Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
| >>>>> "Mike" == Mike Peters <[EMAIL PROTECTED]> writes:
|
| Mike> The problem arises because the Turkish language has two letter
| Mike> i's, a 'normal' one and a dotless i. A 'normal', English capital
| Mike> i is formed with Shift-dotless i (dotless i is in the position
| Mike> of a normal i on the keyboard) and Shift-i (dotted i is next to
| Mike> the return key on my keyboard) gives a dotted capital i (at
| Mike> least in programs that support Turkish). When I open a file,
| Mike> let's say Help=>Tutorial, it tries to read from book.layout but
| Mike> it gets stuck at the line 'Input stdclass.inc' and so then tries
| Mike> to read from artical.layout and the program exits claiming it
| Mike> doesn't recognise the same Input tag in that file. After much
| Mike> experimentation I eventually found that, by changing Input in
| Mike> book.layout to input, lyx recognises the tag and goes on to read
| Mike> from stdclass.inc but again it gets hung up on the next tag
| Mike> containing a capital i (ParIdent in this case).
|
| It seems that, when the turkish locale is in force, the tolower()
| function does not match 'I' to 'i' anymore. Does setting the
| environment variable LC_COLLATE to "C" help?
I suspect that we will not have a good solution for this until we can
begin using Standar C++ locale support and read all files with the
classic locale ('C')
Lgb
