Apache OpenOffice (AOO) Bugzilla – Issue 76698
Writer would not correctly display words in particular font (attached).
Last modified: 2013-08-07 14:38:26 UTC
Please see attached screenshot "izhitsa1.PNG" - Writer does not display correctly words in that font (circled in red). Competing word processors (AbiWord and MSWord) display the words just fine (circled in green). I am also attaching "izh____c.ttf" and "izhitsa.doc".
Created attachment 44688 [details] Illustrating screenshot
Created attachment 44689 [details] ttf used in document
Created attachment 44690 [details] Problematic file
ES->HDU: same problem in MS Word (but there, there's a fallback to a Serif font while OOo fallbacks to a Sans Serif font). I think it's a font problem. In this font, the Cyrillic characters are mapped to the Latin area (which is not correct). Thus, writing with a latin keyboard shows "faked" Cyrillic characters. When typing with a Cyrillic keyboard mapping, another font is used (the same you see in the screenshot). What do you think?
I agree with ES that the problem is with the fake-unicode support of the font. The font claims that it basically only supports latin and doesn't even hint at that this fake unicode might be an alias for cyrillic. Only the RTF-documents indicates a connection between the font and cyrillic: Style 28 in the RTF-document switches the font encoding: {\s28\f4\fs24\lang1049Normal;} @FME: I'm not sure if we want to bother with legacy pseudo-unicode conventions. If we want to then the solution is to translate the portion's text from unicode back to the requested encoding (in this case lang1049) before calling into VCL.
IMHO it is not worth the effort to make OOo aware of these fake fonts. Instead use something like http://docs.openoffice.ru/~doc/ooextras/cyrtools1.2.uno.zip to convert the document to the coding the broken font uses (or even better, just get a real font instead) - you might use Thessalonica as well to convert from one special font-encoding to another one (mainly deals with greek though) Attached screenshot shows the document after running Cyrillic to latin using the abovementioned extension.
Created attachment 44710 [details] screenshot after the conversion to match the font's fake charset
fme: I agree with hdu and cloph. At least we should lower the priority.
AbiWord handles the font just fine ...