Dear Hans, Thank your very much for your attention and your remarks about the sample Faarsi file I sent. Thanks also for your valuable time you spend on such questions from non experts… On 7 janv. 09, at 09:38, Hans Hagen wrote:
[…] i can look into it if i have samples + fonts + expected output
I uploaded a few fonts, as well as three sample tex files and the corresponding PDF outputs on http://www.kavian.fr/tex/mkiv-persian/ The files contain the same persian text with a few words in english in order to show how mkiv behaves wrongly compared with XeTeX and XeConTeXt (which are fine). It is possible that the fonts which are used are not totally compliant, but I don't see why XeTeX and XeConTeXt (that is XeTeX + ConTeXt mkii) give good outputs. To begin with, here are a couple of issues with mkiv: 1) When using "Arabic short vowels" such as U+064F (ARABIC DAMMA) U+0650 (ARABIC KASRA) U+0651 (ARABIC SHADDA) the ligatures between characters breaks down. For instance compare the typesetting of the following words in XeTeX, XeConTeXt and mkiv (actually, in mkiv, this wrong behavior depends also on the font…): مغان. مُغان. تولد. تَوّلد 2) With the font IranNastaliq, when in the midst of persian words one writes an english word, then the words after the english one are garbled in mkiv, while in XeTeX and XeConTeXt there is no such problem. With my best regards: OK