When retrival from a chunked remote fails, display the error that occurred when downloading the chunk

Rather than the error that occurred when trying to download the unchunked
content, which is less likely to actually be stored in the remote.

Sponsored-by: Boyd Stephen Smith Jr. on Patreon

CHANGELOG

@@ -10,6 +10,10 @@ git-annex (8.20211012) UNRELEASED; urgency=medium
   * test: Put gpg temp home directory in system temp directory,
     not filesystem being tested.
   * Avoid crashing tilde expansion on user who does not exist.
+  * When retrival from a chunked remote fails, display the error that
+    occurred when downloading the chunk, rather than the error that
+    occurred when trying to download the unchunked content, which is less
+    likely to actually be stored in the remote.
 
  -- Joey Hess <id@joeyh.name>  Mon, 11 Oct 2021 14:09:13 -0400
 

Remote/Helper/Chunked.hs

@@ -286,9 +286,9 @@ retrieveChunks retriever u vc chunkconfig encryptor basek dest basep enc encc
 	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)
+	firstavail pe currsize ((k:ks):ls)
 		| k == basek = getunchunked
-			`catchNonAsync` (\e -> firstavail (Just e) currsize ls)
+			`catchNonAsync` (\e -> firstavail (Just (pickerr e)) currsize ls)
 		| otherwise = do
 			let offset = resumeOffset currsize k
 			let p = maybe basep
@@ -302,10 +302,15 @@ retrieveChunks retriever u vc chunkconfig encryptor basek dest basep enc encc
 							fromMaybe 0 $ fromKey keyChunkSize k
 						getrest p h iv sz sz ks
 			case v of
-				Left e
-					| null ls -> throwM e
-					| otherwise -> firstavail (Just e) currsize ls
+				Left e -> firstavail (Just (pickerr e)) currsize ls
 				Right r -> return r
+	  where
+		-- Prefer an earlier exception to a later one, because the
+		-- more probable location is tried first and less probable
+		-- ones later.
+		pickerr e = case pe of
+			Just pe' -> pe'
+			Nothing -> e
 
 	getrest _ _ iv _ _ [] = return (Right iv)
 	getrest p h iv sz bytesprocessed (k:ks) = do

doc/bugs/Improvements_to_S3_glacier_integration.mdwn

@@ -115,3 +115,5 @@ ongoing-request="false", expiry-date="Mon, 26 Apr 2021 00:00:00 GMT"
 ### Have you had any luck using git-annex before? (Sometimes we get tired of reading bug reports all day and a lil' positive end note does wonders)
 
 Yes, for loads of stuff. It's awesome, thanks!
+
+> [[closed|done]], see my comment --[[Joey]]

doc/bugs/Improvements_to_S3_glacier_integration/comment_5_9cbe83bbade15b9146d033ceb5d8b05d._comment

@@ -0,0 +1,19 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 5"""
+ date="2021-10-14T16:17:14Z"
+ content="""
+I have added a note to the S3 documentation about `DEEP_ARCHIVE` and the
+glacier special remote.
+
+I have made git-annex display the exception for the more likely chunked 
+location, rather than the less likely unchunked location, when retrieving
+from both locations fails. Although it's still possible for there to be
+situations where the exception if displays is not for the location where
+the content actually is. Eg, if the chunk size of the remote has
+changed over time.
+
+I think that todo is basically talking about the same desire to make the S3
+remote support these glacier-style storage classes, in one way or another, 
+and so I think this bug report can be closed as otherwise a duplicate of it.
+"""]]

doc/special_remotes/S3.mdwn

@@ -44,14 +44,17 @@ the S3 remote.
   
   When using Amazon S3,
   if the remote will be used for backup or archival,
-  and so its files are Infrequently Accessed, "STANDARD_IA" is a
+  and so its files are Infrequently Accessed, `STANDARD_IA` is a
   good choice to save money (requires a git-annex built with aws-0.13.0).
   If you have configured git-annex to preserve
-  multiple [[copies]], also consider setting this to "ONEZONE_IA"
+  multiple [[copies]], also consider setting this to `ONEZONE_IA`
   to save even more money.
 
+  Amazon S3's `DEEP_ARCHIVE` is similar to Amazon Glacier. For that,
+  use the [[glacier]] special remote, rather than this one.
+
   When using Google Cloud Storage, to make a nearline bucket, set this to
-  "NEARLINE". (Requires a git-annex built with aws-0.13.0)
+  `NEARLINE`. (Requires a git-annex built with aws-0.13.0)
 
   Note that changing the storage class of an existing S3 remote will
   affect new objects sent to the remote, but not objects already