get --incomplete: New option to resume any interrupted downloads.

8 files changed, 73 insertions, 52 deletions

diff --git a/Annex/Content.hs b/Annex/Content.hs
index 28ca99544..5e7dd322b 100644
--- a/Annex/Content.hs
+++ b/Annex/Content.hs
@@ -35,9 +35,11 @@ module Annex.Content (
 	thawContent,
 	dirKeys,
 	withObjectLoc,
+	staleKeysPrune,
 ) where
 
 import System.IO.Unsafe (unsafeInterleaveIO)
+import qualified Data.Set as S
 
 import Common.Annex
 import Logs.Location
@@ -663,3 +665,37 @@ dirKeys dirspec = do
 		, return []
 		)
 
+{- Looks in the specified directory for bad/tmp keys, and returns a list
+ - of those that might still have value, or might be stale and removable.
+ - 
+ - Also, stale keys that can be proven to have no value
+ - (ie, their content is already present) are deleted.
+ -}
+staleKeysPrune :: (Git.Repo -> FilePath) -> Bool -> Annex [Key]
+staleKeysPrune dirspec nottransferred = do
+	contents <- dirKeys dirspec
+	
+	dups <- filterM inAnnex contents
+	let stale = contents `exclude` dups
+
+	dir <- fromRepo dirspec
+	liftIO $ forM_ dups $ \t -> removeFile $ dir </> keyFile t
+
+	if nottransferred
+		then do
+			inprogress <- S.fromList . map (transferKey . fst)
+				<$> getTransfers
+			return $ filter (`S.notMember` inprogress) stale
+		else return stale
+
+{- Finds items in the first, smaller list, that are not
+ - present in the second, larger list.
+ - 
+ - Constructing a single set, of the list that tends to be
+ - smaller, appears more efficient in both memory and CPU
+ - than constructing and taking the S.difference of two sets. -}
+exclude :: Ord a => [a] -> [a] -> [a]
+exclude [] _ = [] -- optimisation
+exclude smaller larger = S.toList $ remove larger $ S.fromList smaller
+  where
+	remove a b = foldl (flip S.delete) b a
diff --git a/CmdLine/GitAnnex/Options.hs b/CmdLine/GitAnnex/Options.hs
index 38fa93090..64435c2e9 100644
--- a/CmdLine/GitAnnex/Options.hs
+++ b/CmdLine/GitAnnex/Options.hs
@@ -62,6 +62,9 @@ keyOptions =
 		"operate on specified key"
 	]
 
+incompleteOption :: Option
+incompleteOption = flagOption [] "incomplete" "resume previous downloads"
+
 -- Options to match properties of annexed files.
 annexedMatchingOptions :: [Option]
 annexedMatchingOptions = concat
diff --git a/CmdLine/Seek.hs b/CmdLine/Seek.hs
index 3166ab83d..664c12b62 100644
--- a/CmdLine/Seek.hs
+++ b/CmdLine/Seek.hs
@@ -27,6 +27,7 @@ import CmdLine.Action
 import Logs.Location
 import Logs.Unused
 import Annex.CatFile
+import Annex.Content
 
 withFilesInGit :: (FilePath -> CommandStart) -> CommandSeek
 withFilesInGit a params = seekActions $ prepFiltered a $
@@ -163,42 +164,42 @@ withNothing :: CommandStart -> CommandSeek
 withNothing a [] = seekActions $ return [a]
 withNothing _ _ = error "This command takes no parameters."
 
-{- If --all is specified, or in a bare repo, runs an action on all
- - known keys.
+{- Handles the --all, --unused, --key, and --incomplete options,
+ - which specify particular keys to run an action on.
  -
- - If --unused is specified, runs an action on all keys found by
- - the last git annex unused scan.
+ - In a bare repo, --all is the default.
  -
- - If --key is specified, operates only on that key.
- -
- - Otherwise, fall back to a regular CommandSeek action on
+ - Otherwise falls back to a regular CommandSeek action on
  - whatever params were passed. -}
 withKeyOptions :: Bool -> (Key -> CommandStart) -> CommandSeek -> CommandSeek
 withKeyOptions auto keyop fallbackop params = do
 	bare <- fromRepo Git.repoIsLocalBare
 	allkeys <- Annex.getFlag "all"
 	unused <- Annex.getFlag "unused"
+	incomplete <- Annex.getFlag "incomplete"
 	specifickey <- Annex.getField "key"
 	when (auto && bare) $
 		error "Cannot use --auto in a bare repository"
-	case	(allkeys, unused, null params, specifickey) of
-		(False  , False , True       , Nothing)
+	case	(allkeys, unused, incomplete, null params, specifickey) of
+		(False  , False , False     , True       , Nothing)
 			| bare -> go auto loggedKeys
 			| otherwise -> fallbackop params
-		(False  , False , _          , Nothing) -> fallbackop params
-		(True   , False , True       , Nothing) -> go auto loggedKeys
-		(False  , True  , True       , Nothing) -> go auto unusedKeys'
-		(False  , False , True       , Just ks) -> case file2key ks of
+		(False  , False , False     , _          , Nothing) -> fallbackop params
+		(True   , False , False     , True       , Nothing) -> go auto loggedKeys
+		(False  , True  , False     , True       , Nothing) -> go auto unusedKeys'
+		(False  , False , True      , True       , Nothing) -> go auto incompletekeys
+		(False  , False , False     , True       , Just ks) -> case file2key ks of
 			Nothing -> error "Invalid key"
 			Just k -> go auto $ return [k]
-		_ -> error "Can only specify one of file names, --all, --unused, or --key"
+		_ -> error "Can only specify one of file names, --all, --unused, --key, or --incomplete"
   where
-	go True _ = error "Cannot use --auto with --all or --unused or --key"
+	go True _ = error "Cannot use --auto with --all or --unused or --key or --incomplete"
 	go False a = do
 		matcher <- Limit.getMatcher
 		seekActions $ map (process matcher) <$> a
 	process matcher k = ifM (matcher $ MatchingKey k)
 		( keyop k , return Nothing)
+	incompletekeys = staleKeysPrune gitAnnexTmpObjectDir True
 
