alx/mu4e
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

utils: add utf8_clean helper

Browse Source 
To morph text into valid utf8 if it isn't already.
This commit is contained in:
Dirk-Jan C. Binnema
2025-10-25 16:53:57 +03:00
committed by Seth Ladygo
parent 78539e8900
commit 5c6642c9b9
2 changed files with 38 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
lib/utils/mu-utils.hh
View File

@ -164,6 +164,24 @@ static inline bool contains_unbroken_script(const std::string& str) {
return contains_unbroken_script(str.c_str());
}
/**
* If the string is already valid utf8, return it
* otherwise, return a valid utf8 version
*
* @param str some string
*
* @return a utf8-string
*/
static inline std::string utf8_clean(std::string&& str) {
if (!g_utf8_validate(str.c_str(), static_cast<gssize>(str.length()), {})) {
gchar* clean{g_utf8_make_valid(
str.c_str(), static_cast<gssize>(str.length()))};
str = clean;
g_free(clean);
}
return std::move(str);
}
/**
* Flatten a string -- down-case and fold diacritics.
*
@ -172,7 +190,7 @@ static inline bool contains_unbroken_script(const std::string& str) {
* @return a flattened string
*/
std::string utf8_flatten(const char* str);
inline std::string
static inline std::string
utf8_flatten(const std::string& s) {
return utf8_flatten(s.c_str());
}