Document LC_CTYPE for language data

[jeroen]

Environment

• Tesseract Version: any
• Platform: all

There is a bug report about a problem with loading asian training data, which that might be a bigger problem. It is currently not possible to load asian languages jpn, kor, etc on a system where LANG and locale are set to en_US.UTF-8, which is often the default.

A workaround is to set LC_CTYPE to C before calling api->Init() but it is unclear why this is needed and what side effects this has. Also I don't know if it is safe to set it back to en_US.UTF-8 afterwards.

Some documentation on this would be great.

[zdenop]

Problem is that people do not read doc. Original issue was documented at "code.google era":

[jeroen]

The link you quote talks about LC_NUMERIC. I think this is different issue than LC_CTYPE for asian languages?

[stweil]

Is this a duplicate of issue #1250?