incremental verify for chunked remotes

Simply feed each chunk in turn to the incremental verifier.

When resuming an interrupted retrieve, it does not do incremental
verification. That would need to read the file, up to the resume point,
and feed it to the incremental verifier. That seems easy to get wrong.
Also it would mean extra work done before the transfer can start. Which
would complicate displaying progress, and would perhaps not appear to the
user as if it was resuming from where it left off. Instead, in that
situation, return UnVerified, and let the verification be done in a
separate pass.

Granted, Annex.CopyFile does manage all that, but it's not complicated
by dealing with chunks too.

Sponsored-by: Dartmouth College's DANDI project

CHANGELOG

@@ -10,8 +10,7 @@ git-annex (8.20210804) UNRELEASED; urgency=medium
     git-annex's own progress display.
   * Several special remotes verify content while it is being retrieved,
     avoiding a separate checksum pass. They are: S3, bup, ddar,
-    and gcrypt (with a local repository). This optimisation is not yet
+    and gcrypt (with a local repository).
-    available when chunks are used.
 
  -- Joey Hess <id@joeyh.name>  Tue, 03 Aug 2021 12:22:45 -0400
 

Remote/Helper/Chunked.hs

@@ -269,33 +269,28 @@ retrieveChunks retriever u vc chunkconfig encryptor basek dest basep enc encc
 		-- that are likely not there.
 		iv <- startVerifyKeyContentIncrementally vc basek
 		tryNonAsync (getunchunked iv) >>= \case
-			Right Nothing -> return UnVerified
+			Right r -> finalize r
-			Right (Just iv') ->
+			Left e -> go (Just e) 
-				ifM (liftIO $ finalizeIncremental iv')
+				=<< chunkKeysOnly u chunkconfig basek
-					( return Verified
+	| otherwise = go Nothing 
-					, return UnVerified
+		=<< chunkKeys u chunkconfig basek
-					)
-			Left e -> do
-				go (Just e) =<< chunkKeysOnly u chunkconfig basek
-				return UnVerified
-	| otherwise = do
-		go Nothing =<< chunkKeys u chunkconfig basek
-		return UnVerified
   where
 	go pe cks = do
 		let ls = map chunkKeyList cks
 		currsize <- liftIO $ catchMaybeIO $ getFileSize (toRawFilePath dest)
 		let ls' = maybe ls (setupResume ls) currsize
 		if any null ls'
-			then noop -- dest is already complete
+			then finalize Nothing -- dest is already complete
-			else firstavail pe currsize ls'
+			else finalize =<< firstavail pe currsize ls'
 
 	firstavail Nothing _ [] = giveup "unable to determine the chunks to use for this remote"
 	firstavail (Just e) _ [] = throwM e
 	firstavail pe currsize ([]:ls) = firstavail pe currsize ls
 	firstavail _ currsize ((k:ks):ls)
-		| k == basek = void (getunchunked Nothing)
+		| k == basek = do
-			`catchNonAsync` (\e -> firstavail (Just e) currsize ls)
+			iv <- startVerifyKeyContentIncrementally vc basek
+			getunchunked iv
+				`catchNonAsync` (\e -> firstavail (Just e) currsize ls)
 		| otherwise = do
 			let offset = resumeOffset currsize k
 			let p = maybe basep
@@ -303,36 +298,42 @@ retrieveChunks retriever u vc chunkconfig encryptor basek dest basep enc encc
 				offset
 			v <- tryNonAsync $
 				retriever (encryptor k) p $ \content ->
-					bracketIO (maybe opennew openresume offset) hClose $ \h -> do
+					bracket (maybe opennew openresume offset) (liftIO . hClose . fst) $ \(h, iv) -> do
-						void $ retrieved Nothing (Just h) p content
+						iv' <- retrieved iv (Just h) p content
 						let sz = toBytesProcessed $
 							fromMaybe 0 $ fromKey keyChunkSize k
-						getrest p h sz sz ks
+						getrest p h iv' sz sz ks
 			case v of
 				Left e
 					| null ls -> throwM e
 					| otherwise -> firstavail (Just e) currsize ls
 				Right r -> return r
 
-	getrest _ _ _ _ [] = noop
+	getrest _ _ iv _ _ [] = return iv
-	getrest p h sz bytesprocessed (k:ks) = do
+	getrest p h iv sz bytesprocessed (k:ks) = do
 		let p' = offsetMeterUpdate p bytesprocessed
 		liftIO $ p' zeroBytesProcessed
-		retriever (encryptor k) p' $ 
+		iv' <- retriever (encryptor k) p' $ 
-			void . retrieved Nothing (Just h) p'
+			retrieved iv (Just h) p'
-		getrest p h sz (addBytesProcessed bytesprocessed sz) ks
+		getrest p h iv' sz (addBytesProcessed bytesprocessed sz) ks
 
 	getunchunked iv = retriever (encryptor basek) basep $
 		retrieved iv Nothing basep
 
-	opennew = openBinaryFile dest WriteMode
+	opennew = do
+		iv <- startVerifyKeyContentIncrementally vc basek
+		h <- liftIO $ openBinaryFile dest WriteMode
+		return (h, iv)
 
 	-- Open the file and seek to the start point in order to resume.
 	openresume startpoint = do
 		-- ReadWriteMode allows seeking; AppendMode does not.
-		h <- openBinaryFile dest ReadWriteMode
+		h <- liftIO $ openBinaryFile dest ReadWriteMode
-		hSeek h AbsoluteSeek startpoint
+		liftIO $ hSeek h AbsoluteSeek startpoint
-		return h
+		-- No incremental verification when resuming, since that
+		-- would need to read up to the startpoint.
+		let iv = Nothing
+		return (h, iv)
 
 	{- Progress meter updating is a bit tricky: If the Retriever
 	 - populates a file, it is responsible for updating progress
@@ -349,6 +350,13 @@ retrieveChunks retriever u vc chunkconfig encryptor basek dest basep enc encc
 		p'
 			| isByteContent content = Just p
 			| otherwise = Nothing
+	
+	finalize Nothing = return UnVerified
+	finalize (Just iv) = 
+		ifM (liftIO $ finalizeIncremental iv)
+			( return Verified
+			, return UnVerified
+			)
 
 {- Writes retrieved file content to the provided Handle, decrypting it
  - first if necessary.

doc/todo/OPT__58_____34__bundle__34___get_+_check___40__of_checksum__41___in_a_single_operation/comment_13_a19823409648828b8b9f7a0f2fa7d935._comment

@@ -0,0 +1,10 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 13"""
+ date="2021-08-11T18:16:41Z"
+ content="""
+Some special remotes now support incremental update. So far, limited to
+ones that use the byteRetriever interface.
+Others, that use fileRetriever, including external special remotes, still
+need work.
+"""]]