From 357f37023f997a9a9bfdd45061e1f58a2d1b1776 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rimas=20Misevi=C4=8Dius?= <rmisev3@gmail.com>
Date: Thu, 23 Nov 2023 23:24:11 +0200
Subject: [PATCH] Disallow null characters in URL <-> path conversions

Check for null characters in url_from_file_path and path_from_file_url
functions.
---
 include/upa/url.h        | 9 +++++++++
 include/upa/url_result.h | 1 +
 include/upa/util.h       | 5 +++++
 test/test-url.cpp        | 8 ++++++++
 4 files changed, 23 insertions(+)

diff --git a/include/upa/url.h b/include/upa/url.h
index 0dd5147..dc930fe 100644
--- a/include/upa/url.h
+++ b/include/upa/url.h
@@ -3090,6 +3090,10 @@ inline url url_from_file_path(StrT&& str, file_path_format format = file_path_fo
         }
     }
 
+    // Check for null characters
+    if (util::contains_null(pointer, last))
+        throw url_error(validation_errc::null_character, "Path contains null character");
+
     // make URL
     detail::append_utf8_percent_encoded(pointer, last, *no_encode_set, str_url);
     return url(str_url);
@@ -3164,6 +3168,11 @@ inline std::string path_from_file_url(const url& file_url, file_path_format form
             }
         }
     }
+
+    // Check for null characters
+    if (util::contains_null(path.begin(), path.end()))
+        throw url_error(validation_errc::null_character, "Path contains null character");
+
     return path;
 }
 
diff --git a/include/upa/url_result.h b/include/upa/url_result.h
index a8adb98..43b5edf 100644
--- a/include/upa/url_result.h
+++ b/include/upa/url_result.h
@@ -84,6 +84,7 @@ enum class validation_errc {
     file_url_unsupported_host,      ///< UNC path cannot have "." hostname
     file_url_invalid_unc,           ///< Invalid UNC path in file URL
     file_url_not_windows_path,      ///< Not a Windows path in file URL
+    null_character,                 ///< Path contains null character
 };
 
 /// @brief Check validation error code indicates success
diff --git a/include/upa/util.h b/include/upa/util.h
index b9b1ea3..970e9aa 100644
--- a/include/upa/util.h
+++ b/include/upa/util.h
@@ -125,6 +125,11 @@ inline void append_ascii_lowercase(std::string& dest, const CharT* first, const
 
 // Finders
 
+template <class InputIt>
+inline bool contains_null(InputIt first, InputIt last) {
+    return std::find(first, last, '\0') != last;
+}
+
 template <class CharT>
 UPA_CONSTEXPR_17 bool has_xn_label(const CharT* first, const CharT* last) {
     if (last - first >= 4) {
diff --git a/test/test-url.cpp b/test/test-url.cpp
index e027093..65aa40f 100644
--- a/test/test-url.cpp
+++ b/test/test-url.cpp
@@ -605,6 +605,8 @@ TEST_CASE("url_from_file_path") {
         // non absolute path
         CHECK_THROWS_AS(upa::url_from_file_path("path"), upa::url_error);
         CHECK_THROWS_AS(upa::url_from_file_path("\\\\h\\p", upa::file_path_format::posix), upa::url_error);
+        // null character
+        CHECK_THROWS_AS(upa::url_from_file_path(std::string{ "/p\0", 3 }, upa::file_path_format::posix), upa::url_error);
     }
     SUBCASE("Windows path") {
         // https://learn.microsoft.com/en-us/windows/win32/fileio/naming-a-file
@@ -644,6 +646,8 @@ TEST_CASE("url_from_file_path") {
         // unsupported pathes
         CHECK_THROWS_AS(upa::url_from_file_path("\\\\?\\Volume{b75e2c83-0000-0000-0000-602f00000000}\\Test\\Foo.txt"), upa::url_error);
         CHECK_THROWS_AS(upa::url_from_file_path("\\\\.\\Volume{b75e2c83-0000-0000-0000-602f00000000}\\Test\\Foo.txt"), upa::url_error);
+        // null character
+        CHECK_THROWS_AS(upa::url_from_file_path(std::string{ "/C:/p\0", 6 }, upa::file_path_format::posix), upa::url_error);
     }
 }
 
@@ -660,6 +664,8 @@ TEST_CASE("path_from_file_url") {
         CHECK(path_from_file_url("file:///path", upa::file_path_format::posix) == "/path");
         // POSIX path cannot have host
         CHECK_THROWS_AS(path_from_file_url("file://host/path", upa::file_path_format::posix), upa::url_error);
+        // null character
+        CHECK_THROWS_AS(path_from_file_url("file:///p%00", upa::file_path_format::posix), upa::url_error);
     }
     SUBCASE("Windows path") {
         CHECK(path_from_file_url("file:///C:", upa::file_path_format::windows) == "C:\\");
@@ -685,6 +691,8 @@ TEST_CASE("path_from_file_url") {
         CHECK_THROWS_AS(path_from_file_url("file://///host/", upa::file_path_format::windows), upa::url_error);
         // Unsupported "." hostname
         CHECK_THROWS_AS(path_from_file_url("file://./name", upa::file_path_format::windows), upa::url_error);
+        // null character
+        CHECK_THROWS_AS(path_from_file_url("file:///C:/p%00", upa::file_path_format::posix), upa::url_error);
     }
     SUBCASE("Native path") {
 #ifdef _WIN32