Handle ports, paths, and credentials correctly

lib/Misc/URL.php

@@ -6,7 +6,37 @@
 declare(strict_types=1);
 namespace JKingWeb\Arsse\Misc;
 
+/**
+ * A collection of functions for manipulating URLs
+ */
 class URL {
+    /** User component */
+    const P_USER = 1;
+    /** Password component */
+    const P_PASS = 2;
+    /** Path segment component */
+    const P_PATH = 3;
+    /** Full query component  */
+    const P_QUERY = 4;
+
+    /** Normalizes an absolute URL
+     * 
+     * Normalizations performed are:
+     * 
+     * - Lowercasing scheme
+     * - Lowercasing host names
+     * - IDN normalization (IDN rather than punycode is returned)
+     * - IPv6 address normalization
+     * - Resolution of relative path segments
+     * - Discarding empty path segments
+     * - Discarding empty queries
+     * - %-encoding normalization
+     * - Fragment discarding
+     * 
+     * @param string $url The URL to normalize. Relative URLs are returned unchanged
+     * @param string $u Username to add to the URL, replacing any existing credentials
+     * @param string $p Password to add to the URL, if a username is specified
+     */
     public static function normalize(string $url, string $u = null, string $p = null): string {
         extract(parse_url($url));
         if (!isset($scheme) || !isset($host) || !strlen($host)) {
@@ -14,15 +44,15 @@ class URL {
         }
         $out = strtolower($scheme)."://";
         if (strlen($u ?? "")) {
-            $out .= self::normalizePart($u, self::P_USER, false);
+            $out .= self::normalizePart(rawurlencode($u), self::P_USER, false);
             if (strlen($p ?? "")) {
-                $out .= ":".self::normalizePart($p, self::P_PASS, false);
+                $out .= ":".self::normalizePart(rawurlencode($p), self::P_PASS, false);
             }
             $out .= "@";
-        } elseif (strlen($username ?? "")) {
+        } elseif (strlen($user ?? "")) {
-            $out .= self::normalizePart($username, self::P_USER);
+            $out .= self::normalizePart($user, self::P_USER);
-            if (strlen($password ?? "")) {
+            if (strlen($pass ?? "")) {
-                $out .= ":".self::normalizePart($username, self::P_PASS);
+                $out .= ":".self::normalizePart($pass, self::P_PASS);
             }
             $out .= "@";
         }
@@ -31,11 +61,8 @@ class URL {
         } else {
             $out .= self::normalizeHost($host);
         }
-        if (isset($path)) {
+        $out .= isset($port) ? ":$port" : "";
-            $out .= self::normalizePath($path);
+        $out .= self::normalizePath($path ?? "");
-        } else {
-            $out .= "/";
-        }
         if (isset($query) && strlen($query)) {
             $out .= "?".self::normalizePart($query, self::P_QUERY);
         }
@@ -57,10 +84,22 @@ class URL {
         return $addr;
     }
 
+    /** Normalizes the whole path segment to remove empty segments and relative segments */
     protected static function normalizePath(string $path): string {
-        // stub
+        $parts = explode("/", $path);
-        return $path;
+        $out = [];
+        foreach($parts as $p) {
+            switch ($p) {
+                case "":
+                case ".":
+                    break;
+                case "..":
+                    array_pop($out);
+                    break;
+                default:
+                    $out[] = self::normalizePart($p, self::P_PATH);
+            }
+        }
+        return str_replace("//", "/", "/".implode("/", $out).(substr($path, -1) === "/" ? "/" : ""));
     }
-
-
 }

tests/cases/Misc/TestURL.php

@@ -15,17 +15,49 @@ class TestURL extends \JKingWeb\Arsse\Test\AbstractTest {
     }
     
     /** @dataProvider provideNormalizations */
-    public function testNormalizeAUrl(string $in, string $exp) {
+    public function testNormalizeAUrl(string $url, string $exp, string $user = null, string $pass = null) {
-        $this->assertSame($exp, URL::normalize($in));
+        $this->assertSame($exp, URL::normalize($url, $user, $pass));
     }
 
     public function provideNormalizations() {
         return [
             ["/",                             "/"],
             ["//example.com/",                "//example.com/"],
-            ["http://example.com/", "http://example.com/"],
             ["http://[::1]/",                 "http://[::1]/"],
+            ["http://example.com/",           "http://example.com/"],
             ["HTTP://example.com/",           "http://example.com/"],
+            ["http://example.com",            "http://example.com/"],
+            ["http://example.com:/",          "http://example.com/"],
+            ["HTTP://example.com:80/",        "http://example.com:80/"],
+            ["HTTP://example.com:80",         "http://example.com:80/"],
+            ["http://example.com/?",          "http://example.com/"],
+            ["http://example.com?",           "http://example.com/"],
+            ["http://example.com/#fragment",  "http://example.com/"],
+            ["http://example.com#fragment",   "http://example.com/"],
+            ["http://example.com?#",          "http://example.com/"],
+            ["http://example.com/?key=value", "http://example.com/?key=value"],
+            ["http://example.com/",           "http://user:pass@example.com/", "user", "pass"],
+            ["http://example.com/",           "http://user@example.com/", "user"],
+            ["http://user:pass@example.com/", "http://user:pass@example.com/"],
+            ["http://user@example.com/",      "http://user@example.com/"],
+            ["http://user:pass@example.com/", "http://u:p@example.com/", "u", "p"],
+            ["http://user:pass@example.com/", "http://u@example.com/", "u"],
+            ["http://user:pass@example.com/", "http://user:pass@example.com/", "", "p"],
+            ["http://example.com/",           "http://example.com/", "", "p"],
+            ["http://example.com/path",       "http://example.com/path"],
+            ["http://example.com/path/",      "http://example.com/path/"],
+            ["http://example.com/path/.",     "http://example.com/path"],
+            ["http://example.com/path/./",    "http://example.com/path/"],
+            ["http://example.com/path/..",    "http://example.com/"],
+            ["http://example.com/path/../",   "http://example.com/"],
+            ["http://example.com/a/b/..",     "http://example.com/a"],
+            ["http://example.com/a/b/../",    "http://example.com/a/"],
+            ["http://example.com/../",        "http://example.com/"],
+            ["http://example.com////",        "http://example.com/"],
+            ["http://example.com/a/./b/",     "http://example.com/a/b/"],
+            ["http://example.com/a/../b/",    "http://example.com/b/"],
+            ["http://example.com/.a/",        "http://example.com/.a/"],
+            ["http://example.com/..a/",       "http://example.com/..a/"],
         ];
     }
 }