filebackend: Added supported for retrieving file metadata/headers

* This can be useful for carrying over metadata when copying files around
* Also fixed a bug in sanitizeHdrs() for Swift (broken content-disposition)

Change-Id: I4534e9acac2b306086797b3677f85c05b98e39fc

4 files changed, 138 insertions, 3 deletions

diff --git a/includes/filebackend/FileBackend.php b/includes/filebackend/FileBackend.php
index bb21f1b5e492..f5d63b9df78e 100644
--- a/includes/filebackend/FileBackend.php
+++ b/includes/filebackend/FileBackend.php
@@ -104,6 +104,10 @@ abstract class FileBackend {
 	/** @var FileJournal */
 	protected $fileJournal;
 
+	/** Flags for supported features */
+	const ATTR_HEADERS  = 1;
+	const ATTR_METADATA = 2;
+
 	/**
 	 * Create a new backend instance from configuration.
 	 * This should only be called from within FileBackendGroup.
@@ -201,6 +205,27 @@ abstract class FileBackend {
 	}
 
 	/**
+	 * Get the a bitfield of extra features supported by the backend medium
+	 *
+	 * @return integer Bitfield of FileBackend::ATTR_* flags
+	 * @since 1.23
+	 */
+	public function getFeatures() {
+		return 0;
+	}
+
+	/**
+	 * Check if the backend medium supports a field of extra features
+	 *
+	 * @return integer Bitfield of FileBackend::ATTR_* flags
+	 * @return bool
+	 * @since 1.23
+	 */
+	final public function hasFeatures( $bitfield ) {
+		return ( $this->getFeatures() & $bitfield ) === $bitfield;
+	}
+
+	/**
 	 * This is the main entry point into the backend for write operations.
 	 * Callers supply an ordered list of operations to perform as a transaction.
 	 * Files will be locked, the stat cache cleared, and then the operations attempted.
@@ -902,6 +927,26 @@ abstract class FileBackend {
 	abstract public function getFileContentsMulti( array $params );
 
 	/**
+	 * Get metadata about a file at a storage path in the backend.
+	 * If the file does not exist, then this returns false.
+	 * Otherwise, the result is an associative array that includes:
+	 *   - headers  : map of HTTP headers used for GET/HEAD requests (name => value)
+	 *   - metadata : map of file metadata (name => value)
+	 * Metadata keys and headers names will be returned in all lower-case.
+	 * Additional values may be included for internal use only.
+	 *
+	 * Use FileBackend::hasFeatures() to check how well this is supported.
+	 *
+	 * @param array $params
+	 * $params include:
+	 *   - src    : source storage path
+	 *   - latest : use the latest available data
+	 * @return Array|bool Returns false on failure
+	 * @since 1.23
+	 */
+	abstract public function getFileXAttributes( array $params );
+
+	/**
 	 * Get the size (bytes) of a file at a storage path in the backend.
 	 *
 	 * @param array $params Parameters include:
diff --git a/includes/filebackend/FileBackendMultiWrite.php b/includes/filebackend/FileBackendMultiWrite.php
index 1c9832de74ba..1b2860a201da 100644
--- a/includes/filebackend/FileBackendMultiWrite.php
+++ b/includes/filebackend/FileBackendMultiWrite.php
@@ -567,6 +567,11 @@ class FileBackendMultiWrite extends FileBackend {
 		return $this->backends[$this->masterIndex]->getFileStat( $realParams );
 	}
 
+	public function getFileXAttributes( array $params ) {
+		$realParams = $this->substOpPaths( $params, $this->backends[$this->masterIndex] );
+		return $this->backends[$this->masterIndex]->getFileXAttributes( $realParams );
+	}
+
 	public function getFileContentsMulti( array $params ) {
 		$realParams = $this->substOpPaths( $params, $this->backends[$this->masterIndex] );
 		$contentsM = $this->backends[$this->masterIndex]->getFileContentsMulti( $realParams );
@@ -645,6 +650,10 @@ class FileBackendMultiWrite extends FileBackend {
 		return $this->backends[$this->masterIndex]->getFileList( $realParams );
 	}
 
