From: Joey Hess <joeyh@joeyh.name>
Date: Tue, 22 Dec 2020 16:00:11 +0000 (-0400)
Subject: improve storage of redundant ContentIdentifiers
X-Git-Tag: archive/raspbian/10.20250416-2+rpi1~1^2~98^2~299^2~10
X-Git-Url: https://dgit.raspbian.org/?a=commitdiff_plain;h=06ef1b7d68ee2b149e663f2362e5685303d0bd3a;p=git-annex.git

improve storage of redundant ContentIdentifiers

When a ContentIdentifier is already recorded, don't add it to the log
again, and avoid updating the log.
---

diff --git a/Logs/ContentIdentifier.hs b/Logs/ContentIdentifier.hs
index 38f904ae2b..16c3969155 100644
--- a/Logs/ContentIdentifier.hs
+++ b/Logs/ContentIdentifier.hs
@@ -32,12 +32,16 @@ recordContentIdentifier :: RemoteStateHandle -> ContentIdentifier -> Key -> Anne
 recordContentIdentifier (RemoteStateHandle u) cid k = do
 	c <- liftIO currentVectorClock
 	config <- Annex.getGitConfig
-	Annex.Branch.change (remoteContentIdentifierLogFile config k) $
-		buildLog . addcid c . parseLog
+	Annex.Branch.maybeChange (remoteContentIdentifierLogFile config k) $
+		addcid c . parseLog
   where
-	addcid c l = changeMapLog c u (cid :| contentIdentifierList (M.lookup u m)) l
+	addcid c v
+		| cid `elem` l = Nothing -- no change needed
+		| otherwise = Just $ buildLog $
+			changeMapLog c u (cid :| l) v
 	  where
-		m = simpleMap l
+		m = simpleMap v
+		l = contentIdentifierList (M.lookup u m)
 
 -- | Get all known content identifiers for a key.
 getContentIdentifiers :: Key -> Annex [(RemoteStateHandle, [ContentIdentifier])]