Std::string NLS_LANG character sets

Question

I'm an OCI user but not a pure one. Because I use the free available OTL (Oracle Template Library) from S.Kuchin which is a wrapper around OIC I hope this not off topic here.
The library offers the possibility to read database strings from VARCHAR2 fields to
a std::string. I know that oracle does character converting at client side controlled
via NLS_LANG environment variable.
It's clear to me that reading database strings to std::string is no problem for
one byte character sets liike ISO-8859-1. But how about when I let point NLS_LANG
to UTF-8 or chinese character set e.g. ZHT16BIG5 ?
Is it still safe to read the result to a std::string ?
For example I have a database with default characterset AMERICAN_AMERICA.WE8ISO8859P15. I stored some German umlaut in some
table. I wrote a small program using OTL and set NLS_LANG to UTF8 on client side.
I fetched the data from server to client, stored the data in a std::string, pushed them in a file and yes the data were stored as UTF8.
Is it really so simple or is it dangerous to read the UTF8-converted data to
std::string ? Wat is the common rule ? When may and when may I not read the data
to std::string ?

Answer 1

Hello,
I think you'll have more accurate answer in the Globalization Support Forum:
Globalization Support
Best regards,
Jean-Valentin