Dear LyX developers,
testing the fix for https://www.lyx.org/trac/ticket/9681, I came across a
problem leading to run-together words in Greek and Cyrillic documents.
E.g. the Greek splash.lyx example contains the phrase
“επεξεργαστής εγγράφων.”
which comes out as
”επεξεργαστήςεγγράφων.“
in the default output (tested with LyX 2.2 and 2.4dev). This is due to
the forced change of the final sigma and the fact that the LaTeX Internal
Character Representation (unicodesymbols substitutions) are not
terminated (λ μ becomes \textlambda \textmu instead of \textlambda{}
\textmu{}).
With LyX 2.4dev, Greek and Cyrillic words run into each other if they occur
text with language English (or any other Latin-using language) (see the
newly added textgreek-textcyr-test.lyx file in autotests/export/latex/).
I prepared two alternative patches and attached them to #9681.
The second patch also fixes
// FIXME: modifying i here is not very nice...
in Paragraph::Private::writeScriptChars().
As this is a rather large change, I'd prefer a second pair of eyes having a
look before committing.
Günter
