CODEWITHSUNDEEP
c++encodingutf-16icunana
inzanez
inzanez

Reputation: 366

C++ using ICU and Nana GUI Library - String conversion?

I just did some successful tests with ICU in C/C++. I need to parse different CSV files with different encodings (might be UTF-8, UTF-16LE, ), do some modifications on the data and finally output everything as UTF-8 into a file. That's why my choice fell for ICU. Character set detection works pretty well usually, character handling and conversion to UTF-8 too.

Now I wanted to integrate that library part that does CSV loading, manipulation and so on with a GUI library, Nana. Nana seems to use std::string and std::wstring.

As ICU stores all data internally as UTF-16, so either I got UChars or UnicodeStrings when working with ICU. But how could I use either of them with Nana, that doesn't 'integrate' with ICU? Any way to transform UChar arrays to wstring, or a UnicodeString to wstring?

Didn't find any hints in the ICU documentation, so...maybe somebody else already worked on that?

Upvotes: 1

Views: 520

Answers (2)

qPCR4vir
qPCR4vir

Reputation: 3571

Most nana functions expect std::string encoded in UTF-8.

You could use the ICU functions that take or return char * to do the conversion to UTF-8.

A few of nana functions, like widget::caption have overloads for std::wstring expected to be encoded in UTF-16 (in windows) or UTF-32 (in Linux) which could be used to pass to the OS what could be the string with the native character type and encoding. In case you need conversions nana offers nana::charset which can manage (explicitly or implicitly) some of the most frequently needed conversions from/to UTF-8/UTF-16/UTF-32.

If you experiment passing static_cast<wchar_t *>(some_UChar*) to nana, please tell us about the result. I can't test.

The nana documentation about Unicode treatment urgently need to be updated (mea culpa)

Upvotes: 1

Serge Ballesta
Serge Ballesta

Reputation: 149075

According to ICU documentation, a UChar array is an array of 16 bits wide characters... meaning a wchar_t array in common implementations. That means that provided wchar_t is 16 bits wide in your system, you can safely cast the result of the getTerminatedBuffer() function to a const wchar_t * and either use it directly as a C wide chararcter string, or use it to build a std::wstring.

Upvotes: 0

Related Questions