Properly implement media type detection

lib/Parser/Construct.php

@@ -8,6 +8,7 @@ namespace JKingWeb\Lax\Parser;
 
 use JKingWeb\Lax\Collection;
 use JKingWeb\Lax\Date;
+use JKingWeb\Lax\Url;
 
 trait Construct {
     /** Trims plain text and collapses whitespace */
@@ -30,7 +31,7 @@ trait Construct {
 
     /** Tests whether a string is a valid e-mail address
      * 
-     * Accepts IDN hosts and (with PHP 7.1 and above) Unicode localparts
+     * Accepts IDN hosts and Unicode localparts
      */
     protected function validateMail(string $addr): bool {
         $out = preg_match("/^(.+?)@([^@]+)$/", $addr, $match);
@@ -39,13 +40,12 @@ trait Construct {
         }
         $local = $match[1];
         $domain = $match[2];
-        // PHP's filter_var does not accept IDN hosts, so we have to perform an IDNA transformat first
-        $domain = idn_to_ascii($domain, \IDNA_NONTRANSITIONAL_TO_ASCII, \INTL_IDNA_VARIANT_UTS46); // settings for IDNA2008 algorithm (I think)
-        if ($domain===false) {
-            return false;
+        // PHP's filter_var does not accept IDN hosts, so we have to perform an IDNA transformation first
+        $domain = idn_to_ascii($domain, \IDNA_NONTRANSITIONAL_TO_ASCII | \IDNA_CHECK_BIDI | \IDNA_CHECK_CONTEXTJ, \INTL_IDNA_VARIANT_UTS46); // settings for IDNA2008 algorithm (I think)
+        if ($domain !== false) {$addr = "$local@$domain";
+            return (bool) filter_var($addr, \FILTER_VALIDATE_EMAIL, \FILTER_FLAG_EMAIL_UNICODE);
         }
-        $addr = "$local@$domain";
-        return (bool) filter_var($addr, \FILTER_VALIDATE_EMAIL, \FILTER_FLAG_EMAIL_UNICODE);
+        return false;
     }
 
     protected function parseDate(string $date): ?Date {
@@ -64,6 +64,38 @@ trait Construct {
         return $out ?: null;
     }
 
+    protected function parseMediaType(string $type, ?Url $url = null): ?string {
+        if (preg_match('<^\s*([0-9a-z]+(?:/[!#$%&\'\*\+\-\.^_`|~0-9a-z]+))(?:\s|;|$)>i', $type, $match)) {
+            /* NOTE: The pattern used here is a subset of what is 
+                technically allowed by RFC 7231: the "type" portion
+                is supposed to be as general as the "subtype" portion,
+                but in practice only alphabetic types have ever been
+                registered, making a more specific pattern more
+                practically useful for detecting media types.
+
+                See:
+                <https://tools.ietf.org/html/rfc7231#section-3.1.1.1>
+                <https://tools.ietf.org/html/rfc7230#section-3.2.6>
+
+                Additionally, types without subtypes are accepted as
+                we foresee the general type still being useful to
+                feed processors.
+            */
+            return strtolower($match[1]);
+        }
+        if ($url) {
+            $file = substr($url->getPath(), (int) strrpos($url->getPath(), "/"));
+            $ext = strrpos($file, ".");
+            if ($ext !== false) {
+                $ext = substr($file, $ext + 1);
+                if (strlen($ext)) {
+                    return ($this->mime ?? ($this->mime = new \Mimey\MimeTypes))->getMimeType($ext);
+                }
+            }
+        }
+        return null;
+    }
+
     protected function empty($o): bool {
         return !array_filter((array) $o, function($v) {
             return !is_null($v) && (!$v instanceof Collection || sizeof($v) > 0);

lib/Parser/JSON/Construct.php

@@ -19,7 +19,7 @@ trait Construct {
      * 
      * Returns null otherwise
      */
-    protected function fetchMember(string $key, string $type, \stdClass $obj = null) {
+    protected function fetchMember(string $key, string $type, ?\stdClass $obj = null) {
         $obj = $obj ?? $this->data;
         if (!isset($obj->$key)) {
             return null;
@@ -34,7 +34,7 @@ trait Construct {
     }
 
     /** Returns an object member as a resolved and normalized URL */
-    protected function fetchUrl(string $key, \stdClass $obj = null): ?Url {
+    protected function fetchUrl(string $key, ?\stdClass $obj = null): ?Url {
         $url = $this->fetchMember($key, "str", $obj);
         try {
             return (!is_null($url)) ? new Url($url, $this->url) : null;
@@ -43,13 +43,19 @@ trait Construct {
         }
     }
 
+    /** Returns a media type from an object member or from a URL's file name when possible */
+    protected function fetchType(string $key, ?Url $url, ?\stdClass $obj = null): ?string {
+        $type = $this->fetchMember($key, "str", $obj) ?? "";
+        return $this->parseMediaType($type, $url);
+    }
+
     /** Returns an object member as a parsed date */
-    protected function fetchDate(string $key, \stdClass $obj = null): ?Date {
+    protected function fetchDate(string $key, ?\stdClass $obj = null): ?Date {
         return $this->parseDate($this->fetchMember($key, "str", $obj) ?? "");
     }
 
     /** Returns a plain-text string object member wrapped in a Text object */
-    protected function fetchText(string $key, \stdClass $obj = null): ?Text {
+    protected function fetchText(string $key, ?\stdClass $obj = null): ?Text {
         $t = $this->fetchMember($key, "str", $obj);
         if (!is_null($t)) {
             return new Text($t);

lib/Parser/JSON/Entry.php

@@ -147,15 +147,10 @@ class Entry implements \JKingWeb\Lax\Parser\Entry {
             if ($url) {            
                 $m = new Enclosure;
                 $m->url = $url;
-                $m->type = $this->fetchMember("mime_type", "str", $attachment);
+                $m->type = $this->fetchType("mime_type", $url, $attachment);
                 $m->title = $this->fetchMember("title", "str", $attachment);
                 $m->size = $this->fetchMember("size_in_bytes", "int", $attachment);
                 $m->duration = $this->fetchMember("duration_in_seconds", "int", $attachment);
-                // detect media type from file name if no type is provided
-                if (!$m->type) {
-                    $ext = substr();
-                    $m->type = ($this->mime ?? ($this->mime = new \Mimey\MimeTypes))->getMimeType($ext);
-                }
                 $out[] = $m;
             }
         }