Add Add url::get_part_pos function

upa-url · Dec 31, 2024 · 337ae54 · 337ae54
1 parent 7360681
commit 337ae54
Show file tree

Hide file tree

Showing 3 changed files with 82 additions and 10 deletions.
diff --git a/include/upa/url.h b/include/upa/url.h
@@ -542,19 +542,38 @@ class url {
     /// @return `true` if URL is valid, `false` otherwise
     [[nodiscard]] bool is_valid() const noexcept;
 
+    /// @brief Gets the start and end position of the specified URL part
+    ///
+    /// Returns the start and end position of the part (as defined in
+    /// https://url.spec.whatwg.org/#url-representation) in the string returned by the
+    /// `get_href()`, `href()` or `to_string()` functions.
+    ///
+    /// * `get_part_pos(upa::url::SCHEME)` - get a URL's **scheme** position
+    /// * `get_part_pos(upa::url::USERNAME)` - get a URL's **username** position
+    /// * `get_part_pos(upa::url::PASSWORD)` - get a URL's **password** position
+    /// * `get_part_pos(upa::url::HOST)` - get a URL's **host** position
+    /// * `get_part_pos(upa::url::PORT)` - get a URL's **port** position
+    /// * `get_part_pos(upa::url::PATH)` - get a URL's **path** position
+    /// * `get_part_pos(upa::url::QUERY)` - get a URL's **query** position
+    /// * `get_part_pos(upa::url::FRAGMENT)` - get a URL's **fragment** position
+    ///
+    /// @param[in] t URL's part
+    /// @return the start and end position of the specified URL part
+    [[nodiscard]] std::pair<std::size_t, std::size_t> get_part_pos(PartType t) const;
+
     /// @brief Gets URL's part (URL record member) as string
     ///
     /// Function to get ASCII string of any URL's part (URL record member) defined here:
     /// https://url.spec.whatwg.org/#url-representation
     ///
-    /// * `get_part_view(upa::url::SCHEME)` - get a URL’s **scheme** string
-    /// * `get_part_view(upa::url::USERNAME)` - get a URL’s **username** string
-    /// * `get_part_view(upa::url::PASSWORD)` - get a URL’s **password** string
-    /// * `get_part_view(upa::url::HOST)` - get a URL’s **host** serialized to string
-    /// * `get_part_view(upa::url::PORT)` - get a URL’s **port** serialized to string
-    /// * `get_part_view(upa::url::PATH)` - get a URL’s **path** serialized to string
-    /// * `get_part_view(upa::url::QUERY)` - get a URL’s **query** string
-    /// * `get_part_view(upa::url::FRAGMENT)` - get a URL’s **fragment** string
+    /// * `get_part_view(upa::url::SCHEME)` - get a URL's **scheme** string
+    /// * `get_part_view(upa::url::USERNAME)` - get a URL's **username** string
+    /// * `get_part_view(upa::url::PASSWORD)` - get a URL's **password** string
+    /// * `get_part_view(upa::url::HOST)` - get a URL's **host** serialized to string
+    /// * `get_part_view(upa::url::PORT)` - get a URL's **port** serialized to string
+    /// * `get_part_view(upa::url::PATH)` - get a URL's **path** serialized to string
+    /// * `get_part_view(upa::url::QUERY)` - get a URL's **query** string
+    /// * `get_part_view(upa::url::FRAGMENT)` - get a URL's **fragment** string
     ///
     /// @param[in] t URL's part
     /// @return URL's part string; it is empty if part is empty or null

diff --git a/src/url.cpp b/src/url.cpp
@@ -1,4 +1,4 @@
-// Copyright 2016-2023 Rimas Misevičius
+// Copyright 2016-2024 Rimas Misevičius
 // Distributed under the BSD-style license that can be
 // found in the LICENSE file.
 //
@@ -27,8 +27,21 @@ const uint8_t kPartStart[url::PART_COUNT] = {
 
 } // namespace detail
 
+// Gets start and end position of the specified URL part
 
-  // URL's part flag masks
+std::pair<std::size_t, std::size_t> url::get_part_pos(PartType t) const {
+    if (t == SCHEME)
+        return { 0, part_end_[SCHEME] };
+    const std::size_t e = part_end_[t];
+    if (e) {
+        std::size_t b = part_end_[t - 1];
+        if (b < e) b += detail::kPartStart[t];
+        return { b, e };
+    }
+    return { norm_url_.size(), norm_url_.size() };
+}
+
+// URL's part flag masks
 
 const unsigned url::kPartFlagMask[url::PART_COUNT] = {
     SCHEME_FLAG,

diff --git a/test/test-url.cpp b/test/test-url.cpp
@@ -8,6 +8,20 @@
 #include "test-utils.h"
 #include <unordered_map>
 
+// Conversion to doctest String
+// https://github.com/doctest/doctest/blob/master/doc/markdown/stringification.md
+
+namespace doctest {
+    template <class T1, class T2>
+    struct StringMaker<std::pair<T1, T2>> {
+        static String convert(const std::pair<T1, T2>& value) {
+            std::stringstream strs;
+            strs << "(" << value.first << ", " << value.second << ")";
+            const auto str = strs.str();
+            return { str.data(), static_cast<String::size_type>(str.length()) };
+        }
+    };
+}
 
 template<class T>
 void discard(T&&) {}
@@ -462,6 +476,32 @@ TEST_CASE("url::has_opaque_path") {
 
 // URL parts
 
+TEST_CASE("url::get_part_pos") {
+    upa::url url{ "s://u:p@h:1/?q#f" };
+
+    CHECK(url.get_part_pos(upa::url::SCHEME) == std::make_pair(std::size_t(0), std::size_t(1)));
+    CHECK(url.get_part_pos(upa::url::USERNAME) == std::make_pair(std::size_t(4), std::size_t(5)));
+    CHECK(url.get_part_pos(upa::url::PASSWORD) == std::make_pair(std::size_t(6), std::size_t(7)));
+    CHECK(url.get_part_pos(upa::url::HOST) == std::make_pair(std::size_t(8), std::size_t(9)));
+    CHECK(url.get_part_pos(upa::url::PORT) == std::make_pair(std::size_t(10), std::size_t(11)));
+    CHECK(url.get_part_pos(upa::url::PATH) == std::make_pair(std::size_t(11), std::size_t(12)));
+    CHECK(url.get_part_pos(upa::url::QUERY) == std::make_pair(std::size_t(13), std::size_t(14)));
+    CHECK(url.get_part_pos(upa::url::FRAGMENT) == std::make_pair(std::size_t(15), std::size_t(16)));
+
+    url.hash("");
+    CHECK(url.get_part_pos(upa::url::FRAGMENT) == std::make_pair(std::size_t(14), std::size_t(14)));
+
+    url.search("");
+    CHECK(url.get_part_pos(upa::url::QUERY) == std::make_pair(std::size_t(12), std::size_t(12)));
+    CHECK(url.get_part_pos(upa::url::FRAGMENT) == std::make_pair(std::size_t(12), std::size_t(12)));
+
+    url.parse("s://u:p@h:1");
+    CHECK(url.href() == "s://u:p@h:1");
+    CHECK(url.get_part_pos(upa::url::PATH) == std::make_pair(std::size_t(11), std::size_t(11)));
+    CHECK(url.get_part_pos(upa::url::QUERY) == std::make_pair(std::size_t(11), std::size_t(11)));
+    CHECK(url.get_part_pos(upa::url::FRAGMENT) == std::make_pair(std::size_t(11), std::size_t(11)));
+}
+
 TEST_CASE("url::is_empty and url::is_null") {
     upa::url url;