diff options
author | Vasudev Kamath <vasudev@copyninja.info> | 2016-10-23 21:09:40 +0530 |
---|---|---|
committer | Vasudev Kamath <vasudev@copyninja.info> | 2016-10-23 21:09:40 +0530 |
commit | eda4b75b1cff83336e87da90efca9fd6b4ced2c7 (patch) | |
tree | 491317ce50b5d19bc434ccc4b448d1bc70520177 /louloulibs/utils/encoding.hpp | |
parent | 716c40e4ec45f8d538695225f4f06d541d959084 (diff) | |
parent | 0f14fe83ef53b08bd8fa09670c82f4996c329bdc (diff) | |
download | biboumi-upstream/3.0.tar.gz biboumi-upstream/3.0.tar.bz2 biboumi-upstream/3.0.tar.xz biboumi-upstream/3.0.zip |
New upstream version 3.0upstream/3.0
Diffstat (limited to 'louloulibs/utils/encoding.hpp')
-rw-r--r-- | louloulibs/utils/encoding.hpp | 43 |
1 files changed, 43 insertions, 0 deletions
diff --git a/louloulibs/utils/encoding.hpp b/louloulibs/utils/encoding.hpp new file mode 100644 index 0000000..586edd8 --- /dev/null +++ b/louloulibs/utils/encoding.hpp @@ -0,0 +1,43 @@ +#pragma once + + +#include <string> + +namespace utils +{ + /** + * Return the size, in bytes, of the next UTF-8 codepoint, based on + * the given char. + */ + std::size_t get_next_codepoint_size(const unsigned char c); + /** + * Returns true if the given null-terminated string is valid utf-8. + * + * Based on http://en.wikipedia.org/wiki/UTF-8#Description + */ + bool is_valid_utf8(const char* s); + /** + * Remove all invalid codepoints from the given utf-8-encoded string. + * The value returned is a copy of the string, without the removed chars. + * + * See http://www.w3.org/TR/xml/#charsets for the list of valid characters + * in XML. + */ + std::string remove_invalid_xml_chars(const std::string& original); + /** + * Convert the given string (encoded is "encoding") into valid utf-8. + * If some decoding fails, insert an utf-8 placeholder character instead. + */ + std::string convert_to_utf8(const std::string& str, const char* encoding); +} + +namespace xep0106 +{ + /** + * Decode and encode inplace. + */ + void decode(std::string&); + void encode(std::string&); +} + + |