intermediate changes

ref:d5f945ecdc1f5af1ad57e12787c6b8ed1a9f0f12

7 files changed, 214 insertions, 0 deletions

diff --git a/library/cpp/case_insensitive_string/case_insensitive_char_traits.cpp b/library/cpp/case_insensitive_string/case_insensitive_char_traits.cpp
new file mode 100644
index 0000000000..14e6d1d51f
--- /dev/null
+++ b/library/cpp/case_insensitive_string/case_insensitive_char_traits.cpp
@@ -0,0 +1,34 @@
+#include "case_insensitive_char_traits.h"
+#include "case_insensitive_string.h"
+
+#include <util/string/escape.h>
+
+int TCaseInsensitiveCharTraits::compare(const char* s1, const char* s2, std::size_t n) {
+    while (n-- != 0) {
+        if (to_upper(*s1) < to_upper(*s2)) {
+            return -1;
+        }
+        if (to_upper(*s1) > to_upper(*s2)) {
+            return 1;
+        }
+        ++s1;
+        ++s2;
+    }
+    return 0;
+}
+
+const char* TCaseInsensitiveCharTraits::find(const char* s, std::size_t n, char a) {
+    auto const ua(to_upper(a));
+    while (n-- != 0) {
+        if (to_upper(*s) == ua)
+            return s;
+        s++;
+    }
+    return nullptr;
+}
+
+TCaseInsensitiveString EscapeC(const TCaseInsensitiveString& str) {
+    const auto result = EscapeC(str.data(), str.size());
+    return {result.data(), result.size()};
+}
+
diff --git a/library/cpp/case_insensitive_string/case_insensitive_char_traits.h b/library/cpp/case_insensitive_string/case_insensitive_char_traits.h
new file mode 100644
index 0000000000..2717893c10
--- /dev/null
+++ b/library/cpp/case_insensitive_string/case_insensitive_char_traits.h
@@ -0,0 +1,30 @@
+#pragma once
+
+#include <contrib/libs/libc_compat/string.h>
+
+#include <string>
+
+struct TCaseInsensitiveCharTraits : private std::char_traits<char> {
+    static bool eq(char c1, char c2) {
+        return to_upper(c1) == to_upper(c2);
+    }
+
+    static bool lt(char c1, char c2) {
+        return to_upper(c1) < to_upper(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);
+
+    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;
+
+private:
+    static char to_upper(char ch) {
+        return std::toupper((unsigned char)ch);
+    }
+};
diff --git a/library/cpp/case_insensitive_string/case_insensitive_string.cpp b/library/cpp/case_insensitive_string/case_insensitive_string.cpp
new file mode 100644
index 0000000000..16c0f5ff7a
--- /dev/null
+++ b/library/cpp/case_insensitive_string/case_insensitive_string.cpp
@@ -0,0 +1,22 @@
+#include "case_insensitive_string.h"
+
+#include <library/cpp/digest/murmur/murmur.h>
+
+size_t THash<TCaseInsensitiveStringBuf>::operator()(TCaseInsensitiveStringBuf str) const noexcept {
+    TMurmurHash2A<size_t> hash;
+    for (size_t i = 0; i < str.size(); ++i) {
+        char lower = std::tolower(str[i]);
+        hash.Update(&lower, 1);
+    }
+    return hash.Value();
+}
+
+template <>
+void Out<TCaseInsensitiveString>(IOutputStream& o, const TCaseInsensitiveString& p) {
+    o.Write(p.data(), p.size());
+}
+
+template <>
+void Out<TCaseInsensitiveStringBuf>(IOutputStream& o, const TCaseInsensitiveStringBuf& p) {
+    o.Write(p.data(), p.size());
+}
diff --git a/library/cpp/case_insensitive_string/case_insensitive_string.h b/library/cpp/case_insensitive_string/case_insensitive_string.h
new file mode 100644
index 0000000000..443de3e5f9
--- /dev/null
+++ b/library/cpp/case_insensitive_string/case_insensitive_string.h
@@ -0,0 +1,37 @@
+#pragma once
+
+#include "case_insensitive_char_traits.h"
+
+#include <util/generic/strbuf.h>
+#include <util/generic/string.h>
+#include <util/generic/hash.h>
+#include <util/string/split.h>
+
+using TCaseInsensitiveString = TBasicString<char, TCaseInsensitiveCharTraits>;
+using TCaseInsensitiveStringBuf = TBasicStringBuf<char, TCaseInsensitiveCharTraits>;
+
+template <>
+struct THash<TCaseInsensitiveStringBuf> {
+    size_t operator()(TCaseInsensitiveStringBuf str) const noexcept;
+};
+
+template <>
