This question, which is a prosecution of the debate in Misplaced Accents in Cyrillic Text answered by @egreg, wishes to focus on the mechanism of combination of a combining accent (e.g. combining acute accets 0301) with an uppercase.
When you have to add an acute accent to a cyrillic uppercase Lua/Xe-LaTex joints the given capital letter to the combining acute accent unicode character (U+0301), because in the Unicode chart there is no given glyph for any cyrillic acute accented capital vowel.
This combination comes however with an error for a variety of font sets, as the following code demonstrates:
\documentclass[12pt,a5paper,openright,twoside]{memoir}
\usepackage{fontspec}
\usepackage{xunicode}
\def\CyrillicAccents{А а А́ а́ \\ Е е Е́ е́ \\ И и И́ и́ \\ О о О́ о́ \\ У у У́ у́ \\ Ы ы Ы́ ы́ \\ Э э Э́ э́ \\ Ю ю Ю́ ю́ \\ Я я Я́ я́}
\begin{document}
\fontspec{Linux Libertine}
\noindent\CyrillicAccents
\end{document}
The acute accent does not match the height of the capital letters.
This might be caused by an error in the font set, but if the same set of accented vowels is edited in MS Word or Pages (or Open Office, I guess) the result is without bugs:
It seems that XeLaTeX does not find the correct informations to correctly place the accents, while other editors do.
Which is the method of joining a letter to a combining character? How to control the parameters of this process?
