fast mode

Add --fast flag, that can enable less expensive, but also less thurough versions of some commands.

* Add --fast flag, that can enable less expensive, but also less thurough
  versions of some commands.
* fsck: In fast mode, avoid checking checksums.
* unused: In fast mode, just show all existing temp files as unused,
  and avoid expensive scan for other unused content.

Annex.hs

@@ -40,6 +40,7 @@ data AnnexState = AnnexState
 	, repoqueue :: GitQueue.Queue
 	, quiet :: Bool
 	, force :: Bool
+	, fast :: Bool
 	, defaultbackend :: Maybe String
 	, defaultkey :: Maybe String
 	, toremote :: Maybe String
@@ -56,6 +57,7 @@ newState gitrepo allbackends = AnnexState
 	, repoqueue = GitQueue.empty
 	, quiet = False
 	, force = False
+	, fast = False
 	, defaultbackend = Nothing
 	, defaultkey = Nothing
 	, toremote = Nothing

Backend/SHA.hs

@@ -80,9 +80,10 @@ keyValue size file = do
 checkKeyChecksum :: SHASize -> Key -> Annex Bool
 checkKeyChecksum size key = do
 	g <- Annex.gitRepo
+	fast <- Annex.getState Annex.fast
 	let file = gitAnnexLocation g key
 	present <- liftIO $ doesFileExist file
-	if not present
+	if (not present || fast)
 		then return True
 		else do
 			s <- shaN size file

Command/Unused.hs

@@ -44,7 +44,6 @@ perform = do
 
 checkUnused :: Annex Bool
 checkUnused = do
-	showNote "checking for unused data..."
 	(unused, staletmp) <- unusedKeys
 	let unusedlist = number 0 unused
 	let staletmplist = number (length unused) staletmp
@@ -81,17 +80,27 @@ number n (x:xs) = (n+1, x):(number (n+1) xs)
 unusedKeys :: Annex ([Key], [Key])
 unusedKeys = do
 	g <- Annex.gitRepo
-	present <- getKeysPresent
-	referenced <- getKeysReferenced
-	tmps <- tmpKeys
 	
-	let (unused, staletmp, duptmp) = calcUnusedKeys present referenced tmps
+	fast <- Annex.getState Annex.fast
+	if fast
+		then do
+			showNote "fast mode enabled; assuming all temporary files are unused"
+			tmps <- tmpKeys
+			return ([], tmps)
+		else do
+			showNote "checking for unused data..."
+			present <- getKeysPresent
+			referenced <- getKeysReferenced
+			tmps <- tmpKeys
 	
-	-- Tmp files that are dups of content already present can simply
-	-- be removed.
-	liftIO $ forM_ duptmp $ \t -> removeFile $ gitAnnexTmpLocation g t
+			let (unused, staletmp, duptmp) = calcUnusedKeys present referenced tmps
 
-	return (unused, staletmp)
+			-- Tmp files that are dups of content already present
+			-- can simply be removed.
+			liftIO $ forM_ duptmp $ \t -> removeFile $
+				gitAnnexTmpLocation g t
+
+			return (unused, staletmp)
 
 calcUnusedKeys :: [Key] -> [Key] -> [Key] -> ([Key], [Key], [Key])
 calcUnusedKeys present referenced tmps = (unused, staletmp, duptmp)

Options.hs

@@ -22,6 +22,8 @@ commonOptions :: [Option]
 commonOptions =
 	[ Option ['f'] ["force"] (NoArg (setforce True))
 		"allow actions that may lose annexed data"
+	, Option ['F'] ["fast"] (NoArg (setfast True))
+		"avoid slow operations"
 	, Option ['q'] ["quiet"] (NoArg (setquiet True))
 		"avoid verbose output"
 	, Option ['v'] ["verbose"] (NoArg (setquiet False))
@@ -31,5 +33,6 @@ commonOptions =
 	]
 	where
 		setforce v = Annex.changeState $ \s -> s { Annex.force = v }
+		setfast v = Annex.changeState $ \s -> s { Annex.fast = v }
 		setquiet v = Annex.changeState $ \s -> s { Annex.quiet = v }
 		setdefaultbackend v = Annex.changeState $ \s -> s { Annex.defaultbackend = Just v }

debian/changelog

@@ -5,6 +5,11 @@ git-annex (0.20110321) UNRELEASED; urgency=low
     with git-annex 0.24 or earlier.) The code is believed to work on
     Linux, FreeBSD, and OSX; check compile-time messages to see if it
     is not enabled for your OS.
+  * Add --fast flag, that can enable less expensive, but also less thurough
+    versions of some commands.
+  * fsck: In fast mode, avoid checking checksums.
+  * unused: In fast mode, just show all existing temp files as unused,
+    and avoid expensive scan for other unused content.
 
  -- Joey Hess <joeyh@debian.org>  Tue, 22 Mar 2011 16:52:00 -0400
 

doc/git-annex.mdwn

@@ -278,6 +278,12 @@ Many git-annex commands will stage changes for later `git commit` by you.
   Force unsafe actions, such as dropping a file's content when no other
   source of it can be verified to still exist. Use with care.
 
+* --fast
+
+  Enables less expensive, but also less thorough versions of some commands.
+  What is avoided depends on the command. A fast fsck avoids calculating
+  checksums; a fast unused only shows temp files and not other unused files.
+
 * --quiet
 
   Avoid the default verbose logging of what is done; only show errors