+struct THash<TCaseInsensitiveString> : THash<TCaseInsensitiveStringBuf> {};
+
+namespace NStringSplitPrivate {
+
+    template<>
+    struct TStringBufOfImpl<TCaseInsensitiveStringBuf> {
+        /*
+         * WARN:
+         * StringSplitter does not use TCharTraits properly.
+         * Splitting such strings is explicitly disabled.
+         */
+        // using type = TCaseInsensitiveStringBuf;
+    };
+
+    template<>
+    struct TStringBufOfImpl<TCaseInsensitiveString> : TStringBufOfImpl<TCaseInsensitiveStringBuf> {
+    };
+
+} // namespace NStringSplitPrivate
diff --git a/library/cpp/case_insensitive_string/case_insensitive_string_ut.cpp b/library/cpp/case_insensitive_string/case_insensitive_string_ut.cpp
new file mode 100644
index 0000000000..49f9c59c95
--- /dev/null
+++ b/library/cpp/case_insensitive_string/case_insensitive_string_ut.cpp
@@ -0,0 +1,65 @@
+#include "case_insensitive_string.h"
+
+#include <util/generic/string_ut.h>
+
+class TCaseInsensitiveStringTest : public TTestBase, private TStringTestImpl<TCaseInsensitiveString, TTestData<char>> {
+public:
+    UNIT_TEST_SUITE(TCaseInsensitiveStringTest);
+    UNIT_TEST(TestOperators);
+    UNIT_TEST(TestOperatorsCI);
+
+    UNIT_TEST_SUITE_END();
+};
+
+UNIT_TEST_SUITE_REGISTRATION(TCaseInsensitiveStringTest);
+
+Y_UNIT_TEST_SUITE(TCaseInsensitiveStringTestEx) {
+    Y_UNIT_TEST(BasicTString) {
+        TCaseInsensitiveString foo("foo");
+        TCaseInsensitiveString FOO("FOO");
+        TCaseInsensitiveString Bar("Bar");
+        TCaseInsensitiveString bAR("bAR");
+
+        UNIT_ASSERT_EQUAL(foo, FOO);
+        UNIT_ASSERT_EQUAL(Bar, bAR);
+
+        constexpr TCaseInsensitiveStringBuf foobar("foobar");
+        UNIT_ASSERT(foobar.StartsWith(foo));
+        UNIT_ASSERT(foobar.StartsWith(FOO));
+        UNIT_ASSERT(foobar.EndsWith(Bar));
+        UNIT_ASSERT(foobar.EndsWith(bAR));
+        UNIT_ASSERT(foobar.Contains(FOO));
+        UNIT_ASSERT(foobar.Contains(Bar));
+    }
+
+    Y_UNIT_TEST(BasicStdString) {
+        using TCaseInsensitiveStdString = std::basic_string<char, TCaseInsensitiveCharTraits>;
+        using TCaseInsensitiveStringView = std::basic_string_view<char, TCaseInsensitiveCharTraits>;
+
+        TCaseInsensitiveStdString foo("foo");
+        TCaseInsensitiveStdString FOO("FOO");
+        TCaseInsensitiveStdString Bar("Bar");
+        TCaseInsensitiveStdString bAR("bAR");
+
+        UNIT_ASSERT_EQUAL(foo, FOO);
+        UNIT_ASSERT_EQUAL(Bar, bAR);
+
+        constexpr TCaseInsensitiveStringView foobar("foobar");
+        UNIT_ASSERT(foobar.starts_with(foo));
+        UNIT_ASSERT(foobar.starts_with(FOO));
+        UNIT_ASSERT(foobar.ends_with(Bar));
+        UNIT_ASSERT(foobar.ends_with(bAR));
+        //TODO: test contains after C++23
+    }
+
+/*
+    Y_UNIT_TEST(TestSplit) {
+        TCaseInsensitiveStringBuf input("splitAmeAbro");
+        TVector<TCaseInsensitiveStringBuf> expected{"split", "me", "bro"};
+
+        TVector<TCaseInsensitiveStringBuf> split = StringSplitter(input).Split('a');
+
+        UNIT_ASSERT_VALUES_EQUAL(split, expected);
+    }
+*/
+}
diff --git a/library/cpp/case_insensitive_string/ut/ya.make b/library/cpp/case_insensitive_string/ut/ya.make
new file mode 100644
index 0000000000..b209d4571e
--- /dev/null
+++ b/library/cpp/case_insensitive_string/ut/ya.make
@@ -0,0 +1,9 @@
+OWNER(eeight)
+
+UNITTEST_FOR(library/cpp/case_insensitive_string)
+
+SRCS(
+    case_insensitive_string_ut.cpp
+)
+
+END()
diff --git a/library/cpp/case_insensitive_string/ya.make b/library/cpp/case_insensitive_string/ya.make
new file mode 100644
index 0000000000..dfca415e6b
--- /dev/null
+++ b/library/cpp/case_insensitive_string/ya.make
@@ -0,0 +1,17 @@
+OWNER(eeight)
+
+LIBRARY()
+
+SRCS(
+    case_insensitive_char_traits.cpp
+    case_insensitive_string.cpp
+)
+
+PEERDIR(
+    contrib/libs/libc_compat
+    library/cpp/digest/murmur
+)
+
+END()
+
+RECURSE_FOR_TESTS(ut)