[XeTeX] XeTeX 0.9999.1 released

Khaled Hosny khaledhosny at eglug.org
Thu Apr 4 18:00:23 CEST 2013


On Thu, Apr 04, 2013 at 11:13:46AM +0900, Dohyun Kim wrote:
> 2013/4/4 Khaled Hosny <khaledhosny at eglug.org>:
> > On Thu, Apr 04, 2013 at 07:34:23AM +0900, Akira Kakuto wrote:
> >> Hi Khaled,
> >>
> >> >One more question about vertical layout, the old XeTeX code was
> >> >positioning vertical text somehow displaced on the baseline (compared to
> >> >horizontal text), is this a desirable feature or a mere side effect?
> >>
> >> Attached is an example created by pTeX, which is
> >> known to typset vertical text beautifully.
> >> Please compare XeTeX's results, by mixing usual
> >> horizontal text.
> >
> > Both old a new output don't seem quite right (one is displaced upwards
> > and the other is displaced downwards) unless using the built in Latin of
> > the CJK fonts (in fonts that rotate Latin glyphs in vertical layout), so
> > I restored the old behaviour for sake of backward compatibility. Next
> > year I hope to look into the text directionality as a whole and may be
> > XeTeX will get real vertical layout.
> >
> 
> Wow! It works quite well now.  Thanks a lot.
> 
> I know from my experience that to compensate the baseline difference
> between horizontal and vertical texts, moving vertical texts by 0.5ex
> upwards is optimal.  But I also honor your decision to keep backward
> compatibility.

Ideally we should use BASE table to adjust base lines, but that needs a
bit more work, so lets keep things like this for now.

> Anyway, the only issue remaining is regarding the sign of x_offset.
> To move right in real vertical mode is actually to move upwards in
> xetex, so it seems that we have to reverse the plus sign of x_offset.
> In fact I know a font that has vertical gpos feature using x_offset,
> and this font works correctly when the following patch has been
> applied.  In the same logic, it seems that the sign of x_advance
> should be reversed as well.  But I am not sure as I do not have any
> font to test this.

Thanks, applied. I’ll keep x_advance until we have fonts to test.

Regards,
Khaled


More information about the XeTeX mailing list