Improve url parsing to handle some urls containing illegal [] characters in their paths.

Ie, "https://archive.org/download/zoom-2/Zoom - Release 2 (1996)(Active Software)[!].iso"

Command/AddUrl.hs

@@ -139,7 +139,7 @@ startWeb relaxed optfile pathdepth s = go $ fromMaybe bad $ parseURI urlstring
   where
 	(urlstring, downloader) = getDownloader s
 	bad = fromMaybe (error $ "bad url " ++ urlstring) $
-		parseURI $ escapeURIString isUnescapedInURI urlstring
+		Url.parseURIRelaxed $ urlstring
 	go url = case downloader of
 		QuviDownloader -> usequvi
 		_ -> 

Utility/Url.hs

@@ -263,7 +263,17 @@ download' quiet url file uo = do
 
 {- Allows for spaces and other stuff in urls, properly escaping them. -}
 parseURIRelaxed :: URLString -> Maybe URI
-parseURIRelaxed = parseURI . escapeURIString isAllowedInURI
+parseURIRelaxed s = maybe (go escapemore) Just $ go isAllowedInURI
+  where
+	go f = parseURI $ escapeURIString f s
+	{- Some characters like '[' are allowed in eg, the address of
+	 - an uri, but cannot appear unescaped elsewhere in the uri.
+	 - If parsing fails with those characters unescaped, fall back
+	 - to escaping them too.
+	 -}
+	escapemore '[' = False
+	escapemore ']' = False
+	escapemore c = isAllowedInURI c
 
 hAcceptEncoding :: CI.CI B.ByteString
 hAcceptEncoding = "Accept-Encoding"

debian/changelog

@@ -28,6 +28,8 @@ git-annex (5.20150529) UNRELEASED; urgency=medium
   * Fix bug that prevented enumerating locally present objects in repos
     tuned with annex.tune.objecthash1=true.
     Fixes: unused, object count in info, unannex.
+  * Improve url parsing to handle some urls containing illegal []
+    characters in their paths.
 
  -- Joey Hess <id@joeyh.name>  Sat, 30 May 2015 02:07:18 -0400