Disregard stderr output from identify command. Fixes issue #206

src/support/z_media_identify.erl

@@ -161,11 +161,11 @@ identify_file_os(unix, File, OriginalFilename) ->
 %% @doc Try to identify the file using image magick
 identify_file_imagemagick(ImageFile) ->
     CleanedImageFile = z_utils:os_filename(ImageFile ++ "[0]"),
-    Result    = os:cmd("identify -quiet " ++ CleanedImageFile),
+    Result    = os:cmd("identify -quiet " ++ CleanedImageFile ++ " 2>/dev/null"),
     % ["test/a.jpg","JPEG","3440x2285","3440x2285+0+0","8-bit","DirectClass","2.899mb"]
     % sometimes:
     % test.jpg[0]=>test.jpg JPEG 2126x1484 2126x1484+0+0 DirectClass 8-bit 836.701kb 0.130u 0:02
-    Line1 = first_non_warning(string:tokens(Result, "\r\n")),
+    Line1 = hd(string:tokens(Result, "\r\n")),
     Words = string:tokens(Line1, " "),
     WordCount = length(Words),
     Words1 = if
@@ -194,16 +194,6 @@ identify_file_imagemagick(ImageFile) ->
             {error, "unknown result from 'identify': '"++Line1++"'"}
     end.
 
-    % Even with '-quiet' we still can get pdf warnings, strip them.
-    first_non_warning([]) ->
-        [];
-    first_non_warning([Line|Lines]) ->
-        case string:strip(Line, left) of
-            [] -> first_non_warning(Lines);
-            "****" ++ _ -> first_non_warning(Lines);
-            _ -> Line
-        end.
-
 
 %% @spec mime(String) -> MimeType
 %% @doc Map the type returned by ImageMagick to a mime type