avoid beware of the leopard situation

* Display a warning message when a remote uses a protocol, such as
  git://, that git-annex does not support. Silently skipping such a
  remote was confusing behavior.

  It sets annex-ignore, so the warning is only displayed once.

* Also display a warning message when a remote, without a known uuid,
  is located in a directory that does not currently exist, to avoid
  silently skipping such a remote.

  This is a bit more debatable, since git-annex get will say,
  try making repository available. And since it does not set annex-ignore,
  the warning will be displayed repeatedly. It's also an extreme edge case,
  I don't think I've ever seen it happen in real life.

CHANGELOG

@@ -1,3 +1,14 @@
+git-annex (8.20200502) UNRELEASED; urgency=medium
+
+  * Display a warning message when a remote uses a protocol, such as
+    git://, that git-annex does not support. Silently skipping such a
+    remote was confusing behavior.
+  * Also display a warning message when a remote, without a known uuid,
+    is located in a directory that does not currently exist, to avoid
+    silently skipping such a remote.
+
+ -- Joey Hess <id@joeyh.name>  Mon, 04 May 2020 12:46:11 -0400
+
 git-annex (8.20200501) upstream; urgency=medium
 
   * Improve git-annex's ability to find the path to its program,

Remote/Git.hs

@@ -149,15 +149,15 @@ gitSetup (Enable _) Nothing _ _ _ = error "unable to enable git remote with no s
 configRead :: Bool -> Git.Repo -> Annex Git.Repo
 configRead autoinit r = do
 	gc <- Annex.getRemoteGitConfig r
-	u <- getRepoUUID r
+	hasuuid <- (/= NoUUID) <$> getRepoUUID r
 	annexignore <- liftIO $ getDynamicConfig (remoteAnnexIgnore gc)
-	case (repoCheap r, annexignore, u) of
+	case (repoCheap r, annexignore, hasuuid) of
 		(_, True, _) -> return r
 		(True, _, _)
-			| remoteAnnexCheckUUID gc -> tryGitConfigRead autoinit r
+			| remoteAnnexCheckUUID gc -> tryGitConfigRead autoinit r hasuuid
 			| otherwise -> return r
-		(False, _, NoUUID) -> configSpecialGitRemotes r >>= \case
+		(False, _, False) -> configSpecialGitRemotes r >>= \case
-			Nothing -> tryGitConfigRead autoinit r
+			Nothing -> tryGitConfigRead autoinit r False
 			Just r' -> return r'
 		_ -> return r
 
@@ -244,8 +244,8 @@ repoAvail r
 
 {- Tries to read the config for a specified remote, updates state, and
  - returns the updated repo. -}
-tryGitConfigRead :: Bool -> Git.Repo -> Annex Git.Repo
+tryGitConfigRead :: Bool -> Git.Repo -> Bool -> Annex Git.Repo
-tryGitConfigRead autoinit r 
+tryGitConfigRead autoinit r hasuuid
 	| haveconfig r = return r -- already read
 	| Git.repoIsSsh r = storeUpdatedRemote $ do
 		v <- Ssh.onRemote NoConsumeStdin r
@@ -258,9 +258,12 @@ tryGitConfigRead autoinit r
 			Left _ -> configlist_failed
 	| Git.repoIsHttp r = storeUpdatedRemote geturlconfig
 	| Git.GCrypt.isEncrypted r = handlegcrypt =<< getConfigMaybe (remoteAnnexConfig r "uuid")
-	| Git.repoIsUrl r = return r
+	| Git.repoIsUrl r = do
-	| otherwise = storeUpdatedRemote $ liftIO $ 
+		set_ignore "uses a protocol not supported by git-annex" False
-		readlocalannexconfig `catchNonAsync` (const $ return r)
+		return r
+	| otherwise = storeUpdatedRemote $
+		liftIO readlocalannexconfig
+			`catchNonAsync` const failedreadlocalconfig
   where
 	haveconfig = not . M.null . Git.config
 
@@ -339,6 +342,13 @@ tryGitConfigRead autoinit r
 		s <- Annex.new r
 		Annex.eval s $ check `finally` stopCoProcesses
 		
+	failedreadlocalconfig = do
+		unless hasuuid $ case Git.remoteName r of
+			Nothing -> noop
+			Just n -> do
+				warning $ "Remote " ++ n ++ " cannot currently be accessed."
+		return r
+		
 	configlistfields = if autoinit
 		then [(Fields.autoInit, "1")]
 		else []
@@ -898,7 +908,7 @@ mkState r u gc = do
 			rv <- liftIO newEmptyMVar
 			let getrepo = ifM (liftIO $ isEmptyMVar rv)
 				( do
-					r' <- tryGitConfigRead False r
+					r' <- tryGitConfigRead False r True
 					let t = (r', extractGitConfig FromGitConfig r')
 					void $ liftIO $ tryPutMVar rv t
 					return t

doc/bugs/annex_get_fails_with___39__not_available__39___for_git__58____47____47___protocol.mdwn

@@ -175,3 +175,6 @@ Already posted above
 ### 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)
 
 Well... it did work for git's local protocol...
+
+> [[done]], this was not really a bug, but I have added a warning message
+> --[[Joey]]

doc/bugs/annex_get_fails_with___39__not_available__39___for_git__58____47____47___protocol/comment_2_874c2c2dc99362b559c8fe1715d3c0c4._comment

@@ -0,0 +1,26 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 2"""
+ date="2020-05-04T16:13:02Z"
+ content="""
+git-annex cannot possibly support operation over the git protocol, because
+that protocol does not have any way to add an extension such as git-annex
+to it.
+
+So git-annex currently completely ignores git:// repositories, the same as
+it would any git repository using some other foreign protocol.
+(eg rsync or a protocol added by a git extension)
+
+This is actually identical behavior to if the remote had "url =
+/media/whatever" and the drive was unmounted after the repo was cloned.
+git-annex is not able to discover a uuid for the remote, and so it does
+not know the remote contains the content, and so it does not say anything
+about not being able to get the content from the remote, because it
+never tried. IOW, this is a "beware of the leopard" situation.
+
+So, it would better for this to be handled the same way that a ssh remote
+that lacks git-annex-shell is handled, by explicitly setting annex-ignore
+on first use and displaying a warning. (Although it should not set
+annex-ignore for the /media/whatever case, because the drive may later get
+mounted again.)
+"""]]