[XeTeX] Isomath, Mathspec and Siunitx

[le.farfadet.spatial at free.fr]

Ulrike,

Ulrike Fischer <news3 at nililand.de> wrote:

> That's wrong. You don't get bold italic in your \empf{\bf ..}
> because of the (obsolete and here wrong) \bf. Use \bfseries instead.

Alright. I have tried it. Indeed, I have obtained bold italic.

> No. It is lmodern because isomath defines it to point to this font
> (\math... commands always points to specific fonts) and because
> mathspec/fontspec doesn't overwrite this definition (\mathbfit is
> not a standard command). If you don't want to use unicode-math you
> will have to do the redeclaration yourself:

I have tried it and I have then actually obtained Linux Libertine bold italic
font in mathematical environment. But, I have then lost Greek symbols. If I am
using "\setmathsfont(Digits,Latin,Greek)[Numbers={Lining, Proportional}]{Linux
Libertine O}", then I obtain an error message:

      ! LaTeX Error: Command `\Gamma' already defined.

Obviously, this is about redefining "\Gamma" macro. Of course "\setmathfont"
needs to redefine it, the question is which other package does too? Isomath, I
guess.

   Anyway, I would rather being able to use unicode-math.

> In this case it is due to the definition of \ln (and independant of
> siunitx). \ln uses the operator font. mathspec loads as default
> fontspec with option "no-math", so fontspec doesn't change the
> operator font. You can use \usepackage[math]{mathspec}.

It has worked pretty well, thank you.

   Here is the current state of my example:

\documentclass{article}

\usepackage{isomath}
\usepackage[math]{mathspec}
\usepackage{siunitx}

\defaultfontfeatures{Scale=MatchLowercase}
\setromanfont[Mapping=tex-text]{Linux Libertine O}
\setsansfont[Mapping= tex-text]{DejaVu Sans}
\setmonofont{DejaVu Sans Mono}
\setmathfont{Linux Libertine O}
\setmathsfont(Digits,Latin)[Numbers={Lining, Proportional}]{Linux
  Libertine O}

\begin{document}
\sisetup{detect-all, detect-inline-family=text, detect-inline-weight=text,
  detect-display-math=true}

\SI{6}{\meter \per \second} $\SI{6}{\meter \per \second}$

\emph{\bfseries{u}} $\mathbfit{u}$ $\mathbfit{\tau}$

ln $\ln$
\end{document}

   Best regards.

                             Yoann