miniutf is a C++ implementation of several basic Unicode manipulation functions.
UTF-8, UTF-16, UTF-32 (UCS-4)
miniutf can convert between UTF-8 (
std::string), UTF-16 (
UTF-32 / UCS-4 (
std::u32string). The C++11 standard library provides UTF-8 and -16 encoders
and decoders (
codecvt_utf16), but as of late 2013, libstdc++ doesn't
implement them. Miniutf's conversion functions also provide validity checking and can insert
replacement characters if invalid input is found.
miniutf implements conversion to NFC and NFD as defined in Unicode TR15. It does not implement NFKC or NFKD.
miniutf implements collation as defined by the Default Unicode Collation Element Table, level 1 (Unicode TR10). This requires a large data table, which adds to binary size, so it's in a separate source and header file. The collation function can be used for case- and accent-insensitive searching and sorting.
Unicode defines a one-to-one lowercase translation for each codepoint. (This is needed for Dropbox's internal use, but should be avoided otherwise. One-to-one lowercasing does not always match the lowercase rules of a given language, e.g. German eszett, Turkish dotless i.)
miniutf requires a recent C++11 compiler and standard library, such as Clang 3.3+ or GCC 4.8+.
MIT (see LICENSE.txt)