[XeTeX] Problems with thickness of \frac rule and width of accents (\hat) with XeLaTeX

drgz drgz syrehue at hotmail.com
Mon Aug 16 10:59:46 CEST 2010

Thanks for the detailed answer Ulrik, although it seems like there's not much I can do but learn LuaLaTeX in order to get the documents as I want them to be. :)
Would it be possible to get the LuaLaTeX code you used in order to compile the document as a starting-point, as I've never seen or tried it before?

Best regards,

> Date: Mon, 16 Aug 2010 00:15:35 +0200
> From: ulrik.vieth at arcor.de
> To: xetex at tug.org
> Subject: Re: [XeTeX] Problems with thickness of \frac rule and width of accents (\hat) with XeLaTeX
> Hi,
> I tested it with both XeLaTeX and LuaLaTeX (both from TL2010 pretest).
> In short, the problem only occurs in XeLaTeX, but not in LuaLaTeX,
> despite using the same macro packages and fonts for both engines.
> I do not really understand the problem with the fraction rule thickness.
> It probably should be the same on both engines (despite conceptual
> differences), as it should be based on the same OpenType parameter
> RadicalRuleThickness, but it seems to be different nonetheless:
> XeLaTeX:
> .......\kern4.7461
> .......\rule(0.47998+0.0)x*
> .......\kern2.97542
> LuaLaTeX:
> .......\kern3.43199
> .......\rule(0.79199+0.0)x*
> .......\kern2.98799
> The LuaTeX value of 0.792pt at 12pt matches the scaled value of 0.66pt
> at 10pt of the OpenType font parameter for XITS Math.
> The XeTeX value of 0.48pt at 12pt matches the scaled value of 0.4pt
> at 10pt, which is suspiciously similar to the default rule thickness
> of CM fonts. I don't know how this comes about. I'm not sure if this
> is an engine bug or if it could be caused by a macro problem.
> As for the overly wide hat, this is obviously an engine bug in XeTeX.
> The behavior is clearly wrong and the problem does not occur in LuaTeX.
> I believe I have seen a similar problem before in XeTeX, which had
> been fixed by Jonathan during a lunch break at EuroTeX 2009 last year.
> I wonder if an old bug has resurfaced now or if anything went wrong
> merging the bug fixes (or if it was a different bug that was fixed).
> Regards, Ulrik
> P.S: On second thought, I'm not sure any more if this was really
> the same bug that was fixed or just a similar one. It could have been
> about the positioning of math accents (e.g. centering \dot and \ddot)
> rather than the width of wide math accents, but it also had to do
> with an incorrect calculation of the size based on the subscripts.
> On 08/15/2010 09:37 PM, drgz drgz wrote:
> > Anyhow, I'm playing around with XeLaTeX (the newest version coming with
> > MikTeX) trying to compile some documents. And almost everything is as it
> > should be, except the thickness of the rule in \frac (does not get any
> > thicker) and the width of the \hat accent (which gets wider than it should).
> > I've attached a link to a compiled PDF which illustrates what I
> > experience - see http://folk.ntnu.no/mitrevsk/temp/xelatex/ex.pdf
> >
> > No matter which OTF font with unicode-math support I select I experience
> > this problem when I compile with XeLaTeX. Is there any way to fix this ?
> --------------------------------------------------
> Subscriptions, Archive, and List information, etc.:
>   http://tug.org/mailman/listinfo/xetex
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://tug.org/pipermail/xetex/attachments/20100816/249d55a1/attachment.html>

More information about the XeTeX mailing list