summaryrefslogtreecommitdiff
path: root/includes/LinkFilter.php
diff options
context:
space:
mode:
Diffstat (limited to 'includes/LinkFilter.php')
-rw-r--r--includes/LinkFilter.php115
1 files changed, 69 insertions, 46 deletions
diff --git a/includes/LinkFilter.php b/includes/LinkFilter.php
index d552c696..340ae8f3 100644
--- a/includes/LinkFilter.php
+++ b/includes/LinkFilter.php
@@ -35,9 +35,9 @@ class LinkFilter {
/**
* Check whether $content contains a link to $filterEntry
*
- * @param $content Content: content to check
- * @param string $filterEntry domainparts, see makeRegex() for more details
- * @return Integer: 0 if no match or 1 if there's at least one match
+ * @param Content $content Content to check
+ * @param string $filterEntry Domainparts, see makeRegex() for more details
+ * @return int 0 if no match or 1 if there's at least one match
*/
static function matchEntry( Content $content, $filterEntry ) {
if ( !( $content instanceof TextContent ) ) {
@@ -58,7 +58,7 @@ class LinkFilter {
*
* @param string $filterEntry URL, if it begins with "*.", it'll be
* replaced to match any subdomain
- * @return String: regex pattern, for preg_match()
+ * @return string Regex pattern, for preg_match()
*/
private static function makeRegex( $filterEntry ) {
$regex = '!http://';
@@ -84,16 +84,28 @@ class LinkFilter {
*
* Asterisks in any other location are considered invalid.
*
- * @param string $filterEntry domainparts
- * @param $prot String: protocol
- * @return Array to be passed to DatabaseBase::buildLike() or false on error
+ * This function does the same as wfMakeUrlIndexes(), except it also takes care
+ * of adding wildcards
+ *
+ * @param string $filterEntry Domainparts
+ * @param string $protocol Protocol (default http://)
+ * @return array Array to be passed to DatabaseBase::buildLike() or false on error
*/
- public static function makeLikeArray( $filterEntry, $prot = 'http://' ) {
+ public static function makeLikeArray( $filterEntry, $protocol = 'http://' ) {
$db = wfGetDB( DB_MASTER );
- if ( substr( $filterEntry, 0, 2 ) == '*.' ) {
+
+ $target = $protocol . $filterEntry;
+ $bits = wfParseUrl( $target );
+
+ if ( $bits == false ) {
+ // Unknown protocol?
+ return false;
+ }
+
+ if ( substr( $bits['host'], 0, 2 ) == '*.' ) {
$subdomains = true;
- $filterEntry = substr( $filterEntry, 2 );
- if ( $filterEntry == '' ) {
+ $bits['host'] = substr( $bits['host'], 2 );
+ if ( $bits['host'] == '' ) {
// We don't want to make a clause that will match everything,
// that could be dangerous
return false;
@@ -101,55 +113,66 @@ class LinkFilter {
} else {
$subdomains = false;
}
- // No stray asterisks, that could cause confusion
- // It's not simple or efficient to handle it properly so we don't
- // handle it at all.
- if ( strpos( $filterEntry, '*' ) !== false ) {
- return false;
- }
- $slash = strpos( $filterEntry, '/' );
- if ( $slash !== false ) {
- $path = substr( $filterEntry, $slash );
- $host = substr( $filterEntry, 0, $slash );
- } else {
- $path = '/';
- $host = $filterEntry;
- }
+
// Reverse the labels in the hostname, convert to lower case
// For emails reverse domainpart only
- if ( $prot == 'mailto:' && strpos( $host, '@' ) ) {
+ if ( $bits['scheme'] === 'mailto' && strpos( $bits['host'], '@' ) ) {
// complete email address
- $mailparts = explode( '@', $host );
+ $mailparts = explode( '@', $bits['host'] );
$domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
- $host = $domainpart . '@' . $mailparts[0];
- $like = array( "$prot$host", $db->anyString() );
- } elseif ( $prot == 'mailto:' ) {
- // domainpart of email address only. do not add '.'
- $host = strtolower( implode( '.', array_reverse( explode( '.', $host ) ) ) );
- $like = array( "$prot$host", $db->anyString() );
+ $bits['host'] = $domainpart . '@' . $mailparts[0];
+ } elseif ( $bits['scheme'] === 'mailto' ) {
+ // domainpart of email address only, do not add '.'
+ $bits['host'] = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
} else {
- $host = strtolower( implode( '.', array_reverse( explode( '.', $host ) ) ) );
- if ( substr( $host, -1, 1 ) !== '.' ) {
- $host .= '.';
+ $bits['host'] = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
+ if ( substr( $bits['host'], -1, 1 ) !== '.' ) {
+ $bits['host'] .= '.';
}
- $like = array( "$prot$host" );
+ }
- if ( $subdomains ) {
- $like[] = $db->anyString();
- }
- if ( !$subdomains || $path !== '/' ) {
- $like[] = $path;
- $like[] = $db->anyString();
+ $like[] = $bits['scheme'] . $bits['delimiter'] . $bits['host'];
+
+ if ( $subdomains ) {
+ $like[] = $db->anyString();
+ }
+
+ if ( isset( $bits['port'] ) ) {
+ $like[] = ':' . $bits['port'];
+ }
+ if ( isset( $bits['path'] ) ) {
+ $like[] = $bits['path'];
+ } elseif ( !$subdomains ) {
+ $like[] = '/';
+ }
+ if ( isset( $bits['query'] ) ) {
+ $like[] = '?' . $bits['query'];
+ }
+ if ( isset( $bits['fragment'] ) ) {
+ $like[] = '#' . $bits['fragment'];
+ }
+
+ // Check for stray asterisks: asterisk only allowed at the start of the domain
+ foreach ( $like as $likepart ) {
+ if ( !( $likepart instanceof LikeMatch ) && strpos( $likepart, '*' ) !== false ) {
+ return false;
}
}
+
+ if ( !( $like[count( $like ) - 1] instanceof LikeMatch ) ) {
+ // Add wildcard at the end if there isn't one already
+ $like[] = $db->anyString();
+ }
+
return $like;
}
/**
- * Filters an array returned by makeLikeArray(), removing everything past first pattern placeholder.
+ * Filters an array returned by makeLikeArray(), removing everything past first
+ * pattern placeholder.
*
- * @param array $arr array to filter
- * @return array filtered array
+ * @param array $arr Array to filter
+ * @return array Filtered array
*/
public static function keepOneWildcard( $arr ) {
if ( !is_array( $arr ) ) {