 prepFiltered :: (FilePath -> CommandStart) -> Annex [FilePath] -> Annex [CommandStart]
 prepFiltered a fs = do
diff --git a/Command/Get.hs b/Command/Get.hs
index dcd7e367a..d39b3890f 100644
--- a/Command/Get.hs
+++ b/Command/Get.hs
@@ -21,7 +21,8 @@ cmd = [withOptions getOptions $ command "get" paramPaths seek
 	SectionCommon "make content of annexed files available"]
 
 getOptions :: [Option]
-getOptions = fromOption : autoOption : jobsOption : annexedMatchingOptions ++ keyOptions
+getOptions = fromOption : autoOption : jobsOption : annexedMatchingOptions
+	++ incompleteOption : keyOptions
 
 seek :: CommandSeek
 seek ps = do
diff --git a/Command/Unused.hs b/Command/Unused.hs
index c92ece2d5..4f844081a 100644
--- a/Command/Unused.hs
+++ b/Command/Unused.hs
@@ -9,7 +9,6 @@
 
 module Command.Unused where
 
-import qualified Data.Set as S
 import Control.Monad.ST
 import qualified Data.Map as M
 
@@ -18,7 +17,6 @@ import Command
 import Logs.Unused
 import Annex.Content
 import Logs.Location
-import Logs.Transfer
 import qualified Annex
 import qualified Git
 import qualified Git.Command
@@ -174,18 +172,6 @@ excludeReferenced refspec ks = runfilter firstlevel ks >>= runfilter secondlevel
 	firstlevel = withKeysReferencedM
 	secondlevel = withKeysReferencedInGit refspec
 
-{- Finds items in the first, smaller list, that are not
- - present in the second, larger list.
- - 
- - Constructing a single set, of the list that tends to be
- - smaller, appears more efficient in both memory and CPU
- - than constructing and taking the S.difference of two sets. -}
-exclude :: Ord a => [a] -> [a] -> [a]
-exclude [] _ = [] -- optimisation
-exclude smaller larger = S.toList $ remove larger $ S.fromList smaller
-  where
-	remove a b = foldl (flip S.delete) b a
-
 {- A bloom filter capable of holding half a million keys with a
  - false positive rate of 1 in 1000 uses around 8 mb of memory,
  - so will easily fit on even my lowest memory systems.
@@ -313,28 +299,6 @@ withKeysReferencedInGitRef a ref = do
 	tKey False = fileKey . takeFileName . decodeBS <$$>
 		catFile ref . getTopFilePath . DiffTree.file
 
-{- Looks in the specified directory for bad/tmp keys, and returns a list
- - of those that might still have value, or might be stale and removable.
- - 
- - Also, stale keys that can be proven to have no value are deleted.
- -}
-staleKeysPrune :: (Git.Repo -> FilePath) -> Bool -> Annex [Key]
-staleKeysPrune dirspec nottransferred = do
-	contents <- dirKeys dirspec
-	
-	dups <- filterM inAnnex contents
-	let stale = contents `exclude` dups
-
-	dir <- fromRepo dirspec
-	liftIO $ forM_ dups $ \t -> removeFile $ dir </> keyFile t
-
-	if nottransferred
-		then do
-			inprogress <- S.fromList . map (transferKey . fst)
-				<$> getTransfers
-			return $ filter (`S.notMember` inprogress) stale
-		else return stale
-
 data UnusedMaps = UnusedMaps
 	{ unusedMap :: UnusedMap
 	, unusedBadMap :: UnusedMap
diff --git a/debian/changelog b/debian/changelog
index a5fd44aab..0cc6b7a2b 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -8,6 +8,7 @@ git-annex (5.20150529) UNRELEASED; urgency=medium
     changes to network connections, as was already done with
     network-manager and wicd.
     Thanks to Sebastian Reuße for the patches.
+  * get --incomplete: New option to resume any interrupted downloads.
 
  -- Joey Hess <id@joeyh.name>  Sat, 30 May 2015 02:07:18 -0400
 
diff --git a/doc/git-annex-get.mdwn b/doc/git-annex-get.mdwn
index 4f5917534..a79f1342a 100644
--- a/doc/git-annex-get.mdwn
+++ b/doc/git-annex-get.mdwn
@@ -32,6 +32,19 @@ or transferring them from some kind of key-value store.
   Enables parallel download with up to the specified number of jobs
   running at once. For example: `-J10`
 
+* `--incomplete`
+
+  Resume any incomplete downloads of files that were started and
+  interrupted at some point previously. Useful to pick up where you left
+  off ... when you don't quite remember where that was.
+
+  These incomplete files are the same ones that are
+  listed as unused temp files by [[git-annex-unused]](1).
+
+  Note that the git-annex key will be displayed when downloading,
+  as git-annex does not know the associated file, and the associated file
+  may not even be in the current git working directory.
+
 * `--all`
 
   Rather than specifying a filename or path to get, this option can be
diff --git a/doc/todo/get_--incomplete.mdwn b/doc/todo/get_--incomplete.mdwn
index 5b013cb36..783108beb 100644
--- a/doc/todo/get_--incomplete.mdwn
+++ b/doc/todo/get_--incomplete.mdwn
@@ -4,3 +4,5 @@ download.
 
 `git annex get --incomplete` could do this. (With or without --from to
 specify which remote to get from.) --[[Joey]]
+
+> [[done]] --[[Joey]]