+	public function getFeatures() {
+		return $this->backends[$this->masterIndex]->getFeatures();
+	}
+
 	public function clearCache( array $paths = null ) {
 		foreach ( $this->backends as $backend ) {
 			$realPaths = is_array( $paths ) ? $this->substPaths( $paths, $backend ) : null;
diff --git a/includes/filebackend/FileBackendStore.php b/includes/filebackend/FileBackendStore.php
index fe3a068456b1..7980d100675e 100644
--- a/includes/filebackend/FileBackendStore.php
+++ b/includes/filebackend/FileBackendStore.php
@@ -655,10 +655,15 @@ abstract class FileBackendStore extends FileBackend {
 				$this->cheapCache->set( $path, 'sha1',
 					array( 'hash' => $stat['sha1'], 'latest' => $latest ) );
 			}
+			if ( isset( $stat['xattr'] ) ) { // some backends store headers/metadata
+				$stat['xattr'] = self::normalizeXAttributes( $stat['xattr'] );
+				$this->cheapCache->set( $path, 'xattr',
+					array( 'map' => $stat['xattr'], 'latest' => $latest ) );
+			}
 		} elseif ( $stat === false ) { // file does not exist
 			$this->cheapCache->set( $path, 'stat', $latest ? 'NOT_EXIST_LATEST' : 'NOT_EXIST' );
-			$this->cheapCache->set( $path, 'sha1', // the SHA-1 must be false too
-				array( 'hash' => false, 'latest' => $latest ) );
+			$this->cheapCache->set( $path, 'xattr', array( 'map' => false, 'latest' => $latest ) );
+			$this->cheapCache->set( $path, 'sha1', array( 'hash' => false, 'latest' => $latest ) );
 			wfDebug( __METHOD__ . ": File $path does not exist.\n" );
 		} else { // an error occurred
 			wfDebug( __METHOD__ . ": Could not stat file $path.\n" );
@@ -697,6 +702,39 @@ abstract class FileBackendStore extends FileBackend {
 		return $contents;
 	}
 
+	final public function getFileXAttributes( array $params ) {
+		$path = self::normalizeStoragePath( $params['src'] );
+		if ( $path === null ) {
+			return false; // invalid storage path
+		}
+		$section = new ProfileSection( __METHOD__ . "-{$this->name}" );
+		$latest = !empty( $params['latest'] ); // use latest data?
+		if ( $this->cheapCache->has( $path, 'xattr', self::CACHE_TTL ) ) {
+			$stat = $this->cheapCache->get( $path, 'xattr' );
+			// If we want the latest data, check that this cached
+			// value was in fact fetched with the latest available data.
+			if ( !$latest || $stat['latest'] ) {
+				return $stat['map'];
+			}
+		}
+		wfProfileIn( __METHOD__ . '-miss' );
+		wfProfileIn( __METHOD__ . '-miss-' . $this->name );
+		$fields = $this->doGetFileXAttributes( $params );
+		$fields = is_array( $fields ) ? self::normalizeXAttributes( $fields ) : false;
+		wfProfileOut( __METHOD__ . '-miss-' . $this->name );
+		wfProfileOut( __METHOD__ . '-miss' );
+		$this->cheapCache->set( $path, 'xattr', array( 'map' => $fields, 'latest' => $latest ) );
+		return $fields;
+	}
+
+	/**
+	 * @see FileBackendStore::getFileXAttributes()
+	 * @return bool|string
+	 */
+	protected function doGetFileXAttributes( array $params ) {
+		return array( 'headers' => array(), 'metadata' => array() ); // not supported
+	}
+
 	final public function getFileSha1Base36( array $params ) {
 		$path = self::normalizeStoragePath( $params['src'] );
 		if ( $path === null ) {
@@ -1625,11 +1663,34 @@ abstract class FileBackendStore extends FileBackend {
 					$this->cheapCache->set( $path, 'sha1',
 						array( 'hash' => $val['sha1'], 'latest' => $val['latest'] ) );
 				}
+				if ( isset( $val['xattr'] ) ) { // some backends store headers/metadata
+					$stat['xattr'] = self::normalizeXAttributes( $stat['xattr'] );
+					$this->cheapCache->set( $path, 'xattr',
+						array( 'map' => $val['xattr'], 'latest' => $val['latest'] ) );
+				}
 			}
 		}
 	}
 
