Hi!
After discovering that I cannot enter in LyX all the required composed
characters for my Sanskrit needs, I have decided to use some Perl script
for entering them.
Everything was fine until today when I discovered some weird LyX
behaviour when trying to index my book with Xindy:
Here are some of the characters I need for Sanskrit.
\i \d{h}
\i \d{t}
\i \d{d}
\i \d{r}
\i \d{n}
\i \.{m}
\i \'{s}
\i \={a}
\i \={\i}
Those made up with underdot accent + letter can be entered directly in
LyX, while e.g. those with macron (\={*}) cannot.
If I manually enter them as e.g.: \i \={a}, they display strange on the
screen (there is no macron accent in latin-2 encoding), but fine in DVI
and on the paper.
Strange thing is that if I enter \i \={i} for "i- macron", and load
file, everything is OK, until I save file again within LyX.
Afterwards, instead of \i \={i}, LyX saves it internally as:
\i \={\i}.
I have discovered it when I found out in Xindy log file that my regex
doesn't match since they compare {\i} with {i}.
Is it normal behaviour in LyX or bug?
I consider it a bug since *.tex file has i-macron presented as \={\i}.
Sincerely,
Sasa
