That was strange, as the file was in C:KDEetcfontsfonts.conf . I opened Process Monitor to see where it was looking for the file, and it turned out it was looking in "D:kdekde-mingwetcfontsfonts.conf", where I never had an installation of KDE. So, I created those directories in D: and fc-list.exe was able to load the fonts.conf from there. But, it displayed another problem: if the directory in fonts.conf is listed as default:

Code: Select all

%WINDIR%fonts

it didn't work. So, I changed that to:

Code: Select all

C:/windows/fonts

and fc-list.exe finally listed the fonts, and Okular loaded the text perfectly fine. I hope that this annoying problem can be resolved.

megara wrote:and fc-list.exe finally listed the fonts, and Okular loaded the text perfectly fine. I hope that this annoying problem can be resolved.

I have to say thanks again for that solution - the paths hardcoded are of course the build paths. Since we mainly do package KDE, our fontconfig packages are not in the best shape - I will update fontconfig in the next time though including your findings.

regards

SaroEngels wrote:I have to say thanks again for that solution - the paths hardcoded are of course the build paths. Since we mainly do package KDE, our fontconfig packages are not in the best shape - I will update fontconfig in the next time though including your findings.

regards

Thanks for the quick reply! I hope you will find a way to use the WINDIR system variable in ".fonts", so guys with Windows installed in D: or E: won't have any problems. And keep up the good work - Okular one of the best media viewers for any platform, so I was glad to be able to use it on Windows as well.

I had the same problem here. The method in the first post did give me the fonts back. Although something goes wrong when Okular displays the following PDF:

http://www.tnw.utwente.nl/organisatie/S ... 010708.pdf

If I look at the properties dialog ( Font Tab ) I see that two out of the three paths refer to the wrong font file, which is the cause of the PDF displaying incorrectly. The arial font is linked the marlett.ttf file (which is not a human-readable font) and the Arial,Bold is linked to the Microsoft Yahei Bold (msyhbd.ttf) file (which is human readable). Since most of the document is written in Arial the document is unreadable.

Any thoughts on why this goes wrong (Could it have something to do with fontconfig creating these mistakes in the cache ?) and how it should be fixed ?

Kind regards,

Ben

benvdh wrote:I had the same problem here. The method in the first post did give me the fonts back. Although something goes wrong when Okular displays the following PDF:

http://www.tnw.utwente.nl/organisatie/S ... 010708.pdf

If I look at the properties dialog ( Font Tab ) I see that two out of the three paths refer to the wrong font file, which is the cause of the PDF displaying incorrectly. The arial font is linked the marlett.ttf file (which is not a human-readable font) and the Arial,Bold is linked to the Microsoft Yahei Bold (msyhbd.ttf) file (which is human readable). Since most of the document is written in Arial the document is unreadable.

Any thoughts on why this goes wrong (Could it have something to do with fontconfig creating these mistakes in the cache ?) and how it should be fixed ?

Kind regards,

Ben

I just checked and here on my (debug) installation it works (even without the first changes). I am not sure what this could mean though.
SE

Hey,

The only thing I could add is that I'm currently using 4.2 stable (MinGW) on windows 7 beta. And that, as I mentioned before, fc-list.exe also seems to give a bit of a weird output when run from windows powershell. A bit like a text with the wrong encoding.

Kind regards,

Ben