 	/**
+	 * Normalize file headers/metadata to the FileBackend::getFileXAttributes() format
+	 *
+	 * @param array $xattr
+	 * @return array
+	 * @since 1.22
+	 */
+	final protected static function normalizeXAttributes( array $xattr ) {
+		$newXAttr = array( 'headers' => array(), 'metadata' => array() );
+		foreach ( $xattr['headers'] as $name => $value ) {
+			$newXAttr['headers'][strtolower( $name )] = $value;
+		}
+		foreach ( $xattr['metadata'] as $name => $value ) {
+			$newXAttr['metadata'][strtolower( $name )] = $value;
+		}
+		return $newXAttr;
+	}
+
+	/**
 	 * Set the 'concurrency' option from a list of operation options
 	 *
 	 * @param array $opts Map of operation options
diff --git a/includes/filebackend/SwiftFileBackend.php b/includes/filebackend/SwiftFileBackend.php
index 528889b69bf6..037b1c3c3043 100644
--- a/includes/filebackend/SwiftFileBackend.php
+++ b/includes/filebackend/SwiftFileBackend.php
@@ -142,6 +142,10 @@ class SwiftFileBackend extends FileBackendStore {
 		$this->srvCache = $this->srvCache ?: new EmptyBagOStuff();
 	}
 
+	public function getFeatures() {
+		return ( FileBackend::ATTR_HEADERS | FileBackend::ATTR_METADATA );
+	}
+
 	protected function resolveContainerPath( $container, $relStoragePath ) {
 		if ( !mb_check_encoding( $relStoragePath, 'UTF-8' ) ) { // mb_string required by CF
 			return null; // not UTF-8, makes it hard to use CF and the swift HTTP API
@@ -179,6 +183,8 @@ class SwiftFileBackend extends FileBackendStore {
 					continue; // blacklisted
 				} elseif ( preg_match( '/^(x-)?content-/', $name ) ) {
 					$headers[$name] = $value; // allowed
+				} elseif ( preg_match( '/^content-(disposition)/', $name ) ) {
+					$headers[$name] = $value; // allowed
 				}
 			}
 		}
@@ -189,7 +195,7 @@ class SwiftFileBackend extends FileBackendStore {
 				$part = trim( $part );
 				$new = ( $disposition === '' ) ? $part : "{$disposition};{$part}";
 				if ( strlen( $new ) <= 255 ) {
-					$res = $new;
+					$disposition = $new;
 				} else {
 					break; // too long; sigh
 				}
@@ -986,6 +992,20 @@ class SwiftFileBackend extends FileBackendStore {
 		$this->cheapCache->set( $path, 'stat', $val );
 	}
 
+	protected function doGetFileXAttributes( array $params ) {
+		$stat = $this->getFileStat( $params );
+		if ( $stat ) {
+			if ( !isset( $stat['xattr'] ) ) {
+				// Stat entries filled by file listings don't include metadata/headers
+				$this->clearCache( array( $params['src'] ) );
+				$stat = $this->getFileStat( $params );
+			}
+			return $stat['xattr'];
+		} else {
+			return false;
+		}
+	}
+
 	protected function doGetFileSha1base36( array $params ) {
 		$stat = $this->getFileStat( $params );
 		if ( $stat ) {