Use cmbright with unicode text under XeLaTeX

Question

XeLaTeX has been my preferred LaTeX engine because it straightforwardly handles text in Unicode (UTF-8 encoding). I know this can be done using \usepackage[utf8]{inputenc} with other engines, too, but it appears to me that the handling in XeLaTeX is more uniform, less of a hack; and I'm also fond of its capability to use OpenType fonts.

Now in a new document I chose to use Computer Modern Bright via \usepackage{cmbright} but found that I cannot access characters by directly typing in Unicode. Meaning "---" works, but "—" does not, the same for "\´a" vs "á" etc. First I thought this is because I did not load fontspec, and it's true that as soon as I load the package, unicode works – just not with Computer Modern Bright. If I load fontspec before cmbright I get Latin Modern Roman, if after: Latin Modern Sans.

Is there any way to make this work?

Answer 1

Use the open type version of the fonts:

\documentclass{article}
\usepackage{fontspec}
\setmainfont[ItalicFont=CMU Bright Oblique]{CMU Bright}
\begin{document}
abc  \itshape abc

\end{document}