Go back to libc locale and use c_regex_traits

Unfortunately, cygwin does not support c++ locales.

src/highlighters.cc

@@ -884,7 +884,7 @@ void expand_unprintable(const Context& context, HighlightFlags flags, DisplayBuf
                 {
                     auto coord = it.coord();
                     Codepoint cp = utf8::read_codepoint<utf8::InvalidPolicy::Pass>(it, end);
-                    if (cp != '\n' and not std::isprint((wchar_t)cp, std::locale{}))
+                    if (cp != '\n' and not iswprint((wchar_t)cp))
                     {
                         if (coord != atom_it->begin())
                             atom_it = ++line.split(atom_it, coord);

src/main.cc

@@ -707,8 +707,7 @@ int run_pipe(StringView session)
 
 int main(int argc, char* argv[])
 {
-    try { std::locale::global(std::locale("")); }
+    setlocale(LC_ALL, "");
-    catch (std::runtime_error&) { setlocale(LC_ALL, ""); }
 
     set_signal_handler(SIGSEGV, signal_handler);
     set_signal_handler(SIGFPE,  signal_handler);

src/ranked_match.cc

@@ -34,7 +34,6 @@ using Utf8It = utf8::iterator<const char*>;
 
 static int count_word_boundaries_match(StringView candidate, StringView query)
 {
-    std::locale locale;
     int count = 0;
     Utf8It query_it{query.begin(), query};
     Codepoint prev = 0;
@@ -42,10 +41,8 @@ static int count_word_boundaries_match(StringView candidate, StringView query)
     {
         const Codepoint c = *it;
         const bool is_word_boundary = prev == 0 or
-                                      (!std::isalnum((wchar_t)prev, locale) and
+                                      (!iswalnum((wchar_t)prev) and iswalnum((wchar_t)c)) or
-                                       std::isalnum((wchar_t)c, locale)) or
+                                      (iswlower((wchar_t)prev) and iswupper((wchar_t)c));
-                                      (std::islower((wchar_t)prev, locale) and
-                                       std::isupper((wchar_t)c, locale));
         prev = c;
 
         if (not is_word_boundary)

src/regex.cc

@@ -8,7 +8,7 @@ namespace Kakoune
 using Utf8It = RegexUtf8It<const char*>;
 
 Regex::Regex(StringView re, flag_type flags) try
-    : boost::wregex{Utf8It{re.begin(), re}, Utf8It{re.end(), re}}, m_str(re.str())
+    : RegexBase{Utf8It{re.begin(), re}, Utf8It{re.end(), re}}, m_str(re.str())
 {} catch (std::runtime_error& err) { throw regex_error(err.what()); }
 
 String option_to_string(const Regex& re)

src/regex.hh

@@ -17,8 +17,10 @@ struct regex_error : runtime_error
     {}
 };
 
+using RegexBase = boost::basic_regex<wchar_t, boost::c_regex_traits<wchar_t>>;
+
 // Regex that keeps track of its string representation
-struct Regex : boost::wregex
+struct Regex : RegexBase
 {
     Regex() = default;
 
@@ -36,6 +38,10 @@ private:
 template<typename It>
 using RegexUtf8It = utf8::iterator<It, wchar_t, ssize_t>;
 
+template<typename It>
+using RegexIteratorBase = boost::regex_iterator<RegexUtf8It<It>, wchar_t,
+                                                boost::c_regex_traits<wchar_t>>;
+
 namespace RegexConstant = boost::regex_constants;
 
 template<typename Iterator>
@@ -70,19 +76,18 @@ struct MatchResults : boost::match_results<RegexUtf8It<Iterator>>
 };
 
 template<typename Iterator>
-struct RegexIterator : boost::regex_iterator<RegexUtf8It<Iterator>>
+struct RegexIterator : RegexIteratorBase<Iterator>
 {
-    using ParentType = boost::regex_iterator<RegexUtf8It<Iterator>>;
     using Utf8It = RegexUtf8It<Iterator>;
     using ValueType = MatchResults<Iterator>;
 
     RegexIterator() = default;
     RegexIterator(Iterator begin, Iterator end, const Regex& re,
                   RegexConstant::match_flag_type flags = RegexConstant::match_default)
-        : ParentType{Utf8It{begin, begin, end}, Utf8It{end, begin, end}, re, flags} {}
+        : RegexIteratorBase<Iterator>{Utf8It{begin, begin, end}, Utf8It{end, begin, end}, re, flags} {}
 
-    const ValueType& operator*() const { return *reinterpret_cast<const ValueType*>(&ParentType::operator*()); }
+    const ValueType& operator*() const { return *reinterpret_cast<const ValueType*>(&RegexIteratorBase<Iterator>::operator*()); }
-    const ValueType* operator->() const { return reinterpret_cast<const ValueType*>(ParentType::operator->()); }
+    const ValueType* operator->() const { return reinterpret_cast<const ValueType*>(RegexIteratorBase<Iterator>::operator->()); }
 };
 
 inline RegexConstant::match_flag_type match_flags(bool bol, bool eol, bool bow, bool eow)

src/unicode.hh

@@ -29,7 +29,7 @@ enum WordType { Word, WORD };
 template<WordType word_type = Word>
 inline bool is_word(Codepoint c)
 {
-    return c == '_' or std::isalnum((wchar_t)c, std::locale{});
+    return c == '_' or iswalnum((wchar_t)c);
 }
 
 template<>
@@ -68,8 +68,8 @@ inline CharCategories categorize(Codepoint c)
     return CharCategories::Punctuation;
 }
 
-inline Codepoint to_lower(Codepoint cp) { return std::tolower((wchar_t)cp, std::locale{}); }
+inline Codepoint to_lower(Codepoint cp) { return towlower((wchar_t)cp); }
-inline Codepoint to_upper(Codepoint cp) { return std::toupper((wchar_t)cp, std::locale{}); }
+inline Codepoint to_upper(Codepoint cp) { return towupper((wchar_t)cp); }
 
 inline char to_lower(char c) { return c >= 'A' and c <= 'Z' ? c - 'A' + 'a' : c; }
 inline char to_upper(char c) { return c >= 'a' and c <= 'z' ? c - 'a' + 'A' : c; }