I see... one option would be to provide your own FNT file (which is actually a plain text file) to map coordinates within the PNG file to the corresponding characters. If you open it within a text editor, you would see something like this:
Code: Select all
info face="DejaVu Sans" size=15 bold=0 italic=0 charset="" unicode=1 stretchH=100 smooth=1 aa=2 padding=0,0,0,0 spacing=1,1 outline=0 common lineHeight=15 base=12 scaleW=1790 scaleH=16 pages=1 packed=0 alphaChnl=1 redChnl=0 greenChnl=0 blueChnl=0 page id=0 file="font_dejavu_sans_0.png" chars count=191 char id=32 x=1632 y=0 width=4 height=15 xoffset=0 yoffset=0 xadvance=4 page=0 chnl=15 char id=33 x=1471 y=0 width=5 height=15 xoffset=0 yoffset=0 xadvance=5 page=0 chnl=15 char id=34 x=1401 y=0 width=6 height=15 xoffset=0 yoffset=0 xadvance=6 page=0 chnl=15 (... and so on ...)
An alternative would be to create a custom version of the 'font_conv' utility that would convert the image in a more direct way, though unless you're using some fixed width font, you still would need some way to specify the offsets.
That's weird... perhaps I should have included the DLL in the pngq directory, as I had done with 'libpng13.dll'.
Anyway, until I fix this, one option would be to donload the DLL directly from sourceforge: http://prdownloads.sourceforge.net/libp ... p?download ; just add the copy of 'zlib1.dll' that comes inside this zip to the 'pngnq-0.5-i386-win' directory.