Add TCaseInsensitiveAsciiString

Followup for rXXXXXX - further optimize ascii-only case insensitive strings
1fca7889a074a191eadce12247bdd6dd18b75ab2

1 files changed, 40 insertions, 15 deletions

diff --git a/library/cpp/case_insensitive_string/case_insensitive_char_traits.h b/library/cpp/case_insensitive_string/case_insensitive_char_traits.h
index 2717893c10..fdf5f3a6c0 100644
--- a/library/cpp/case_insensitive_string/case_insensitive_char_traits.h
+++ b/library/cpp/case_insensitive_string/case_insensitive_char_traits.h
@@ -2,29 +2,54 @@
 
 #include <contrib/libs/libc_compat/string.h>
 
+#include <util/string/ascii.h>
+
 #include <string>
 
-struct TCaseInsensitiveCharTraits : private std::char_traits<char> {
-    static bool eq(char c1, char c2) {
-        return to_upper(c1) == to_upper(c2);
-    }
+namespace NPrivate {
+    template <typename TImpl>
+    struct TCommonCaseInsensitiveCharTraits : private std::char_traits<char> {
+        static bool eq(char c1, char c2) {
+            return TImpl::ToCommonCase(c1) == TImpl::ToCommonCase(c2);
+        }
 
-    static bool lt(char c1, char c2) {
-        return to_upper(c1) < to_upper(c2);
-    }
+        static bool lt(char c1, char c2) {
+            return TImpl::ToCommonCase(c1) < TImpl::ToCommonCase(c2);
+        }
 
-    static int compare(const char* s1, const char* s2, std::size_t n);
+        static const char* find(const char* s, std::size_t n, char a);
 
-    static const char* find(const char* s, std::size_t n, char a);
+        using std::char_traits<char>::assign;
+        using std::char_traits<char>::char_type;
+        using std::char_traits<char>::copy;
+        using std::char_traits<char>::length;
+        using std::char_traits<char>::move;
+    };
+} // namespace NPrivate
 
-    using std::char_traits<char>::assign;
-    using std::char_traits<char>::char_type;
-    using std::char_traits<char>::copy;
-    using std::char_traits<char>::length;
-    using std::char_traits<char>::move;
+struct TCaseInsensitiveCharTraits : public ::NPrivate::TCommonCaseInsensitiveCharTraits<TCaseInsensitiveCharTraits> {
+    static int compare(const char* s1, const char* s2, std::size_t n);
 
 private:
-    static char to_upper(char ch) {
+    friend ::NPrivate::TCommonCaseInsensitiveCharTraits<TCaseInsensitiveCharTraits>;
+
+    // XXX return unsigned char. Current impl depends on char signedness, and if char is signed,
+    // TCaseInsensitiveCharTraits::compare returns different result from std::char_traits<char>::compare for non-ascii strings.
+    static char ToCommonCase(char ch) {
         return std::toupper((unsigned char)ch);
     }
 };
+
+struct TCaseInsensitiveAsciiCharTraits : public ::NPrivate::TCommonCaseInsensitiveCharTraits<TCaseInsensitiveCharTraits> {
+    // WARN: does not work with null bytes (`compare("ab\0c", "ab\0d", 4)` returns 0).
+    static int compare(const char* s1, const char* s2, std::size_t n) {
+        return ::strncasecmp(s1, s2, n);
+    }
+
+private:
+    friend ::NPrivate::TCommonCaseInsensitiveCharTraits<TCaseInsensitiveAsciiCharTraits>;
+
+    static unsigned char ToCommonCase(char ch) {
+        return AsciiToLower(ch);
+    }
+};