Copy with a reflink when exporting a tree to a directory special remote

Remote.Directory makes a temp file, then calls this, and since the temp
file exists, it prevented probing if CoW works.

Note that deleting the empty file does mean there's a small window for a
race. If another process is also exporting to the remote, that could let it
make the same temp file. However, the temp filename actually has the
processes's pid in it, which avoids that being a problem.

This may have been a reversion caused by commits around
63d508e885, but I haven't gone back and
tested to be sure. The directory special remote had supposedly supported
CoW for this going back to about half a year before that.

Sponsored-by: Graham Spencer on Patreon

Annex/CopyFile.hs

@@ -31,17 +31,15 @@ newCopyCoWTried = CopyCoWTried <$> newEmptyMVar
 
 {- Copies a file is copy-on-write is supported. Otherwise, returns False.
  -
- - The destination file must not exist yet, or it will fail to make a CoW copy,
- - and will return false.
+ - The destination file must not exist yet (or may exist but be empty), 
+ - or it will fail to make a CoW copy, and will return false.
  -}
 tryCopyCoW :: CopyCoWTried -> FilePath -> FilePath -> MeterUpdate -> IO Bool
 tryCopyCoW (CopyCoWTried copycowtried) src dest meterupdate =
 	-- If multiple threads reach this at the same time, they
 	-- will both try CoW, which is acceptable.
 	ifM (isEmptyMVar copycowtried)
-		-- If dest exists, don't try CoW, since it would
-		-- have to be deleted first.
-		( ifM (doesFileExist dest)
+		( ifM destfilealreadypopulated
 			( return False
 			, do
 				ok <- docopycow
@@ -62,6 +60,22 @@ tryCopyCoW (CopyCoWTried copycowtried) src dest meterupdate =
 	docopycow = watchFileSize dest meterupdate $
 		copyCoW CopyTimeStamps src dest
 	
+	dest' = toRawFilePath dest
+
+	-- Check if the dest file already exists, which would prevent
+	-- probing CoW. If the file exists but is empty, there's no benefit
+	-- to resuming from it when CoW does not work, so remove it.
+	destfilealreadypopulated = 
+		tryIO (R.getFileStatus dest') >>= \case
+			Left _ -> return False
+			Right st -> do
+				sz <- getFileSize' dest' st
+				if sz == 0
+					then tryIO (removeFile dest) >>= \case
+						Right () -> return False
+						Left _ -> return True
+					else return True
+
 data CopyMethod = CopiedCoW | Copied
 
 {- Copies from src to dest, updating a meter. Preserves mode and mtime.

CHANGELOG

@@ -9,6 +9,7 @@ git-annex (10.20230322) UNRELEASED; urgency=medium
     the view branch.
   * Windows: Support urls like "file:///c:/path"
   * addurl, importfeed: Fix failure when annex.securehashesonly is set.
+  * Copy with a reflink when exporting a tree to a directory special remote.
 
  -- Joey Hess <id@joeyh.name>  Thu, 23 Mar 2023 15:04:41 -0400
 

doc/bugs/use_reflink_to_export_to_directory_special_remote.mdwn

@@ -40,3 +40,5 @@ After running, a `btrfs filesystem du ...` tells me that the `repo` and `import`
 ---
 
 In general: Thank you so much for this wonderful piece of software, I'm using it for years now and manage virtually everything with it (audio, video, pictures, important documents, papers, …).
+
+> [[fixed|done]] --[[Joey]]