syntax-based normalization

fix #8, fix boostorg#65

include/boost/url/detail/impl/pct_encoding.ipp

@@ -81,6 +81,106 @@ key_equal_encoded(
     }
 }
 
+int
+pct_decode_compare_unchecked(
+    string_view lhs,
+    string_view rhs) noexcept
+{
+    auto consume_one = []
+        (string_view::iterator& it,
+         char &c,
+         std::size_t& n)
+    {
+        if(*it != '%')
+        {
+            c = *it;
+            ++it;
+        }
+        else
+        {
+            pct_decode_unchecked(
+                &c,
+                &c + 1,
+                string_view(it, 3));
+            it += 3;
+        }
+        ++n;
+    };
+
+    std::size_t lhs_n = 0;
+    std::size_t rhs_n = 0;
+    auto lhs_it = lhs.begin();
+    auto rhs_it = rhs.begin();
+    char lhs_c = '\0';
+    char rhs_c = '\0';
+    while (lhs_it < lhs.end() &&
+           rhs_it < rhs.end())
+    {
+        consume_one(lhs_it, lhs_c, lhs_n);
+        consume_one(rhs_it, rhs_c, rhs_n);
+        if ( lhs_c < rhs_c )
+            return -1;
+        else if ( rhs_c < lhs_c )
+            return +1;
+    }
+    if ( lhs_n == rhs_n )
+        return 0;
+    if ( lhs_n < rhs_n )
+        return -1;
+    return +1;
+}
+
+int
+pct_decode_icompare_unchecked(
+    string_view lhs,
+    string_view rhs) noexcept
+{
+    auto consume_one = []
+        (string_view::iterator& it,
+         char &c,
+         std::size_t& n)
+    {
+        if(*it != '%')
+        {
+            c = grammar::ascii_tolower(*it);
+            ++it;
+        }
+        else
+        {
+            pct_decode_unchecked(
+                &c,
+                &c + 1,
+                string_view(it, 3));
+            c = grammar::ascii_tolower(c);
+            it += 3;
+        }
+        ++n;
+    };
+
+    std::size_t lhs_n = 0;
+    std::size_t rhs_n = 0;
+    auto lhs_it = lhs.begin();
+    auto rhs_it = rhs.begin();
+    char lhs_c = '\0';
+    char rhs_c = '\0';
+    while (lhs_it < lhs.end() &&
+           rhs_it < lhs.end())
+    {
+        consume_one(lhs_it, lhs_c, lhs_n);
+        consume_one(rhs_it, rhs_c, rhs_n);
+        if ( lhs_c < rhs_c )
+            return -1;
+        else if ( rhs_c < lhs_c )
+            return +1;
+    }
+    if ( lhs_n == rhs_n )
+        return 0;
+    if ( lhs_n < rhs_n )
+        return -1;
+    return +1;
+}
+
+
 } // detail
 } // urls
 } // boost

include/boost/url/detail/impl/remove_dot_segments.hpp

@@ -0,0 +1,43 @@
+//
+// Copyright (c) 2022 alandefreitas (alandefreitas@gmail.com)
+//
+// Distributed under the Boost Software License, Version 1.0. (See accompanying
+// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+//
+// Official repository: https://github.com/CPPAlliance/url
+//
+
+#ifndef BOOST_URL_DETAIL_IMPL_REMOVE_DOT_SEGMENTS_HPP
+#define BOOST_URL_DETAIL_IMPL_REMOVE_DOT_SEGMENTS_HPP
+
+namespace boost {
+namespace urls {
+namespace detail {
+
+template<class Allocator>
+std::basic_string<char,
+                  std::char_traits<char>,
+                  Allocator>
+remove_dot_segments(
+    string_view s,
+    Allocator const& a,
+    bool remove_leading_dot_dot)
+{
+    using string_type = std::basic_string<
+        char,
+        std::char_traits<char>,
+        Allocator>;
+    string_type output(s.size(), '_', a);
+    char* it = &output[0];
+    char* end = it + output.size();
+    std::size_t n = remove_dot_segments(
+        it, end, s, remove_leading_dot_dot);
+    output.resize(n);
+    return output;
+}
+
+}
+}
+}
+
+#endif