[XeTeX] Kerning flaw with MinionPro & XeTex

[Alessandro Ceschini]

Yeah, as I told you I'd taken a look in FontForge and it came out font
designers had anticipated such combination in the kerning pairs table.
In effect it works in PangoView but I can't figure out how to direct the
output toward a pdf, so I can't show you the right kerning, anyhow there
should be more space, that's obvious.
There should be some bug in XeTeX, I guess it doesn't look for kerning
pairs that involve alternate glyphs.
Best wishes,

Le samedi 05 février 2011 à 22:26 +0100, Adam Twardoch (List) a écrit :
> Sorry, I replied prematurely. I did have to change my document to
> 
> \documentclass{article}
> 
> \usepackage{fontspec}
> 
> \usepackage{polyglossia}
> 
> \setmainlanguage{Serbian}
> 
> \usepackage{xunicode}
> 
> \usepackage{xltxtra}
> 
> \setmainfont[Script=Cyrillic,Language=Serbian]{Minion Pro}
> 
> \begin{document}
> 
> \fontsize{24}{24}\selectfont Убить!
> 
> \end{document}
> 
> 
> to get the Serbian variant, and you're right.
> 
> I've also tested it with a plain XeTeX variant:
> 
> 
> \font\samplefont = "Minion Pro:script=cyrl;language=SRB" at 72bp
> 
> \samplefont
> 
> \setbox0 \hbox{Убить!}
> 
> \shipout\box0
> 
> \end
> 
> 
> The font has the kerning definition:
> 
> @kc38_first_29 = [\uni040E \uni0423 \uni0474 ];
> @kc38_second_16 = [\uni0435 \uni043E \uni0441 \uni0444 \uni0451 \uni0454
> \uni0473 \uni0431.ital ];
> pos @kc38_first_29 @kc38_second_16 -118;
> 
> and the kerning pair in question is \uni0423\uni0431.ital
> 
> So it should be kerned with the value -118. Unfortunately, it does seem
> like the rendered kerning value is deeper than that.
> 
> Best,
> Adam
> 
>